Differential detection rates of clinically significant prostate cancer in transition zone versus peripheral zone lesions: Implications for transperineal magnetic resonance imaging-targeted biopsy strategy.

1
Introduction
Prostate cancer remains one of the most commonly diagnosed malignancies in men worldwide, with tissue biopsy continuing to serve as the definitive diagnostic method. Improvements in biopsy techniques have focused on enhancing the detection of clinically significant prostate cancer (csPCa) while reducing the overdiagnosis of indolent tumors and their associated complications. Magnetic resonance imaging (MRI) has significantly transformed prostate cancer diagnostics by allowing precise identification and targeted sampling of suspicious lesions that may be overlooked by conventional systematic biopsy methods.1,2
A landmark study by Siddiqui et al found that MRI/ultrasound fusion-guided biopsy identified 30% more high-risk cancers and 17% fewer low-risk cancers compared with standard systematic biopsy in a cohort of over 1,000 men.3 These results played a major role in the broader use of MRI-targeted biopsy techniques, which follow the guidelines of the Prostate Imaging Reporting and Data System (PI-RADS). The most recent version, PI-RADS v2.1, released in 2019, offers a standardized approach to MRI interpretation in prostate cancer evaluation.4 Emerging data support the use of both targeted and systematic biopsy in combination, as relying solely on targeted biopsy may fail to identify some csPCa cases.5,6 The prostate's anatomical structure significantly influences cancer detection, with approximately 70% of cancers originating in the peripheral zone (PZ) and about 20% in the transition zone (TZ).7 These zones differ in tumor characteristics and detection reliability. Previous studies have shown that TZ lesions tend to have lower biopsy positivity rates than PZ lesions, especially in the midgland and base regions,8 possibly due to variations in histology or technical challenges in sampling.
The transperineal biopsy approach has gained increasing preference due to its favorable safety profile and enhanced diagnostic reach.9 Compared to the transrectal route, the transperineal method substantially lowers infection risk and provides improved access to the anterior and apical prostate regions, which are common sites for TZ tumors.10, 11, 12 Evidence from large-scale studies, including the TRANSLATE trial, has further validated the diagnostic benefits of transperineal biopsy performed under local anesthesia.13 Nonetheless, the most effective targeting strategies for TZ lesions remain unclear. Factors such as benign prostatic hyperplasia, variable tissue composition, and technical challenges may impede precise localization and sampling of TZ lesions.14,15 Further investigation is needed to clarify the association between lesion location, PI-RADS score, and the yield of targeted biopsy in order to optimize diagnostic workflows and inform clinical decision-making.
The present study aimed to assess the predictive role of lesion location—comparing TZ-only versus PZ-involved lesions—in detecting clinically significant prostate cancer through MRI-targeted biopsy. We hypothesized that TZ-only lesions would show lower csPCa detection rates than lesions involving the PZ, and that this association may differ by PI-RADS category.

2
Material and methods
2.1
Study population and design
This retrospective, single-center study included 338 consecutive patients who underwent biparametric MRI (bpMRI)-guided targeted biopsy between April 2023 and October 2024. The study protocol was reviewed and approved by the Institutional Review Board of Juntendo University (Approval No [13-089]). As this was a retrospective analysis using existing clinical data, the requirement for individual informed consent was waived. In line with institutional and ethical policies, an opt-out approach was employed, with study details publicly disclosed and patients given the opportunity to refuse participation. Eligible participants were those with suspicious lesions on bpMRI, assigned PI-RADS scores of 3–5, and for whom targeted and systematic biopsy results were available. Only patients with complete clinical and imaging datasets—including PSA values, prostate volume, and lesion characteristics—were included in the analysis. All bpMRI scans were performed using either 1.5-Tesla or 3-Tesla MRI systems. The imaging protocol comprised T2-weighted and diffusion-weighted imaging sequences, following the PI-RADS v2.1 standards. Experienced radiologists interpreted all imaging studies. All PI-RADS assessments were performed by five board-certified radiologists with at least 8 years of clinical experience in prostate MRI interpretation. Target lesions were classified based on their PI-RADS scores (3, 4, or 5) and anatomical location—specifically, whether they were located in the TZ or PZ. Lesions confined entirely to the TZ were labeled “TZ only,” while those involving the PZ—either entirely or spanning the TZ and PZ—were categorized as “PZ included.” This binary classification was selected to align with clinical decision-making regarding biopsy strategy and anticipated diagnostic yield.

2.2
Biopsy procedure
All patients underwent targeted and systematic biopsies via the transperineal route under general anesthesia, utilizing transrectal ultrasound guidance in combination with MRI/ultrasound fusion technology. The procedures were performed using the BioJet® system (D&K Technologies, Kanalweg, Germany) and the ARIETTA 70® ultrasound system (Hitachi Aloka Medical Ltd., Japan). Targeted biopsies included 1–4 cores per lesion, determined by the size of the lesion. This approach is consistent with recent evidence demonstrating that while obtaining 5 cores per lesion maximizes detection rates, approximately 95% of clinically significant cancers are detected by the first three cores.16 When multiple target lesions (PI-RADS ≥3) were identified, targeted biopsies were performed for each lesion separately. Systematic biopsies followed a 16-core template to ensure thorough sampling of the entire prostate with 4 cores obtained from each peripheral zone (PZ) bilaterally and 4 cores from each transition zone (TZ) bilaterally, including 2 cores from the anterior aspect of each TZ. Histopathological evaluation was carried out by experienced genitourinary pathologists according to the guidelines of the International Society of Urological Pathology.

2.3
Outcome definition
The primary outcome was the detection of csPCa in the targeted biopsy cores, defined as a Gleason score of 3 + 4 or greater (ISUP Grade Group ≥2). For analytical purposes, biopsy results were categorized according to the presence or absence of csPCa in targeted and systematic biopsy samples. A positive targeted biopsy was defined as the detection of csPCa within the targeted cores, regardless of findings from the systematic biopsy. Candidate predictor variables were chosen based on prior studies and clinical significance. These included the anatomical location of the target lesion (classified as “TZ only” for lesions confined to the TZ and “PZ included” for those involving the PZ), PSA density (calculated by dividing serum PSA level by prostate volume), prostate volume (measured in milliliters using T2-weighted MRI), maximum lesion diameter (measured in millimeters on T2-weighted MRI), and PI-RADS score (categorized as 3, 4, or 5, with PI-RADS 3 serving as the reference category in the analysis).

2.4
Statistical analysis
2.4.1
Descriptive statistics
Continuous variables are reported as both median (range) and mean ± standard deviation. To account for the potential non-normal distribution often seen in clinical datasets, data distribution was assessed by comparing means and medians, and both were reported. Categorical variables are summarized as frequencies and percentages.

2.4.2
Univariate screening
All candidate predictor variables were examined through univariate analysis to assess their association with the primary outcome—detection of csPCa in targeted biopsy cores. For binary variables, 2 × 2 contingency tables were used to calculate odds ratios (ORs) with 95% confidence intervals (CIs). The chi-squared test was used to determine statistical significance. For continuous variables, Welch's t-test was employed to compare groups with and without csPCa detected in targeted biopsy, as this test does not require equal variances and is robust against moderate deviations from normality. All variables showed statistical significance (P < 0.001) and met the inclusion threshold (P < 0.20) for the multivariable analysis.

2.4.3
Multivariable logistic regression
A multivariable logistic regression model was constructed using maximum likelihood estimation. The outcome variable was the presence of csPCa in the targeted biopsy cores. All clinically relevant variables identified through univariate analysis were included in the model. The PI-RADS score was analyzed as a categorical variable, with PI-RADS 3 designated as the reference group.

2.4.4
Statistical software
All statistical analyses were performed using EZR (Easy R), version 1.68 (Jichi Medical University, Saitama, Japan), a freely accessible graphical interface for R (The R Foundation for Statistical Computing, Vienna, Austria).17 A two-sided P-value of <0.05 was considered statistically significant for all comparisons.

3
Results
3.1
Study population characteristics
A total of 338 patients met the eligibility criteria and were included in the final analysis. Targeted biopsies were positive for csPCa in 138 patients (40.8%), while 200 patients (59.2%) had negative results from the targeted biopsy. The overall cancer detection rate (positive findings in either targeted or systematic biopsy) was 52.4%, with 177 patients diagnosed. Patient characteristics stratified by lesion location (TZ-only vs PZ-included lesions) are presented in Table 1. Among 338 patients, 173 (51.2%) had TZ-only lesions and 165 (48.8%) had PZ-included lesions. PSA density had a median value of 0.161 (range, 0.014–2.885) and showed considerable variability, with a mean of 0.232 ± 0.233, indicating a right-skewed distribution. Prostate volume had a median of 38.0 ml (range, 11.0–131.0 ml) and a mean of 43.7 ± 21.0 ml, also reflecting a right-skewed pattern. The maximum lesion diameter had a median of 10.0 mm (range, 3.0–40.0 mm) and a mean of 11.5 ± 5.7 mm, approximating a normal distribution.
TZ-only lesions were associated with significantly larger prostate volumes (47.1 ± 21.4 ml vs 40.3 ± 20.1 ml, P = 0.003). The total number of target biopsy cores obtained (representing the cumulative cores taken from all targeted lesions per patient) was comparable between groups (P = 0.444), with a median of 3 cores in both groups. The proportion of patients with ≥3 cores obtained was also similar between TZ-only and PZ-included lesions (70.5% vs 63.0%, P = 0.166). Other patient characteristics, including age (P = 0.551), PSA levels (P = 0.342), PSA density (P = 0.381), and maximum lesion diameter (P = 0.434), showed no significant differences between lesion location groups. Previous prostate biopsy history was comparable between TZ-only and PZ-included groups (P = 0.125).

3.2
Targeted and systematic biopsy by lesion location
Detection rates for clinically significant prostate cancer were compared between different biopsy approaches, stratified by lesion location (Table 2). Cases positive on both targeted and systematic biopsy were observed in 21.4% (37/173) of TZ-only lesions and 50.3% (83/165) of PZ-included lesions, with an overall rate of 35.5% (120/338). Systematic biopsy detected csPCa in 32.4% (56/173) of TZ-only lesions and 62.4% (103/165) of PZ-included lesions, with an overall rate of 47.0% (159/338). Targeted biopsy detected csPCa in 27.2% (47/173) of TZ-only lesions and 55.2% (91/165) of PZ-included lesions, with an overall rate of 40.8% (138/338).
Significant differences were observed between TZ-only and PZ-included lesions across all biopsy methods (P < 0.001). Agreement analysis revealed discordant results in both anatomical zones. Targeted biopsy detected cases not identified by systematic biopsy in 5.8% (10/173) of TZ-only lesions and 4.8% (8/165) of PZ-included lesions. Systematic biopsy detected cases not identified by targeted biopsy in 11.0% (19/173) of TZ-only lesions and 12.1% (20/165) of PZ-included lesions. Both methods were negative in 61.8% (107/173) of TZ-only lesions and 32.7% (54/165) of PZ-included lesions.

3.3
Target biopsy positivity by PI-RADS and location
The distribution of PI-RADS was as follows: 42 patients (12.4%) had PI-RADS 3 lesions, 218 patients (64.5%) had PI-RADS 4 lesions, and 78 patients (23.1%) had PI-RADS 5 lesions. In terms of lesion location, 173 patients (51.2%) had lesions classified as TZ-only, while 165 patients (48.8%) had PZ-included lesions. Fig. 1 illustrates the association between PI-RADS score, lesion location, and biopsy outcomes. A consistent trend was observed, with PZ-included lesions showing higher target biopsy positivity rates across all PI-RADS categories. The most notable difference was seen in PI-RADS 4 lesions, where target positivity was 53.8% for PZ-included lesions, compared to 24.8% for TZ-only lesions. Detailed stratified results are shown in Table 3. For PI-RADS 3 lesions, the overall detection rate of csPCa was 21.4%, with a target positivity of 7.1%. PI-RADS 4 lesions had an overall csPCa detection rate of 51.8% and a target positivity rate of 40.4%. PI-RADS 5 lesions exhibited the highest detection rates, with 70.5% overall and 60.3% target positivity. Table 4 displays the ORs comparing TZ-only and PZ-included lesions across the PI-RADS categories. The greatest OR was observed in PI-RADS 4 (OR = 3.55), indicating the most substantial advantage of PZ inclusion in this intermediate-risk group. PI-RADS 5 also showed a benefit (OR = 1.44), while PI-RADS 3 had a high OR (OR = 3.4) despite relatively low absolute detection rates.

3.4
Univariate analysis
Univariate screening analysis revealed that all evaluated variables were significantly associated with target biopsy positivity (Table 5). Lesions classified as TZ-only exhibited a significant negative association with target biopsy positivity, as indicated by Pearson's chi-squared test (χ2 = 27.38, df = 1, P < 0.001). Among the continuous variables, PSA density showed the strongest correlation with target biopsy positivity, with a significantly higher mean in the positive group (mean difference, +0.19; t = −6.86, df = 166.1, P < 0.001). Prostate volume was also significantly associated, with a lower average volume observed in patients with positive target biopsies (mean difference, −19.22 ml; t = 10.13, df = 326.2, P < 0.001). Maximum lesion diameter demonstrated a significant positive association as well, with larger lesions seen in the positive group (mean difference, +2.39 mm; t = −3.73, df = 250.4, P < 0.001). PI-RADS classification also showed a strong univariate association with target biopsy positivity. Compared to PI-RADS 3, PI-RADS 4 lesions had a significantly higher detection rate (χ2 = 15.66, df = 1, P < 0.001), while PI-RADS 5 lesions demonstrated an even stronger association (χ2 = 29.54, df = 1, P < 0.001). These results reinforce the predictive value of higher PI-RADS scores. All variables fulfilled the inclusion threshold (P < 0.20) for the multivariable analysis, with each showing high statistical significance (P < 0.001).

3.5
Multivariable logistic regression model
The final multivariable logistic regression model incorporated all six predictor variables (Table 6). Lesions classified as TZ-only were found to be a strong independent negative predictor of target biopsy positivity (OR = 0.367; 95% CI, 0.207–0.650; P < 0.001), indicating that TZ-only lesions had 63% lower odds of a positive targeted biopsy compared to PZ-included lesions. Among the continuous variables, PSA density demonstrated the strongest association, with each unit increase linked to a markedly higher likelihood of positivity (OR = 35.2; 95% CI, 3.84–323.0; P = 0.002). Prostate volume was inversely associated with biopsy positivity; each 1 ml increase corresponded to a 5.5% reduction in odds (OR = 0.945; 95% CI, 0.923–0.967; P < 0.001). Maximum lesion diameter also showed a positive association, with each 1 mm increase associated with a 9% increase in odds (OR = 1.090; 95% CI, 1.010–1.170; P = 0.020). The PI-RADS score remained a significant predictor with notable effect sizes. Relative to PI-RADS 3, PI-RADS 4 lesions had 5.49 times higher odds of positivity (95% CI, 1.44–20.9; P = 0.013), while PI-RADS 5 lesions showed 6.8 times higher odds (95% CI, 1.51–30.5; P = 0.013). Assessment of multicollinearity using variance inflation factors (VIF) indicated no problematic relationships among the predictor variables. VIF values ranged from 1.02 to 1.43, with the highest VIF observed for maximum lesion diameter, remaining well below the commonly accepted threshold of 2.5. The multivariable logistic regression model demonstrated a significant improvement over the null model (χ2 = 154.79, df = 6, P < 0.001)

4
Discussion
This study found that TZ-only lesions had significantly lower positivity rates compared to PZ-included lesions in MRI/ultrasound fusion-guided targeted biopsy. In multivariate analysis, the TZ-only location was identified as an independent negative predictor of targeted biopsy positivity (OR = 0.367; 95% CI, 0.207–0.650; P < 0.001), underscoring the diagnostic difficulties associated with targeting TZ lesions. These findings align with previous studies that have reported similar challenges in detecting TZ lesions. Hoeks et al observed reduced detection sensitivity for TZ cancers compared to PZ cancers using 3T MRI, attributing the difference in part to the heterogeneity of tissue caused by benign prostatic hyperplasia.15 Likewise, Oto et al noted that even with diffusion-weighted and contrast-enhanced MRI, distinguishing cancerous from benign tissue in the TZ remains difficult.14 In our study, the most substantial difference was seen in PI-RADS 4 lesions, with PZ-included lesions showing a 53.8% positivity rate compared to 24.8% for TZ-only lesions. This highlights the relevance of anatomical location within the PI-RADS 4 category and suggests that more cautious interpretation may be warranted when evaluating TZ lesions. Supporting this, Susan et al reported notably lower detection rates of clinically significant cancers in TZ lesions compared to PZ lesions during MRI-targeted biopsy.18 They proposed that this could be due to TZ tumors being located more anteriorly and farther from the end-firing TRUS probe, which reduces lesion visibility with increasing distance from the transrectal entry point. Although our study utilized a transperineal approach—which theoretically offers better access to anterior regions—TZ lesions still demonstrated significantly lower detection rates than PZ lesions. This points to potential limitations in lesion visibility on MRI. A combination of factors is likely responsible for the reduced detection rates in TZ lesions. First, the TZ is a common site for age-related benign prostatic hyperplasia, which complicates distinguishing malignant lesions on MRI.19 Second, TZ cancers are often lower grade, which may contribute to their reduced visibility on MRI.20 These observations underscore the need to refine diagnostic approaches for TZ lesions. Potential strategies include assigning greater weight to systematic biopsy and more proactive use of clinical indicators such as PSA density. Notably, our study identified PSA density as one of the most robust predictors of targeted biopsy positivity (OR = 35.2, P = 0.002), indicating its potential value in evaluating TZ lesions.
This study has several noteworthy limitations. First, as a retrospective analysis from a single institution, the possibility of selection bias cannot be eliminated. Second, the pathological evaluation relied on biopsy samples rather than whole-mount prostatectomy specimens, which may not accurately capture the exact location and volume of cancer. Third, although multiple radiologists interpreted the MRI scans, interreader variability was not assessed.21 Fourth, in our binary classification scheme, 16 patients with lesions spanning both TZ and PZ were grouped under the PZ-included category due to their limited number. While this allowed sufficient statistical power, it may have introduced heterogeneity into the PZ-included group. These mixed-location lesions might represent an intermediate subtype with detection rates between those of pure TZ and pure PZ lesions, although their small number precluded separate analysis. Fifth, our analysis did not account for the specific anatomical location of lesions within the apex-to-base dimension of the prostate. Given that transperineal biopsy approaches may have varying access to apical versus basal regions, future studies should consider stratifying lesions by both zonal location (TZ vs PZ) and apex-to-base positioning to better understand potential sampling biases and their impact on detection rates. Lastly, since only the transperineal biopsy approach was employed, a direct comparison with the transrectal route could not be made. Future studies should focus on determining optimal biopsy techniques for TZ lesions through prospective multicenter trials. It is especially important to investigate whether integrating clinical variables—such as PSA density, lesion size, and PI-RADS score—can enhance cancer detection in this anatomically and radiologically complex zone.22 Moreover, the use of artificial intelligence for image interpretation and advancements in MRI technology may further aid in identifying TZ lesions.23, 24, 25
In conclusion, this study demonstrated that TZ lesions exhibited markedly lower positivity rates than PZ lesions in MRI-targeted biopsy. These results highlight the critical role of lesion location in prostate cancer detection and indicate the need for a more tailored approach when evaluating TZ lesions. Clinicians should be aware of these diagnostic challenges and consider incorporating additional systematic biopsies and thorough evaluation of clinical parameters in the management of patients with TZ lesions.

Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request, subject to institutional review board approval and appropriate data use agreements.

Contributions
All authors have significantly contributed to the study and are in agreement with the content of the manuscript. Each author's contribution is as follows: N.N. and S.H designed and directed the project. N.N. processed the experimental data, performed the analysis. N.N. wrote the manuscript, and T.S., K.T., T.K., T.O., M.N., H.I. and S.H. revised it.

Declaration of Generative AI and AI-assisted technologies in the writing process
No generative AI or AI-assisted technologies were used in the writing process of this manuscript.

Conflict of interest
There is no conflict of interest.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.