Kaffey Zaphanlene, Mirbahaeddin Sarah, Wahid Kareem, Kamel Serageldin, Vouffo Michael, Otun Adegbenga O, Belal Zayne, Aponte Wesson Ruth A, Carriere Patrick P, Dede Cem, Maniakas Anastasios, Goepfert Ryan P, Garden Adam S, Lee Anna, Choi Karen Y, Christensen Joani M, Lincoln Christie, Manzar Gohar S, Dudzinski Stephanie O, Thomas Rehema, Mehrens Hunter, Cardoso Richard C, Schellingerhout Dawid, Watson Erin, Chen Melissa M, Lai Stephen Y, Fuller Clifton D, Moreno Amy C, Humbert-Vidan Laia
Division of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, USA; Division of Radiology, The University of Texas MD Anderson Cancer Center, Houston, USA.
Division of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, USA.
Radiother Oncol. 2025 Jul;208:110917. doi: 10.1016/j.radonc.2025.110917. Epub 2025 May 2.
BACKGROUND & PURPOSE: Osteoradionecrosis (ORN) of the jaw is a severe complication affecting up to 15% of head and neck cancer patients treated with radiotherapy. The ClinRad system, endorsed by ASCO/ISOO/MASCC, incorporates radiographic features for ORN severity classification, but variability in imaging use and specialty expertise may impact diagnostic accuracy. This study benchmarks physician performance in diagnosing and staging ORN across specialties and imaging modalities.
MATERIALS & METHODS: A retrospective diagnostic validation study was conducted at MD Anderson Cancer Center, involving 20 physicians from oral oncology, radiation oncology, surgery, and neuroradiology. Participants reviewed 85 de-identified imaging sets (CT and orthopantogram (OPG)) from 30 patients with confirmed ORN, diagnosing and staging cases using the ClinRad system. ROC analysis assessed diagnostic accuracy, while intra- and inter-observer agreement was measured using Cohen's and Fleiss kappa statistics.
Paired CT-OPG imaging significantly improved diagnostic performance across specialties (p < 0.001), with AUC values ranging from 0.79 (residents) to 0.98 (surgeons). However, inter- and intra-rater agreement remained low, with median Fleiss kappa values of 0.22, 0.13, and 0.05 for ClinRad stages 0/1, 2, and 3, respectively. No specialty demonstrated significantly superior diagnostic accuracy (p > 0.05).
This study establishes a benchmark for radiographic ORN detection, revealing diagnostic variability across specialties. Findings emphasize the need for standardized imaging protocols, interdisciplinary training, and multimodal imaging to improve diagnostic accuracy.
颌骨放射性骨坏死（ORN）是一种严重并发症，在接受放射治疗的头颈癌患者中发生率高达15%。美国临床肿瘤学会（ASCO）/国际口腔肿瘤学会（ISOO）/多学科癌症支持治疗学会（MASCC）认可的ClinRad系统纳入了用于ORN严重程度分类的影像学特征，但成像使用的差异和专业知识可能会影响诊断准确性。本研究对各专业和成像方式在ORN诊断和分期方面的医生表现进行了基准测试。
在MD安德森癌症中心进行了一项回顾性诊断验证研究，涉及来自口腔肿瘤学、放射肿瘤学、外科和神经放射学的20名医生。参与者查看了30例确诊ORN患者的85套去识别化影像集（CT和全景曲面断层片（OPG）），并使用ClinRad系统对病例进行诊断和分期。ROC分析评估诊断准确性，观察者内和观察者间一致性使用科恩（Cohen）和弗莱iss卡方统计量进行测量。
配对的CT-OPG成像显著提高了各专业的诊断性能（p < 0.001）,AUC值范围从0.79（住院医师）到0.98（外科医生）。然而，评分者间和评分者内一致性仍然较低，ClinRad 0/1期、2期和3期的弗莱iss卡方值中位数分别为0.22、0.13和0.05。没有哪个专业表现出显著更高的诊断准确性（p > 0.05）。
本研究建立了影像学ORN检测的基准，揭示了各专业之间的诊断差异。研究结果强调需要标准化成像方案、跨学科培训和多模态成像以提高诊断准确性。
