Cancer detection rate of prebiopsy MRI with subsequent systematic and targeted biopsy are superior to non-targeting systematic biopsy without MRI in biopsy naïve patients: a retrospective cohort study

Prostate cancer (PCa) is the most common male malignancy and the second most common cause of male cancer-related death [1]. It is usually diagnosed based on systematic transrectal ultrasound (TRUS)-guided random biopsies of the prostate gland. However, a significant number of transrectal biopsies are negative for cancer, yielding inaccurate results [2‐4]. The cancer detection rate with a standard TRUS-guided prostate biopsy is only 20~ 40% [5]. Furthermore, TRUS-guided transrectal biopsies have a limited ability to sample the anterior prostate [6].

The ideal prostate biopsy goal would be to identify clinically significant PCa and minimize the detection of indolent disease. The growing availability of prostate multiparametric magnetic resonance imaging (mpMRI), novel functional imaging modalities, and increased standardization have created an opportunity for the detection, localization, and staging of PCa [7]. Screening patients using mpMRI may avoid the morbidity associated with a biopsy if no lesions are seen [8]. In addition, targeted biopsies should identify more clinically significant PCa than non-targeted TRUS-guided biopsies [9]. High PCa detection rates have been reported using magnetic resonance imaging (MRI)-targeted biopsies, both in patients with prior negative TRUS biopsies and in biopsy-naïve patients [9‐13]. However, it is not clear whether prebiopsy MRI with a subsequent targeted biopsy is superior to the traditional systematic non-targeted TRUS biopsy in biopsy-naïve patients [14‐16]. Therefore, this study examined whether prebiopsy mpMRI with the subsequent addition of targeted biopsies for suspicious lesions could improve the PCa detection rate in biopsy-naïve patients. These results were compared with those of a standard cohort of patients who underwent systematic non-targeted TRUS biopsies without MRI.

In this study, the detection rate of clinically significant PCa was significantly higher in the MRI cohort than the non-MRI cohort, whereas the clinically insignificant PCa detection rate was similar in both groups (Table 1). Performing MRI was an independent predictor of PCa detection (Table 3). These results suggest that prebiopsy MRI has the potential to improve biopsy outcomes in biopsy-naïve patients. We believe that a prebiopsy MRI has at least two advantages: patients can be selected more efficiently and a targeted biopsy of the index lesion can be added.

Traditionally, the decision regarding whether a prostate biopsy should be performed has been based mainly on the PSA, digital rectal examination findings, and age, which leads to inaccurate results. MRI can provide more precise information about the likelihood of the presence of PCa and prostate volume before biopsy. The prostate volume is negatively associated with the cancer detection rate, as shown in this (Table 3) and previous studies [20] and prostate volume was significantly smaller in the MRI cohort than the non-MRI cohort in this study (Table 1). These suggest that prostate biopsies might not be recommended in patients with a large prostate volume and/or normal MRI. Patient selection may partially explain the higher cancer detection rate in the MRI cohort. The combination of MRI findings and other biomarkers may better determine which patients should undergo prostate biopsy. Recently, we reported that the combination of PIRADS-2 score and PSA density was useful for decision-making before a prostate biopsy [21]. In that study, no patients with a PIRADS-2 score of ≤ 3 and PSA density of < 0.15 ng/mL/cm³ were diagnosed with clinically significant PCa. In the present study, 27 patients had a PIRADS-2 score of ≤ 3 and PSA density of < 0.15 ng/mL/cm³. Of these, no patients were diagnosed with clinically significant PCa (data not shown).

In this study, targeted biopsies based on MRI detected PCa more efficiently than systematic biopsies (Table 2). Furthermore, some patients in the MRI cohort were diagnosed with PCa or upgraded by the addition of targeted biopsies (Fig. 3). In particular, targeted biopsy was beneficial for the diagnosis of anterior cancer because most of the tumors detected by targeted biopsies, but missed by systematic biopsies, were located in the anterior zone. Collectively, the addition of a targeted biopsy might improve the biopsy outcome. However, the effectiveness of targeted biopsy in this study may have been underestimated because the urologist was not blinded to the suspicious lesions on MRI, so that these could be targeted in systematic biopsies. This suggestion is supported by the fact that the cancer-positive core rate in systematic biopsies in the MRI cohort was significantly higher than that in the non-MRI cohort (Table 2). However, performing the study with the urologists blinded to the MRI during the systematic biopsies would be preferred to evaluate the effectiveness of systematic and targeted biopsies accurately. The cancer detection rate of 55.3% in the MRI cohort in this study was somewhat lower than that reported in previous series that performed targeted biopsies with or without systematic biopsies (56–64%) [10, 13, 22, 23]. Furthermore, one third of the tumors were missed by targeted biopsies, but detected by systematic biopsies (Fig. 3). Three techniques have been reported for targeted biopsies: MRI-guided in-bore biopsy; MRI-TRUS fusion-guided biopsy; and cognitive registration. Although few direct comparisons have been performed, MRI-guided in-bore and MRI-TRUS fusion-guided biopsies likely yield a higher detection rate of clinically significant PCa compared with cognitive registration [13, 14, 22‐24]. The cognitive registration technique was used in this study, which may have caused the somewhat low PCa detection rate. However, MRI-guided in-bore and MRI-TRUS fusion-guided biopsies have limited availability and are complex and costly to introduce and/or perform; therefore, they are not feasible for routine use so far. These suggest that targeted biopsies using cognitive registration are more practical and may improve the performance of prostate biopsy.

This study found a correlation between the PIRADS-2 score and PCa detection rate, especially in clinically significant PCa. ROC analysis revealed an AUC of 0.801 which was superior to that of 0.738 in the three-point scale; the sensitivity, specificity, PPV, and NPV were 0.87, 0.75, 0.81, and 0.82, respectively. A meta-analysis assessing the performance of mpMRI for detecting PCa found a specificity of 0.88 (95% CI, 0.82–0.92), sensitivity of 0.74 (95% CI, 0.66–0.81), and NPV of 0.64–0.94, which is consistent with our results. Recently, PIRADS was annotated, revised, and published as a second version, PIRADS-2, to define standards of high-quality clinical service for mpMRI, including image creation and reporting [17, 18]. Kuru et al. performed ROC analysis for PCa detection using the PIRADS score and found an excellent specificity of 0.90–0.98 [25]. Grey et al. [26] reported that a ROC analysis of clinically significant PCa yielded an AUC of 0.88–0.89 and sensitivity, specificity, PPV, and NPV of 0.95–0.97, 0.6, 0.58–0.61, and 0.97–0.98, respectively. The difference in the ability to predict biopsy outcome among studies may be due to differences in the MRI machines and protocols used, as well as to variations in the PIRADS scoring, biopsy protocols, and patient characteristics.

Two recent randomized controlled trials (RCTs) compared the outcomes between prebiopsy MRI with the addition of a targeted biopsy and a conventional TRUS-guided random biopsy [15, 16]. One used a MRI-TRUS fusion targeted biopsy, while the other used the cognitive registration technique, similar to our study. Both found that the PCa detection rate was similar between the MRI and control groups, which is not consistent with our results. Selection bias in our retrospective study could have been the major reason for the difference versus the two RCTs, and patient selection after MRI may have been one of the major causes of the higher cancer detection rate in our MRI cohort. There were also differences in the study protocols, sample sizes, and biopsy protocols among studies. Furthermore, the PCa detection rate of 54–57% for the conventional TRUS-guided biopsy of the two RCTs was higher than that in our study, and another reported study [5].

This study had some limitations. First, the analysis was retrospective and patient selection bias may have been present, as described above. However, it appears based on the statistics that the two cohorts were very similar other than prostate size. Second, PIRADS scoring was not performed at the same time as the biopsy. Third, some patients underwent 1.5-Tesla MRI, whereas others underwent 3-Tesla MRI, reflecting technological changes. However, the PCa detection rate did not differ significantly between 1.5-Tesla and 3-Tesla MRI (data not shown). Fourth, there is no widely accepted definition of clinically significant PCa. Fifth, the promise of targeted biopsies is to reduce the number of total biopsies. However, our study suggests that systematic biopsies should not be omitted when cognitive fusion transperineal biopsies are performed because one third of the tumors were missed by targeted biopsies but detected by systematic biopsies (Fig. 3). Sixth, several radiologists were involved in reporting the MRIs and several urologists performed biopsies. However, this could be deemed an advantage, since clinical effectiveness of prebiopsy mpMRI was demonstrated despite this heterogeneity. Finally, only one radiologist was involved in the MRI review of the PIRADS-2 scoring in each institution; therefore, inter-observer reliability could not be assessed.

Prebiopsy mpMRI with subsequent systematic plus targeted biopsies could yield a more clinically significant PCa detection rate than a non-targeted TRUS-guided biopsy in biopsy-naïve patients. PIRADS-2 scoring is useful for predicting biopsy outcome. However, large prospective studies are needed to confirm our results.