WCLC 2025 - Posters & ePosters-P1.04.23 Real-World Predictive Performance of Lung-RADS in the Evaluation of Part-Solid Nodules

Pdf Summary

This study evaluated the real-world predictive performance of Lung-RADS in assessing part-solid nodules (PSNs) detected via low-dose lung cancer screening CT scans, focusing on its diagnostic accuracy and risk stratification capabilities. PSNs often represent early invasive adenocarcinoma and have a malignancy rate of up to 63%, with the size of the solid component correlating with invasiveness. Lung-RADS stratifies risk based on total nodule size and the size of the solid component, categorizing PSNs into levels 2, 3, 4A, 4B, and 4X.<br /><br />Data from 4,382 CT scans at a single tertiary institution between May 2017 and February 2025 identified 78 PSNs, with 53 nodules in 46 patients included after exclusions. Malignancy was confirmed by histopathology or imaging follow-up. The study found the malignancy rate across Lung-RADS categories was higher than expected in lower categories, with 24.5% of nodules malignant overall. Sensitivity was high (90%) when using category 3 as the positive threshold, but specificity was low (18.8%), indicating many false positives. ROC analysis showed moderate predictive performance, with the solid component size cutoff of 3.8 mm yielding better specificity (68.75%) and accuracy (71.43%) than total size alone.<br /><br />The differentiation between categories 4A and 4B did not significantly improve risk stratification. The authors noted limitations due to the small sample size and called for cautious interpretation. The study highlights that while Lung-RADS is sensitive in detecting malignancy in PSNs, its specificity is limited, and further refinement may be needed to improve clinical decision-making for these lesions.<br /><br />In summary, Lung-RADS is a useful tool for risk stratification of PSNs found on lung cancer screening CTs, but its predictive accuracy, especially specificity, can be improved, and solid component measurements are crucial in malignancy risk assessment.

Asset Subtitle

Seulgi You

Meta Tag

Speaker Seulgi You

Topic Screening and Early Detection

Keywords

Lung-RADS

part-solid nodules

PSNs

lung cancer screening

low-dose CT

malignancy rate

solid component size

risk stratification

diagnostic accuracy

ROC analysis