Department of Pathology, The University of Hong Kong, Queen Mary Hospital, 11/F, Block T, 102 Pokfulam Road, HKSAR, Hong Kong.
Department of Pathology, CUHK Medical Centre, 9 Chak Cheung Street, Shatin, New Territories, HKSAR, Hong Kong.
Sci Rep. 2022 Apr 28;12(1):6965. doi: 10.1038/s41598-022-11009-x.
Deep myxoid soft tissue lesions have posed a diagnostic challenge for pathologists due to significant histological overlap and regional heterogeneity, especially when dealing with small biopsies which have profoundly low accuracy. However, accurate diagnosis is important owing to difference in biological behaviors and response to adjuvant therapy, that will guide the extent of surgery and the need for neo-adjuvant therapy. Herein, we trained two convolutional neural network models based on a total of 149,130 images representing diagnoses of extra skeletal myxoid chondrosarcoma, intramuscular myxoma, low-grade fibromyxoid sarcoma, myxofibrosarcoma and myxoid liposarcoma. Both AI models outperformed all the pathologists, with a significant improvement of accuracy up to 97% compared to average pathologists of 69.7% (p < 0.00001), corresponding to 90% reduction in error rate. The area under curve of the best AI model was on average 0.9976. It could assist pathologists in clinical practice for accurate diagnosis of deep soft tissue myxoid lesions, and guide clinicians for precise and optimal treatment for patients.
深部黏液样软组织病变由于组织学上的显著重叠和区域性异质性，给病理学家的诊断带来了挑战，尤其是在处理准确性极低的小活检时。然而，准确的诊断很重要，因为不同的生物学行为和对辅助治疗的反应不同，这将指导手术的范围和新辅助治疗的需求。在此，我们基于代表 extraskeletal myxoid chondrosarcoma、intramuscular myxoma、low-grade fibromyxoid sarcoma、myxofibrosarcoma 和 myxoid liposarcoma 诊断的总共 149130 张图像，训练了两个卷积神经网络模型。这两个 AI 模型的表现都优于所有病理学家，与平均准确率为 69.7%（p<0.00001）的病理学家相比，准确率显著提高，达到了 97%，相应的错误率降低了 90%。最佳 AI 模型的曲线下面积平均为 0.9976。它可以帮助病理学家在临床实践中准确诊断深部软组织黏液样病变，并为临床医生为患者提供精确和最佳的治疗提供指导。
