Maintenance Dostarlimab After Chemoradiation Misses PFS End Point in High-Risk Locally Advanced Cervical Cancer

Maintenance dostarlimab-gxly (Jemperli) following chemoradiation did not result in a statistically significant improvement in progression-free survival (PFS) vs chemoradiation followed by observation in patients with high-risk locally advanced cervical cancer, according to interim data from the phase 2 GEICO 78-C/ATOMICC trial (NCT03833479) that were presented at the 2026 ESMO Gynaecological Cancers Annual Congress.¹

At the time of the analysis, dostarlimab (n = 88) demonstrated a 46% reduction in the risk of disease progression or death vs observation (n = 46; stratified HR, 0.64; 95% CI, 0.31-1.31; 1-sided P = .11). The 12-, 24-, 36-, and 48-month PFS rates in the dostarlimab arm were 82% (95% CI, 74%-90%), 79% (95% CI, 70%-87%), 77% (95% CI, 68%-86%), and 74% (95% CI, 63%-84%), respectively. The respective rates in the observation arm were 80% (95% CI, 68%-92%), 75% (95% CI, 62%-88%), 67% (95% CI, 52%-83%), and 62% (95% CI, 45%-79%).

A largely homogeneous treatment effect in favor of dostarlimab was seen in the subgroup analysis, lead study author, Ana Oaknin, MD, said. A larger treatment effect was observed in patients with positive lymph nodes, but the 95% CI crossed 1 in all the subgroups that were evaluated because of the low number of events.

“In patients with high-risk locally advanced cervical cancer, PFS at the interim analyses exceeded expectations in both arms and numerically favored dostarlimab, although not reaching statistical significance,” Oaknin, head of the Gynaecological Cancer Programme at the Vall d’Hebron Institute of Oncology in Barcelona, Spain, said in a presentation of the data.

High unmet need remains for women with high-risk locally advanced cervical cancer following concurrent chemoradiation, with an approximate 2-year recurrence rate of 50% in this population.

Oaknin noted that when the trial was designed in 2018, chemoradiation was standard for this population, the effect of adding immunotherapy was unknown, and the FIGO 2009 staging system was still valid.

Who were the patients enrolled in the GEICO 78-C/ATOMICC trial?

The open-label randomized, phase 2 trial enrolled patients with high-risk locally advanced cervical cancer who had achieved complete or partial response after standard concurrent chemoradiation with curative intent.^1,2 Prior immune checkpoint inhibition was not allowed, and an ECOG performance status of 0 or 1 was required. Specific staging requirements mandated patients who were FIGO 2009 stage IB2, IIA2, or IIB and pelvic lymph node positive; FIGO 2009 stage IIIA, IIIB, or IVA; or any FIGO 2009 stage and para-aortic lymph node positive.

Chemoradiation must have been completed within 12 weeks before consenting to treatment.

Patients were randomly assigned 2:1 to maintenance dostarlimab at 500 mg every 3 weeks for 4 cycles, equating to 1000 mg every 6 weeks for up to 2 years in total, or observation.¹

The primary end point was PFS per RECIST 1.1 criteria. Secondary end points included overall survival (OS), patient-reported outcomes, and safety.

Patients were stratified by histology (squamous vs adenosquamous/adenocarcinoma), response to chemoradiation (complete vs partial response), FIGO 2009 stage/nodal status (IB2, IIA2, or IIB and pelvic lymph node positive vs IIIA, IIIB, or IVA vs any FIGO 2009 stage and para-aortic lymph node positive).

How was the GEICO 78-C/ATOMICC trial designed?

The trial was designed with the assumption that 2-year PFS would be 45% in the control arm and 59% in the dostarlimab arm, corresponding to a HR of 0.66. Accounting for 10% dropout and a 1:2 randomization ratio, 132 evaluable patients were required. The prespecified final analysis after 74 PFS events was projected to provide 80% power to detect superiority of the experimental arm at a 1-sided stratified log-rank alpha of 0.2. The first patient was enrolled on August 14, 2019, and the last patient on June 27, 2023. A protocol amendment was subsequently implemented to conduct an interim analysis after at least 32 PFS events, which occurred in 2025. The final analysis is planned after 5 years of follow-up or 74 PFS events, whichever occurs first, with completion anticipated in 2028.

The 134 randomized patients were enrolled across 22 Spanish sites. In the observation arm (n = 46), median age was 53 years (range, 28-80), and in the dostarlimab arm (n = 88), median age was 48 years (range, 24-78). Most patients in both arms had an ECOG performance status of 0 (78% vs 77%, respectively) and squamous histology (83% vs 81%). By FIGO 2009 stage, most patients were stage IB to IIB (54% vs 65%), with smaller proportions having stage III (35% vs 33%) or stage IVA disease (11% vs 2%). Pelvic lymph node involvement was present in 67% and 69% of patients, respectively, and para-aortic lymph node involvement in 30% and 26%.

Regarding response to prior chemoradiation, 67% and 70% of patients achieved a complete response, while 33% and 30% achieved a partial response. Three-dimensional high-risk clinical target volume brachytherapy was administered to 76% and 81% of patients, respectively. Median cisplatin duration was 4.6 weeks (IQR, 4.1-5.3) in the observation arm and 4.3 weeks (IQR, 4.1-5.1) in the dostarlimab arm.

A total of 172 patients were screened, of whom 134 were randomized. Among the 38 screen failures, the most common reasons for exclusion were ineligibility (n = 20), withdrawal of consent (n = 15), physician decision (n = 2), and initiation of another therapy (n = 1). Randomized patients were assigned to either observation (n = 46) or dostarlimab (n = 88). Of those assigned to dostarlimab, 87 received at least 1 dose; 1 patient withdrew consent prior to treatment.

At the September 17, 2025, data cutoff, with a median follow-up of 40.6 months, 87 patients had discontinued dostarlimab treatment, including those who completed 2 years of therapy (n = 49), experienced disease progression (n = 16), experienced toxicity-related events (n = 13), withdrew consent (n = 5), discontinued per physician decision (n = 2), were lost to follow-up (n = 1), or died (n = 1). Ten patients in the dostarlimab arm and 6 in the observation arm discontinued from the study; 78 and 40 patients, respectively, remained in follow-up at the time of data cutoff.

What did the OS and safety data show with dostarlimab maintenance therapy?

A preliminary look at OS with only 13% (n = 18 of 134) of events having occurred at the time of analysis revealed an HR of 0.94 (95% CI, 0.34-2.58). The 12-, 24-, 36-, and 48-month OS rates in the dostarlimab arm were 98% (95% CI, 94%-100%), 90% (95% CI, 84%-97%), 86% (95% CI, 78%-94%), and 83% (95% CI, 74%-92%), respectively. The respective rates in the observation arm were 93% (95% CI, 86%-100%), 91% (95% CI, 82%-99%), 88% (95% CI, 77%-98%), and 88% (95% CI, 77%-98%).

With respect to treatment exposure and safety, the median duration of dostarlimab therapy was 24.7 months (IQR, 8.4-25.1). Adverse effects (AEs) occurred in 97% (n = 84) and 96% (n = 44) of patients in the dostarlimab and observation arms, respectively; 85% (n = 74) were treatment related in the dostarlimab arm. Grade 3 or greater AEs were reported in 36% (n = 31) and 24% (n = 11) of patients in the dostarlimab and observation arm; 18% (n = 16) were treatment related in the dostarlimab arm. Only 1 grade 5 AE occurred in the dostarlimab arm.

Dose delays occurred in 45% (n = 39) of patients in the dostarlimab arm, owing to non-hematologic toxicity (n = 30; 34%), hematologic toxicity (n = 5; 6%), and infusion-related reaction (n = 1; 1%).

“Longer follow-up is needed to determine the final outcomes and clarify the potential benefit from dostarlimab as maintenance therapy for high-risk locally advanced cervical cancer,” Oaknin concluded.

Disclosures: Oaknin disclosed personal fees for advisory boards (AbbVie, Agenus, AstraZeneca, BioNTech, BMS, Clovis Oncology, Corcept Therapeutics, Daiichi Sankyo, Debiopharm International, Deciphera Pharmaceuticals, Eisai, Exelixis, F. Hoffmann-La Roche, Genmab, GSK, Immunogen, Ipsen, Itheos, Medison Pharma Trading AG, Mersana Therapeutics, MSD, Myriad Genetics, Novocure, OncXerna Therapeutics Inc, Ottimo, Pharma&, PharmaMar, Regeneron, Seagen/Pfizer, Shattuck Labs, Stemline Therapeutics, Sutro Biopharma, TORL Therapeutics, Zentalis, Zymeworks); travel/accommodation (AstraZeneca, PharmaMar, Roche); research funding to institution (AbbVie Deutschland, Advaxis Inc, Aeterna Zentaris, Amgen, Aprea Therapeutics AB, BMS, Clovis Oncology Inc, Eisai, F. Hoffmann-La Roche, Immunogen, MSD de España, Millennium Pharmaceuticals, PharmaMar, Regeneron Pharmaceuticals, Tesaro); and non-financial interests: leadership roles for ESMO and GCIG; GYN track chair for ESMO 2019; member of ASCO, ESMO, GCIG, GOG and SEOM.

References

1. Oaknin A, Estevez Garcia P, Rubio Perez MJ, et al. Maintenance dostarlimab after chemoradiation for patients with high-risk locally advanced cervical cancer: GEICO 78-C/ATOMICC randomized phase 2 trial. Presented at: 2026 ESMO Gynaecological Cancers Congress; June 17-19, 2026; Copenhagen, Denmark. Abstract 26O.
2. TSR-042 as maintenance therapy for patients with high-risk locally advanced cervical cancer after chemo-radiation (ATOMICC). ClinicalTrials.gov. Updated June 17, 2026. Accessed June 22, 2026. https://clinicaltrials.gov/study/NCT03833479