Miao Lei, Jiang JiuMing, Li JianWei, Liu Hui, Hu SiJie, Zhang HuanHuan, Gong LiHua, Zhang YuHeng, Wang SiCong, Quan GuangNan, Li Xiao, Li Meng
Department of Interventional Therapy, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, 100021, China.
Department of Radiology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, 100021, China.
Eur Radiol. 2025 Jun 26. doi: 10.1007/s00330-025-11770-3.
To determine the diagnostic potential of histogram features derived from synthetic magnetic resonance imaging (SyMRI) for the differentiation of benign and malignant soft-tissue tumors (STTs).
Seventy-two patients with STTs of the extremities or trunk (29 benign, 43 malignant) were prospectively enrolled. Quantitative histogram features, extracted from T1, T2, and proton density (PD) images, were obtained by delineating the three-dimensional volume of interest (VOI). Patients were divided into a benign group and a malignant group according to their pathological results. Logistic regression was used to construct diagnostic models. Receiver operating characteristic (ROC) analysis was performed to assess diagnostic performance.
Several histogram features derived from SyMRI, such as PD_Energy, T1_Skewness, and T2_Kurtosis, were significantly elevated in patients with malignant STTs (p < 0.05). All the models constructed on the basis of SyMRI histogram features performed comparably to or better than the Clinical_Model (a model based on conventional MRI features). The Combined_Model (a model integrating histogram features from the PD, T1, and T2 maps) has the highest AUC (0.946).
Whole-tumor histogram analysis via SyMRI can distinguish benign from malignant STTs well. The Combined model, incorporating features from the PD, T1, and T2 models, is effective. Future efforts should focus on external validation across multiple institutions and prospective prognostic validation studies.
Question Differentiating between benign and malignant STTs is crucial for effective treatment, but current imaging methods struggle to reliably distinguish these tumor types. Findings Whole-tumor histogram analysis using SyMRI identified significant differences in quantitative features between benign and malignant STTs with high diagnostic accuracy. Clinical relevance This non-invasive imaging approach enhances diagnostic precision, reducing the need for biopsies and assisting clinicians in developing tailored treatment strategies for soft-tissue tumor patients.
确定源自合成磁共振成像（SyMRI）的直方图特征对鉴别良性和恶性软组织肿瘤（STT）的诊断潜力。
前瞻性纳入72例四肢或躯干STT患者（29例良性，43例恶性）。通过勾勒三维感兴趣体积（VOI），从T1、T2和质子密度（PD）图像中提取定量直方图特征。根据病理结果将患者分为良性组和恶性组。采用逻辑回归构建诊断模型。进行受试者操作特征（ROC）分析以评估诊断性能。
源自SyMRI的几个直方图特征，如PD_能量、T1_偏度和T2_峰度，在恶性STT患者中显著升高（p < 0.05）。基于SyMRI直方图特征构建的所有模型的表现与临床模型（基于传统MRI特征的模型）相当或更好。联合模型（整合来自PD、T1和T2图谱的直方图特征的模型）具有最高的曲线下面积（AUC，0.946）。
通过SyMRI进行的全肿瘤直方图分析能够很好地区分良性和恶性STT。结合PD、T1和T2模型特征的联合模型是有效的。未来的工作应集中在多机构的外部验证和前瞻性预后验证研究上。
问题鉴别良性和恶性STT对有效治疗至关重要，但目前的成像方法难以可靠地区分这些肿瘤类型。发现使用SyMRI进行全肿瘤直方图分析发现良性和恶性STT在定量特征上存在显著差异，诊断准确性高。临床意义这种非侵入性成像方法提高了诊断精度，减少了活检的必要性，并有助于临床医生为软组织肿瘤患者制定个性化治疗策略。
