ROCSAN Study Step I Misses 16-Week ORR End Point in Endometrial or Ovarian Carcinosarcoma


Key Takeaways

  • Niraparib alone or with dostarlimab showed ORRs of 3.8% and 12.0%, respectively, in carcinosarcoma patients post-platinum chemotherapy.
  • Median PFS was less than 3 months for all treatment arms, highlighting the poor prognosis in this patient population.
Isabelle Ray-Coquard, MD, PhD

Isabelle Ray-Coquard, MD, PhD

Niraparib (Zejula) alone or in combination with dostarlimab-gxly (Jemperli) did not result in an objective response rate (ORR) over 20% in patients with endometrial or ovarian carcinosarcoma after at last 1 line of platinum-based chemotherapy, according to preliminary phase 2 results from the phase 2/3 ROCSAN study (NCT03651206).

The data, which were presented during the 2024 ESMO Gynaecological Cancers Congress, showed that the 16-week ORR achieved with niraparib monotherapy (n = 26) was 3.8%. With dostarlimab plus niraparib (n = 25), the 16-week ORR was 12.0%; with chemotherapy (n = 13), the ORR at this time point was 15.4%. In all patients (n = 64), the ORR was 9.4%. Lower 90% confidence interval limits fell below the P0 hypothesis in all arms.

Regarding secondary end points, the best ORRs in the niraparib alone, dostarlimab/niraparib, and chemotherapy arms were 3.8%, 20.0%, and 15.4%, respectively. Moreover, the disease control rates in these respective arms were 26.9%, 52%, and 30.8%. Notably, 73.1% of those given niraparib monotherapy, 48.0% of those given the doublet, and 69.2% of those given chemotherapy experienced disease progression or died before 8 weeks.

Moreover, at a median follow-up of 11.2 months, the median progression-free survival (PFS) with single-agent niraparib was 2.0 months (95% CI, 1.9-2.2), and the 2-month PFS rate was 57.7% (95% CI, 36.8%-73.9%). With dostarlimab plus niraparib, the median PFS was 2.7 months (95% CI, 1.9-3.7) and the 2-month PFS rate was 60.0% (95% CI, 38.4%-76.1%). With chemotherapy, the median PFS was 1.9 months (95% CI, 1.7-3.6), and the PFS rate at 2 months was 38.5% (95% CI, 14.1%-62.8%).

“The ROCSAN study demonstrated the feasibility of a randomized clinical trial in a rare and poor-prognostic population, such as gynecologic carcinosarcoma, adding to a large translational research program supported by a European Commission grant,” Isabelle Ray-Coquard, MD, PhD, of the Centre Leon Berard, in Lyon, France, said during a presentation of the data. “ROCSAN part I study reports acceptable and good safety profile of the combination of dostarlimab and niraparib in the pretreated [patiets with] carcinosarcoma resistant to platinum-based chemotherapy. Despite the low number of patients, this trial confirms the poor prognosis of these patients with a median PFS of less than 3 months in all 3 treatment arms. The first phase did not meet its primary objective regarding the primary end point of ORR at 16 weeks.”

Historically, single-agent chemotherapy in the form of doxorubicin or paclitaxel has been frequently used for patients with carcinosarcoma who received frontline platinum-based chemotherapy. However, this approach led to response rates under 20%, PFS of less than 4 months, and overall survival (OS) under 1 year. Recent data from pivotal trials like KEYNOTE-755 (NCT03517449), RUBY-1 (NCT03981796), NRG GY018 (NCT03914612), and DUO-E (NCT04269200) have shown that combination regimens comprised of PD-1/PD-L1 inhibitors and chemotherapy or lenvatinib (Lenvima) have been efficacious in patients with mismatch repair–proficient endometrial cancer. In patients with endometrial carcinosarcoma, antiangiogenic monotherapy or paired with PD-1 inhibitors did not show activity.

“Based on available data, niraparib and dostarlimab therapies could potentially bring benefit to carcinosarcoma patients post–platinum-based chemotherapy considering the high unmet medical need,” Ray-Coquard said.

The ROCSAN trial enrolled patients with metastatic or recurrent endometrial or ovarian carcinosarcoma who received at least 1 prior line of platinum-based chemotherapy and relapsed within 1 year of their last cycle of chemotherapy. They could not have received more than 4 lines of chemotherapy, and tumor samples were mandatory. In phase 2, participants (n = 63) would be randomly assigned 2:2:1 to receive niraparib monotherapy, dostarlimab plus niraparib, or chemotherapies. For the interim analysis, investigators sought to establish the “best experimental arm” between single-agent niraparib and dostarlimab plus niraparib. Once established, the research will enter phase 3, where participants (n = 133) will be randomized 2:1 to the best arm (niraparib alone or the doublet) vs chemotherapies, and the final analysis will focus on OS and will include those from phase 2.

The data shared during the meeting were from the phase 2 portion of the research. Key stratification factors included number of prior lines of chemotherapy (1 vs 2 to 3), FIGO stage at diagnosis (I to II vs III to IV), type of carcinosarcoma (ovarian vs endometrial), and performance status (0 to 1 vs 2). The primary end point of the study was 4-month ORR per RECIST 1.1 criteria and secondary end points comprised safety and tolerability; best ORR, duration of response, PFS, and PFS2 by RECIST 1.1 criteria; TTST; and patient-reported outcomes.

Ray-Coquard explained that after an internal analysis is performed, an external data monitoring committee could recommend to select the optimal experimental arm for phase 3 evaluation between single-agent niraparib and the doublet, and the recommendation will “be based on the benefit-risk ratio, taking into account the observed response rate and the safety profile of each arm.” For each of the arms, investigators utilized a single-stage design based on Fisher’s exact test to identify phase 2 sample size, “with a 10% unacceptable RR-4M and a 30% targeted RR-4M, assuming a 10% 1-sided alpha and more than 90% power.”

She added that if at least 5 of the 25 patients achieved RR-4M, then the corresponding arm would be “promising.” If neither of the 2 arms were determined to be promising, the study would stop due to lack of efficacy signal. If one of the 2 arms is promising, only that arm will move into phase 3. If both arms are promising, the largest estimated response rate would determine which arm would be selected for phase 3 exploration.

A total of 64 patients were recruited between July 15, 2020, and June 15, 2022. In the chemotherapy arm, patients could have received investigator’s choice of doxorubicin (38.5%), paclitaxel (30.8%), gemcitabine (23.1%), pegylated liposomal doxorubicin (7.7%), or topotecan (0%). In all patients, 29.7% had stage I to II disease at diagnosis, and 70.3% had stage III to IV disease. Most patients had endometrial carcinosarcoma (82.8%) and the remaining 17.2% had ovarian carcinosarcoma. Moreover, 65.6% of patients previously received 1 line of chemotherapy, and 34.4% received 2 to 3 prior lines.

The median patient age at the time of randomization was 70.0 years (range, 34.0-84.0). The best response to prior chemotherapy was complete or partial response for 51.5% of patients and stable or progressive disease for 26.6% of patients; this information was missing for 21.9% of patients. The delay between last chemotherapy line and randomization was a median of 1.4 months (range, 0.3-5.7). About half (51.6%) of patients had at least 1 clinical symptom at inclusion.

Additional data showed that at a median follow-up of 11.2 months, the median OS was 6.7 months (95% CI, 3.8-9.6) with niraparib, 6.5 months (95% CI, 3.9-12.4) with the doublet, and 4.5 months (95% CI, 3.0-not evaluable) with chemotherapy. This was a non-comparative assessment.

“Considering the treatment duration, as you can imagine, it’s very short,” Ray-Coquard said. For niraparib alone, it was a median of 29.5 days (range, 14.0-229.0); for the doublet, it was a median of 70.0 days (range, 7.0-364.0) for niraparib and 84.0 days (range, 42.0-379.0) for dostarlimab; and for chemotherapy, it was 64.0 days (range, 50.0-190.0). At least 1 treatment interruption was needed for 30.8% of those given niraparib monotherapy; 48.0% with niraparib in the combination; and 24.0% for dostarlimab; and 38.5% for chemotherapy.

“The majority of patients stopped treatment for progressive disease. Only 1 patient completed treatment as per protocol, and right now, we have only 3 patients ongoing treatment—1 in the niraparib arm and 2 in the combination arm,” Ray-Coquard said.

All patients experienced at least 1 adverse effect (AE). In the niraparib arm, 61.5% of patients experienced at least 1 AE that was grade 3 or higher; 44.0% with niraparib for the combination, 8% for dostarlimab; and 30.8% for doxorubicin and 15.4% for gemcitabine. Grade 5 events occurred in 7.7% of those in the niraparib monotherapy arm, 4.0% of those in the doublet arm, and 7.7% of those in the chemotherapy arm. No grade 5 AEs were related to study treatment, Ray-Coquard noted.

The main toxicities observed in the niraparib monotherapy, doublet, and chemotherapy arms were anemia (46.2%; 44.0%; 30.8%), thrombocytopenia (23.1%; 24.0%; 0%), nausea (15.4%; 8.0%; 30.8%), and asthenia or fatigue (38.4%; 36.0%; 46.2%).

“Translational research analyses are ongoing, and results are awaited to explain the potential role for this combination in the carcinosarcoma population,” Ray-Coquard concluded. “Alternative therapies are urgently needed for [these] patients.”


Ray-Coquard I, Bellesoeur A, Fabbro M, et al. ENGOT-EN8/ROCSAN: an academic multicentric randomized phase II/III evaluating dostarlimab in combination with niraparib versus niraparib alone compared to chemotherapy in the treatment of endometrial/ovarian carcinosarcoma after at least one line of platinum-based chemotherapy, preliminary results of the phase II part. Presented at: 2024 ESMO Gynaecological Cancers Congress; June 20-22, 2024; Florence, Italy. 34O.

Related Videos
Sheldon M. Feldman, MD
Rita Mukhtar, MD
Lajos Pusztai, MD, DPhil
Hope S. Rugo, MD
Maurie Markman, MD
Karl Semaan, MD, MSc
Bradley McGregor, MD, discusses findings from a phase 1b study of abemaciclib  in clear cell renal cell carcinoma.
Toni K. Choueiri, MD
Neil J. Shah, MBBS
Mary Philip, MD, PhD