Negligible impact of perifissural nodules in an AI-first reader workflow from UK lung screening trial

This study quantitatively assessed the impact of perifissural nodules (PFNs) on radiologist workload in an AI-first reader workflow, using CT data from the UK Lung Cancer Screening (UKLS) trial. AVIEW LCS (Coreline Soft, v.39.14) was used to automatically detect nodules ≥100mm³ in 1,252 low-dose CT (LDCT) scans. Typical PFNs ≥100mm³ as the sole finding requiring radiologist review accounted for only 1.9% of cases (24/1,252), and none of the 57 detected typical PFNs were malignant. These results demonstrate that the additional radiologist burden attributable to PFNs in an AI-first workflow is negligible, and that a volume-based AI classification system is sufficiently safe and efficient in a lung cancer screening setting. Jaeyoun Yi (이재연) from Coreline Soft is listed as a co-author.