Negative multiparametric magnetic resonance imaging for prostate cancer: further outcome and consequences

Since the introduction and widespread use of MRI targeted prostate biopsies, detection rates increased over the past years [1]. Thus, EAU-guidelines on prostate cancer recommend to perform multiparametric MRI (mpMRI) prior to prostate biopsy [2]. But there remains great dispute over biopsy necessity in MRI-negative patients with clinical suspicion for PCa. On one hand, patients and clinicians strongly desire diagnostic clarification which mostly leads to a systematic prostate biopsy when MRI results show negative. On the other hand, there is fear on both sides of overevaluation and overdiagnosis which leads to unnecessary biopsies or overtreatment of non-clinically significant PCa (ISUP = 1). Therefore, there is a strong demand for additional prognostic factors in MRI-negative patients with suspicion of PCa to take a more differentiated decision toward prostate biopsy. Alongside total PSA levels, PSA density, free/total PSA ratio and PSA doubling time are additional parameters typically used for PCa management and surveillance. But there are no sufficient data that recommend either of those factors for diagnostic use prior to prostate biopsy [3‐6], due to multiple variations (such as prostate volume, different intervals between PSA determination, variables in PSA measurement, instability of PSA at different temperatures). In addition, these factors were not evaluated in patients with negative MRI. Evaluation of blood- or urine-based biomarkers is still in an experimental phase and not widely accessible. Therefore, further investigation of patients with negative mpMRI and clinical suspicion for PCa is needed. In this study, patients’ outcomes after initial negative mpMRI were collected over 3 years, to investigate undetected clinically significant PCa (csPCa) and subsequent therapy, such as prostate cancer treatment, deobstruction therapy and further diagnostic evaluation via MRI.

With prostate cancer being the most common malign tumor disease in men, there is great demand for more diagnostic certainty in patients with negative mpMRI to avoid over diagnostic. At our medical center, 15.65% (216 eligible patients out of 1380 overall) between 2018 and 2020 showed negative mpMRI, from which 75% received prostate biopsy with only 5.1% of them showing clinically significant prostate cancer. Previous studies showed negative predictive values (NPV) of mpMRI, ranging from 76 to 99% [7‐23]. Thus, the negative prognostic value of mpMRI at our institute is sufficient and congruent to other data. But the remaining uncertainty on both sides (patient and clinician) leads to anxiety and subsequently to overevaluation and unnecessary biopsies. Clinical and laboratory factors such as family history, digital rectal examination, total PSA, PSA density, PSA velocity, Biomarkers (e.g., PCA3, PHI, 4Kscore) and BRCA mutation analysis can support the decision making progress regarding prostate cancer evaluation with prostate biopsy and help in the argumentation against overtreatment. Unfortunately, most of these factors are still being investigated and partly show incongruent data (in case of total PSA, age or prostate volume) [8, 10, 15, 21, 24, 25].

Due to the wide accessibility, PSA density is one of the most investigated predictive factors in addition to mpMRI and can elevate the NPV of mpMRI significantly [17, 24]. Schoots et al. developed a guidance tool for biopsy decision making in biopsy-naïve patients dependent on PSA Density [26]. They recommend not performing biopsy if PSA Density is < 0.20 ng/ml/cm³ and PI-RADS Score is 1 or 2 in low and intermediate risk category [26]. Our data show a significant positive correlation (OR 1624, CI [15.9; 165689]), p = 0.001) in patients with higher PSA density for the occurrence of PCa (Table 3). Interestingly, prostate volume and total PSA level prior to mpMRI showed to be independent predictive factors for PCa (Table 3). Therefore, patients with smaller prostate volume and higher PSA showed PCa significantly more often. As PSA density is calculated from total PSA in relation to prostate volume, it was surprising to see that PSA density was not an independent predictive factor. Interestingly total PSA was slightly higher (6.55 ng/ml) in non-PCa patients (in comparison with 6.4 ng/ml) (Table 3). This effect might be caused by patient selection, as only patients with elevated PSA-level received MRI evaluation. Free PSA also showed to be an additional independent predictive factor (Table 3). Free PSA usually is related with bound PSA to create the f/t-PSA-quotient. The lower the free PSA, the higher the probability of PCa. In this study, free PSA level was significantly lower (1.08 ng/ml) in patients with PCa than in non-PCa (1.47 ng/ml) and therefore, congruent to this relation.

Age is a commonly known risk factor for PCa. Consistently to that age showed to be an independent predictive factor for PCa as well (67 versus 66 years) (Table 3). Considering decision making for prostate biopsy, only age could be identified as an independent predictive factor (Table 2). Total PSA prior to mpMRI and PSA density showed significant predicational value in univariate analysis, which, however, could not be reproduced in multivariate analysis (Table 2). Investigation of secondary findings described in the initial mpMRI showed that patients who were subsequently diagnosed with PCa showed signs of prostatitis significantly more often than men without (41.2 versus 17.6%; OR 3.26, CI [1.47; 7.17], p = 0.004) (Table 3). Other secondary findings considered (BPH, post-infectious findings and calcifications) did not turn out to be statistically significant risk factors. It must be noted that only the presence or absence of an incidental finding was recorded in the initial mpMRI as part of the data collection. Whether the secondary findings corresponded in their localization to a subsequently diagnosed prostate carcinoma was not further examined. Other studies found prostatitis to be a risk factor for undetected prostate cancer [24, 27, 28]. However, the current data are not sufficient to make a general recommendation for a systematic biopsy in the case of a negative mpMRI with signs of prostatitis.

These data may reflect the existing uncertainty of treating physicians to decide against an unnecessary biopsy. However, considering the data from Table 3, the decision against a biopsy could be made easier due to the independent predictive power of age, PSA, free PSA, prostate volume and PI-RADS lesion in relation to the occurrence of prostate cancer.

In summary, our study provides significant data that support the predictive value of the mpMRI as well as clinical and PSA-related factors, specifically in patients with negative mpMRI. With further investigation, these parameters could become mandatory in the decision process regarding prostate biopsies in cases with negative mpMRI to elevate the negative predictive value of mpMRI even more. In selected patient groups (older patients, higher prostate volume, higher free PSA and lower PSA-Density), prostate biopsy after negative mpMRI can be avoided due to lower risk of undetected prostate cancer. But counseling of these patients regarding residual risk for undetected prostate cancer and an offer for a prostate biopsy should be mandatory. Biomarkers could provide more diagnostic certainty. However, many candidates still require extensive research and are bound to be less accessible in the near future, because PSA and mpMRI are current fundamental diagnostics for PCa. Additional imaging procedures as high frequent ultrasound may help to reduce these problems even more.

Our study also has some limitations. Data acquisition was retrospective, and study population was relatively small. Also, MRI machines and PI-RADS Versions varied in the observation period. Furthermore, prostate biopsies differed in biopsy counts. There is no fixed definition of csPCa, which limits the comparability to other studies.

In addition, it is unclear how many undetected PCa could have been missed in the none biopsy group (52 out of 216). This reflects the real-life scenario in which patients and clinicians find themselves in. This group of patients who never received a biopsy either rejected biopsy or showed stable follow-up (constant PSA, no PI-RADS lesions in following mpMRI, etc.). It is also arguable if a 12-fold systematic biopsy is sufficient as “gold standard” as a significant proportion of PCa is missed by this diagnostic procedure. In our opinion, our endpoint “no diagnosis of PCa” is reasonable, since the risk of undetected clinically significant prostate carcinoma in this specific group of patients is low. Therefore, we think it is justifiable not to perform prostate biopsy, if sufficient follow-up care is guaranteed. A major concern remains that the follow-up time of 21 months might not be long enough to fully answer this question without performing biopsy in all patients.

Further information on the entire patient cohort, including men with PIRADS 3–5, would be of value. However, the goal of this specific study was to depict real-time scenarios where some patients had unclear MRI results. Specifically, we aimed to dig deeper in those patients with no PIRADS lesion or PIRADS 1 and 2 lesions. Therefore, we conducted telephone interviews within our patient cohort. Due to economic and staff related shortcomings, it was not possible to perform this detailed work up for all patients. Therefore, we do not have data of the 1120 mpMRI which showed suspicious lesions (PI-RADS ≥ 3). Therefore, we cannot investigate, on how many of these patients subsequently had prostate cancer.

We think that our follow-up is comparatively consistent due to telephone interviews on all patients who were not followed-up in our clinic. We understand that our study suffers from lack of standardization, randomization, controls and selection bias, due to the retrospective analysis of real-world data.

The additional evaluation of clinical factors such as total and free PSA, age and prostate volume in combination with mpMRI could increase the diagnostic certainty in patients with negative mpMRI with regard to non-significant prostate cancers in order to reduce unnecessary prostate biopsy. Patients with negative MRI are in sum at low risk of csPCa.