Folie 1

1
Abstract 4091
Resectability and agreement between surgeons
Review of CT- and MRI- scans of the CELIM study
(Multicenter randomized trial of
cetuximab/FOLFOX versus cetuximab/FOLFIRI in
unresectable liver metastases).
Wolf Bechstein,1 Hauke Lang,2 Claus-Henning
Köhne,3 Fabio Parisi,4 Hans-Rudolf Raab,3 Andrea
Frilling,5 Ralf Konopke,6 Jürgen Weitz,7
Christian Stroszczynski,6 Gunnar Folprecht6
1University Hospital Frankfurt, Germany,
2University Hospital Mainz, Germany, 3 Klinikum
Oldenburg, Oldenburg, Germany, 4New York
University, N.Y., U.S., 5University Hospital
Essen, Germany, 6University Hospital Carl Gustav
Carus, Dresden, Germany, 7University of
Heidelberg, Dpt. of Surgery, Heidelberg, Germany
2
Background

• Resection of liver metastases provides favorable
  long-term survival (Adam Ann Surg 2004)
• Resectability of colorectal liver metastases
  depends on technical resectability and prognostic
  factors
• Number of liver metastases is an important
  prognostic factor and pts with gt 4 liver mets
  were excluded from neoadjuvant trial for
  resectable liver metastases (Nordlinger, Lancet
  2007)
• In primary non-resectable liver metastases,
  resection rate correlates with response to
  chemotherapy
• Few data are available on achievement of
  resectability due to chemotherapy and on the
  agreement between surgeons regarding resectabilty
• Cetuximab increases response rates when added to
  FOLFIRI or FOLFOX (Van Cutsem NEJM 2009,
  Bokemeyer JCO 2009)
• The CELIM study compared tumor response and
  resectability rates in patients with unresected
  liver metastases receiving neoadjuvant treatment
  with cetuximab plus FOLFIRI or FOLFOX6

3
Patient selection

• Patients with non-resectable colorectal liver
  metastases
• Definition of non-resectability
• 5 liver metastases and/or
• liver metastases that are technically
  non-resectable defined by local surgeon in
  cooperation with local radiologist (amount of
  functional liver tissue remaining, infiltration
  of non-resectable structures)
• Expected resectability after response to
  chemotherapy was not an inclusion criterion
• No extrahepatic disease
• Karnofsky PS 80 and adequate hepatic, renal,
  and bone marrow function
• Metastases histologically confirmed
• Patients with simultaneous liver metastases were
  eligible if the primary tumor was resected 1
  month prior to chemotherapy
• Informed consent no prior chemotherapy (except
  adjuvant chemotherapy 6 months ago) no
  concurrent immunotherapy, chemotherapy or hormone
  therapy no previous malignancy other than
  colorectal cancer, basal cell carcinoma, or
  pre-invasive carcinoma of the cervix no
  inflammatory bowel disease no relevant coronary
  heart disease

4
Patients with non-resectable colorectal liver
metastases(technically non-resectable / 5
liver metastases)without extrahepatic metastases
closed early

• Stratification
• technically non-resectable / 5 liver
  metastases
• staging with PET
• EGFR IHC
• FOLFOX6 oxaliplatin 100 mg/m², 5-FU 4002400
  mg/m², FA 400 mg/m²
• FOLFIRI irinotecan 180 mg/m², 5-FU 4002400
  mg/m², FA 400 mg/m²
• Cetuximab 400 mg/m², then 250 mg/m² weekly

5-FU, 5-fluorouracil FA, folinic acid EGFR,
epidermal growth factor receptor IHC,
immunohistochemistry PET, positron emission
tomography.
5
Patients with non-resectable colorectal liver
metastases(technically non-resectable / 5
liver metastases)without extrahepatic metastases
Biopsy EGFR screening
Randomization
FOLFOX6 cetuximab FOLFIRI cetuximab
Primary endpoint Response
EGFR, epidermal growth factor receptor.
6
Patient characteristics
    FOLFOX6 FOLFIRI All
    cetuximab cetuximab patients
n56 n55 n111 
       
Median age (y.) 65.1 62.0 63.3
Sex      
male 64 64 64
KRAS status (n99)      
wild-type   70 71 71
Primary tumor site      
rectal cancer 38 51 44
Adjuvant chemotherapy      
yes 11 22 16
7
Patient characteristics
  FOLFOX6 FOLFIRI All
  cetuximab cetuximab patients
n56 n55  n111
Number liver metastases      
lt5   23 31 27
5-10   55 49 52
gt10   20 15 17
NA   2 5 4
Prior liver resection      
yes 16 9 13
NA, not available
8
Efficacy confirmed response
  FOLFOX6 FOLFIRI All
  cetuximab cetuximab patients
n53 n53 n106
CR/PR 68 57 62
95 CI 54-80 42-70 52-72
SD 28 30 29
PD 4 13 8
Responses confirmed by 2nd CT scan according to
RECIST or by resection
Chi square test for comparison between
FOLFOX6cetuximab vs FOLFIRIcetuximab p would
be 0.23 CR, complete response PR, partial
response SD, stable disease PD, progressive
disease, CT, computed tomography
9
Confirmed response by subgroups
  KRAS KRAS EGFR EGFR
  wild-type mutant IHC IHC -
n67 n27 n77 n29
CR/PR 70 41 60 69
95 CI 58-81 22-61 48-71 49-85
Responses confirmed by 2nd CT scan according to
RECIST or by resection
Chi square test for comparison between KRAS
wild-type vs KRAS mutant p lt 0.01 CR, complete
response PR, partial response SD, stable
disease PD, progressive disease CT, computed
tomography EGFR, epidermal growth factor
receptor IHC, immunohistochemistry CI,
confidence interval
10
Liver resections
  FOLFOX6 FOLFIRI All
  cetuximab cetuximab patients
n53 n53 n106
R0 resections 38 30 34
R1-resection or resection with RFA 2 8 5
RFA 9 6 8
Total R0 / R1 resection / RFA 49 43 46
RFA, radio frequency ablation
Resections in pts with KRAS wild-type tumors
22/67 pts (33)
11
Perioperative morbidity / mortality
  Chemotherapy cetuximab Chemotherapy cetuximab
  Chemotherapy cetuximab Chemotherapy cetuximab
n45 () n45 ()
All morbidity 16 (36)
Median stay in hospital 13 days 13 days
Median stay on intensive care unit 2.0 days 2.0 days
Mechanical ventilation gt 1 day 4 (9)
Operative revisions 3 (7)
Bleeding 2 (4)
Hepatic failure 1 (2)
Pleural effusion 6 (13)
Wound infection 4 (9)
Biliary leakage 2 (4)
Renal failure 1 (2)
Urinary tract infection 2 (4)
Two post-operative deaths (4) Gram-negative
sepsis 8 days postop (right hemihepatectomy,
FOLFOX cetuximab arm) Multiorgan failure 75
days postop (two-staged liver resection, FOLFOX
cetuximab arm)
12
Blinded surgical review

• A blinded surgical review performed for CT/MRI at
  baseline and at 4 months
• CT / MRI scans were presented by a radiologist to
  5-6 liver expert surgeons of participating
  centers in two workshops.
• CT, computed tomography MRI, magnetic resonance
  imaging

13
Surgical review

• CT / MRI scans were evaluated by the surgeons
  without knowing when the scan was taken (before
  or after chemotherapy) and without clinical data
• Surgeons allocated scans toresection /
  exploration / chemotherapy preferred /
  non-resectable
• Surgeons were blinded to the votes of the other
  participants.
• 181 scans reviewed / 171 scans evaluable
• Paired scans (baseline and follow-up) available
  for 68/106 ptsActual resection rate in this
  subgroup 34, response rate 60
• Following imaging review22/68 scans (32)
  judged resectable at baseline41/68 scans (60)
  judged resectable at follow-up (plt0.01)

14
Waterfall plot of resectability at baseline
Votes for resectable are in green, for
borderline resectable, surgical exploration
recommended in light green, chemotherapy
preferred in yellow and for unresectable in
red. Actual R0 resected cases are marked with
, R resected cases and patients with
radiofrequency ablation -
15
Waterfall plot of resectability after chemotherapy
Resectability according to imaging increased by
28 (32 ? 60) plt0.01
Votes for resectable are in green, for
borderline resectable, surgical exploration
recommended in light green, chemotherapy
preferred in yellow and for unresectable in
red. Actual R0 resected cases are marked with
, R resected cases and patients with
radiofrequency ablation -
16
Patterns of voting of the individual surgeons
Votes for resectable/exploration are in
green, chemotherapy preferred in yellow and for
unresectable in red. Blinded review of
patient MRI and CT scans at baseline and
follow-up revealed large variation between
reviewers in the decision making process.
17
Agreement and critical disagreement
1 2 3 4 5 6 7
1   73 (135) 60 (136) 60 (132) 61 (62) 71 (65) 78 (64)
2 4 (135)   64 (166) 60 (161) 70 (69) 64 (89) 67 (86)
3 3 (136) 2 (166)   55 (164) 56 (68) 76 (92) 62 (89)
4 15 (132) 17 (161) 5 (164)   76 (67) 62 (91) 63 (88)
5 8 (62) 14 (69) 3 (68) 18 (67)      
6 5 (65) 2 (89) 4 (92) 4 (91)     65 (89)
7 2 (64) 2 (86) 2 (89) 15 (88)   4 (89)  
Reviewer
Agreement
Critical disagreement
Agreement (grey/green) show the percentage of
agreement between two individual surgeons with
the categories resection/exploration,
chemotherapy preferred and unresectable.
Total rate in all decisions is 64.5 Critical
disagreement means the proportion of contrary
votes (one surgeon for resection/exploration,
the other for unresectable) The total rate in
all pairs is 6.8 The numbers in brackets mean
the evaluable images per surgical pair.
18
Conclusions

• High response rates induced by cetuximab plus
  either FOLFOX or FOLFIRI
• 70 confirmed response in KRAS wild-type patients
• Resections among patients with initially
  non-resectable liver metastases
• 34 R0 liver resection
• 46 R0 or R1 liver resection and/or RFA
• Perioperative morbidity/mortality comparable to
  experience from literature
• Resectability according to imaging review
  improved significantly following treatment with
  cetuximab plus FOLFOX or FOLFIRI
• Cetuximab plus FOLFOX or FOLFIRI are good options
  for conversional chemotherapy for KRAS wild-type
  patients(Europe, not approved in U.S.)
• Although, as demonstrated here, there is
  heterogeneity between different surgeons in
  treatment decisions, this rarely results in the
  need for a surgical second opinion within
  experienced centers

19
We thank...

• All patients and their relatives
• All investigators at the study sites University
  Hospital Dresden Klinikum Oldenburg University
  Hospital Vienna University Hospital Tübingen
  University Hospital Göttingen University
  Hospital Munich Rechts der IsarKlinikum
  Passau Krankenhaus der Barmherzigen Brüder
  TrierUniversity Hospital / NCT Heidelberg
  University Hospital Würzburg University
  Hospital Frankfurt Klinikum CelleUniversity
  Hospital Essen Klinikum Magdeburg University
  Hospital Mannheim Klinikum AscherslebenKlinikum
  Essen-Mitte
• The companies which supported this
  studyMerck-Serono, Sanofi-Aventis, and Pfizer