Abstract
Background
The comparative diagnostic performance of biparametric MRI (bpMRI) versus multiparametric MRI (mpMRI) for clinically significant prostate cancer (csPCa) continues to be debated. This study aimed to compare mpMRI and bpMRI in detecting csPCa across prostate-specific antigen (PSA) strata and identify supplementary tools comparable to dynamic contrast-enhanced (DCE) imaging.
Methods
Images were evaluated using mpMRI-based mp-PI-RADS and bpMRI-based bp-PI-RADS and simplified PI-RADS (S-PI-RADS) schemes. The lesion volume (LV) was manually segmented by a radiologist using ITK-SNAP software on high b-value DWI images. The diagnostic performance was assessed via receiver operating characteristic (ROC) curve analysis. The differences of T2WI-score, DCE assessment and LV between csPCa and non-csPCa in peripheral zone (PZ) with DWI category 3 were compared.
Results
For overall PSA, mp-PI-RADS and bp-PI-RADS showed comparable AUCs (0.889 vs. 0.882; P > 0.05). When PSA ≤ 10 ng/ml, mp-PI-RADS exhibited the highest specificity (91.0% vs. bp-PI-RADS: 64.4%, S-PI-RADS: 75.0%) and PPV (73.0% vs. bp-PI-RADS: 47.7%, S-PI-RADS: 52.5%). When PSA > 10 ng/ml, S-PI-RADS demonstrated higher sensitivity (91.6% vs. mp-PI-RADS: 83.2%, bp-PI-RADS: 81.2%) and F1-score (0.873 [0.822–0.924] vs. mp-PI-RADS: 0.832 [0.778–0.886], bp-PI-RADS: 0.831 [0.777–0.885]). Among DWI category 3 PZ lesions, neither DCE nor T2WI significantly stratified csPCa risk (P = 0.657 and P = 0.424), whereas LV ≥ 0.5 cm³ showed markedly higher csPCa risk (83.8% vs. 45.8%; P < 0.001).
Conclusions
While mpMRI and bpMRI exhibit comparable overall diagnostic performance but context-dependent strengths: mpMRI demonstrates higher specificity for avoiding unnecessary biopsies when PSA ≤ 10 ng/ml, whereas bpMRI (particularly S-PI-RADS) maximizes sensitivity for csPCa detection when PSA > 10 ng/ml. LV is anticipated to serve as a complementary radiological biomarker at the absence of DCE.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12880-025-01884-x.
Keywords: Multiparametric MRI, Biparametric MRI, Clinically significant prostate cancer, Comparison, Simplified prostate imaging report and data system
Introduction
The latest data on malignant tumor surveillance, released by the National Cancer Center in 2022, reveals that the proportion of prostate cancer (PCa) among male malignancies has increased from 3.5 to 5.3%, with its incidence consistently ranking sixth among male malignancies [1]. The clinical adoption of multiparametric MRI (mpMRI) as a non-invasive triage tool for detecting clinically significant PCa (csPCa) prior to biopsy has become well-established in diagnostic protocols [2, 3]. In 2012, the Prostate Imaging Report and Data System (PI-RADS) was introduced to standardize mpMRI scanning protocols, image interpretation, and reporting [4]. The initial version exhibited substantial interobserver variability in interpretation, attributed to excessive reference sequences and suboptimal integration of sequence-specific scoring. Consequently, the second edition emphasizes the significance of T2-weighted imaging (T2WI) and diffusion-weighted imaging (DWI) sequences, while diminishing the role of dynamic contrast-enhanced imaging (DCE-imaging) to a binary score for qualitative assessment of focal enhancement [5]. In version 2.1 of PI-RADS (PI-RADS v2.1) released in 2019 [6], DCE-imaging was only used to upgrade peripheral zone (PZ) DWI category 3 lesions to category 4, without impacting transition zone (TZ) lesions.
The clinical utility of DCE-imaging as an adjunct to conventional T2WI and DWI remains contentious in diagnostic practice. Beyond the debate over its added value, DCE implementation poses technical challenges requiring advanced MRI systems and specialized operational expertise. These practical considerations have prompted growing consensus within the urological imaging community that DCE may be omitted from routine protocols, giving rise to the biparametric MRI (bpMRI) paradigm [7–9]. The efficacy of bpMRI in detecting csPCa has been demonstrated by extensive clinical validation, with its diagnostic accuracy comparable to that of mpMRI [10, 11]. Omitting DCE not only simplifies scanning, reduces imaging time, and lowers the cost of medical treatment but also mitigates the potential risk of systemic renal fibrosis and renal failure associated with contrast medium [12]. Nevertheless, neither traditional PI-RADS nor its bpMRI adaptation provides granular therapeutic guidance, particularly regarding indeterminate category 3 lesions. Scialpi et al. developed a simplified bpMRI-based PI-RADS (S-PI-RADS), which categorizes PI-RADS 3 lesions into subgroups 3a (< 0.5 cm3) and 3b (≥ 0.5 cm3) based on lesion volume (LV), and recommends biopsy for lesions classified as 3b or higher [13].
Prostate-specific antigen (PSA) serves as a critical biomarker in PCa screening and biopsy decision-making. Clinically, the PSA level ranging from 4 to 10 ng/ml is referred to as the diagnostic gray zone, wherein csPCa may be detected, but most patients (approximately 70%) may undergo unnecessary biopsies [14]. When PSA levels exceed 10 ng/ml, the detection rate of prostate cancer (PCa) increases significantly, reaching approximately 62%: 46% for PSA levels between 10 and 20 ng/mL, 76% for 20–50 ng/mL, and 93% for > 50 ng/ml, which strongly indicates the need for biopsy consideration [15]. However, the level of PSA can be influenced by various factors such as prostate massage, benign prostatic hyperplasia (BPH), prostatitis, and others. Consequently, relying solely on PSA for diagnosis may result in a higher misdiagnosis rate. Han et al.‘s study revealed that bpMRI exhibits significantly superior performance compared to mpMRI in gray-zone PSA patients, achieving an AUC of 0.884 with 88.4% specificity [16]. This paradigm-shifting finding contrasts with earlier investigations, potentially attributable to refined patient stratification through rigorous PSA-level categorization - a methodological enhancement absent in previous trial designs.
Due to the limited availability of comparison among mpMRI and bpMRI in detecting csPCa across various PSA stratifications, and given the emerging clinical demand for validating S-PI-RADS diagnostic frameworks, this study aims to (1) compare and analyze the diagnostic efficacy of bpMRI-based PI-RADS (bp-PI-RADS) and S-PI-RADS as well as mpMRI-based PI-RADS (mp-PI-RADS) in case with PSA ≤ 10 ng/ml and PSA > 10 ng/ml, respectively; (2) investigate a supplementary quantitative tool comparable to DCE-imaging to resolving diagnostic uncertainty.
Methods
Study population
1347 patients who underwent prostate MRI examinations from institution 1 (September 2022 - December 2024) and from institution 2 (January 2021 - December 2023) were collected continuously. 737 patients were excluded according to the following criteria: (1) lack of histopathological results or clinical data (n = 563); (2) biopsies or other treatments prior to MRI (n = 107); (3) absence of DCE-imaging in the MRI scanning protocol (n = 46) and (4) inadequate quality of MRI images that did not meet the evaluation requirements (n = 21). The final analytical cohort comprised 610 eligible patients. Figure 1 delineates the complete patient selection workflow.
Fig. 1.
The flowchart depicting the criteria for patient inclusion and exclusion
MRI technique
The baseline mpMRI of both agencies was conducted using phased-array body coils on 3.0T scanners (Philips Ingenia and Philips Achieva, respectively). The scan parameters were set in accordance with the recommendations of PI-RADS v2.1 [6], which included T2WI, high b-value DWI (b = 1500 s/mm2 for Institution 1 and b = 2000 s/mm² for Institution 2) and corresponding apparent diffusion coefficient (ADC) maps, as well as axial DCE sequences covering the prostate and seminal vesicles. During DCE scanning, the Gd-containing contrast medium was administered at a dose of 0.1 mmol/kg with an injection rate ranging from 2 ~ 3 ml/s, while maintaining a time resolution of 10 ~ 15 s. The MRI plans of the two agencies are detailed in Table 1.
Table 1.
Detailed sequence parameters for prostate multi-parametric MRI in both institution
| Parameter | Institution 1 | Institution 2 | ||||
|---|---|---|---|---|---|---|
| T2WI | DWI | DCE-MRI | T2WI | DWI | DCE-MRI | |
| TR (ms) | 5639 | 4000 | 4.2 | 4569 | 3997 | 3.2 |
| TE (ms) | 125 | 50 | 0 | 91 | 70 | 1.53 |
| Thickness (mm) | 3 | 3 | 3 | 3 | 3 | 3 |
| Interslice gap (mm) | 0 | 0 | 0 | 0 | 0 | 0 |
| Matrix size | 320 × 280 | 184 × 184 | 256 × 256 | 320 × 320 | 128 × 128 | 224 × 224 |
| FOV (mm×mm) | 200 × 200 | 203 × 203 | 300 × 270 | 200 × 200 | 200 × 200 | 250 × 250 |
| No. of signals acquired | 2 | 2 | 1 | 2 | 3 | 1 |
| b factor (s/mm2) | 0, 50, 1500 | 0, 50, 800, 2000 | ||||
TR Repetition time, TE Time echo, FOV Field of view, T2WI T2-Weighted Imaging, DCE dynamic contrast-enhanced
Image analysis
The MRI images were independently assessed by two radiologists with 5 years (reader 1) and 3 years (reader 2) of experience in interpreting prostate MRI. While informed of patient age, both interpreters remained blinded to PSA values and pathological results. During the initial mp-PI-RADS assessment, readers sequentially reviewed T2WI, DWI, ADC, and DCE sequences, assigning individual scores for each sequence. The bp-PI-RADS category was simultaneously derived from the same session by integrating T2WI and DWI scores while excluding DCE data. S-PI-RADS categories were subsequently assigned using LV measurements from the same dataset. As previous studies have indicated that the incremental value of DCE imaging over T2WI is limited [17, 18], we sought to implement minor modifications to the standard bp-PI-RADS protocol: for PZ lesions with a DWI category of 3, the final score was determined based on T2WI. Specifically, if the T2WI score was ≥ 4, the overall category was upgraded to 4; however, other category lesions were remained consistent with mp-PI-RADS. In S-PI-RADS scheme, PI-RADS 4 and 5 lesions were classified as category 4. DWI was the dominant sequence of both PZ and TZ lesions. Based on the LV, the DWI category 3 lesions were further divided into 3a (< 0.5 cm3) and 3b (≥ 0.5 cm3). The disputed lesions were resolved through negotiation between the two readers, with the assistance of a senior radiologist (reader 3) who possesses 10 years of experience in prostate MRI interpretation. Supplementary Figure S1 illustrates the overall scoring process for the three schemes.
The LVs were manually segmented slice-by-slice by Reader 1 using ITK-SNAP software (version 3.4.0), with all contours drawn directly on high b-value DWI images. Subsequently, 65 lesions were randomly segmented by reader 2 to assess the reproducibility of LV segmentation. Meanwhile, reader 3 assumed the responsibility of supervising and reviewing their segmentation work. The MRI-prostate volume was calculated using the formula: maximum anteroposterior diameter × maximum transverse diameter × maximum length diameter × 0.52 (cm3).
Radiological-pathological correlation
Histopathological confirmation via transrectal ultrasound (TRUS)-cognitive fusion targeted biopsy served as the reference standard. Two experienced ultrasound physicians, each with over 10 years of experience, reviewed MRI images prior to the procedure in order to facilitate the location of suspicious lesions during TRUS examination. Target lesions received 2–3 core samplings, complemented by 10-core systematic biopsy (apex, mid-gland, and base regions bilaterally). The positions of all cores were recorded to correspond with the MRI image. If the lesion evaluated by the reader was located in the same area as the pathological report, it was considered a radiological-pathological match.
The biopsy cores were assessed by full-time urogenital pathologists with more than 12-years working experience. According to consensus conference of the International Society of Urological Pathology (ISUP) [19], csPCa was defined as a tumor with Gleason score (GS) ≥ 7, while tumors with GS = 6 were defined as clinically insignificant PCa (ciPCa) and grouped together with BPH in the non-csPCa group.
Statistical analysis
Statistical analysis was conducted at the patient level, using SPSS (v22.0; IBM), MedCalc (v15.0; MedCalc Software), and R (v4.1.0; R Foundation). Non-normally distributed continuous variables (confirmed by Shapiro-Wilk test, P < 0.05) are reported as median (interquartile range [IQR]). Interreader agreement for PI-RADS assessments was evaluated by weighted Kappa test. For the detection of csPCa at overall PSA (undifferentiated PSA level), PSA ≤ 10 ng/ml and PSA > 10 ng/ml, the diagnostic efficacy of mp-PI-RADS, bp-PI-RADS and S-PI-RADS was assessed using receiver operating characteristic (ROC) curve and area under the curve (AUC). The optimal threshold of each scheme was determined by the maximum Youden index, and calculated sensitivity, specificity, accuracy, positive predictive value (PPV) and negative predictive value (NPV). F1-score was adopted to evaluate the performance of balancing sensitivity and PPV, and 1000-bootstrapping was used to evaluate statistical significance. Delong test was used to compare the differences of AUC. Mann-Whitney U test was employed to compare the differences of T2WI-score, DCE evaluation, LV between csPCa and non-csPCa in PZ lesions with DWI category 3. Kruskal-Wallis test with Dunn’s post-hoc analysis was used for comparing LV differences across lesions of varying histological grades. Bland-Altman analysis assessed inter-reader variability in LV measurements, reporting mean difference (bias) and 95% limits of agreement (95%LoA = mean difference ± 1.96 × standard deviation). The P-values for comparisons among multiple groups were adjusted using the Bonferroni correction method, P < 0.05 was considered to be statistically significant.
Results
Baseline data
Among 610 patients, 43.8% (267/610) were identified as csPCa and 56.2% (343/610) were non-csPCa. Specifically, there were 41.5% (253/610) patients with PSA ≤ 10 ng/ml and 58.5% (357/610) patients with PSA > 10 ng/ml. In PSA ≤ 10 ng/ml and PSA > 10 ng/ml group, the proportions of csPCa were 25.7% (65/253) and 56.6% (202/357), respectively, and the difference was statistically significant (P < 0.001). Detailed clinical and pathological baseline characteristics are systematically summarized in Table 2.
Table 2.
The patient’s baseline clinicopathological data at overall PSA level, PSA ≤ 10 ng/ml and PSA > 10 ng/ml
| Characteristics | Overall PSA level | PSA ≤ 10 ng/ml | PSA > 10 ng/ml | P value |
|---|---|---|---|---|
| Total, n (%) | 610 (100%) | 253 (41.5%) | 357 (58.5%) | |
| age, years, median (IQR) | 70 (64, 75) | 68 (63, 73) | 71 (65, 77) | < 0.001a |
| fPSA/tPSA, median (IQR) | 0.13 (0.09, 0.19) | 0.16 (0.12, 0.21) | 0.12 (0.08, 0.18) | < 0.001a |
| PV, cm3, median (IQR) | 50.8 (35.7, 73.3) | 46.9 (33.5, 67.3) | 52.2 (37.8, 77.6) | 0.003a |
| PSAD, ng/ml/cm3, median (IQR) | 0.23 (0.13, 0.45) | 0.13 (0.09, 0.20) | 0.36 (0.22, 0.80) | < 0.001a |
| Location, n (%) | 0.006b | |||
| PZ | 259 (42.5%) | 91 (14.9%) | 168 (27.5%) | |
| TZ | 351 (57.5%) | 162 (26.6%) | 189 (31.0%) | |
| Lesion size, mm, median (IQR) | 13.9 (9.3, 20.1) | 11.4 (7.8, 15.2) | 16.1 (10.7, 22.8) | < 0.001a |
| LV(cm3), median (IQR) | 0.51 (0.21, 1.40) | 0.33 (0.15, 0.69) | 0.85 (0.31, 2.26) | < 0.001a |
| ISUP grade, n (%) | < 0.001b | |||
| Benign | 286 (46.9%) | 159 (62.8%) | 127 (35.6%) | |
| 1 | 57 (9.3%) | 29 (11.5%) | 28 (7.8%) | |
| 2 | 50 (8.2%) | 24 (9.5%) | 26 (7.3%) | |
| 3 | 59 (9.7%) | 17 (6.7%) | 42 (11.8%) | |
| 4 | 83 (13.6%) | 16 (6.3%) | 67 (18.8%) | |
| 5 | 75 (12.3%) | 8 (3.2%) | 67 (18.8%) |
PSA: prostate-specific antigen, PV Prostate volume, PSAD PSA density, TZ Transition zone, PZ Peripheral zone, LV lesion volume; ISUP International Society Of Urological Patheology
a Mann-Whitney U test, b Chi-square test
Reader-to-reader consistency
The two readers assigned identical categories to 74.1% (452/610) patients by the mp-PI-RADS scheme, while there were 81.0% (494/610) and 87.7% (535/610) patients with matching categories by the bp-PI-RADS and S-PI-RADS protocols respectively (Fig. 2). The weighted Kappa values were 0.733, 0.792 and 0.810 respectively. The Bland-Altman scatter plot demonstrates that the mean volume difference of the 65 lesions measured by both readers is 0.02 ± 0.08 cm3. The scatter of 92.3% (60/65) falls within the upper and lower limits of the 95%LoA (Fig. 2). The extremely small measurement standard deviation and high limits of agreement compliance rate confirming excellent volumetric reproducibility.
Fig. 2.
The categories and specific numbers of mp-PI-RADS, bp-PI-RADS and S-PI-RADS evaluated by two radiologists. mp-PI-RADS: Kappa coefficient = 0.733 (a); bp-PI-RADS: Kappa coefficient = 0.792 (b); S-PI-RADS: Kappa coefficient = 0.810 (c); the evaluation consistency of both bpMRI schemes was superior to that of mpMRI. The Bland-Altman analysis revealed that the mean error in lesion volume measurement was merely 0.02 ± 0.08 cm³, indicating excellent repeatability and reliability (d)
Comparison of three schemes for detecting csPCa
For overall PSA, PSA ≤ 10 ng/ml, and PSA > 10 ng/ml, the diagnostic thresholds of mp-PI-RADS and S-PI-RADS were category 4 and 3b. However, bp-PI-RADS observed a threshold of category 3 when the PSA ≤ 10 ng/ml, while it was 4 for overall PSA or PSA > 10 ng/ml. The detection rates and diagnostic efficacy calculated based on the corresponding thresholds for the three schemes are shown in Fig. 3 and Table 3.
Fig. 3.
The csPCa detection rates in each assessment category of mp-PI-RADS, bp-PI-RADS, and S-PI-RADS at overall PSA level, PSA ≤ 10 ng/ml and PSA > 10 ng/ml
Table 3.
Comparison of diagnostic efficacy of mp-PI-RADS, bp-PI-RADS and S-PI-RADS at different PSA levels
| Variables | cut-off | SEN, % | SPE, % | ACC, % | PPV, % | NPV, % | F1-score | AUC | P value* |
|---|---|---|---|---|---|---|---|---|---|
| Overall PSA level | |||||||||
| mp-PI-RADS | 4 | 80.1 | 85.4 | 83.1 | 81.1 | 84.7 | 0.806 (0.752–0.860) |
0.889 (0.863–0.914) |
0.285a |
| bp-PI-RADS | 4 | 77.9 | 84.8 | 81.8 | 80.0 | 83.1 | 0.789 (0.734–0.844) |
0.882 (0.856–0.908) |
0.072b |
| S-PI-RADS | 3b | 88.8 | 75.5 | 81.3 | 73.8 | 89.7 | 0.806 (0.752–0.860) |
0.853 (0.824–0.882) |
0.018c |
| PSA ≤ 10 ng/ml | |||||||||
| mp-PI-RADS | 4 | 70.8 | 91.0 | 85.7 | 73.0 | 90.0 | 0.719 (0.642–0.796) |
0.882 (0.836–0.928) |
0.306a |
| bp-PI-RADS | 3 | 93.8 | 64.4 | 71.9 | 47.7 | 96.8 | 0.632 (0.548–0.716) |
0.867 (0.819–0.914) |
0.129b |
| S-PI-RADS | 3b | 80.0 | 75.0 | 76.3 | 52.5 | 91.6 | 0.634 (0.552–0.716) |
0.818 (0.764–0.872) |
0.021c |
| PSA > 10 ng/ml | |||||||||
| mp-PI-RADS | 4 | 83.2 | 78.7 | 81.2 | 83.6 | 78.2 | 0.832 (0.778–0.886) |
0.874 (0.837–0.910) |
1.000a |
| bp-PI-RADS | 4 | 81.2 | 81.3 | 81.2 | 85.0 | 76.8 | 0.831 (0.777–0.885) |
0.872 (0.836–0.909) |
1.000b |
| S-PI-RADS | 3b | 91.6 | 76.1 | 84.9 | 83.3 | 87.4 | 0.873 (0.822–0.924) |
0.866 (0.827–0.904) |
1.000c |
SEN sensitivity, SPE specificity, ACC accuracy, PPV positive predictive value, NPV negative predictive value, AUC areas under the curve
*, The differences in AUCs among the bp-PI-RADS, S-PI-RADS and mp-PI-RADS were compared using the Delong test; a mp-PI-RADS vs. bp-PI-RADS, b bp-PI-RADS vs. S-PI-RADS, c mp-PI-RADS vs. S-PI-RADS. P-values were adjusted using the Bomferroni correction
For overall PSA, the diagnostic efficacy of mp-PI-RADS and bp-PI-RADS was comparable (AUC: 0.889 vs. 0.882, P = 0.285). While S-PI-RADS demonstrated a sensitivity of 88.8%, its overall performance stability was limited (with specificity and PPV at only 75.5% and 73.8% respectively), and the AUC was comparable with bp-PI-RADS (0.853 vs. 0.882, P = 0.072) and significantly lower compared to mp-PI-RADS (0.853 vs. 0.889, P = 0.018).
For PSA ≤ 10 ng/ml, mp-PI-RADS detected csPCa at a rate of 70.8% (46/65) with a PPV of 73.0% and a negative biopsy rate of 27% (17/63). In comparison, bp-PI-RADS and S-PI-RADS had detection rates of 93.8% (61/65) and 80.0% (52/65), respectively, with PPVs of 47.7% and 52.5%, and negative biopsy rates of 52.3% (67/128) and 47.5% (47/99). Compared with bp-PI-RADS and S-PI-RADS, mp-PI-RADS demonstrates a more favorable balance between the detection of csPCa and the reduction of unnecessary biopsies, with F1-scores of 0.719 (0.642–0.796) vs. 0.632 (0.548–0.716) and 0.634 (0.552–0.716), respectively (P = 0.012 and 0.015) (Fig. 4). Of 19 missed csPCa by mpMRI, 73.7% (14/19) were ISUP 2 with median LV = 0.41 cm³, which potentially suitable for active surveillance. Only 5 (26.3%) were ISUP ≥ 3.
Fig. 4.
Radar chart comparing the diagnostic performance of mp-PI-RADS, bp-PI-RADS and S-PI-RADS. For overall PSA, mp-PI-RADS and bp-PI-RADS showed higher AUC and more balanced performance compared to S-PI-RADS. For PSA ≤ 10 ng/ml, mp-PI-RADS better balanced csPCa detection and avoids unnecessary biopsies due to its high positive predictive value and F1-score. For PSA > 10 ng/ml, S-PI-RADS revealed the highest csPCa detection rate and positive predictive value, with a higher F1-score than mp-PI-RADS and bp-PI-RADS
For PSA > 10 ng/ml, despite there was no significant difference in the diagnostic efficacy between mp-PI-RADS and bp-PI-RADS and S-PI-RADS (AUCs were 0.874, 0.872 and 0.866 respectively, P all > 0.05), S-PI-RADS demonstrated higher sensitivity (91.6% [185/202] vs. mp-PI-RADS: 83.2% [168/202], bp-PI-RADS: 81.2% [164/202]) and the numerically highest F1-score (0.873, [0.822–0.924] vs. mp-PI-RADS: 0.832 [0.778–0.886], bp-PI-RADS: 0.831 [0.777–0.885]). For PPV, S-PI-RADS was slightly lower than mp-PI-RADS and bp-PI-RADS (83.3% vs. 83.6%, 85.0%).
Characteristics of DWI category 3 lesions in PZ
A total of 85 PZ lesions exhibited DWI category 3, comprising 53 csPCa and 32 non-csPCa. DCE and T2WI-score did not significantly stratify the lesions. DCE upgraded 54 lesions, with 61.1% (33/54) being csPCa, compared to 64.5% (20/31) in the non-upgraded group (P = 0.657). T2WI-score upgraded 30 lesions, with 56.7% (17/30) being csPCa, compared to 65.5% (36/55) in the non-upgraded group (P = 0.424). LV upgraded 37 lesions, with 83.8% (31/37) being csPCa, compared to 45.8% (22/48) in the non-upgraded group (P < 0.001), proved significantly discriminatory. Similar results were observed in both PSA ≤ 10 ng/ml and PSA > 10 ng/ml subgroups; detailed stratification patterns illustrated in Table 4.
Table 4.
Characteristics of DWI category 3 lesions in peripheral zone
| Characteristics | Overall PSA level | PSA ≤ 10 ng/ml | PSA > 10 ng/ml | ||||||
|---|---|---|---|---|---|---|---|---|---|
| csPCa (n = 53) | non-csPCa (n = 32) | P value | csPCa (n = 19) | non-csPCa (n = 20) | P value | csPCa (n = 34) | non-csPCa (n = 12) | P value | |
| DCE, n (%) | 0.657a | 0.062b | 0.083a | ||||||
| - | 20 (64.5%) | 11 (35.5%) | 7 (77.8%) | 2 (22.2%) | 13 (59.1%) | 9 (40.9%) | |||
| + | 33 (61.1%) | 21 (38.9%) | 12 (40.0%) | 18 (60.0%) | 21 (87.5%) | 3 (12.5%) | |||
| T2WI-score, n (%) | 0.424a | 0.501b | 0.964c | ||||||
| < 4 | 36 (65.5%) | 19 (34.5%) | 14 (53.8%) | 12 (46.2%) | 22 (75.9%) | 7 (24.1%) | |||
| ≥ 4 | 17 (56.7%) | 13 (43.3%) | 5 (38.5%) | 8 (61.5%) | 12 (70.6%) | 5 (29.4%) | |||
| LV, n (%) | < 0.001a | 0.031b | 0.028c | ||||||
| < 0.5 | 22 (45.8%) | 26 (54.2%) | 11 (37.9%) | 18 (62.1%) | 11 (57.9%) | 8 (42.1%) | |||
| ≥ 0.5 | 31 (83.8%) | 6 (16.2%) | 8 (80.0%) | 2 (20.0%) | 23 (85.2%) | 4 (14.8%) | |||
PSA prostate-specific antigen, csPCa clinically significant prostate cancer, T2WI T2-Weighted Imaging, DCE dynamic contrast-enhanced, LV lesion volume
a Chi-square test, b Fisher exact test, c Yates’ correction test
The correlation between LV and histological grade
For overall PSA, benign lesions were smaller than ISUP grade 1, 2, and 3–5 PCa (median volume [IQR]: 0.25 [0.12, 0.49] cm³ vs. 0.45 [0.19, 1.13] cm³ vs. 0.83 [0.42, 1.13] cm³ vs. 2.28 [0.75, 3.04] cm³, all P < 0.01) (Fig. 5). Across PSA risk levels, higher ISUP grade PCa correlated with larger LVs, particularly for ISUP grade 3–5 PCa compared to grade 1 (overall PSA: 2.28 [0.75, 3.04] cm³ vs. 0.45 [0.19, 1.13] cm³, P < 0.001; PSA ≤ 10 ng/ml: 0.79 [0.49, 1.57] cm³ vs. 0.32 [0.13, 1.00] cm³, P = 0.002; PSA > 10 ng/ml: 1.61 [0.92, 3.53] cm³ vs. 0.60 [0.30, 1.39] cm³, P < 0.001).
Fig. 5.
Violin plot comparing the volume differences among various histological types. (a): overall PSA; (b): PSA ≤ 10 ng/ml; (c): PSA > 10 ng/ml. In general, the volume of benign lesions was notably smaller compared to that of csPCa lesions. Additionally, PCa lesions with higher ISUP grade exhibited a larger volume than those with a lower ISUP grade. *** P < 0.001, ** P < 0.01, * P < 0.05
Discussion
This study was purposed to investigate the disparity between mpMRI and bpMRI in the diagnosis of csPCa under different PSA background, in order to provide individualized prostate MRI scanning for patients with PSA ≤ 10 ng/ml and PSA > 10 ng/ml. The bpMRI-based schemes exhibited certain enhancement in inter-observer consistency when compared to the mpMRI-based scheme. Our data reject a blanket equivalence between mpMRI and bpMRI. Instead, they reveal a PSA-dependent trade-off. In PSA gray area, given the low prevalence of csPCa, it is imperative to minimize unnecessary biopsies for patients whenever possible. mpMRI’s high specificity (91.0%) minimizes unnecessary biopsies but misses 29.2% of csPCa (primarily small ISUP 2 lesions suitable for active surveillance). When PSA > 10 ng/ml, the risk of csPCa is really high, making it imperative to optimize csPCa detection rates. S-PI-RADS showed the highest sensitivity and F1-score among the tested schemes, and its PPV was marginally lower than mp-PI-RADS, which may reduce missed csPCa in this high-risk population without significantly increasing negative biopsies. Though its overall diagnostic performance remains comparable to standard approaches, S-PI-RADS offers practical benefits (e.g., no contrast requirement, lower assessment divergence). Moreover, among T2WI-score, DCE evaluation and LV, only LV showed a significant risk stratification effect for DWI category 3 lesions in PZ. This suggests that LV may serve as a valuable supplementary tool for risk stratification when DCE is not available.
The role of DCE-imaging is gradually being diminished with the continuous updates of PI-RADS, but the conclusions of various studies regarding the retention of DCE are inconclusive. Tamada et al. [20] compared the diagnostic efficacy of PI-RADS v2.1 based on mpMRI and bpMRI. Their results showed that mpMRI exhibited higher sensitivity than bpMRI, while the specificity showed an inverse relationship. Consistent findings were observed in our study for both the overall PSA and PSA > 10 ng/ml cohorts. Notably, when PSA ≤ 10 ng/ml, the optimal threshold for bp-PI-RADS was lowered from category 4 to 3, resulting in excessively high sensitivity and markedly reduced specificity. This inconsistent performance profile is unlikely to be suitable for effectively reducing negative biopsies. Druskin et al. [21] conducted MRI-TRUS fusion targeted biopsies for focal prostate lesions with PI-RADS scores of 3, 3 + 1, and 4, revealing an increased proportion of DCE-positive lesions with GS ≥ 7. The results of Xu et al. [22] also showed a significant disparity in DCE results between csPCa and non-csPCa among patients with PI-RADS scores ≥ 3 (particularly PI-RADS 4). Additionally, DCE-imaging can offer assistance beyond the scoring criteria in differentiating PZ anterior edge lesions, enhancing visualization of lesions on T2WI or DWI sequences, and facilitating observation of the lesion capsule [23].
Numerous studies also question the diagnostic value of DCE-imaging, suggesting it may be redundant when provided high-quality T2WI and DWI images [24, 25]. While positive DCE results may enhance the detection rate of csPCa in controversial lesions within the PZ, they could lead to over-biopsy. In this study, within both the overall PSA and the PSA > 10 ng/ml cohorts, the DCE-upgraded group showed a higher risk of csPCa compared to non-csPCa, but also a higher proportion of missed csPCa in the non-upgraded group, with no significant stratification effect between groups. When PSA ≤ 10 ng/ml, the DCE-upgraded group demonstrated a lower csPCa risk relative to non-csPCa, whereas the non-upgraded group showed a higher proportion of missed csPCa. The study conducted by Song et al. [26] demonstrated that while the detection rate of csPCa using bpMRI was lower compared to mpMRI, its diagnostic performance remained within a high range. Additionally, the significantly reduced inspection time associated with bpMRI will contribute to meeting the growing demand and promoting widespread adoption of MRI screening for PCa.
The primary limitations of bpMRI include reduced sensitivity and the inability to accurately assess PI-RADS 3 lesions, which undermines confidence in performing biopsies. Incorporating PSA-derived indicators such as PSAD and fPSA/tPSA may enhance the biopsy strategy. Several studies have demonstrated that combining bpMRI with PSAD yields superior results compared to mpMRI and significantly improves the sensitivity of bpMRI [16, 27, 28]. On the other hand, due to the uncertainty of PI-RADS 3 lesions, many urologists choose direct biopsy and find that csPCa prevalence is notably high [29]. The findings from Choi et al. [30] showed that the biopsy threshold was set at PI-RADS 3, and the csPCa detection rates of PI-RADS category 3 lesion by 2 readers using bpMRI and mpMRI were 60%, 58.3% and 65.2%, 60%, respectively. Collectively, these findings suggest that DCE may have limited essential value for stratifying PI-RADS 3 lesions, with emerging tools offering more reliable risk stratification when DCE is unavailable.
LV has been demonstrated as an effective imaging biomarker for detecting PCa and predicting histological aggressiveness, with a strong correlation between imaging evaluation and tumor histological volume [31, 32]. The proportion of csPCa in lesions with volume < 0.5 cm3 (about 2%) is comparatively lower compared to that of ciPCa, which is considered typically indolent – exhibiting minimal invasive potential and stability post-diagnosis [33]. The results of Wang et al. [34] showed that the AUC of S-PI-RADS in diagnosing csPCa was 0.906, which was equivalent to that of mp-PI-RADS (AUC = 0.919, Z = 1.145, P > 0.05). Scialpi et al. [35] discovered that in patients with multi-focal PCa, S-PI-RADS exhibited significantly superior performance and demonstrated excellent inter-observer consistency in detecting the index lesion compared to mp-PI-RADS. Our results revealed that benign lesions consistently exhibited smaller volumes than PCa lesions. Furthermore, higher ISUP grade PCa correlated significantly with larger lesion volumes, suggesting LV may provide insight into tumor heterogeneity and biological aggressiveness. In terms of LV measurement, S-PI-RADS only mentions the use of length, width, and height measurements followed by calculation using the quasi-ellipsoidal formula. However, due to the different morphological changes of the lesions, especially in irregular (not elliptical) shaped lesions, accurately obtaining the volume of the lesion can be challenging, potentially resulting in significant discrepancies between measured and actual volumes. In this study, a segmentation software was employed to completely isolate the lesion and automatically obtain its volume with high repeatability across different readers.
Several limitations of this study should be acknowledged. Firstly, as only patients confirmed by histopathology were included, our cohort’s higher csPCa prevalence may overestimate diagnostic performance. Future prospective studies should include all MRI-triggered biopsies, including PI-RADS 1–2 cases followed clinically. Secondly, DCE may still contribute to lesion detection and characterization beyond scoring; validation of automated LV measurement and evaluation of “LV + DCE” integration in decision models are necessary. Thirdly, cognitive fusion biopsies may not adequately sample lesions with tiny volume, this may limit the validity of the correlation between lesion size and csPCa rate. Subsequent works should use in-bore MRI biopsy for sub-centimeter lesions to eliminate targeting confounders. Finally, variations in b values used for DWI images between different institutions may impact the consistency in assessing DWI categories.
Conclusion
mpMRI and bpMRI are not universally interchangeable. Clinical implementation should be guided by PSA context: mpMRI may accurately avoid biopsy when PSA ≤ 10 ng/ml, while bpMRI (especially S-PI-RADS) may increase detection rate when PSA > 10ng/ml. LV is anticipated to serve as a complementary radiological biomarker at the absence of DCE.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
Not applicable.
Abbreviations
- bpMRI
Biparametric MRI
- mpMRI
Multiparametric MRI
- csPCa
Clinically significant prostate cancer
- DCE
Dynamic contrast-enhanced imaging
- PSA
Prostate-specific antigen
- LV
Lesion volume
- ROC
Receiver operating characteristic
- PZ
Peripheral zone
- TZ
Transition zone
- DWI
Diffusion weighted imaging
- AUC
Area under the curve
- ADC
Apparent diffusion coefficient
- BPH
Benign prostatic hyperplasia
- PI-RADS
Prostate Imaging Report and Data System
- RP
Radical prostatectomy
- TRUS
Transrectal ultrasound
- ISUP
International Society of Urological Pathology
- ciPCa
Clinically insignificant PCa
Author contributions
PJ: Conceptualization, Methodology, Writing - Original Draft, Visualization. ZD and KL: Data collection, Investigation. FH and YL: interpretation, Software, Investigation. LY: Data curation, Validation. GS: Supervision. LS: Supervision, Resources. XW: Writing - Review & Editing, Project administration, Funding acquisition.
Funding
This research was supported by Natural Science Foundation of Zhejiang Province under Grant No.LGF22H220006, Medical Health Science and Technology Project of Zhejiang Province under Grant 2021KY549 and 2022KY100.
Data availability
Interested users may request access to these data, where institutional approvals along with signed data use agreements and/or material transfer agreements may be needed/negotiated. Derived result data supporting the findings of this study are available upon reasonable requests.
Declarations
Ethics approval and consent to participate
This retrospective study was approval from the ethics committees of Zhejiang Cancer Hospital and exemption from the requirement for written informed consent (IRB-2024-432). All procedures in studies with human participants complied with the ethical standards of the institutional/national research committee and the 1964 Helsinki Declaration and its amendments.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Han B, Zheng R, Zeng H, et al. Cancer incidence and mortality in china, 2022. J Natl Cancer Cent. 2024;4(1). 10.1016/j.jncc.2024.01.006. [DOI] [PMC free article] [PubMed]
- 2.Stabile A, Giganti F, Rosenkrantz AB, et al. Multiparametric MRI for prostate cancer diagnosis: current status and future directions. Nat Rev Urol. 2020;17(1):41–61. 10.1038/s41585-019-0212-4. [DOI] [PubMed] [Google Scholar]
- 3.Sun Z, Wang K, Kong Z, et al. A multicenter study of artificial intelligence-aided software for detecting visible clinically significant prostate cancer on MpMRI. Insights Imaging. 2023;14(1):72. 10.1186/s13244-023-01421-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Barentsz JO, Richenberg J, Clements R, et al. ESUR prostate MR guidelines 2012. Eur Radiol. 2012;22(4):746–57. 10.1007/s00330-011-2377-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Weinreb JC, Barentsz JO, Choyke PL, et al. Eur Urol. 2016;69(1):16–40. 10.1016/j.eururo.2015.08.052. PI-RADS Prostate Imaging - Reporting and Data System: 2015, Version 2. [DOI] [PMC free article] [PubMed]
- 6.Turkbey B, Rosenkrantz AB, Haider MA, et al. Eur Urol. 2019;76(3):340–51. 10.1016/j.eururo.2019.02.033. Prostate Imaging Reporting and Data System Version 2.1: 2019 Update of Prostate Imaging Reporting and Data System Version 2. [DOI] [PubMed]
- 7.Stanzione A, Imbriaco M, Cocozza S, et al. Biparametric 3T magnetic resonance imaging for prostatic cancer detection in a biopsy-naïve patient population: a further improvement of PI-RADS v2? Eur J Radiol. 2016;85(12):2269–74. 10.1016/j.ejrad.2016.10.009. [DOI] [PubMed] [Google Scholar]
- 8.Greer MD, Shih JH, Lay N, Turkbey B, et al. Validation of the dominant sequence paradigm and role of dynamic Contrast-enhanced imaging in PI-RADS version 2. Radiology. 2017;285(3):859–69. 10.1148/radiol.2017161316. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Woo S, Suh CH, Kim SY, et al. Head-to-Head comparison between biparametric and multiparametric MRI for the diagnosis of prostate cancer: A systematic review and Meta-Analysis. AJR Am J Roentgenol. 2018;211(5):W226–41. 10.2214/AJR.18.19880. [DOI] [PubMed] [Google Scholar]
- 10.Thaiss WM, Moser S, Hepp T, et al. Head-to-head comparison of biparametric versus multiparametric MRI of the prostate before robot-assisted transperineal fusion prostate biopsy. World J Urol. 2022;40(10):2431–8. 10.1007/s00345-022-041201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Gan JM, Kikano EG, Smith DA, et al. Clinically significant prostate cancer detection after a negative prebiopsy MRI examination: comparison of biparametric versus multiparametric MRI. AJR Am J Roentgenol. 2022;218(5):859–66. 10.2214/AJR.21.26569. [DOI] [PubMed] [Google Scholar]
- 12.Gatti M, Faletti R, Calleris G, et al. Prostate cancer detection with biparametric magnetic resonance imaging (bpMRI) by readers with different experience: performance and comparison with multiparametric (mpMRI). Abdom Radiol (NY). 2019;44(5):1883–93. 10.1007/s00261-019-01934-3. [DOI] [PubMed] [Google Scholar]
- 13.Scialpi M, Martorana E, Scialpi P, et al. S-PI-RADS and PI-RRADS for biparametric MRI in the detection of prostate cancer and Post-treatment local recurrence. Anticancer Res. 2023;43(1):297–303. 10.21873/anticanres.16163. [DOI] [PubMed] [Google Scholar]
- 14.Schröder FH, Hugosson J, Roobol MJ, et al. Screening and prostate cancer mortality: results of the European randomised study of screening for prostate cancer (ERSPC) at 13 years of follow-up. Lancet. 2014;384(9959):2027–35. 10.1016/S0140-6736(14)60525-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Philip J, Manikandan R, Javlé P, et al. Prostate cancer diagnosis: should patients with prostate specific antigen > 10ng/mL have stratified prostate biopsy protocols? Cancer Detect Prev. 2009;32(4):314–8. 10.1016/j.cdp.2008.12.004. [DOI] [PubMed] [Google Scholar]
- 16.Han C, Liu S, Qin XB, et al. MRI combined with PSA density in detecting clinically significant prostate cancer in patients with PSA serum levels of 4 ~ 10ng/mL: biparametric versus multiparametric MRI. Diagn Interv Imaging. 2020;101(4):235–44. 10.1016/j.diii.2020.01.014. [DOI] [PubMed] [Google Scholar]
- 17.Kuhl CK, Bruhn R, Krämer N, et al. Abbreviated biparametric prostate MR imaging in men with elevated prostate-specific antigen. Radiology. 2017;285(2):493–505. 10.1148/radiol.2017170129. [DOI] [PubMed] [Google Scholar]
- 18.De Visschere P, Lumen N, Ost P, et al. Dynamic contrast-enhanced imaging has limited added value over T2-weighted imaging and diffusion-weighted imaging when using PI-RADSv2 for diagnosis of clinically significant prostate cancer in patients with elevated PSA. Clin Radiol. 2017;72(1):23–32. 10.1016/j.crad.2016.09.011. [DOI] [PubMed] [Google Scholar]
- 19.Epstein JI, Amin MB, Reuter VE, et al. Contemporary Gleason grading of prostatic carcinoma: an update with discussion on practical issues to implement the 2014 international society of urological pathology (ISUP) consensus conference on Gleason grading of prostatic carcinoma. Am J Surg Pathol. 2017;41(4):e1–7. 10.1097/PAS.0000000000000820. [DOI] [PubMed] [Google Scholar]
- 20.Tamada T, Kido A, Yamamoto A, et al. Comparison of biparametric and multiparametric MRI for clinically significant prostate cancer detection with PI-RADS version 2.1. J Magn Reson Imaging. 2021;53(1):283–91. 10.1002/jmri.27283. [DOI] [PubMed] [Google Scholar]
- 21.Druskin SC, Ward R, Purysko AS, et al. Dynamic contrast enhanced magnetic resonance imaging improves classification of prostate lesions: A study of pathological outcomes on targeted prostate biopsy. J Urol. 2017;198(6):1301–8. 10.1016/j.juro.2017.07.011. [DOI] [PubMed] [Google Scholar]
- 22.Xu L, Zhang G, Shi B, et al. Comparison of biparametric and multiparametric MRI in the diagnosis of prostate cancer. Cancer Imaging. 2019;19(1):90. 10.1186/s40644-019-0274-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zawaideh JP, Sala E, Shaida N, et al. Diagnostic accuracy of biparametric versus multiparametric prostate MRI: assessment of contrast benefit in clinical practice. Eur Radiol. 2020;30(7):4039–49. 10.1007/s00330-020-06782-0. [DOI] [PubMed] [Google Scholar]
- 24.Di Campli E, Delli Pizzi A, Seccia B, et al. Diagnostic accuracy of biparametric vs multiparametric MRI in clinically significant prostate cancer: comparison between readers with different experience. Eur J Radiol. 2018;101:17–23. 10.1016/j.ejrad.2018.01.028. [DOI] [PubMed] [Google Scholar]
- 25.Mahajan M, Gupta V, Gupta P, et al. Evaluation of clinically significant prostate cancer using biparametric magnetic resonance imaging: an evolving concept. J Cancer Res Ther. 2022;18(6):1640–5. 10.4103/jcrt.JCRT_1313_20. [DOI] [PubMed] [Google Scholar]
- 26.Song J, Zhao C, Zhang F, et al. The diagnostic performance in clinically significant prostate cancer with PI-RADS version 2.1: simplified BpMRI versus standard MpMRI. Abdom Radiol (NY). 2023;48(2):704–12. 10.1007/s00261-022-03750-8. [DOI] [PubMed] [Google Scholar]
- 27.Lee SJ, Oh YT, Jung DC, et al. Combined analysis of biparametric MRI and prostate-Specific antigen density: role in the prebiopsy diagnosis of Gleason score 7 or greater prostate cancer. AJR Am J Roentgenol. 2018;211(3):W166–72. 10.2214/AJR.17.19253. [DOI] [PubMed] [Google Scholar]
- 28.Kim MJ, Park SY. Biparametric magnetic resonance Imaging-Derived nomogram to detect clinically significant prostate cancer by targeted biopsy for index lesion. J Magn Reson Imaging. 2022;55(4):1226–33. 10.1002/jmri.27841. [DOI] [PubMed] [Google Scholar]
- 29.Venderink W, van Luijtelaar A, Bomers JGR, et al. Results of targeted biopsy in men with magnetic resonance imaging lesions classified equivocal, likely or highly likely to be clinically significant prostate cancer. Eur Urol. 2018;73(3):353–60. 10.1016/j.eururo.2017.02.021. [DOI] [PubMed] [Google Scholar]
- 30.Choi MH, Kim CK, Lee YJ, et al. Prebiopsy biparametric MRI for clinically significant prostate cancer detection with PI-RADS version 2: A multicenter study. AJR Am J Roentgenol. 2019;212(4):839–46. 10.2214/AJR.18.20498. [DOI] [PubMed] [Google Scholar]
- 31.Yilmaz EC, Shih JH, Belue MJ, et al. Prospective evaluation of PI-RADS version 2.1 for prostate cancer detection and investigation of multiparametric MRI-derived markers. Radiology. 2023;307(4):e221309. 10.1148/radiol.221309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Yang DD, Lee LK, Tsui JMG, et al. AI-derived tumor volume from multiparametric MRI and outcomes in localized prostate cancer. Radiology. 2024;313(1):e240041. 10.1148/radiol.240041. [DOI] [PubMed] [Google Scholar]
- 33.Liddell H, Jyoti R, Haxhimolla HZ. mp-MRI prostate characterised PIRADS 3 lesions are associated with a low risk of clinically significant prostate Cancer - A retrospective review of 92 biopsied PIRADS 3 lesions. Curr Urol. 2015;8(2):96–100. 10.1159/000365697. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Wang G, Yu G, Chen J, et al. Can high b-value 3.0 T biparametric MRI with the simplified prostate image reporting and data system (S-PI-RADS) be used in biopsy-naïve men? Clin Imaging. 2022;88:80–6. 10.1016/j.clinimag.2021.06.024. [DOI] [PubMed] [Google Scholar]
- 35.Scialpi M, Martorana E, Torre R, et al. Index lesion detection in multifocal prostate cancer: simplified PI-RADS biparametric MRI vs PI-RADS v2.1 multiparametric MRI. Clin Imaging. 2023;94:108–15. 10.1016/j.clinimag.2022.11.013. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Citations
- Weinreb JC, Barentsz JO, Choyke PL, et al. Eur Urol. 2016;69(1):16–40. 10.1016/j.eururo.2015.08.052. PI-RADS Prostate Imaging - Reporting and Data System: 2015, Version 2. [DOI] [PMC free article] [PubMed]
- Turkbey B, Rosenkrantz AB, Haider MA, et al. Eur Urol. 2019;76(3):340–51. 10.1016/j.eururo.2019.02.033. Prostate Imaging Reporting and Data System Version 2.1: 2019 Update of Prostate Imaging Reporting and Data System Version 2. [DOI] [PubMed]
Supplementary Materials
Data Availability Statement
Interested users may request access to these data, where institutional approvals along with signed data use agreements and/or material transfer agreements may be needed/negotiated. Derived result data supporting the findings of this study are available upon reasonable requests.