Comparing Risk Prediction Models in Lung Cancer Screening: Focus on Black Smokers
Posted: Tuesday, December 8, 2020
An article published in the Journal of Thoracic Oncology compared two lung cancer screening models, as the current model is based solely on age and smoking history. Lawrence E. Feldman, MD, of the University of Illinois, Chicago, and colleagues focused on the current U.S. Preventive Services Task Force (USPSTF) lung cancer screening eligibility criteria and the PLCOm2012 risk prediction model. They found the PLCOm2012 model more effectively selected Black ever-smokers for screening.
“We found that the PLCOm2012 model was significantly more sensitive in selecting lung cancer patients as being eligible for screening for the entire cohort,” stated coauthor Mary M. Pasquinelli, DNP, also of the University of Illinois, in an International Association for the Study of Lung Cancer press release. “Broader use of this model in racially diverse populations may help overcome disparities in lung cancer screening and outcomes.”
A total of 883 lung cancer cases diagnosed between 2010 and 2019 were identified from an urban medical center serving a racially and ethnically diverse population, the authors noted. Researchers retroactively assessed cases for screening eligibility based on the USPSTF criteria versus the PLCOm2012 model. Data were analyzed for risk threshold based on age, education level, body mass index, cancer history, race and ethnicity, among other factors.
The racial/ethnic makeup of the ever-smokers examined in this study consisted of 258 Whites (29.2%), 497 Blacks (56.3%), 69 Hispanics (7.8%), 24 Asians (2.7%), and 35 other (4.0%). Compared with White ever-smokers, Black ever-smokers had lower pack-years (37.3 vs. 48.2; P < .001), were more likely to be current smokers (65.6% vs. 57.0%; P < .001), and had a lower mean body mass index (25.8% vs. 27.1%; P = .004). They also found that in the Black cohort, the PLCOm2012 model had substantially higher sensitivity than the USPSTF criteria at thresholds of 1.51%, 1.70%, and 2.00% (P < .0001).
Disclosure: For full disclosures of the study authors, visit www.jto.org.
