Using artificial intelligence (AI) to screen cancer patients for clinical trials is effective and efficient, according to research published in the Journal of the National Cancer Institute.
Researchers found that AI screening generally had high accuracy, sensitivity, and specificity. In addition, AI screening proved more efficient than manual screening.
The researchers reviewed 10 studies of AI use in the clinical trial enrollment process. The studies included 19 datasets and more than 50,000 patients. The sample sizes ranged from 96 to 48,124, and the studies were conducted in the United States, Europe, and Australia.
Published AI algorithms were used in 4 studies. For 5 studies, researchers developed their own algorithms. For 1 study, researchers used an algorithm created for a previous project.
Across 13 evaluable datasets, the accuracy of AI algorithms ranged from 62.8% to 100%. Across 17 datasets, the sensitivity of the algorithms ranged from 46.7% to 100%.
Across 16 datasets, the specificity of the algorithms ranged from 59.2% to 100%. However, all but 1 of the evaluable datasets had accuracy, sensitivity and specificity values higher than 80%.
The positive predictive value ranged from 12.6% to 100% across 17 datasets. The negative predictive value ranged from 87.2% to 100% across 16 datasets.
Eight datasets encompassing a total of 40,447 patients had sufficient data for a meta-analysis. The summary sensitivity of the AI algorithms was 90.5%, and the summary specificity was 99.3%.
Data on time savings were reported for 4 studies, and all 4 showed that AI screening was more efficient than manual screening. One study showed a 78% time savings with AI. Another study showed that AI performed screening in 1.15 hours that would have taken 263 days if done manually.
“AI appears to have comparable, if not superior, performance to manual screening for oncology clinical trials enrollment,” the researchers wrote. “As well, AI is efficient, requiring less time and resources to enroll patients; therefore, AI should be further investigated and implemented for this application.”
Chow R, Midroni J, Kaur J, et al. Use of artificial intelligence for cancer clinical trial enrolment: A systematic review and meta-analysis. J Natl Cancer Inst. Published online January 23, 2023. doi:10.1093/jnci/djad013
This article originally appeared on Cancer Therapy Advisor