Abstract 2298: Understanding tumor biology complexity in the advanced cancer setting: PreciGENE® platform predictions correlate with exceptional responses to cancer treatment

Abstract

Abstract Cancer is a disease caused by an elaborate network of genetic and environmental factors, and its biologic complexity renders difficult the selection of optimized treatment options. New tools are needed to allow oncologists to quickly distinguish, among the plethora of possible combinations, those regimens that best fit the patient's specificities and increase the chances of survival. The PreciGENE analytical platform integrates genomic, transcriptomic and proteomic data to provide fully personalized therapeutic options. Here we present an analysis of 70 exceptional responders to cancer treatment for whom a molecular profile was published in the literature between February 2010 and June 2017. We aimed to evaluate the performance of the PreciGENE tool in the determination of optimized regimens for cancer therapy. Methods: A PubMed search was conducted using the keywords "exceptional response" and "cancer." All cases published in the period defined and presenting tumor molecular data were kept for analysis. Molecular and drug descriptions were entered into the PreciGENE system that computed a score (%), with a high score representing a better fit between treatment and tumor profile. Correlation between the predictive score and the clinical response observed was evaluated using a Mann-Whitney test: complete or partial responses (CR/PR) were compared to stable or progressive diseases (SD/PD). Results: A total of 202 treatment lines were retrieved, including 70 successful and 132 unsuccessful lines of therapy. Patients received an average of 1.6 ([CI95%] =1.5-1.8) treatment lines and molecular profiles retrieved an average of 3.7 [2.7-4.7] alterations per tumor. 84/202 (42%) regimens included at least two therapeutic agents. Among the regimens that led to an exceptional response (CR/PR), 31/70 (44%) were combination therapies. The predictive scores obtained for regimens that led to a positive outcome were significantly greater than those of regimens that had failed (mean [CI95%] = 60 [52-68]% for CR/PR vs 14 [10-19]% for SD/PD; p=.0001). Using a score threshold of 25% (corresponding to a therapeutic regimen matching 25% of the pathogenic alterations detected in the tumor), the PreciGENE platform results presented a sensitivity of 84% and a specificity of 77% for the prediction of the clinical response. The negative predictive value was 90%; the positive predictive value was 66%. Conclusion: The algorithm used by the PreciGENE decision-support platform correctly ranked the treatment response in 70 published cancer patients having received 202 different regimens and presented an exceptional response to at least one therapeutic regimen. In this retrospective study, we show that such systems may empower oncologists in their choice of treatment. Prospective studies of the use of decision-support platforms in advanced cancer are warranted. Citation Format: Amelie Boichard, Timothy V. Pham, Stephane B. Richard, Razelle Kurzrock. Understanding tumor biology complexity in the advanced cancer setting: PreciGENE® platform predictions correlate with exceptional responses to cancer treatment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2298.