基于临床标准超声的肝脂肪变性分类深度学习模型的开发
Development of a Deep Learning Model for Classification of Hepatic Steatosis from Clinical Standard Ultrasound.
作者信息
Kaffas Ahmed El, Bhatraju Krishna Chaitanya, Vo-Phamhi Jenny M, Tiyarattanachai Thodsawit, Antil Neha, Negrete Lindsey M, Kamaya Aya, Shen Luyao
机构信息
Department of Radiology, University of California San Diego School of Medicine, La Jolla, CA, USA.
Department of Radiology, University of California San Diego School of Medicine, La Jolla, CA, USA; Department of Radiology, Stanford University School of Medicine, Stanford, CA, USA.
出版信息
Ultrasound Med Biol. 2025 Feb;51(2):242-249. doi: 10.1016/j.ultrasmedbio.2024.09.020. Epub 2024 Nov 12.
OBJECTIVE
Early detection and monitoring of hepatic steatosis can help establish appropriate preventative measures against progression to more advanced disease. We aimed to develop a deep learning (DL) program for classification of hepatic steatosis from standard-of-care grayscale ultrasound (US) images.
METHODS
In this single-center retrospective study, we utilized grayscale US images from January 1, 2010, to October 23, 2022, labeled with magnetic resonance imaging (MRI) proton density fat fraction (MRI-PDFF) to develop a DL multi-instance program for differentiating normal (S0) from steatotic liver (S1/2/3) and normal/mild steatosis (S0/1) from moderate/severe steatosis (S2/3). Diagnostic performances were assessed with area under the receiver operating characteristic curves (AUC), sensitivity, specificity and balanced accuracy with 95% confidence interval (CI).
RESULTS
A total of 403 patients with 403 US exams were included: 171 (42%) were normal (S0: MRI-PDFF <5%), 154 (38%) had mild steatosis (S1: MRI-PDFF 5-17.4%), 29 (7%) had moderate steatosis (S2: MRI-PDFF >17.4%-22.1%) and 49 (12%) had severe steatosis (S3: MRI-PDFF >22.1%). The dataset was split to include 322 patients in train/validation and 81 patients in a holdout test set (kept blind). The S0 versus S1/2/3 model achieved 81.3% (95% CI 72.1-90.5) AUC, 81.1% (70.6-91.6) sensitivity, 71.4% (54.7-88.2) specificity and 76.3% (66.4-86.2) balanced accuracy. The S0/1 versus S2/3 model achieved 95.9% (89-100) AUC, 87.5% (71.3-100) sensitivity, 96.9% (92.7-100) specificity and 92.2% (83.8-100) balanced accuracy. A multi-class model achieved a sensitivity of 71.4% (54.7-88.2) for S0, 67.6% (52.5-82.7) for S1 and 87.5% (71.3-100) for S2/3; specificity for the same model was 81.1% (70.6-91.6) for S0, 77.3% (64.9-89.7) for S1 and 96.9% (92.7-100) for S2/3.
CONCLUSION
Our DL program offered high sensitivity and accuracy in detecting and categorizing hepatic steatosis from standard-of-care ultrasound.
目的
早期检测和监测肝脂肪变性有助于制定适当的预防措施，以防止病情进展为更严重的疾病。我们旨在开发一种深度学习（DL）程序，用于根据标准护理灰度超声（US）图像对肝脂肪变性进行分类。
方法
在这项单中心回顾性研究中，我们利用了2010年1月1日至2022年10月23日的灰度US图像，这些图像用磁共振成像（MRI）质子密度脂肪分数（MRI-PDFF）进行标记，以开发一个DL多实例程序，用于区分正常肝脏（S0）和脂肪变性肝脏（S1/2/3），以及正常/轻度脂肪变性（S0/1）和中度/重度脂肪变性（S2/3）。使用受试者操作特征曲线下面积（AUC）、敏感性、特异性和95%置信区间（CI）的平衡准确性来评估诊断性能。
结果
共纳入403例患者的403次US检查：171例（42%）为正常（S0：MRI-PDFF<5%），154例（38%）有轻度脂肪变性（S1：MRI-PDFF 5-17.4%），29例（7%）有中度脂肪变性（S2：MRI-PDFF>17.4%-22.1%），49例（12%）有重度脂肪变性（S3：MRI-PDFF>22.1%）。数据集被分为训练/验证组的322例患者和保留测试组的81例患者（保持盲态）。S0与S1/2/3模型的AUC为81.3%（95%CI 72.1-90.5），敏感性为81.1%（70.6-91.6），特异性为71.4%（54.7-88.2），平衡准确性为76.3%（66.4-86.2）。S0/1与S2/3模型的AUC为95.9%（89-100），敏感性为87.5%（71.3-100），特异性为96.9%（92.7-100），平衡准确性为92.2%（83.8-100）。一个多类模型对S0的敏感性为71.4%（54.7-88.2），对S1的敏感性为67.6%（52.5-82.7），对S2/3的敏感性为87.5%（71.3-100）；同一模型对S0的特异性为81.1%（70.6-91.6），对S1的特异性为77.3%（64.9-89.7），对S2/3的特异性为96.9%（92.7-100）。
结论
我们的DL程序在根据标准护理超声检测和分类肝脂肪变性方面具有高敏感性和准确性。
Zotero 插件