Muthukumaran Chitambaram EH6043 114223584 MPH UCC dissertation

DEPARTMENT OF EPIDEMIOLOGY AND PUBLIC HEALTH Title: Ocular cancer incidence in Ireland – histological subtypes, staging and treatment outcomes between 1992 and 2012 By Muthukumaran Chitambaram Student Number: 114223584 Head of Department: Prof. Ivan Perry Project Supervisor: Dr. Zubair Kabir Project Tutor: Ms. Shelly Chakraborty PhD Student Report on research completed and submitted in fulfilment of the examination for the Masters of Public Health at University College Cork, September 2015. [1] Declaration form University College Cork Department of Epidemiology and Public Health Title of Thesis: Ocular cancer incidence in Ireland – histological subtypes, staging and treatment outcomes between 1994 and 2012 Name of Student: Muthukumaran Chitambaram Student Number: 114223584 I declare that the content of this assignment is all my own work. Where the work of others has been used to augment my assignment it has been referenced accordingly. Signed: CT. Muthukumaran Date: 02/10/2015 Word count: Approximately 13,752 words [2] Acknowledgements For the ancestors who paved the path before me upon whose shoulders I stand. A special thanks to my father Mr. Chitambaram Periyaiya, mother Ms. Rajeswari Chitambaram, wife Ms. Prema Muthukumaran and daughters Charulata and New born who have always supported and prayed for me throughout my education and life. My deep gratitude goes first to my supervisor Dr. Zubair Kabir, who expertly guided me throughout my MPH dissertation by understanding me and bringing the best out of me. I am extremely grateful for the assistance and support throughout the course he provided. I am using this opportunity to thank all the professors and lecturers who taught me in my postgraduation. I would like to express my sincere gratitude to the Department of Epidemiology and Public Health, University College Cork for letting me fulfil my dream of being a student here. I take this opportunity to thank my friends Ms. Laura McCarthy and Mr. Mahesh Dayalan and my tutor Ms. Shelly Chakraborty for the timely advice and support they provided. I am thankful for the aspiring guidance provided by Dr. Mahalakshmi Prasad in many difficult situations. A special thanks to Dr. Sanjiv Aggarwal cardiologist for always supporting me in all my works from the beginning of my career. I would also like to thank my previous institutions I worked and studied in particular Cancer Institute (WIA), Adyar. A special thanks to my radiation oncology professors Dr. A. Vasanthan and Dr. G. Selvaluxmy for the knowledge they have provided. Thank you Muthukumaran [3] Table of content Chapter Page A. List of Tables and Figures………………………………………………………………......7 B. List of Abbreviations………………………………………………………………………10 C. Glossary......................................................................................................................….....12 D. Abstract……………………………………………………………………………………13 1. Aim, objectives and rationale………………………………………………………………14 1.1. Rationale and importance to public health……………………………………………….14 1.2. General aims……………………………………………………………………………..16 1.3. Objectives………………………………………………………………………………..17 2. Literature Review………………………………………………………………………….18 2.1. Introduction……………………………………………………………………………...18 2.1.1. Ocular neoplasm according to morphological types…………………………………..19 2.1.2. Histology………………………………………………………………………………21 2.2. Objectives of the literature review………………………………………………………21 2.3. Search strategy and study selection criteria……………………………………………..22 2.4. Methods of searching the literature………………………………………………………23 2.4.1. Electronic database searching………………………………………………………….23 2.4.2. Book searching………………………………………...………………………………23 2.4.3. Reference list searching………………………………………………………………..24 [4] 2.5. Literature data extraction and quality appraisal methods………………………………...25 2.5.2. Summary and conclusion of the literature review……………………………………...29 3. Data and Methodology……………………………………………………………………..32 3.1. NCRI background information…………………………………………………………..32 3.2. Administrative structure…………………………………………………………………34 3.3. Data collection and management………………………………………………………...35 3.4. Ethical aspects……………………………………………………………………………37 3.5. Data used for analysis……………………………………………………………………37 3.6. Data cleaning…………………………………………………………………………….38 3.7. Descriptive statistics……………………………………………………………………..39 3.8. Analysis plan for the specific objectives…………………………………………………39 3.8.1. Objective 1……………………………………………………………………………..39 3.8.2. Objective 2……………………………………………………………………………..41 3.8.3. Objective 3……………………………………………………………………………..41 4. Results……………………………………………………………………………………..42 4.1. Descriptive statistics……………………………………………………………………. 42 4.2. Results for the specific objectives……………………………………………………….47 4.2.1. Age-Standardized incidence rates……………………………………………………..47 4.2.2. Kaplan-Meier survival curve with cos proportional hazard analysis……….…………51 5. Discussion and conclusion…………………………………………………………………64 [5] 5.1. Retinoblastoma incidence…………………………………………………..……………65 5.2. Ocular Melanoma Incidence……………………………………………….….…………65 5.3. Ocular melanoma and Gender…………………………………………….….…………..66 5.4. Screening and ocular cancer…………………………………………………….…….….66 5.5. Ocular cancer survival rate by morphology………………………………………………67 5.6. Ocular cancer survival rates by age-group……………………………………………….67 5.7. Strengths and limitations of the study……………………………………………….…...68 5.8. Implication for public health policy and recommendations for the future……….………68 5.9. Conclusions……………………………………………………………………………...69 6. References…………………………………………………………………………………70 7. Appendices………………………………………………………………………………...76 [6] A. List of Tables and Figures Table Title Page Number 1 Inclusion & exclusion criteria. 2 Search terms used in advanced search of database 3 Quality assessment summary 4 Data used and description 5 Distribution of new Ocular Cancer cases in the Republic of Ireland (1994-2012) by socio-demographic characteristics across five grouped time-periods 6 Distribution of new specific morphologic types of ocular cancer cases in the Republic of Ireland (1994-2012) by sociodemographic characteristics 7 Distribution of new specific morphologic types of ocular cancer cases in the Republic of Ireland (1994-2012) by clinical characteristics 8 Age-Standardized Incidence Rates (per 100,000) of Total Ocular Cancer cases in the Republic of Ireland (1994-2012) by age-groups across five grouped time-periods 9 Age-Standardized Incidence Rates (per 100,000) of Total Ocular Cancer cases in the Republic of Ireland (1994-2012) by gender across five grouped time-periods 10 Age-Standardized Incidence Rates (per 100,000) of Total Ocular Cancer cases in the Republic of Ireland (1994-2012) [7] by specific morphologic types across five grouped timeperiods 11 Age-Standardized Incidence Rates (per 100,000) of Melanoma cases in the Republic of Ireland (1994-2012) by age-groups across five grouped time-periods 12 Age-Standardized Incidence Rates (per 100,000) of Melanoma cases in the Republic of Ireland (1994-2012) by gender across five grouped time-periods 13 Age-Standardized Incidence Rates (per 100,000) of Retinoblastoma cases in the Republic of Ireland (1994-2012) by age-groups across five grouped time-periods 14 Age-Standardized Incidence Rates (per 100,000) of Retinoblastoma cases in the Republic of Ireland (1994-2012) by gender across five grouped time-periods 15 Univariate and Multivariable Cox Proportional Hazards Regression analysis with covariates [8] Figure Title Page Number 1 Flow chart diagram for search strategy 2 The governance structure of the National Cancer Registry Ireland 3 Source: National Cancer Registry Ireland 4 Kaplan-Meier of 1994-2012 Ocular Cancer cases by gender 5 Kaplan-Meier of 1994-2012 Ocular Cancer cases by agegroup 6 Kaplan-Meier of 1994-2012 Ocular Cancer cases by morphologic types 7 Kaplan-Meier of 1994-2012 Ocular Cancer cases by smoking status 8 Kaplan-Meier of 1994-2012 Ocular Cancer cases by anatomic sites (ICDs) 9 Kaplan-Meier of 1994-2012 Ocular Cancer cases by timeperiods 10 Kaplan-Meier of 1994-2012 Ocular Cancer cases by gender and age-groups 11 Kaplan-Meier of 1994-2012 Ocular Cancer cases by residual 12 Kaplan-Meier of 1994-2012 specific morphologic ocular cancer cases by age-groups 13 Kaplan-Meier of 1994-2012 specific morphologic ocular cancer cases by gender [9] B. List of Abbreviations NCRI- National Cancer Registry Ireland ASR- Age-standardised rates HSE- Health services executives SCC- Squamous cell carcinoma CASP-Critical Appraisal Skills Programme US- United States UK- United Kingdom TROs-Tumour registration officers HR- Hazards ratios CI- Confidence intervals KM- Kaplan-Meier WHO- World Health Organisation UV- Ultra violate radiation DNA- Deoxyribonucleic acid ERMS- Embryonal rhabdomyosarcoma COMS- Collaborative ocular melanoma study HIV- Human immunodeficiency virus HPV- Human papilloma virus OSSN- Ocular surface squamous neoplasia [10] GP- General Practitioner [11] C. Glossary Confidence interval- indicates that there is a 95% probability that the effect of treatment in the whole population lies within the stated range (Peat, Barton and Elliott, 2008). Confounders- Patient characteristics that may affect the results a study is trying to measure. Hazards ratio- The risk of the event in a study group divided by the risk of the event in a reference group. Incidence- the number of new cases of disease within a given time period. Log rank test- Statistic used to compare the evens observed and expected between two studies. p-Value- Probability that a difference between study groups would have occurred if null hypothesis was true. Variable- any characteristic or attributable that can be measured. Melanoma- A type of cancer that develops from the melanocytes (MedicineNet, 2015). Squamous cell cancer- A type of cancer of epithelial cell origin. Retinoblastoma- A cancer that rapidly develops from immature cells of retina. Age-Standardized rate- A technique used to allow populations to be compared when the age profile of the populations are quite different. [12] D. Abstract Background: Ocular cancer is a significant public health problem worldwide as it affects the vision and psychology of the patients and their family. In Ireland from 1994 to 2012 a total of 861 ocular cases has been diagnosed and treated, out of which 327 died and 534 living with significant morbidity. Methods: Population-based data, diagnosed with ocular cancer (ICD 10 codes C690- C699), (N=861), was obtained from the National Cancer registry Ireland. We analysed morphology wise incidence trend of ocular neoplasm from the year 1994-2012, and calculated the Age standardized rates for the years. Survival rates were also examined using Kaplan-Meier curve and cox proportional hazard with different covariates for the years 1994-2012. Results: Increased (163%) age specific retinoblastoma incidence is noted during the study period 1994-2012. In Ireland 78% of all ocular cancer is melanoma in morphology and older age group is four to five times more prone to get melanoma. There is a relative increase of 29% in male melanoma incidence during the years 1994-2012. Survival rates decrease in older age groups in 30-60 age-group a four time (HR=3.76; CI=1.81-7.79: p=0.00) and for above 60 agegroup a nine times (Hr=9.32: CI= 4.60-18.89; p=0.00) is noted. Conclusion: Further research is needed in the field of retinoblastoma and Male ocular melanoma as there is a significant increase in incidence rates. Screening in under five age for retinoblastoma should also be studied for its effect on preventing loss of vision and eye ball. Awareness should be created on artificial tanning and UV in ocular melanoma prevention. [13] Chapter 1 1 Aims, objectives and rationale 1.1 Rationale and importance to public health Ocular neoplasm refers to a cancerous growth in any part of the eye. Eye cancers can be divided into primary and secondary (TheFreeDictionary.com, 2015). Cancer affecting the inside of the eye are intraocular and those affecting the outside of the eye are extraocular (cancer, 2015). In Ireland 19,200 cancer cases registered annually according to 2009 to 2011 data. The incidence rate of invasive cancers is 425 case per 100,000 population per year. Every year there is 0.5% increase in incidence as per 2008-2010 data. Ocular cancers are listed in other invasive cancers which accounts for 37.3 per 100,000 population (Ncri.ie, 2015). In Ireland cancer of eye accounts between 26 and 69 cases a year from 1994 to 2012(Ncri.ie, 2015). Eye cancer morphology, prevalence, treatment and outcome differs for each and every country. Melanoma dominates in most of the countries followed by lymphoma and squamous cell carcinoma. In Ireland National Cancer Registry of Ireland (NCRI) has materials regarding melanoma which is categorised in skin melanoma (Ncri.ie, 2015). Other than that there is very less known about the other types and outcome in Ireland. Ultraviolet (UV) radiation is a major risk factor for most skin cancers. Sunlight, tanning lamps and tanning beds are also sources of UV rays (Cancer.org, 2015). Exposure to UV rays is a risk factor for skin cancer. UV rays damage the DNA of skin cells. Skin cancers start when this damage affects the DNA of genes that control skin cell growth. UVA and UVB rays can damage the DNA and cause skin cancer. Melanoma of the eye has also been linked to sun exposure in some studies (Cancer.org, 2015). Skin cancer incidence is strongly associated with the regional distribution and with the UVB radiation (Chang et al., 2010). The UV index is a measure of the level of UV radiation reaching the earth’s surface. Summer in Ireland 2014 UV [14] index was measured. A total of 181 days from April to September was measured for UV rays. It is found that the south and west of the Ireland had higher UV ray level. UV level index >=3 in Cork is 172 days, in Dublin 147 days, in Galway 163 days in Limerick 162 days and in Donegal 150 days. Cork had 95% days with high UV index during the 2014 summer (Cancer.ie, 2015). Melanoma is largely a sporadic event, risk factors include iris colour, ability to tan, northern European ancestry, and rarely family history. Melanosis is also associated with melanoma. Early diagnosis I s associated with best prognosis (Eyewiki.aao.org, 2015). Childhood ocular cancer which are strictly confined to young age group. Retinoblastoma and embryonal rhabdomyosarcoma have genetic aetiology and occurs in young age group. Retinoblastoma is the most common intraocular malignancy affecting children. The incidence varies from 3.4 to 42.6 cases per million live births by country. RB1 gene is a tumour suppressing gene which is affected in this disease. Red Reflex Examination in neonates, infants and children is recommended for early diagnosis as a screening. Early diagnosis can maximize the patient’s visual prognosis and survival. Present illness, family history and ophthalmic examination are critical for diagnosis (Eyewiki.aao.org, 2015). Awareness, screening and early diagnosis helps to save the eye. Rhabdomyosarcoma is the most common childhood sarcoma of soft tissue. Common age of occurrence is 7 to 8 years of age. RMS develops in the orbital soft tissues. The disease may be spontaneous or associated with familial syndromes, p53 tumour suppressor gene, or congenital malformation. Ophthalmoscopy examination and family history helps the diagnosis (Eyewiki.aao.org, 2015). The American Academy of Paediatrics policy statement on Red Reflex Examinations in Neonates, Infants, and Children recommends that all neonates, infants, and children should have an examination of the red reflex before discharge from the neonatal nursery and at all [15] subsequent routine health supervision visits. Early diagnosis of retinoblastoma can maximize the patient’s visual prognosis as well as survival rate (Eyewiki.aao.org, 2015). This study of ocular cancer incidence, risk factors, and morphology is an initial step for planning a screening program and creating an awareness among public and professionals. This study is an initial process to give a general knowledge about the incidence, risk demography, outcome associated with different morphology, survival time, recurrence. The result of this research will provide local evidence in terms of pattern of ocular cancer, morphology, and risk group which can provide an additional insight into existing ocular cancer diagnosis, and treatment methods in health services in Ireland. This thesis is structured as follows: this section details a brief background to the thesis and outlines the aims and objectives. Following on from this, section two provides a critical review of all relevant literature relating to ocular cancer incidence and survival. Information on the data collection procedures adopted by the NCRI team are described in section three. This section also provides a detailed description of the methodology employed in the secondary analysis of the dataset. Section four then provides a detailed analysis of the results of the statistical tests applied, while section five concludes with a discussion of the findings, examining their implications and limitations, and the potential for future research. 1.2 General aim A cross sectional study to study the incidence from (1994-2012) and survival rate of ocular cancer in Ireland using National Cancer Registry Population level data. Stratified analysis will be used to access the values for variables age-group, gender and morphology. Secondary aim of the study is to do a cross-sectional comparison of ocular cancer morphology using individual-level data received from National Cancer Registry Ireland. [16] 1.3 Objectives  To analyse yearly incidence, Age-Standardised incidence of ocular cancer in republic of Ireland between 1994 and 2012.  To stratify estimated age standardised incidence by gender, Age group, and Morphology.  To determine the age, gender and morphological wise survival rates in ocular cancer from 1994 to 2012.  To examine the association of life-style risk factors (sex, age-group, smoking, morphology, tumour site and year of incidence) with ocular neoplasm (1994 -2012). [17] Chapter 2 2. Literature Review 2.1. Introduction Ocular neoplasm or eye cancer refers to a cancerous growth in any part of the eye. Some eye cancers are primary, while other represent metastases from primary cancers elsewhere in the body (TheFreeDictionary.com, 2015). Eye cancer is one of the rarest cancers. Melanoma of the eye is the most common type of eye cancer among adults and retinoblastoma in children. In United Kingdom eye cancer amounts to 430 every year (Nhs.uk, 2015). In United States every year 2,580 new eye cancer cases are diagnosed and contributes to 270 deaths due to the disease (Cancer.org, 2015). In Ireland 69 new cases diagnosed during the year 2012(Ncri.ie, 2015). Melanoma accounts for 70% of all ocular cancer according to the studies in United States (Cancer.org, 2015). In Ireland 50 eye cancer case where registered during the year 2009 and the total number of cancer cases that year is 19,200. According to this eye cancer accounts for only 0.26 % of all cancer cases (Ncri.ie, 2015). In Ireland 90% of primary invasive melanoma cases arises in skin and 5% in the eye (Anon, 2015). According to the NCRI cancer trend January 2011 melanoma skin shows an increasing trend in all age group. Age and incidence is positively associated in skin melanoma. In Europe Scandinavian countries have higher risk than other European countries and Ireland is the 4th highest in the list during 2008. According to 2008 study there is a 5.8% for males and 308% for females increase annually in melanoma incidence in Ireland. Age standardised incidence rate are also increasing by 4.5% in men and 2.0% in female annually in Ireland. Ocular melanoma metastatic disease was diagnosed at a 25% and 34% rate at 5 years and 10 years respectively in the collaborative ocular melanoma study (COMS) with liver affected in 89% of the time(Eyewiki.aao.org, 2015). The mortality rate diagnosis of [18] metastatic disease was 80% at first year and 92% at second year (Development of Metastatic Disease After Enrollment in the COMS Trials for Treatment of Choroidal Melanoma, 2005). Human eye consists of sclera a white part, cornea a clear bulging surface in front of the eye, Iris a heavily pigmented muscle, two chambers one anterior and one posterior to iris, Ciliary body holds the lens and controls the movement, choroid lies between the retina and sclera, Vitreous humoral which fills space between lens and retina, retina is located in the back of the eye ball and the eye lid which is anterior to all and protects the eye (Nei.nih.gov, 2015). Cancer of the eyes is broadly divided into two category as intraocular and extraocular. The most common intraocular tumour is melanoma followed by lymphoma in adult and retinoblastoma in children. Intraocular melanoma has two different kinds of cells which is spindle cell and epithelioid cells. Lymphoma is of two kinds one is Hodgkin and other is nonHodgkin (Cancer.org, 2015). Extra ocular or tumour around the eye ball may affect conjunctiva, cornea, eyelid, adnexa, muscle, nerves and skin around the eyeball. Melanoma skin, basal cell carcinoma, rhabdomyosarcoma and squamous cell cancer are the few which affects the extra ocular parts (Cancer.org, 2015), (Cancer.Net, 2012) 2.1.1. Ocular neoplasm according to morphological type Adenocarcinoma is a cancerous tumour of epithelial cell origin that has glandular origin and characteristics. Risk factor for adenocarcinoma differs for each and every site. Strong sunlight, smoking, male sex, older age group, and genetic are some of the risk factors for adenocarcinoma (Cancer.org, 2015). Adenocarcinoma of small intestine has a 55% five years survival for stage I and 5% five year survival for stage IV disease (Cancer.org, 2015). Embryonal rhabdomyosarcoma (ERMS) is derived from mesenchymal precursors of mesodermal origin. ERMS accounts for 70 – 80% of all RMS tumours, and usually occurs in [19] young children (Atlasgeneticsoncology.org, 2015). Cells of this tumour have lost a small piece of chromosome 11 that came from mother, and it has been replaced by a second copy of that part of the chromosome from the father. This seems to make the IGF2 gene on chromosome 11 overactive (Cancer.org, 2015). Orbital rhabdomyosarcoma is the most common childhood soft tissue sarcoma accounting for approximately 5% of all childhood cancers in United States. Survival has increased due to the treatment modalities which is mostly combined. Embryonal cell type of rhabdomyosarcoma has a 94% five year survival rate (Eyewiki.aao.org, 2015). Retinoblastoma is a rare form of cancer that rapidly develops from the immature cells of retina. It is the most common malignant tumour of the eye in children. Retina is the specialized lightsensitive tissue at the back of the eye that detects light and colour. The most common first sign of retinoblastoma is a visible whiteness in the pupil called “cat’s eye reflex” or leukocoria. Mutation in the RB1 gene are responsible for most cases of retinoblastoma. A small percentage of retinoblastoma are caused by deletion in the region of chromosome 13 that contains the RB1 gene. In addition RB1 affected children usually also have intellectual disability, slow growth, and distinctive facial features such as prominent eyebrows, short nose with a broad nasal bridge, and ear abnormalities. People with germinal retinoblastoma may have a family history of the disease, and they are at risk of passing on the gene to next generation (Genetics Home Reference, 2015). Treatment options depend on the size of the tumour. Small tumours may be treated by laser surgery or cryotherapy. Large tumours and local tumours are treated with radiotherapy. Chemotherapy may be needed if the tumour has spread beyond the eye. The eye may need removal (enucleation) if the tumour does not respond to other treatments (Updated by: Adam S. Levy, 2015). Melanoma affects all parts of the eye uvea, ocular, choroid, ciliary body, iris, eyelid, and conjunctiva. Melanoma is more common in skin and has the same risk factor. UV rays exposure, ability to tan, and colour of the iris are some risk factors for melanoma (Holly, Aston [20] and Char, 1990). Treatment for melanoma differs according to the size, greater the size more chance of enucleation. Most of the patients are treated with combined radiotherapy and chemotherapy. Metastasis occurs at 5, 10, and 20 years in 6%, 12% and 20%. Outcome of the disease is not treatment specific and is mostly associated with virulence, site and age of the patient. Radiotherapy is a preferred method of treatment so as to save the globe. 2.1.2 Histology Most common ocular cancer in adult is Melanoma. Incidence of melanoma varies from 55%75% of all eye cancers. Retinoblastoma is the most common childhood tumour in eye. Incidence is estimated to be 1 in 18,000 to 30,000 live birth (Atlasgeneticsoncology.org, 2015). Squamous cell carcinoma of the eye amounts for 1-2.8 per 100,000 population in United States (Wikipedia, 2015). Adenocarcinoma, embryonal rhabdomyosarcoma, mixed epithelioid and spindle cell carcinoma, sarcoma are other some cancer which affects eye (Wikipedia, 2015). 2.2. Objective of the literature review The purpose of the study is to examine time trend and morphology wise incidence, and survival of ocular cancer. The main objective of this literature reviews was to systematically collect and critically summarise information available from relevant papers related to ocular cancer trends globally. The literature review focused also on risk factors which can be avoided and of public health importance. The literature review primarily assessed population based evidence which analysed trend and morphology related to ocular cancer. [21] We gathered evidence that evaluated how morphology, age, and sex effects ocular cancer trend, outcome and treatment. 2.3. Search strategy and study selection criteria A systematic reviewing method was employed to garner the relevant literature for the research objective. A detailed search strategy was developed and implemented to find all relevant articles/texts. Appropriate inclusion and exclusion criteria were set to target and focus the search (Table 1). Search terms were found by running a scoping search and identify the relevant indexing terms for relevant articles. The terms were modified according to electronic database being used. The terms used is given in the table below (Table2). The search method we used is searching and reviewing the Literature byMary Ebeling and Julie Gibbs to search electronic database, Book, reference searching and hand searching (Corwin.com, 2015). Table 1: Inclusion & exclusion criteria. 1 2 3 4 5 Inclusion Criteria Full text, English language and human studies. Studies after 1990 Published literature/ grey literature Primary aim of the study/ research related to the topic- studies that primarily focus on ocular cancer trends, incidence, therapeutics, interventions and outcome. Hospital/ population based studies Studies that focus on morphology associated with outcome, treatment and recurrence. [22] Exclusion Criteria Foreign language, non-human. Studies before 1990 Primary aim of the study/ research not related to the topic. Case reports studies Studies other than ocular cancer and lymphoma. 2.4. Methods of searching the literature We followed the search methods step by step. In addition to conventional searches of electronic database (PubMed, Embase/Medline, Cocnarane library) searches of other electronic sources were also conducted to locate relevant Irish based evidence for the literature. Google search and oncology books were also searched. Unrelated articles and journals came in search of these databases were firstly eliminated by checking titles and abstracts using the inclusion criteria previously defined. Full texts of the remaining studies were then obtained for further appraisal. 2.4.1 Electronic database searching The three electronic databases listed were individually searched for relevant articles. Also outlines the combinations of search terms used. Specific keywords, according to Boolean logic principle, Mesh term, and search limitations methods were applied. Table 2: Search terms used in advanced search of database. Database Strategy #1 #2 Medline via PubMed #1 AND #2 Eye Retinoblastoma* OR Adenocarcinoma* OR Epithelioid cell melanoma* OR Melanoma* OR Spindle cell melanoma* OR embryonal Rhabdomyosarcoma* Melanoma NOS* OR mixed epithelioid and spindle cell* OR Sarcoma* OR Squamous cell carcinoma 2.4.2. Book searching Directly searching MD Anderson Manual of Medical Oncology 2011 (Kantarjian, Wolff and Koller, 2011) and Clinical Radiation Oncology 3rd edition by Leonard L (Gunderson, Tepper and Bogart, 2012). Gunderson Books for relevant material on Ocular neoplasm done. [23] 2.4.3 Reference list searching The reference lists of articles we searched for ocular neoplasm were cross-checked to identify additional relevant articles for the review. The articles were also used to verify the validity of the selected article. Figure 1: Flow chart diagram for search strategy Identified studies through electronic database search with PubMed, Cochrane Library, Medline and Google scholar Information collected through other sources (n= 729) Studies removed for duplicate (n=182) Studies remained with full article (n=547) Inclusion and exclusion criteria applied (n=385) Articles with full text (n=162) Irrelevant articles and removed (n=146) Studies included in literature review (n=16) the studies include two systematic reviews and one metaanalysis NCRI, WHO, Melanoma skin foundation, HSE, CSO. Two books. [24] 2.5. Literature data extraction and quality appraisal methods The quality of the each paper included in this review was evaluated, using Critical Appraisal Skills Programme (CASP) checklists established by Oxford University Public Health Resource Unit (Critical Appraisal Skills Programme (CASP), 2015). The CASP toolkit allowed for individual evaluation of the literature review. Weakness and strengths in every study also determined. Table 3. 2.5.1. Table 3: Quality assessment summary Title Super selective ophthalmic artery infusion of melphalan for intraocular retinoblastoma : preliminary results from 140 treatments (Venturi et al., 2012). Author Design Acta Clinical trial Ophthalmologi ca Scandinavica Foundation 2012 Risk of CHODICK, G cataract 2009 extraction among adult retinoblastoma survivors(Cho dick, 2009) Trilateral retinoblastoma : a systematic review and meta-analysis (de Jong et al., 2014). DE JONG, M. C., KORS, W. A., DE GRAAF, P., CASTELIJNS, J. A., KIVELÄ, T. AND MOLL, A. C. 2014 Sampling June 2008 to October 2010, 38 patients (18 women, 20 men; age range at first treatment, 7 months to 22 years) with 41 eyes with retinoblastom a Retrospectiv Seven e cohort hundred fiftystudy three subjects (828 eyes) were available for analysis for an average of 32 years of follow-up a systematic 90 articles review and with 174 metapatients analysis included in the metaanalysis. [25] Strength Clinical trial, good follow up, precise estimation of outcome, outcome clearly specified, benefits worth the harms and cost. Cataract extraction as result. Measuring the outcome. 32 years of follow-up Limitation Nonrandom ised selection. Exclusion of infiltrating RB, anterior chamber invasion, glaucoma, haemorrhag e. PRISMA guidelines in selectin the studies, Good statistical methods Lack of histopathol ogical proof of the disease in some might be a bias. Lack of UV ray exposure details, work related effects details. Combined proton beam radiotherapy and transpupillary thermotherapy for large uveal melanomas: a randomized study of 151 patients (Desjardins et al., 2006). Effects of radiotherapy on uveal melanomas and adjacent tissues (Groenewald, Konstantinidis and Damato, 2012). Desjardins, Randomized L.,Lumbroso- control trial Le Rouic, 2006 151 randomly assigned uveal melanoma between February 1999 to April 2003 Clinical trial, 38 months of follow-up. Randomis ation. All clinical outcome considered . Very less follow-up time, there is no compare group or study are result mentioned. GROENEWA Review LD, C., KONSTANTI NIDIS, L. AND DAMATO, B 2012 57 Articles have been reviewed for outcome and complication s. 57 Articles, all complicati ons, effects, dose limits and treatment patterns are explained 8.7 years of mean follow-up. Specified and pathologic ally documente d lesions. No statistical significanc e mentioned, no detailed report on each and every study given Tumour thickness and orbital involveme nt proved by pathologic al report. Clear outcome measurem ent. Confoundin g factors like age and sex was not mentioned. Primary iris CONWAY, R. Retrospectiv melanoma: M e nondiagnostic 2001 comparative features and case series outcome of conservative surgical treatment (Conway, 2001). 51 cases diagnosed as iris melanoma and was treated with local resection from 1980 – 2000 Orbital exenteration in 95 cases of primary conjunctival malignant melanoma (Paridaens et al., 1994). 95 orbital exentration case of malignant melanoma PARIDAENS, A. D., MCCARTNE Y, A. C., MINASSIAN, D. C. AND HUNGERFO RD, J. L. 1994 Retrospectiv e clinicopathol ogical study [26] Deals only with iris melanoma. Other than surgical methods were not mentioned. Rhabdomyosa rcoma: the experience of the pediatric unit of Kasr El-Aini Center of Radiation Oncology and Nuclear Medicine (NEMROCK) (from January 1992 to January 200126(EL-AAL, 2006) . Plaque radiotherapy in the management of scleralinvasive conjunctival squamous cell carcinoma: an analysis of 15 eyes (Arepalli et al., 2014). Topical mitomycin-C for treatment of partiallyexcised ocular surface squamous neoplasia (Rahimi et al., 2015). Screening for Metastasis From Choroidal Melanoma: The Collaborative Ocular Melanoma Study Group Report Abd El-Aal Retrospectiv HH, Habib EE, e study Mishrif MM. 55 paediatric rhabdomyosa rcoma diagnosed during the years 1992 to 2001 Risk stage wise and advanced multidisciplinar y treatment Small sample size, lack of generalisab ility of the treatment because of the cost. AREPALLI, Interventiona Interventiona S., KALIKI, l case series l case series S., SHIELDS, involving 15 C. L., patients with EMRICH, J., histopatholog KOMARNIC ically KY, L. AND confirmed SHIELDS, J. scleral and/or A. intraocular 2014 invasion of SCC Multiple outcome measurem ents. Hospital based and valid data. Retrospecti ve study. Lack of good statistical methods. RAHIMI, F., Clinical trial ALIPOUR, F., GHAZIZADE H HASHEMI, H. AND HASHEMIAN , M. 2009 Interventio nal prospectiv e study. Proven pathology. A good statistical measure. Very less number of cases. No comparativ e studies have been mentioned. Prospectiv e longitudin al study. Valid laboratory methods. No alternative methods discussed fo sensitivity and specifcity. Marie Diener- Prospective West, Sandra longitudinal M. Reynolds, Donna J. Agugliaro, Robert Caldwell. 2004 [27] 17 pathologicall y proved ocular surface squamous cell carcinoma between 2004 – 2006 All the 2,320 patients enrolled in COMS trial had LFT done (Diener-West, 2004). Uveal melanoma in relation to ultraviolet light exposure and host factors (Holly, Aston and Char, 1990). Sun exposure predicts risk of ocular melanoma in Australia (Vajdic et al., 2002). Intermittent and chronic ultraviolet light exposure and uveal melanoma: a meta-analysis (Shah et al., 2005). Epidemiology of ocular surface squamous neoplasia in Africa (Gichuhi et al., 2013). Para meningeal rhabdomyosar coma in paediatric age: results of a pooled analysis from North HOLLY, E. Case control A., ASTON, D. A. AND CHAR, D. H. 407 cases of melanoma diagnoses between 1978 and 1987, and 870 controls Multiple risk factors, good statistical analysis. Nonrandom and highly selective cases. A retrospectiv e study. VAJDIC, C. M., KRICKER, A., GIBLIN, M., MCKENZIE, J., AITKEN, J., GILES, G. G. AND ARMSTRON G, B. K. SHAH, C. P., WEIS, E., LAJOUS, M., SHIELDS, J. A. AND SHIELDS, C. L 2005 Case control 290 clinically study diagnosed melanoma cases and 916 controls Blinding, Bias due to prospectiv less study e analysis, population. confoundi ng are included in analysis. MetaAnalysis 12 studies which provided clear odds ratio a d standard error for the exposure and outcome is included. Multiple exposure, free of bias, clear analysis of data. GICHUHI, S., SAGOO, M. S., WEISS, H. A. AND BURTON, M. J. 2013 Systematic and nonsystematic review A fixed effect model with HIV and smoking used for metaanalysis. J. H. M. Pooled Merks, G. L. analysis De Salvo, C. Bergeron, G. Bisogno, A. De Paoli, A. Ferrari, A. Rey 10 studies on RMS with factors PM involvement, age, size, invasiveness, nodes treated with RT 2014 [28] Only 12 studies from 133 studies considered. No effort was made to get the relevant study analysis made. Multiple Lack of outcome, exact study HIV, UV method has exposure, introduced HPV, a selection smoking bias. all discussed K-M Bias in methods, selection of multiple studies. outcome and exposure. American and European cooperative groups (Parameninge al rhabdomyosar coma in pediatric age: results of a pooled analysis from North American and European cooperative groups, 2014). 2.5.2. Summary and conclusion of the literature review Ocular neoplasm is being studied from the year 1944 till now. Around 15,948 studies has been done so far (Ncbi.nlm.nih.gov, 2015). Most of the studies are done with single case or with less number of cases due to less prevalence. Meta-analysis is about melanoma and its risk factor (UV) and the systematic review is about squamous cell carcinoma. None of the papers covers the ocular neoplasm as whole. Meta-analysis gives an inconclusive evidence on melanoma with UV rays. A meta-analysis was conducted (2005) by Shah, C.P(Shah et al., 2005) discusses the exposures like welding, outdoor leisure, birth latitude, and UV exposure with melanoma and found positive welding (OR 2.05), and no significance (OR 1.37) with UV ray exposure and outdoor leisure activity. There are many studies which correlates UV with skin melanoma and ocular neoplasm with significant effect. Elizabeth A. Holly’s (Holly, Aston and Char, 1990) study on uveal melanoma in relationship to ultraviolet light exposure and host factors found green, grey, blue and hazel eye colour, north European skins, ability to tan, sun and welding burns, large [29] nevi are all risk factors for uveal melanoma. Claire M. Vajdic’s study(Vajdic et al., 2002) sun exposure predicts risk of ocular melanoma in Australia found that choroidal and ciliary melanoma are strong positive for prolonged sun exposure. The study is inconclusive for iris and conjunctival melanoma. A D Paridaens study (Paridaens et al., 1994) orbital exenteration in 95 cases of primary conjunctival malignant melanoma advocates exenteration should be reserved as a palliative procedure for advanced stage of neoplastic disease. The study also advocates surgery combined with chemotherapy and radiotherapy to preserve the globe. Marie Diener-west (Diener-West, 2004) studied screening for metastasis from choroidal melanoma for the liver metastasis. 2320 patients enrolled for COMS study in that 714 patients had metastasis in that 675 died, he found that LFT is not sensitive enough to find the liver metastasis. This study gives an over view about the metastasis and survival. Systematic review(Gichuhi et al., 2013) (2013) by Gichuhi,s. focused on squamous cell carcinoma and compared the age specific rates of various countries and discussed about the risk factors such as HIV, smoking, and Human Papilloma Virus (HPV). HIV (OR= 6.17) and direct day light exposure, outdoor occupation (OR=1.7) found to have strong association and smoking 2 studies (OR= 1.4) had a comparatively weaker association than the other three. As HIV is an immune compromising disease we could understand the squamous cell carcinoma is associated with age and decreasing immunity. Arepalli S (Arepalli et al., 2014) studied the plaque radiotherapy in the management of scleral-invasive conjunctival squamous cell carcinoma and found out its effective in curing the disease and as an alternative for enucleation. Brachytherapy and electron beam therapy are the radiotherapy methods used to treat SCC successfully. Rahimi F (Rahimi et al., 2015) studied the effect of topical mitomycin-c for treatment of partially excised ocular surface squamous neoplasm. The mitomycin-C had 94% [30] nonrecurrence for 41 months without any major side effects. For the SCC and OSSN first and preferred choice of non-invasive is mitomycin-C. Marcus C de jong (2014) systematic review and meta-analysis (de Jong et al., 2014) explains the treatment advancement and increased survival after 1995 in the trilateral retinoblastoma. In trilateral RB 5 year survival after treatment has increased from 6% to 57% because of conventional and high dose chemotherapy with stem cell rescue. In this he has clearly stated that none of the trilateral RB survived before 1995 due to lack of treatment. Carlo Venturi’s study (Venturi et al., 2012) on super selective ophthalmic artery infusion of melphalan for intraocular retinoblastoma also shows the advancement in treatment and administration and survival. Chodick. C studied the risk of cataract extraction among adult retinoblastoma survivors explains the same, out of 1729 RB cases diagnosed from 1914 to 1984 only 1,169 were alive. In this population due to radiotherapy 51 cataract extraction happened. Pooled analysis(Parameningeal rhabdomyosarcoma in pediatric age: results of a pooled analysis from North American and European cooperative groups, 2014) by J.H.M. Merks (2014) on Para meningeal RMS in paediatric age with the North American and European cooperative groups concentrated on the para meningeal involvement which is considered as poor prognostic factor. 10 studies with radiotherapy for RMS was analysed for the event free survival, and overall survival. Patients treated with radiotherapy had a good EFS (62.6) and OS (66.2) compared with non RT treated respectively 40.8%. This study reinfused the use of radiotherapy in RMS and also have shown a good prognosis in para meningeal involvement patients if treated with radiotherapy. Kasr El-Aini centre of radiation oncology paediatric unit treated 55 case of embryonal RMS with various regimes of chemotherapy and conventional radiotherapy. Despite the advances in the therapy 30% of the cases experienced progression or relapse. [31] Chapter 3 3. Data and Methodology This secondary analysis for the thesis used the data from National Cancer Registry Ireland. The data available was both population and individual based. The available data from the NCRI is for the period 1994 to 2012. As the ocular cancer affects all the age group we got the data from 0 to death or till the end of the study (2012). Fixed pre-specified intervals (cohort –based survival analysis), 1994-1997, 1998-2001, 2002-2005, 2006-2009 and 2009-2012, also the data of the year of death ends at 2012 where used(Ncri.ie, 2015). WHO standard diagnostic tool ICD used for ocular cancer (ICD10 code C69) also includes the sites (C690 Conjunctiva, C691 Cornea, C692 Retina, C693 Choroid, C694 Ciliary body, C695 Lacrimal gland and Duct, C696 Orbit and C699 Eye, unspecified. 3.1 NCRI background information In 1991 Minister for health established the National Cancer Registry which was funded by the Department of health. The board was set up to collect and record all cancer data and has been on its work since 1994. The functions of the board were laid down in 1991 legislation which came to amendment in 1996. Key duties of the board are as follows.  To identify, collect, record, classify store and analyse information regarding cancer relating incidence and prevalence of Ireland.  To collect, classify, record and store information in relation to every newly diagnosed cancer patients information in relation to the cancer.  To promote and facilitate collected data for approved research and in health planning and management.  To annually publish registries activities [32]  To advice, inform, assist the Minister in relation to cancer related aspects. Cross border co-operation is done with Northern Ireland where the Cancer Registry is a member of Ireland-Northern Ireland –National Cancer Institute Cancer Consortium. The Cancer consortium was established in 1999 at the Stormont Parliament Building in Belfast. As a joint effort the two registries announced the findings of the entire island of Ireland findings on May 2001 which was the first All-Ireland Cancer Statistics report. The consortium also published collaborative reports on September 2004 April 2009(Ncri.ie, 2015). The National Cancer Registry Ireland has 52 staff members presently. Throughout the country NCRI has 19 members and others are based at registry’s headquarters in Cork. The NCRI collect a wide range of information for every new cancer case. Name, address, gender, date of birth, socioeconomic status, cancer location, type, stage and the treatment received for the same are collected. All the information’s are collected from the records maintained in the hospital medical records. The death certificate is used to collect data in the event of mortality. Trained Tumour Registration Officers (TROs) who are based in the hospital collect most of the data for NCRI. This is an active data collection, and all new cases, registration details, patient information’s and treatment details are collected. Hospital pathology report provides almost 85% of the information on cancer. Some of the hospitals provide electronic data to the registry. [33] 3.2. Administrative structure Figure 2: The governance structure of the National Cancer Registry Ireland is given below. Department of Health National Cancer Registry Board Director Dr Harry Comber (Acting) Corporate services Research & analysis Data management & tumour registration officers [34] Information technology 3.3. Data collection and Management National Cancer Registry Ireland collects population based data from the information management unit at the Department of Health. The statistical information is collected from Public Health Information System (PHIS). The data gives information of the population, age group, and overall population of every county and Ireland as a whole. The registry uses the age-specific population and computes the age-specific rates. The crude rats for the cancer is also compute with the use of sum of age-specific population rather than using the overall population. The PHIS and the health board uses the same sum of county populations. The NCRI data are a few years behind because of many reason. Mostly all the case are identified within 12 months of the diagnostic date. As the cancers are diagnosed through various sources like pathology report, X-rays, and scans, it takes long time to collect all report and put together to get the accurate information. With the limited resources strict quality control measures are used in collecting the information. The national cancer registry keeps Patients name and address to track the patient if he attends other hospital, to know the outcome, and to identify the risk areas. The cancer registry is an independent agency and do not share any information in any circumstances (Ali, 2014). Cancer Factsheets is a one page summary of statistics for each and every cancer which includes the number of new cases, mortality rate for the cancer in Ireland. Presently 2009 to 2012 annual average incidence and mortality is available. The factsheets are frequently updated and will be available online (Ncri.ie, 2015). The figure below shows how the NCRI collects the information, and the availability of data sources can be seen. [35] Source: National Cancer Registry Ireland Figure 3: availability of data sources of NCRI [36] 3.4. Ethical aspects Ethical approval was obtained from The Clinical Research Committee of the Cork Teaching Hospitals, UCC on 14th of April 2015. 3.5. Data used for the analysis The variables used for the study from the dataset for this analysis are as follows. Table 4: Data used and description Variable Name Year Sex Survival time Patients status Age Marital status Smoking status Presentation Diagnostic method Morphology type Site of cancer Affected side Residue Treatment Recurrence Prefix YOI SEX SURV-time Description Year of incidence. Gender (sex of patient) M= male F= female Survival in days from date of diagnosis to death or 31/12/2012 (whichever is earliest) VITAL_STAT 2=dead, 1= alive (at the end of follow-up) AGE_GP Five year age group at diagnosis MARITAL Marital status of the patient: (M) Married, (D) Divorced, (E) separated, (S) single (W) widowed (Z) other and unknown. SMOKER Smoking status of the patient © current (N) Never smoked, (X) Ex-Smoker, (Z) Unknown PRES Method of presentation (A) Autopsy, (C1) Screening unspecified, (C2) Screening organised, (I) incidental, (S) symptomatic, (Z) unknown. DIAG Method of diagnosis (H) Histology, (L) clinical ( R) Radiology (Z) unknown MORP_TYPE Morphological subtype. ICD10 ICD10 cancer code. SIDE Eye affected (L=Left, R=Right, B= Bothe, Z=Unknown) RESIDUAL Extent of the residual tumour after treatment Surgery, chemo, Type of treatment given. radio RECUR_time Time in days from diagnosis to recurrence. [37] 3.6. Data Cleaning Dataset preparing for analysis  Year of incidence (YOI) all the 19 years divided into 5 groups 1994 to 1997 (1997), 1998 to 2001 (2001), 2002 to 2005 (2005), 2006 to 2009 (2009), and 2010 to 2012 the last 3 years (2012).  Vital statistics (VITAL_STAT) was given as string variable so it was renamed as vital and used. No change is made in the values (Ats.ucla.edu, 2015).  Sex was renamed as gender and Male (M) converted to 1, Female (F) converted to 2 and used.  Age group (AGE_GP) which was again a string variable renamed as age group. The available content was in five years group which was combined and made three group as 1 (0 to 29), 2 (30 to 59) and 3 (60 and above).  Marital status (MARITAL) string removed and changed to numerical value. Married as 1, Divorced as 2, Separated as 3, Single never married as 4, Widowed as 5 and unknown as 6.  Smoker was again converted into numerical values Current smoker as 1, Never smoked as 2, Ex-smoker as 3, and unknown as 4.  Presenting method PRES was divided into three category from six. Screening unspecified and opportunistic as C, symptoms kept as it is and unknown and incidental as Z. No case was in Autopsy, and organised screening so left as such.  Diagnostic methods were in 13 categories, which was made four all the laboratory methods cytology, histology put in H, all clinical methods as L, all radiological methods into R, and unknown as Z.  Morphological types (MORPH_TYPE) was given in their original names as adenocarcinoma, squamous, melanoma and retinoblastoma. All that was changed to [38] numerical value and combined. All melanomas of various types as category 1, Retinoblastoma as 2 and all other negligible and unknown combined and named as 3.  Residual after treatment (RESIDUAL) no residue kept as same, all microscopic and macroscopic residual named as 1 and unknown as X. 3.7. Descriptive statistics In the descriptive statistics basic demographic information will be calculated using the data. Age, sex, marital status, smoking habit, diagnostic method, presenting methods, treatment, morphology, ICD10, sides, residual, treatment methods surgery, radiotherapy, and chemotherapy all the frequencies will be calculated. A separate descriptive analysis will be done with morphological type. 3.8. Analysis plan for the specific objectives Data analysis was conducted using STATA version IC 13.0 (64-bits) software and the Excel 2013. 3.8.1 Objective 1 The main objective is to determine cancer incidence by age, as the risk of cancer increases exponentially with the increasing age. Different population age group have different burden of disease which cannot be represented be crude incidence rate. The age standardised is the true representation of the burden of each age group. Age- standardised rate is a summary of that age specific group rate using a standard population (Cso.ie, 2015). The age wise population for each and every age group and every year was collected from the CSO website. Using the Stata every year incidence, age group wise incidence and gender age group wise incidence was computed. The calculation is a weighted average of age-specific rates (Iarc.fr, 2015). [39] Age-standardised rate (SR) = (SUM (ri * Pi))/SUM Pi ri is the age-group specific rate for age group I in the population being studied. Pi is the population of age group I in the standard population. The result is always expressed as rate per 100,000 person-years. Age group less than 30, 30 to 60 and above 60 are the groups to be computed. Age-standardized rates for all the three age groups, in that male and female separate for each and every age group and all this rates to be calculated for all the five year groups and tabulated. Every age group population for every year was taken from CSO and added to get the desired population for our age groups. The same age standardised rates were also be computed for each and every morphology type, for every age group and for both the sex. This gives the age standardised incidence rate for each and every ocular cancer morphology for every year for every 100,000 population. The age-standardised rates for every cancer for different age group will give an idea of the incidence of the age group. Excel will be used and the age standardised rates will be tabulated and graph will be made to see the graphical representation of every morphology, age group, and gender and year wise incidence. The graphical representation will give a clear idea about the incidence of each and every year with all the parameters. Age standardised rate will be calculated for <30, 30 to 60 and 60+ age group in total for the years 1997, 2001, 2005, 2009, and 2012, again this will be divided into Male and female for the same age group, the male and female will be further divided into morphology melanoma, retinoblastoma and other ocular cancers. The same values will be transferred to excel and the graph will be made as again as mentioned above and plotted. [40] 3.8.2 Objective 2 The Kaplan-Meier estimate is also known as product limit estimator, is used to estimate the survival function from life time data. The KM curve can take the censored data into account. The survival analysis can evaluate the data by follow-up time rather than calendar time. The KM curve can also provide the patients at risk so that we understand the time and the number of patients. In this analysis KM curve will be plotted for gender with the analysis time. Second KM curve will be plotted against different age group. A separate KM curve for the morphology type will be plotted to see the difference in survival. And KM curve for smoking, ICD10 code, residual after treatment, age and morphology, and gender and morphology will also be made to see the difference. For each and every KM curve a Log rank test will be computed to see the difference between the groups observed and estimated events. All the potential confounder for ocular cancer will be considered and a KM curve will be plotted for each of them. 3.8.3 Objective 3 Cox Proportional Hazard models will be done to estimate the risk over time, to get the proportional hazards. Age groups, smoking groups, marital status, morphology, ICD10 all will be regressed separately to find the significant groups. A multivariate cox regression will also be done with all the variables and will be fitted and interpreted. A table with every variable and groups, with P-value and confidence interval and Hazard ratio will be produced. [41] Chapter 4 4 Results 4.1. Descriptive statistics The population based data received from NCRI for the study had 861 individuals who had ocular neoplasm. The period of study is 1994 to 2012 a total of 19 years. The population age is 0 to death during the diagnosis to death. The basic demographic characteristics of the study population were divided into five time periods (1994-1997, 1998-2001, 2002-2005, 2006-2009, and 2010-2012) are presented below (table 5). The study population has been presented with the factors sex, age group, marital status, smoking status, and morphology. Table 5: Distribution of new Ocular Cancer cases in the Republic of Ireland (1994-2012) by socio-demographic characteristics across five grouped time-periods Years 19941997 No (%) 19982001 No (%) 20022005 No (%) 20062009 No (%) 20102012 No (%) Total 78 (47.56%) 76 (47.50%) 110 (55.56%) 92 (50.00%) 84 (54.19%) 440 (51.10%) 86 (52.44%) 84 (52.50%) 88 (44.44%) 92 (50.00%) 71 (45.81%) 421 (48.90%) 16 (9.76%) 20 (12.50%) 18 (09.09%) 28 (15.22%) 23 (14.84%) 105 (12.20%) 30-60 59 (35.98%) 44 (27.50%) 73 (36.87%) 54 (29.35%) 49 (31.61%) 279 (32.40%) >60 89 (54.27%) 96 (60.00%) 107 (54.04%) 102 (55.43%) 83 (53.55%) 477 (55.40%) 126 (76.83%) 119 (74.38%) 162 (81.82%) 142 (77.17%) 124 (80.00%) 673 (78.16%) Patients Gender Male Female Age < 30 Morphology Melanoma [42] No (%) Retinoblastoma 8 (4.88%) 13 (8.13%) 10 (5.05%) 16 (8.70%) 17 (10.97%) 64 (7.43%) Other cancers 30 (18.29%) 28 (17.50%) 26 (13.13%) 26 (14.13%) 14 (90.3%) 124 (14.40%) 88 (53.66%) 81 (50.63%) 116 (58.59%) 84 (45.65%) 88 (56.77%) 457 (53.08%) Divorced 0 (0.00%) 1 (0.63%) 1 (0.51%) 2 (1.09%) 3 (1.94%) 7 (0.81%) Separated 2 (1.22%) 1 (0.63%) 4 (2.02%) 5 (2.72%) 1 (0.65%) 13 (1.51%) Single (never 29 Married) (17.68%) 31 (19.38%) 39 (19.70%) 52 (28.26%) 39 (25.16%) Widowed 36 (21.95%) 29 (18.13%) 27 (13.64%) 22 (11.96%) 12 (7.74%) 9 (5.49%) 17 (10.63%) 11 (5.56%) 19 (10.33%) 12 (7.74%) 24 (15.00%) . 34 (17.17%) 24 (13.04%) 13 (8.39%) 125 (14.52%) 71 (45.81%) 365 (42.39%) 19 (12.26%) 66 (7.67%) 52 (33.55%) 305 (35.42%) Marital Status Married Unknown/other Smoking Status Current 30 (18.29%) 190 (22.07%) 126 (14.63%) 68 (7.90%) Never smoked 78 (47.56%) 62 (38.75%) Ex-Smoker 17 (10.37%) 10 (6.25%) Unknown 39 (23.78%) 64 (40.00%) 79 (39.90) 75 (40.76%) 6 (3.03%) 14 79 (7.61%) (39.90%) 71 (38.59%) 56 (34.15%) 74 (46.25%) 120 (60.61%) 141 (76.63) 143 (92.26%) 534 (62.02%) 108 (65.85%) 164 (19.05%) 86 (53.75%) 160 (18.58%) 78 (39.39%) 198 (23.00%) 43 (23.37%) 184 (21.37%) 12 (7.74%) 155 (18.00%) 327 (37.98%) 861 (100%) Vital Statistics Alive Dead Total A total of 861 ocular cancer patients were diagnosed to have ocular neoplasm. Males 440 (51.10%), and females 421 (48.90%) and shares almost equally. Below 30 age group 105 (12.20%), the 30 to 60 years age group 279 (32.40%) and the above 60 age group amount for [43] more than half of the cases 447 (55.40%). Morphology type melanoma of all types’ accounts for almost 80% of the study population 673 (78.16%), Retinoblastoma 64 (7.43%) and other cancer 124 (14.40%). Smoking status unknown 305 (35.42%) and never smoked 365 (42.39%) group share more the current smokers 125 (14.52%) and EX-smoker 66 (7.67%). Patient’s life status shows a significant trend in the year groups. The death during the 1994-1997 is 108 (65%) at the end of the study during the years 2010 to 2012 it is only 12 (7.74%). Total patients alive during the study period is 534 (62.02%) and dead (327 (37.98%). Distribution of newly diagnosed cases during the study period has been divided according to the morphology and given below (table 6). Melanomas (M-49.18% & F-50.82%) and retinoblastoma (M-51.56% & F-48.44%) have almost equal gender population, the other ocular cases males 76 (61.29%) and females (38.71%). In the age group above 60 population melanoma 402 (59.73%) and the other cancers 75 (60.48%) were found. The retinoblastoma all 64 cases are confined to less than 30 years age population. In particular all the retinoblastomas were found to be in below 5 years of age group. Vital status in melanoma 409 (60.77%), retinoblastoma 63 (98.44%) and other cancers 62 (50.00%) were alive. Table 6: Distribution of new specific morphologic types of ocular cancer cases in the Republic of Ireland (1994-2012) by socio-demographic characteristics Melanomas No (%) Retinoblastomas Others No (%) No (%) Gender Male 331 (49.18%) 33 (51.56%) Female 342 (50.82%) 31 (48.44%) Age <30 21 (3.12%) 30-60 250 (37.15%) 0 (0.00%) 64 (100%) [44] Total No (%) 76 (61.29%) 440 (51.10%) 48 (38.71%) 421 (48.90%) 20 (16.13%) 105 (12.20%) 29 (23.39%) >60 402 (59.73%) 0 (0.00%) 75 (60.48%) 279 (32.40%) 477 (55.40%) Vital Statistics Alive 409 (60.77%) 63 (98.44%) Dead 264 (39.23%) 1 (1.56%) Marital Status Married 410 (60.92%) 0 (0.00%) Divorced 7 (1.04%) 0 (0.00%) Separated 9 (1.34%) 0 (0.00%) Single (never 94 (13.86%) 64 (100%) Married) Widowed 100 (14.86%) 0 (0.00%) Unknown/other 53 (7.88%) 0 (0.00%) 62 (50.00%) 534 (62.02%) 62 (50.00%) 327 (37.98%) 47 (37.90%) 457 (53.08%) 0 (0.00%) 7 (0.81%) 4 (3.23%) 13 (1.51%) 32 (25.81%) 190 26 (20.97%) (22.07%) 15 (12.10%) 126 (14.63%) 68 (7.90%) Smoking Status Current 108 (16.05%) 0 (0.00%) Never smoked 267 (39.67%) 57 (89.06%) Ex-Smoker 59 (8.77%) Unknown 239 (35.51%) 7 (10.94%) 59 (47.58%) 66 (7.67%) 673 (78.16%) 64 (7.43%) 305 (35.42%) 861 (100%) Total 0 (0.00%) 17 (13.71%) 125 (14.52%) 41 (33.06%) 365 7 (5.65%) (42.39%) 124 (14.40%) Melanoma was found to be high in never smoked 267 (39.67%) and with unknown smoking status 239 (35.51%). Clinical Characteristics associated with the cancer morphology has been plotted below (Table 7). Anatomical site choroidal plexus 424 (63.00%) and ciliary body 166 (24.67%) amounts for most of the melanoma cases. All the retinoblastoma 64 (100%) cases were confined to retina [45] alone. Other cancer morphology has a fair distribution in all sites more in conjunctiva 27 (21.77%) and lacrimal gland 28 (22.58%). Table 7: Distribution of new specific morphologic types of ocular cancer cases in the Republic of Ireland (1994-2012) by clinical characteristics Melanomas No (%) Retinoblastomas No (%) Others No (%) Total No (%) Anatomic sites Conjunctiva 36 (5.35%) 0 (0.00%) 27 (21.77%) 63 (7.32%) Cornea 3 (0.45%) 0 (0.00%) 4 (3.23%) 7 (0.81%) Retina 8 (1.19%) 64 (100%) 5 (4.03%) 77 (8.94%) Choroid 424 (63.00%) 0 (0.00%) 22 (17.74%) 446 (51.80%) Ciliary body 166 (24.67%) 0 (0.00%) 15 (12.10%) 181 (21.02%) gland 0 (0.00%) 0 (0.00%) 3 (2.42%) 3 (0.35%) 8 (1.19%) 0 (0.00%) 28 (22.58%) 36 (4.18%) 28 (4.16%) 0 (0.00%) 20 (16.13%) 48 (5.57%) 3 (0.45%) 1 (1.56%) 0 (0.00%) 4 (0.46%) 607 (90.19%) 61 (95.31%) 99 (79.84%) 767 (89.08%) Unknown Diagnostic Methods Histology 63 (9.36%) 2 (3.13%) 25 (20.16%) 90 (10.45%) 421 (62.74%) 52 (81.25%) 61 (50.00%) 534 (62.31%) Clinical 108 (16.10%) 8 (12.50%) 25 (20.49%) 141 (16.45%) Radiology 135 (20.12%) 4 (6.25%) 19 (15.57%) 158 (18.44%) Other Treatment Surgery 7 (1.04%) 0 (0.00%) 17 (13.93%) 24 (2.80%) 394 (58.54%) 53 (82.81%) 40 (32.26%) 487 (56.56%) Chemotherapy 65 (9.66%) 28 (43.75%) 15 (12.10%) 108 (12.54%) Lacrimal & duct Orbit Eye unspecified Presentation Screening (unorganised) Symptomatic [46] Radiotherapy 204 (30.31%) Residual after Rx No residue 175 (32.29%) Residue (micro & Macro) 97 (17.90%) Unknown 270 (49.82%) Total 673 (78.16%) 3 (4.69%) 37 (29.84%) 244 (28.34%) 38 (70.37%) 17 (18.89%) 230 (33.53%) 4 (7.41%) 31 (34.44%) 132 (19.24%) 12 (22.22%) 64 (7.43%) 42 (46.67%) 124 (14.40%) 324 (47.23%) 861 (100%) Most of the cases presented with symptoms 767 (89.08%) and none of the cases were found in organised screening. Only four cases have been identified during opportunistic screening. Most of the diagnosis were made by histological 534 (62.31%) findings, clinically 141 (16.45%) and radiologically 158 (18.44%) were diagnosed. In total surgery 487 (56.56%), chemotherapy 108 (12.54%) and radiotherapy 244 (28.34%) performed as treatment. Combined therapy chemo and surgery, radiotherapy surgery 67 (13.76%) were performed. Radiotherapy & chemotherapy combined was performed in 43 patients. Some 15 patients have received all three modality of treatment. After treatment histologically (micro and macroscopic) confirmed residual were present in 132 (19.24%) cases, and in 324 (47.23%) cases residual could not be accessed. 4.2. Results for the specific objectives 4.2.1 Age-Standardized Incidence rates Table 8: Age-Standardized Incidence Rates (per 100,000) of Total Ocular Cancer cases in the Republic of Ireland (1994-2012) by age-groups across five grouped time-periods. AGE (in years) <30 30-60 >60 1994-1997 Rates/100,000 1998-2001 Rates/100,000 2002-2005 Rates/100,000 2006-2009 Rates/100,000 2010-2012 Rates/100,000 0.22 1.11 4.03 0.28 0.71 4.07 0.24 1.06 4.12 0.36 0.71 3.39 0.40 0.87 3.67 [47] Age-Standardized incidence rates for total ocular cancer case by age group for the five grouped time period is give above (Table 8). During the all grouped period above 60 year population had a higher SR which is above 3.5 per 100,000 population. The above 60 age population is four to eight times more likely to get ocular cancer than the other. The rates in <30 age group have gone up from 0.22/100,000 in 1994-1997 to 0.40/100,000 in 2010-2012 (a relative increase of 81%). In contrast in the rates in 30 to 60 years age group have declined from 1.11/100,000 in 1994-1997 to 0.87/100,000 in 2010-2012 (a relative decrease of 21%). The rates in above 60 age group have declined from 4.03/100,000 in 1994-1997 to 3.67/100,000 (a relative decrease of 9%). Table 9: Age-Standardized Incidence Rates (per 100,000) of Total Ocular Cancer cases in the Republic of Ireland (1994-2012) by gender across five grouped time-periods GENDE R Men women 1994-1997 Rates/100,00 0 1998-2001 Rates/100,00 0 2002-2005 Rates/100,00 0 2006-2009 Rates/100,00 0 2010-2012 Rates/100,00 0 1.08 1.18 1.00 1.07 1.04 1.30 1.01 1.01 1.23 1.04 Age-standardized incidence rate by gender for the grouped years plotted above (Table 9). The rates in men have gone up from 1.08/100,000 in1994-1997 to 1.23/100,000 in 2010-2012 (a relative increase of 13%). In contrast rates in women has gone down from 1.18/100,000 in 1994-1997 to 1.04/100,000 in 2010-2012 (a relative decline of 12%). [48] Table 10: Age-Standardized Incidence Rates (per 100,000) of Total Ocular Cancer cases in the Republic of Ireland (1994-2012) by specific morphologic types across five grouped time-periods 1994-1997 Rates/100,0 00 1998-2001 Rates/100,0 00 2002-2005 Rates/100,0 00 2006-2009 Rates/100,0 00 2010-2012 Rates/100,0 00 MORPHOLOG Y Melanoma 0.87 0.77 0.95 0.68 0.91 Retinoblastoma 0.06 0.06 0.06 0.09 0.12 Others* 0.21 0.18 0.15 0.14 0.10 Others*= Adenocarcinoma, Embryonal rhabdomyosarcoma, Sarcoma, Squamous cell carcinoma and unspecified. Morphology wise age-standardization rates for the ocular cancer is plotted below (Table 10). Melanoma remains high with around 1 for 100,000 population during all the time period. Retinoblastoma was the lowest in morphology which is 0.06 to 0.12 for 100,000 population. Other cancers combined together age-standardized rate was around 0.1 for 100,000 population. The rates in Melanoma have gone up from 0.87/100,000 in 1994-1997 to 0.91/100,000 in 20102012 (a relative increase of 4%). The rates in retinoblastoma have also gone up from 0.06/100,000 in 1994-1997 to 0.12/100,000 in 2010-2012 (a relative increase of 100%) it has doubled. In contrast rates of other cancer have declined from 0.21/100,000 in 1994-1997 to 0.10/100,000 in 2010-2012 (a relative decrease of 52%). Table 11: Age-Standardized Incidence Rates (per 100,000) of Melanoma cases in the Republic of Ireland (1994-2012) by age-groups across five grouped time-periods 1994-1997 1998-2001 2002-2005 2006-2009 2010-2012 Rates/100,000 Rates/100,000 Rates/100,000 Rates/100,000 Rates/100,000 AGE (in years) <30 0.06 30-60 0.96 >60 3.22 0.07 0.60 3.30 0.04 0.94 3.62 [49] 0.09 0.66 2.82 0.05 0.83 3.27 Melanoma case specific age standardization rates for the grouped year for all age groups are plotted above (Table 11). Above 60 age group have a higher incidence rate than the other two groups and is almost three to four times higher than the 30 to 60 age group and is more than 50 times higher than the below 30 age group. Below 30 age group people incidence for the given time is 0.06 for 100,000 population. During the given periods all the rates and age group remained almost same with very little change. Table 12: Age-Standardized Incidence Rates (per 100,000) of Melanoma cases in the Republic of Ireland (1994-2012) by gender across five grouped time-periods 1994-1997 1998-2001 2002-2005 2006-2009 2010-2012 Rates/100,000 Rates/100,000 Rates/100,000 Rates/100,000 Rates/100,000 Gender Men women 0.79 0.94 0.65 0.87 1.00 0.91 0.76 0.80 1.02 0.79 Age-Standardized rates of melanoma gender wise given for the grouped time period given above (Table 12). The Melanoma rates in men have gone up from 0.79/100,000 in 1994-1997 to 1.02 in 2010-2012 (a relative increase of 29%). In contrast the melanoma rates in female have declined from 0.94/100,000 in 1994-1997 to 0.79/100,000 in 2010-2012 (a relative decline of 16%). Table 13: Age-Standardized Incidence Rates (per 100,000) of Retinoblastoma cases in the Republic of Ireland (1994-2012) by age-groups across five grouped time-periods 1994-1997 1998-2001 2002-2005 2006-2009 2010-2012 Rates/100,000 Rates/100,000 Rates/100,000 Rates/100,000 Rates/100,000 AGE (in years) <30 0.11 30-60 0.00 >60 0.00 0.18 0.00 0.00 0.13 0.00 0.00 [50] 0.21 0.00 0.00 0.29 0.00 0.00 Retinoblastoma is a below five years disease and with the table given (Table 13) it is clear that the other age groups 30-60 and above 60 is not affected. The retinoblastoma rates have gone up from 0.11/100,000 in 1994-1997 to 0.29/100,000 in 2010-2012 (a relative increase of 163%). Table 14: Age-Standardized Incidence Rates (per 100,000) of Retinoblastoma cases in the Republic of Ireland (1994-2012) by gender across five grouped time-periods 1994-1997 1998-2001 2002-2005 2006-2009 2010-2012 Rates/100,000 Rates/100,000 Rates/100,000 Rates/100,000 Rates/100,000 Gender Men women 0.08 0.03 0.09 0.07 0.06 0.06 0.10 0.08 0.09 0.16 Retinoblastoma age-standardized rates by gender is given above (Table 14). The retinoblastoma rates in men remained almost same during the all given years. In contrast the women retinoblastoma rates have gone up from 0.03/100,000 in 1994-1997 to 0.16/100,000 in 2010-2012 (a relative increase of 433%). 4.2.2. Kaplan-Meier survival curve with cox proportional hazard analysis Survival analysis was done with the same data and KM curves were plotted for variables of significance and explained below. The study subject were 861 and 9 of them were dropped on the day when they got diagnosed. The remained population was 852. During the study period there was 318 deaths observed. The earliest observed was on the day 1 and the last observed or the maximum time observed is 6937 days. The total analysis time for the study is 1956270 days. The cumulative incidence for the total tome period is 0.00016 and the median survival time is 4552. [51] Ocular cancer incidence by gender Kaplan-Meier survival curve of ocular cancer by gender plotted is given below (Figure 4). Male and female do not have different chance of survival from ocular cancer (Log-rank test: χ2 = 0.12; p =0.73). Figure 4: Kaplan-Meier of 1994-2012 Ocular Cancer cases by gender 0.00 0.25 0.50 0.75 1.00 Kaplan-Meier survival estimates by gender 0 Number at risk gender = 0 419 gender = 1 433 2000 4000 analysis time in days 6000 8000 199 204 85 75 19 19 0 0 Female [52] Male Ocular cancer by age groups Km curve survival curve plotted for the different age group is given below (Figure-5). The graph shows clear demarcation between the rates of survival amongst the three age-groups. The below 30 age group shows significantly higher survival than the other two groups. The 3060 age group shows better chance of survival from ocular cancer than the above 60 age group. The graph clearly shows that all the three age groups have different chance of survival. (Logrank test: χ2 = 94.39; p=0.00). Kaplan-Meier survival estimates by age group 0.00 0.25 0.50 0.75 1.00 Figure 5: Kaplan-Meier of 1994-2012 Ocular Cancer cases by age-group 0 Number at risk agegroup = 1 105 agegroup = 2 277 agegroup = 3 470 2000 4000 analysis time in days 6000 8000 55 152 196 33 72 55 10 19 9 0 0 0 <30 age Above 60 age [53] 30-60 age Ocular cancer by morphology The KM survival curve plotted against the morphological type is plotted below (Figure 6). Retinoblastoma shows better survival than the other two groups. The curves of melanoma and other cancers crosses in many areas, and the Log rank may not find the difference between the groups (Log-rank test χ2 = 29.60; p=0.00). The cox regression model also indicates that retinoblastoma (HR=0.03; CI= 0.00-0.23; p=0.00) has 97% better prognosis than the melanoma, and the other cancers have 16% poor prognosis than Melanoma. From the graph it is very clear the retinoblastoma has a very good survival and have only one death. Kaplan-Meier survival estimates by morphology 0.00 0.25 0.50 0.75 1.00 Figure 6: Kaplan-Meier of 1994-2012 Ocular Cancer cases by morphologic types 0 Number at risk type = 1 673 type = 2 64 type = 3 115 2000 4000 analysis time in days 6000 8000 317 33 53 115 22 23 25 6 7 0 0 0 Melanoma Other cancers [54] Retinoblastoma Ocular cancer by smoking status The KM curve shows the survival estimates of ocular cancer by smoking status (Figure 7). The graph shows the survival of EX-smoker and current smoker is poorer than other groups, and the never smoker and unknown groups survival are almost similar (Log-rank test, χ2 = 8.83; p=0.03) and the curve crosses in many place. The cox regression done with the smoking status does not show any significant changes in between the groups (Table 15). 0.00 0.25 0.50 0.75 1.00 Figure 7: Kaplan-Meier of 1994-2012 Ocular Cancer cases by smoking status Number at risk SMOKER = 1 SMOKER = 2 SMOKER = 3 SMOKER = 4 Kaplan-Meier survival estimates bt smoking status 0 2000 4000 analysis time in days 6000 8000 124 365 66 297 63 170 23 147 23 72 9 56 6 19 3 10 0 0 0 0 Current Ex-smoker [55] Never Unknown Ocular cancer by anatomical site (ICD10) The KM curve shows the survival estimates by the involved sites according to ICD10 classification. From the graph it is clear that the retinal and ciliary body have a significantly better survival than the other sites (Log-rank test χ2 =34.02; p=0.00). The cox regression run also showed the same retina (HR=0.17; CI 0.07-0.39; p=0.00), Ciliary body (HR=0.64, CI 0.42-0.98; p=0.04) (Table 15). The death rates by ocular cancer site retina is 83% better than conjunctiva, and the death rate from ciliary body is 36 % better than the conjunctiva. Figure 8: Kaplan-Meier of 1994-2012 Ocular Cancer cases by anatomic sites (ICDs) Conjunctiva Cornea Retina Choroid Ciliary body Lacrimal Gland & duct 0.00 0.50 1.00 0.25 0.75 0.00 0.50 1.00 0.25 0.75 Kaplan-Meier survival estimates by sites (ICD10) 0 0.00 0.50 1.00 0.25 0.75 Orbit 0 2000 4000 6000 Eye, unspecified 8000 0 2000 4000 6000 analysis time in days Graphs by ICD10 [56] 8000 2000 4000 6000 8000 Ocular cancer by year of incidence The KM curve shows the survival estimate by year of incidence (Figure9). From the given graph it is clear that all the year groups have similar survival (Log-rank test χ2 = 5.82; p=0.21). The cox regression done with year of incidence shows a better survival for the 2006-2009 year of incidence (HR=0.67; CI=0.45-0.97; p=0.04). The 2006-2009 year of incidence population had a 33% better survival than the 1994-1997 group. 0.00 0.25 0.50 0.75 1.00 Figure 9: Kaplan-Meier of 1994-2012 Ocular Cancer cases by time-periods Kaplan-Meier survival estimates by year of incidence Number at risk YOI = 1997 YOI = 2001 YOI = 2005 YOI = 2009 YOI = 2012 0 2000 4000 analysis time in days 6000 8000 162 155 197 183 155 106 107 144 46 0 75 82 3 0 0 38 0 0 0 0 0 0 0 0 0 YOI = 1994-1997 YOI = 2002-2005 YOI = 2010-2012 [57] YOI = 1998-2001 YOI = 2006-2009 Ocular cancer by gender and age group The KM curve shows the survival estimates of male and female of all the three age-groups (Figure 10). From the graph it is clear the below 30 years of age have a good survival rates than the other four groups. In that below 30 years male have slightly better survival than the below 30 years female (Log-rank test χ2 =95.02; p=0.00). The cox regression also indicates the same result. The 30 to 60 years age-group (HR=3.76; CI 1.81-7.79; p=0.00) had 3 times poor survival and above 60 age group (HR=9.32; CI=4.60-18.90; p=0.00) had 9 time poor survival than the below 30 years age-group. Figure 10: Kaplan-Meier of 1994-2012 Ocular Cancer cases by gender and age-groups 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Kaplan-Meier survival estimates gender & age-group 0 Female, <30 years Female, 30-60 years Female, 60+ age Male, <30 years Male, 30-60 years Male, 60+ age 2000 4000 6000 8000 0 2000 4000 6000 analysis time in days Graphs by gender and agegroup [58] 8000 0 2000 4000 6000 8000 Ocular cancer by residual status The KM curve shows the survival estimate by the state of residue after treatment (Figure 11). From the graph it is clear that the group whom the residual status cannot be accessed has a poor survival than the no residual group and residual group (Log-rank test χ2 =6.82; p=0.03). The cox regression also shows significance for unknown residual status (HR=1.44; CI=1.06-1.95; p= 0.02) which shows that the unknown residual status death rate is 44% higher than the no residual group (Table 15). Figure 11: Kaplan-Meier of 1994-2012 Ocular Cancer cases by residual 0.00 0.25 0.50 0.75 1.00 Kaplan-Meier survival estimates by residual 0 Number at risk RESIDUAL = 0 230 RESIDUAL = 1 132 RESIDUAL = 2 323 2000 4000 analysis time in days 6000 8000 112 84 113 37 26 36 1 0 2 0 0 0 RESIDUAL = No residue RESIDUAL = Unknown [59] RESIDUAL = Residue + Ocular cancer by morphology and age group The KM curve shows the survival estimates by morphology for every age group (Figure 12). The graphs below clearly shows most of the groups are different form on another (Log-rank test χ2 =121.36; p=0.00). Retinoblastoma under 30 years have a better prognosis than the other all groups (HR=0.11; CI=0.01-0.96; p=0.46) in contrast the above 60 age group of melanoma and other cancers have very poor prognosis (HR=4.99; CI=2.31-10.74; p=0.00). It is clear that the above 60 age-group prognosis of any morphology is five times poorer than the below 30 age-group. Figure 12: Kaplan-Meier of 1994-2012 specific morphologic ocular cancer cases by agegroups Kaplan-Meier survival estimates by Morphology & age-group Melanoma, 30-60 years Melanoma, >60 years Retinoblastoma, < 30 years Other cancers, <30 years Other cancers, 30-60 years 0.00 0.50 1.00 0.25 0.75 0.00 0.50 1.00 0.25 0.75 Melanoma, < 30 age 0.00 0.50 1.00 0.25 0.75 0 2000 4000 6000 Other cancers, > 60 years 0 2000 4000 6000 8000 analysis time in days Graphs by type and agegroup [60] 8000 0 2000 4000 6000 8000 Ocular cancer by morphology and gender The KM curve shows the survival analysis by morphology and gender (Figure 13). Both male and female retinoblastoma groups show better survival compared to the other groups (Logrank test χ2 =31.01; p=0.00). The cox regression done with morphology and gender also shows that the retinoblastoma have a better prognosis than the other group (HR=0.03; CI=0.00-0.23; p=0.00). It is clear the death rate of retinoblastoma is 97% better than the melanoma. Figure 13: Kaplan-Meier of 1994-2012 specific morphologic ocular cancer cases by gender 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 Kaplan-Meier survival estimates by morphology & gender 0 Melanoma, Female Melanoma, Male Retinoblastoma, Female Retinoblastoma, Male Other cancers, Female Other cancers, Male 2000 4000 6000 8000 0 2000 4000 6000 analysis time in days Graphs by type and gender [61] 8000 0 2000 4000 6000 8000 Multivariable analysis Univariate analysis and multivariate analysis has been done with the covariates to adjust the potential confounding and fond the variable which is significantly associated with the survival. In the univariate analysis the 30-60 age-group shows three times more risk as compared to below 30 age group (HR=3.67; CI=1.81-7.79; P=0.00). The above 60 age-group is nine time more risk than the below 30 age-group (HR=9.32; CI4.60-18.89; P=0.00). Never married group showed 44% lower risk than the married group (HR=0.56; CI=0.40-0.78; P=0.00). The widow group showed 63% higher risk than the married group (HR=1.63; CI=1.24-2.14; P=0.00). In the cancer sites retina showed 83% lesser risk (HR=0.17; CI=0.08-0.39; P=0.00) and ciliary body showed 36% lesser risk (HR=0.64; CI=0.42-0.98; P= 0.04) than the conjunctiva involvement. The residual status which is unknown/ not able to access had 44% higher risk (HR=1.44; CI=1.07-1.95; P=0.02) than the no residual group. Table 15: Univariate and Multivariable Cox Proportional Hazards Regression analysis with covariates Covariates Female Male < 30 age 30-60 age > 60 age Married Divorced Separated Single Widowed Unknown Current smoker Never smoked Ex-smoker Unknown Conjunctiva Cornea Retina Univariate Analysis HR (95%CI) P- Value Reference 0.96 (0.77-1.20) 0.73 Reference 3.76 (1.81-7.79) 0.00 9.32(4.60-18.89) 0.00 Reference 1.27 (0.31-5.13) 0.73 0.69 (0.25-1.85) 0.46 0.56 (0.40-0.78) 0.00 1.63 (1.24-2.14) 0.00 0.89 (0.55-1.45) 0.64 Reference 0.77 (0.56-1.05) 0.10 1.32 (0.85-2.05) 0.21 0.82 (0.60-1.14) 0.24 Reference 0.70 (0.21-2.29) 0.56 0.17 (0.08-0.39) 0.00 [62] Multivariate analysis HR (95% CI) P-Value 0.94 (0.71-1.25) 0.69 1.74 (0.62-4.91) 4.90 (1.80-13.35) 0.29 0.00 1.36 (0.33-5.63) 0.67 (0.21-2021) 0.93 (0.60-1.44) 1.12 0.79-1.60 0.68 (0.37-1.24) 0.67 0.52 0.75 0.51 0.20 0.69 (0.46-1.03) 1.02 (0.58-1.80) 0.72 (0.48-1.06) 0.07 0.93 0.10 0.89 (0.25-3.10) 1.38 (0.55-3.44) 0.85 0.49 Choroid Ciliary body Lacrimal G&D Orbit Eye, unspecified No residue Residue present Residue unknown 0.91 (0.62-1.33) 0.64 (0.42-0.98) 1.53 (0.37-6.39) 0.99 (0.54-1.81) 1.23 (0.73-2.09) Reference 1.06 (0.73-1.53) 1.44 (1.07-1.95) 0.63 0.04 0.56 0.99 0.43 1.11 (0.68-1.80) 0.76 (0.44-1.29) Omitted 1.75 (0.77-3.97) 1.30 (0.65-2.57) 0.67 0.31 0.76 0.02 0.88 (0.60-1.29) 1.27 (0.93-1.75) 0.52 0.14 0.18 0.45 Multivariable analysis done and all the potential confounders adjusted and the results are shown in the table (table 15). In this model only the age group above 60 have retained the effect, and is still a potential risk factor for the survival and is five times more risk than the other age groups (HR=4.90;CI =1.80-13.35, P-0.00). The proportional-hazards assumption done with the above model (X2 (19) =14.36; P=0.76). This shows that the null hypothesis cannot be rejected for this model of assumption. [63] Chapter 5 5. Discussion and conclusion Discussion The main aim of the study was to examine the age-standardized rates of ocular neoplasm and the survival rates in the Republic of Ireland over 19 year period from 1994-2012 using the National Cancer Registry Ireland (NCRI) incidence data. The data from NCRI had 861 ocular cancer incident cases and (861+17) 878 (17 bilateral) eyes over a period of 19 years. The main significant finding are: 1. The risk of death by ocular neoplasm was increased significantly (HR=9.32; CI=4.60-18.89; p=0.00) with older age (above 60 years of age) compared to the 30 to 60 years of age group and below 30 years of age group, implying that the risk is around 9 times higher than those below 30 years of age. 2. Retinoblastoma, a genetic disorder has seen 100% increase in the age-standardized incidence rate between 1994 and 2012. The rates in retinoblastoma have also gone up from 0.06/100,000 in 1994-1997 to 0.12/100,000 in 2010-2012 (a relative increase of 100%) it has doubled. 3. Ocular Melanoma in males has seen 29% increase from 0.79/100,000 in 1994-1997 to 1.02 in 2010-2012. Organised screening played no role in diagnosis of ocular cancers during the given period. Ocular melanoma which is caused by prolonged exposure to ultra violet (UV) rays contributes to almost 78% of all eye cancers in the Republic of Ireland. Ocular lymphoma which is a disease mostly associated with immune deficiency and old age is not recorded in found in Republic of Ireland during the 1994 and 2012. Ocular cancer is always listed in the other cancers which occupies a least portion (less than 7%) of all cancers. Ocular cancer in advanced stages leads to loss of eye sight and loss of eye [64] ball and disfiguration. Recent research and treatment is aimed at preserving the sight and anatomical structures, the same is discussed in this chapter in conclusion. 5.1. Retinoblastoma incidence Retinoblastoma a child-hood disease, which affects the under 5 age- group population is a genetic disorder which runs in family. In the NCRI data we have 64 proven retinoblastoma cases during the period 1994-2012. Eight cases (4.88%) were recorded during the four years from 1994-1997, which increased to 17 cases (10.97%) in 2010-2012. Overall the agestandardized retinoblastoma incidence rates increased from 0.06 for 100,000 population to 0.12 for 100,000 population in 2010-2012. A relative increase of 100% was noted during this 19 year period. Age specific incidence rates of retinoblastoma cases were 0.11 cases for 100,000 in population below 30 years age during 1994-1997 which increased to 0.29 for 100,000 during 2010-2012 years, a relative increase of 163% was noted in the below 30 age group population. As the retinoblastoma is highly confined to below 5 years of age group if calculated for the age group the incidence will be higher than this. Worldwide, the cumulative lifetime incidence rate of retinoblastoma is 1 per 18000 to 30000 live births (ABRAMSON and SCHEFLER, 2004). Most of the children lose their sight and eyeball as it involves retina-a vital site for vision. 5.2. Ocular Melanoma Incidence Ocular melanoma is associated with the Ultra Violet (UV) radiation, time of stay outside, occupations like welding and heat. Scandinavian countries have a higher risk of melanoma compared to other countries. North European skin colour, blue and light Iris, navies, and ability to tan are considered risk factors for melanoma. In United States the incidence is reported as 0.5 for 100,000 population where as in Europe it is 0.7 for 100,000 population (Ocularmelanoma.org, 2015). In United Kingdom 70% of all eye cancer is melanoma during [65] the year 2000 (Huerta, 2001). In Africa the incidence is around 0.02-0.04 per 100,000 population and in Asia it is 0.01 for 100,000 population (Futuremedicine.com, 2015). Uveal melanoma incidence in United States among various race during 1992-2000 0.03 (black), 0.03 (Asian), 0.16 (Hispanic), and 0.6 (non-Hispanic white) per 100,000 population (Hu et al., 2005). In Ireland 78 % of all ocular cancers are melanoma, and the standardized rates stay between 0.68 and 0.91 per 100,000 population in the given time period 1994 and 2012. The result proves the Irish (non-Hispanics) in United States are also more prone for ocular melanoma. Older age group (above 60 years of age) are four times prone to ocular melanoma than the adult population (30-60 years of age). Age standardized rates remain around 2.28-3.62 for 100,000 above 60 years population during the years 1994-2012. Whereas for the 30-60 years population the standardized rates remained around 0.60-0.96 for 100,000 population. 5.3. Ocular melanoma and Gender During the study period 673 melanoma cases where noted in that 331 (49.18%) males and 342 (50.82%) females. From the given years melanoma in men have gone up from 0.79 for 100,000 population to 1.02 for 100,000 population, whereas in female it have gone down from 0.94 for 100,000 population to 0.79 for 100,000 population in the years 1994-2012. There is a relative increase of 29% in male melanoma incidence and a 16% decrease in female melanoma incidence in the study years (1994-2012). Worldwide the same is noted in the gender, an increase in male ocular melanoma incidence and decreased female melanoma incidence. 5.4. Screening and ocular cancer Presentation of ocular cancer to the hospital or to GPs was also studied with the data provided by NCRI. Of the total 861 cases, 767 (89%) of the cases reported to the hospital after the symptoms appear (90% melanoma 95% retinoblastoma, and 79% other cancers). Only one of 64 retinoblastoma was identified in unorganised screening. Unorganised screening [66] (opportunistic and screening for other purpose) have found 4 cases (0.46%) cases in the 19 years. Ocular specific screening has found no cases during the study period 1994-2012. Early diagnosis of ocular cancer can prevent visual loss, and eye ball (Eyewiki.aao.org, 2015). In United States all neonates have been screened for retinoblastoma as a routine practise to prevent patients losing the eye ball (enucleation). In advanced ocular cases enucleation is performed to prevent the spread of cancer and as a treatment of choice, which leaves a permanent disfiguration and psychological effect. The American Academy of Paediatrics recommends red reflex examination in neonates, infants and children before discharge and during all subsequent routine health visits (Aapos.org, 2015). 5.5. Ocular cancer survival rate by morphology Survival rates in melanoma and other cancers are poor than the retinoblastoma (figure 6). In total 64 retinoblastoma cases only one died and all the 63 are alive. The log rank run with morphology shows the same that there is difference in survival between morphological subtypes. (χ2 =29.60; p=0.00). In cox regression the retinoblastoma survival is 93% (HR=0.03; CI=0.00-0.23; p=0.00) better than melanoma, and the other cancers are 16% poorer than the melanoma. Worldwide the survival of retinoblastoma is around 95% and in Ireland it is 98.44% during the study period 1994-2012. 5.6. Ocular cancer survival rates by age-group Ocular cancer risk of survival increases by age, overall. In the below 30 years age group, 92.38% were alive, in 30-60 years age-group 71.68% were alive and in above 60 years of age group 49.69% were alive. Below 30 years of age, the survival is best compared to those who were older (Log-rank test χ2= 94.39; p=0.00). Ocular cancer survival is poorer in the 30-60 (HR=3.76; CI=1.81-7.79; p=0.00) and above 60 age group (HR=9.32; CI=4.60-18.89; p=0.00). [67] Age group 30-60 years are 4 time more in survival risk than below 30 years of age and the above 60 years age group are 9 time more riskier than the below 30 years of age group. 5.7. Strengths and limitations of the study This is the first study to identify the survival rate of ocular cancer in the Irish population for maximum of a 19 years follow-up period 1994-2012., using the most comprehensive nationally representative ocular cancer data in Ireland. Morphology wise analysis is one of the strengths of the study. Morphology wise age-standardized incidence and survival analysis was done for a better understanding. The main limitation of this study is the lack of information on staging, loss of sight and loss of eyeball in the data. The study is a secondary analysis and the lack the information on employment (UV) exposure. One of the other limitations is the small number of cases each year that precludes from undertaking robust statistical modelling techniques. 5.8. Implication for public health policy and recommendations for the future Ocular cancer public health policy has two dimensions according to morphology. Screening for retinoblastoma which is a very easy procedure can be implemented by the government as per The American Academy of paediatrician’s recommendations. Every child who visit GP, and hospital should be examined for retinoblastoma and risk group screening should be organised and genetic tests to be conducted. Public health policy in melanoma should involve, weather report, news channels and government agencies and screening. Government should warn people of using artificial tanners. Whether reports and news should have the UV index of the day and warnings. Screening should be organised for the individuals with blue iris, light skin and excessive navies in the body. [68] Furthermore studies should be done on county level UV levels and melanoma incidence. For retinoblastoma there should be studies on the genetic factors and genetic matching before childbirth for high risk individuals. 5.9. Conclusions These detailed population-based results of the Republic of Ireland found some interesting facts of ocular cancer incidence and survival rate. We found there is an increase in male ocular melanoma incidence between the years 1994-2012 with no good explanations. Likewise we found a 100% increase in retinoblastoma incidence, again with no good explanation. Survival rates in older age group is poorer than the adult and younger population during the years 19942012. Further research is needed in the field of retinoblastoma and Male ocular melanoma as there is a significant increase in incidence rates. Screening in under five age for retinoblastoma should also be studied for its effect on preventing loss of vision and eye ball. Awareness should be created on artificial tanning and UV in ocular melanoma prevention. [69] Chapter 6 Referencing Aapos.org, (2015). Retinoblastoma — AAPOS. [online] Available at: http://www.aapos.org/terms/conditions/93 [Accessed 29 Sep. 2015]. ABRAMSON, D. and SCHEFLER, A. (2004). UPDATE ON RETINOBLASTOMA. Retina, 24(6), pp.828-848. Ali, H. (2014). Incidence and Survival Rates in Oral Cancer in the Republic of Ireland: A timetrend analysis, 1994-2009. MPH. The University College Cork, Cork, Ireland. Anon, (2015). [online] Available at: http://www.ncri.ie/publications/cancer-trends-andprojections/cancer-trends-melanoma [Accessed 12 Aug. 2015]. Arepalli, S., Kaliki, S., Shields, C., Emrich, J., Komarnicky, L. and Shields, J. (2014). Plaque Radiotherapy in the Management of Scleral-Invasive Conjunctival Squamous Cell Carcinoma. JAMA Ophthalmol, 132(6), p.691. Atlasgeneticsoncology.org, (2015). Eye tumors: an overview. [online] Available at: http://atlasgeneticsoncology.org/Tumors/EyeTumOverviewID 5272.html [Accessed 20 Aug. 2015]. Atlasgeneticsoncology.org, (2015). Soft tissue tumors: Embryonal rhabdomyosarcoma. [online] Available at: http://atlasgeneticsoncology.org/Tumors/EmbryoRhabdomyoI D5193.html [Accessed 15 Aug. 2015]. Ats.ucla.edu, (2015). Stata FAQ: How can I quickly convert many string variables to numericvariables?. [online] Available at: http://www.ats.ucla.edu/stat/stata/faq/destring.htm [Accessed 15 Sep. 2015]. cancer, T. (2015). Types of eye cancer | Cancer Research UK. [online] Cancerresearchuk.org. Available at: http://www.cancerresearchuk.org/aboutcancer/type/eye-cancer/about/types-of-eye-cancer [Accessed 6 Aug. 2015]. Cancer.ie, (2015). Irish sun strong enough to cause damage and increase skin cancer risk on 90% of days last summer. [online] Available at: http://www.cancer.ie/content/irish-sun-strong-enough-causedamage-and-increase-skin-cancer-risk-90-days-lastsummer#sthash.ReJDbOPA.dpbs [Accessed 24 Aug. 2015]. Cancer.Net, (2012). Eye Cancer Overview. [online] Available http://www.cancer.net/cancer-types/eye-cancer/overview [Accessed 10 Aug. 2015]. [70] at: Cancer.org, (2015). Does UV radiation cause cancer?. [online] Available at: http://www.cancer.org/cancer/cancercauses/radiationexposure andcancer/uvradiation/uv-radiation-does-uv-cause-cancer [Accessed 24 Aug. 2015]. Cancer.org, (2015). Survival rates of small intestine adenocarcinoma, by stage. [online] Available at: http://www.cancer.org/cancer/smallintestinecancer/detailedgui de/small-intestine-cancer-survival-rates [Accessed 25 Aug. 2015]. Cancer.org, (2015). What are the risk factors for small intestine adenocarcinoma?. [online] Available at: http://www.cancer.org/cancer/smallintestinecancer/detailedgui de/small-intestine-cancer-risk-factors [Accessed 15 Aug. 2015]. Cancer.org, (2015). What is eye cancer?. [online] Available at: http://www.cancer.org/cancer/eyecancer/detailedguide/eyecancer-what-is-eye-cancer [Accessed 10 Aug. 2015]. Cancer.org, (2015). What is ultraviolet (UV) radiation?. [online] Available at: http://www.cancer.org/cancer/cancercauses/sunanduvexposure /skincancerpreventionandearlydetection/skin-cancerprevention-and-early-detection-what-is-u-v-radiation [Accessed 24 Aug. 2015]. Chang, N., Feng, R., Gao, Z. and Gao, W. (2010). Skin cancer incidence is highly associated with ultraviolet-B radiation history. International Journal of Hygiene and Environmental Health, 213(5), pp.359-368. Chodick, G. (2009). Risk of Cataract Extraction Among Adult Retinoblastoma Survivors. Archives of Ophthalmology, 127(11), p.1500. Conway, R. (2001). Primary iris melanoma: diagnostic features and outcome of conservative surgical treatment. British Journal of Ophthalmology, 85(7), pp.848-854. Corwin.com, (2015). [online] Available at: http://www.corwin.com/upmdata/18485_05_Gilbert(Researching)_Ch_04_WATERMARK ED [Accessed 26 Aug. 2015]. Critical Appraisal Skills Programme (CASP), (2015). Critical Appraisal Skills Programme (CASP). [online] Available at: http://www.caspuk.net/#!checklists/cb36 [Accessed 22 Aug. 2015]. Cso.ie, (2015). Population by Age Group, Sex and CensusYear - StatBank - data and statistics. [online] Available at: http://www.cso.ie/px/pxeirestat/Statire/SelectVarVal/Define.a sp?Maintable=CDD02&Planguage=0 [Accessed 15 Sep. 2015]. [71] de Jong, M., Kors, W., de Graaf, P., Castelijns, J., Kivelä, T. and Moll, A. (2014). Trilateral retinoblastoma: a systematic review and meta-analysis. The Lancet Oncology, 15(10), pp.1157-1167. Desjardins, L., Lumbroso-Le Rouic, L., Levy-Gabriel, C., Dendale, R., Delacroix, S., Nauraye, C., Est&egrave;ve, M., Plancher, C. and Asselain, B. (2006). Combined Proton Beam Radiotherapy and Transpupillary Thermotherapy for Large Uveal Melanomas: A Randomized Study of 151 Patients. Ophthalmic Res, 38(5), pp.255-260. Development of Metastatic Disease After Enrollment in the COMS Trials for Treatment of Choroidal Melanoma. (2005). Archives of Ophthalmology, 123(12), p.1639. Diener-West, M. (2004). Screening for Metastasis From Choroidal Melanoma: The Collaborative Ocular Melanoma Study Group Report 23. Journal of Clinical Oncology, 22(12), pp.2438-2444. EL-AAL, A. (2006). Rhabdomyosarcoma: the experience of the pediatric unit of Kasr El-Aini Center of Radiation Oncology and Nuclear Medicine (NEMROCK) (from January 1992 to January 2001). Journal of the Egyptian National Cancer Institute, 18(1). Eyewiki.aao.org, (2015). Ocular Melanoma - EyeWiki. [online] Available http://eyewiki.aao.org/Ocular_Melanoma#Prognosis [Accessed 25 Aug. 2015]. at: Eyewiki.aao.org, (2015). Retinoblastoma - EyeWiki. [online] Available at: http://eyewiki.aao.org/Retinoblastoma [Accessed 24 Aug. 2015]. Eyewiki.aao.org, (2015). Rhabdomyosarcoma - EyeWiki. [online] Available at: http://eyewiki.aao.org/Rhabdomyosarcoma [Accessed 24 Aug. 2015]. Eyewiki.aao.org, (2015). Uveal Melanoma - EyeWiki. [online] Available at: http://eyewiki.aao.org/Uveal_Melanoma [Accessed 24 Aug. 2015]. Futuremedicine.com, (2015). Incidence, prevalence and epidemiology of ocular melanoma, Ocular Melanoma: Advances in Diagnostic and Therapeutic Strategies, Future Medicine. [online] Available at: http://www.futuremedicine.com/doi/abs/10.2217/fmeb2013.13 .176 [Accessed 29 Sep. 2015]. Genetics Home Reference, (2015). Retinoblastoma. [online] Available at: http://ghr.nlm.nih.gov/condition/retinoblastoma [Accessed 16 Aug. 2015]. Gichuhi, S., Sagoo, M., Weiss, H. and Burton, M. (2013). Epidemiology of ocular surface squamous neoplasia in Africa. Trop Med Int Health, 18(12), pp.1424-1443. [72] Groenewald, C., Konstantinidis, L. and Damato, B. (2012). Effects of radiotherapy on uveal melanomas and adjacent tissues. Eye, 27(2), pp.163-171. Gunderson, L., Tepper, J. and Bogart, J. (2012). Clinical radiation oncology. Philadelphia, Pa.: Saunders/Elsevier. Holly, E., Aston, D. and Char, ,. (1990). Uveal melanoma in relation to ultraviolet light exposure and host factors. Cancer research, 50(18). Hu, D., Yu, G., McCormick, S., Schneider, S. and Finger, P. (2005). Population-Based Incidence of Uveal Melanoma in Various Races and Ethnic Groups. American Journal of Ophthalmology, 140(4), pp.612.e1-612.e8. Huerta, C. (2001). Incidence of ocular melanoma in the general population and in glaucoma patients. Journal of Epidemiology & Community Health, 55(5), pp.338-339. Iarc.fr, (2015). IARC Publications - PDFs online - Cancer Epidemiology - Cancer Incidence in Five Continents Vol. IX. [online] Available at: http://www.iarc.fr/en/publications/pdfs-online/epi/sp160 [Accessed 15 Sep. 2015]. Kantarjian, H., Wolff, R. and Koller, C. (2011). The MD Anderson manual of medical oncology. New York: McGraw-Hill. Kim, J., Abramson, D. and Dunkel, I. (2007). Current Management Strategies for Intraocular Retinoblastoma. Drugs, 67(15), pp.2173-2185. MedicineNet, Ncbi.nlm.nih.gov, (2015). (2015). Melanoma. [online] Available at: http://www.medicinenet.com/script/main/art.asp?articlekey=4 342 [Accessed 1 Oct. 2015]. Home - PubMed - NCBI. [online] Available at: http://www.ncbi.nlm.nih.gov/pubmed [Accessed 31 Aug. 2015]. Ncri.ie, (2015). Cancer factsheets | National Cancer Registry Ireland. [online] Available at: http://www.ncri.ie/factsheets [Accessed 15 Sep. 2015]. Ncri.ie, (2015). Cancer in Ireland 1994-2011: Annual report of the National Cancer Registry | National Cancer Registry Ireland. [online] Available at: http://www.ncri.ie/publications/statistical-reports/cancerireland-1994-2011-annual-report-national-cancer-registry [Accessed 9 Aug. 2015]. Ncri.ie, (2015). Cancer trends - Melanoma | National Cancer Registry Ireland. [online] Available at: http://www.ncri.ie/publications/cancer-trendsand-projections/cancer-trends-melanoma [Accessed 9 Aug. 2015]. [73] Ncri.ie, (2015). Incidence statistics | National Cancer Registry Ireland. [online] Available at: http://www.ncri.ie/data/incidence-statistics [Accessed 15 Sep. 2015]. Ncri.ie, (2015). What we do | National Cancer Registry Ireland. [online] Available at: http://www.ncri.ie/about/history [Accessed 14 Aug. 2015]. Ncri.ie, (2015). What we do | National Cancer Registry Ireland. [online] Available at: http://www.ncri.ie/about/history [Accessed 15 Sep. 2015]. Nei.nih.gov, (2015). Diagram of the Eye | National Eye Institute. [online] Available at: https://nei.nih.gov/health/eyediagram [Accessed 10 Aug. 2015]. Nhs.uk, (2015). Eye cancer NHS Choices. [online] Available http://www.nhs.uk/conditions/cancer-of-theeye/pages/definition.aspx [Accessed 12 Aug. 2015]. at: Ocularmelanoma.org, (2015). Ocular Melanoma Foundation | OMF - disease. [online] Available at: http://www.ocularmelanoma.org/disease.htm [Accessed 29 Sep. 2015]. Parameningeal rhabdomyosarcoma in pediatric age: results of a pooled analysis from North American and European cooperative groups. (2014). Annals of Oncology, 25(3), pp.756-756. Paridaens, A., McCartney, A., Minassian, D. and Hungerford, J. (1994). Orbital exenteration in 95 cases of primary conjunctival malignant melanoma. British Journal of Ophthalmology, 78(7), pp.520-528. Peat, J., Barton, B. and Elliott, E. (2008). Statistics workbook for evidence-based health care. Chichester, West Sussex, UK: Wiley-Blackwell/BMJ Books. Rahimi, F., Alipour, F., Ghazizadeh Hashemi, H. and Hashemian, M. (2015). Topical mitomycin-C for treatment of partially-excised ocular surface squamous neoplasia. Archives of Iranian medicine, 12(1). Shah, C., Weis, E., Lajous, M., Shields, J. and Shields, C. (2005). Intermittent and Chronic Ultraviolet Light Exposure and Uveal Melanoma. Ophthalmology, 112(9), pp.1599-1607. Stata.com, (2015). Data Analysis and Statistical Software | Stata. [online] Available at: http://www.stata.com/ [Accessed 22 Aug. 2015]. TheFreeDictionary.com, (2015). Eye Cancer. [online] Available at: http://medicaldictionary.thefreedictionary.com/Eye+Cancer [Accessed 6 Aug. 2015]. Updated by: Adam S. Levy, a. (2015). Retinoblastoma: MedlinePlus Medical Encyclopedia. [online] Nlm.nih.gov. Available at: http://www.nlm.nih.gov/medlineplus/ency/article/001030.htm [Accessed 16 Aug. 2015]. [74] Vajdic, C., Kricker, A., Giblin, M., McKenzie, J., Aitken, J., Giles, G. and Armstrong, B. (2002). Sun exposure predicts risk of ocular melanoma in Australia. International Journal of Cancer, 101(2), pp.175182. Venturi, C., Bracco, S., Cerase, A., Cioni, S., Galluzzi, P., Gennari, P., Vallone, I., Tinturini, R., Vittori, C., De Francesco, S., Caini, M., D’Ambrosio, A., Toti, P., Renieri, A. and Hadjistilianou, T. (2012). Superselective ophthalmic artery infusion of melphalan for intraocular retinoblastoma: preliminary results from 140 treatments. Acta Ophthalmologica, 91(4), pp.335-342. Wikipedia, (2015). Conjunctival squamous cell carcinoma. [online] Available at: https://en.wikipedia.org/wiki/Conjunctival_squamous_cell_ca rcinoma [Accessed 20 Aug. 2015]. [75] 7. Appendices Appendix 1: CASP toolkits used for critical appraisal and quality assessment of each study included in the literature review section. 1. Appraisal tool for clinical trial 2. Appraisal tool for systematic reviews 3. Appraisal tool for cohort studies. CRITICAL APPRISAL SKILLS PROGRAMME Making sense of evidence 11 questions to help you make sense of a trial How to use this appraisal tool Three broad issues need to be considered when appraising the report of a randomised controlled trial:  Are the results of the trial valid?  What are the results?  Will the results help locally? The 11 questions on the following pages are designed to help you think about these issues systematically. The first two questions are screening questions and can be answered quickly. If the answer to both is yes, it is worth proceeding with the remaining questions. There is some degree of overlap between the questions, you are asked to record a yes, no or can’t tell to most of the questions. A number of prompts are given after each question. These are designed to remind you why the question is important. Record your reasons for your answers in the spaces provided Screening Questions [76] Consider: An issue can be ‘focused’ In terms of  The population studied  The intervention given  The comparator given  The outcomes considered randomised? Consider:  How was this carried out, some methods  may produce broken allocation concealment  Was the allocation concealed from researchers? Detailed questions personnel blinded? Consider:  Health workers could be; clinicians, nurses etc  Study personnel – especially outcome assessors Consider: Look at  Other factors that might affect the outcome such as age,  sex, social class, these may be called baseline characteristics 5. A were the groups treated equally? the trial properly accounted for at its [77] conclusion? Consider:  Was the trial stopped early?  Were patients analysed in the groups to which  they were randomised? 7. How large was the treatment effect? Consider:  What outcomes were measured?  Is the primary outcome clearly specified?  What results were found for each outcome?  Is there evidence of selective reporting of outcomes? 8. How precise was the estimate of the treatment effect? Consider:  What are the confidence limits?  Were they statistically significant? (or to the local population?) Consider:  Do you have reason to believe that your population  of interest is different to that in the trial  If so, in what way? considered? Consider: [78]  Is there other information you would like to have seen?  Was the need for this trial clearly described? Consider:  Even if this is not addressed by the trial, what do you think? Making sense of evidence 10 questions to help you make sense of a review How to use this appraisal tool Three broad issues need to be considered when appraising the report of a systematic review: • Are the results of the review valid? (Section A) • What are the results? (Section B) • Will the results help locally? (Section C) The 10 questions on the following pages are designed to help you think about these issues systematically. The first two questions are screening questions and can be answered quickly. If the answer to both is “yes”, it is worth proceeding with the remaining questions. There is some degree of overlap between the questions, you are asked to record a “yes”, “no” or “can’t tell” to most of the questions. A number of prompts are given after each question. These are designed to remind you why the question is important. Record your reasons for your answers in the spaces provided. Screening Questions HINT: An issue can be ‘focused’ In terms of • The population studied • The intervention given • The outcome considered HINT: ‘The best sort of studies’ would [79] • Addressthe reviews question • Have an appropriate study design (usually RCTs for papers evaluating interventions) Detailed questions studies were included? HINT: Look for • Which bibliographic databases were used • Follow up from reference lists • Personal contact with experts • Search for unpublished as well as published studies • Search for non-English language studies the quality of the included studies? HINT: The authors need to consider the rigour of the studiesthey have identified. Lack of rigour may affect the studies’ results. (“All that glistersis not gold” Merchant of Venice – Act II Scene 7) was it reasonable to do so? HINT: Consider whether • The results were similar from study to study • The results of all the included studies are clearly displayed • The results of the different studies are similar • The reasonsfor any variations in results are discussed 6. What are the overall results of the review? [80] HINT: Consider • If you are clear about the review’s‘bottom line’ results • What these are (numerically if appropriate) • How were the results expressed (NNT, odds ratio etc 7. How precise are the results? HINT: Look at the confidence intervals, if given HINT: Consider whether • The patients covered by the review could be sufficiently different to your population to cause concern • Your local setting is likely to differ much from that of the review HINT: Consider whether • Is there other information you would like to have seen HINT: Consider • Even if this is not addressed by the review, what do you think? Making sense of evidence 12 questions to help you make sense of cohort study How to use this appraisal tool Three broad issues need to be considered when appraising a cohort study:  Are the results of the study valid? (Section A)  What are the results? (Section B)  Will the results help locally? (Section C) The 12 questions on the following pages are designed to help you think about these issues systematically. The first two questions are screening questions and can be answered quickly. If the answer to both is “yes”, it is worth [81] proceeding with the remaining questions. There is some degree of overlap between the questions, you are asked to record a “yes”, “no” or “can’t tell” to most of the questions. A number of italicised prompts are given after each question. These are designed to remind you why the question is important. Record your reasons for your answers in the spaces provided Screening Questions 1. Did the stud HINT: A question can be ‘focused’ In terms of  The population studied  The risk factors studied  The outcomes considered  Is it clear whether the study tried to detect a beneficial or harmful effect? HINT: Look for selection bias which might compromise the generalisibility of the findings:  Was the cohort representative of a defined population?  Was there something special about the cohort?  Was everybody included who should have been included? Detailed questions minimise bias? HINT: Look for measurement or classification bias:  Did they use subjective or objective measurements?  Do the measurements truly reflect what you want them to (have they been validated)?  Were all the subjects classified into exposure groups using the same procedure [82] minimise bias? HINT: Look for measurement or classification bias:  Did they use subjective or objective measurements?  Do the measures truly reflect what you want them to (have they been validated)?  Has a reliable system been established for detecting all the cases (for measuring disease occurrence)?  Were the measurement methods similar in the different groups?  Were the subjects and/or the outcome assessor blinded to exposure (does this matter)? confounding factors? List the ones you think might be important, that the author missed. confounding factors in the design and/or analysis? List: enough? enough? 7. What are the results of this study? HINT: Consider  What are the bottom line results? [83]  Have they reported the rate or the proportion between the exposed/unexposed, the ratio/the rate difference?  How strong is the association between exposure and outcome (RR,)?  What is the absolute risk reduction (ARR)? 8. How precise are the results? HINT: Look for the range of the confidence intervals, if given. 9. Do you believe the results? HINT: Consider  Big effect is hard to ignore!  Can it be due to bias, chance or confounding?  Are the design and methods of this study sufficiently flawed to make the results unreliable?  Bradford Hills criteria (e.g. time sequence, dose-response gradient, biological plausibility, consistency) HINT: Consider whether  A cohort study was the appropriate method to answer this question  The subjects covered in this study could be sufficiently different from your population to cause concern  Your local setting is likely to differ much from that of the study  You can quantify the local benefits and harms available evidence? 12. What are the implications of this study for practice? HINT: Consider [84]  One observational study rarely provides sufficiently robust evidence to recommend changes to clinical practice or within health policy decision making  For certain questions observational studies provide the only evidence  Recommendations from observational studies are always stronger when supported by other evidence [85] Appendices 2 [86]