Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Gina Brown Pent Mon

Download as pdf or txt
Download as pdf or txt
You are on page 1of 69

Advances in diagnosis and

imaging technology: MRI


Gina Brown
Department of Radiology
Royal Marsden Hospital
Imperial College, London

gina.brown@rmh.nhs.uk
Advances in Imaging
Technology
Rectal Cancer
Professor Gina Brown
Royal Marsden Hospital
Imperial College
UK gina.brown@rmh.nhs.uk
The Royal Marsden

The promised future of


imaging….
• Perfusion and tumour • Tumour microenvironment
permeability • Biology of tumour
• Diffusion and tumour • Angiogenesis in tumours
cellularity
• Tumour cellularity and
• PET-CT and tumour
proliferation
metabolism
• Likely response to treatment?
• Textural analysis
• Biomarker for treatments?
• Hypoxia imaging
The Royal Marsden

Quantitative Biomarkers in
Imaging
• Predicting prognosis – stage assessment
• Predicting response
• Assessing response
• Independent prognostic / predictive
imaging biomarkers to determine
treatment
The Royal Marsden

Quantitative DCE T1-weighted MRI


Ktrans
Ve
Tumour

conc
SI/ Gd
Muscle

Quantitative estimates Time

• Ktrans (efflux constant)


( )
Kinetic Modelling
( K trans / EES )( t t 0 )
2
ai e e mi ( t t0 )
C (t ) = D.K trans
• Ve (interstial space) m (K i =1 i trans / EES )

• T1AUC Tofts and Kermode


Prognosis?
The Royal Marsden

Permeability
Perfusion
Diffusion

Pre-treatment tumour
Permeability, Perfusion and
Diffusion predict response to George M, Dzik-Jurasz ASK et al, BJS 2001
Dzik-Jurasz ASK et al, Lancet 2002
chemoradiotherapy
“the mean is significantly higher
The Royal Marsden

for responders….”
What does data such as
this this mean for an
individual patient?
How can this be
translated to clinical
practice?
The Royal Marsden
The Royal Marsden

Conclusions ROC analysis for K trans post-


treatment, revealed that K trans has
drawn for post an AUC of 0.7941 (0.5764, 1.0119)
in predicting pCR.
hoc data
analysis A K trans of 0.3 emerged as the best
cut-off for distinguishing pCR from
non-pCR.

Was this validated in a prospective


dataset? – we’re still waiting for the
prospective data breakthrough….
The Royal Marsden

Another post hoc analysis


The Royal Marsden

Using ROC analysis to derive


a cut-off….

“We used the ad-hoc analysis because when starting the study no earlier
data was available to set the cutoff value.”
Before clinical implementation of the DCE-MR for response assessment
a larger prospective multicenter study with a predefined cutoff value will
need to be performed…..
The Royal Marsden

Examples from PET-CT


publications
The Royal Marsden

PET SUV validation did not work as well

• ROC curve cut-off of


48% for pathological
responders vs
nonresponders

Janssen et al
Int J Radio. Biol. Phys
2011
The Royal Marsden

FDG-PET assessment of SUV max at 2 weeks

Metabolic response

Post Radiotherapy
Acute inflammatory
changes
The Royal Marsden

DIFFUSION WEIGHTED MRI IN


ASSESSING RECTAL CANCER
The Royal Marsden
Assessing response: the role
The Royal Marsden

of new technologies?
Method Prospectively validated against DFS
outcomes
MRI DWI No – many retrospective quantitative cut-offs and
qualitative assessments – none prospectively
validated

DCE-MRI No – many retrospective values proposed –


none validated
PET-CT No – but retrospective SUV cut-offs proposed –
unverified prospectively
The Royal Marsden
The Royal Marsden

Checklist of Biomarker
Recommendations
1. Has the proposal laid out proof of a
clinical need i.e. proof of
outcomes/consequence.
2. What is the true degree of
error/deficiency and consequence in
current non-quantitative imaging or non
imaging test ?
The Royal Marsden

Checklist of Biomarker
Recommendations
3. Does the biomarker being tested have a
predefined threshold/criteria or range for
positive vs negative test or is it exploratory –
if exploratory (ie shown a correlation)
4. Have the authors then completed the study
by defining and then validating thresholds in
an independent dataset
The Royal Marsden

Checklist of Biomarker
Recommendations
5. Has the biomarker been tested
against other clinical variables and
other imaging methods of
assessment and found to be an
independent predictor compared with
the current best standards of care?
The Royal Marsden

Checklist of Biomarker
Recommendations
6. Has the imaging measure been
tested against PFS/DFS/survival
outcomes and shown to be
independent of existing clinical and
imaging non quantitative tests?
The Royal Marsden

Developments in MRI based


management of Rectal Cancer
• The unimportance of nodal status
• The importance of extramural depth of spread as a biomarker
• Recognition and effective preoperative treatment of mrCRM involvement
• Recognition and effective preoperative treatment of mrEMVI – impact on
survival
• Staging and assessment of low rectal cancer
• Staging and assessment of Early Rectal Cancer
• Using the post treatment MRI TRG assessment as a biomarker for further
preoperative treatment stratification
The Royal Marsden

Are lymph nodes still a


biomarker for local recurrence?

The Dukes era 1932 - 1990s


The TME era 2000 - present
The Royal Marsden

Randomised trial evidence unimportance of


nodal status on local recurrence
• For a good quality total
mesorectal excision that is
circumferential resection
margin negative – there is no
difference – CR07 5-6% LR
(Quirke et al Lancet
Oncology) rates irrespective
of node status, Lancet 2009
• OCUM trial (Germany) and LR and stage III disease, is linked to poor quality surgery
Quicksilver trial (Canada) Where surgical quality is good and CRM is clear, lymph
nodes are not associated with LR
Lymph Nodes are not the cause of CRM involvement
Limitations of the TNM – T3 category forms
80% of rectal cancers
• Jass (St Marks, UK) : – independent
prognostic significance
• Harrison (Tennessee, USA): prognostic score
depth of spread in mm
• Cawthorne (Guildford, UK): depth of spread
significance
• Merkel and Hermanek (Erlangen, Germany) :
● T3 subclassification T3<5mm
• T3a <1mm
• T3b>1-5mm,
• T3c>5-15mm
Erlangen: >800 patients
T3>5mm
• T3d>15mm (TNM staging system 1993 pT3<5mm N any same survival as
supplement) T2 85%
pT3>5mm N any, 54% survival
“measuring extramural depth is the least subjective and
most reliable of all the observations by radiologists”

295/311 (95 %) patients who underwent primary surgery.


The mean difference between MRI and histopathology assessment of
tumor EMD was -0.046 mm, SD = 3.85 mm, the 95 % CI was -0.487 to
0.395 mm.
MRI and histopathology assessment of tumor spread are considered
equivalent to within 0.5 mm (qR). Radiology 2007
The Royal Marsden

MERCURY trial
• 2002-2003
• 11 international centres (30 radiologists)
• 295 patients undergoing primary surgery
• Policy to avoid pre and post operative radiotherapy
for mrCRM clear, mrEMVI negative, T3b or less
rectal cancers, regardless of N stage
The Royal Marsden

Outcomes for MRI good prognosis rectal cancers: regardless of N stage

Taylor et al, MERCURY


Annals of Surgery 2011
The Royal Marsden

mrCRM involvement defined as tumour spread (continuous or discontinuous) within


1mm of the TME plane (bounded by mesorectal fascia and intersphincteric plane).
Independent risk factor for local recurrence (Hazard Ratio 3.5)

mrCRM clear local recurrence 7% versus 20% if mrCRM involved


pCRM clear local recurrence 6.5% versus 26% if pCRM involved
The Royal Marsden

Measuring size of nodes worsens results – overstaging


and overtreatment of low risk patients
• node positive if either irregular border or
mixed signal intensity.
• Metastases demonstrated in 51/56 nodes
(91%, 95% CI 81% to 96%) with either an
irregular border or a mixed intensity signal.
• only 9/225 nodes (4%, CI 2.1% to 7.4%) with
smooth borders and a uniform signal
contained metastases irrespective of size.
• Size of node bears no relationship to
malignant risk
Brown et al Radiology 2003
The Royal Marsden

MRI detected Lymph Nodes close to the mesorectal fascia are


not associated with pCRM involvement (Shihab et al, BJS 2010)

• Involvement of CRM by lymph node


metastases alone is uncommon (1.3% of all
patients in MERCURY series).
• Caution when recommending neoadjuvant
therapy based solely on an MRI-detected
lymph node close to the mesorectal fascia.
The Royal Marsden

MR CRM prediction for low rectal cancers:


TME plane safety
• 1. MRI Low Rectal Stage 1: tumour on MRI images
appears confined to bowel wall (intact muscularis
propria of the internal sphincter).
• 2. MRI Low Rectal Stage 2: tumour on MRI partially
replaces the muscle coat but there is >1mm distance to
TME/intersphincteric plane. Above sphincter it is
confined to the mesorectum.
• 3. MRI Low Rectal Stage 3: invading into the
intersphincteric plane or lying within 1mm of levator
muscle above the level of the sphincter complex.
• 4. MRI Low Rectal Stage 4: invading the external anal
sphincter and infiltrating/ extending beyond the levators
+/- invading adjacent organ.

“Shihab et al: MRI staging of low rectal cancer." Eur Radiol 19(3): 643-650.
The Royal Marsden

Primary Surgery for Low


Rectal Cancers
• Almost half (44·4%, 124/279) of study participants had a ‘safe’
mrLRP and no adverse MRI features. The recommended
management was to proceed straight to surgery with an
intersphincteric resection, adhering to this guidance (50%) led to
a clear 16 pCRM in 98% of cases.
• When MRI low-risk patients were offered CRT or an ELAPE -
this resulted in a numerically higher pCRM involvement.
Additional treatment and more radical surgery did not result in a
benefit to the patient and may represent overtreatment.

Battersby, N. J., et al Prospective Validation of a Low Rectal Cancer Magnetic Resonance


Imaging Staging System and Development of a Local Recurrence Risk Stratification Model: The
MERCURY II Study. Ann Surg. 2015
The Royal Marsden

Results from 19 sites recruiting to MERCURY


MRI Tool for predicting risk of pCRM
involvement 5%
No Risk Factors
mr ‘Unsafe’ plane 2% pCRM risk
12% 15%

MRI 31

Height <4cm 26 53% 31 mrEMVI


5%
4% 25
9% 12%
MRI invading edge
Anterior
4%
TME Mesorectal plane
For coloanal anastomosis/ intersphincteric
>1mm of intersphincteric plane clear
Beyond TME ELAPE plane
<1mm intersphincteric plane clear
Beyond TME exenterative planes
The Royal Marsden

When is a node not a node?


The Royal Marsden

• Poor interobserver agreement for


EMVI
• Large variations in reporting rates
10% -50% - underreporting
widespread
• Lack of agreement of definitions
The Royal Marsden

UK Royal College of
Pathologists Guidelines 2014
• pEMVI detection rates should be
>30%
• For units with <30%, should use
elastin staining
• MRI is reliable as method of detecting
EMVI and can be used to audit
pathology
• Lone arteriole sign, to improve
detection of discontinuous vascular
deposits
The Royal Marsden

Characteristic features of
mrEMVI• Expansion of
extramural vessels
by tumour
• Serpiginous /
tubular extension
of tumour signal
MRI for detection of extramural vascular invasion
in rectal cancer.
AJR Am J Roentgenol 191(5): 1517-1522.
mrEMVI is associated with pelvic
The Royal Marsden

sidewall tumour deposits


Smith et al. “Prognostic significance of MRI-detected Extramural Vascular Invasion."
The Royal Marsden

BJS. 2008

MRI-EMVI score & Outcome


n=135. Median follow-up=3·12 (0·9-5·7) years.
MRI-EMVI score= 0-2
100 MRI-EMVI score= 3-4
80
% Relapse-

71%
60
40 p = 0·0015
free

32%
20
0
0 1 2 3 4 5 6
Time since operation (Years)
The Royal Marsden

MRI detected more persistent EMVI post CRT than pathology

Chand M, Evans J, Swift RI, et al. Prognostic Significance of


Postchemoradiotherapy High-Resolution MRI and Histopathology
Detected Extramural Venous Invasion in Rectal Cancer. Ann Surg. 2014.
The Royal Marsden
The Royal Marsden

A good prognosis tumour?


Looks like a T1sm3
The Royal Marsden
Discontinuous EMVI in ERC
And a pelvic sidewall tumour deposit
mrEMVI
• mrEMVI seen in 40% of
rectal cancers
• Detected more readily
than by pathology
• Independent risk factor
for CRM involvement,
local and distant
recurrence
The Royal Marsden

Lymph nodes versus extranodal deposits


These tumours have entirely
different prognostic outcomes
Stage II (T3N0) Stage III (T3N1) Stage I (T1N0)

mrT3dN0EMVI pos mrT3aN1CRM-ve mrT1 EMVI deposit,


CRM+:CRT+chemo + CRM+ve, Preoperative CRT
beyond TME surgery
Primary TME surgery and ELAPE
The Royal Marsden

Assessing response
Method Prospectively validated against DFS
outcomes
MRI DWI No – many retrospective quantitative cut-offs and
qualitative assessments – none prospectively
validated

DCE-MRI No – many retrospective values proposed –


none validated
PET-CT No – but retrospective SUV cut-offs proposed –
unverified prospectively
mrVolume assessment Yes: >80% volume reduction
mrTRG Yes : TRG1-5 validated prospectively and
against outcomes
mrT and mrN stage validated prospectively and against outcomes
The Royal Marsden

Timing after CRT? When is


maximum response reached?

Baseline 6 weeks 12 weeks


mrT4 ymrT3b ymrT2
Final Pathology: ypT2N0
The Royal Marsden

MR TRG
The Royal Marsden TRG and Survival
(Patel et al JCO 2011)
MRI TRG 1-3
72% at 5 yrs

MRI TRG 4-5


27% at 5 yrs

p=0.001
HR 3.28 (95%CI; 1.22–8.80).
The Royal Marsden

Royal Marsden n=208


patients

Yu et al , ESMO World GI Congress, Barcelona 2015


The Royal Marsden

EXPERT C trial for patients at high risk


of local and distant failure
– Tumors within 1 mm of mesorectal fascia
(ie, potential circumferential resection
margin involvement)
– T3 c (extramural spread 5-15 mm) and
T3 d (extramural spread >15 mm),
regardless of N stage
– MRI T4a or T4b disease regardless of N
stage
– Low rectal cancer with tumor bordering
the intersphincteric/ distal TME plane on
MRI
– Tumors with MRI extramural venous
invasion (mrEMVI)
Overall Survival by TRG (1-2 v 3 v 4-5)
The Royal Marsden

after Chemo-Radiotherapy
in EXPERT-C trial (both arms)
mrTRG1-2
89.8%

67.5%
65.9%

mrTRG1-2, 39.8%
mrTRG3 , 29%
mrTRG4-5, 31%
The Royal Marsden

mrTRG is a prognostic
(and predictive) biomarker
– Shows good interobserver radiology agreement and
reproducibility
– MERCURY trial (JCO 2011 – multiple radiologists)
– EXPERT-C trial
– GEMCAD study (17 radiologists)
– CORE study (interobserver agreement)
– Identified 40% of patients with mrTRG1/2 – 89.8% overall
survival. Compared with only 8.8% patients with pathologic CR.
– Therefore mrTRG could be justified as a more clinically relevant
endpoint
The Royal Marsden

mrTRG is a prognostic
(and predictive) biomarker
• Shows good interobserver radiology agreement and reproducibility
● MERCURY trial (JCO 2011 – multiple radiologists)
● EXPERT-C trial
● GEMCAD study (17 radiologists)
● CORE study (interobserver agreement)
● MERCURY 2 trial – risk factor for CRM involvement
• In EXPERT C trial identified 40% of patients with mrTRG1/2 – 89.8% overall
survival. Compared with only 15% pathologic CR rate (90% survival).
• Therefore mrTRG could be justified as a more clinically relevant endpoint
TRIGGER :
randomised
phase III trial
patients will be
randomised to
an mrTRG based
treatment
strategy
Training of
Radiologists to
undertake
mrTRG
The Royal Marsden

MRI Trials and the Colorectal Patient Pathway


The Royal Marsden
The Royal Marsden

www.slideshare.net/ginabrown3
• MRI reporting templates
• MRI high resolution technique
• How to identify mrEMVI
• Details of workshops for surgeons
and radiologists
Reporting Minimum Standards
Baseline assessment of Rectal cancer MRI report
Primary tumour Lymph node assessment
The primary tumour is demonstrated as an [ Annular | Semi-annular | Ulcerating | | Polypoidal | Only benign reactive and no suspicious nodes shown [N0]
Mucinous] mass with a [nodular / smooth] infiltrating border. [ ] mixed signal/irregular border nodes [N1/N2]
Extramural venous invasion: [ No evidence ] [ Evidence]
The distal edge of the luminal tumour arises at a height of [ ] mm from anal verge: [ ] Small [ ]Medium [ ]Large vein invasion is present
The distal edge of the tumour lies [ ]mm [Above,at, below] the top of the puborectalis sling CRM
The tumour extends craniocaudally over a distance of [ ] mm The closest circumferential resection margin is at o’clock
The proximal edge of tumour lies [above at below] the peritoneal reflection The closest CRM is from [Direct spread of tumour] [Extramural venous invasion] [Tumour
deposit]
Invading edge of tumour extends from [ to ] O’clock
Minimum tumour distance to mesorectal fascia: mm [CRM clear ] [CRM involved]
Tumour is [confined to] [extends through] the muscularis propria:
Peritoneal deposits: [ No evidence] [ Evidence]
Extramural spread is [ ] mm Pelvic side wall lymph nodes:
mrT stage: [T1 ] [ T2 ] [ T3a] [ T3b ] [ T3c] [ T3d ] [T4visceral ] [T4 [ None] [ Benign] [ Malignant mixed signal/irreg border]
peritoneal] Location: [Obturator fossa • R •L ] . [External Iliac Nodes • R •L] .[ Internal iliac • R •L ]

Tumour is [present] [not present] the level of the puborectalis sling at this level: Summary: MRI Overall stage: T N M [CRM clear] , [ CRM involved ] , [ EMVI
[Tumour is confined to the submucosal layer/part thickness of muscularis propria indicating that the positive] [EMVI negative],[PSW positive ] [PSW negative]
intersphincteric plane/mesorectal plane is safe and intersphincteric APE or ultra low TME is No adverse features eligible for primary surgery
possible] High risk safe margins for preoperative therapy : eligible for Serenade, Marvel
[Tumour extends through the full thickness of the muscularis propria : intersphincteric Poor prognosis unsafe margins eligible for preoperative chemoradiotherapy: eligible for 6 vs 12
plane/mesorectal plane is unsafe, Extralevator APE. is indicated for radial clearance] trial
[Tumour extends into the intersphincteric plane : intersphincteric plane/mesorectal plane is unsafe, Low Rectal <6cm – eligible for the Low Rectal Study.
therefore an extralevator APE. is indicated for radial clearance]
[Tumour extends into the external sphincter : intersphincteric plane/mesorectal plane is unsafe.]
[ Tumour extends into adjacent [prostate/vagina/bladder/sacrum] : exenterative procedure will be
required

Additional comments:

.
Reporting Template Post Treatment
Post Treatment Assessment MRI Rectal Cancer Lymph nodes:
Comparison is made with the previous examination of: • None /Only benign reactive [N0]
• The treated tumour: shows no fibrosis,TRG5 • Present number mixed signal/irregular border [N1/N2]
• Less than <25% fibrosis, predominant tumour signal, TRG4
• 50% tumour/fibrosis, TRG 3 Extramural venous invasion: [• No evidence • Evidence]
•>75% fibrosis, minimal tumour signal intensity,TRG2 [• Small • Medium • Large]
•low signal fibrosis only no intermediate tumour signal TRG1 CRM
Closest circumferential resection margin: [ ]O’clock
The distal edge of the luminal tumour arises at a height of [ ] mm from anal verge: Closest CRM is from [ Direct spread of tumour • Extramural venous invasion • Tumour deposit]
The distal edge of the tumour lies [ ]mm [Above, at, below] the top of the puborectalis sling Minimum tumour distance to mesorectal fascia: [ ]mm [ • CRM clear • CRM involved]
compared with []mm previously
The tumour extends craniocaudally over a distance of [ ] mm compared with [ ]mm previously Peritoneal deposits: [• No evidence • Evidence ]
The proximal edge of tumour lies [above at below] the peritoneal reflection
The invading edge of treated tumour extends from [ to ] O’clock Pelvic side wall lymph nodes: • None • Benign • Malignant
Tumour signal is [Confined to / Extends through the muscularis propria.] [Location: Obturator fossa • R •L . External Iliac Nodes •R •L. Inf Hypogastric •R •L ]
Fibrotic signal is [ Confined to / Extends through muscularis propria.]
Extramural spread: [ ]mm for tumour signal [ ]for fibrotic stroma Summary: y MRI Overall stage ymrT ymr N M , TRG
• Low/intermediate risk, CRM clear, TRG 1-2, EMVI negative
yMR T stage: • T1 • T2 • T3a • T3b • T3c • T3d •T4 visceral •T4 peritoneal • High prognosis, CRM pos or TRG4/5 or EMVI positive

Treated tumour [is/ is not] present at or below the puborectalis sling TRG1-2 low tumour – eligible for consideration for deferral of surgery
• tumour signal/fibrosis extends into the submucosal layer/part thickness of muscularis propria :
intersphincteric plane/mesorectal plane is safe intersphincteric APE or ultra low TME possible, CRM
is safe
• tumour signal/fibrosis extends through the full thickness of muscularis propria : intersphincteric
plane/mesorectal plane is unsafe, for extralevator APE.
• tumour signal/fibrosis extends into external sphincter : intersphincteric plane/mesorectal plane is
unsafe:for extralevator APE
•tumour signal/fibrosis extends into beyond external sphincter into [prostate/vagina ] : intersphincteric
plane / mesorectal plane is unsafe, for extralevator APE.

You might also like