Single-agent carboplatin produced a confirmed prostate-specific antigen (PSA) decline of at least 50% in 41% of patients with homologous recombination repair (HRR)-altered metastatic castration-resistant prostate cancer (mCRPC) in the phase 2 CIPHER trial. Responses differed across molecular subgroups, while hematologic adverse events were frequent.
The results were published online in ESMO Open on September 23, 2026, in the original article, “Carboplatin in patients with metastatic castration-resistant prostate cancer harboring somatic or germline homologous recombination repair gene mutations: phase II single-arm trial (CIPHER).”
Authors: R. Jain, N. Vasudeva, H. Baskarane, A. Arora, A. Sharma, A. Kumar, K.P. Haresh, S. Kaushal, S.A. Shamim, B. Nayak, R.K. Sahoo, M. Divakar, A. Seth, A. Sharma, S. Bakhshi, and A. Batra.
Why CIPHER Was Conducted
HRR alterations can leave tumor cells vulnerable to treatments that damage DNA. Although poly (ADP-ribose) polymerase (PARP) inhibitors have shown benefit in selected patients with HRR-altered mCRPC, the authors noted that cost limits access to these treatments in many resource-constrained settings. Platinum chemotherapy offers another way to target tumors with impaired DNA repair, but prospective evidence for carboplatin in genomically selected mCRPC has been limited.
CIPHER was designed to assess the antitumor activity and tolerability of carboplatin in men with mCRPC harboring a deleterious somatic or germline HRR alteration. Patients had previously received docetaxel and/or a novel androgen receptor pathway inhibitor. The open-label, single-arm study was conducted at two cancer centers in India.
Study Design and Patients
Eligible patients had progressive metastatic castration-resistant prostate adenocarcinoma, an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1, and adequate bone marrow and renal function. They received intravenous carboplatin at an area under the curve of 5 every 21 days. Treatment continued until disease progression, unacceptable toxicity, or a decision by the investigator or patient to stop.
The primary endpoint was a confirmed PSA50 response, defined as a PSA decline of at least 50% from baseline, verified by a second measurement at least three weeks later without intervening radiographic or clinical progression. Secondary endpoints included radiographic progression-free survival (rPFS), time to objective response, duration of response, safety, and health-related quality of life. Objective response rate was assessed as an exploratory endpoint. Radiographic assessments were performed with prostate-specific membrane antigen (PSMA) PET-CT at baseline and every three months.
Enrollment began in September 2023 and closed in July 2025. Of 213 patients screened, 44 met the genomic eligibility criteria. Five withdrew consent before starting treatment, leaving 39 patients in the efficacy and safety population. The planned enrollment of 49 patients was not reached because of slow accrual during the predefined study period.
Seventeen treated patients had BRCA1 or BRCA2 alterations, nine had ATM alterations, and 13 had other HRR alterations. Sixteen were classified as having germline alterations and 23 as having somatic alterations. The median age was 65 years, and 21 patients had received at least two prior lines of therapy. Patients received a median of four carboplatin cycles.
PSA and Radiographic Responses
A confirmed PSA50 response was observed in 16 of 39 patients (41.0%; 95% confidence interval [CI], 25.6%–57.9%). The median time to PSA50 response was 2.4 months. PSA responses varied by alteration. Two of six patients with BRCA1 alterations (33%) and seven of 11 with BRCA2 alterations (63%) achieved a PSA50 response. Responses were also reported in two of nine patients with ATM alterations (22%) and both patients with PALB2 alterations. Several of these groups contained very few patients, limiting interpretation of the differences between them.
At a median follow-up of 14.71 months, 31 patients were assessable for radiographic response. Ten had a partial response, giving an objective response rate (ORR) of 32.3% among radiographically assessable patients (95% CI, 16.7%–51.4%). Using all 39 treated patients as the denominator, the ORR was 25.6% (95% CI, 13.0%–42.1%). Seven of the 31 assessable patients had stable disease, for a disease control rate of 54.8%.
The median time to objective response was 3.48 months, and the median duration of response was 6.93 months (95% CI, 5.55 months to not reached). Among patients with BRCA1/2 alterations who achieved a partial response, the median duration of response had not been reached at a median follow-up of 15.57 months in the BRCA1/2 subgroup. Five patients remained on treatment at the data cutoff.
Radiographic responses occurred in six of 17 patients with BRCA1/2 alterations (35.3%) and one of nine with ATM alterations (11.1%). Individual patients with RAD50, PALB2, and FANCL alterations also had radiographic responses. No objective responses were observed among patients with CDK12, RAD51D, or MRE11 alterations.
Progression and Exploratory Findings
Median rPFS in the overall cohort was 4.53 months (95% CI, 3.19–7.62). By molecular subgroup, median rPFS was 4.53 months for patients with BRCA1/2 alterations, 3.06 months for those with ATM alterations, and 5.72 months for those with other HRR alterations. The difference between these subgroups was not statistically significant.
In exploratory analyses, visceral metastasis remained associated with shorter rPFS after multivariable adjustment (adjusted hazard ratio, 3.21; 95% CI, 1.23–8.39). Patients previously exposed to a taxane had a lower PSA50 response rate than taxane-naive patients—30.0% versus 52.6%—and a shorter median rPFS of 3.2 versus 6.0 months. The authors advised caution in interpreting these exploratory findings.
An exploratory comparison by alteration origin found PSA50 responses in 10 of 16 patients with germline alterations (62.5%), compared with six of 23 classified as having somatic alterations (26.1%). The difference remained statistically significant after adjustment, but the difference in rPFS did not. Interpretation is also limited by the study’s sequential testing approach: patients enrolled on the basis of a somatic alteration did not subsequently undergo germline testing.
Safety and Quality of Life
Treatment-emergent adverse events occurred in 38 of 39 patients (97.4%), and grade 3 or higher events occurred in 23 (59.0%). Hematologic events were the most frequent. Anemia was reported in 37 patients (94.9%), including 18 (46.2%) with grade 3 or higher anemia. Thrombocytopenia occurred in 31 patients (79.5%), including eight (20.5%) with grade 3 or higher events. Neutropenia occurred in 27 patients (69.2%), including four (10.3%) with grade 3 or higher events. Three patients (7.7%) developed febrile neutropenia.
Fatigue was the most common nonhematologic event, affecting 30 patients (76.9%). Dose delays occurred in 61.5% of patients, dose reductions in 38.5%, and treatment discontinuation because of toxicity in six patients (15.3%). Ten patients died during the study period. Investigators considered four deaths treatment-related: two intracranial hemorrhages, one subdural hematoma, and one exacerbation involving chronic obstructive pulmonary disease and lower respiratory tract infection.
Patient-reported outcomes showed no significant change over time in the Functional Assessment of Cancer Therapy-Prostate total score. Global health status and physical, role, emotional, and cognitive functioning scores on the EORTC questionnaire showed no statistically significant change over time. Social functioning declined by cycle 4, while incontinence aid and hormonal symptom scores increased over time.
What the Results Mean
CIPHER adds prospective evidence of carboplatin activity in HRR-altered mCRPC, with confirmed PSA50 responses in 41% of treated patients. PSA and radiographic responses were more frequent in the BRCA1/2 group than in the ATM group, but the study was too small to establish definitive differences between alterations.
Its single-arm design does not permit a direct comparison with PARP inhibitors or other systemic treatments. Enrollment fell short of the planned sample size, and the molecular subgroups were small and heterogeneous. The study also used PSMA PET-CT to assess progression, which may affect comparisons with trials that used conventional imaging.
The authors concluded that single-agent carboplatin may be a pragmatic option for selected patients with HRR-altered mCRPC, particularly where access to PARP inhibitors is limited. Its observed activity must be considered alongside frequent hematologic toxicity and the four deaths investigators considered treatment-related.
