International Ovarian Tumour Analysis (IOTA) Phase 5

A multicentre study to examine the short and long term outcomes of the
conservative management of benign-looking adnexal masses and the pre-
operative characterisation of ovarian tumours

STUDY CO-ORDINATORS

Tom Bourne, Lil Valentin, Dirk Timmerman

Contact details:
Dirk Timmerman, MD, PhD
Department of Obstetrics and Gynaecology, University Hospitals Leuven,
Herestraat 49, B-3000 Leuven, BELGIUM.
Telephone: + 32 16 344201 (office) Fax: + 32 16 344205
+ 32 16 344215 (secretary)
E-mail: dirk.timmerman@uzleuven.be

STEERING COMMITTEE

Dirk Timmerman (University Hospitals, KU Leuven)

Tom Bourne (Imperial College, London)
Antonia C. Testa (Università Cattolica di Sacro Cuore, Roma)
Lil Valentin (University of Lund / Malmö)
Ben Van Calster (KU Leuven)
Sabine Van Huffel (ESAT-SISTA, KU Leuven)
Ignace Vergote (University Hospitals, KU Leuven)

Summary

The medium to long term behaviour of benign-looking adnexal masses that do not undergo surgery is
unknown. It is possible for these masses to undergo malignant transformation, rupture or torsion.
Furthermore they may undergo changes in volume and/or morphology that may or may not predict any of
these behaviours. To date, no research has rigorously investigated the long-term behaviour of such
masses. Consequently, there are no evidence-based guidelines on the optimal management of the majority
of adnexal tumours. It is therefore not surprising that clinical practice is highly variable, with some
clinicians preferring to operate on virtually any mass. When a clinician decides not to operate, the time
intervals selected for follow up scans is often arbitrarily chosen. On the other hand, we do have some
convincing data to suggest that simple cysts are rarely malignant and so it is generally thought that
operating on these common tumours is probably not necessary and simply increases costs and morbidity.
Developing new insights into the natural history of benign looking conservatively managed ovarian
masses would potentially change the management of thousands of women, by avoiding surgery or even
further surveillance for some and detecting cancer earlier or even preventing it for others.
In this international multicentre study IOTA phase 5 we aim to develop the optimal evidence-based
algorithm for the management of all adnexal tumours in order to improve the detection of ovarian cancer
while at the same time reducing the number of unnecessary operations. At least three thousand patients
with an adnexal mass will undergo an ultrasound examination and if no operation is needed they will be
followed up for at least 5 years. At each visit the investigator will assess the tumour and decide whether
surgery is necessary based on the available information and local protocols. Survival and logistic
regression analysis will be used to develop decision aids to assist clinicians in making decisions regarding
surgery and follow up.

Relation with other IOTA studies

The IOTA study (International Ovarian Tumour Analysis) is a multicentre collaborative project for
the pre-operative characterisation of ovarian tumours..

IOTA phase 1: The first phase of IOTA was conducted between 1999 and 2002. Several new
mathematical models were developed based on the prospectively collected data of 1066 patients with a
persisting adnexal tumour from 9 European centres (1). Between 2002 and 2005 three centres
continued the prospective collection in order to be able to perform an internal validation of
mathematical models developed in IOTA phase 1. In this so-called IOTA phase 1b study a dataset of
507 new patients was prospectively collected in 3 out of the 9 original IOTA centres (2). All models
proved to perform excellently with areas under the ROC curves of more than 0.94.

IOTA phase 2: The second phase of IOTA consisted of an external validation of the models and this
was conducted between 2005 and 2007. The diagnostic algorithms were prospectively validated on
1938 patients with adnexal tumours in 19 centres in Belgium, Italy, UK, Sweden, Poland, Czech
Republic, Canada, and China (3). A first analysis showed that overall performance of the logistic
regression models was excellent (area under the ROC curve 0.94). We concluded that a subgroup of
“uncertain” tumours needs a reliable second stage test in order to help even experienced ultrasound
examiners.

IOTA phase 3: The third phase of the IOTA study started in 2010.
The aim was to validate the added value of mathematical models as new diagnostic tool in the
prediction of ovarian cancer in clinical practice in centres that were involved in IOTA phase 1 or 2. It
is a temporal validation of IOTA mathematical models as a first stage examination. However in cases
where the prediction is unreliable, we aim to further improve the predictive performance of this
diagnostic tool with second stage tests, such as new sets of tumor markers, proteomics and three-
dimensional Power Doppler ultrasonography.

IOTA phase 4: Randomised controlled trial in 7 London hospitals. Clinical implementation of IOTA
logistic regression models LR2 vs. the routinely used Risk of Malignancy Index. Assessment of efficacy,
referral pattern and costs.

Introduction
There is very little evidence on which to base a recommendation on how apparently benign looking
adnexal masses should be managed. Because the natural history of such adnexal masses is not known,
and because of the fear of “missing” ovarian cancer, many adnexal masses are currently surgically
removed, even if they do not manifest any signs of malignancy. This is not optimal, because every
surgical procedure is associated with risks of both short-term and long-term complications, for example
pulmonary embolism, deep vein thrombosis, and bowel perforation or obstruction (4). Furthermore we do
not know if benign ovarian lesions impact on fertility, although we know that surgery on ovaries may
cause adhesions, which in turn may cause infertility, chronic pelvic pain and bowel obstruction. We do
not know the frequency of these complications, nor do we know how often benign adnexal masses are
associated with complications such as torsion if they are not removed. There is some evidence, however,
that expectant management of presumed ovarian dermoid cysts less than 6 cm is safe and does not seem
to interfere with pregnancy or delivery (5-7). The available data also suggest that expectant management
of simple cysts less than 5 cm in post-menopausal women is a safe strategy (8,9).

On the other hand ovarian cancer is associated with a high mortality rate and significant morbidity. It is
the fifth leading cause of cancer-related deaths (10). The lifetime risk of developing ovarian cancer is
around 1 in 50 to 1 in 70 (11,12). Every year more than 200,000 new cases are diagnosed worldwide. T he
disease has a poor prognosis, with five-year relative survival strongly depending on disease stage (12,13).
For example, Cancer Research UK reports survival rates of 73% for stage I versus 16% for stage IV
ovarian cancer.

New diagnostic strategies to decrease mortality are needed, as treatment advances have not decreased
mortality over the past 20 years (14). Effective screening programmes may help, but current candidate
tests remain unsatisfactory (15). A crucial issue is that ovarian cancer is typically asymptomatic in its
early stages. Screening algorithms have generally resulted in high sensitivity at the cost of a large number
of false positives. As a result a large number of surgical interventions are made in order to find relatively
few cancers. However although one study has suggested that there is no benefit in removing benign
ovarian tumors (16), there are no conclusive data to inform us regarding the long term behaviour of
presumed benign ovarian cysts left in situ. Should benign cysts have malignant potential, then a policy of
removing such masses may have a significant impact on mortality from this disease.
Research directed towards the use of diagnostic tests and models to predict malignancy in ovarian
tumours has focused on masses that have been subsequently surgically removed in order to provide a clear
histological end point. Clinicians decide whether to operate on an ovarian mass depending on a number
of factors. These may include the subjective characterisation of the mass using ultrasound, the use of
simple models such as the risk of malignancy index, the age of the patient, the serum CA 125 level and
the presence or absence of symptoms such as pain. The management of cysts that are not removed
surgically is not evidence based and often subject to wide variation. In the absence of rigorous follow up
data, we do not know how many false negative results for cancer are associated with these cysts, or if they
sometimes undergo malignant transformation. We will only gain this knowledge by long term systematic
follow up of a large cohort of ovarian cysts.

A number of studies have focused on the prediction of malignancy in surgically removed masses (for
overviews, see references 2,17-19). There is strong scientific evidence that subjective evaluation of a
mass using ultrasound by an experienced examiner is a very good method for discriminating between
benign and malignant adnexal masses (2,3,20-21), and that a correct histological diagnosis can be
suggested on the basis of ultrasound findings in many cases (22,23). We have previously established the
International Ovarian Tumor Analysis (IOTA) group to develop and validate prediction models based on
large, multi-centre datasets with standardised definitions and data collection procedures (25). The aim was
to develop robust models to predict malignancy that performed well, and were widely generalisable. In
doing so we aimed to overcome the shortcomings of earlier studies such as small sample sizes, single
centre recruitment, and lack of standardised data collection. These models (1, 25-27) successfully passed
temporal and external validation (2,3,28). Following these validation studies, we selected two logistic
regression models for further study. The first (LR1) is a model with 12 predictors, the second (LR2)
contains only six predictors. Even in postmenopausal women conservative management and sonographic
follow up of incidental unilocular and multilocular cysts <7cm may be a valuable option (29).

Objectives

The general aim of this study is the development of the optimal algorithm for the management of all
adnexal masses. This can be broken down into different specific objectives: 1) to study the occurrence of
complications such as rupture, torsion, or malignancy in patients with benign looking conservatively
treated masses; 2) to test the published IOTA diagnostic models for predicting that a mass is malignant at
first visit or benign (either on the basis of histology following surgery or by the absence of malignant
features on an ultrasound scan one year after the initial visit), and to predict complications (e.g. ocurrence
of malignancy and other) during long-term follow-up; 3) to investigate factors that may be related to the
need for surgery during long-term follow up; 4) to study the natural history of conservatively treated
benign looking masses and to establish descriptive curves of the longitudinal changes seen in parameters
from conservatively managed benign tumors (e.g. change in diameter, size of any solid component,
number of papillations, or color score). We hope these curves will allow us to determine if any particular
growth pattern is associated with complications or malignancy

Related to these four objectives, we aim to carry out the following analyses: 1) descriptive analysis of
complications overall and by participating center (anonymised) and overall Kaplan-Meier curves for the
need for surgery among benign looking masses; 2) estimation of discriminatory ability of LR1 for
malignancy at the initial visit using the c-index and ROC curves and for complications during long-term
follow-up using the hazard ratio and c-index within the context of survival analysis; 3) survival analysis to
investigate predictors of the need for surgery during long-term follow-up of non-operated masses; 4) the
development of longitudinal curves of the changes seen in the characteristics of non-operated masses
using longitudinal analysis techniques such as mixed models and functional linear discriminant analysis
(FLDA).

Methods

Study design
International multicenter prospective observational cohort study

Eligible for inclusion

 Any woman at least 18 years old with an adnexal mass.
 Any mass with benign ultrasound morphology may be suitable for conservative management.
 Pregnant patients can be included, but their data will be analysed separately.
Exclusion Criteria
.
 Cysts that are deemed to be clearly physiological and less than 3 cm in maximum diameter are
not eligible for inclusion.
 Any cyst with features of malignancy is excluded from the conservative management
 The denial or withdrawal of oral informed consent

Official approval by the Ethics Committee

The multicentre project IOTA phase 5 will be submitted to the Ethics Committee of the University
Hospitals Leuven as main investigating centre as well as in each participating centre.
The study will be performed in accordance with generally accepted standards of Good Clinical
Practice and the investigators will adhere to all applicable laws and regulations governing the conduct
of clinical trials, including but not limited to the ICH Harmonized Tripartite Guidelines for Good
Clinical Practice and the Declaration of Helsinki (2008).

Insurance policy
This multicentre international study is initiated by the University Hospitals Leuven, Belgium. Each
participating centre outside Belgium is fully responsible for patient care within its own hospital in
agreement with local laws. Each centre is also responsible for all legal aspects of patient care and for
its own insurance for all matters related to this study.

Financial Support

The IOTA phase 3 project is supported by an Applied Biomedical Research grant (Toegepast
Biomedisch Onderzoek, TBM) from the Flanders Institute for Scientific and Technological Research:
IWT Flanders, Belgium (IWT-TBM 070706). This grant covers costs of central data collection,
proteomic analysis, analysis of new tumour markers and statistical analyses. For IOTA phase 5 we
received a research grant for a doctoral researcher by the Flemish Fund for Scientific Research (FWO
Vlaanderen 06260, IOTA5).

There is no financial compensation for principal investigators nor patients.

Definition of benign ultrasound morphology

This is defined on the basis of subjective assessment of ultrasound findings by an experienced
ultrasound examiner. Only lesions where the ultrasound examiner is certain or almost certain that the
lesion is benign can be managed conservatively. The management of benign masses will be decided
according to local protocols.

Number of study patients and recruitment period

This is an observational study and therefore a sample size cannot be calculated. We aim to collect at
least 3000 women with an adnexal mass and at least 1000 women with an adnexal mass managed
conservatively. We plan an initial recruitment period of eighteen months. Patients will be followed up for
at least 5 years, unless surgical intervention is necessary.
Follow-up
Ultrasound (and clinical) follow up will be organised by the ultrasound examiner who entered the
patient into the study. Follow-up will be after 3 months (maximal range 1-4 months), 6 months
(maximal range 4-8 months) and then every 12 months (maximal range 10-14 months).
Although measurement of serum CA 125 levels is encouraged, it is not mandatory for inclusion in the
study. If CA125 is measured it should be recorded in the study screen and preferably measured on
each visit.

Duration of follow-up
A yearly analysis will be carried out in order to evaluate acute complications. The duration of follow
up will be for at least 5 years and is not limited as long as the patient is compliant with the study and
the study is ongoing.
Departments (e.g. radiology departments) that are not involved with clinical decision making about
plans for follow up or surgery cannot participate in the full IOTA 5 study. In these centres data can be
prospectively collected as an observational study. Only patients with appropriate outcome measures
(i.e. follow up ultrasonography after one or more years or patients with complete details of clinical
history or surgical procedures) will be included in any statistical analysis.

Collection of clinical data

Family history: Number of first degree relatives with ovarian cancer (0-...)

Medical history: Personal history of ovarian cancer and breast cancer

Age (years)
Previous hysterectomy (yes/no)
Previous oophorectomy (yes/no)
Contraception (drop down list) (None/oral combined contraceptive
pill/progestogen only pill/ patch/vaginal ring/Mirena coil/copper IUD)
Hormonal therapy (yes, no).

Is the patient currently wishing to conceive? (yes/no)

Menopausal status (pre- or postmenopausal)
History of subfertility? Yes/no
History of ovarian stimulation for subfertility? Yes/no

For ALL patients before menopause two extra questions pop up:

Patient is currently pregnant? (No/Yes)

Patient became pregnant during the last year? (No/Yes)

 If Yes: Outcome of pregnancy (it should be possible to enter more than one date, should there
be more than one pregnancy during follow-up):
 o Ongoing pregnancy
o Miscarriage; date: …
o Ectopic; date: …
o Termination; date: …
o Delivery; date: …
o Complications from the lesion during pregnancy? No/Yes (pop up list):
 Acute pain
 Chronic pain
 Suspected torsion
 Infection
 Haemorrhage related to the cyst
 Cyst rupture
 Required surgery
 Other, please specify: ….
o Complications from the lesion during delivery? No/Yes (pop up list):
 Acute pain
 Suspected torsion
 Haemorrhage
 Cyst rupture
 Obstructed labour
 Malpresentation (e.g. breech or unstable lie)
 Other, please specify: ….….

Ultrasound examination
A standardized ultrasound examination following the IOTA protocol is carried out.

All ultrasound variables are included in the dedicated software. In the database 0 always means NO
and 1 always means YES.
The adnexal lesion is that part of an ovary or of an adnexal mass that is judged by ultrasonography to
be not consistent with normal physiology. This can be a persistent unilocular cyst, surrounded by
normal looking ovarian stroma with some follicles. In this case the whole ovary containing the cyst is
the ‘ovary’, whereas the unilocular cyst is the ‘lesion’. Both are measured and the cyst is described as
being ‘unilocular’ and not ‘unilocular-solid’. In other cases the lesion is separate from the ovary (e.g.
hydrosalpinx). Again, both ovary and lesion are measured separately. In other cases no normal
ovarian stroma is seen. In these cases the lesion and the ovary are undistinguishable and the
measurement of lesion and ovary will be the same.

Measurements (in mm): The ovary in two perpendicular planes

The lesion in two perpendicular planes
The volume of the tumor is calculated from the three diameters in two perpendicular planes

 The presence of ascites (i.e.fluid outside the pouch of Douglas) is noted (yes/no).
 Fluid in the pouch of Douglas is measured in the sagittal plane (the largest anteroposterior diameter
is given).
(see Figure)

 An incomplete septum (as seen in hydrosalpinges) is defined as a thin strand of tissue running
across the cyst cavity from one internal surface to the contralateral side, butis not complete in some
scanning planes. If a cyst only has incomplete septa, it is unilocular, despite the fact that in certain
sections the cyst appears to be multilocular.
 Solid means echogenicity suggesting the presence of tissue (e.g. the myometrium, the ovarian
stroma, myomas, fibromas). Blood clots and the presence of solid tissue can be distinguished by
looking for internal movement when gently pushing the structure with the transducer. The presence
of blood flow (with the appropriate color Doppler settings) is diagnostic for solid tissue. The
absence of flow is not definitive . In cases of doubt the lesion should be classified as solid.
 Solid papillary projections are defined as any solid projections into the cyst cavity from the cyst
wall greater than or equal to 3 mm in height

Base
Height

If it is unsure whether solid papillary projections or an incomplete septum are present, the ‘worse case
scenario’ is used. E.g. ‘cogwheel excrescences’ and ‘beads-on-a-string’ (as seen in hydrosalpinges)
should be classified as papillary excrescences if their height is greater than or equal to 3 mm. The
‘white ball’ in a dermoid (i.e. Rokitansky node), should not be classified as a solid papillary
projection.
The ‘sludge’ on the internal walls of endometriotic cysts is not regarded as a papillary projection. In
these cases the internal walls are usually ‘irregular’.

 The number of separate papillary projections is noted (1/2/3/more).

 The presence of flow within some of these projections is noted (yes/no).
 Solid papillary projections are described as being ‘smooth’ or ‘irregular’ (e.g. cauliflower-like).
In some cases it is difficult to judge whether it is a papillary projection and from which point to
measure the projection. In these cases it may be helpful to use an imaginary line as shown in the
following schematic drawing:
All lesions are qualitatively classified into one of 5 categories:
1. unilocular (a unilocular cyst without septa and without solid parts or papillary structures).
Normal ovarian stroma is not regarded as ‘solid’ (e.g. a peritoneal cyst, containing a normal
ovary, is unilocular and not unilocular-solid).

incomplete septum; e.g. in

hydrosalpinx)
2. unilocular cyst with solid component (a unilocular cyst with a measurable solid component or at
least one papillary structure). This category may include pyo- or hydrosalpinges with the so-
called ‘beads-on-a-string’ or ‘cogwheel’ appearance if  3 mm. If the solid part contains very
small cysts the mass might be unilocular-solid (see below).

::

3. multilocular (a cyst with at least one septum but no measurable solid components or papillary
projections). The ‘lesion’ is measured as indicated by the arrows.
4. multilocular with solid component (a multilocular cyst with a measurable solid component or at
least one papillary structure)

5. solid (a tumour where the solid components comprise 80% or more of the tumour when
assessed in a two-dimensional section).

(solid tumour with an irregular cyst wall)

A solid tumour may contain papillary projections protruding into the small cysts.
Quantitative assessment of morphology

 In cystic-solid tumours the largest solid component is measured separately (in three perpendicular
planes). The solid component is noted as being smooth or irregular (e.g. cauliflower-like). In some
cases a solid papillary projection is the largest solid component and thus the papillary projection is
recorded both as papillary projection and as solid component.
 The internal wall is also noted as being smooth or irregular.

Smooth Smooth Irregular

If there is a solid papillary projection, then the wall is irregular by definition.

 The external wall of tumors are not examined unless they are solid.
 In cases of solid tumours the description of the internal wall being smooth or irregular is usually
not applicable but the outline of the tumour is described as smooth or irregular.
 If there is any irregularity in either the inner wall of any cyst or in the outer wall of a solid tumour
or on the surface or echogenicity of a solid component, the lesion is described as ‘irregular’.
 The dominant feature of the cystic contents is described as anechoic (black), low-level echogenic
(homogeneous low level echogenic as seen in mucinous tumours), ‘ground glass’ appearance
(homogeneously dispersed echogenic cystic contents, as often seen in endometriotic cysts),
hemorrhagic (with internal thread-like structures, representing fibrin strands; it is possible to
describe the echogenicity as star-shaped, cobweb-like or jelly-like) or mixed echogenic (as often
seen in teratomas) (see images attached).

::::::
:

::::::
:

Anechoic : : :level
Low ::: Ground glass Hemorrhagic Hemorrhagic
:

::::::
:

Mixed Mixed (old blood-fluid level Mixed (e.g. abscess)

or fat-fluid level)

 The presence of acoustic shadows, defined as loss of acoustic echo behind a sound-absorbing
structure, is noted as well. Solid tumours are identified by the appearance of the internal texture,
by the absence of internal movement when moving the transducer or by colour Doppler imaging
(presence of central flow).
 In solid tumours the dominant feature of any cystic contents is described only if it can be assessed.
 ‘Ovarian crescent sign’, defined as the presence of normal ovarian tissue adjacent to an adnexal
tumour. (“absent” or “present”, mandatory new variable for phase 3 and 5)
 Ultrasound evidence of metastases (e.g. “omental cake” or peritoneal tumoural implants). (“absent”
or “present”, mandatory new variable for phase 3 and 5)
Colour Doppler imaging and blood flow indices
Subsequently, the entire tumor is surveyed by CDI. The power, gain and pulse repetition frequency
are initially adjusted for maximum sensitivity of low blood flow states. The lowest velocity signals
are filtered out by gradually increasing the pulse repetition frequency and flow analysis is concentrated
on the highest velocity signals. A subjective semiquantitative assessment of the amount of blood flow
(area and colour scale) within the septa, cyst walls, or solid tumor areas is made: a score of 1 is given
when no blood flow can be found in the lesion; a score of 2 is given when only a small amount of flow
can be detected; 3 is given when moderate flow is present and 4 is given when the adnexal mass
appears highly vascular with marked blood flow using colour Doppler (abundant flow). This colour
score refers only to the colour Doppler image and not to Doppler shift spectrum. It is given for the
tumour as a whole (not for a solid part or a septum only, but for the whole tumour). Multiple
photographic prints are made of relevant structures and Doppler signals.

Quality control

Several informative images or volumes of all adnexal masses should be made. Preferably, these are
stored digitally. Photographs or video are acceptable as well.

Subjective assessment

After ultrasonographic examination of the mass the investigator gives his subjective assessment of the
mass:
A: Malignant or benign or borderline?
B: Probability of malignancy: 1 = benign
(=level of certainty) 2 = probably benign
3 = uncertain
4 = probably malignant
5 = malignant
C: Self impression: presumed histological diagnosis (e.g. dermoid, serous cystadenoma,
endometrioma, abscess…)

Surgical intervention
Surgery is performed according to local protocols. The reason for surgery, e.g. symptoms (pain,
discomfort or pressure symptoms), raised serum CA125 levels or changes in the morphology or
volume of the mass is recorded in the study screen.

Study screen
An astraia study screen will be used which will permit the entry of multiple scans per patient.

All centres will receive the IOTA 5 study screen.

At initial set up centres can choose between the two options:
 1. Full IOTA 5 study (with planning of appropriate follow up and conservative
management whenever feasible).
2. Observational study only (e.g. in radiology departments that are not involved in
management decisions)

Recorded variables (entered in the astraia study screen).

Patient data
(click one option from list below)
 New patient with diagnosis of adnexal mass
 New patient who was already in follow up in your centre for adnexal mass before she was
enrolled to the IOTA 5 study. How many months in follow up? …
 Follow up scan of patient that is already enrolled to the IOTA 5 study before

Ultrasound
 Spontaneous resolution of the adnexal mass (no further details are entered)
 Adnexal mass present (fill in all variables below)

 12 variables described in LR 1:
o Age
o personal history of ovarian cancer
o personal history of breast cancer
o Max diameter of lesion
o Max diameter of solid component
o Presence of ascites
o Presence of blood flow within papillary projection
o Irregular internal cyst walls
o Presence of a purely solid tumour
o Colour score (1/2/3/4)
o Presence of acoustic shadows
o Current hormonal therapy
o Presence of pain during the examination

 as well as simple rules, RMI and other variables:

o Type of tumour (unilocular/unilocular-solid/multilocular /multiloc-solid/solid)
o Ovarian crescent sign
o Cyst content
o Incomplete septum
o Mobility: mobile/reduced mobility/completely fixed
o Number of locules
o Number of papillations (0/1/2/3/more)
o Size of ovary
o Bilateral tumour
o Evidence of metastases
o Menopausal status: premenopausal/postmenopausal
 o Serum CA 125 result (not mandatory)

 Symptoms during the last year before ultrasound scan (multiple options are possible)

o Pelvic pain
o Postmenopausal bleeding
o increased abdominal size
o persistent abdominal distention (bloating)
o appetite loss
o constipation
o diarrhoea
o urinary urgency
o urinary frequency
o weight changes
o dyspareunia
o Other: please specify: ….

 For centres participating at the full IOTA 5 study only: Suggested management recorded by
examiner: “What type of management do you propose for this patient based on ultrasound and
clinical data?”
o Conservative management without follow up
o Conservative management with follow up as specified in the protocol
o Surgery by a gynaecologist or general surgeon
o Surgery by an oncological surgeon

Current status of the patient (this new tab should come before the tab “Histology”). A fixed
query could be made to automatically ask the investigator about the status of all patients that
were not rescanned 55 weeks after their previous scan as soon as the investigator opens the
IOTA 5 study screen.

 Lost to follow up (no other pop-up)

 Patient stopped participating to the study (please specify why : …..) (no other pop-up)
 Patient withdrew her consent (data cannot be used for statistical analysis and no reason is
aked) (no other pop-up)
 Surgery performed (pop-up of fields about operation below)

If Surgery
 Date of operation:

 Type of operation: cystectomy or oophorectomy or staging etc

o Laparotomy with vertical incision
o Laparotomy with horizontal incision
o Operative laparoscopy
o Diagnostic laparoscopy
o Primary chemotherapy
  Indication for operation (more than one possibility may be thicked)
o Suspicion of malignancy based on ultrasound
o Suspicion of malignancy based on other information if so what?
o Malignancy cannot be excluded
o Acute pain
o Chronic pain
o Suspected torsion
o Fertility concerns
o Patient request
o Increase in size of the tumour
o Change in morphology of the tumour
o Increase in CA 125 level
o Indicated by other imaging technique (CT, MRI…)
o Other doctor recommended operation. Please specify the reason: …
o “en passent” removal of the mass when patient was operated for another indication
o Other: Specify: ….

 Findings at operation (more than one option may be ticked)

o No complications of the tumour
o Torsion
o Rupture
o Inflammation/Infection
o Adhesions
o Bleeding from tumour
o Metastatic cancer
o Other complications of tumour: specify: ….
o Other non-gynaecological pathology (e.g. appendicitis): specify: ….

 Complications during operation (within one week of surgery). (more than one option may be
ticked) :
o Conversion from laparoscopy to laparotomy
o Bowel perforation
o Bleeding requiring transfusion
o Embolism, deep venous thrombosis
o Wound Infection
o Peritonitis
o Other: Specify: ….

 Histological diagnosis (pop up list as before) with open text area (“Details”)

If follow up examination:

 Complications of adnexal mass: no /yes

 If yes : date of first complication: ….
  And pop up list (you can select more than one):
o Acute pain
o Chronic pain
o Suspected torsion
o Infection
o Haemorrhage
o Rupture
o Other, please specify: ….

 Death: date
 Death directly or indirectly relatd to adnexal mass?
 cause of death (

 Autopsy findings: specify: ….

Not needed for IOTA phase 5: second stage tests

Consent / information leaflet

Information leaflets are at the discretion of the participating centres.

Approval of the local Ethical Committee for clinical studies is necessary.

Serum tumour markers

Serum CA 125 measurements or other tumour markers are performed locally, using a CA 125 II
immunoradiometric assay

Tissue collection

Preferably the whole tumour should be removed. However, representative biopsies may be sufficient
(e.g. in advanced ovarian cancer or endometrioma).

Tumour classification

Tumours are classified according to the criteria recommended by the International Federation of
Gynecology and Obstetrics (FIGO). In malignant tumours the degree of differentiation is included.

Statistical analysis

For the first objective, a descriptive analysis of complications on all data and stratified for
participating center (anonymized) will be performed, as well as an overall Kaplan-Meier curve of
‘complication free survival’. Complication free survival is defined as the time to the need for surgery
during long-term follow-up. Patients that did not need surgery at the end of the follow-up period are
right censored at the time of the last examination.
For the second objective, the discriminatory ability of LR1 and the polytomous model to detect
malignancy at the initial visit will be assessed. To this end, the logit of the risk of malignancy given by
LR1, logit(LR1), is used to predict malignancy using logistic regression, and the performance assessed
with the odds ratio, the c-index and a ROC curve. Linearity of the effect of logit(LR1) will be assessed
using spline functions. This analysis is planned one year after the end of the inclusion phase of the
trial. Non-operated masses will be classified as benign at the initial visit if there is an absence of a
clinical diagnosis of cancer after 1 year of follow up. In addition, the ability of LR1 to predict
complications during long-term follow-up will be assessed. This will be done using Cox proportional
hazards regression with the logit(LR1) as predictor of complication free survival. Performance will be
assessed using the hazard ratio and the c-index within the context of survival analysis. Linearity of the
effect of logit(LR1) will be investigated using Schoenfeld residuals.

For the third objective, a multivariable survival analysis will be undertaken using Cox proportional
hazards regression or more complex alternatives based on support vector machines (30). The models
will be penalized to prevent overfitting, given that not enough events (i.e., patients that need surgery
during long term follow-up) are expected for the number of available predictor variables. Internal
validation will be assessed through bootstrapping rather than a split of the data in training and test sets
(31).

For the fourth objective, longitudinal normative curves will be derived of the changes seen in the
characteristics of non-operated masses. Longitudinal analysis techniques such as mixed models,
longitudinal support vector machines, and functional linear discriminant analysis (FLDA) (32-34).

Study supervision

Central supervision: the Steering Committee is responsible for the protocol, quality control,
interim analyses of the data and final analysis and reporting of the study.

Local supervision: the Principal Investigators are responsible for the data collection in their
centres.

Dirk Timmerman is responsible for the co-ordination of the overall IOTA project and the
contact between the centres.

Sabine Van Huffel and Bart De Moor are responsible for the data management and the
development of new algorithms, in collaboration with Ben Van Calster, Lieveke Ameye and
Kirsten Van Hoorde.

Publication policy

The steering committee is responsible for publication of the data in scientific journal s. As
such the members are co-authors in all resulting clinically relevant papers, to which they
made significant contributions. By the time of the final analysis the principal investigators
have to have contributed at least 50 cases to the study. They are co-authors, according to the
number of patients they contributed to the study (depending on the journal’s restriction of the
number of co-authors) on condition that they contribute to writing the papers and read and
approve the final version.

Purely mathematical papers without clinical relevance related to the study data are published
by S. Van Huffel, B. De Moor and co-workers at ESAT with reference to the IOTA group and
the inclusion of as many as possible of the clinical contributors.

The Katholieke Universiteit Leuven represented by its department K.U.LEUVEN RESEARCH &
DEVELOPMENT, having its office in 3000 Leuven, Minderbroedersstraat 8A – box 5105, Belgium,
VAT number BE 419.052.173 holds intellectual property rights that might result from the IOTA
project.
References

1. Timmerman D, Testa AC, Bourne T, Ferrazzi E, Ameye L, Konstantinovic ML, Van Calster B, Collins
WP, Vergote I, Van Huffel S, Valentin L. Logistic regression model to distinguish between the benign and
malignant adnexal mass before surgery: a multicenter study by the International Ovarian Tumor Analysis
Group. J Clin Oncol 2005; 23:8794-8801.
2. Van Holsbeke C, Van Calster B, Testa AC, Domali E, Lu C, Van Huffel S, Valentin L, Timmerman D.
Prospective internal validation of mathematical models to predict malignancy in adnexal masses: results
from the Internation Ovarian Tumor Analysis study. Clin Cancer Res 2009; 15:684-691.
3. Timmerman D, Van Calster B, Testa AC, Guerriero S, Fischerova D, Lissoni AA, Van Holsbeke C, Fruscio
R, Czekierdowski A, Jurkovic D, Savelli L, Vergote I, Bourne T, Van Huffel S, Valentin L. Ovarian cancer
prediction in adnexal masses using ultrasound based logistic regression models: a temporal and external
validation study by the IOTA group. Ultrasound Obstet Gynecol 2010; 36(2):226-34.
4. Menon U, Gentry-Maharaj A, Hallett R, Ryan A, Burnell M, Sharma A, Lewis S, Davies S, Philpott S,
Lopes A, Godfrey K, Oram D, Herod J, Williamson K, Seif MW, Scott I, Mould T, Woolas R, Murdoch
J, Dobbs S, Amso NN, Leeson S, Cruickshank D, McGuire A, Campbell S, Fallowfield L, Singh N,
Dawnay A, Skates SJ, Parmar M, Jacobs I. Sensitivity and specificity of multimodal and ultrasound
screening for ovarian cancer, and stage distribution of detected cancers: results of the prevalence screen
of the UK Collaborative Trial of Ovarian Cancer Screening(UKCTOCS). Lancet Oncol. 2009
Apr;10(4):327-40.
5. Hoo WL, Yazbek J, Holland T, Mavrelos D, Tong EN, Jurkovic D. Expectant management of
ultrasonically diagnosed ovarian dermoid cysts: is it possible to predict outcome? Ultrasound Obstet
Gynecol. 2010 Aug;36(2):235-40.
6. Caspi B, Appelman Z, Rabinerson D, Zalel Y, Tulandi T, Shoham Z. The growth pattern of ovarian
dermoid cysts: a prospective study in premenopausal and postmenopausal women. Fertil Steril. 1997
Sep;68(3):501-5. PubMed PMID: 9314922.
7. Caspi B, Levi R, Appelman Z, Rabinerson D, Goldman G, Hagay Z. Conservative management of
ovarian cystic teratoma during pregnancy and labor. Am J Obstet Gynecol. 2000 Mar;182(3):503-5.
PubMed PMID: 10739498.
8. Valentin L, Akrawi D. The natural history of adnexal cysts incidentally detected at transvaginal
ultrasound examination in postmenopausal women. Ultrasound Obstet Gynecol. 2002 Aug;20(2):174-
80.
9. Valentin L, Skoog L, Epstein E. Frequency and type of adnexal lesions in autopsy material from
postmenopausal women: ultrasound study with histological correlation. Ultrasound Obstet Gynecol.
2003 Sep;22(3):284-9.
10. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics 2009. CA Cancer J Clin 2009; 59:225-
249.
11. Cancer Research UK a. UK ovarian cancer incidence statistics. http://info.cancerresearchuk.
org/cancerstats/types/ovary/incidence/ (accessed Oct 7, 2009).
12. Horner MJ, Ries LAG, Krapcho M, Neyman N, Aminou R, Howlader N, Altekruse SF, Feuer EJ, Huang L,
Mariotto A, Miller BA, Lewis DR, Eisner MP, Stinchcomb DG, Edwards BK (eds.) SEER Cancer statistics
review 1975-2006, NCI, Bethesda, MD.
13. Cancer Research UK b. UK Ovarian cancer survival statistics. http://info.cancerresearchuk.
org/cancerstats/types/ovary/survival/ (accessed Oct 7, 2009).
14. Engel J, Eckel R, Schubert-Fritschle G, et al. Moderate progress for ovarian cancer in the last 20 years:
prolongation of survival, but no improvement in cure rate. Eur J Cancer 2002; 38:2435-45.
15. Vergote I, Amant F, Ameye L, Timmerman D. Screening for ovarian carcinoma: not quite there yet
(editorial). Lancet Oncology 2009; 10:308-309.
16. Crayford TJB, Campbell S, Bourne TH, Rawson HJ, Collings WP. Benign ovarian cysts and ovarian
cancer: a cohort study with implications for screening. Lancet 2000; 355:1060-1063.
17. Timmerman D. The use of mathematical models to evaluate pelvic masses: can they beat an expert
operator? Best Pract Res Clin Obstet Gynaecol 2004; 18:91-104.
18. Van Holsbeke C, Van Calster B, Valentin L, Testa AC, Ferrazzi E, Dimou I, Lu C, Moerman Ph, Van
Huffel S, Vergote I, Timmerman D. External validation of mathematical models to distinguish between
benign and malignant adnexal tumors: a multicenter study by the International Ovarian Tumor Analysis
Group, Clin Cancer Res 2007; 13:4440-4447.
19. Geomini P, Kruitwagen R, Bremer GL, Cnossen J, Mol BW. The accuracy of risk scores in prediting
ovarian malignancy: a systematic review. Obstet Gynecol 2009; 113:384-394.
20. Valentin L, Jurkovic D, Van Calster B, Testa A, Van Holsbeke C, Bourne T, Vergote I, Van Huffel S,
Timmerman D. Adding a single CA 125 measurement to ultrasound imaging performed by an
experienced examiner does not improve preoperative discrimination between benign and malignant
adnexal masses. Ultrasound Obstet Gynecol. 2009 Sep;34(3):345-54. PubMed PMID: 19585547.
21. Van Calster B, Timmerman D, Bourne T, Testa AC, Van Holsbeke C, Domali E, Jurkovic D, Neven P,
Van Huffel S, Valentin L. Discrimination between benign and malignant adnexal masses by specialist
ultrasound examination versus serum CA-125. J Natl Cancer Inst. 2007 Nov 21;99(22):1706-14. Epub
2007 Nov 13. PubMed PMID: 18000221.
22. Valentin L. Pattern recognition of pelvic masses by gray-scale ultrasound imaging: the contribution of
Doppler ultrasound. Ultrasound Obstet Gynecol. 1999 Nov;14(5):338-47.
23. Sokalska A, Timmerman D, Testa AC, Van Holsbeke C, Lissoni AA, Leone FP, Jurkovic D, Valentin
L. Diagnostic accuracy of transvaginal ultrasound examination for assigning a specific diagnosis to
adnexal masses. Ultrasound Obstet Gynecol. 2009 Oct;34(4):462-70.
24. Timmerman D, Valentin L, Bourne TH, Collins WP, Verrelst H, Vergote I. Terms, definitions and
measurements to describe the sonographic features of adnexal tumors: a consensus opinion from the
International Ovarian Tumor Analysis (IOTA) group. Ultrasound Obstet Gynecol 2000; 16:500-505.
25. Van Calster B, Timmerman D, Lu C, Suykens JAK, Valentin L, Van Holsbeke C, Amant F, Vergote I, Van
Huffel S. Preoperative diagnosis of ovarian tumors using Bayesian kernel-based methods. Ultrasound
Obstet Gynecol 2007; 29:496-504.
26. Van Calster B, Timmerman D, Nabney IT, Valentin L, Testa AC, Van Holsbeke C, Vergote I, Van Huffel
S. Using Bayesian neural networks with ARD input selection to detect malignant ovarian masses prior t o
surgery. Neural Comput Applic 2008; 17:489-500.
27. Ameye L, Valentin L, Testa AC, Van Holsbeke C, Domali E, Van Huffel S, Vergote I, Bourne T,
Timmerman D. A scoring system to differentiate malignant from benign masses in specific ultrasound -
based subgroups of adnexal tumors. Ultrasound Obstet Gynecol 2009; 33:92-101.
28. Van Holsbeke C, Van Calster B, Bourne T, Melis GB, Testa AC, Guerriero S, Fruscio R, Lissoni AA,
Czekierdowski A, Savelli L, Van Huffel S, Valentin L, Timmerman D. External validation of diagno stic
models to predict the risk of malignancy in adnexal masses. Submitted.
29. Leone FP, Crepaldi A, Marciante C, Cetin I. Sonographic follow-up of unilocular >5 cm and multilocular
ovarian cysts <7 cm in post-menopausal women: preliminary results. OC21.05. Ultrasound Obstet Gynecol
2011; 38 (Suppl. 1): 40.
30. Van Belle V, Pelckmans K, Suykens JAK, Van Huffel S. Additive survival least squares support vector
machines. Stat Med 2010; 29:296-308.
31. Steyerberg EW, Harrell FE Jr, Borsboom GJ, Eijkemans MJ, Vergouwe Y, Habbema JDF. J Clin
Epidemiol 2001; 54:774-781.
32. Verbeke G, Molenberghs G. Linear mixed models for longitudinal data. Springer: New York, 2000.
33. Luts J, Molenberghs G, Verbeke G, Van Huffel S, Suykens JAK. A mixed effects least squares support
vector machine model for classification of longitudinal data. Internal Report 10 -249, ESAT-SISTA,
K.U.Leuven (Leuven, Belgium), 2010.
34. James GM, Hastie TJ. Functional linear discriminant analysis for irregularly sampled curves. J R Stat Soc B
2001;63:533–550.