An automatic diagnosis model of otitis media with high accuracy rate using transfer learning.

Introduction: Chronic Suppurative Otitis Media (CSOM) and Middle Ear Cholesteatoma are two common chronic otitis media diseases that often cause confusion among physicians due to their similar location and shape in clinical CT images of the internal auditory canal. In this study, we utilized the transfer learning method combined with CT scans of the internal auditory canal to achieve accurate lesion segmentation and automatic diagnosis for patients with CSOM and middle ear cholesteatoma. Methods: We collected 1019 CT scan images and utilized the nnUnet skeleton model along with coarse grained focal segmentation labeling to pre-train on the above CT images for focal segmentation. We then fine-tuned the pre-training model for the downstream three-classification diagnosis task. Results: Our proposed algorithm model achieved a classification accuracy of 92.33% for CSOM and middle ear cholesteatoma, which is approximately 5% higher than the benchmark model. Moreover, our upstream segmentation task training resulted in a mean Intersection of Union (mIoU) of 0.569. Discussion: Our results demonstrate that using coarse-grained contour boundary labeling can significantly enhance the accuracy of downstream classification tasks. The combination of deep learning and automatic diagnosis of CSOM and internal auditory canal CT images of middle ear cholesteatoma exhibits high sensitivity and specificity.

The value of CT-based radiomics nomogram in differential diagnosis of different histological types of gastric cancer.

To establish and verify a nomogram based on computed tomography (CT) radiomics analysis to predict the histological types of gastric cancer preoperatively for patients with surgical indications. A sum of 171 patients with gastric cancer were included into this retrospective study. The least absolute shrinkage and selection operator (LASSO) was used for feature selection while the multivariate Logistic regression method was used for radiomics model and nomogram building. The area under curve (AUC) was used for performance evaluation in this study. The radiomics model got AUCs of 0.755 (95% CI 0.650-0.859), 0.71 (95% CI 0.543-0.875) and 0.712 (95% CI 0.500-0.923) for histological prediction in the training, the internal and external verification cohorts. The radiomics nomogram based on radiomics features and Carbohydrate antigen 125 (CA125) showed good discriminant performance in the training cohort (AUC: 0.777; 95% CI 0.679-0.875), the internal (AUC: 0.726; 95% CI 0.5591-0.8933) and external verification cohort (AUC: 0.720; 95% CI 0.5036-0.9358). The calibration curve of the radiomics nomogram also showed good results. The decision curve analysis (DCA) shows that the radiomics nomogram is clinically practical. The radiomics nomogram established and verified in this study showed good performance for the preoperative histological prediction of gastric cancer, which might contribute to the formulation of a better clinical treatment plan.