Mismatch repair–deficient (dMMR) tumors beyond those in the rectum may respond to PD-1 blockade in the neoadjuvant setting, offering the option of organ preservation in early-stage cancer regardless of the tumor type, researchers from Memorial Sloan Kettering (MSK) Cancer Center reported at the 2025 American Association for Cancer Research (AACR) Annual Meeting.¹ Their findings were simultaneously published in TheNew England Journal of Medicine.²

In 2022, the same researchers reported that 100% of patients with dMMR rectal tumors achieved clinical complete responses and avoided surgical resection after neoadjuvant treatment with the anti–PD-1 agent dostarlimab-gxly.³ Because of its historic outcomes, the study was presented during a special session and received a standing ovation at that year’s ASCO Annual Meeting.

“These findings are very important for patients with early-stage dMMR tumors, because it’s likely they do not need surgery or radiation if they are treated first with immunotherapy for a sufficient time.”

— Andrea Cercek, MD

Tweet this quote

Both studies were led by Andrea Cercek, MD, an attending and Section Head of Colorectal Cancer at MSK, and LuisDiaz, Jr, MD, an attending and Head of Solid Tumor Oncology at MSK. “That study was in just 14 patients…. Given that exciting data and the durability of those responses, we wanted to ask whether we could extend this beyond rectal cancer,” Dr. Cercek said, prefacing the current study at the AACR meeting.

She explained that although dMMR as a biomarker occurs more frequently in cancers of the colon and rectum, dMMR status is found across many solid cancers. “In collaboration with Dr. Diaz, we sought to determine how effectively immunotherapy could induce tumor elimination in a broad range of early-stage dMMR cancers and whether these patients with complete clinical responses could then forgo surgical resection.”

Luis Diaz, Jr, MD

In the current study, 80% of all patients achieved a clinical complete response, including 100% of the rectal cancer cohort and 65% of those with other solid dMMR tumors. Most patients were managed nonoperatively, and fewer than 5% had recurrences, Dr. Diaz reported.

In the phase II trial, a total of 103 patients with stage II to III resectable dMMR cancers were treated with dostarlimab for 6 months. Those who achieved a clinical complete response could choose not to undergo resection or any other treatment. The first cohort included 49 patients with rectal cancer, and the second cohort included 54 patients with nonrectal cancers, including gastroesophageal, hepatobiliary, genitourinary, and gynecologic malignancies.

Both cohorts shared the primary endpoint of clinical complete response after dostarlimab and exploratory endpoints involving circulating tumor DNA (ctDNA). Baseline tumor mutational burden and microsatellite instability scores were similar between the two cohorts, and 95% of patients had detectable ctDNA prior to immunotherapy.

Key Findings

Of the 49 patients in the rectal cohort, 100% achieved complete clinical responses, 100% elected nonoperative management, and 96% were recurrence-free at 2 years (median follow-up = 30 months). Among the 54 patients with nonrectal tumors, 35 (65%) had complete clinical responses, of whom 33 elected nonoperative management; the 2-year recurrence-free survival rate was 85% (median follow-up = 15 months). Altogether, of the 84 patients in both cohorts who experienced complete clinical responses, 82 chose to skip surgery, Dr. Cercek noted.

By tumor type, aside from the responses demonstrated in rectal tumors, the best complete response rates were observed in urothelial tumors (100%), hepatobiliary tumors (100%), colon cancers (82%), and gastric cancers (57%), with lower response rates in the other tumor types.

The investigators are closely evaluating the pathology of the incomplete responders to look for clues that may explain their outcomes. They found that several tumors had developed unrelated MMR-proficient tissue, which does not respond to immunotherapy; others had been exposed to previous anti–PD-1 therapy. It is possible that incomplete responders may respond better to combination checkpoint blockade, longer duration, or in combination with chemotherapy, Dr. Cercek said.

Tumor-informed ctDNA levels were detectable at baseline in 87% of patients, and on-therapy levels correlated with complete and incomplete responses. More than 60% of patients with incomplete responses were ctDNA-positive not only at the end of treatment but also as early as 3 months and even 6 weeks after treatment. Approximately 20% of complete responders were ctDNA-positive at 6 weeks, with levels declining by 3 months and disappearing by the end of treatment.

At the time of this reporting, four patients’ complete responses had lasted 5 years. “These findings are very important for patients with early-stage dMMR tumors, because it’s likely they do not need surgery or radiation if they are treated first with immunotherapy for a sufficient time,” Dr. Diaz said.

At a press briefing, Dr. Cercek indicated she believes the ability to elicit these responses in dMMR solid tumors is a class effect of anti–PD-1/L1 agents and is probably not unique to dostarlimab. 

DISCLOSURE: Dr. Cercek has served as an advisor for Amgen, Agenus, AbbVie, Daiichi Sankyo, Janssen, Merck, GSK, Pfizer, 3T Biosciences, Summit, UroGen, and Regeneron; and has a patent pending related to this study. Dr. Diaz has been compensated for a leadership role with Epitope; owns stock in Amgen, Delfi Diagnostics, Epitope, Neophore, Seer, and Zydecom Corporation; has held a consultant or advisory role with Neophore, Seer, and Zydecom; holds several patents and/or has patents pending; and has received other compensation from Blackstone and Innovatus Capital Partners.

REFERENCES

1. Cercek A, Foote MB, Shia J, et al: Nonoperative management of mismatch repair deficient tumors. 2025 AACR Annual Meeting. Abstract CT003. Presented April 27, 2025.

2. Cercek A, Foote MB, Rousseau B, et al: Nonoperative management of mismatch repair–deficient tumors. N Engl J Med. April 27, 2025 (early release online).

3. Cercek A, Lumish M, Sinopoli J, et al: PD-1 blockade in mismatch repair–deficient, locally advanced rectal cancer. N Engl J Med 386:2363-2376, 2022.

EXPERT POINT OF VIEW

Michael Overman, MD, Professor of Gastrointestinal Medical Oncology at The University of Texas MD Anderson Cancer Center, noted that neoadjuvant PD-1 blockade in rectal dMMR cancers has been incorporated into treatment guidelines and said the findings reported by Cercek et al bolster support for them.

Michael Overman, MD

“The preferred approach is immune checkpoint therapy for up to 6 months, but there have been some open questions, and I think this study helps to address them,” Dr. Overman said. “With its number of patients and the outcomes, this study sets a pretty clear framework for us of treating patients with a single agent for 6 months. The data are very good for treating 6 months with the single agent; that seems to make sense. And, in regard to assessing clinical response, we use rectal MRIs and endoscopy, and I would propose that tumor-informed ctDNA, as shown in this study, is helpful and should be something we incorporate into our management.”

Dr. Overman also emphasized the importance of determining mismatch repair/microsatellite stability status by immunohistochemistry as part of treatment decision-making, with confirmation by next-generation sequencing. “This is a critical first step in identifying patients” for whom this organ-sparing approach may be appropriate “and should be included in discussion with your surgeon,” he said.

Furthermore, Dr. Overman called the recurrence-free survival outcomes at 2 years—96% for the rectal cohort and 85% for the nonrectal cohort—“both excellent” but acknowledged that 15 months of follow-up for the latter is too short to draw conclusions. The cause of the lower rates of complete clinical response in most nonrectal solid tumor types is not clear, he added, though it may reflect differences in disease biology or alterations in gene pathways. “There are potentially many hypotheses,” he noted.

As for extending neoadjuvant anti–PD-1 blockade and nonoperative management to patients with nonrectal dMMR solid tumors, Dr. Overman commented: “I think this is a valid approach…, with the caveat that this is really up to the patient’s goals and based on a risk/benefit discussion with the patient’s treating physician.” 

DISCLOSURE: Dr. Overman has served as a consultant to Roche, BMS, MedImmune, Merck, Amgen, Takeda, Janssen, Pfizer, Array, Gritstone, 3DMed, Nouscom, Atreca, and Bayer.