Related Experiment Video
Updated: Jun 25, 2025
A Three-Dimensional Spheroid Model to Investigate the Tumor-Stromal Interaction in Hepatocellular Carcinoma
Published on: September 30, 2021
Deep learning and digital pathology powers prediction of HCC development in steatotic liver disease
Takuma Nakatsuka1, Ryosuke Tateishi1, Masaya Sato1,2
1Department of Gastroenterology, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan.
Insights
A deep learning (DL) model accurately predicts hepatocellular carcinoma (HCC) in patients with steatotic liver disease using whole-slide images. This AI tool identifies early cancer signs beyond traditional fibrosis assessment.
Area of Science:
- Digital pathology
- Artificial intelligence in medicine
- Hepatology
Background:
- Identifying high-risk patients for hepatocellular carcinoma (HCC) in steatotic liver disease (SLD) is challenging.
- Current methods may not fully capture early signs of hepatocarcinogenesis.
Purpose of the Study:
- To develop and validate a deep learning (DL) model for predicting HCC development in SLD patients.
- To utilize hematoxylin and eosin-stained whole-slide images for HCC risk prediction.
Main Methods:
- A DL model using deep convolutional neural networks was trained on whole-slide images from SLD patients.
- Paired and non-paired cases were used for training and validation, respectively.
- Model performance was evaluated using accuracy and area under the curve (AUC).
Main Results:
- The DL model achieved 81.0% accuracy and 0.80 AUC in training, and 82.3% accuracy and 0.84 AUC in validation.
- Performance was comparable to logistic regression models based on fibrosis stage.
- The DL model identified HCC development in patients with mild fibrosis and highlighted key pathological features.
Conclusions:
- The DL model shows potential for early hepatocarcinogenesis detection in SLD patients.
- It can identify subtle pathological features beyond fibrosis staging.
- This AI approach may improve risk stratification for HCC in SLD.
Background And Aims:
Identifying patients with steatotic liver disease who are at a high risk of developing HCC remains challenging. We present a deep learning (DL) model to predict HCC development using hematoxylin and eosin-stained whole-slide images of biopsy-proven steatotic liver disease.
Approach And Results:
We included 639 patients who did not develop HCC for ≥7 years after biopsy (non-HCC class) and 46 patients who developed HCC <7 years after biopsy (HCC class). Paired cases of the HCC and non-HCC classes matched by biopsy date and institution were used for training, and the remaining nonpaired cases were used for validation. The DL model was trained using deep convolutional neural networks with 28,000 image tiles cropped from whole-slide images of the paired cases, with an accuracy of 81.0% and an AUC of 0.80 for predicting HCC development. Validation using the nonpaired cases also demonstrated a good accuracy of 82.3% and an AUC of 0.84. These results were comparable to the predictive ability of logistic regression model using fibrosis stage. Notably, the DL model also detected the cases of HCC development in patients with mild fibrosis. The saliency maps generated by the DL model highlighted various pathological features associated with HCC development, including nuclear atypia, hepatocytes with a high nuclear-cytoplasmic ratio, immune cell infiltration, fibrosis, and a lack of large fat droplets.
Conclusions:
The ability of the DL model to capture subtle pathological features beyond fibrosis suggests its potential for identifying early signs of hepatocarcinogenesis in patients with steatotic liver disease.
