Immunotherapy Doubts Fading in GI Cancers

Immunotherapy approaches are showing early signs of activity against a range of gastrointestinal cancers, defying the skeptical view that these tumors would not respond to the emerging agents succeeding in other malignancies.

Christopher R. Heery, MD

Immunotherapy approaches are showing early signs of activity against a range of gastrointestinal (GI) cancers, defying the skeptical view that these tumors would not respond to the emerging agents succeeding in other malignancies.

That was part of the message that Christopher R. Heery, MD, an immunotherapy research leader at the National Cancer Institute (NCI), delivered during a presentation at the 1st Annual School of Gastrointestinal Oncology™ that Physicians’ Education Resource (PER) hosted in New York City on April 23.

Antibodies aimed at the CTLA-4 or PD-1/PD-L1 immune checkpoints have generated responses in gastric, pancreatic, colorectal (CRC), and hepatocellular carcinoma, said Heery, who is director of the Clinical Trials Group at the NCI’s Laboratory of Tumor Immunology and Biology.

The responses, however, have been lower than in other tumor types, and Heery emphasized that more research is needed to understand the underlying biology of GI cancers. Studies thus far suggest that immune responses can be primed or induced, and several trials into combinations of checkpoint blockade agents and vaccines are getting underway.

Milestones in GI Cancers

“For most cancers, we can’t just do checkpoint inhibition,” said Heery. “We’ll need something else to generate T-cell responses. We’ll need more studies to understand what are the mechanisms of a given tumor that need to be targeted to make this all happen.”Heery said recent research into agents targeting PD-1/PD-L1 in gastric cancer and in CRC tumors with mismatch-repair (MMR) deficiency have shown the potential for checkpoint blockade strategies.

In a phase I study, durvalumab (MEDI4736) elicited a 7% response rate (2 of 28 patients), according to an interim analysis.1

In a phase Ib study, pembrolizumab (Keytruda) monotherapy generated a 30.8% overall response rate among 39 patients with PD-L1—expressing recurrent or metastatic adenocarcinoma of the stomach or gastroesophageal junction.2 PD-L1 expression was considered positive if ≥1% of tumor cells or any part of the tumor stroma stained positive on archived tumor samples using a prototype immunohistochemistry assay.

Another milestone for immunotherapy in GI cancers came in 2015 when researchers demonstrated a 40% response rate after pembrolizumab treatment among 10 patients with heavily pretreated CRC who harbored genetic defects in DNA MMR pathways.3 There were no responses among patients whose tumors were MMR proficient.

Among 17 participants with noncolorectal MMR-deficient GI cancers who received pembrolizumab, the objective response rate was 47%, according to updated results of that study presented during the 2016 Gastrointestinal Cancers Symposium.4 These included patients with ampullary, pancreatic, biliary, small bowel, and gastric cancers. Four participants achieved a complete response.

MMR-deficient tumors harbor thousands of mutations that may produce neoantigens that can be recognized and targeted by T cells, leading to the hypothesis that immune augmentation with PD-1/PD-L1 blockade could be highly effective in MMR-deficient tumors.

Another useful immune-related marker is the level of tumor-infiltrating lymphocytes (TILs), which can be measured at diagnosis and is predictive of outcomes in colorectal cancer, Heery said. In 2006, Galon et al posited that in situ analysis of CD3 TILs was a better predictor of survival than standard histopathological methods.5

Focus on Microenvironment

Heery said the research showed that a higher density of TILs correlated with a lower risk of recurrence almost independent of the stage at diagnosis.The success of therapies aimed at immune checkpoints has established that these agents act on T cells, Heery said.

“Immune checkpoint inhibitors require underlying T-cell activation,” he said. “If there aren’t T cells trying to kill the tumor, using a checkpoint inhibitor is essentially a waste of time.”

Only a minority of tumors have natural T cells including 50% of melanomas, 20% to 30% of renal cell carcinomas, and 10% to 20% of lung cancer and CRC, according to Heery.

Yet, although CRCs are in a similar range as lung cancers for the prevalence of natural T cells, the checkpoint blockade agents have been less successful in CRC than they have been in lung cancer.

“At baseline, a lot of patients with colon cancer have dense T-cell infiltrate but in metastatic disease that number seems to go down significantly,” said Heery, adding that more research is needed to understand the dynamics.

At the same time, Heery said that immune responses could be induced in patients with GI cancers but that a change in tactics would be needed. He said anticancer vaccines typically are evaluated in patients who have received multiple lines of therapy but that in fact “vaccines work best with minimal disease.”

Combinations of checkpoint blockade agents with vaccines and/or standard therapies such as chemotherapy, targeted therapies, or radiation are under consideration.

Heery said 1 trial scheduled to get underway at Georgetown University is a small phase I/II study of durvalumab combined with the PANVAC pox viral vaccine, capecitabine, and bevacizumab. Another study would investigate avelumab, which also targets PD-L1, in combination with FOLFOX-A (FOLFOX plus bevacizumab) and a carcinoembryonic antigen (CEA)—based vaccine.

“What we’re going to need to do is understand on a per-patient basis what the inhibitory factors within that tumor are,” said Heery. “If we can tip the balance so that we activate more T cells and get them into the tumor, we might be able to overcome this lack of immune cells in the tumor.”

“The goal is to take a T-cell poor tumor and drive it to have T-cell infiltrates,” Heery noted.


  1. Segal NH, Hamid O, Hwu W, et al. A phase I multi-arm dose-expansion study of the anti-programmed cell death-ligand-1 (PD-L1) antibody MEDI4736: Preliminary data. Ann Oncol. 2014;25(suppl 4): iv361-iv372. 10.1093/annonc/mdu342.
  2. Muro K, Bang Y, Shankaran V, et al. A phase 1b study of pembrolizumab (Pembro; MK-3475) in patients (Pts) with advanced gastric cancer. Presented at: ESMO 2014 Congress; September 26-30, 2014; Madrid, Spain. LBA15.
  3. Le DT, Uram JN, Wang H, et al. PD-1 blockade in tumors with mismatch-repair deficiency. N Engl J Med. 2015;372(26):2509-2520.
  4. Le DT, Uram JN, Wang H, et al. PD-1 blockade in mismatch repair deficient non-colorectal gastrointestinal cancers. J Clin Oncol. 2016;34(suppl 4S; abstr 195).
  5. Galon J, Costes A, Sanchez-Cabo F, et al. Type, density, and location of immune cells within human colorectal tumors predict clinical outcome. Sci Transl Med. 2006;313(5795):1960-1964.

<<< View more from the 2016 School of Gastrointestinal Oncology