Background Extracting metastatic information from previous radiologic-text reports is important, however, laborious annotations have limited the usability of these texts. We developed a deep-learning model for extracting primary lung cancer sites and metastatic lymph nodes and distant metastasis information from PET-CT reports for determining lung cancer stages. Methods PET-CT reports, fully written in English, were acquired from two cohorts of patients with lung cancer who were diagnosed at a tertiary hospital between January 2004 and March 2020. One cohort of 20,466 PET-CT reports was used for training and the validation set, and the other cohort of 4190 PET-CT reports was used for an additional-test set. A pre-processing model (Lung Cancer Spell Checker) was applied to correct the typographical errors, and pseudo-labelling was used for training the model. The deep-learning model was constructed using the Convolutional-Recurrent Neural Network. The performance metrics for the prediction model were accuracy, precision, sensitivity, micro-AUROC, and AUPRC. Results For the extraction of primary lung cancer location, the model showed a micro-AUROC of 0.913 and 0.946 in the validation set and the additional-test set, respectively. For metastatic lymph nodes, the model showed a sensitivity of 0.827 and a specificity of 0.960. In predicting distant metastasis, the model showed a micro-AUROC of 0.944 and 0.950 in the validation and the additional-test set, respectively. Conclusion Our deep-learning method could be used for extracting lung cancer stage information from PET-CT reports and may facilitate lung cancer studies by alleviating laborious annotation by clinicians.

Blood and fluid analysis is extensively used for classifying the etiology of pleural effusion. However, most studies focused on determining the presence of a disease. This study classified pleural effusion etiology employing deep learning models by applying contrastive-loss. Patients with pleural effusion who underwent thoracentesis between 2009 and 2019 at the Asan Medical Center were analyzed. Five different models for categorizing the etiology of pleural effusion were compared. The performance metrics were top-1 accuracy, top-2 accuracy, and micro-and weighted-AUROC. UMAP and t-SNE were used to visualize the contrastive-loss model’s embedding space. Although the 5 models displayed similar performance in the validation set, the contrastive-loss model showed the highest accuracy in the extra-validation set. Additionally, the accuracy and micro-AUROC of the contrastive-loss model were 81.7% and 0.942 in the validation set, and 66.2% and 0.867 in the extra-validation set. Furthermore, the embedding space visualization in the contrastive-loss model exhibited typical and atypical effusion results by comparing the true and false positives of the rule-based criteria. Therefore, classifying the etiology of pleural effusion was achievable using the contrastive-loss model. Conclusively, visualization of the contrastive-loss model will provide clinicians with valuable insights for etiology diagnosis by differentiating between typical and atypical disease types.