PSA-density, DRE, and PI-RADS 5: potential surrogates for omitting biopsy?

Objective In contrast to other malignancies, histologic confirmation prior treatment in patients with a high suspicion of clinically significant prostate cancer (csPCA) is common. To analyze the impact of extracapsular extension (ECE), cT-stage defined by digital rectal examination (DRE), and PSA-density (PSA-D) on detection of csPCA in patients with at least one PI-RADS 5 lesion (hereinafter, “PI-RADS 5 patients”). Materials and methods PI-RADS 5 patients who underwent MRI/Ultrasound fusion biopsy (Bx) between 2016 and 2020 were identified in our institutional database. Uni- and multivariable logistic-regression models were used to identify predictors of csPCA-detection (GGG ≥ 2). Risk models were adjusted for ECE, PSA-D, and cT-stage. Corresponding Receiver Operating Characteristic (ROC) curves and areas under the curve (AUC) were calculated. Results Among 493 consecutive PI-RADS 5 patients, the median age and PSA was 69 years (IQR 63–74) and 8.9 ng/ml (IQR 6.0–13.7), respectively. CsPCA (GGG ≥ 2) was detected in 405/493 (82%); 36/493 patients (7%) had no cancer. When tabulating for PSA-D of > 0.2 ng/ml/cc and > 0.5 ng/ml/cc, csPCA was found in 228/253 (90%, PI-RADS5 + PSA-D > 0.2 ng/ml/cc) and 54/54 (100%, PI-RADS5 + PSA-D > 0.5 ng/ml/cc). Finally, a model incorporating PSA-D and cT-stage achieved an AUC of 0.79 (CI 0.74–0.83). Conclusion In PI-RADS 5 patients, PSA-D and cT-stage emerged as strong predictors of csPCA at biopsy. Moreover, when adding the threshold of PSA-D > 0,5 ng/ml/cc, all PI-RADS 5 patients were diagnosed with csPCA. Therefore, straight treatment for PCA can be considered, especially if risk-factors for biopsy-related complications such as obligatory dual platelet inhibition are present. Supplementary Information The online version contains supplementary material available at 10.1007/s00345-024-04894-6.


Introduction
Some patients with a high suspicion of prostate cancer (PCA) wish to skip prostate biopsy and directly initiate PCAtreatment such as androgen deprivation therapy (ADT) or radical prostatectomy (RP).This strategy of imaging-based treatment is a common practice in other malignancies such as renal cell carcinoma [1].Besides patients´ preferences, a subset of patients with a high suspicion of PCA is also at high-risk for biopsy-related complications such as patients with obligatory dual platelet inhibition or hematological conditions.In these cases, the added value of biopsy prior to treatment compared with biopsy-risks appear marginal.
Interestingly, Meissner et al. recently reported a small series of patients without prostate biopsy before undergoing RP.In this study, all patients harbored a PI-RADS score ≥ 4 and a PSMA-PET score ≥ 4 (on a five-point Likert scale) [2].In accordance with the PRIMARY-trial [3], the authors concluded, that highly selected men with positive MRI and intense PSMA-uptake may avoid confirmatory biopsy, but prospective research is warranted.However, PSMA-PET/ CT is not recommended for the primary diagnosis of PCA in current guidelines [4][5][6], related to high costs and is not widely available.
In contrast to PSMA-PET/CT, digital rectal examination (DRE), prostate-specific antigen (PSA), and multiparametric magnetic resonance imaging (mpMRI) are available before MRI/Ultrasound fusion prostate biopsy (Bx).While MRI has a high negative predictive value for clinically significant PCA (csPCA, GGG ≥ 2) [4], within the PRECISION trial [7], 6% of patients with PI-RADS 5 lesions harbored no PCA at biopsy.These findings are consistent with those of other trials [8].Therefore, mpMRI alone appears to be insufficient for omitting biopsies.
Therefore, we analyzed in a subgroup of patients with at least one PI-RADS 5 lesion (hereinafter, named "PI-RADS 5 patients"), whether the prediction of csPCA may be easily enhanced adding base clinical parameters, such as DRE and PSA-D.Finally, we tested if biopsies could be omitted in selected cases based on these findings.

Patients
Within our institutional database, we identified patients who had undergone MRI/Ultrasound fusion prostate biopsy (Bx) for suspected PCA between 2016 and 2020.In general, prostate biopsy was recommended according to national and international guidelines [4,6].The indication for mpMRI was based on clinical suspicion of PCA and was initiated by the referring physician.Targeted imaging-guided biopsy was performed in patients with a PI-RADS score ≥ 3 [12].Based on high clinical suspicion of PCA, also PI-RADS 2 lesions were targeted in selected patients.Patients with a previous positive biopsy, prior prostate surgery or pelvic radiation therapy, and those receiving hormone therapy or 5-alpha-reductase inhibitors were excluded from the analysis (suppl.Figure 2: Consort flow diagram).The clinical T-stage, defined by the findings at DRE, was prospectively assigned by the attending urologist according to the TNMsystem [4].All patients provided informed consent for the procedure and retrospective data analysis.

mpMRI protocol
If performed in-house, 3.0 Tesla MRI was conducted using an Ingenia system (Philips Medical System, Best, The Netherlands) with a phased-array coil, according to the common European Society of Urogenital Radiology [13] and PI-RADS v2 guideline recommendations [12,14].Imaging was analyzed by a dedicated uro-radiologist with > 10 years of experience in prostate-MRI [12].Patients were assigned the highest PI-RADS scores.I.e.patients with at least one PI-RADS 5 lesion were classified as "PI-RADS 5 patients".Extracapsular extension (ECE) was defined as overt ECE, capsular bulging or irregularity, broad capsular contact (> 1 cm), filling of the retro-prostatic angle, and asymmetry or invasion of the neurovascular bundles.If external imaging was available, second-reading by the aforementioned uroradiologist was performed prior to biopsy.
A minimum of two TBx cores were obtained from each lesion, depending on its location and size.After TBx was completed, a systematic biopsy (SBx) of the residual prostate was performed to achieve whole gland coverage including the complete apex [15].All cores were individually sampled, documented, and analyzed.

Outcomes of interest
The primary outcome of interest was the detection of csPCA in PI-RADS 5 patients.csPCA was defined as any prostate cancer with a Gleason Grade Group (GGG) ≥ 2. Patients' baseline characteristics (age, number of previous prostate biopsies, PSA level, DRE, prostate volume), biopsy parameters (number of biopsy cores taken, number of positive cores according to histopathology, detection of csPCA), and mpMRI results (maximum PI-RADS score, suspicion of ECE) were analyzed.To calculate the PSA-density (PSA-D), the serum PSA level was divided by the prostate volume measured by mpMRI, using the ellipsoid formula.All data were prospectively stored in an institutional database (FileMaker Pro 10; FileMaker Inc., Santa Clara, USA).

Statistical analysis
Descriptive statistics included frequencies and proportions for categorical variables.Medians and interquartile ranges (IQR) were reported for continuously-coded variables.We relied on univariable and multivariable logistic-regression model analyses to test significant predictors for csPCA at biopsy and to build csPCA-predictive MRI-based risk-models in the subgroup of PI-RADS 5 patients.The covariates for adjustment were predefined and consisted of ECE (yes vs. no), PSA-D (continuously coded), and cT-stage (cT1 vs. cT2).Receiver Operating Characteristic (ROC) curves were drawn for the models, and the corresponding areas under the curve (AUC) were compared using the DeLong-test.All tests were two-sided, and the significancelevel was set at p < 0.05.The R software environment for statistical computing and graphics (version 3.4.3,R Foundation for Statistical Computing) was used for all statistical analyses.

Discussion
The added value of MRI and consecutive fusion biopsies for the diagnosis of csPCA has been demonstrated in large trials [7,8,16].Therefore, our study on the diagnostic influence of easy and inexpensive factors on the detection of csPCA has several important findings.
First, PSA-density and DRE remain cornerstones of riskassessment, especially in patients with high suspicion of PCA on mpMRI.This finding is in accordance with previous studies [9] and guideline recommendations [4].Specifically, the link between PSA-density and csPCA has been reported before [10,17].While it is known that the addition of PSA-D to PI-RADS improves the overall predictive performance [18,19], we validated this in a high-risk cohort of PI-RADS 5 patients.
Second, DRE is not devoid of limitations, and is prone to interobserver variability and errors [20].However, at a highvolume center and in patients with highly suspect lesions, it achieved significant and clinically relevant predictor status (OR: 5.86, p < 0.001).For instance, in the PI-RADS 5 patients with a PSA-D between 0.2 and 0.5 ng/ml/cc, a positive DRE improved csPCA-detection from 175/200 (88%) to 57/59 (97%) (Table 2).These results confirm those of previous studies, such as the ERSPC-trial [11].Therefore, we observed an impressive congruence of highly suspicious  findings in palpation, suspect laboratory and imaging results for PCA detection.In other words, if one can feel, see, and measure a lesion suspected for cancer properly in the lab, it will be most likely a malignancy.This research supports the biopsy-free strategy proposed by Meissner et al. for patients with high PSA in small prostates (e.g., 25 cc prostate volume with PSA of 12,5 ng/ ml), even without PSMA-PET/CT [2].Specifically, 54/54 patients with PI-RADS 5 and PSA-D > 0.5 ng/ml/cc had csPCA.In general, the potential harm of biopsies is small and potentially neglectable [21][22][23].Especially patients, for whom radical prostatectomy is the preferred treatment option, are rarely unable to undergo biopsy at adequate risk.The effect of upstaging at RP due to treatment delay is in fusion biopsy rather irrelevant [24].Biopsy results are indispensable for precise treatment planning, such as the necessity of lymphadenectomy and choosing staging/ treatment modalities.While the combination of PI-RADS 5 and PSA-D > 0.5 ng/ml/cc did predict csPCA very accurately, these patients presented with a heterogeneous spectrum of Gleason grade groups.For instance, GGG 2 and GGG 5 were present in 24/54 (44%) and 10/54 (19%) patients, respectively (Table 2).In this light, the biopsy-omitting approach should be considered mainly for well-informed, locally advanced high-risk PCA patients, accepting the risk of suboptimal and often over-intensified treatment.In surgery, this might be the sacrifice of any primary nerve-sparing approach and an unnecessarily high rate of lymphadenectomies.In radiotherapy, adjuvant ADT duration may be prolonged excessively due to the uncertainties without biopsy (i.e., up to 3 years for high-risk versus 6 months for intermediate-risk patients) [4,5].However, direct initiation of ADT for instance appears reasonable in patients with a double high-risk constellation based on findings from the field of urology and other comorbidities, i.e. high-risk for csPCA and high-risk for biopsy-related complications.Age is a known risk factor for csPCA [25] and upstaging at RP [26] as well as for comorbidities such as cardio-vascular disease [27] with,e.g., the necessity for dual platelet inhibition.Therefore, especially in this cohort of patients the added value of a biopsy-vs.imaging-guided approach appears marginal.
The potential ability to skip biopsies in (double) highrisk constellations contrasts with the heterogeneity in the literature for omitting biopsies in PI-RADS 3 patients with a certain PSA-D cutoff [28].Sensitivity of mpMRI as well as PSA-D increase with higher-grade, larger cancers [29,30].Therefore, strategies to omit biopsies based on these parameters are more reasonable on the high-suspicion end of the scale.
Third, diligent work and review of cases are of utmost importance.The simple combination of PSA-D and DRE had an AUC of 0.79 in PI-RADS 5 patients.Only 7 of 493 (1%) patients with PI-RADS 5 had normal prostatic tissue, while most negative biopsies showed some type of inflammation.
Despite its strengths, our study is not devoid of limitations.First, specimens from radical prostatectomy and long-term oncologic follow-up of negative biopsies were not available as reference standards.Second, as it is the very nature of retrospective analysis of a tertiary care referral center, the potential different levels of selection bias should not be underestimated.Furthermore, this explorative study reported data from highly experienced biopsy-operators, genitourinary radiologists, and pathologists.Especially for PSA-D calculations with limited standardization of prostate volume measurement, and heterogeneity in MRI reading and measurement within different institutions, our data may not be representative of different settings.Through advances in biopsy technologies, such as the transperineal approach, biopsies can be offered to almost all patients who are eligible for active treatment.This has and will decrease the incentive to omit the biopsy, making our research less relevant.In general, we do not propose abolishing the biopsy in "clear cases" per se, but offer the patients who wish to avoid the diagnostic delay, discomfort, and potential complications of prostate biopsy the opportunity to choose themselves.

Conclusions
PSA-D and DRE were strong predictors of csPCA at biopsy in PI-RADS 5 patients.Therefore, these patients who have received a pathology report negative for PCA should be followed-up carefully.Conversely, biopsies may be omitted prior treatment in well-informed patients with a high suspicion of locally advanced high-risk PCA, especially if risk-factors for biopsy-related complications such as obligatory dual platelet inhibition are present.These patients, however, must accept the risk of suboptimal or overtreatment by this biopsy-free approach.

Table 2
Detection rates of prostate cancer within different subgroups and stratified according to Gleason Grade Groups (GGG)