Feasibility of lung cancer prediction from low-dose CT scan and smoking factors using causal models
Thorax Mar 31, 2019
Raghu VK, et al. - Researchers examined 218 subjects with lung cancer or benign nodules to construct an efficient tool that can predict early lung cancer using clinical, demographic and low-dose CT (LDCT) characteristics. The training cohort included 92 subjects from the Pittsburgh Lung Screening Study cohort. The Lung Cancer Causal Model (LCCM) was developed using learnt probabilistic graphical models. For validation of LCCM, a separate cohort of 126 subjects was used. A greatly enhanced predictive accuracy was obtained by including LDCT scan features, and an improved detection of cancer was offered by LCCM over existing methods (like the Brock parsimonious model). In the validation cohort, LCCM enabled detection of 30% of the benign nodules with no risk of misclassifying cancer nodules. Overall, LCCM represents a promising tool to predict lung cancer.
-
Exclusive Write-ups & Webinars by KOLs
-
Daily Quiz by specialty
-
Paid Market Research Surveys
-
Case discussions, News & Journals' summaries