基于机器学习的MR radiomics分析可以帮助改善PI-RADS v2在临床相关前列腺癌中的诊断性能(英文)

2017-04-03 09:27:04 admin 19

Machine learning-based analysis of MR radiomics can help to improve the diagnosticperformance of PI-RADS v2 in clinically relevant prostate cancer.

Eur Radiol. 2017 Apr 3

Wang J1Wu CJ2Bao ML3Zhang J2Wang XN2Zhang YD4.

Author information

    1.Center for Medical Device Evaluation, CFDA, Beijing, China, 100044.

    2.Department of Radiology, the First Affiliated Hospital with Nanjing Medical University, 300, Guangzhou Road, Nanjing, Jiangsu Province, China, 210009.

    3.Department of Pathology, the First Affiliated Hospital with Nanjing Medical University, Nanjing, China, 210009.

    4.Department of Radiology, the First Affiliated Hospital with Nanjing Medical University, 300, Guangzhou Road, Nanjing, Jiangsu Province, China, 210009. 

Abstract

OBJECTIVE


To investigate whether machine learning-based analysis of MR radiomics can help improve the performance PI-RADS v2 in clinically relevant prostate cancer (PCa).

METHODS


This IRB-approved study included 54 patients with PCa undergoing multi-parametric (mp) MRI before prostatectomy. Imaging analysis was performed on 54 tumours, 47 normal peripheral (PZ) and 48 normal transitional (TZ) zone based on histological-radiological correlation. Mp-MRI was scored via PI-RADS, and quantified by measuring radiomic features. Predictive model was developed using a novel support vector machine trained with: (i) radiomics, (ii) PI-RADS scores, (iii) radiomics and PI-RADS scores. Paired comparison was made via ROC analysis.

RESULTS


For PCa versus normal TZ, the model trained with radiomics had a significantly higher area under the ROC curve (Az) (0.955 [95% CI 0.923-0.976]) than PI-RADS (Az: 0.878 [0.834-0.914], p < 0.001). The Az between them was insignificant for PCa versus PZ (0.972 [0.945-0.988] vs. 0.940 [0.905-0.965], p = 0.097). When radiomics was added, performance of PI-RADS was significantly improved for PCa versus PZ (Az: 0.983 [0.960-0.995]) and PCa versus TZ (Az: 0.968 [0.940-0.985]).

CONCLUSION


Machine learning analysis of MR radiomics can help improve the performance of PI-RADS in clinically relevant PCa.

KEY POINTS


Machine-based analysis of MR radiomics outperformed in TZ cancer against PI-RADS. • Adding MR radiomics significantly improved the performance of PI-RADS. • DKI-derived Dapp and Kapp were two strong markers for the diagnosis of PCa.

KEYWORDS


Machine learning; Multi-parametric MRI; Prostate Imaging Reporting and Data System v2; Prostate cancer; Support vector machine


慧影医疗科技(北京)有限公司

地点:北京市海淀区中关村东升科技园B2-C103

电话:400-890-9020

邮箱:radcloud@huiyihuiying.com

关闭
图片
图片
  • 人工智能诊断云平台