Prior readouts from KEYNOTE-B96 showed that the study met its primary end point of progression-free survival (PFS) in both PD-L1–positive and all-comer populations, and its secondary end point of OS in the PD-L1–positive population.2 Second interim analysis results presented during the 2025 ESMO Congress confirmed that pembrolizumab (Keytruda) plus weekly paclitaxel with or without bevacizumab (Avastin) significantly improved PFS regardless of PD-L1 status and improved OS in the PD-L1–positive subgroup. In the intention-to-treat (ITT) population, the median PFS was 8.3 months vs 6.4 months with the pembrolizumab (n = 322) vs placebo (n = 321) regimens, respectively (HR, 0.73; 95% CI, 0.62-0.86).1
“Several factors should be considered when deciding [whether] to use [this regimen or mirvetuximab soravtansine-gynx (Elahere)] first in sequence, but this regimen should certainly be incorporated into our treatment algorithm for patients with PROC,” Colombo emphasized to OncLive®during the meeting.
In her interview, Colombo discussed how unique aspects of the KEYNOTE-B96 trial design set it up for success where other chemoimmunotherapy trials have failed; reported data from both the first and second interim analyses; and discussed where this regimen might fit into the larger treatment armamentarium for platinum-resistant, recurrent ovarian cancer.
Colombo is a medical oncologist and director of the Gynecologic Oncology Program, chair of the Division of Medical Gynaecologic Oncology, and director of the Ovarian Cancer Center at the European Institute of Oncology, IRCCS, and a faculty member in the Department of Medicine and Surgery at the University of Milan-Bicocca in Italy.
OncLive: ICIs have historically shown limited activity in PROC. How does KEYNOTE-B96 build upon or differ from previous studies evaluating PD-1/PD-L1 inhibitors in this population?
Colombo:When we started [KEYNOTE-B96], we decided to use 1 type of chemotherapy, which was weekly paclitaxel. This could be the reason why our study is positive compared with the others, because there was previous evidence that the use of metronomic chemotherapy with pembrolizumab and bevacizumab may enhance the immune response. For that reason, compared with other studies that were investigating different form of chemotherapy, we selected only weekly paclitaxel. [The rationale for weekly paclitaxel use also stemmed from] evidence from the [phase 3 AGO-OVAR 2.29/ENGOT-ov34 study (NCT03353831)]. Overall, that study was not positive, but in the subgroup of patients treated with weekly paclitaxel, investigators could identify a clear trend in favor of the addition of atezolizumab [Tecentriq] to weekly paclitaxel [and bevacizumab]. Therefore, there was both biological rationale and clinical evidence [indicating] that this combination is effective.
How was the KEYNOTE-B96 study designed?
We included patients with platinum-resistant, recurrent ovarian cancer with up to 2 prior lines of chemotherapy and a confirmed diagnosis of epithelial ovarian cancer. Stratification factors [included] the use of bevacizumab, the [tumor] region, and PD-L1 combined positive score [CPS]. Patients were [randomly assigned] to receive either pembrolizumab or placebo with weekly paclitaxel, with or without bevacizumab.
We [enrolled] 643 patients from 187 sites [across] 25 countries, and almost all these patients received treatment.1 At the time of [the second interim] analysis, [91.9% of patients had] discontinued [treatment], mainly due to progression.
We had 2 interim analyses and 1 final analysis. [The primary objectives of] these first 2 analyses were PFS in the CPS of 1 or higher population and in the ITT population. Then we had 2 secondary hypotheses, which were OS in these 2 populations.
What efficacy data have been presented from the KEYNOTE-B96 study?
At the time of the first interim analysis, a clear PFS benefit was demonstrated with pembrolizumab in the CPS of 1 and higher population [HR, 0.72 (95% CI, 0.58-0.89; P = .0014)] and ITT population [HR, 0.70 (95% CI, 0.58-0.84; P < .0001)]. This [benefit] was statistically significant. We reached the primary end point in both populations at the first interim analysis.
The second interim analysis confirmed the data for PFS, but more importantly, demonstrated a clear benefit in OS. That is the most relevant aspect of this study, because this is a difficult-to-treat patient population. To see a 4-month benefit in median OS is a big achievement.
The control arm in our study performed particularly well. We saw a median PFS of 7.2 months [95% CI, 6.2-8.1] in [patients with a] CPS of 1 or higher, which is [long compared with the median PFS seen in] the control arms in other studies. Even the median OS of 14.0 months [from the second interim analysis] is [relatively] long. Despite this strong control arm, we demonstrated a [4.2]-month advantage in median OS [with pembrolizumab].
Were there notable differences in treatment benefit between the PD-L1–positive vs PD-L1–negative subgroups in KEYNOTE-B96?
A PFS benefit [with pembrolizumab] was seen in both [patient] populations, and [a news release] announced that an OS benefit was also seen in the ITT population.2 [Findings from the final analysis] have not yet been reported in detail. [At the 2025 ESMO Congress], I only reported data in the CPS 1 and higher population because, at the time of the second interim analysis, statistical significance was found only in this population.
What should be known about the safety profile of the pembrolizumab regimen in KEYNOTE-B96?
The incidence of grade 3 or higher adverse effects [AEs] or serious AEs [AEs] leading to discontinuation was numerically slightly higher in the pembrolizumab arm. However, we also observed that the median treatment exposure was longer in the pembrolizumab arm, at 33 weeks compared with 28 weeks [in the placebo arm]. That may be the reason why we saw a numerically higher incidence of serious AEs.
If we focus specifically on the immune-mediated AEs, they were, as expected, more frequent in the pembrolizumab arm. However, they were generally low grade and were mainly endocrinopathies, [which were] manageable with dose adjustments or the use of concomitant medication. Overall, there was nothing unexpected and nothing that cannot be managed easily in our daily practice.
What do the KEYNOTE-B96 data suggest about the potential role for pembrolizumab-based regimens in the current treatment paradigm for PROC?
We are now in a world of antibody-drug conjugates [ADCs]. Everybody is looking at ADCs as a new drug that we should use for patients with platinum resistance. They are effective, but nobody expected a positive result from our study. Now we’ll [need to] learn how to integrate this information into our algorithm.
There are several considerations that we can [be mindful of when selecting between the KEYNOTE-B96 regimen and an ADC, including] the number of prior lines of therapy or [a patient’s] positivity for a receptor. For instance, if we are [considering] mirvetuximab soravtansine, high expression of folate receptor alpha [would be required], and only approximately 35% of patients [with PROC meet this criterion].
There are many other ADCs on the horizon and many of them will not need biomarkers. When they become available, it will be difficult to determine the best sequence. We have so many trials now ongoing with different ADCs [containing] the same payload in the same population. We will hopefully have a lot of positive trials soon, but we will not know which ADC to choose and which sequence to follow. In this respect, having an alternative like [the KEYNOTE-B96 regimen] would be of value for our patients, because this is a completely different approach [than ADCs]. It will be easily integrated into the new algorithm and will offer patients more options. We need a lot of options in this population.
References
- Colombo N, Zsiros E, Sebastianelli A, et al. Pembrolizumab vs placebo plus weekly paclitaxel ± bevacizumab in platinum-resistant recurrent ovarian cancer: results from the randomized double-blind phase 3 ENGOT-ov65/KEYNOTE-B96 study. Presented at: European Society of Medical Oncology Congress 2025; October 17-20, 2025; Berlin, Germany. Abstract LBA3.
- Merck announces phase 3 KEYNOTE-B96 trial met secondary endpoint of overall survival (OS) in all comers population of patients with platinum-resistant recurrent ovarian cancer. News release. Merck. October 16, 2025. Accessed November 4, 2025. https://www.merck.com/news/merck-announces-phase-3-keynote-b96-trial-met-secondary-endpoint-of-overall-survival-os-in-all-comers-population-of-patients-with-platinum-resistant-recurrent-ovarian-cancer/
