Michal R. Tomaszewski, Ph.D. from H. Lee Moffitt Cancer Center and Research Institute speaks about ASCO 2021 Abstract – Imaging-based patient inclusion model for clinical trial performance optimization.
Link to Abstract:
https://meetinglibrary.asco.org/record/196629/abstract
Background information:
The development of novel cancer medicines in the era of precision medicine is heavily reliant on the appropriate identification of target patient populations. We believe that computational imaging data analysis may be utilized to produce a quantitative population enrichment method in clinical trials, and we want to provide a framework for this analysis.
Methodologies:
This concept was investigated in patients with soft-tissue sarcoma (STS) who were enrolled in a randomized Phase III clinical study (SARC021) to see how effective evofosfamide (Evo), a hypoxia triggered prodrug, was in conjunction with doxorubicin (Dox). SARC021 was notable for failing to reach its survival goal (PMC7771354). We looked at whether a radiomic analysis and relevant clinical covariates-based inclusion/exclusion model may have resulted in a substantial therapeutic advantage of the Evo+Dox combination over standard Dox monotherapy. A total of 163 radiomics characteristics were collected from 303 patients’ lung metastases in the SARC021 experiment, which were separated into demographically matched training and test sets. The most repeatable traits were discovered using a stability study. To distinguish between the two treatment groups, univariable and multivariable models were used.
The following are the outcomes:
Based on a model-derived risk score threshold, a customised enrichment framework was built for tailored patient selection. Short Run Emphasis, a radiomic characteristic, was found to be the most informative. An enriched subgroup (42 percent) of patients had longer OS in Evo+Dox vs. Dox groups [p = 0.01, Hazard Ratio (HR) = 0.57 (0.36-0.90), outperforming a clinical-only strategy. The substantial survival difference was validated in an independent test set using the same model and threshold value (p = 0.002, HR = 0.29 (0.13-0.63), 38 percent patients included). The Table shows the split of Dox+Evo treatment benefit based on the proportion of patients included in the model. This method also revealed which patients were most likely to benefit from doxorubicin alone.
Final Thoughts:
The study proposes a radiomic strategy for patient enrichment in clinical trials based on a quantitative score that is the first of its type. We’ve shown that if the innovative model had been used to select individuals for inclusion in the SARC021 study, the primary survival goal for patients with metastatic STS would have been fulfilled.