虚拟磁共振弹性成像与连续时间随机游走模型预测临床显著性前列腺癌的效能评估
本研究回顾性纳入107例疑似前列腺病变患者，比较vMRE、CTRW、常规DWI及临床指标在鉴别临床显著性前列腺癌（csPCa，Gleason评分≥3+4）中的表现。多因素分析确定μ、α、Dm和PSAD为独立预测因子，联合模型AUC达0.98，显著优于各单一影像模态及单独变量。vMRE和CTRW均优于常规DWI，联合面板可作为无创高性能生物标志物辅助减少不必要穿刺。
要点
- 联合μ、α、Dm和PSAD的模型AUC为0.98，显著优于单一影像参数。
- vMRE（μ）与CTRW（α+β+Dm）诊断效能均优于常规DWI（ADC）。
- csPCa组μ、PSAD、PI-RADS、TPSA更高，α、β、Dm、ADC更低。
关键影像表现/征象
- csPCa组μ、PSAD、PI-RADS、TPSA高于非csPCa组，α、β、Dm、ADC低于非csPCa组。
- 多因素分析显示μ、α、Dm和PSAD为独立预测因子。
- 联合模型AUC为0.98，显著优于vMRE（μ）的0.85、CTRW（α+β+Dm）的0.92、DWI（ADC）的0.78。
- 联合模型亦显著优于单独α（AUC=0.87）、β（AUC=0.74）、Dm（AUC=0.82）、PSAD（AUC=0.65）、PI-RADS（AUC=0.84）和TPSA（AUC=0.76），所有P<0.05。
诊断标准与关键数值
- csPCa定义：Gleason评分≥3+4。
- 联合模型（μ+α+Dm+PSAD）诊断csPCa的AUC为0.98。
- vMRE（μ）AUC为0.85，CTRW（α+β+Dm）AUC为0.92，DWI（ADC）AUC为0.78。
- 单独参数AUC：α为0.87，β为0.74，Dm为0.82，PSAD为0.65，PI-RADS为0.84，TPSA为0.76。
临床应用与阅片要点
- 在疑似前列腺癌患者中，可结合vMRE的μ值和CTRW的α、Dm参数辅助鉴别csPCa。
- 联合PSAD可进一步提升预测效能，AUC达0.98，有助于筛选可避免穿刺的患者。
- vMRE和CTRW参数均优于常规ADC，阅片时可关注μ升高及α、Dm降低提示csPCa可能。
局限性
- 回顾性研究设计。
- 单中心样本量有限（107例）。
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
OBJECTIVE: To evaluate the predictive performance of virtual magnetic resonance elastography (vMRE) and continuous-time random walk (CTRW) models for clinically significant prostate cancer (csPCa, Gleason score ≥ 3 + 4).
METHODS: A total of 107 patients with suspected prostate lesions were retrospectively enrolled. Quantitative parameters from vMRE (µ), CTRW (α, β, Dm), conventional DWI (ADC), and clinical indicators were compared between csPCa and non-csPCa groups. Logistic regression was used to identify independent predictors and construct a combined model, with efficacy validated by receiver operating characteristic curve (AUC) and DeLong test.
RESULTS: The csPCa group had higher µ, PSAD, PI-RADS, and TPSA, and lower α, β, Dm, and ADC. Multivariate analysis identified µ, α, Dm, and PSAD as independent predictors. The combined model integrating these four parameters yielded the optimal diagnostic performance with an AUC of 0.98, which was significantly superior to single imaging modalities and individual variables, including vMRE (µ), CTRW (α + β + Dm), DWI (ADC), α, β, Dm, PSAD, PI-RADS, and TPSA (AUC = 0.85, 0.92, 0.78, 0.87, 0.74, 0.82, 0.65, 0.84, 0.76, and 0.80, respectively; all P < 0.05).
CONCLUSION: Both vMRE and CTRW are superior to conventional DWI in differentiating csPCa. The combined µ + α + Dm + PSAD panel serves as a non-invasive, high-performance biomarker for accurate csPCa prediction, which may help reduce unnecessary prostate biopsies and optimize clinical decision-making.
引用
- Zhang M, Zhao E, Li X, et al. Evaluating the efficacy of virtual magnetic resonance elastography and continuous-time random walk models in predicting clinically significant prostate cancer. Abdom Radiol (NY). 2026. PMID: 42240643. doi:10.1007/s00261-026-05549-3
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。