JAMA Netw Open. 2026 Jul 1;9(7):e2623166. doi: 10.1001/jamanetworkopen.2026.23166.
ABSTRACT
IMPORTANCE: Pembrolizumab plus chemotherapy, with or without bevacizumab, improves overall survival (OS) in advanced cervical cancer (ACC), but platinum choice and bevacizumab use remain discretionary.
OBJECTIVE: To compare OS by platinum agent and bevacizumab use among patients with ACC receiving first-line pembrolizumab plus chemotherapy in a clinical setting.
DESIGN, SETTING, AND PARTICIPANTS: This comparative effectiveness study used a retrospective multinational cohort with propensity score matching in the TriNetX Global Collaborative Network. Participants included adults with ACC (International Statistical Classification of Diseases, 10th Revision, Clinical Modification, code C53) treated between January 1, 2015, and October 31, 2024. Patients with prior cancer or contraindications to bevacizumab therapy were excluded. Matching was performed on age, race and ethnicity, body mass index, smoking status, and prior chemoradiotherapy. Data were analyzed from November 19 to 28, 2024.
EXPOSURES: Cisplatin or carboplatin, with or without bevacizumab, both combined with first-line pembrolizumab plus paclitaxel.
MAIN OUTCOMES AND MEASURES: Two comparisons were analyzed: cisplatin vs carboplatin and bevacizumab vs no bevacizumab. The primary outcome was OS, defined as time from treatment initiation to death from any cause. Secondary outcomes were adverse events of special interest, including fistula, bowel perforation, and pulmonary embolism.
RESULTS: Among the 1931 patients receiving first-line pembrolizumab with paclitaxel (mean [SD] age, 54.0 [13.2] years), 623 matched pairs of cisplatin-treated patients (n = 719) and carboplatin-treated patients (n = 1212) were included. The mean (SD) age was 51.2 (12.7) years for cisplatin-treated patients and 50.8 (12.7) years for carboplatin-treated patients; the median follow-up was 10.8 (IQR, 3.7-18.0) months and 9.9 (IQR, 2.5-17.2) months, respectively. The median OS was 25.1 (IQR, 9.2 to not reached) months in both groups (HR, 0.98 [95% CI, 0.81-1.18]; P = .35). Bevacizumab-treated patients (n = 803) and those not receiving bevacizumab (n = 667) yielded 455 matched pairs. The mean (SD) age was 53.6 (13.3) years in the bevacizumab-treated group and 54.3 (13.3) years in the group not receiving bevacizumab; the median follow-up was 11.3 (IQR, 4.4-18.2) months and 8.4 (IQR, 1.6-15.1) months, respectively. Twelve-month OS was 74.8% (95% CI, 70.4%-79.2%) vs 64.5% (95% CI, 59.8%-69.2%), respectively; 24-month OS was 56.2% (95% CI, 50.9%-61.5%) vs 54.0% (95% CI, 48.7%-59.3%), respectively. No increased risk of fistula (odds ratio [OR], 0.94 [95% CI, 0.63-1.40]; P = .76), bowel perforation (OR, 1.21 [95% CI, 0.52-2.82]; P = .67), or pulmonary embolism (OR, 0.62 [95% CI, 0.38-1.01]; P = .052) was observed.
CONCLUSIONS AND RELEVANCE: In this comparative effectiveness study of patients with ACC in a clinical setting, cisplatin and carboplatin treatment were associated with similar effectiveness when combined with pembrolizumab. Bevacizumab was associated with early OS benefit, which attenuated over time. These findings warrant prospective validation.
PMID:42446878 | DOI:10.1001/jamanetworkopen.2026.23166
