Differential Diagnosis of Image Discovered Small Pulmonary Nodules (SPN): A Real-World Study
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Objective
The present investigation explored the disease spectrum of SPN (≤2cm) and attempted to establish the potential predictive models for non-benign small pulmonary nodules (NBSPN) identified from CT screening in a real-world perspective study.
Methods
A retrospective analysis was conducted on 6166 patients with SPN detected via chest CT scans at Fuzhou University Affiliated Provincial Hospital from January 2017 to September 2022. R language and SPSS software were used for data analyses.
Results
Of the 6,166 patients with SPN, 954 (15.47%) had their diagnosis confirmed pathologically. Among these nodules, NBSPN accounted for 88.7%. In NBSPN, the precursor glandular lesion (PGL) accounted for 14.18%, including 5.83% with atypical adenomatous hyperplasia (AAH) and 94.17% with adenocarcinoma in situ (AIS). Lung adenocarcinoma (LUAD) accounted for 85.82%, of which 32.07% were microinvasive adenocarcinoma (MIAD) and 67.93% invasive adenocarcinoma (IAD). Using the ROC model, none of the classical clinical factors had predictive value to the nature of SPN, including age, smoking, underlying conditions, CEA level, family history etc. Interestingly, the risk factors for NBSPN predicted by the ROC model include: female (OR: 1.842, 95%CI: 1.086-3.125, P=0.024), pure ground-glass opacities(pGGO) (OR: 5.243, 95%CI: 2.640-10.411, P<0.001), and part-solid (PS) (OR: 5.643, 95%CI: 1.970-16.167, P=0.001) in chest imaging, collectively delivering a highly significant AUC value at 0.748 in training set and 0.799 in validation set for distinguishing NBSPN from BSPN.
Conclusion
Conventional clinical data are not sufficient to differentiate NBSPN from BSPN. Factors with high predictive values for NBSPN include female, pGGO and PS.
