The goal of this observational study is to improve the intraoperative diagnosis accuracy of invasiveness for small-sized lung adenocarcinoma by combining multi-modal information. The main question it aims to answer is whether multi-modal information have great value of prediction on the invasiveness for small-sized lung adenocarcinoma. Since a promising limited resection is largely based on intraoperative frozen section diagnosis, there is a growing demand on the high-accuracy of timely pathology diagnosis. The multi-modal information of participants will be collected retrospectively.

Inclusion Criteria:

CT examination within 3 months before surgery Patients with operable clinical stage I lung cancer No previous treatment in the lungs or any other organ

≥ 20 years and ≤ 80 years old Tumor less than 3cm in diameter on thin-slice (0.625-1 mm) CT images Lung adenocarcinoma confirmed by surgical resection and histopathological diagnosis

Exclusion Criteria:

Marked artifacts on CT images History of preoperative treatment Incomplete clinical information or DICOM images History of other malignant tumors Lung cancer associated with cystic airspaces

From January 2018 to December 2022, about 3000 patients with lung adenocarcinoma were enrolled in this retrospective study.