Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota, USA.
Independent Researcher, Chelsea, Massachusetts, USA.
Gut. 2021 Jul;70(7):1335-1344. doi: 10.1136/gutjnl-2020-322821. Epub 2020 Oct 7.
The diagnosis of autoimmune pancreatitis (AIP) is challenging. Sonographic and cross-sectional imaging findings of AIP closely mimic pancreatic ductal adenocarcinoma (PDAC) and techniques for tissue sampling of AIP are suboptimal. These limitations often result in delayed or failed diagnosis, which negatively impact patient management and outcomes. This study aimed to create an endoscopic ultrasound (EUS)-based convolutional neural network (CNN) model trained to differentiate AIP from PDAC, chronic pancreatitis (CP) and normal pancreas (NP), with sufficient performance to analyse EUS video in real time.
A database of still image and video data obtained from EUS examinations of cases of AIP, PDAC, CP and NP was used to develop a CNN. Occlusion heatmap analysis was used to identify sonographic features the CNN valued when differentiating AIP from PDAC.
From 583 patients (146 AIP, 292 PDAC, 72 CP and 73 NP), a total of 1 174 461 unique EUS images were extracted. For video data, the CNN processed 955 EUS frames per second and was: 99% sensitive, 98% specific for distinguishing AIP from NP; 94% sensitive, 71% specific for distinguishing AIP from CP; 90% sensitive, 93% specific for distinguishing AIP from PDAC; and 90% sensitive, 85% specific for distinguishing AIP from all studied conditions (ie, PDAC, CP and NP).
The developed EUS-CNN model accurately differentiated AIP from PDAC and benign pancreatic conditions, thereby offering the capability of earlier and more accurate diagnosis. Use of this model offers the potential for more timely and appropriate patient care and improved outcome.
自身免疫性胰腺炎（AIP）的诊断具有挑战性。AIP 的超声和横断面成像表现与胰腺导管腺癌（PDAC）非常相似，且 AIP 的组织取样技术并不理想。这些局限性常常导致诊断延迟或失败，从而对患者的管理和预后产生负面影响。本研究旨在创建一种基于内镜超声（EUS）的卷积神经网络（CNN）模型，该模型经过训练可区分 AIP 与 PDAC、慢性胰腺炎（CP）和正常胰腺（NP），并具有足够的性能实时分析 EUS 视频。
使用来自 AIP、PDAC、CP 和 NP 病例 EUS 检查的静态图像和视频数据数据库来开发 CNN。采用闭塞热图分析来识别 CNN 在区分 AIP 与 PDAC 时重视的超声特征。
从 583 名患者（146 例 AIP、292 例 PDAC、72 例 CP 和 73 例 NP）中提取了总计 1174461 个独特的 EUS 图像。对于视频数据，CNN 每秒处理 955 个 EUS 帧，其表现为：99%的敏感性，98%的特异性，用于区分 AIP 与 NP；94%的敏感性，71%的特异性，用于区分 AIP 与 CP；90%的敏感性，93%的特异性，用于区分 AIP 与 PDAC；90%的敏感性，85%的特异性，用于区分 AIP 与所有研究条件（即 PDAC、CP 和 NP）。
开发的 EUS-CNN 模型可准确地区分 AIP 与 PDAC 和良性胰腺疾病，从而提供更早和更准确的诊断能力。该模型的使用具有提供更及时和适当的患者护理以及改善预后的潜力。
