Division of Gastroenterology, Washington University School of Medicine, 660 South Euclid Ave., Campus, Box 8124, Saint Louis, MO, 63110, USA.
Mallinckrodt Institute of Radiology, Washington University School of Medicine, Saint Louis, MO, USA.
Surg Endosc. 2022 Jun;36(6):4199-4206. doi: 10.1007/s00464-021-08752-z. Epub 2021 Oct 15.
Identifying patients likely to have CDL is an important clinical dilemma because endoscopic retrograde cholangiopancreatography (ERCP), carries a 5-7% risk of adverse events. The purpose of this study was to compare the diagnostic test performance of the 2010 and 2019 ASGE criteria used to help risk stratify patients with suspected CDL.
Consecutive patients evaluated for possible CDL from 2013 to 2019 were identified from surgical, endoscopic, and radiologic databases at a single academic center. Inclusion criteria included all patients who underwent ERCP and/or cholecystectomy with intraoperative cholangiogram (IOC) for suspected CDL. We calculated the diagnostic test performance of criteria from both guidelines and compared their discrimination using the receiver operator curve. Univariate and multivariate analysis was used to identify the strongest component predictors.
1098 patients [age 57.9 ± 19.0 years, 62.8% (690) F] were included. 66.3% (728) were found to have CDL on ERCP and/or IOC. When using the 2019 guidelines, the sensitivity, specificity, PPV, NPV, and accuracy are 65.8, 78.9, 86.3, 54.1, and 70.4%, respectively. Using the 2010 guidelines, the sensitivity, specificity, PPV, NPV, and accuracy are 50.5, 78.9, 82.5, 44.8, and 60.1%, respectively. The AUC for high-risk criteria using the 2019 guidelines [0.726 (0.695, 0.758)] was greater than for the 2010 guidelines [0.647 (0.614, 0.681)]. The key difference providing the increased discrimination was the inclusion of stones on any imaging modality, which increased the sensitivity to 55.0% from 29.1%. Not including CDL on imaging or cholangitis, a dilated CBD was the strongest individual predictor of CDL on multivariate analysis (OR 3.70, CI 2.80, 4.89).
Compared to 2010, the 2019 high-risk criterion improves diagnostic test performance, but still performs suboptimally. Less invasive tests, such as EUS or MRCP, should be considered in patients with suspected CDL prior to ERCP.
识别可能患有胆总管结石（CDL）的患者是一个重要的临床难题，因为内镜逆行胰胆管造影术（ERCP）有 5-7%的不良事件风险。本研究旨在比较 2010 年和 2019 年 ASGE 标准用于帮助 CDL 疑似患者进行风险分层的诊断测试性能。
从一家学术中心的外科、内镜和放射学数据库中连续确定了 2013 年至 2019 年期间因疑似 CDL 而接受评估的患者。纳入标准包括所有接受 ERCP 和/或胆囊切除术并在术中行胆管造影术（IOC）的患者。我们计算了两个指南中标准的诊断测试性能，并使用接收器工作特征曲线比较了它们的区分能力。使用单变量和多变量分析来确定最强的组成预测因素。
共纳入 1098 例患者[年龄 57.9±19.0 岁，62.8%（690 例）为女性]。66.3%（728 例）的患者在 ERCP 和/或 IOC 上发现 CDL。使用 2019 年指南时，敏感性、特异性、PPV、NPV 和准确性分别为 65.8%、78.9%、86.3%、54.1%和 70.4%。使用 2010 年指南时，敏感性、特异性、PPV、NPV 和准确性分别为 50.5%、78.9%、82.5%、44.8%和 60.1%。使用 2019 年指南的高危标准的 AUC [0.726（0.695，0.758）]大于使用 2010 年指南的 AUC [0.647（0.614，0.681）]。提高区分能力的关键区别是将任何影像学上的结石包括在内，从而将敏感性从 29.1%提高到 55.0%。在多变量分析中，不包括影像学上的 CDL 或胆管炎，扩张的 CBD 是 CDL 的最强个体预测因素（OR 3.70，CI 2.80，4.89）。
与 2010 年相比，2019 年的高危标准提高了诊断测试性能，但仍然表现不佳。在疑似 CDL 的患者进行 ERCP 之前，应考虑使用超声内镜或磁共振胰胆管造影等侵袭性较小的检查。
