A Randomized Phase II Trial of Cabozantinib and Cabozantinib Plus Nivolumab Versus Standard Chemotherapy in Patients with Previously Treated Non-Squamous NSCLC
Categories (click each to see list of all clinical trials associated with that category): Thoracic (ONC)
Current Status: Open
Phase: II
Principal Investigator: Sharma, Bhavina
Eligibility: https://www.clinicaltrials.gov/study/NCT04310007?term=%09NCT04310007&rank=1
Summary
Primary Outcome:
1. PFS for pt pop. w/ NSCLC [Time Frame: After 58 events (From time of randomization to documented DZ progression [site review of imaging] or death from any cause, whichever occurs first)]
The final analysis of PFS will be performed when the full info of PFS is reached (i.e., 58 events for each of the 2 primary comparisons). The 2 primary comparisons of PFS will each use a logrank test stratified on the RSF w/ a one-sided type I error rate of 10%. Est of PFS including medians and CI, will be calculated using the KM method. CPhM, stratified, on the same fact, will be used to est the Tx HR. In the event that both primary PFS comparisons of the combination of nivolumab w/ cabozantinib and cabozantinib alone versus the standard chemotherapy are statistically significant, PFS will be compared between the 2 experimental arms (combination of w/w/ cabozantinib versus vs cabozantinib alone), using a stratified log rank test w/ a one-sided type I error rate of 10%.
Secondary Outcome:
1. OS for each arm [Time Frame: From time of randomization to documented DZ progression (site review of imaging) or death from any cause, whichever occurs first, assessed up to 3 years]
The final analysis of OS will be performed when the full info of PFS is reached (i.e., 58 events for each of the 2 primary comparisons). OS will be tested in a hierarchical fashion if the primary comparison of PFS between Tx arms is statistically significant, using a log rank test stratified on the RSF w/ a one-sided type I error rate of 10%. Est of OS, including medians and CI, will be calculated using the KM method. CPhM, stratified, on the same fact, will be used to est the Tx HR. OS will be further tested in a hierarchical fashion if the comparison of PFS between the 2 experimental arms is statistically significant. Point est of all EP will be accompanied by the corresponding 2-sided 80% CI. Subset analyses of PFS and OS by Tx arm will be est and compared w/in subsets.
2. Best objective response for each arm [Time Frame: Every 6 weeks for the first year on study, and then every 12 weeks after year one]
Will be evaluated via Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 criteria. Response rates (complete response [CR] + partial response [PR]) and toxicity will be compared using Fisher's exact tests w/ a one-sided type I error rate of 10%; multivariable logistic regression modeling will be used to adjust for the effect of any covariates that are associated w/ these categorical outcomes. Modeling proc will impl backward selection; variables significant at the 0.10 level in the univariate setting will be chosen for inclusion in an initial full model, and at each step the least significant variable will be removed from the model. Only those covariates w/ p < 0.05 will remain in any final models, unless there are fact ID by the study team as crucial to model int. Point est of all EP will be accompanied by the corresponding 2-sided 80% CI.
3. Toxicity profile of monotherapy w/ cabozantinib, and the combination of nivolumab and cabozantinib [Time Frame: Up to 3 years after completion of study Tx]
Will be determined using the Common Terminology Criteria for Adverse Events (CTCAE) criteria. Response rates CR + pathologic response [PR]) and toxicity will be compared using Fisher's exact tests w/ a one-sided type I error rate of 10%; multivariable logistic regression modeling will be used to adjust for the effect of any covariates that are associated w/ these categorical outcomes. Modeling proc will impl backward selection; variables significant at the 0.10 level in the univariate setting will be chosen for inclusion in an initial full model, and at each step the least significant variable will be removed from the model. Only those covariates w/ p < 0.05 will remain in any final models, unless there are fact ID by the study team as crucial to model int. Point est of all EP will be accompanied by the corresponding 2-sided 80% CI.