Zheng Bowen, Mo Futian, Shi Xiaoran, Li Wenhao, Shen Quanyou, Zhang Ling, Liao Zhongjian, Fan Cungeng, Liu Yanping, Zhong Junyuan, Qin Genggeng, Tao Jie, Lv Shidong, Wei Qiang
Department of Urology, Nanfang Hospital, Southern Medical University, Guangzhou, Guangdong 510515, China (B.Z., F.M., X.S., S.L., Q.W.).
School of Automation, Guangdong University of Technology, Guangzhou, Guangdong 510006, China (W.L., Q.S., J.T.); Guangdong Provincial Key Laboratory of Intelligent Decision and Cooperative Control, Guangzhou, Guangdong 510006, China (W.L., Q.S., J.T.); Guangdong-Hong Kong Joint Laboratory for Intelligent Decision and Cooperative Control, Guangzhou, Guangdong 510006, China (W.L., Q.S., J.T.).
Acad Radiol. 2025 May;32(5):2709-2722. doi: 10.1016/j.acra.2024.12.012. Epub 2025 Jan 2.
To develop an automatic deep-radiomics framework that diagnoses and stratifies prostate cancer in patients with prostate-specific antigen (PSA) levels between 4 and 10 ng/mL.
A total of 1124 patients with histological results and PSA levels between 4 and 10 ng/mL were enrolled from one public dataset and two local institutions. An nnUNet was trained for prostate masks, and a feature extraction module identified suspicious lesion masks. Radiomics features were extracted from the biparametric magnetic resonance imaging using these masks. Machine learning models were developed to diagnose prostate cancer (PCa), clinically significant PCa (csPCa), and high-risk csPCa based on radiomics and clinical features. The models were evaluated in both internal and external cohorts. The best model was further compared with PSA density (PSAD), free to total PSA (F/T PSA), and Prostate Imaging Reporting and Data System (PI-RADS) scores in the external cohort.
The models based on both radiomics and clinical features outperformed those based on radiomics or clinical features alone. The top-performing models achieved areas under the curve of 0.80, 0.88, and 0.83 on internal testing, and 0.79, 0.80, and 0.82 on external testing for diagnosing PCa, csPCa, and high-risk csPCa. Our deep-radiomics model surpassed PSAD, F/T PSA, and PI-RADS scores in an external cohort. Decision curve analysis indicated that our model offers greater net benefit than these methods.
The deep-radiomics model automatically segments prostate and suspicious lesions, diagnoses, and stages of PCa in patients with PSA levels between 4 and 10 ng/mL. Our method addresses the shortcomings of manual segmentation and inconsistency, delivering outstanding performance. It provides multilevel predictions to assist clinical decision-making and benefit patients with gray zone PSA.
开发一种自动深度放射组学框架，用于对前列腺特异性抗原（PSA）水平在4至10 ng/mL之间的前列腺癌患者进行诊断和分层。
从一个公共数据集和两个本地机构招募了1124例组织学结果明确且PSA水平在4至10 ng/mL之间的患者。使用nnUNet训练前列腺掩码，一个特征提取模块识别可疑病变掩码。利用这些掩码从双参数磁共振成像中提取放射组学特征。基于放射组学和临床特征开发机器学习模型，以诊断前列腺癌（PCa）、临床显著性前列腺癌（csPCa）和高危csPCa。在内部和外部队列中对模型进行评估。在外部队列中将最佳模型与PSA密度（PSAD）、游离PSA与总PSA（F/T PSA）以及前列腺影像报告和数据系统（PI-RADS）评分进行进一步比较。
基于放射组学和临床特征的模型优于仅基于放射组学或临床特征的模型。在内部测试中，表现最佳的模型在诊断PCa、csPCa和高危csPCa时的曲线下面积分别为0.80、0.88和0.83，在外部测试中分别为0.79、0.80和0.82。我们的深度放射组学模型在外部队列中超过了PSAD、F/T PSA和PI-RADS评分。决策曲线分析表明，我们的模型比这些方法具有更大的净效益。
深度放射组学模型可自动分割前列腺和可疑病变，对PSA水平在4至10 ng/mL之间的前列腺癌患者进行诊断和分期。我们的方法克服了手动分割的缺点和不一致性，表现出色。它提供多层次预测以辅助临床决策，使处于PSA灰色地带的患者受益。
