Treatment Sequencing in Platinum-Resistant Epithelial Ovarian Cancer After MIRASOL, KEYNOTE-B96, and ROSELLA.

Platinum-resistant epithelial ovarian cancer has long been managed as a clinically defined state with limited therapeutic differentiation. This narrative review aims to clarify how recent Phase III evidence should inform treatment allocation and sequencing after the MIRASOL, KEYNOTE-B96, and ROSELLA trials. Historically, treatment relied on non-platinum single-agent chemotherapy, with bevacizumab added for selected patients after AURELIA; however, this approach did not overcome short progression-free survival, limited response durability, cumulative toxicity, and the absence of robust biological selection. Recent randomized evidence has changed this landscape. MIRASOL established mirvetuximab soravtansine as a folate receptor alpha (FRα)-directed antibody-drug conjugate (ADC) for assay-defined FRα-positive platinum-resistant disease. KEYNOTE-B96 showed that pembrolizumab added to weekly paclitaxel, with or without bevacizumab, improves outcomes in programmed death ligand 1 (PD-L1) combined positive score (CPS)-positive platinum-resistant recurrent epithelial ovarian cancer. ROSELLA demonstrated that relacorilant, a selective glucocorticoid receptor (GR) antagonist, improves the activity of nanoparticle albumin-bound paclitaxel (nab-paclitaxel) in a population selected by clinical criteria rather than by a conventional tumor biomarker. Together, these trials define three distinct evidence anchors: FRα-targeted ADC therapy, PD-L1 CPS-selected chemoimmunotherapy, and tumor-biomarker-independent GR antagonism with nab-paclitaxel. They do not establish a simple hierarchy of preferred regimens. Instead, treatment sequencing should integrate biomarker eligibility, prior bevacizumab and poly (ADP-ribose) polymerase inhibitor exposure, taxane feasibility, ocular and immune-related toxicity risks, corticosteroid requirements, disease tempo, and patient priorities. The central challenge is to convert these validated options into rational, patient-specific sequencing strategies across biomarker-directed and tumor-biomarker-independent treatments.