An Magnetic Resonance Imaging–directed Targeted-plus-perilesional Biopsy Approach for Prostate Cancer Diagnosis: “Less Is More”

Take Home Message This retrospective study found arguments to limit systematic biopsies only to ipsilateral prostate lobe sectors directly adjacent to the magnetic resonance imaging (MRI)-positive lesion. An MRI-directed targeted-plus-perilesional biopsy approach detects clinically significant prostate cancer cases equivalent to the recommended practice of an MRI-directed targeted-plus-systematic biopsy approach, while reducing overdiagnosis rates and the number of biopsy cores.

CI 6.6-21.7%). If a targeted-plus-perilesional biopsy approach would have been performed, the number of biopsy cores per patient would have been reduced significantly (a mean difference of 5.2; 95% CI 4.9-5.6, p < 0.001). Conclusions: An MRI-directed targeted-plus-perilesional biopsy approach detected almost all csPCa cases, while limiting overdiagnosis and reducing the number of biopsy cores. Prospective clinical trials are needed to substantiate the withholding of nonperilesional SBx in men with unilateral lesion(s) on MRI. Patient summary: Limiting systematic biopsies to the proximity of the suspicious area on magnetic resonance imaging helps detect an equivalent number of aggressive cancers and fewer indolent cancers. These findings may help patients and physicians choose the best biopsy approach.

Introduction
Magnetic resonance imaging (MRI) of the prostate is increasingly being used in men suspected of having prostate cancer (PCa) to target prostate biopsies, either by cognitive or software-assisted fusion approaches [1,2]. Even though recent studies have shown that the use of targeted biopsy (TBx) increases the detection of clinically significant PCa (csPCa), still a substantial proportion of csPCa is missed by TBx alone [3,4]. It remains unknown whether csPCa is missed due to the limited sensitivity of MRI, the suboptimal image fusion, the biopsy technique and strategy, or a combination of these. Therefore, complementary systematic biopsy (SBx) is still recommended by the European Association of Urology (EAU) PCa guidelines to maximize the detection of csPCa [1].
Complementary SBx increases the detection of csPCa; however, it also increases the detection of clinically insignificant PCa (ciPCa) [4][5][6]. Given its very low metastatic potential and consequently very low risk of cancerspecific mortality, detection of ciPCa is generally considered overdiagnosis, and is associated with overtreatment and a high patient burden [7]. Furthermore, the increased number of biopsy cores, when performing both TBx and SBx, is associated with a longer procedure time, more pronounced patient discomfort, increased workload for pathologists, and increased costs [8][9][10]. In view of the abovementioned disadvantages, the traditional whole-gland SBx template is currently being disputed [6,11,12].
Ideally, prostate biopsy strategies maximize the detection of csPCa, while minimizing the detection of ciPCa. Considering that most men benefit diagnostically from increased sampling of index lesions, limiting SBx to the region around the index lesion could potentially effectuate this [13,14]. This study aims to retrospectively evaluate the diagnostic performance of TBx + perilesional SBx, compared with TBx + standard SBx, in men with MRI-positive unilateral lesion(s) suggestive of being csPCa.

MRI/ultrasound-fusion biopsy and histopathology
Transperineal prostate biopsy procedures were performed using a freehand cognitive fusion approach based on the Ginsburg protocol [17]. mapped the lesion of interest in the Ginsburg protocol. In addition to the reference of TBx + standard SBx, a hypothetical biopsy template was defined: TBx + perilesional SBx ( Fig. 1 and Supplementary Fig. 1). Based on pre-existing literature, perilesional SBx cores were defined as all nontargeted biopsy cores taken from the ipsilateral prostate lobe sectors directly adjacent to the MRI-positive index lesion [13].
Core-needle biopsy specimens were potted and reported separately according to the sector these originated from. All specimens were prospectively analyzed by experienced uropathologists and reported according to the International Society of Urological Pathology (ISUP) consensus for grading of PCa [18]. Clinically significant PCa was defined as Gleason score 3 + 4, equivalent to ISUP grade group (GG) 2. Additionally, a second stricter definition of csPCa was used, equivalent to ISUP GG 3.

Statistical analysis
Categorical variables were reported as frequency distributions and percentages, and continuous variables were expressed as medians with interquartile ranges (IQRs).

Results
A total of 235 men were retrospectively included for analysis. Patient characteristics are presented in Table 1.  (Table 3). In the case of discrepancy in the detection of csPCa between TBx + standard SBx and TBx + perilesional SBx approaches, the highest ISUP GG was found in the contralateral sectors directly adjacent to the MRI-positive lesion (ie, if the MRI-positive lesion is located in the right posterior sector, the highest ISUP GG is found in the contralateral adjacent posterior sector; Supplementary Table 1). Moreover, all the detected nonperilesional csPCa cases involved ISUP GG 2 cancers. Consequently, if a stricter definition of csPCa was used, equivalent to ISUP GG 3, a TBx + perilesional SBx approach would have correctly detected 47/47 (100.0%; 95% CI 92.5-100.0%) csPCa cases (Supplementary Table 2).
Compared with TBx, standard and perilesional SBx minimally increased the csPCa detection rate, as six of 95 (6.3%) csPCa diagnoses were found systematically (Supplementary Table 3). Particularly in men with PI-RADS 4 and 5 lesions, standard and perilesional SBx were of limited or no added value. In contrast, standard and perilesional SBx were of greater added value in men with PI-RADS 3 lesions (Supplementary Table 4).    Using a TBx + perilesional SBx approach, detection of ciPCa would have been reduced substantially. Limiting SBx to ipsilateral prostate lobe sectors directly adjacent to the MRI-positive lesion would have reduced the systematically found ciPCa from 14 to three (difference 12.8%; 95% CI 6.6-21.7%) of 86 men. Nonperilesional SBx detected small ciPCa lesions in these 11 cases, as in most cases only one or two biopsy cores showed positive findings.

Number of biopsy cores
When performing only perilesional SBx in addition to TBx, the total number of biopsy cores per patient decreased by 38.5% (IQR 30.0-50.0%); a median number of 7 (IQR 6-9) biopsy cores were taken using the TBx + perilesional SBx approach. Compared with the TBx + standard SBx approach, this resulted in a significant reduction of 5.2 (p < 0.001; 95% CI 4.9-5.6) biopsy cores per patient on average.

Discussion
Our findings suggest that a TBx + perilesional SBx approach would be a good alternative that reduces the detection of ciPCa and the total number of biopsy cores, while preserving the detection of csPCa. Moreover, based on the data presented, a further reduction in the number of biopsies could be reached in men with PI-RADS 4 and 5 lesions.
To maximize the detection of csPCa, standard SBx as a supplement to TBx is currently still recommended by EAU guidelines on PCa diagnosis [1,11,12,19]. The addition of SBx, however, not only improves the detection of csPCa, but also coincides with an increased detection of ciPCa [20,21]. Possible explanations for missing csPCa on TBx are related to imprecise lesion registration (underestimation of tumor volume) and targeting errors due to cognitive fusion inaccuracies [19,21,22]. Considering that most of the csPCa missed by TBx are found in sextants adjacent to MRIpositive lesions, limiting SBx to the vicinity of MRI-positive lesions has been suggested as an alternative to standard ''whole-prostate'' SBx.
Given the lack of consensus on how to implement a TBx + perilesional SBx approach, different interpretations exist in the current literature, differing in both the suggested numbers and the locations of perilesional SBx [13]. Owing to the high level of skill and expertise required for previously proposed templates, applicability has been limited so far. This study's proposed TBx + perilesional SBx approach used equivalent numbers of biopsy cores to those reported in previously published studies [6,12,23], but differed in the proposed locations. Limiting SBx to the ipsilateral prostate lobe sectors directly adjacent to the MRIpositive lesion, which requires less skill and expertise, will improve the applicability and usefulness compared with previously proposed approaches.
The csPCa detection rate of TBx within this cohort is notable. By performing only TBx and thus omitting SBx altogether, 89/95 (93.7%; 95% CI 86.8-97.6%) cases of csPCa would have been classified correctly. This is substantially higher than that reported in the existing literature, where detection rates of csPCa for TBx range between 67% and 89% [13]. The relatively high number of TBx (five biopsy cores; IQR four to six) taken per patient within this cohort,   making it comparable with an extended TBx or focal saturation approach, could be the explanation [11]. Increasing the number of TBx events also overcomes the problem of imprecise lesion registration and targeting errors, rendering complementary perilesional SBx redundant. However, good image quality is of paramount importance. Poor image quality greatly complicates the registration and targeting of lesions. Ideally, before implementing a TBx (+perilesional SBx) approach, image quality should be ensured by using a scoring system (Prostate Imaging Quality [PI-QUAL] score or in-house scoring system) [24]. This study did not use a scoring system but rather a binomial determination of the image quality: insufficient (no PI-RADS score) or sufficient (a PI-RADS score).
Preventing overdiagnosis of ciPCa also prevents overtreatment and the associated impaired quality of life due to urinary, sexual, or gastrointestinal side effects. Limiting the number of SBx events substantially reduces overdiagnosis rates; 12.8% of men in our present study would have avoided a diagnosis of ciPCa. Additionally, reducing the number of biopsy cores would also reduce workload for pathologists, shorten procedure time, and minimize patient discomfort [9,10]. Considering the fact that perilesional SBx had limited or no added value in PI-RADS 4 and 5 lesions within this cohort, it could be argued that the number of biopsies can be reduced further in these subgroups. By performing only perilesional SBx in men with PI-RADS 3 lesions and limiting prostate biopsies in men with PI-RADS 4 and 5 lesions to TBx only, the reduction in the total number of biopsies per patient could be increased further to 61.5% (IQR 50.0-70.0%). However, because a hypothetical biopsy template was used in this study, we were unable to provide the necessary data to substantiate these arguments.
Our study is not devoid of limitations. To start, since we used a prospective database in which transperineal prostate biopsies were reviewed, definitive histopathological verification from prostatectomy specimens was lacking. Although the Ginsburg protocol reliably detects 97% of csPCa later found at prostatectomy [25], the lack of definitive histopathology limits the findings of this study. Owing to daily practice, MRI scans were performed on two different scanners and interpreted by different radiologists, which may have influenced the findings. The accuracy of the MRI reading may vary with changes in technical equipment and by the interpretation of different observers [26]. All prostate MRI scans were read by seven experienced radiologists, although the experience in reading prostate MRI varied among the radiologists. All seven radiologists have read sufficient MRI scans to be experienced enough to be an independent reader, and six out of seven have read at least 300 MRI scans [27]. Interpretation by experienced but nonexpert readers may have influenced the findings. This cohort used biparametric MRI acquisition rather than multiparametric MRI acquisition. The lack of dynamic contrast enhanced (DCE) sequences may have affected MRI interpretation; in the presence of a DCE sequence, a PI-RADS 3 lesion might have been interpreted as a PI-RADS 4 lesion. In addition, the proposed TBx + perilesional SBx approach relies on the presence of an MRI-positive lesion, which limits its applicability; men with negative MRI, but still a high clinical suspicion, need different biopsy approaches. Another limitation that merits discussion is the therapeutic impact of such a TBx + perilesional SBx approach. It is unknown whether it is better to reduce the number of biopsy cores or improve prostate mapping, to perform a more confident and safer treatment. Finally, given the retrospective nature of this study, results need to be interpreted with caution. Future prospective cohort studies are therefore needed to further investigate and optimize this promising diagnostic biopsy approach.

Conclusions
Limiting systematic biopsies only to ipsilateral prostate lobe sectors directly adjacent to the MRI-positive lesion, thereby avoiding contralateral systematic biopsies altogether, has shown to be a promising diagnostic biopsy approach in biopsy-naïve men suspected of having csPCa. Increased sampling of the index lesion correctly detects almost all csPCa cases, while limiting the overdiagnosis of ciPCa. Future studies, using prostatectomy specimens as histopathological verification, focusing on its applicability and therapeutic impact are needed to facilitate the adoption of this proposed biopsy approach. Other: None.
Financial disclosures: Marinus J. Hagens certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.
Funding/Support and role of the sponsor: None.