Deep Learning for Differentiating Pulmonary Metastasis from Primary Lung Cancer Constructed on Frozen Sections for Intraoperative Diagnosis.

The surgical procedures for primary lung cancer and metastatic lung tumors differ. An intraoperative diagnosis is important for surgical procedure decisions. This study digitized frozen sections for intraoperative diagnosis, developing and evaluating deep learning models to differentiate lung cancer from metastatic tumors. A total of 1,668 slides from 1,458 patients with lung cancer or metastatic tumors who underwent surgery with an intraoperative diagnosis at a single institution were included. Lung cancer comprised 1,170 slides, while metastatic tumors comprised 498. Models were constructed to calculate the prediction probability using attention-based multiple-instance learning. Diagnostic performance was assessed using accuracy and area under the curve for each histological type; the area under the curve was 0.888 (0.871-0.904) with an accuracy of 80.1%. Lung adenocarcinoma prediction accuracy was favorable at 87.5% (95.5% for the lepidic pattern and 88.7% for the papillary pattern), while that for squamous cell carcinoma was 62.0%. Colon cancer was the most common metastatic tumor, with an accuracy of 89.3%, followed by soft tissue sarcoma with an accuracy of 82.6%. When limited to adenocarcinoma of lung cancer and metastatic tumors, the accuracy was 85.9%, while that for squamous cell carcinoma was 63.6%. The results indicate that deep learning can provide reasonable support in differentiating metastatic lung tumors from primary lung cancer on frozen sections, with performance varying by histological subtype. Deep learning may assist in intraoperative decision-making under time and resource constraints, though further multi-institutional validation is needed.