A multicenter proof-of-concept study on deep learning-based intraoperative discrimination of primary central nervous system lymphoma.

Accurate intraoperative differentiation of primary central nervous system lymphoma (PCNSL) remains pivotal in guiding neurosurgical decisions. However, distinguishing PCNSL from other lesions, notably glioma, through frozen sections challenges pathologists. Here we sought to develop and validate a deep learning model capable of precisely distinguishing PCNSL from non-PCNSL lesions, especially glioma, using hematoxylin and eosin (H&E)-stained frozen whole-slide images. Also, we compared its performance against pathologists of varying expertise. Additionally, a human-machine fusion approach integrated both model and pathologic diagnostics. In external cohorts, LGNet achieved AUROCs of 0.965 and 0.972 in distinguishing PCNSL from glioma and AUROCs of 0.981 and 0.993 in differentiating PCNSL from non-PCNSL lesions. Outperforming several pathologists, LGNet significantly improved diagnostic performance, further augmented to some extent by fusion approach. LGNet's proficiency in frozen section analysis and its synergy with pathologists indicate its valuable role in intraoperative diagnosis, particularly in discriminating PCNSL from glioma, alongside other lesions.

Prognostic value of the immunohistochemical score based on four markers in head and neck squamous cell carcinoma.

Purpose: Head and neck squamous cell carcinoma (HNSCC) ranks sixth among all cancers globally regarding morbidity, and it has a poor prognosis, high mortality, and highly aggressive properties. In this study, we established a model for predicting prognosis based on immunohistochemical (IHC) scores. Methods: Data on 402 HNSCC cases were collected, the glmnet Cox proportional hazards model was used, risk factors were analyzed for predicting the prognosis of survival, and the IHC score was established. We used the IHC score to predict disease-free survival (DFS) using training and independent validation cohorts, including 264 cases in total. Additionally, the accuracy of the IHC score and the TNM system (8th edition) was compared. A DFS prediction nomogram was established by combining the prognostic factors. Results: The IHC scores included CK, Ki-67, p16, and p40 staining intensity. The concordance index and the Kaplan-Meier survival analysis showed that the IHC scores had high predictive power for HNSCC. Our results showed that the IHC score is an independent factor that can predict prognosis in a multivariate Cox regression analysis. When predicting DFS, the IHC score had a significantly higher value for the area under the ROC curve (AUC) than that of the TNM system. A nomogram was established and included the IHC score, age, tumor location, and the TNM stage. The calibration curves exhibited high consistency between the prognosis predicted by our nomogram and the actual prognosis. Conclusions: The IHC score was more accurate than the eighth edition of the TNM system in predicting HNSCC prognosis. Therefore, combining the two methods can facilitate individualized patient consultation and care.

A deep learning model and human-machine fusion for prediction of EBV-associated gastric cancer from histopathology.

Epstein-Barr virus-associated gastric cancer (EBVaGC) shows a robust response to immune checkpoint inhibitors. Therefore, a cost-efficient and accessible tool is needed for discriminating EBV status in patients with gastric cancer. Here we introduce a deep convolutional neural network called EBVNet and its fusion with pathologists for predicting EBVaGC from histopathology. The EBVNet yields an averaged area under the receiver operating curve (AUROC) of 0.969 from the internal cross validation, an AUROC of 0.941 on an external dataset from multiple institutes and an AUROC of 0.895 on The Cancer Genome Atlas dataset. The human-machine fusion significantly improves the diagnostic performance of both the EBVNet and the pathologist. This finding suggests that our EBVNet could provide an innovative approach for the identification of EBVaGC and may help effectively select patients with gastric cancer for immunotherapy.