Immune-Related End Points Consistent With Traditional Measurements of Dostarlimab in Non-Endometrial dMMR Solid Tumors

Dostarlimab achieved durable antitumor responses in patients with non-endometrial mismatch repair–deficient solid tumors treated in an expansion cohort of the phase 1 GARNET trial, according to interim findings presented during the ESMO World Congress on Gastrointestinal Cancer.

Dostarlimab (Jemperli), a monoclonal antibody, achieved durable antitumor responses in patients with non-endometrial mismatch repair–deficient (dMMR) solid tumors treated in an expansion cohort of the phase 1 GARNET trial, according to interim findings presented during the ESMO World Congress on Gastrointestinal Cancer 2021.1

The data show that dostarlimab also has a manageable safety profile when used as treatment of non-endometrial dMMR solid tumors with various histologies. The efficacy results were also consistent with results of the primary end point of objective response rate (ORR) by RECIST v1.1 per blinded independent review committee (BIRC), a measurement that does not require confirmation of progressive disease (PD).

Dostarlimab is FDA approved for the treatment of adult patients with dMMR recurrent or advanced endometrial cancer that has progressed on or following prior treatment with a platinum-containing regimen. The drug also has approval from the European Medicines agency for this indication.2 These approvals were granted based on primary results from the GARNET study. Although the drug can currently be used to treat patients in real-world settings, there are questions about whether RECIST v1.1 criteria undervalues the clinical benefit of immunotherapies.1

GARNET is a dose-escalation and cohort expansion study to assess the efficacy and safety of dostarlimab in patients with advanced solid tumors. The study is conducted in 3 parts to find the appropriate dose of the drug (Part 1), execute a safety run-in analysis (Part 2A), and explore the efficacy and safety in an expansion cohort (Part 2B). In Part 2B, 5 different cohorts were evaluated included patients with dMMR endometrial cancer (cohort A1), those with MMR proficient endometrial tumors (cohort A2), patients with non–small cell lung cancer (cohort E), patients with non-endometrial dMMR, or POLε-mutated pan solid tumors (cohort F), and those with advanced, relapsed, high-grade serous, endometrioid, or clear cell ovarian, fallopian tube, or primary peritoneal cancer (cohort G).

The interim date presentation given by Thierry Andre during the World Congress on Gastrointestinal Cancer 2021 focused on cohort F, whose data cutoff date was March 1, 2020. The cohort was assessed for the secondary end points of the GARNET study which include immune-related (ir) ORR, ir duration of response (DOR), and ir disease control rate (DCR). The end points were all assessed per investigator assessment by irRECIST to account for the potential shortcomings of the RECIST v1.1 criteria. It was hypothesized that using irRECIST might avoid early treatment discontinuation in the event of pseudoprogression.

Patients in cohort F were required to have PD following systemic therapy and have no satisfactory alternative treatment options as confirmed by 2 scans submitted prior to the first dose of dostarlimab. Patients enrolled to the cohort who had colorectal cancer must have also had PD or presented intolerance to fluoropyrimidine, oxaliplatin, and irinotecan. All patients included in cohort F were required to be naïve to treatment with PD-L1 inhibitor therapy.

Screening for the cohort was conducted based on results from g immunohistochemistry (IHC), polymerase chain reaction, or next-generation sequencing tests. Eligibility to treatment with dostarlimab in this cohort analysis was determined by MMR IHC results.

A total of 106 met the requirements to be assessed in cohort F. Baseline characteristics showed that the median age of the group was 61.5 years (range 24-85 years). The population was 55% male and 45% female. The majority of the patients had a baseline ECOG performance status of 1 (60%) and the remaining 40% had a score of 1.

Disease characteristics screened at baseline showed that 65% of the cohort had colorectal cancer, 11% had cancer of the small intestine, 8% had gastric and gastroesophageal junction cancer, 4% had pancreatic carcinoma, 2% had ovarian cancer, and 2% had hepatocellular carcinoma. The remaining 8% of patients had other tumor types. The disease stage was metastatic for nearly all of the patients included in the analysis (97%), and the remaining 3% had locally advanced disease.

In terms of prior therapy, 45% of patients had 2 prior lines of treatment. Twenty-four percent received 1 prior line of therapy, and 24% received 3 prior lines of therapy. Only 8% of patients received 4 or more prior lines of therapy. All patients in the cohort received prior chemotherapy, and 87% had previously undergone surgery. Twenty percent of the cohort received prior radiotherapy.

Median follow-up for the analysis was 12.4 months. Patients received 500 mg of dostarlimab once every 3 weeks from cycle 1 to 4 followed by dostarlimanb 1000 mg once every 6 weeks from cycle 5 through 6. Treatment continued until disease progression or unacceptable toxicity.

The results showed an irORR of 45.0% based on the investigator's assessment by irRECIST. The result was an improvement compared with the RECIST v1.1 by BICR data of 38.7%. Notably. irORR per irRECIST by investigator assessment include ir complete response (CR) in 5.5% of patients, ir partial responses (PRs) in 39.4%, and ir stable disease (SD) in 23.9% of patients. About 23% of these patients had PD, and 6.4% of the arm were not evaluable for response and the irDCR was 69.7%. The median for those assessed by investigators per irRECIST was not reached (95% CI, 1.74+ to 21.88+ months).

In comparison, CRs were observed in 7.5% of patients who were assessed by BICR per RECIST v1.1. PRs were observed in 31.1% of patients, 24.5% had SD, and 30.2% had PD, and 6.6% were not evaluable for response. The DCR in this group was 63.2%. The median DOR in this RECIST v1.1 group was also not reached (95% CI, 1.74+ to 21.88+ months).

Thierry noted during his presentation that responses to dostarlimab were observed across all the tumor types evaluated. Further, the responses were durable with 87.8% of responders having remained in response as of the data cutoff date.

The safety portion of the analysis included 173 patients with non-endometrial dMMR solid tumors who received at least 1 dose of dostarlimab. Any-grade treatment-emergent adverse events (AEs) were seen in 96.5% of the population. Treatment-related AEs (TRAEs) of any grade occurred in 68.8% of patients.Grade ≥3TEAEs were observed in 49.1% of those assessed and grade ≥3 TRAEs were seen in 11.6%.Treatment-related serious AEs occurring in 7.5%.

Of the TRAE observed in the study patients, 4.6% of them led to discontinuation of treatment, and 1.25 led to death.

Dostarlimab did not show any new safety signals during the analysis and the majority of the TEAEs observed were low grade and considered to be manageable.


1. Andre T, Berton D, Curigliano G, et al. Analysis of the immune-related (ir) endpoints of the mismatch repair–deficient (dMMR) non-endometrial solid cancers cohort from the GARNET study. Presented at: ESMO World Congress on Gastrointestinal Cancer 2021; June 30 -July 3, 2021. Abstract SO-9.

2. FDA grants accelerated approval to dostarlimab-gxly for dMMR endometrial cancer. News release. FDA. April 22, 2021. Accessed July 1, 2021.