Lung cancer screening intervals based on cancer risk

Abstract

ObjectivesAs low-dose CT screening is gaining acceptance, focus is on increasing the efficiency of screening. One major consideration is to reduce the total number of annual rounds by increasing the interval between screening rounds. It has been suggested that longer intervals could be used for individuals who are at lower risk of lung cancer. In this study, we explored whether eligible participants in a program of LDCT screening who are at lower risk of lung cancer have less aggressive cancers than those at higher risk. MethodsWe retrospectively identified 118 participants in I-ELCAP database between 1992–2019 who had been screened using HIPAA-compliant protocols and had solid lung cancers diagnosed on an annual round of screening, 7–18 months after the prior round. Volume doubling time (VDT) for each cancer was calculated. Estimated risk of developing lung cancer was calculated using PLCOM2012 model. The strength of the relationship between VDT and individual PLCOM2012 scores was assessed by Pearson(r) and Spearman (ρ) correlation coefficients. ResultsVDTs were significantly different by cell-type (p < 0.0001); median VDT for small cell was 34.0 days, followed by other cell-types (61.8 days), squamous-cell (73.3 days), and adenocarcinoma (135.7 days). The median VDT for the 78 (66.1 %) Stage I lung cancers was significantly longer than the 40 Stage II + lung cancers (101.4 days vs. 45.5 days, p < 0.0001). None of the established lung cancer risk indicators (age, pack-years of smoking, or PLCOM2012 scores) were significant predictors of VDT or lung cancer stage. ConclusionNo significant relationship was demonstrated between risk of developing lung cancer (measured by risk models, age or smoking history) and lung cancer aggressiveness (measured by VDT, cell-type and Stage). This suggests that there is no evidence for determining intervals between repeat screenings using risk-based characteristics. It does not, however, exclude the possibility that future models may establish such a relationship.