PHILADELPHIA-"We tend to think of colorectal cancer (CRC), gastric cancer, and pancreatic cancer as homogenous diseases with monolithic approaches to therapy, but in reality we have a very large, heterogeneous population of patients whose individual risks we have just begun to understand. Few patients get on study, and we base all of our standard treatments on those small numbers of patients. That leaves many questions unanswered," stated Daniel G. Haller, MD, of the University of Pennsylvania in Philadelphia. Patients have not yet begun to be selected prospectively for trials based on simple measures such as performance status, polymorphisms that predict toxicity, or predictive markers for response. The question Dr. Haller asks is should studies be continued simply because there is a response without considering the fact that agents may have become more toxic and more expensive? Colorectal Cancer The major issues in chemotherapy for metastatic CRC are whether there are optimal first- and second-line regimens, advantages of single-agent vs combination chemotherapy, patient selection, selection of drugs based on predictive markers, duration of therapy (continuous vs fixed), sequencing of chemotherapy, and optimal endpoints for clinical trials (overall survival, time to disease progression, or clinical benefit). Avenues of investigation for treatment of metastatic CRC include the desire to substitute oral fluoropyrimidines for infusional agents because of their suitability for combination regimens with oxaliplatin(Drug information on oxaliplatin) (Eloxatin) and irinotecan(Drug information on irinotecan) (Camptosar) and use with irradiation. According to Dr. Haller, the key new agents for investigation are probably epidermal growth factor receptor inhibitors such as cetuximab(Drug information on cetuximab) (Erbitux) and antiangiogenesis agents such as bevacizumab(Drug information on bevacizumab) (Avastin). "Unless we've been living in a crypt, we all know about the N9741 data showing that FOLFOX (fluorouracil [5-FU], leucovorin, oxaliplatin) is a good candidate for first-line use," Dr. Haller stated (see Table 1). FOLFOX was similarly effective for second-line use (see Table 2). Researchers trying to determine the optimal sequence for deploying the various 5-FU/irinotecan and 5-FU/ oxaliplatin regimens have thus far not identified any significant difference in the sequences tested. "There really is no difference between FOLFOX-6 and FOLFIRI (5-FU/leucovorin/irinotecan) in the meaningful endpoint of response," Dr. Haller added. "The main differences are in toxicity." Molecular profiling studies are seeking prognostic markers to identify patients with good survival who do not need adjuvant chemotherapy and predictive markers to identify patients who will benefit from therapy. The biologic markers of prognosis in CRC most likely to enter trials are thymidylate synthase levels, microsatellite instability, and allelic loss in chromosome 18. "Looking at our progress in firstline treatment of metastatic CRC, we find that response rate and survival have both increased," Dr. Haller revealed. In adjuvant treatment, the MOSAIC trial comparing infusional 5-FU/ leucovorin with or without oxaliplatin (FOLFOX-4) showed that FOLFOX produced improvements in 3-year disease-free survival (77.8% vs 72.9%, P < . 01). This difference is both statistically significant and clinically relevant, according to Dr. Haller. Other trials have shown that an improvement in 3-year survival translates into a 5-year-survival advantage. The question of how to identify subgroups of patients with different potentials for benefit or toxicity remains an important one. A recent pooled analysis of data from adjuvant rectal cancer trials showed that T and N stages have independent prognostic significance. Single risk factors of N+ disease confined to the wall or N- disease extending beyond the wall without adherence or invasion are both associated with survival and disease control. This analysis concluded that the new T and N stages best reflect prognosis and should be used to present information to patients. Gastric Cancer The direction for treatment of patients with gastric cancer is less clear. The V 325 phase III trial attempted to learn whether adding docetaxel(Drug information on docetaxel) (Taxotere) to the standard regimen of cisplatin(Drug information on cisplatin) and 5-FU was beneficial. Interim analysis suggests significant improvements in response rate, progression- free survival, and overall survival from the three-drug regimen. "This is a very active combination but it is not the easiest to give. We have had a lot of discussion about the fact that a quarter of the patients, many of whom were responding, stopped therapy because they saw no point in continuing," Dr. Haller said. With regard to adjuvant chemotherapy, interpreting results from previous individual trials is complicated by the fact that many used suboptimal chemotherapy regimens. Meta-analyses are consistent with possible survival benefits. "The de facto standard, as with rectal cancer, has become chemoradiation," Dr. Haller stated. "The question remains: Does every one of these patients need radiation therapy, or are some going to benefit simply from better surgery and adjuvant chemotherapy?" The MAGIC trial of neoadjuvant epirubicin(Drug information on epirubicin) (Ellence), cisplatin, 5-FU (ECF), surgery, and postoperative ECF vs surgery alone showed a small survival advantage for the combination regimen. In a slightly larger trial, with better patient selection, Dr. Haller believes that this would have been a definitively positive result. Pancreatic Cancer Gemcitabine(Drug information on gemcitabine) (Gemzar) combinations have attracted much attention in the treatment of patients with pancreatic cancer. "All of these have had promising phase II data and not-so- promising phase III data," Dr. Haller said. "One of the issues is that we may not know how to give gemcitabine most effectively. It is a pro-drug and may not reach its full potential when given over 30 minutes rather than prolonged infusion." Data on the irinotecan/gemcitabine combination presented at the American Society of Clinical Oncology (ASCO) 2003 meeting suggested that although there was a better response rate with the combination than with gemcitabine alone, it did not translate into a survival advantage. A phase III trial of gemcitabine/cisplatin was similarly disappointing, Dr. Haller concluded.