Department of Radiology, Ghent University Hospital, Corneel Heymanslaan 10, 9000, Ghent, Belgium.
Department of Urology, Ghent University Hospital, Corneel Heymanslaan 10, 9000, Ghent, Belgium.
Eur J Radiol. 2019 May;114:92-98. doi: 10.1016/j.ejrad.2019.02.031. Epub 2019 Feb 25.
To investigate which clinical and radiological characteristics can predict clinically significant prostate cancer (csPCa) in PI-RADS 3 lesions. To investigate which clinical and radiological characteristics influence the clinician to biopsy a PI-RADS 3 lesion.
mpMRI PI-RADS 3 lesions scored by 1 out of 3 highly specialized radiologists in a single high-volume center during the period March 2015 to August 2017 were investigated. This score was based on T2 weighted and diffusion weighted imaging (DWI) sequences. Clinical characteristics of all patients with PI-RADS 3 lesions were collected from medical records. Radiological characteristics were collected from radiology reports. Some radiological characteristics such as apparent diffusion coefficient (ADC) in a region of interest at the tumor site and ADC at a site contralateral to the tumor site were calculated on DWI sequences. Cox regression analysis was performed to identify which characteristics could predict csPCa in PI-RADS 3 lesions and which characteristics could influence the behavior of a clinician whether or not to biopsy a PI-RADS 3 lesion.
csPCa could be detected in 31 out of 131 patients with PI-RADS 3 lesions (22.9%). A lower median prostate volume (p = 0.015) and a lower ratio of ADC of the tumor on ADC of the contralateral prostate (ADCT/ADCCLP) (p < 0.001) significantly predisposed for csPCa in multivariate logistic regression. For peripheral zone lesions, a diagnostic model with biopsy of only those PI-RADS 3 lesions with a prostate volume <44 cc and a ratio of ADCT/ADCCLP < 70% showed a sensitivity for detection of csPCa of 59% with a specificity of 88%. (area under the curve 0.780) A suspicious rectal examination (p = 0.011) and the mentioning of prostatitis on the MRI report (p = 0.020) influenced clinicians to biopsy a PI-RADS 3 lesion positively and negatively respectively. For transition zone lesions, previous negative biopsies (p = 0.044) predisposed for csPCa.
Prostate volume and the ratio of ADC tumor on ADC of the contralateral prostate have the potential to predict csPCa in PI-RADS 3 lesions with a sensitivity of 59% and specificity of 88%. A suspicious rectal examination and the mentioning of prostatitis on the MRI report influenced the decision of clinicians to biopsy a PI-RADS 3 lesion.
探讨 PI-RADS 3 病变中哪些临床和影像学特征可预测临床显著前列腺癌(csPCa)。探讨哪些临床和影像学特征影响临床医生对 PI-RADS 3 病变进行活检。
研究了 2015 年 3 月至 2017 年 8 月期间,由 3 名高度专业化的放射科医生在单一高容量中心对 PI-RADS 3 病变进行评分的 MRI 前列腺影像报告和数据系统(PI-RADS)评分 3 分的 mpMRI 病变。该评分基于 T2 加权和弥散加权成像(DWI)序列。从病历中收集所有 PI-RADS 3 病变患者的临床特征。从放射学报告中收集影像学特征。在肿瘤部位的感兴趣区域和肿瘤对侧部位的表观弥散系数(ADC)等一些影像学特征,在 DWI 序列上进行计算。采用 COX 回归分析,确定哪些特征可预测 PI-RADS 3 病变中的 csPCa,以及哪些特征可影响临床医生对 PI-RADS 3 病变进行活检的决策。
在 131 例 PI-RADS 3 病变患者中,31 例(22.9%)可检测到 csPCa。多变量逻辑回归显示,前列腺体积中位数较低(p=0.015)和肿瘤 ADC 与对侧前列腺 ADC 比值(ADCT/ADCCLP)较低(p<0.001)显著增加了 csPCa 的风险。对于外周带病变,仅对前列腺体积<44 cc 且 ADCT/ADCCLP 比值<70%的 PI-RADS 3 病变进行活检的诊断模型,检测 csPCa 的敏感性为 59%,特异性为 88%(曲线下面积 0.780)。直肠指检可疑(p=0.011)和 MRI 报告中提到前列腺炎(p=0.020)分别对临床医生进行 PI-RADS 3 病变活检产生积极和消极影响。对于移行带病变,既往阴性活检(p=0.044)增加了 csPCa 的风险。
前列腺体积和肿瘤 ADC 与对侧前列腺 ADC 比值有潜力预测 PI-RADS 3 病变中的 csPCa,其敏感性为 59%,特异性为 88%。直肠指检可疑和 MRI 报告中提到前列腺炎影响临床医生对 PI-RADS 3 病变进行活检的决策。