David E. Cohn

Background: High-grade serous ovarian carcinoma (HGSOC) is the most common and lethal subtype of epithelial ovarian cancer and is frequently diagnosed at advanced stages. Because currently available blood-based biomarkers have limited performance in early-stage disease, there is a need to identify circulating biomarker candidates associated with early-stage HGSOC. In this retrospective multi-institutional case–control study, we evaluated whether serum extracellular vesicle (EV)-associated protein signatures distinguish early-stage HGSOC from healthy controls. Methods: Serum samples (n = 252) were obtained retrospectively from multiple institutions and included healthy controls and patients with early- and advanced-stage HGSOC. EV-associated proteins were profiled using liquid chromatography–tandem mass spectrometry (LC–MS/MS) and proximity extension assay (PEA) to identify candidate proteins enriched in early-stage HGSOC. Selected candidates were evaluated by enzyme-linked immunosorbent assay (ELISA), and tissue-level expression was examined in early-stage HGSOC specimens. A multimarker combination model was generated using a smoothed empirical estimate of hyper-volume under the manifold (SHUM) approach and internally assessed by leave-one-out cross-validation. Results: Ten EV-associated serum proteins were prioritized on the basis of differential expression and fold change and were confirmed to be expressed in early-stage HGSOC tissues. In ELISA-based analyses, the combined 10-protein EV panel distinguished early-stage HGSOC from healthy controls with an area under the curve (AUC) of 0.99 in the study dataset, whereas MUC16 (CA-125) showed substantially lower performance in this comparison. The SHUM-based model yielded a true-positive rate of 0.971, a false-positive rate of 0.057, and a Matthews correlation coefficient of 0.915 in the analyzed cohort. Several candidate proteins were differentially enriched in EV fractions but not in matched whole serum. Conclusions: Serum EV-associated proteins are altered in early-stage HGSOC and define a multi-protein signature associated with this disease state in a retrospective case–control setting. These findings support further evaluation of EV-based biomarker candidates in clinically representative and prospectively collected cohorts that include benign gynecologic conditions, symptomatic patients, and pre-diagnostic samples.

Figure S2: Figure 7: Mean ± SD concentration versus time profiles of AR (ADC), total antibody, DM4, and DM4-Me after administration of AR at 2.2 mg/kg IV x 1 plus Bev 10mg/kg IV.

Representativeness of Study Participants

Treatment-related AEs (Any grade)

A randomized phase II study of bevacizumab with weekly anetumab ravtansine or weekly paclitaxel in platinum-resistant or refractory high grade ovarian cancer NCI trial#10150.

Figure S3. Relationship between anetumab ravtansine AUC0-inf (A) and DM4 AUC0-168h (B) and best antitumor response.

: Pharmacokinetic parameters of AR after administration of 2.2 mg/kg AR at 2.2 mg/kg IV x 1 plus Bev 10mg/kg IV.

Purpose: Mesothelin (MSLN) is highly expressed in high-grade serous/endometrioid ovarian cancers (HGOC). Anetumab ravtansine (AR) is an antibody–drug conjugate directed at the MSLN antigen with a tubulin polymerization inhibitor. We assessed the safety, activity, and pharmacokinetics of the combination AR/bevacizumab (Bev; ARB) versus weekly paclitaxel/Bev (PB) in patients with platinum-resistant/refractory HGOC (prrHGOC). Patients and Methods: Following a run-in phase I study to assess ARB safety, patients with prrHGOC with centrally confirmed MSLN-positive expression were randomized to ARB or PB (weekly paclitaxel 80 mg/m2 with Bev 10 mg/kg biweekly). Patients were stratified by platinum resistance/refractory and prior Bev. The primary endpoint was progression-free survival (PFS), and secondary endpoints were overall response rate, safety, and blood-based angiome biomarker assessment. A futility analysis was planned after 35 PFS events. Results: The combination of Bev (10 mg/kg) biweekly with AR (2.2 mg/kg) weekly was well tolerated. About phase II results, MSLN positivity was 88%, and 57 patients were randomized (28 ARB and 29 PB). Forty-two percentage of patients received prior Bev, and 23% were platinum-refractory. At futility analysis, the median PFS was 5.3 and 12.7 months for ARB and PB, respectively [P = 0.03; HR = 2.02 (1.06–3.86)]. The overall response rate was 21% with ARB and 65% with PB. The most common treatment-related grade ≥3 adverse events were anemia (18%) with ARB and neutropenia (24%) with PB. Higher baseline levels of circulating IL6 were associated with worse PFS, and its levels decreased with PB treatment. Conclusions: Our study stopped at interim analysis highlighting the benefit of PB in prrHGOC as the standard of care.