Deep learning for precision grading of Schistosoma japonicum-induced liver fibrosis in ultrasound images.

A deep learning system accurately grades Schistosoma japonicum-induced liver fibrosis in ultrasound images, with 93.9% of predictions within 0.5 grades of experts.

• Why it matters: Liver fibrosis from schistosomiasis poses a significant health challenge in endemic areas, and current grading methods are subjective and reliant on specialist expertise, limiting widespread screening.
• What they did: Researchers developed and tested a deep learning model using a multicentre dataset of 167,702 labeled ultrasound images on a 36-level scale, mapping to four clinical grades, with an independent test set of 16,811 images.
• The result: The system achieved a mean absolute error of 0.116 and processed images in under 400 milliseconds on portable devices, enabling scalable, consistent, and rapid screening in resource-limited settings.