目的 探讨基于双模态超声(二维超声联合剪切波弹性超声)与双参数MRI(T2WI联合ADC)的影像组学特征,构建并验证一个用于术前无创预测前列腺癌(PCa)的联合模型。方法 选取2023-01至2025-10在航天中心医院就诊的前列腺占位患者共232例,其中PCa 127例,良性105例。所有患者均接受双模态超声及双参数MRI检查。手动勾画病灶后,提取高通量影像组学特征。采用LASSO回归进行特征筛选,并整合临床独立风险因素(年龄、PSA密度)。随后,构建并比较临床模型、超声影像组学模型、MRI影像组学模型及联合模型。采用五折交叉验证评估性能。结果 联合模型在验证集中展现出最优的诊断效能,其曲线下面积(AUC)为0.92 (95% CI: 0.84~0.97),显著优于临床模型(AUC=0.76)、超声模型(AUC=0.78)及MRI模型(AUC=0.85)(所有P<0.05)。决策曲线分析证实其在较宽阈值范围内能提供更大的临床净获益。结论 本研究构建的多模态影像组学与临床因素的联合模型,可以提升PCa术前无创诊断的准确率,具有良好的稳健性与临床适用性。
Abstract
Objective To construct and validate a combined model for preoperative noninvasive prediction of prostate cancer (PCa) using radiomic features from dual-modal ultrasound (B-mode+shear wave elastography) and biparametric MRI (T2WI+ADC). Methods A total of 232 patients with prostate mass lesions who visited the Aerospace Center Hospital from January 2023 to October 2025 were selected, including 127 cases of PCa and 105 cases of benign conditions. All patients underwent dual-modal ultrasound and biparametric MRI examinations. After manual delineation of the lesions, high-throughput radiomics features were extracted. Feature selection was performed using LASSO regression, and independent clinical risk factors (age, PSA density) were integrated. Four models were constructed and compared: a clinical model, an ultrasound radiomics model, an MRI radiomics model, and a combined model. The performance was evaluated via five-fold cross-validation. Results The combined model demonstrated the best diagnostic efficacy in the validation set, with an AUC of 0.92 (95% CI: 0.84-0.97), significantly higher than the AUCs of the clinical model (0.76), the ultrasound radiomics model (0.78), and the MRI radiomics model (0.85) (all P<0.05). Decision curve analysis confirmed its greater net clinical benefit across a wide threshold range. Conclusions The validated combined model integrating multimodal radiomics and clinical factors can significantly improve the accuracy of preoperative noninvasive diagnosis of PCa, demonstrating robust performance and high clinical applicability.
关键词
前列腺癌 /
影像组学 /
超声 /
核磁共振 /
机器学习 /
预测模型
Key words
prostate cancer /
radiomics /
ultrasound /
MRI /
machine learning /
predictive model
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] Pinsky P F, Parnes H. Screening for prostate cancer[J]. N Engl J Med,2023,388: 1405-1414.
[2] Cancer facts & figures 2024[EB/OL]. Accessed: December 16, 2024: https://www.cancer.org/research/ cancer-factsstatistics/all-cancer-facts-figures/2024-cancer-facts-figures.htm
[3] Hugosson J, Månsson M, Wallström J, et al. Prostate cancer screening with PSA and MRI followed by targeted biopsy only[J]. N Engl J Med, 2022, 387: 2126-2137.
[4] Cho H, Oh C K, Cha J, et al. Association of serum prostate-specific antigen (PSA) level and circulating tumor cell-based PSA mRNA in prostate cancer[J]. Prostate Int, 2022,10(1):14-20.
[5] Teramoto Y, Numbere N, Wang Y, et al. Clinical significance of perineural invasion by prostate cancer detected on needle core biopsy[J]. Am J Clin Pathol, 2023,159:116-119.
[6] Nedelcu A, Oerther B, Benkendorff A, et al. PI-RADS Version 2.1 for Prostate MRI interpretation: associations of study quality and cancer detection metrics-a systematic review and meta-analysis[J]. Am J Roentgenol, 2025.Doi: 10.2214/AJR.25.33583.
[7] Lim C S, Abreu-Gomez J, Thornhill R, et al. Utility of machine learning of apparent diffusion coefficient (ADC) and T2-weighted (T2W) radiomic features in PI-RADS version 2.1 category 3 lesions to predict prostate cancer diagnosis[J]. Abdom Radiol (NY), 2021,46(12):5647-5658.
[8] Shiva M, Wei C, Molana H, et al. Costeffectiveness of prostate cancer detection in biopsy-naïve men: ultrasound shear wave elastography vs. multiparametric diagnostic magnetic resonance imaging[J]. Healthcare (Basel),2022,10(2):254.
[9] Keskin E T, Kaplanoglu V, Senocak C, et al. Transrectal shear wave elastography for detection of prostate cancer[J]. Urologia, 2023,90(2):230-235.
[10] Cavallo A U, Stanzione A, Ponsiglione A, et al.Prostate cancer MRI methodological radiomics score: a EuSoMII radiomics auditing group initiative[J]. Eur Radiol, 2025,35(3):1157-1165.
[11] Liang L, Zhi X, Sun Y, et al. A Nomogram based on a multiparametric ultrasound radiomics model for discrimination between malignant and benign prostate lesions[J]. Front Oncol, 2021,11:610785.
[12] Sun Y, Fang J, Shi Y, et al. Machine learning based on radiomics features combing B-mode transrectal ultrasound and contrast-enhanced ultrasound to improve peripheral zone prostate cancer detection[J]. Abdom Radiol (NY), 2024 ,49(1):141-150.
[13] Stabile A, Giganti F, Rosenkrantz A B, et al. Multiparametric MRI for prostate cancer diagnosis: current status and future directions[J]. Nat Rev Urol, 2020,17(1):41-61.
[14] Brancato V, Aiello M, Basso L, et al. (2021) Evaluation of a multiparametric MRI radiomic-based approach for stratification of equivocal PI-RADS 3 and upgraded PI-RADS 4 prostatic lesions[J]. Sci Rep,2021,11(1):643.
[15] Giambelluca D, Cannella R, Vernuccio F, et al. (2021) PI-RADS 3 lesions: role of prostate mri texture analysis in the identification of prostate cancer[J]. Curr Probl Diagn Radiol, 2021,50(2):175-185.
[16] Israel B, Leest M V, Sedelaar M, et al. Multiparametric magnetic resonance imaging for the detection of clinically significant prostate cancer: what urologists need to know. Part 2: interpretation[J]. Eur Urol, 2020,77(4):469-480.
[17] Jiang W, Lai B, Li X, et al. Integrating shear wave elastography and multiparametric MRI for accurate prostate cancer diagnosis[J]. Am J Cancer Res, 2025,15(1):348-362.
[18] Liu D, Lin W, Wei C, et al. Exploring the potential clinical value of shear wave elastography in prostate cancer: a meta-analysis[J]. BMC Cancer, 2025,25(1):1032.
[19] Zaurito P, Stabile A, Montorsi F, et al. The prognostic role of prostate MRI in prostate cancer patients[J]. Curr Opin Urol, 2025,35(4):418-425.
[20] Morote J, Borque-Fernando Á, Triquell M, et al. A clinically significant prostate cancer predictive model using digital rectal examination prostate volume category to stratify initial prostate cancer suspicion and reduce magnetic resonance imaging demand[J]. Cancers (Basel), 2022,14(20):5100.
基金
国家自然科学基金项目(62371010),北京市海淀区卫生健康发展科研培育计划(HP2024-32-507004)
PDF(1922 KB)
