BOT/BAL in recurrent ovarian cancer: survival estimates and selection limits

October 3, 2026: extended survival follow-up

The IGCS poster reports an ovarian-cancer cohort from the Phase 1b C-800-01 program, using a December 13, 2025 data cutoff. The headline 48% three-year survival is a Kaplan–Meier estimate among 35 efficacy-evaluable patients, not a raw proportion alive among 44 treated patients. This report preserves those different denominators and the older cutoff. It is an extended analysis of an existing cohort, not a new randomized ovarian-cancer trial or an approval. [1] [2]

Fc-enhanced CTLA-4 blockade with PD-1 inhibition

Botensilimab is an Fc-enhanced anti-CTLA-4 antibody intended to improve immune priming and Fc-receptor engagement, including effects on suppressive regulatory T cells within tumors. Balstilimab blocks PD-1 signaling. The combination seeks complementary activity in tumors that have responded poorly to conventional checkpoint monotherapy. This mechanism does not demonstrate that every ovarian tumor acquires an effective antitumor immune response; antigenicity, immune exclusion and prior treatment can still constrain activity. [2]

BOT was administered at 1 or 2 mg/kg every six weeks with BAL 3 mg/kg every two weeks. Median delivered doses were one for BOT, range 1–11, and three for BAL, range 1–50. These exposure counts help explain why durable effects after limited treatment are of interest, but they do not establish a minimum effective dose or that early discontinuation is beneficial. The poster supplies no quantitative receptor occupancy, exposure–response model or randomized contribution-of-components analysis.

Population, dose cohorts and the nine-patient gap

NCT03860272 enrolled patients with progression after available standard therapy or no appropriate standard option. The ovarian cohort included 44 treated women, median age 66 years, range 37–79; 22 had ECOG performance status 1. Median prior lines were four, range 1–17, and 29 had at least four. All had received platinum, 34 bevacizumab, 25 a PARP inhibitor, six a PD-(L)1 inhibitor and two mirvetuximab. Twelve were platinum-sensitive and 32 platinum-resistant or refractory at baseline. [2] [4]

Twenty-one received BOT 1 mg/kg and 22 received 2 mg/kg; one patient crossed from BOT monotherapy to combination treatment, with the combination baseline used for the analysis. Efficacy required at least one post-baseline six-week scan and included 35 patients. Excluding nine treated patients can select those who survived or remained evaluable long enough for assessment. The poster does not fully resolve why each was excluded or supply an intention-to-treat sensitivity analysis. There was no randomization, blinded control or prespecified confirmatory multiplicity hierarchy.

Response and survival: the original poster values

Eight of 35 patients responded: one complete and seven partial responses, ORR 23%, 95% CI 10–40. Median DOR was 9.7 months, CI 2.8–not estimable. Disease control at six weeks was 21/35, or 60%, CI 42–76, and clinical benefit at 24 weeks was 11/35, or 31%, CI 17–49. These thresholds describe different durations and must not be presented as interchangeable response rates. [2]

Median PFS was 2.8 months, CI 1.4–5.5, while median OS was 14.8 months, CI 12.1–not estimable. Estimated 24- and 36-month OS were both 48%, CI 27–65. Median observed follow-up was 10.9 months, range 1.3–38.7; reverse-Kaplan–Meier follow-up was 25.2 months, CI 10.9–33.4. The difference reflects distinct methods, not inconsistent patient histories. A survival tail with wide confidence limits is compatible with meaningful benefit in some patients, but without a control it cannot determine the treatment-attributable survival gain.

Subgroups are overlapping exploratory observations

Platinum-sensitive patients had ORR 3/10, 30%, CI 7–65, and platinum-resistant/refractory patients 5/25, 20%, CI 7–41. Estimated three-year OS was 45%, CI 11–75, and 47%, CI 23–68, respectively. Among 23 with at least four prior lines, ORR was 4/23, 17%, CI 5–39. Among 17 who had progressed on a PARP inhibitor, ORR was 2/17, 12%, CI 2–36, with estimated three-year OS 61%, CI 29–82. The wide intervals and overlapping membership preclude ranking these categories as independently validated sensitive populations. [2]

Eleven of 44 treated patients were alive and off treatment at last follow-up, including six responders. That 25% raw descriptive proportion has a different denominator and time definition from the 48% three-year survival estimate. Neither is an appropriate external comparator for a different trial. Later ovarian disclosures from the same study should be appended as new dated analyses, not counted as independent replications of this cohort.

Immune-mediated toxicity is clinically consequential

The poster reports treatment-related immune-mediated adverse events, not a complete all-cause TEAE inventory. Any such event occurred in 30/44, 68%, and grade 3 or higher events in 14/44, 32%. Diarrhea/colitis occurred in 19/44, 43%, including nine grade 3 or higher cases, 20%. No treatment-related deaths were reported. This cannot be expanded into a claim of no deaths from any cause or no serious adverse events. [2]

Immune gastrointestinal toxicity can require treatment interruption and immunosuppression, and may affect dose exposure and interpretability of outcomes. Exact all-cause SAE, AE-discontinuation and steroid-use counts were not supplied in the poster table. Safety comparisons with historical checkpoint regimens would require the same grading rules, follow-up and denominator. The original poster has been visually checked so that its safety categories are not silently relabeled as all TEAEs.

Funding and program distinction

Agenus develops both antibodies. Its July 2026 private placement raised approximately $85 million gross; possible additional warrant proceeds are contingent and must not be counted as cash already available. The company is also running ROBBIN in neoadjuvant microsatellite-stable colon cancer. That Phase 3 program is a different indication, treatment context and evidence chain; it does not make this ovarian cohort Phase 3. Manufacturing arrangements and funding support execution but cannot resolve the ovarian cohort’s lack of randomized control or incomplete safety reporting. [3]

Sources / References

1. Agenus · IGCS announcement · October 3, 2026
2. Porter et al. · original IGCS 2026 poster 969
3. Agenus · Q2 2026 results and subsequent financing
4. ClinicalTrials.gov · C-800-01