ABSTRACT
Objective
PI‐RADS 3 lesions on prostate multiparametric MRI (mpMRI) are indeterminate findings with debated management. We evaluated their radiologic evolution on follow‐up mpMRI and correlated imaging changes with biopsy outcomes.
Methods
We retrospectively identified 166 men with at least one PI‐RADS 3 lesion on initial mpMRI and at least one follow‐up mpMRI without intervening treatment. Lesions were classified at follow‐up as regressed (downgraded to PI‐RADS ≤ 2), stable (remained PI‐RADS 3), or progressed (upgraded to PI‐RADS 4–5). Lesions were stratified by zonal location. MRI‐targeted biopsy results were reviewed when available (n = 64).
Results
On follow‐up, 52% of lesions regressed, 29% remained stable, and 19% progressed. Progression was significantly more common for peripheral‐zone lesions (25%) than transition‐zone lesions (6%); no central‐zone lesions progressed. Of the biopsied lesions, most yielded benign or low‐grade pathology: 44% were benign, 47% were ISUP 1, 6% had high‐grade PIN or ASAP, and only 3% harbored clinically significant prostate cancer (csPCa; ISUP ≥ 2). Both csPCa cases occurred in lesions that had progressed radiologically. Lesions with imaging progression were significantly more likely to yield cancer on initial or follow‐up biopsy than stable/regressed lesions (83% vs 40%; p = 0.037).
Conclusions
In this real‐world cohort, most PI‐RADS 3 lesions remained indolent: over half regressed, and 19% progressed. Biopsied lesions had a very low rate of csPCa (3%). These results suggest that surveillance with interval MRI rather than immediate biopsy may be a safe strategy for selected men with PI‐RADS 3 lesions.
Keywords: MRI surveillance, multiparametric MRI, PI‐RADS, prostate cancer detection, prostate MRI
1. Introduction
Multiparametric MRI (mpMRI) has become central to prostate cancer evaluation, improving detection of significant tumors while reducing unnecessary biopsies [1, 2]. However, the indeterminate PI‐RADS (Prostate Imaging Reporting and Data System) 3 category remains a management dilemma. PI‐RADS 3 lesions are common (typically apparent in 20%–30% of scans) and carry a non‐negligible cancer risk, but most underlying cancers are low‐risk. For example, Schoots et al. reviewed > 8,000 cases and found 22% of diagnostic mpMRIs had a PI‐RADS 3 index lesion [1]. Validation studies indicate only about 10%–20% of PI‐RADS 3 lesions harbor cancer on biopsy, and far fewer (often < 10%) contain clinically significant cancer (ISUP ≥ 2) [1, 3, 4].
Current guidelines advocate risk‐adapted management for PI‐RADS 3 lesions. Notably, the 2024 EAU prostate cancer guidelines incorporate PSA density (PSAD) into decision‐making for PI‐RADS 3: lesions with PSAD < 0.10 ng/mL/cm3 have only approximately 4% csPCa risk, whereas PSAD > 0.15 approaches 30% csPCa risk. In practice, a low PSAD PI‐RADS 3 lesion can often be safely monitored with repeat MRI, whereas a high PSAD lesion may prompt biopsy [5, 6].
Despite these advances, there is a knowledge gap in the natural history of PI‐RADS 3 lesions which are managed conservatively. Steinkohl et al. followed 141 men with PI‐RADS 3 on initial mpMRI: 76.6% of lesions were reclassified on first follow‐up (either regressed or progressed), Upgrading to PI‐RADS 4 tended to occur around 12 months, suggesting an optimal follow‐up interval of 12 months [3]. Kwe et al. noted that patients whose lesions progressed on follow‐up had markedly higher csPCa detection rates (43% vs 9%) [7]. Nonetheless, reported progression rates vary (often 10%–30%), and previous series are limited by their cohort size or length of follow‐up. Boschheidgen et al. recently proposed that patients with PI‐RADS 3 lesions can safely defer follow‐up MRI for 24 months, since those diagnosed with cancer reliably show grade progression by 12–24 months [8].
We performed a retrospective analysis of 166 patients with PI‐RADS 3 lesions and repeat prostate mpMRI to quantify lesion evolution on serial mpMRI and to correlate changes with biopsy outcomes.
2. Materials and Methods
2.1. Study Design and Cohort
We retrospectively reviewed our institutional prostate mpMRI database (2017–2023), comprising 4,990 unique patients. Unique patients were defined as distinct individual patients counted once, regardless of the number of prostate MRI examinations performed during the study period. Inclusion criteria were: an initial mpMRI showing ≥ 1 PI‐RADS 3 lesion, plus ≥ 1 follow‐up mpMRI performed after a period of surveillance, without any definitive prostate treatment (surgery, radiation, or focal therapy) during the interval. We identified 166 such patients. If a patient had multiple distinct PI‐RADS 3 lesions on the index MRI, each lesion was recorded and tracked individually; analyses by lesion location were therefore lesion‐level, whereas overall outcomes (e.g. whether the patient ever had progression) were assessed at the patient level. Patients were excluded if the baseline MRI had any PI‐RADS ≥ 4 lesion, if no follow‐up MRI was obtained, or if treated between scans.
2.2. MRI Protocol and Lesion Classification
All MRIs were acquired on a 3 Tesla scanner with a 16‐channel phased‐array coil. The standardized prostate mpMRI protocol included multiplanar T2‐weighted sequences, axial diffusion‐weighted imaging (high b‐values up to 1400–2000 s/mm2 with ADC mapping), and dynamic contrast‐enhanced imaging after gadolinium administration. Scans were interpreted in routine clinical practice by four dedicated genitourinary radiologists, each with 10–20 years of experience in genitourinary imaging and prostate MRI interpretation. Each suspicious finding was assigned a PI‐RADS score 1–5 according to PI‐RADS v2.0/v2.1 criteria. For serial examinations, follow‐up reports were reviewed to track previously described PI‐RADS 3 lesions and determine whether the lesion regressed, remained stable, or progressed. When available, prior MRI examinations were used as comparators as part of routine radiologic interpretation. We recorded the maximum diameter of each PI‐RADS 3 lesion and categorized its anatomic location by prostate zone: peripheral zone (PZ), transition zone (TZ), or central zone (CZ). In cases where a lesion spanned zones, the dominant zone was assigned according to the radiology report.
2.3. Lesion Evolution Definitions
Follow‐up MRIs were reviewed to identify changes in the highest PI‐RADS score of each tracked lesion. We defined three categories of PI‐RADS evolution:
Regression: The lesion was downgraded to PI‐RADS 1–2 or was no longer visible on follow‐up.
Stability: The lesion remained PI‐RADS 3 on follow‐up.
Progression: The lesion was upgraded to PI‐RADS 4 or 5 on any follow‐up scan.
For patients with multiple follow‐up MRIs, the highest PI‐RADS category observed at any time during surveillance was used to define evolution. For patients with multiple baseline PI‐RADS 3 lesions, all lesions were recorded; however, the primary clinical outcome was defined at the patient level. Overall patient‐level MRI evolution was assigned according to the most clinically relevant lesion trajectory observed during follow‐up: progression was defined as upgrade of any baseline PI‐RADS 3 lesion to PI‐RADS 4–5; stability was defined as absence of progression with at least one lesion remaining PI‐RADS 3; and regression was defined as downgrade to PI‐RADS 1–2 or disappearance of all baseline PI‐RADS 3 lesions. Secondary zonal analyses were performed at the lesion level and included only baseline PI‐RADS 3 lesions with available zonal classification and evaluable follow‐up evolution.
2.4. Biopsy Data and Analysis
Clinical records were reviewed for any prostate biopsy following the initial MRI. Biopsies were performed at the treating urologist's discretion after the initial MRI or during follow‐up due to rising PSA, patient preference, or MRI changes. When done, both MRI‐targeted biopsies of the lesion(s) and systematic biopsies were obtained. At least two cores per target were sampled, plus standard systematic cores. All specimens were evaluated by genitourinary pathologists, assigning Gleason grade (ISUP Grade Group). We recorded the highest‐Grade Group per lesion; ISUP Grade ≥ 2 (Gleason ≥ 3 + 4) lesions were considered clinically significant prostate cancer (csPCa). Premalignant findings (high‐grade PIN or atypical small acinar proliferation [ASAP]) were also noted.
For lesion‐level analyses of biopsies, we included 80 lesions in 64 patients who underwent biopsy during follow‐up (some patients had more than one PI‐RADS 3 lesion biopsied). We categorized the pathology outcome for each lesion (benign, premalignant, ISUP 1, or ISUP ≥ 2). In addition, we stratified biopsy results by MRI evolution status (progressed vs. stable/regressed) and by lesion location, to explore any association between changes in imaging and lesion location and detection of any cancer or csPCa.
2.5. Kscore Subgroup Analysis
Available 4Kscore results were retrospectively collected from the medical record. The reported 4Kscore percentage and date of testing were recorded and linked to each patient's baseline mpMRI, follow‐up PI‐RADS evolution, and biopsy outcome when available. Given that 4Kscore testing was not performed systematically and was available in only 14 patients, this analysis was considered exploratory. Accordingly, results were summarized descriptively and were not included in the primary statistical comparisons or multivariable analyses.
2.6. Statistics
Descriptive statistics summarized patient age, PSA, prostate volume, PSAD, and follow‐up interval (median with interquartile range). Categorical outcomes (proportions of lesions regressing, stable, and progressing; biopsy results) were compared using chi‐square or Fisher exact tests. Continuous variables were compared with Mann‐Whitney U tests. A two‐sided p < 0.05 was considered statistically significant. All analyses were performed using SPSS v29 (IBM Corp.). This retrospective study was approved by our institutional review board with waiver of informed consent.
3. Results
3.1. Patient and Lesion Characteristics
We included 166 patients (median age 66.4 years, IQR 61.2–70.4) with at least one baseline PI‐RADS 3 lesion and follow‐up MRI. Median PSA was 5.8 ng/mL (IQR 4.42–8.0), median prostate volume was 61.8 mL (IQR 46–82.2), and median PSAD was 0.088 ng/mL/cm3 (IQR 0.063–0.132). Notably, 65% of patients had PSAD < 0.10 at baseline, and only 15% had PSAD > 0.15. The median interval to first follow‐up MRI was 1.48 years (IQR 1.07–2.29); 70% of patients had one repeat scan, while 30% underwent multiple repeat scans (range 2–5) as part of ongoing surveillance. No patient received definitive prostate cancer treatment during the follow‐up period (Table 1a, Figure 1).
Table 1a.
Baseline demographic, clinical, and imaging characteristics of the 166 patients included in the study.
| Variable (median, IQR) | Patients with PI‐RADS 3 (n = 166) |
|---|---|
| Age (years) | 66.4 [61.2, 70.4] |
| PSA (ng/mL) | 5.8 [4.4, 8.0] |
| Prostate volume (mL) | 61.8 [46.0, 82.2] |
| PSA density (ng/mL/cm3) | 0.09 [0.06, 0.132] |
| Time to repeat mpMRI (years) | 1.48 [1.07, 2.29] |
Note: Data are presented as median with interquartile range (IQR).
Abbreviations: PSA = prostate‐specific antigen; PSAD = prostate‐specific antigen density.
Figure 1.
Distribution of time intervals between the baseline mpMRI demonstrating a PI‐RADS 3 lesion and the first follow‐up mpMRI. The median interval was 1.48 years (interquartile range [IQR] 1.07–2.29 years), reflecting real‐world surveillance practices. Most patients underwent repeat imaging within approximately 1–2 years after the initial examination. [Color figure can be viewed at wileyonlinelibrary.com]
For the secondary lesion‐level zonal analysis, 180 evaluable baseline PI‐RADS 3 lesions with available zonal classification and follow‐up evolution were included. Of these, 120 lesions (66.7%) were located in the peripheral zone, 49 (27.2%) in the transition zone, and 11 (6.1%) in the central zone. The median maximal lesion diameter was 7 mm.
3.2. Lesion Evolution on Follow‐up MRI
Using the predefined patient‐level classification, radiologic regression was the most common outcome. Overall, 86/166 patients (51.8%) demonstrated regression to PI‐RADS 1–2 or lesion resolution on follow‐up MRI. Forty‐nine patients (29.5%) remained stable at PI‐RADS 3, and 31 patients (18.7%) progressed to PI‐RADS 4–5. Figure 2a,b illustrates these patient‐level outcomes.
Figure 2.
(a,b) Evolution of PI‐RADS 3 lesions on follow‐up mpMRI. Most lesions regressed or remained stable: 51.5% were downgraded to PI‐RADS 1–2, 29.3% remained PI‐RADS 3, and 18.6% progressed to PI‐RADS 4–5. [Color figure can be viewed at wileyonlinelibrary.com]
3.3. Predictors of Lesion Dynamics
In the secondary lesion‐level zonal analysis, lesion dynamics differed significantly by zonal location (χ 2 = 20.1, p < 0.001). Peripheral zone (PZ) lesions were most prone to upgrade: 30/120 PZ lesions (25.0%) progressed to PI‐RADS 4–5, compared with 3/49 transition zone (TZ) lesions (6.1%) and 0/11 central zone (CZ) lesions (0%). In the PZ, 70/120 lesions (58.3%) regressed and 20/120 (16.7%) remained PI‐RADS 3. In the TZ, 27/49 lesions (55.1%) regressed and 19/49 (38.8%) remained PI‐RADS 3. CZ lesions either remained PI‐RADS 3 (6/11, 54.5%) or regressed (5/11, 45.5%), with no observed progression. Thus, although more than half of both PZ and TZ lesions regressed, PZ lesions accounted for the highest absolute number and proportion of progressors. Baseline lesion size was not significantly different between progressed and non‐progressed lesions (mean diameter 7.5 vs 7.0 mm, p > 0.5; Figure 3). In patient‐level exploratory analyses, age, PSA, and PSA density were not significantly associated with follow‐up PI‐RADS category (Supporting Figure 1).
Figure 3.
Lesion evolution according to zonal location. Peripheral‐zone lesions demonstrated the highest progression rate, whereas no central‐zone lesions progressed. [Color figure can be viewed at wileyonlinelibrary.com]
3.4. Clinical Characteristics of the Biopsy Cohort
Among the 166 men with PI‐RADS 3 lesions who underwent serial mpMRI, biopsy pathology reports were available for 64 patients. Compared with the overall PI‐RADS 3 cohort, clinical characteristics of biopsied patients were not significantly different: median age was 65.8 vs 66.4 years, PSA 5.5 vs 5.8 ng/mL, prostate volume 58 vs 61.8 mL, and PSAD 0.097 vs 0.088 ng/mL2, (all p > 0.05; Table 1b).
Table 1b.
Comparison of baseline demographic, clinical, and imaging characteristics between patients with available prostate biopsy results and the overall PI‐RADS 3 cohort.
| Variable (median, IQR) | Patients with PI‐RADS 3 and prostate biopsy (n = 64) | Patients with PI‐RADS 3 (n = 166) | p value (Mann‐Whitney U) |
|---|---|---|---|
| Age (years) | 65.8 [61.0, 69.5] | 66.4 [61.2, 70.4] | p = 0.25 |
| PSA (ng/mL) | 5.5 [3.9, 7.9] | 5.8 [4.4, 8.0] | p = 0.35 |
| Prostate volume (mL) | 58.0 [39.3, 75.5] | 61.8 [46.0, 82.2] | p = 0.07 |
| PSA density (ng/mL/cm3) | 0.097 [0.06, 0.16] | 0.088 [0.06, 0.132] | p = 0.29 |
Note: Data are presented as median with interquartile range (IQR).
Abbreviations: PSA = prostate‐specific antigen; PSAD = prostate‐specific antigen density.
3.5. Pathology Outcomes
Biopsy results for these 64 patients are summarized in Figure 4a. Clinically significant cancer (ISUP ≥ 2) was detected in 2/64 patients (3.1%). Low‐grade prostate cancer (ISUP 1) was found in 30/64 (46.9%). Premalignant findings (high‐grade PIN or ASAP) were present in 4/64 (6.2%), and 28/64 patients (43.8%) had benign histology consistent with BPH. Both patients with ISUP ≥ 2 disease demonstrated lesion progression to PI‐RADS 4 on follow‐up MRI.
Figure 4.
Biopsy pathology results and PI‐RADS evolution by biopsy category. (a) Biopsy pathology among patients with available pathology results: 44% benign/BPH, 6% premalignant findings (high‐grade PIN/ASAP), 47% ISUP 1, and 3% ISUP ≥ 2. (b) PI‐RADS evolution according to biopsy pathology category. Lesions upgraded on MRI had a higher biopsy cancer rate than stable/regressed lesions (83% vs 40%; p < 0.05), and both csPCa cases were found only in the progressed group. [Color figure can be viewed at wileyonlinelibrary.com]
3.6. MRI Progression versus Biopsy Results
To explore the relationship between imaging changes and the presence of cancer, we compared biopsy results stratified by MRI evolution status. Of the 31 lesions that progressed to PI‐RADS 4–5 on follow‐up, 17 were biopsied; 13 of these (76%) showed cancer (mostly ISUP 1). In contrast, among the 135 lesions that remained PI‐RADS 3 or regressed, 47 were biopsied and only 18 (38%) had any cancer (p = 0.037). Importantly, both csPCa cases were in the group that progressed. None of the stable/regressed lesions that were biopsied harbored Gleason ≥ 3 + 4 (all were benign or ISUP 1). Thus, patients with an upgrade to PI‐RADS 4–5 on surveillance MRI were significantly more prone to cancer detection (Figure 4b).
3.7. Exploratory 4Kscore Analysis
4Kscore results were available for 14 patients. The median 4Kscore was 7.5% (range 1.0–23.0). Lesion evolution in this subgroup included regression in eight patients, stability in 3, and progression in 3. Median 4Kscore values were 5.0%, 21.2%, and 7.0%, respectively. Biopsy pathology was available in six patients and demonstrated benign/BPH findings in 3, ISUP Grade Group 1 disease in 2, and ISUP ≥ 2 disease in 1. The single ISUP ≥ 2 case had a 4Kscore of 4.0% and demonstrated radiologic progression on follow‐up mpMRI. Given the limited sample size and non‐systematic use of 4Kscore testing, these findings were considered exploratory, and no consistent association between 4Kscore and lesion evolution or biopsy outcome was observed.
4. Discussion
4.1. Lesion Evolution and Zonal Patterns
In the current cohort of men with PI‐RADS 3 index lesions followed with serial mpMRI, more than half of lesions regressed over time (52%), approximately one‐third remained stable (29%), and only 19% progressed to a higher PI‐RADS category (4–5). This pattern mirrors prior reports that most PI‐RADS 3 findings are ultimately low risk: for example, one series found 55% of PI‐RADS 3 lesions remained stable over 2 years with only a few progressing to csPCa. Importantly, lesion behavior varied by anatomic zone [9]. Progression was most common in peripheral zone (PZ) lesions (25%), much less so in transition zone (TZ) lesions (6%), and no progression was seen in central zone (CZ) lesions. This zonal gradient may be related to previous findings regarding prostate cancer biology: cancers arising in the TZ or CZ tend to have lower Gleason grade and better outcomes than those in the PZ [9, 10]. The zonal differences observed may suggest that PI‐RADS 3 lesion in the PZ should prompt closer surveillance, whereas similar lesions in the TZ or central gland may be at minimal risk, thus requiring less stringent follow‐up.
4.2. Biopsy Outcomes and Implications
We identified significant cancer in 3% of men and it was only found in the progressive subgroup. Kwe et al. recently confirmed that imaging progression is the strongest predictor of csPCa: in their cohort, 23% of men showed PI‐RADS upgrading on follow‐up MRI, and those with upgrading had a csPCa rate of 43% versus only 9% in those without upgrading. They concluded that stable PI‐RADS 3 lesions may not require biopsy, whereas progression (upgrade to ≥ 4) warrants prompt re‐biopsy. Our series supports this strategy: all csPCa cases emerged in lesions that “upgraded” on MRI. Similarly, the group emphasized that repeat MRI in 6–12 months can safely triage patients if the lesion remains unchanged or regresses, biopsy is unnecessary, but progression indicates biopsy [7].
In the current cohort, the classic risk factors PSA, prostate volume and PSAD were not associated with the PI‐RADS dynamics, suggesting they are independent predictors of outcome. Published nomograms and risk tables integrate PSAD with PI‐RADS: a large series found PSAD, age, and prior biopsy status to be independent predictors of csPCa in PI‐RADS 3 men [11, 12]. In these men a PSAD cutoff of 0.15 ng/mL/cc could avoid most biopsies at the cost of missing only 6.5% of csPCa, echoing the concept that low‐risk patients can be observed [11, 12].
Current guidelines support a risk‐adapted paradigm when deciding which men should undergo biopsy. The EAU Prostate Cancer guidelines emphasize combining MRI findings with clinical parameters to guide biopsy decisions by using a risk‐matrix linking PI‐RADS categories and PSAD tiers to stratify csPCa risk. PI‐RADS 3 lesions with PSAD < 0.15 carry only a 4% csPCa risk [13, 14], whereas a higher PSAD is associated with a higher csPCa risk which may justify a prostate biopsy. The AUA/SUO early‐detection guideline similarly defines MRI PI‐RADS ≥ 3 as “abnormal” and recommends targeted biopsy in context, but also allows adjunctive tools (biomarkers, clinical calculators) to refine the need [15]. This framework aligns with our findings: most men with stable PI‐RADS 3 lesions (often with low PSAD) can defer biopsy, while those with imaging progression or high PSAD proceed to biopsy.
In practice, this risk‐adapted strategy means most men with PI‐RADS 3 lesions may benefit from close imaging follow‐up rather than immediate biopsy. If follow‐up MRI remains PI‐RADS 3 or regresses (especially in a TZ lesion or low PSA‐D setting), surveillance may be continued [1]. Conversely, any PI‐RADS 3 lesion that progresses on repeat MRI warrants a prostate biopsy since all csPCa were identified among patients that progressed [7]. Thus, our real‐world data suggests that guideline‐recommended risk stratification (PSAD, zone, MRI dynamics) can safely reduce overtreatment without missing significant cancers.
Blood‐based biomarkers such as the 4Kscore may provide additional risk stratification in men with equivocal prostate MRI findings, particularly when deciding whether PI‐RADS 3 lesions should undergo biopsy or imaging surveillance. In our cohort, 4Kscore data were available only in a small exploratory subgroup of 14 patients. Within this limited subset, 4Kscore values did not show a consistent association with subsequent MRI evolution or biopsy outcome. Notably, the single patient with ISUP ≥ 2 disease in this subgroup had a low 4Kscore but demonstrated radiologic progression on follow‐up mpMRI. These findings should not be interpreted as evidence against the clinical utility of 4Kscore, given the small sample size and non‐systematic use of the test in routine practice. Rather, they emphasize that in this cohort, MRI dynamics—particularly lesion upgrade to PI‐RADS 4–5—remained the most informative marker of clinically significant disease. Larger prospective studies integrating serial mpMRI, PSA density, zonal location, and blood‐based biomarkers are needed to determine whether combined models can further refine biopsy selection among men with PI‐RADS 3 lesions.
Key strengths of this study include its relatively large, well‐characterized cohort of 166 patients—among the largest to date focusing exclusively on PI‐RADS 3 lesions—and the real‐world nature of the data, derived from routine clinical care. All mpMRI studies were interpreted and followed longitudinally by dedicated genitourinary (GU) radiologists with subspecialty expertise, ensuring consistency and accuracy in lesion identification and PI‐RADS scoring. Lesions were systematically tracked across serial scans, allowing robust assessment of temporal evolution and correlation with histopathology when available.
Limitations include the retrospective study design, which may introduce selection bias. Some patients with higher‐risk features may have undergone early biopsy without serial imaging and thus were excluded from this analysis, potentially leading to an underestimation of progression and csPCa rates. Consequently, the present cohort may represent a somewhat lower‐risk subset of PI‐RADS 3 lesions managed with surveillance MRI in routine clinical practice.
An additional limitation relates to mpMRI quality and longitudinal reproducibility. Although all studies were performed using a standardized institutional prostate MRI protocol and interpreted by dedicated genitourinary radiologists, formal PI‐QUAL scoring was not available for all examinations. Therefore, we could not directly assess whether apparent PI‐RADS regression, stability, or progression was influenced by differences in image quality, artifacts, acquisition parameters, or inter‐scan technical variability rather than true biological change. This is particularly relevant for PI‐RADS 3 lesions, where subtle changes in diffusion restriction, T2 appearance, or enhancement may alter categorization. Recent prospective data by Cannoletta et al. [16] demonstrated that mpMRI quality, assessed by PI‐QUAL, significantly affects diagnostic accuracy and reproducibility of prostate MRI interpretation, emphasizing the importance of MRI quality control when interpreting serial prostate MRI studies.
Histopathologic verification was not available for all patients or lesions, introducing verification bias. Biopsy was performed selectively according to clinical judgment, PSA dynamics, patient preference, and/or MRI changes rather than by protocol. As a result, stable or regressed lesions managed with imaging surveillance alone may have included undetected cancer, although no clinically evident progression was observed during follow‐up. Conversely, progressed lesions were more likely to undergo biopsy, which may have enriched the progressed group for cancer detection. Therefore, biopsy‐related findings should be interpreted as hypothesis‐generating and reflective of real‐world clinical decision‐making rather than as a uniform reference‐standard analysis.
While the cohort size is substantial, the absolute number of lesions that progressed or yielded csPCa remains modest, limiting the power for subgroup analyses, especially by anatomic zone. The follow‐up duration may have missed very late progressions. Nevertheless, the consistent association between MRI upgrade and cancer detection supports the clinical relevance of serial mpMRI evolution in selected men with PI‐RADS 3 lesions.
5. Conclusion
PI‐RADS 3 lesions exhibit heterogeneous behavior: nearly half regress or resolve, while only a minority progress, primarily in the peripheral zone. The detection of clinically significant cancer was exceedingly low (3%) and confined to lesions that progressed on imaging. These findings suggest that a surveillance strategy may be reasonable for selected PI‐RADS 3 lesions and that serial mpMRI and PSA‐based risk assessment may safely obviate biopsy in many men. This real‐world experience aligns with EAU/AUA guidance advocating a risk‐adapted approach integrating PI‐RADS score, PSAD, and zonal location. By applying this strategy, we balance early detection of the few aggressive cancers against avoidance of unnecessary biopsies and overdiagnosis. Future work should continue to refine thresholds (e.g. optimal MRI interval, PSAD cutoffs) and validate these results prospectively.
Author Contributions
Concept and design: Tomer Bashi, Roy Mano, Snir Dekalo. Acquisition of data: Tomer Bashi, Yanai Rubinstein, Adi Kidron, Roy Mano, Yanai Rubinstein, Meitar Atias, Sophie Barnes. Analysis and interpretation of data: Tomer Bashi, Roy Mano, Avi Beri. Drafting of the manuscript: Tomer Bashi. Critical revision of the manuscript for important intellectual content: Roy Mano, Snir Dekalo, Ofer Yossepowitch, Mario Sofer. Statistical analysis: Tomer Bashi. Supervision: Roy Mano, Snir Dekalo, Ofer Yossepowitch.
Funding
The authors have nothing to report.
Ethics Statement
This retrospective study was approved by the Institutional Review Board of Tel Aviv Sourasky Medical Center. The requirement for informed consent was waived due to the retrospective design of the study.
Consent
The requirement for patient informed consent was waived by the Institutional Review Board due to the retrospective design of the study and use of de‐identified data.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Figure S1: Distribution of patient age, PSA, PSA density, and baseline ROI size according to PI‐RADS score on follow‐up MRI. No significant correlation was observed.
Data Availability Statement
The datasets generated and analyzed during the current study are not publicly available due to institutional privacy regulations but are available from the corresponding author upon reasonable request.
References
- 1. Schoots I. G., “MRI in Early Prostate Cancer Detection: How to Manage Indeterminate or Equivocal PI‐RADS 3 Lesions?,” Translational Andrology and Urology 7, no. 1 (2018): 70–82, 10.21037/tau.2017.12.31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Radtke J. P., Wiesenfarth M., Kesch C., et al., “Combined Clinical Parameters and Multiparametric Magnetic Resonance Imaging for Advanced Risk Modeling of Prostate Cancer‐Patient‐Tailored Risk Stratification Can Reduce Unnecessary Biopsies,” European Urology 72, no. 6 (2017): 888–896, 10.1016/j.eururo.2017.03.039. [DOI] [PubMed] [Google Scholar]
- 3. Steinkohl F., Gruber L., Bektic J., et al., “Retrospective Analysis of the Development of PIRADS 3 Lesions over Time: When Is a Follow‐Up MRI Reasonable?,” World Journal of Urology 36, no. 3 (2018): 367–373, 10.1007/s00345-017-2135-0. [DOI] [PubMed] [Google Scholar]
- 4. Junker D., Schäfer G., Edlinger M., et al., “Evaluation of the PI‐RADS Scoring System for Classifying mpMRI Findings in Men With Suspicion of Prostate Cancer,” BioMed Research International 2013 (2013): 252939, 10.1155/2013/252939. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Sigle A., Borkowetz A., von Hardenberg J., et al., “Prediction of Significant Prostate Cancer in Equivocal Magnetic Resonance Imaging Lesions: A High‐Volume International Multicenter Study,” European Urology Focus 9, no. 4 (2023): 606–613, 10.1016/j.euf.2023.01.020. [DOI] [PubMed] [Google Scholar]
- 6. Eldred‐Evans D., Tam H., Sokhi H., et al., “An Evaluation of Screening Pathways Using a Combination of Magnetic Resonance Imaging and Prostate‐Specific Antigen: Results From the IP1‐PROSTAGRAM Study,” European Urology Oncology 6, no. 3 (2023): 295–302, 10.1016/j.euo.2023.03.009. [DOI] [PubMed] [Google Scholar]
- 7. Kwe J., Baunacke M., Boehm K., Platzek I., Thomas C., and Borkowetz A., “PI‐RADS Upgrading as the Strongest Predictor for the Presence of Clinically Significant Prostate Cancer in Patients With Initial PI‐RADS‐3 Lesions,” World Journal of Urology 42, no. 1 (2024): 84. 10.1007/s00345-024-04776-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Boschheidgen M., Schimmöller L., Doerfler S., et al., “Single Center Analysis of an Advisable Control Interval for Follow‐up of Patients With PI‐RADS Category 3 in Multiparametric MRI of The Prostate,” Scientific Reports 12, no. 1 (2022): 6746. 10.1038/s41598-022-10859-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Greene D. R., Wheeler T. M., Egawa S., Dunn J. K., and Scardino P. T., “A Comparison of the Morphological Features of Cancer Arising in The Transition Zone and in The Peripheral Zone of The Prostate,” Journal of Urology 146, no. 4 (1991): 1069–1076, 10.1016/s0022-5347(17)38003-5. [DOI] [PubMed] [Google Scholar]
- 10. King C. R., Ferrari M., and Brooks J. D., “Prognostic Significance of Prostate Cancer Originating From The Transition Zone,” Urologic Oncology: Seminars and Original Investigations 27, no. 6 (2009): 592–597, 10.1016/j.urolonc.2008.05.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Grivas N., Lardas M., Espinós E. L., et al., “Prostate Cancer Detection Percentages of Repeat Biopsy in Patients With Positive Multiparametric Magnetic Resonance Imaging (Prostate Imaging Reporting and Data System/Likert 3–5) and Negative Initial Biopsy. A Mini Systematic Review,” European Urology 82, no. 5 (2022): 452–457, 10.1016/j.eururo.2022.07.025. [DOI] [PubMed] [Google Scholar]
- 12. Bittencourt L. K., Guricova K., Zucker I., Durieux J. C., and Schoots I. G., “Risk‐Based MRI‐Directed Diagnostic Pathway Outperforms Non‐Risk‐Based Pathways in Suspected Prostate Cancer Biopsy‐Naïve Men: A Large Cohort Validation Study,” European Radiology 32, no. 4 (2022): 2330–2339, 10.1007/s00330-021-08407-6. [DOI] [PubMed] [Google Scholar]
- 13. Kortenbach K. C., Løgager V., Thomsen H. S., and Boesen L., “Early Experience in Avoiding Biopsies for Biopsy‐Naïve Men with Clinical Suspicion of Prostate Cancer but Non‐Suspicious Biparametric Magnetic Resonance Imaging Results and Prostate‐Specific Antigen Density & amp;lgt; 0.15 ng/mL(2): A 2‐year Follow‐up Study,” Acta Radiologica Open 11, no. 4 (2022): 20584601221094825, 10.1177/20584601221094825. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Konishi T., Washino S., Okochi T., and Miyagawa T., “Combination of Biparametric Magnetic Resonance Imaging with Prostate‐Specific Antigen Density to Stratify the Risk of Significant Prostate Cancer: Initial Biopsy and Long‐Term Follow‐up Results,” International Journal of Urology 29, no. 9 (2022): 1031–1037, 10.1111/iju.14948. [DOI] [PubMed] [Google Scholar]
- 15. https://www.auanet.org/guidelines-and-quality/guidelines/early-detection-of-prostate-cancer-guideline).
- 16. Cannoletta D., Brembilla G., Longoni M., et al., “Impact of Prostate mpMRI Quality on Diagnostic Accuracy in Prostate Cancer: A Prospective Study,” BJU International 8 (2026): 70373. 10.1111/bju.70373. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1: Distribution of patient age, PSA, PSA density, and baseline ROI size according to PI‐RADS score on follow‐up MRI. No significant correlation was observed.
Data Availability Statement
The datasets generated and analyzed during the current study are not publicly available due to institutional privacy regulations but are available from the corresponding author upon reasonable request.