This special “annual highlights” supplement to Oncology News International (ONI)is a compilation of selected news on important advances in the management ofgastrointestinal cancers over the past year, as reported in ONI. Guest Editor, Dr.James L. Abbruzzese, comments on the reports included herein and discussesdevelopments in the clinical management of GI cancers, with a look at the impactof targeted agents with cytotoxic chemotherapy, first-line and adjuvant therapies foradvanced disease, and the role of statins and COX-2 inhibitors in prevention.
SOUTH SAN FRANCISCO-Adding bevacizumab (Avastin) to5-FU/leucovorin (LV) reduced risk ofdeath by 25% vs 5-FU/LV or irinotecan(CPT-11, Camptosar)/5-FU/LV(IFL), according to a retrospectivecombined analysis."We showed a quite significant survivalin the group when we add bevacizumabto 5-FU/LV," said investigatorRobert D. Mass, MD, director ofOncology, Genentech, Inc., South SanFrancisco, California.The investigators cited several studiesshowing that bevacizumab improvedsurvival time in metastatic colorectalcancer. These include a phaseIII randomized trial (N Engl J Med350(23):2335-2342, 2004) using bevacizumaband IFL, and three randomizedstudies in which bevacizumab wascombined with 5-FU/LV."The difficulty with all of these studiesis that they were all quite small, andthey all had very strong trends in survival,but none of them reached statisticalsignificance," Dr. Mass said.To address this issue, Genentechinvestigators performed a combinedanalysis, in which data for bevacizumab/5-FU/LV-treated patients waspooled and compared with a "combinedcontrol" of 5-FU/LV and IFL(abstract 3616). "It was basically amethod by which we could get morerobust statistics," Dr. Mass said at the40th Annual Meeting of the AmericanSociety of Clinical Oncology.The combined analysis included atotal of 249 patients who received bevacizumab/5-FU/LV (with the bevacizumab5-mg dose) and 241 receivingeither 5-FU/LV alone or IFL. Allthe studies used the standard RoswellPark bolus 5-FU/LV regimen.Dr. Mass reported a 25% reductionin risk of death for patients receivingthe bevacizumab-containing regimenvs 5-FU/LV or IFL alone (hazard ratio[HR] = 0.742, 95% confidence interval,0.59-0.93). Overall survival timewas 17.9 months, vs 14.6 months inthe combined control group (P =.0081)."The survival numbers were notquite significant for any of these trialsby themselves," Dr. Mass said, "butwhen we looked at them as a combinedanalysis, we saw a highly significant[difference]."Progression-free survival time wassignificantly increased in the bevacizumab/5-FU/LV group, from 5.6months to 8.8 months (HR = 0.63, P =.0001). Interpretation of safety differenceswas confounded by inclusion ofIFL-treated patients in the controlgroup, although adverse events were"consistent" with what was reportedin the individual trials, according toinvestigators.Wide Survival BenefitsIn a related study, Genentech investigatorsdescribed a subgroup analysisof survival from the phase III trialof bevacizumab/IFL recently reportedin the New England Journal of Medi-cine. Investigators reported an increasein survival from 15.6 months for IFLto 20.3 months for bevacizumab/IFL(P < .001).To evaluate the effects of baselinerisk factors on survival, they analyzeda number of risk factors, includingperformance status, primary tumorsite, age, sex, race, prior treatment,and duration of metastatic disease, aswell as baseline levels of albumin, alkaline phosphatase, and lactate dehydrogenase.The survival benefit, however, wasseen in all these prespecified subgroups.For example, median survivaltime increased from 14.9 months to21.6 months for patients who had receivedprior radiotherapy, and from15.6 months to 19.9 months for thosewho had not.Adjusting for a number of baselinefactors, investigators determined thatthe addition of bevacizumab resultedin a 34% reduction in hazard of deathvs those patients who had receivedplacebo."It didn't matter whether [patients]were old or young, had low orhigh albumins, or poor vs good performancestatus," Dr. Mass said. "Bevacizumabplus IFL benefited all theprespecified subsets."