The accuracy of mpMRI combined with clinical scales in predicting invasion of capsule and seminal vesicle in prostate cancer
10.3760/cma.j.cn112330-20210602-00309
- VernacularTitle:mpMRI联合临床量表预测前列腺癌包膜外和精囊侵犯的准确性
- Author:
Tianyu XIONG
1
;
Xiaoqi FAN
;
Xiaobo YE
;
Yun CUI
;
Mingshuai WANG
;
Min LI
;
Tao JIANG
;
Yinong NIU
Author Information
1. 首都医科大学附属北京朝阳医院泌尿外科，北京 100020
- Keywords:
Prostatic neoplasms;
Prostate cancer;
Magnetic resonance imaging;
Clinicopathological features;
Prediction
- From:
Chinese Journal of Urology
2022;43(2):122-127
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To explore the accuracy of mpMRI combined with Partin table, MSKCC nomogram and CAPRA score in predicting extracapsular extension and seminal vesicle invasion of prostate cancer.Methods:From January 2016 to June 2021, a total of 178 patients who underwent laparoscopic radical prostatectomy were selected. The average age of patients was (68.3±3.5) years, the average preoperative PSA level was (24.5±7.1)ng/ml, and the average percentage of positive cores in biopsy was 44.3%. The clinical T 1c stage was determined in 67 cases (37.6%), T 2a in 69 cases (38.8%) and T 2b-2c in 42 cases(23.6%). Biopsy Gleason score of 3+ 3=6 was found in 45 cases(25.3%), 3+ 4=7 in 41 cases(23.0%), 4+ 3=7 in 26 cases(14.6%), 8 with different combinations in 36 cases(20.2%), and 9 or 10 in 30 cases(16.9%). According to preoperative PSA level, biopsy Gleason score, clinical stage, age, total biopsy cores and positive cores, the posibility of extracapsular extension and seminal vesicle invasion were predicted using 2012-version Partin table and MSKCC nomogram. CAPRA score of each patient was calculated. The prediction schemes were built as follows: ①mpMRI alone, ②mpMRI combined with Partin scale, ③mpMRI combined with MSKCC nomogram, ④mpMRI combined with CAPRA score. The results of each prediction scheme were compared with postoperative pathological reports. Logistic regression analysis was used to evaluate the relationship between predictive results and postoperative pathological outcomes. The receiver operating characteristic curve of each prediction scheme was drawn. The area under curve was used to compare the predictive accuracy of each combination scheme for the pathological results of prostate cancer. The decision analysis curve of each prediction scheme was drawn. The clinical benefits of each scheme were analyzed by comparing the net return under different risk thresholds. Results:mpMRI predicted extracapsular extension in 21 cases(11.8%) and seminal vesicle invasion in 16 cases(9.0%). The postoperative pathological results reported extracapsular extension in 27 cases(15.2%) and seminal vesicle invasion in 39 cases(21.9%). Logistic regression analysis showed that mpMRI and clinical scales were predictors related to the pathological results of prostate cancer( P<0.05). The receiver operating characteristic curve of each scheme showed that the area under curve for predicting extracapsular extension by using mpMRI, mpMRI combined with Partin table, mpMRI combined with MSKCC nomogram and mpMRI combined with CAPRA score were 0.599, 0.652, 0.763 and 0.780, respectively, and the area under curve for predicting seminal vesicle invasion were 0.607, 0.817, 0.826 and 0.820, respectively. Compared with simple application of mpMRI, except that the scheme of mpMRI combined with Partin table had no obvious advantage in predicting extracapsular extension( P=0.117), any other combined scheme had higher prediction accuracy( P<0.01). mpMRI combined with MSKCC nomogram or CAPRA score was better than mpMRI combined with Partin table in predicting extracapsular invasion ( P<0.01). There was no significant difference in predicting seminal vesicle invasion among these three combination schemes ( P>0.05). The net income of the combined prediction scheme was higher than that of using mpMRI alone under any risk threshold. The scheme of using mpMRI combined with MSKCC nomogram had the highest net income. Conclusions:mpMRI combined with clinical scales has good accuracy in predicting pathological characteristics of prostate cancer in Chinese population. Compared with other schemes in this study, the combination scheme of mpMRI combined with MSKCC nomogram has the highest prediction accuracy.
