103 Artificial intelligence publication trends in reproductive cancers – who is being left behind?. (13th November 2020)

• Record Type: Journal Article
• Authors: Chan, J; Liao, C; Mann, A; Kapp, D; Mysona, D
• Is Part Of: International journal of gynecological cancer. Volume 30(2020)Supplement 3
• Page Start: A49
• Page End: A50
• Publication Date: 2020-11-13
• Subjects: Generative organs, Female -- Cancer -- Periodicals; 616.99465
• DOI: 10.1136/ijgc-2020-IGCS.92
• Languages: English
• ISSNs: 1048-891X

Abstract

Abstract : Introduction: There is an increasing presence and utility of artificial intelligence (AI) in oncology. We proposed to determine the trends of AI publications in screening, diagnosis, surgery, and treatment of reproductive cancers over time. Methods: Using the PubMed database, we used keywords and MeSH terms to index research articles from 1990 to 2019, and the National Cancer Institute's Joinpoint Regression Program for statistical analysis. Results: We identified a significant increase in AI research on all cancer types over the last 30 years from 19 to 1, 829 publications per year. 14, 721 publications were related to AI and cancer, 41% of which discussed diagnosis, 30% treatment, 24% surgery, and 5% screening (p<0.001). Despite having the lowest number of publications, screening had the highest average annual rate of increase at 23.6% (p<0.001) (table 1A ). The numbers of breast and prostate cancer publications were significantly higher than that of gynecologic cancers. Of 5, 808 reproductive cancer and AI publications, prostate cancer comprised 42%, breast 40%, cervical 8%, ovarian 6%, and uterine 5% of all AI research. Of gynecologic cancers, ovarian cancer publications had the highest average yearly increase at 16.1% on joinpoint analysis, whereas the average annual rates of increase in uterine and cervical cancer publications were 14.5% and 10.7% (p<0.001) (table 1B ). Conclusion: Compared to breast and prostate cancers, there are a disproportionately lower number and rate of publications related to gynecologic cancers and AI. Ovarian malignancies were the most widely published compared to uterine and cervical malignancies.