24 ORAL Modulation of adjuvant 5-FU treatment by folinic acid and interferon-alpha in colon and rectal cancer: Comparison of two phase-III trials including 1652 patients

S8

yTNM I II III

Proffered papers: Colorectal cancer II

0

I

II

ns ns 0.0307

0.016 0.005

0.045

20%/98%, for pT1-2 15%/96%, for pT3 64%/48%, and for pT4 86%/ 75%. The difficulty appears to be related to the interpretation of post RCT fibrosis. Conclusion: Downstaging after chemoradiation of a locally advanced tumor on MRI is often underestimated. The prediction on MRI of tumor regression from the MRF is better, with specific morphological features on MRI corresponding to regression from or persistence of invasion of the MRF. The problem remains the interpretation of the dark areas of post RCT fibrosis. 23

ORAL

Adjuvant chemotherapy for stage III (sIII) colon cancer in the canadian atlantic provinces. What is the economic impact of capecitabine versus 5FU/LV? K. Virik 1 , C. Skedgel 2 , T. Younis 1 . 1 QE II Health Sciences Centre, Dalhousie University, Medical Oncology, Halifax, Canada; 2 Capital Health and Dalhousie University, Medicine, Halifax, Canada

Non-responder patients had the worse prognosis (5-years survival 30%). The variable with higher prognostic significance was yN (p < 0.0003), especially if we distinguish N1 by N2 (p < 0.00022). With a mean followup of 36 months, local recurrence rate was 5.4%. Conclusions: the prognosis of patients with advanced rectal cancer submitted to neoadjuvant radiochemotherapy depends on their response to therapy (yTNM) and not on cTNM. 22

ORAL

How good is MRI after neo-adjuvant chemoradiation of locally advanced rectal cancer? G. Beets 1 , R. Vliegen 2 , G. Lammering 3 , E. Dresen 2 , H. Rutten 4 , K. Oei 2 , A. Kessels 5 , A. De Bruïne 6 , M. von Meyenfeldt 1 , R. Beets-Tan 2 . 1 University Hospital Maastricht, Surgery, Maastricht, The Netherlands; 2 University Hospital Maastricht, Radiology, Maastricht, The Netherlands; 3 Maastro Clinic, Radiotherapy, Maastricht, The Netherlands; 4 Catharina Hospital, Surgery, Eindhoven, The Netherlands; 5 University Hospital Maastricht, Epidemiology, Maastricht, The Netherlands; 6 University Hospital Maastricht, Pathology, Maastricht, The Netherlands Purpose: Chemoradiation (CRT) for locally advanced rectal cancer often results in downsizing of the tumor mass. An important clinical question is whether a resection should include all originally involved tissues and organs, or that a more conservative resection is allowed in patients who respond well. The aim of the study is to assess the accuracy of post CRT MRI for the prediction of tumor downstaging and invasion of the mesorectal fascia (MRF), and to identify MRI criteria for tumor regression from the MRF. Patients and Methods: The pre- and post CRT MR images of 64 patients with locally advanced rectal cancer (defined by involved or threatened MRF) were assessed for T stage, invasion of the MRF (confidence level scoring) and morphological pattern of the primary tumor near the MRF. As MRI has been shown to be reliable to image the extent an unirradiated primary tumor, the pre CRT MRI was considered as the reference. The post CRT MRI was compared with the histology of the surgical specimen and sensitivity, specificity, PPV and NPV were calculated and a ROC curve was constructed. Results: Invasion of MRF: Tumor regression from the mesorectal fascia occurred in 34/64 (53%) patients. The area under the ROC curve, sensitivity, specificity, PPV and NPV of post CRT MRI for invasion of the fascia was 0.81, 100%, 32%, 57%, 100% respectively. Three groups of morphological patterns were observed. 1) unchanged tissue strands, unchanged minimal fat pads with or without development of strands correlated with 0% invasion 2) mainly nodular or diffuse iso/hyper-intense infiltration of MRF becoming diffuse hypo-intense infiltration correlated with 38% invasion and 3) mostly unchanged nodular or diffuse iso/hyper- intense infiltration of MRF correlated with 89% invasion of MRF. Downstaging in T-stage: 33/64 (52%) tumors showed downstaging, with a CR rate of 8%. Sensitivity/ specificity of post CRT MRI for CR was

Background: Adjuvant chemotherapy for SIII colon cancer is an accepted standard of care. Oral Capecitabine (Cap) has been shown to be at least equivalent and possibly superior to 5FU/LV (Mayo regimen) with a superior relapse free survival (RFS). At the present time, provincial funding for Oxaliplatin is not uniform as the drug is currently not commercially available in Canada and Health Canada Special Access Program approval is required for its use. Capecitabine is associated with a higher drug cost but improved toxicity profile and appears to be cost-effective (CE). An economic analysis was undertaken to compare these two alternatives and examine their potential budget impact in the Atlantic Provinces (AP) for 2005 onwards. Methods: A cost minimisation analysis was performed in Canadian $ ($) using Cap and 5FU/LV given as per the X-ACT trial. The direct costs including chemotherapy drug acquisition, laboratory investigation, health resources utilisation and the cost of hospitalisation due to adverse events were examined based at the Nova Scotia Cancer Centre. Indirect costs included travel and opportunity cost based on the average provincial wage and participation rates. Complete drug delivery was assumed. A CE model was also constructed. The Markov model developed used a hypothetical cohort of 1000 pts with SIII colon cancer and projected costs and outcomes over a 5 year horizon (discounted at 3% and in $ 2005). All recurrences were modeled to death. The number of patients (pts) with SIII colon cancer suitable for adjuvant chemotherapy in each of the AP was estimated from the literature [1]. Results: The use of CAP is an extra $3739 per patient (pp), principally reflecting the higher drug cost although the indirect costs favour CAP (difference: $2464/patient). However, it is a CE option compared to 5FU/LV. The cost is $3477 pp when accounting for relapses avoided. The projected cost impact for the AP using the model at 5 years for 2007: NL $215K, PEI $58K, NS $393K pts and NB $315K. Sensitivity analyses for key parameters will be provided. Conclusion: CAP has more direct costs but has indirect cost savings. The potential budget impact in the AP is estimated at $981K using the model to account for the impact in decreases in recurrence. Reference: [1] Canadian Cancer Statistics 2005, Canadian Cancer Society. 24

ORAL

Modulation of adjuvant 5-FU treatment by folinic acid and interferon-alpha in colon and rectal cancer: Comparison of two phase-III trials including 1652 patients K. Link 1 , L. Staib 1 , W. Baumann 1 , M. Redenbacher 1 , S. Sander 3 , D. Henne-Bruns 2 , M. Kornmann 2 . 1 Study Group Oncology of Gastrointestinal Tumors, Department of Surgery, Wiesbaden, Germany; 2 University of Ulm, Department of General, Visceral and Transplantation Surgery, Ulm, Germany; 3 University of Ulm, Department of Biometry and Medical Documentation, Ulm, Germany The benefit of adjuvant therapy in lymph node-positive colon cancer was established using 5-fluorouracil (5-FU) and levamisol (LEV). Combined postoperative chemoradiotherapy has been recommended for patients (pts) with locally advanced rectal cancer. 5-FU cytotoxicity can be modulated by folinic acid (FA) or interferon-alpha (INFa). The aim of this study was to compare the efficacy of modulating 5-FU+LEV by either FA or INFa

Proffered papers: Sentinel node biopsy in the adjuvant treatment of colon and rectal cancer. Pts with R0-resected colon (UICC IIb and III) or rectal (UICC II and III) cancer were stratified according to T, N, and participating center and randomized to receive 5-FU+LEV alone or in combination with FA or INFa. Rectal cancer pts additionally received postoperative radiation. A total of 855 eligible pts were enrolled in the colon (FOGT1) and 797 pts in the rectal (FOGT2) cancer trial, respectively. Addition of FA to 5-FU+LEV improved recurrence-free and overall survival (p=0.007 and p=0.004, respectively) in colon cancer pts. There was also a tendency toward an improved 3-year overall survival by addition of FA to 5-FU+LEV for rectal cancer pts with 79% (95% CI: 74-84%) in comparison to 5-FU+LEV alone with 71% (95% CI: 65-77%). In contrast to colon cancer these beneficial effects vanished after 5 years in rectal cancer. Addition of INFa did not alter survival of colon and rectal cancer pts. Toxicity was generally tolerable for 5-FU+LEV and 5-FU+LEV with FA, but significantly higher by adding INFa especially for rectal cancer pts. Addition of INFa was associated with increased toxicity without influencing the outcome and should therefore not be recommended for adjuvant treatment of colon and rectal cancer. Addition of FA increased the 5-year recurrence-free and overall survival rate of colon cancer pts by 9.3 and 11.5 percentage points, respectively, but was without significant effects after 5 years in rectal cancer. Modulation of 5-FU by FA is effective for improving the outcome of colon cancer pts. Due to the fact that the initially observed beneficial effects of modulating 5-FU by FA in rectal cancer did not translate into an improved outcome after 5 years suggests that rectal cancer may be different from colon cancer in regard to chemosensitivity and that more intense regimens are mandatory to improve the prognosis of locally advanced rectal cancer.

PROFFERED PAPERS

Sentinel node biopsy 25

ORAL

A randomized clinical trial on sentinel node biopsy versus axillary lymph node dissection in breast cancer: early results of the Sentinella/GIVOM trial G. Zavagno 1 , P. Del Bianco 2 , G. Scalco 3 , L. Barutta 4 , P. Burelli 5 , P. Carraro 6 , G. Tacchetti 7 , R. Mencarelli 8 , G.L. De Salvo 2 , D. Nitti 1 . 1 University Hospital, Surgery Branch, Padova, Italy; 2 Istituto Oncologico Veneto, Clinical Trials and Biostatistics Unit, Padova, Italy; 3 S Bortolo Hospital, Surgery Branch, Vicenza, Italy; 4 S Martino Hospital, Surgery Branch, Belluno, Italy; 5 General Hospital, Surgery Branch, Conegliano, Italy; 6 Ca’ Foncello Hospital, Surgery Branch, Treviso, Italy; 7 General Hospital, Surgery Branch, Dolo, Italy; 8 University Hospital, Pathology Department, Padova, Italy Background: The sentinel lymph node biopsy (SLNB) is widely used for axillary staging in breast cancer patients. However, all studies report variable rates of false negative results, whose prognostic impact is still unclear. In 1999, GIVOM (Gruppo Interdisciplinare Veneto di Oncologia Mammaria) started a multicentric randomized trial to evaluate whether SLNB is equivalent to axillary lymph node dissection (ALND), in terms of disease free and overall survival. We report the analysis of results at 52.8 months. Methods: Patients with breast cancer < 3 cm were randomized to SLNB followed by ALND (group A) or SNB followed by ALND only if sentinel node was metastatic (group B). Surgeons should have performed at least 15 cases of SLNB followed by ALND as a learning curve before randomizing patients, however a surgical quality control was not planned. The day before surgery, patients underwent a peritumoral subdermal injection of 30-50 MBq of Nanocoll. SN was harvested under the guidance of a γ probe. Histological examination of the SNs was performed using haematoxylin-eosin and immunohistochemical staining. Follow up controls were performed at 6-monthly intervals. Results: 723 patients, enrolled in 18 centers, belonging to either academic or community hospitals, were considered for analysis (365 in group A and 358 in group B). The rate of intra-operative SN identification was similar between the groups (95.1%, 95%CI=92.5-97.0 in group A and 95.0%, 95%CI=92.4-96.9 in group B). In group A, the overall accuracy of the SN status was 91.0%, the negative predictive value 92.6% (95%CI=88.8-95.4), the sensitivity 83.2%. Therefore, the false negative rate (FNR) was 17.8% (18 of 107, 95%CI=10.6-24.6). At a median follow up of 52.8 months, no

S9

clinical axillary recurrences were observed in both groups. Distribution of recurrences and deaths among the two groups are reported in Table 1. Table 1 Event

Arm A (n=365)

Arm B (n=358)

Supraclavicular metastasis Ipsilateral breast Controlateral breast Distant metastasis

0 5 2 14

3 11 1 11

Death due to breast cancer Death from other causes

6 4

10 10

Conclusions: In this study, the sentinel node FNR is higher than average rates reported in the literature. Since this trial has been carried out in a heterogeneous setting and without request of a formal training course for participating surgeons, it well reflects the situation in a non-specialist clinical practice setting. In spite of the high FNR, however, no significant differences in prognosis were found between the two arms at 52.8 months. 26

ORAL

Morbidity comparison of sentinel lymph node procedure versus conventional axillary lymph node dissection for breast cancer patients: results of Sentinella – GIVOM, an Italian randomized clinical trial G.L. De Salvo 1 , P. Del Bianco 1 , R. Marconato 2 , F. Bozza 3 , C. Racano 4 , P. Pietrarota 5 , G. Meneghini 6 , D. Casara 7 , D. Nitti 8 , G. Zavagno 8 . 1 Istituto Oncologico Veneto, Clinical Trials Unit, Padova, Italy; 2 General Hospital, Surgery Branch, Venezia, Italy; 3 General Hospital, Surgery Branch, Padova, Italy; 4 General Hospital, Surgery Branch, Cittadella, Italy; 5 University Hospital, Surgery Branch, Verona, Italy; 6 General Hospital, Surgery Branch, Montecchio Maggiore, Italy; 7 Istituto Oncologico Veneto, Nuclear Medicine, Padova, Italy; 8 University Hospital, Surgery Branch, Padova, Italy Purpose: to compare physical morbidity and health-related quality of life (HRQOL) in breast cancer patients who received standard axillary dissection (ALND) or sentinel lymph node biopsy, followed by axillary dissection only in the case of sentinel-node positivity (sentinel lymph node biopsy: SLNB), within a randomized clinical trial. Patients and Methods: Patients with early breast cancer ≤ 3 cm with a clinically negative axilla were randomly allocated to ALND or SLNB. All patients underwent physical examination every 6 months in order to assess any arm-related symptoms. A subset of patients completed the SF36 quality of life questionnaire and the Psychological General Well Being Index (PGWBI) before randomization, at 6 and 12 months after surgery and yearly thereafter. Results of the first 18 months are reported. Results: Between May 1999 and December 2004, 721 patients entered the trial: 365 patients were randomized to the ALND group and 356 to the SLNB group. Six months after surgery, the SLNB group had significantly less lymph-edema, movement restrictions, pain and numbness with respect to the ALND group, but this difference ceased to be significant at 12 and 18 months, except for numbness that remained significantly less frequent in the SLNB arm (Table 1). 321 patients recruited from March 2001 participated in the quality of life assessment. No differences were found between the two groups in any HRQOL domains.

Table 1. Arm morbidity comparisons by time from surgery

Edema

Limitations

Pain

Numbness

6 months 12 months 18 months 6 months 12 months 18 months 6 months 12 months 18 months 6 months 12 months 18 months

OR (SNP/ALND)

95% CI

p

0.41 0.51 0.59 0.49 0.73 0.61 0.53 0.79 0.81 0.65 0.53 0.37

0.21-0.80 0.27-0.98 0.30-1.18 0.29-0.82 0.40-1.35 0.29-1.28 0.34-0.84 0.48-1.31 0.46-1.43 0.46-0.93 0.35-0.82 0.22-0.63

0.0096 0.0441 0.1364 0.0062 0.3154 0.1945 0.0071 0.3633 0.4696 0.0195 0.0042 0.0002