The Key Laboratory of Cardiovascular Remodeling and Function Research, Chinese Ministry of Education and Chinese Ministry of Health, and The State and Shandong Province Joint Key Laboratory of Translational Cardiovascular Medicine, Qilu Hospital of Shandong University, No. 107 Wenhuaxi Road, Jinan, 250012, People's Republic of China.
Department of Ultrasound, First Affiliated Hospital of Southern University of Science and Technology, The Second Clinical Medical College of Jinan University, Shenzhen People's Hospital, Shenzhen, 518020, People's Republic of China.
Eur Radiol. 2021 Jul;31(7):4991-5000. doi: 10.1007/s00330-020-07561-7. Epub 2021 Jan 6.
To investigate how a DL model makes decisions in lesion classification with a newly defined region of evidence (ROE) by incorporating "explainable AI" (xAI) techniques.
A data set of 785 2D breast ultrasound images acquired from 367 females. The DenseNet-121 was used to classify whether the lesion is benign or malignant. For performance assessment, classification results are evaluated by calculating accuracy, sensitivity, specificity, and receiver operating characteristic for experiments of both coarse and fine regions of interest (ROIs). The area under the curve (AUC) was evaluated, and the true-positive, false-positive, true-negative, and false-negative results with breakdown in high, medium, and low resemblance on test sets were also reported.
The two models with coarse and fine ROIs of ultrasound images as input achieve an AUC of 0.899 and 0.869, respectively. The accuracy, sensitivity, and specificity of the model with coarse ROIs are 88.4%, 87.9%, and 89.2%, and with fine ROIs are 86.1%, 87.9%, and 83.8%, respectively. The DL model captures ROE with high resemblance of physicians' consideration as they assess the image.
We have demonstrated the effectiveness of using DenseNet to classify breast lesions with limited quantity of 2D grayscale ultrasound image data. We have also proposed a new ROE-based metric system that can help physicians and patients better understand how AI makes decisions in reading images, which can potentially be integrated as a part of evidence in early screening or triaging of patients undergoing breast ultrasound examinations.
• The two models with coarse and fine ROIs of ultrasound images as input achieve an AUC of 0.899 and 0.869, respectively. The accuracy, sensitivity, and specificity of the model with coarse ROIs are 88.4%, 87.9%, and 89.2%, and with fine ROIs are 86.1%, 87.9%, and 83.8%, respectively. • The first model with coarse ROIs is slightly better than the second model with fine ROIs according to these evaluation metrics. • The results from coarse ROI and fine ROI are consistent and the peripheral tissue is also an impact factor in breast lesion classification.
通过结合“可解释人工智能”（xAI）技术，研究深度学习（DL）模型如何在新定义的证据区域（ROE）中做出决策进行病变分类。
该研究使用了一个包含 367 名女性的 785 张二维乳腺超声图像数据集。使用 DenseNet-121 对病变是良性还是恶性进行分类。为了进行性能评估，通过计算准确性、敏感性、特异性和接收器工作特征曲线（ROC），对粗感兴趣区（ROI）和细 ROI 实验的分类结果进行评估。评估了曲线下面积（AUC），并报告了测试集中高、中、低相似性的真阳性、假阳性、真阴性和假阴性结果。
输入粗和细超声图像 ROI 的两个模型分别达到了 0.899 和 0.869 的 AUC。粗 ROI 模型的准确性、敏感性和特异性分别为 88.4%、87.9%和 89.2%，细 ROI 模型的准确性、敏感性和特异性分别为 86.1%、87.9%和 83.8%。该 DL 模型捕捉到了医生在评估图像时考虑的高相似度的 ROE。
该研究证明了使用 DenseNet 对有限数量的二维灰度超声图像数据进行乳腺病变分类的有效性。该研究还提出了一种新的基于 ROE 的度量系统，可以帮助医生和患者更好地理解 AI 如何在阅读图像时做出决策，这可能会被整合为早期筛查或对接受乳腺超声检查的患者进行分诊的证据的一部分。
输入粗和细超声图像 ROI 的两个模型分别达到了 0.899 和 0.869 的 AUC。粗 ROI 模型的准确性、敏感性和特异性分别为 88.4%、87.9%和 89.2%，细 ROI 模型的准确性、敏感性和特异性分别为 86.1%、87.9%和 83.8%。
根据这些评估指标，第一个使用粗 ROI 的模型略优于第二个使用细 ROI 的模型。
粗 ROI 和细 ROI 的结果是一致的，外围组织也是乳腺病变分类的一个影响因素。
