RESUMEN
Background and Aims: We aim to develop a diagnostic tool for pathological-image classification using transfer learning that can be applied to diverse tumor types. Methods: Microscopic images of liver tissue with and without hepatocellular carcinoma (HCC) were used to train and validate the classification framework based on a convolutional neural network. To evaluate the universal classification performance of the artificial intelligence (AI) framework, histological images from colorectal tissue and the breast were collected. Images for the training and validation sets were obtained from the Xiamen Hospital of Traditional Chinese Medicine, and those for the test set were collected from Zhongshan Hospital Xiamen University. The accuracy, sensitivity, and specificity values for the proposed framework were reported and compared with those of human image interpretation. Results: In the human-machine comparisons, the sensitivity, and specificity for the AI algorithm were 98.0, and 99.0%, whereas for the human experts, the sensitivity ranged between 86.0 and 97.0%, while the specificity ranged between 91.0 and 100%. Based on transfer learning, the accuracies of the AI framework in classifying colorectal carcinoma and breast invasive ductal carcinoma were 96.8 and 96.0%, respectively. Conclusion: The performance of the proposed AI framework in classifying histological images with HCC was comparable to the classification performance achieved by human experts, indicating that extending the proposed AI's application to diagnoses and treatment recommendations is a promising area for future investigation.
RESUMEN
Hepatitis B virus (HBV) infection represents a public health threat and a challenge for the medical community. Untimely treatment may lead to liver cirrhosis and even liver cancer. At present, the major treatment for hepatitis B e antigen (HBeAg)-positive chronic hepatitis B patients includes administration of interferon-α (IFN-α), which has anti-viral and immunomodulatory effects. Plasmacytoid dendritic cells (pDCs) and Toll-like receptor-9 (TLR-9) have important roles in anti-viral therapy. However, their predictive value regarding the efficacy of IFN-α treatment of HBeAg-positive chronic hepatitis B (CHB) patients has remained elusive. A total of 178 patients with CHB and HBeAg-positive status, who had not received any previous anti-HBV treatment, were enrolled in the present study. All patients were treated with IFN-α. HBV DNA load, hepatitis B surface antigen and serum alanine aminotransferase were measured prior to and following 48 weeks of treatment. According to HBV levels, the patients were divided into a response group and non-responders group. To determine the amount of pDCs, blood dendritic cell antigen 2 (BDCA-2)- and immunoglobulin-like transcript 7 (ILT7)-expressing cells in liver biopsies were detected using immunohistochemistry. TLR-9 expression in peripheral blood mononuclear cells was determined by reverse transcription-quantitative PCR. There was no significant difference in the proportion of pDCs (BDCA-2; ILT7) and TLR-9 mRNA expression between the response group and the non-responders group prior to IFN-α treatment. After IFN-α treatment, BDCA-2, ILT7 and TLR-9 mRNA expression was obviously increased in the response group compared with that in the non-responders group (P<0.05). Increased expression of BDCA-2, ILT7 and TLR-9 mRNA was negatively correlated with HBV DNA (P<0.05). Increased levels of pDCs and TLR-9 were negatively correlated with HBV DNA, and were thus capable of predicting the IFN-α treatment response in patients with CHB and HBeAg-positive status.
