
Marginal radiomics features as imaging biomarkers for pathological invasion in lung adenocarcinoma
This study identified 236 patients from two cohorts who underwent surgery for ground-glass nodules (GGNs). The novel marginal features described, when combined with a radiomics model, could help to differentiate invasive adenocarcinoma (IA) from adenocarcinoma in situ (AIS) and minimally invasive adenocarcinoma (MIA) on preoperative CT scans. Key points Our novel marginal features could improve the existing radiomics model to