Department of Radiology, Duke University Medical Center, DUMC Box 3808, Durham, NC, 27710, USA.
Department of Biostatistics, Gillings School of Global Public Health, The University of North Carolina at Chapel Hill, 135 Dauer Drive, 3101 McGavran-Greenberg Hall, CB #7420, Chapel Hill, NC, 27599, USA.
Abdom Radiol (NY). 2018 Mar;43(3):702-712. doi: 10.1007/s00261-017-1255-8.
Our objective is to determine the accuracy of multiparametric MRI (mpMRI) in predicting pathologic grade of prostate cancer (PCa) after radical prostatectomy (RP) using simple apparent diffusion coefficient metrics and, specifically, whether mpMRI can accurately separate disease into one of two risk categories (low vs. higher grade) or one of three risk categories (low, intermediate, or high grade) corresponding to the new prognostic grade group (PGG) criteria.
This retrospective, HIPAA-compliant, IRB-approved study included 140 patients with PCa who underwent 3 T mpMRI with endorectal coil and transrectal ultrasound-guided (TRUS-G) biopsy before RP. MpMRI was used to classify lesions using a two-tier (low-grade/PGG 1 vs. high-grade/PGG 2-5) or a three-tier system (low-grade/PGG 1 vs. intermediate-grade/PGG 2 vs. high-grade/PGG 3-5). Accuracy of mpMRI was compared against RP for each system.
The predictive accuracy of mpMRI using the two-tier system is higher than when using three-tier system (0.77 and 0.45, respectively). There were similar rates of undergrading between mpMRI and TRUS-G biopsy compared to RP (16% & 21%; respectively); rate of overgrading was higher for mpMRI vs. TRUS-G biopsy compared to RP (42% & 17%, respectively). When mpMRI and TRUS-G biopsy are combined, rate of undergrading is 1.4% and overgrading is 11%.
MpMRI predictive accuracy is higher when using a two-tier vs. a three-tier system, suggesting that advanced metrics may be necessary to delineate intermediate- from high-grade disease. Rates of under- and overgrading decreased when mpMRI and TRUS-G biopsy are combined, suggesting that these techniques may be complementary in predicting tumor grade.
本研究旨在使用简单的表观扩散系数指标，确定多参数 MRI（mpMRI）在预测根治性前列腺切除术后前列腺癌（PCa）病理分级的准确性，具体而言，就是评估 mpMRI 能否准确地将疾病分为低危和高危两个风险类别，或者能否准确地将疾病分为低危、中危和高危三个风险类别，以对应新的预后分级分组（PGG）标准。
这是一项回顾性研究，符合 HIPAA 规定，经过 IRB 批准，纳入了 140 例在 3T 场强下接受了直肠内线圈和经直肠超声引导（TRUS-G）活检的 PCa 患者。使用 mpMRI 对病变进行分类，采用两级（低危/PGG 1 与高危/PGG 2-5）或三级系统（低危/PGG 1 与中危/PGG 2 与高危/PGG 3-5）。比较两种系统下 mpMRI 与 RP 对 PCa 分级的准确性。
使用两级系统时，mpMRI 的预测准确性高于三级系统（分别为 0.77 和 0.45）。与 RP 相比，mpMRI 与 TRUS-G 活检的低估率相似（分别为 16%和 21%）；mpMRI 高估率高于 TRUS-G 活检与 RP（分别为 42%和 17%）。当将 mpMRI 和 TRUS-G 活检结合时，低估率为 1.4%，高估率为 11%。
使用两级系统时，mpMRI 的预测准确性高于三级系统，提示可能需要使用高级指标来区分中危和高危疾病。当将 mpMRI 和 TRUS-G 活检结合时，低估率和高估率降低，提示这两种技术在预测肿瘤分级方面可能具有互补性。
