Department of Pathology, Cochin Hospital, AP-HP, Paris Descartes University, Paris, France.
Am J Surg Pathol. 2012 Aug;36(8):1194-201. doi: 10.1097/PAS.0b013e31825a6308.
The Weiss score is the reference method to distinguish between a benign and a malignant adrenocortical tumor (ACT). A program was initiated to improve the reproducibility of the pathologic diagnosis of ACTs in France through the National INCa-COMETE Network. Twelve pathologists from all Reference Centers of the Network analyzed 50 selected ACTs using a web-based virtual microscopy approach in a blind design, allowing to determine the intraobserver and interobserver reproducibilities of the Weiss system. All ACTs were read twice in random order before and after a coaching meeting organized to harmonize and improve analyses and create an online tutorial. The validity of the virtual approach was first established by comparing the 2 consensuses (virtual and microscopic) obtained for each tumor by 3 pathologists who performed the 2 approaches in a blinded manner. For the "dichotomized Weiss score" (separating malignant ≥3 from benign ≤2 tumors) interobserver reproducibility was "substantial" at the first "virtual" reading (κ = 0.70) and increased at the second "virtual" reading (κ = 0.75). In parallel, 7 of the 9 items of the Weiss system showed improvement. The diagnostic accuracy of the observers as a group, using the modal group score approach, showed an improved sensitivity from 86% to 95% for the diagnosis of malignant ACTs. We show the validity of the virtual microscopy approach and that the program improved the practice of the Weiss system reading and therefore the diagnosis of ACT. This tool can now be extended for other research and/or routine purposes in this rare cancer.
魏斯评分是鉴别良性和恶性肾上腺皮质肿瘤(ACT)的参考方法。通过国家 INCa-COMETE 网络,启动了一项计划,旨在提高法国 ACT 病理诊断的可重复性。网络的所有参考中心的 12 位病理学家使用基于网络的虚拟显微镜方法在盲法设计中分析了 50 例选定的 ACT,以确定魏斯系统的观察者内和观察者间可重复性。在组织了一次协调和改进分析并创建在线教程的辅导会议之前和之后,所有 ACT 均以随机顺序进行了两次重读。首先通过比较 3 位病理学家以盲法进行的两种方法(虚拟和显微镜)获得的每个肿瘤的 2 个共识,确定了虚拟方法的有效性。对于“二分魏斯评分”(将恶性≥3 与良性≤2 的肿瘤分开),第一次“虚拟”阅读时观察者间的可重复性为“中等”(κ=0.70),第二次“虚拟”阅读时可重复性增加(κ=0.75)。同时,魏斯系统的 9 项中的 7 项有所改善。作为一个群体的观察者的诊断准确性,使用模态组评分方法,在诊断恶性 ACT 方面,敏感性从 86%提高到 95%。我们证明了虚拟显微镜方法的有效性,并且该计划提高了魏斯系统阅读的实践,从而提高了 ACT 的诊断。现在,该工具可以扩展到这种罕见癌症的其他研究和/或常规用途。