Our clinical practice experiences showed a 44.70% csPCa DR among 76 biopsy-naive men with PI-RADS v2.1 scores of 4 and 5 in prebiopsy mp-MRI who underwent systematic biopsy followed by cog-MRGB. There were significant statistical differences in PSA level, prostate volume, lesion location and diameter, and PI-RADS score between the patients according to the csPCa outcome, and a higher PI-RADS category led to a higher csPCa DR. In the same way, the csPCa was more observed among PI-RADS 5 category, while the PI-RADS score of 4 contained more benign prostate histopathology.
PI-RADS v2.1 was developed to modify PI-RADS v2 and improve its limitations and inconsistencies, aiming to enhance PI-RADS diagnostic performance. [6, 7] A recent systematic review and meta-analysis of 17 studies demonstrated 59% and 85% pooled csPCa DR for PI-RADS scores of 4 and 5, respectively. The Higher PI-RADS v2.1 category was also significantly associated with a higher cancer DR. Moreover, subgroup analysis did not reveal differences in the cancer DR stratified by the PZ and TZ. [8] In the current study, we reported a csPCa DR of 17.10% for PI-RADS 4 and 68.30% for PI-RADS 5, which is lower than the meta-analysis results, notably for PI-RADS 4. These apparent differences could be attributed to studies from developed countries that most utilized 3T mp-MRI and advanced biopsy techniques. Similarly, our study showed a higher DR in the higher PI-RADS category. In contrast, the csPCa DR is different between the PZ and TZ. (54.90%compared to 34.30%)
Compared to our results, the studies evaluating the PI-RADS v2.1 diagnostic performance by utilizing 1.5T mp-MRI demonstrated an overall csPCa DR of 40.66% and 34.20% in software-assisted MRI-TRUS fusion biopsy (fus-MRGB) and fus-MRGB or cog-MRGB, respectively. [10, 14] Although both 1.5T and 3T MRI can be employed appropriately to provide reliable and adequate diagnostic examinations, 3T MRI can improve imaging resolution by increasing the signal-to-noise ratio. [5] However, a recent meta-analysis emphasized that 1.5T and 3T MRI do not differ statistically in diagnostic performance. [15]
Among previous publications applying PI-RADS v2.1, Chen et al. [16] reported 62.64% and 48.85% overall PCa and csPCa DRs among 174 Chinese men undergoing cog-MRGB according to the 3T MRI target lesions, respectively. The reported DRs were slightly higher than our study, albeit Chen et al. evaluated all PI-RADS categories. Furthermore, a Korean study including 233 patients who received cog-MRGB due to a target lesion located at the PZ in a 3T MRI showed a 32.6% DR for csPCa in the prostate PZ among all PI-RADS categories. [17] In contrast, we reported a 54.90% csPCa DR for the PZ among patients with PI-RADS scores of 4 and 5.
Considering the effect of mp-MRI and magnetic resonance targeted biopsy on increasing the DR of csPCa, cog-MRGB could be a lower-cost and practical option to improve biopsy accuracy in clinical routine, especially in countries without available expensive fusion equipment. [18, 19] Comparing the cog-MRGB, fus-MRGB, and in-bore MRI targeted biopsy (inbore-MRGB) performance for overall PCa detection revealed only a significant difference between cog-MRGB and inbore-MRGB, with no significant superiority for any specific technique in csPCa diagnosis. [20] Later investigation demonstrated that these techniques did not significantly differ in detecting PCa and csPCa. [19] These findings are supported by the meta-analysis of literature comparing cog-MRGB with fus-MRGB. [21] Moreover, Kaufmann et al. [22] reported a significantly higher overall PCa DR in robot-assisted MRI/TRUS fusion and inbore-MRGB compared to cog-MRGB among men with prior negative biopsy without any significant differences in csPCa detection. However, the potential limitations of the cog-MRGB method are undeniable, such as a requirement of implementing TRUS imaging in parallel with mp-MRI findings, the effects of target lesion diameter and anatomical location on biopsy accuracy, and the absence of tracking and recording coordination between the target and biopsy during the biopsy procedure. [18, 23, 24]
In reviewing the literature conducted among the Middle East population, we found a few previous Turkish studies [11–13] evaluating PI-RADS v2.1 diagnostic performance. (Supplementary Table S3) Koparal et al. [11] reported a 24.60% DR of csPCa in the fus-MRGB and systematic biopsy among 1143 biopsy naïve men or men with previous negative biopsies who had PI-RADS v2 and v2.1 scores ≥ 2. Conversely, our study illustrated a higher csPCa DR with a smaller sample size of PI-RADS 4 and 5 categories and the cog-MRGB (24.60% vs. 44.70%). Other study results showed a 63% csPCa DR among 246 patients (a 66% csPCa DR among 200 biopsy-naïve patients) with PI-RADS scores of 4 and 5 in the 3T system according to version 1, 2, and 2.1 criteria, who received in-bore transrectal biopsy. [12] Moreover, their results showed a significantly higher DR for PI-RADS 5 than 4, i.e., 49.7% for PI-RADS 4 and 75% for PI-RADS 5. Furthermore, another study with a 154-participant sample size demonstrated 80% and 82% csPCa DR among patients with PI-RADS v2.1 scores of 4 and 5, utilizing a 1.5T scanner and 12 quadrant systematic biopsies. [13] However, 19 patients with lesions on the TZ underwent cog-MRGB, which showed 11 csPCa. In contrast to these studies, the csPCa DR was lower in the current study, markedly in PI-RADS score of 4. (17.10% compared to 49.70% and 80%)
Our results showed only 34.30% and 17.10% overall PCa and csPCa DRs among PI-RADS 4 categories, respectively. Compared to the studies mentioned before, that was notably lower than expected. Hence, we tried to discuss various probable causes. In our study, PI-RADS 4 patients were more likely to have smaller lesions, mainly located at the TZ of the prostate. (Supplementary Table S2) Despite advancing the prostate mp-MRI evaluation and targeted biopsy techniques, PCa diagnosis remains challenging in the TZ due to the overlapping of benign prostatic hyperplasia features with PCa. [25–27] Therefore, major revisions were implemented in PI-RADS v2.1 to focus on these limitations. [6] Nevertheless, PI-RADS v2.1 did not demonstrate any statistical superiority in sensitivity and specificity for detecting PCa or csPCa in the TZ [27], and there was a high false positive rate for diagnosis csPCa in TZ. [28] Furthermore, the cog-MRGB is still challenging in smaller target lesions at the prostate base or apex, particularly in those located anteriorly. [18, 23, 24] Therefore, the cog-MRGB was recommended to be highly advantageous for sampling diffuse abnormalities or large lesions located at the prostate PZ. [18] Hence, lesion size might contribute to csPCa DR, especially in the PI-RADS score of 4, and the suggested threshold of lesion size was 10 millimeters. [26, 29, 30] In this regard, in our study, the calculated means of lesion diameter significantly differed between PI-RADS 4 and 5, albeit the median of calculated means was 13 millimeters (IQR 11.50–14) among men with PI-RADS 4. Moreover, our patients received a median of 2 additional cognitive cores (IQR 1–3) during the biopsy. According to previous investigations, the extra biopsy cores obtained from MRI target lesions might benefit patients by improving PCa detection. [31–33] In this context, inadequate sampling cognitive cores might diminish our DRs, especially in PI-RADS categories of 4. In light of these probable causes, re-evaluation of pathology or biopsy technique, a repeat of the mp-MRI or biopsy, and active surveillance should be considered for managing these cases based on PCa clinical suspicion. [34]
The noticeable strength of this study was representing the initial experiences of PI-RADS v2.1 implementation among a homogenous biopsy-naïve population from a developing country in the Middle East. Moreover, the same single radiologist, urologist, and pathologist performed all procedures that caused low heterogeneous results.
The current study was also exposed to several limitations. First, the study included a small sample size. However, PCa shows a growing pattern among North African and the Middle Eastern population, and also the Iranian male, the PCa burden is still low, in contrast to Western countries. [3, 4] Thus, a limited sample size was expected. Second, considering the study’s small sample size, we tried to detect the association instead of developing a prediction model for csPCa diagnosis. Third, standard systematic plus cog-MRGB was considered the reference test, and mp-MRI findings did not compare with prostatectomy specimens due to the study’s retrospective design and lack of final pathology results.