CT影像组学鉴别肾上腺富血供腺瘤和嗜铬细胞瘤OA
Value of CT radiomics in discrimination of hypervascular adrenal adenoma from pheo-chromocytoma
目的 探讨CT影像组学模型在肾上腺富血供腺瘤(HAA)及嗜铬细胞瘤(AP)鉴别诊断中的价值.方法 回顾性纳入 2022-01 至 2024-07 在解放军总医院第三医学中心放射诊断科行 CT四期扫描并经病理证实的 79 例患者(HAA 53 例,AP 26 例),收集患者的临床指标、影像特征,并通过科研平台分期像提取影像组学特征,采用F_Test方法对影像组学特征降维和筛选,其中差异有统计学意义的因素纳入二元logistic回归分析及后续建模,最后绘制预测模型的ROC曲线及决策曲线以评估其预测效能和临床应用价值.结果 平扫CT值(OR=1.166,P<0.01)和肿瘤长径(OR=2.226,P<0.01)是鉴别两组的独立危险因素.基于CTpre、动脉期(CTa)、动脉晚期(CTla)、静脉期(CTv)、四期联合、影像特征及临床指标模型验证集的AUC分别为 0.93、0.92、0.88、0.86、0.87、0.88、0.72,基于CTpre影像组学效能最优,其准确度、敏感度、特异度分别为0.86、0.89、0.85.结论 基于CT影像组学模型在鉴别HAA和AP中具有较高的预测性能和应用价值.
Objective To explore the value of the CT radiomics model in the differential diagnosis of hypervascular adrenal adenoma(HAA)and pheochromocytoma(AP).Methods A total of 79 patients(53 HAA and 26 AP)who underwent phase 4 CT scan and were pathologically confirmed in the Department of Radiology of the Third Medical Center of PLA General Hospital from Janu-ary 2022 to July 2024 were retrospectively included.Clinical indicators and radiomics features extracted from CT images through scien-tific research platform were analyzed.Using the F_Test to reduce dimensionality and screen radiomics features,factors with statistically significant differences were included in binary logistic regression analysis and subsequent modeling.Finally,ROC curve and decision curve of the predictive model were plotted to evaluate its predictive performance and clinical application value.Results The value of plain CT scan(OR=1.166,P<0.01)and the long diameter of tumor(OR=2.226,P<0.01)were independent risk factors for the discrimination of the two groups.The AUC based on CTpre,arterial phase(CTa),late arterial phase(CTla),venous phase(CTv),four-phase combination,imaging features,and clinical indicator model validation set were 0.93,0.92,0.88,0.86,0.87,0.88 and 0.72,respectively.The CTpre-based radiomics model was characterized by the best performance,with accuracy,sensitivity,and spe-cificity of 0.86,0.89,and 0.85,respectively.Conclusions The model based om CT radiomics has good predictive performance and application value in the discrimination of HAA from AP.
程远华;陆静;滑蓉蓉;何绪成;张亚楠;王贵生;陈晓霞
636150,四川省宣汉县人民医院放射影像科100039 北京,解放军总医院第三医学中心放射诊断科100039 北京,解放军总医院第三医学中心放射诊断科100039 北京,解放军总医院第三医学中心放射诊断科100039 北京,解放军总医院第三医学中心放射诊断科100039 北京,解放军总医院第三医学中心放射诊断科100039 北京,解放军总医院第三医学中心放射诊断科
临床医学
CT影像组学肾上腺富血供腺瘤嗜铬细胞瘤鉴别
CTradiomicshypervascular adrenal adenomaadrenal pheochromocytomadiscrimination
《武警医学》 2025 (2)
154-158,5
评论