Predictors of real-world progression-free survival (rwPFS) in patients (pts) with epithelial ovarian cancer (EOC) treated with first-line maintenance (1LM) niraparib (nir): Post hoc analysis of the 1NSPIRE chart review study.
Abstract
e17556 Background: The retrospective, US-based multicenter 1NSPIRE chart review study reported a median rwPFS of 21.7 mo in pts with EOC treated with 1LM nir (N=218). This post hoc analysis of 1NSPIRE used machine learning methods to identify factors associated with rwPFS. Methods: Eligible adults with EOC who initiated 1LM nir (index; nir initiation date) between 01Apr2020–01Apr2023 were followed from index to earliest of death, loss to follow-up, or start of data collection (site-specific initiation visit date). Univariable Cox regression analyses were used to identify characteristics associated with rwPFS. Variables with statistically ( P <0.15) or clinically (hazard ratio <0.5 or >2.0) significant results were first included in multivariable analyses. Variables independently associated with rwPFS were then identified by least absolute shrinkage and selection operator (LASSO) and stepwise selection. Predictors of longer rwPFS were identified by survival tree analysis. Results: Univariable analyses showed that age, BRCA/ homologous recombination deficiency (HRD) status or HRD status alone, type of cytoreductive surgery (CRS), residual disease (RD) status, neoadjuvant chemotherapy (CT), number of cycles/duration of and best response to first-line (1L) CT, cancer antigen 125 (CA-125) level at nir initiation, hypertension, and starting nir dose were associated with rwPFS. LASSO and stepwise selection (Table) identified BRCA -mutated ( BRCA m) status, no RD, and normalized CA-125 level to be significantly associated with rwPFS. Survival tree analysis identified CA-125 level, BRCA /HRD, and RD status as the top predictors of longer rwPFS. Conclusions: BRCA m/HRD status, no RD, and normalized CA-125 level at nir initiation were associated with longer rwPFS among pts with EOC who received 1LM nir. These findings align with clinical expectation but should be interpreted with caution due to small covariate subgroups. Covariates n Hazard ratio a (95% CI) P value ComorbiditiesHypertension 89 1.05 (0.69–1.60) 0.81 BRCA /HRD status BRCA mHRd and BRCA wt/ BRCA unkHRp BRCA wt and HRDunk 22495058 13.28 (1.31–8.20)4.07 (1.64–10.07)2.23 (0.91–5.46) 0.01<0.010.08 Type of CRSPrimary onlyInterval only 77102 11.56 (0.83–2.93) 0.17 HRD status closest to nir startNoneVisibleUnknown 1064132 11.67 (1.03–2.71)1.59 (0.88–2.90) 0.040.13 Duration of 1L CT, mo 179 1.02 (0.81–1.29) 0.87 Best response after 1L CTCompletePartialStable 1193723 11.41 (0.83–2.41)0.79 (0.40–1.58) 0.200.51 CA-125 level at nir startNormalized (≤35 U/mL)Elevated (>35 U/mL) 15821 12.25 (1.23–4.13) <0.01 Analysis cohort (N=179). a Higher numbers indicate higher risk of progression. HRd, homologous recombination deficient; HRp, homologous recombination proficient; nir, niraparib; unk, unknown; wt, wild-type.
Article Details
Journal Info
Journal of Clinical Oncology
Lippincott Williams & Wilkins
Authors (10)
Lisa Michelle Landrum
IU Health Simon Cancer Center, Indianapolis, IN
Premal H. Thaker
Washington University School of Medicine, Siteman Cancer Center, St. Louis, MO
Jonathan Lim
The University of Manchester & The Christie NHS Foundation Trust, Manchester, United Kingdom
Julia Moore
GSK, Upper Providence, PA
John Sampalis
McGill University and PPD, part of Thermo Fisher Scientific, Montreal, QC, Canada
Laura Sayegh
PPD, part of Thermo Fisher Scientific, Montreal, QC, Canada
Jean Hurteau
GSK, Waltham, MA
Amanda Golembesky
GSK, Durham, NC
Bobbie Rimel
University of Washington, Fred Hutchinson Cancer Center, Seattle, WA
Fernanda Musa