Comparison of publications between cancer sites in high impact factor journals.

Abstract

1588 Background: Prior studies have shown disparities in federal funding allocation by cancer site after accounting for mortality, incidence, and person-years of life lost. Whether disparities exist in publications in high impact factor journals is unknown. The aim of this study is to analyze research publications in high impact factor journals by cancer site relative to their burden from 2000 to 2018. Methods: Abstracts from the top 55 impact factor journals from 2000 to 2018 were extracted from PubMed. We developed an algorithm to search the title of the abstract to determine whether the abstract was about cancer and identify the cancer type. The algorithm was validated against the gold standard of human review in 1143 abstracts and yielded a sensitivity and a specificity of >95%. An article proportion for each cancer site was calculated (abstracts for a cancer site / total cancer abstracts). Article proportion was compared to cancer site-specific incidence, mortality and lethality from NCI’s SEER database using scatter plots and non-parametric Wilcoxon signed rank test. Results: We identified 128,377 articles; 31,045 (24%) were about cancer. Cancers with the most publications in high impact factor journals were breast (22%), lung (11%), and colorectal (10%) cancer. Breast, lung and colorectal cancer were also well represented in high impact factor journals compared to their incidence, mortality, and lethality. Gynecologic (ovarian, cervical, and endometrial), pancreatic, hepatobiliary, and prostate cancers were poorly represented in all 3 metrics. Breast, lung, and colorectal cancer ranked in the top half of Article to Lethality scores, while the gynecologic, pancreatic, and hepatobiliary cancers were in the bottom half (p <0.05 for all comparisons). Comparison of rankings by lethality versus by funding indicates relative underfunding of the gynecologic, pancreatic, hepatobiliary, and prostate cancers. Conclusions: Research publications in high impact factor journals by cancer site are not proportionate to individual societal cancer burden. Significant disparities exist between breast, lung, and colorectal cancer when compared to gynecologic, pancreatic, hepatobiliary, and prostate cancers relative to their disease burden. Underfunding of these cancer sites likely contributes to this publication gap.[Table: see text]