Quantitative CT parameters and imaging features for differentiating non-invasive from invasive pathology in surgically resected pulmonary micronodules: a retrospective study

Background This study aimed to investigate the diagnostic utility of computed tomography (CT)-derived quantitative parameters combined with qualitative imaging features for differentiating non-invasive from invasive pathology in surgically resected pulmonary micronodules (≤ 5 mm), a subgroup in which conventional imaging characteristics are often subtle or atypical, using histopathology as the reference standard. Methods We retrospectively analyzed 96 surgically resected pulmonary micronodules from 61 patients. Benign lesions, atypical adenomatous hyperplasia (AAH), and adenocarcinoma in situ (AIS) formed the non-invasive group; minimally invasive adenocarcinoma (MIA) and invasive adenocarcinoma (IAC) formed the invasive group. Generalized estimating equations (GEE) accounted for within-patient clustering. Model performance was assessed using receiver operating characteristic analysis and patient-level bootstrap resampling. A sensitivity analysis excluded benign lesions. Results The cohort included 47 non-invasive and 49 invasive nodules. In multivariable GEE, surface area was associated with invasive pathology (OR = 1.016, 95% CI 1.006–1.027; P = 0.002). The multivariable model had an apparent AUC of 0.773 and an optimism-corrected AUC of 0.766; its AUC did not differ significantly from surface area alone (AUC = 0.716; DeLong P = 0.097). After excluding 29 benign lesions, surface area remained associated with invasive pathology (OR = 1.023, 95% CI 1.002–1.043; P = 0.030). Conclusion Surface area was associated with invasive pathology in this highly selected surgical cohort, including after adjustment for within-patient clustering and exclusion of benign lesions. The exploratory model requires independent validation before broader clinical application. Clinical trial number Not applicable.