Testing the Use of the Combination of Selumetinib and Olaparib or Selumetinib Alone Targeted Treatment for RAS Pathway Mutant Recurrent or Persistent Ovarian and Endometrial Cancers, A ComboMATCH Treatment Trial

Progression-free survival

Disease progression will be defined using Response Evaluation Criteria in Solid Tumors (RECIST) version (v)1.1 criteria, as determined by the treating physician. The primary analyses will be based on logrank tests stratified by the stratification factors as recorded at randomization. All enrolled patients will be included, regardless of compliance with their assigned study regimen. Patients will be grouped by their randomized treatment for intention-to-treat analyses. Treatment hazard ratios and 90% confidence intervals will be estimated using proportional hazards models specified with a main-effect for the randomized treatment assignment (experimental versus \[vs\] reference), and stratified using the stratification factors recorded at randomization. Treatment group differences will be graphed using Kaplan-Meier methods.

Incidence of adverse events (AE)

The safety population will support these analyses. The nature, frequency, and degree of toxicity will be tabulated at the System Organ Class and AE-specific term levels using Common Terminology Criteria for Adverse Events v5.0. Each patient will be represented according to the maximum grade observed for each term. Tabulations will show the number and percentage of patients by maximum grade, within the treatment group received, regardless of the randomized treatment assignment.

Objective response rate (ORR) between two arms

ORR is defined as the binomial proportion of evaluable patients with a best overall response of complete response (CR) or partial response (PR) (by RECIST 1.1) within 6 months of the date of the last enrollment. Responses reported by the treating physician will be used for these analyses. The ORR estimates by treatment arm will be supported by their 2-sided, 95% Wilson-Score confidence intervals. The relative odds of response in the experimental arm (vs the reference arm) will be estimated using a multivariable logistic regression model specified with main effects for the randomized treatment assignment and covariate adjustments for the stratification factors reported at baseline.

ORR in crossover patients

Determined by RECIST 1.1. The crossover population will support these analyses. The same evaluation criteria for ORR in the measurable disease population will be applied to the crossover population. These analyses will be done separately for each cohort.

Duration of response of both arms

These analyses will be supported by patients in the measurable disease population who have a best overall response of PR or CR. Treatment group differences in response duration will be graphed using Kaplan-Meier methods and compared using logrank tests, stratified by the stratification factors defined at randomization. The relative hazards of progression or death in the experimental group (vs the reference group) will be estimated using a multivariable proportional hazards regression model specified with main effects for the treatment indicators and covariate adjustments for the stratification factors reported at baseline. These analyses will be done separately for each cohort.

ORR

Defined as the binomial proportion of patients with a best overall response of CR or PR. Defined using RECIST 1.1 criteria. Roughly 10 genomic and transcriptomic measures are expected to present as dichotomous factors. When 3 or more levels are presented, the relatively small sample sizes will likely require collapsing down to 2 levels. If presented as continuous variables, the median of the observed values will classify evaluable patients as having wild-type (WT) or mutated expression.

Resistance to therapy

Defined as the proportion of patients with best overall response of disease progression. Defined using RECIST 1.1 criteria. Roughly 10 genomic and transcriptomic measures are expected to present as dichotomous factors. When 3 or more levels are presented, the relatively small sample sizes will likely require collapsing down to 2 levels. If presented as continuous variables, the median of the observed values will classify evaluable patients as having WT or mutated expression.

Concordance between the diagnostic tumor mutation profile, biopsy mutations, and pre-treatment circulating tumor deoxyribonucleic acid (ctDNA) mutations

Assess the concordance of the diagnostic tumor mutation profile generated by the designated laboratory, the pre-treatment biopsy mutation profile, and the pre-treatment ctDNA mutation profile.