Artificial neural network analysis on computerized transrectal ultrasound in early detection of prostate cancer / 中华泌尿外科杂志

ABSTRACT

Objective To investigate the application of artificial neural network analysis on computerized transrectal ultrasound (ANNAcTRUS) in early detection of prostate cancer.Methods Sixty men with or without prior biopsies, either due to elevated PSA or abnormal digital rectal findings, were included in this study from January 2014 to July 2015.Patient's mean age was (65.6 ± 8.9) years (51-89 years).Their PSA level was (9.8 ± 4.9)μg/L.The patients received the ANNAcTRUS targeted 6-core biopsy.Each patient received six targeted biopsies of suspicious regions, which was identified by ANNAcTRUS online system.Histopathologic examination was further carried out to confirm the result of the targeted biopsies.Results According to the results of ANNAcTRUS,52 of 60 patients received biopsy in ANNAcTRUS group.ANNAcTRUS targeted 6-core biopsy was able to detect prostate cancer in 24 (46.2％) of 52 patients.The distribution of Gleason Score was as follows : 3 + 3 (n =9), 3 + 4 (n =8), 4 + 3 (n =4), 4 +4 (n =2) and 5 +4 (n =1).For patients without prior negative biopsy,ANNAcTRUS targeted 6-core biopsy was able to detect prostate cancer in 17 (51.5％) of 33 patients.Conclusions ANNAcTRUS targeted 6-core biopsy illustrates a higher detection rate of prostate cancer.Furthermore, ANNAcTRUS targeted 6-core biopsy tends to detect low-grade prostate cancer.