OBJECTIVES

: Integration of AI (defined as a technology solution that uses machine-based algorithms to learn from experience, adjust to new inputs, and augment the capabilities of humans) into cancer care has the potential to drive efficiencies and reduce costs of care by helping target resources and efforts. The objective of this study was to assess the perceptions of oncologists regarding the utility and application of AI in cancer care.

METHODS

: A representative sample of community-based oncologists from across the United States convened for live meetings held in January, June and October 2019 to discuss the role of AI in oncology. Responses to survey questions were captured via an audience response system; results were analyzed using descriptive statistics.

RESULTS

: A total of 151 oncologists were surveyed, with 59% reporting their primary specialty as hematology oncology and 75% having been in practice >10 years. Forty percent of respondents defined AI as the augmentation of human capabilities, while 34% selected a definition based on machine-based algorithms, and 14%, predictive analyses. The key areas identified as opportunities for AI to improve care for cancer patients included improving accuracy of diagnostics (54%), helping to determine best treatment paths (49%) and automating administrative tasks so physicians can focus on patients (40%). Among the 99 physicians surveyed regarding their top concerns about AI, responses included the need for additional clinical data demonstrating utility of AI (28%), potential for bias in AI algorithms (23%), and the belief that AI cannot outperform human intelligence/intuition (17%).

CONCLUSIONS

: Oncologists foresee use of AI in cancer detection, reducing administrative burden and improving diagnostic accuracy in the foreseeable future, but need additional data proving its clinical utility and impact on patient outcomes. These insights can help to inform efforts of providers, payors and technology developers engaged in bringing AI to cancer care.