Zhang Hao, Zhang Li, Zhang Chi, Zhu Yan-Hao, Hong Yi-En, Li Lin, Lai Li
Department of Radiology, Dianjiang People's Hospital of Chongqing, Chongqing, 408300, People's Republic of China.
Department of Radiology, Changdu People's Hospital of Xizang, Xizang, 854000, People's Republic of China.
Sci Rep. 2025 Mar 28;15(1):10671. doi: 10.1038/s41598-025-94860-y.
To systematically analyze CT imaging features of hepatic cystic echinococcosis (CE), explore radiological-pathological correlations, and develop a diagnostic algorithm for accurate disease classification. This retrospective study included 48 pathologically confirmed cases of hepatic CE from two medical centers. CT imaging features were analyzed by two experienced radiologists, evaluating lesion characteristics including location, morphology, wall features, and calcification patterns. Imaging findings were correlated with pathological results. A diagnostic algorithm was developed and validated, with inter-observer agreement assessed using Fleiss kappa coefficient. Seven distinct CT imaging patterns were identified, corresponding to different pathological stages: unilocular cystic (25.0%), multivesicular (8.3%), collapsed inner wall (10.4%), partially solidified (10.4%), solidified (16.7%), and calcified (25.0%) types, with complicated cases (4.2%) showing additional features. The proposed diagnostic algorithm achieved 94.0% accuracy (451/480 classifications) in validation testing by ten junior radiologists, with excellent inter-observer agreement (quadratic-weighted Fleiss kappa coefficient = 0.740 [95% CI 0.577-0.902], Gwet's AC2 coefficient = 0.768). Primary diagnostic challenges involved differentiating between CE2 and CE3b lesions, and between CE3b and CE4 lesions. This study explores the correlation between CT imaging patterns and pathological stages of hepatic CE, proposing a validated diagnostic algorithm. The findings provide valuable insights for CE classification, particularly in regions where the disease is emerging or underrecognized.
系统分析肝囊性棘球蚴病（CE）的CT成像特征，探索放射学与病理学的相关性，并开发一种用于准确疾病分类的诊断算法。这项回顾性研究纳入了来自两个医疗中心的48例经病理证实的肝CE病例。由两名经验丰富的放射科医生分析CT成像特征，评估病变特征，包括位置、形态、壁特征和钙化模式。将影像学表现与病理结果进行相关性分析。开发并验证了一种诊断算法，使用Fleiss卡方系数评估观察者间的一致性。识别出七种不同的CT成像模式，对应不同的病理阶段：单房囊性（25.0%）、多房性（8.3%）、内壁塌陷（10.4%）、部分凝固（10.4%）、凝固（16.7%）和钙化（25.0%）类型，复杂病例（4.2%）表现出其他特征。在十名初级放射科医生进行的验证测试中，所提出的诊断算法准确率达到94.0%（480次分类中的451次），观察者间一致性良好（二次加权Fleiss卡方系数 = 0.740 [95% CI 0.577 - 0.902]，Gwet's AC2系数 = 0.768）。主要诊断挑战包括区分CE2和CE3b病变，以及CE3b和CE4病变。本研究探讨了肝CE的CT成像模式与病理阶段之间的相关性，提出了一种经过验证的诊断算法。研究结果为CE分类提供了有价值的见解，特别是在该疾病正在出现或未得到充分认识的地区。
