Tumor 'Profiling' Takes on New Significance in GI Cancer

Two thematic questions are at the forefront of gastrointestinal (GI) cancer research: What role will immunotherapy play in the management of these cancers, and how will further subtyping of tumors allow for personalized selection of therapies?

The data in GI cancers presented at the 2016 annual meeting of the American Society of Clinical Oncology (ASCO) show that the bar is moving forward on both of these issues, but more work needs to be done. Of note, there have been connections and crossovers between these two research endeavors, which complement each other in the development of better treatments for this spectrum of diseases.

Microsatellite Status Key in Current GI Immunotherapy

Immunotherapy remains in a much earlier phase of development for the treatment of GI cancers when compared with other disease types. Part of the reason for this is that the currently available immune checkpoint inhibitors do not appear to have the same level of activity against GI cancers as against other solid tumors. Understanding why this tempered activity exists will be critical to incorporating these therapies into the treatment of these GI diseases.

In colorectal cancer, the utility of immune checkpoint inhibitors seems limited to tumors that are microsatellite instable (microsatellite instability-high [MSI-H]). Unfortunately, only 4% of colorectal cancers are MSI-H. Establishing what role immune checkpoint inhibitors may have in the management of this small tumor subset is currently an area of active research by several ongoing studies. Among them is the CheckMate-142 study, presented at the meeting by Dr Michael Overman.^[1] In this phase 2 study, patients with MSI-H colorectal cancers are treated with either anti-PD-1 monotherapy with nivolumab or with the combination of nivolumab and anti-CTLA-4 therapy with ipilimumab. This study is ongoing, but the interim results presented suggest clinical activity and tolerability of this strategy. In an effort to address the remaining colorectal cancers with an immunotherapy strategy, Dr Johanna Bendell presented the phase 1b study of the combination of MEK inhibition and anti-PD-L1 targeting, using cobimetinib and atezolizumab, respectively.^[2] The strategy here is sound: Tumors are thought to be sensitized to immunotherapy via MEK inhibition, which upregulates MHC I on tumor cells, leading to intratumoral T-cell infiltration and enhanced opportunity for anti-PD-L1 activity. The interim results of this study also suggest that there is a trend towards improved activity in microsatellite-stable colorectal tumors compared with what would be expected from either monotherapy strategy alone. Of note, tumor response is not correlated with PD-L1 expression status.

Extending the microsatellite story beyond colorectal cancer, Dr Dung Le presented data on microsatellite instability as an independent marker of benefit from immune checkpoint inhibitor therapy in patients with solid tumors.^[3] There were small patient numbers presented whose tumors were MSI-H but also some exciting responses from some otherwise difficult-to-treat GI cancers, including pancreatic and biliary carcinomas. If microsatellite stability is validated as a biomarker of immune checkpoint inhibitor efficacy, it will prove exceptionally useful in selecting patients most likely to benefit from such therapies, and it will aide in the development of therapies for those patients whose tumors do not meet such criteria.

In CRC, Location, Location, Location

Several studies on colon cancer considered the effect of tumor location on disease prognosis. Right-sided colon cancers bearing a poorer prognosis than their left-sided counterparts is not a new concept, but these studies are now considering why the disparity exists, on the basis of molecular drivers of the tumors.

Dr Alan Venook reminded us that, embryologically speaking, the right colon and left colon have their origins in different organs, with the midgut giving rise to the right colon and the hindgut giving rise to the left colon. Dr Venook discussed a further analysis of the SWOG/CALBG 80405 trial looking at tumor origin in the right versus left colon as a prognostic measure.^[4] In this analysis (which was not a preplanned analysis of the trial), right-sided colon cancers had a worse prognosis than those arising from the left colon. There was a suggestion that sidedness could translate into a treatment distinction as well, in regard to first-line monoclonal antibodies. Right-sided tumors seemed to do better when treated initially with the anti-VEGF bevacizumab, while left-sided tumors seemed to fare better with the anti-EGFR cetuximab. Because of the nature in which these treatment results were arrived at, it is premature to consider this method of antibody selection as standard. Dr Venook focused instead on the ongoing need to understand the biologic differences between tumors arising from the right versus left side of the colon.

Supporting these findings, Dr Deborah Schrag presented data from a SEER analysis looking into the prognostic information that can be garnered from considering tumor sidedness.^[5] A similar finding of inferior survival among patients with right-sided colon cancers as compared to those with left-sided colon cancers again was described. Understanding the differences that Drs Venook and Schrag describe is key to the ability to exploit them therapeutically.

Dr Michael Lee discussed biologic features that offer potential explanations for these differences, including NRAS and BRAF mutation statuses, tumor methylation, and molecular subtypes.^[6] Clearly, there is wide variability among colon cancers, and the importance of further understanding these tumor subtypes is key. Going forward, it will be important to consider sidedness, and to explore the underlying biologic differences of such sidedness, as a stratification factor for clinical trial design—and eventually for the treatment of colon cancers.

Emerging Options in Gastric and Pancreatic Cancer

The need for better understanding of tumor subtypes to generate treatment was apparent with the first results of the highly anticipated CRITICS trial for the adjuvant management of gastric cancer.^[7] Two different adjuvant standards exist currently for the management of these diseases: perioperative chemotherapy, on the basis of the MAGIC trial; and postoperative chemo/radiation therapies, on the basis of the Intergroup study.^[8,9]

The CRITICS trial, a hybrid of these two approaches, attempts to find a single standard for the adjuvant therapy of gastric cancer. All patients enrolled receive the same initial therapy: three cycles of chemotherapy consisting of a triplet combination of an anthracycline, a platinum, and a fluoropyrimidine, followed by definitive surgery. Postoperatively, patients are randomly assigned to either chemotherapy or chemoradiation combination therapies. At this first look at CRITICS, presented by Dr Marcel Verheji, there were no significant differences between the two study arms. Given our growing understanding of the variability of gastric cancers, multiple standard options should not be surprising.

Going forward, greater strides are likely to be made by stepping away from a one-size-fits-all chemotherapy to a more personalized approach, which augments effective chemotherapy with appropriately selected targeted agents.

In pancreatic cancer, Dr John Neoptolemos, from the University of Liverpool, presented results of the ESPAC-4 study, which found good evidence that a chemotherapy combination may be a new, superior adjuvant regimen in the management of pancreatic cancer.^[10] This study compared gemcitabine monotherapy versus the combination of gemcitabine with capecitabine in the adjuvant setting. Because this trial began in 2008, patients were managed in a classic manner—that is, they went directly to surgery without having neoadjuvant therapy.

The study included "real world" patients, such as those with node-positive disease and R1 surgical resection margins. The median overall survival differences were statistically significant and favored the dual therapy, at 28 months for the combination and 25.5 months for gemcitabine alone. Moreover, at 5 years, survival differences were 28.8 months versus 16.3 months, again favoring the combination arm. The recurrence rates were not reported.

Standard treatment for pancreas cancer remains difficult to define, given the available modalities and sequences of treatment. Nevertheless, it remains the deadliest cancer, and every opportunity to offer a greater benefit to our patients is critical. The data from the ESPAC-4 certainly helps move that bar in a meaningful way.

In the near future, tumor subtyping will refine the therapies offered to our patients. Appreciation for tumor heterogeneity is already apparent in the design of therapeutic trials. Needless to say, unraveling these subtypes will increase the complexity of treatment recommendations, but it will also allow for more precise treatment recommendations for our patients.