基于MRI的前列腺周围脂肪与肿瘤病灶影像组学预测前列腺癌从活检到根治性前列腺切除术的病理升级
本研究纳入314例经根治性前列腺切除术病理确诊的前列腺癌患者,按1:1随机分为训练队列和验证队列,从T2WI和ADC序列提取前列腺周围脂肪和肿瘤的影像组学特征,联合临床数据建立联合模型。联合模型在训练队列预测Gleason评分升级的AUC为0.925(95%CI 0.872-0.979),验证队列AUC为0.937(95%CI 0.891-0.983),校准良好,决策曲线分析显示临床实用性好。
要点
- 联合前列腺周围脂肪、肿瘤病灶影像组学与临床数据的模型可预测活检至根治术的Gleason评分升级。
- 训练队列AUC=0.925(95%CI 0.872-0.979),验证队列AUC=0.937(95%CI 0.891-0.983)。
- 列线图校准良好,决策曲线分析提示临床实用性较好。
核心结论
- 整合前列腺周围脂肪、前列腺癌病灶影像组学特征与临床数据的联合模型在预测前列腺癌患者从活检到根治性前列腺切除术的Gleason评分升级方面表现优异,为Gleason评分升级风险分层提供了一种新的、可靠的无创工具。
研究设计与样本
- 回顾性纳入314例经根治性前列腺切除术病理确诊的前列腺癌患者,按1:1随机分为训练队列(n=157)和验证队列(n=157);所有患者术前接受MRI检查及经直肠超声引导下前列腺穿刺活检;从T2加权成像(T2WI)和表观扩散系数(ADC)序列提取前列腺周围脂肪和肿瘤的影像组学特征;采用单因素和多因素logistic回归筛选独立临床危险因素,并整合前列腺周围脂肪与前列腺癌影像组学特征建立联合模型;使用受试者工作特征(ROC)曲线、校准和决策曲线分析评估模型性能。
诊断标准与关键数值
- 联合模型(整合前列腺周围脂肪、前列腺癌影像组学特征及临床数据)在训练队列预测Gleason评分升级的AUC=0.925(95%CI 0.872-0.979),Hosmer-Lemeshow检验χ2=9.316,P=0.316。
- 联合模型在验证队列预测Gleason评分升级的AUC=0.937(95%CI 0.891-0.983),校准良好,Hosmer-Lemeshow检验χ2=12.871,P=0.116。
临床应用与阅片要点
- 作者认为,整合前列腺周围脂肪、前列腺癌病灶影像组学特征与临床数据的联合模型可用于预测活检至根治性前列腺切除术的Gleason评分升级风险分层。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
RATIONALE AND OBJECTIVES: To assess the predictive value of MRI-based radiomics of periprostatic fat (PPF) and tumor lesions for predicting Gleason score (GS) upgrading from biopsy to radical prostatectomy (RP) in prostate cancer (PCa).
METHODS: A total of 314 patients with pathologically confirmed prostate cancer (PCa) after radical prostatectomy (RP) were included in the study. The patients were randomly assigned to the training cohort (n = 157) and the validating cohort (n = 157) in a 1:1 ratio. All had pre-surgery MRI followed by transrectal ultrasound-guided prostate biopsy. Radiological features were extracted from T2-weighted imaging (T2WI) and apparent diffusion coefficient (ADC) sequences for PPF and tumors. Univariate and multivariate logistic regression identified independent clinical risk factors, and a combined model was established by integrating radiomic features of PPF and PCa. Model performance was assessed using receiver operating characteristic (ROC) curves, calibration, and decision curve analysis.
RESULTS: The combined model, incorporating radiomic features of PPF, PCa, and clinical data, predicted GS upgrading from biopsy to RP excellently (AUC=0.925, 95%CI0.872-0.979) in the training cohort. The Hosmer-Lemeshow test confirmed model fit (χ2 = 9.316, P = 0.316). The nomogram was validated in the validating cohort; it showed good accuracy (AUC= 0.937, 95% CI, 0.891-0.983) and was well calibrated (χ2 = 12.871, P = 0.116). Decision curve analysis indicated good clinical utility of the radiomic nomogram.
CONCLUSION: The combined model incorporating PPF, PCa, and clinical data showed excellent performance in predicting GS upgrading from biopsy to RP in PCa patients. This offers a novel and reliable noninvasive tool for GS upgrading risk stratification.
引用
- Liu WQ, Wei Y, Ke ZB, et al. Radiomics of Periprostatic Fat and Tumor Lesion Based on MRI Predicts the Pathological Upgrading of Prostate Cancer from Biopsy to Radical Prostatectomy. Acad Radiol. 2024. PMID: 39730248. doi:10.1016/j.acra.2024.11.043
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。