An Analysis of MRI-Fusion Prostate Biopsy Results in PI-RADS 3 MRI Findings in a Cohort of Men in a Community Hospital Setting ()
1. Introduction
Urologists face a difficult challenge in deciding who should undergo a biopsy to evaluate for prostate cancer. Multiple strategies have evolved to help decide when a biopsy of the prostate should be performed. These must be balanced against the risks of a biopsy including sepsis, cost, the possibility of false positive or negative results, and discomfort. As MRI has become more frequently used as a diagnostic adjunct, our group decided to examine more closely the characteristics of patients who underwent a prostate biopsy with mpMRI findings of PI-RADS 3. Management of this intermediate risk group has proved challenging to many clinicians.
MRI has become an important adjunct in the decision-making process. Advances including 3 Tesla magnets and improved software have increased the sensitivity and specificity of locating likely sites of prostate cancer. The first version of the Prostate Imaging-Reporting and Data System (PI-RADS v1) was published in 2012. Further refinements have improved accuracy and decreased inter-observer variation since. These advances have helped increase the likelihood of tumor detection and staging [1]. The specific definition of the various PI-RADS findings are as follows:
PI-RADS 1: Clinically significant disease is highly unlikely to be present;
Pi-RADS 2: Clinically significant disease is unlikely to be present;
PI-RADS 3: Clinically significant Cancer is equivocal (intermediate risk);
PI-RADS 4: Clinically significant cancer is likely to be present;
PI-RADS 5: Clinically significant cancer is highly likely to be present.
The mpMRI can be used to help target a suspicious MRI lesion, maximizing the likelihood of the biopsy obtaining tissue from the area of interest [2] [3] [4].
2. Materials and Methods
Exempt status was granted on August 2, 2021, by the Clinical Research Review Committee of Lowell General Hospital [5]. After de-identification, a retrospective analysis was performed of all MRI-fusion guided biopsies performed between October 16, 2017, and November 15, 2021, numbering 320 in total. Patients were excluded unless the maximal PI-RADS score was 3 or higher, and appropriate follow-up information and demographic data were complete. A total of 312 evaluable patients were ultimately included. Data were obtained regarding age, ethnicity, pre-biopsy serum PSA level (ng/ml), maximum PI-RADS score obtained on a multiparametric MRI, and calculated PSA density (PSAD) expressed as ng/ml/cc), and the final pathology results. PSA was determined using a Qualigen Therapeutics FastPack PSA system (Qualigen, Carlsbad, CA). If the initial PSA at the time of referral came from another institution, it was generally repeated using the Qualigen system for consistency. A decision to proceed with an mpMRI, and/or ultimately to an MRI-guided prostate fusion biopsy, was made by the patient and their Urologist after a discussion of the risks and benefits. All Urologists were in the same single specialty department at the hospital. Many patients with a PI-RADS 3 finding ultimately elected to not undergo biopsy. This report focuses only on the patients who were biopsied.
MpMRI studies were performed on a Siemens Magnetom Skyra (3 Tesla) device (Siemens USA, Washington DC), utilizing multiplanar T2W sequences with diffusion-weighted imaging and dynamic contrast-enhanced sequences prior to and following the intravenous administration of Dotarem without the use of an endorectal coil. All studies were read by Radiologists with subspecialty training in abdominal imaging. If the decision was made to proceed with an MRI-guided fusion biopsy, three-dimensional segmentation of the prostate gland with mapping of any prostatic lesions was performed on a separate workstation. All prostate fusion biopsies were performed in a trans-rectal fashion using an ARTEMIS 3D Semi-Robotic Prostate Fusion Biopsy System (InnoMedicus, Ltd, Switzerland) under either general anesthesia or deep sedation by an experienced Urologist. All specimens were reviewed by several Pathologists. Pathologic findings included Gleason score, presence, or absence of perineural invasion, concurrent prostatitis, atypical small acinar cell proliferation and presence or absence of Gleason 4 pattern cribriform morphology (which suggests a poorer prognosis). As the description of cribriform pathology was developed during the study period, our Pathologists re-reviewed all the PI-RADS 3 patients whose biopsies contained Gleason 4 areas for the presence or absence of cribriform change, allowing a more contemporary comparison between studies obtained early in the observation period to later specimens.
3. Results
A total of 312 evaluable patients were identified with PI-RADS scores of 3, 4 or 5, (32 patients with PI-RADS 3, 197 with PI-RADS 4 and 83 with PI-RADS 5). Only maximal PI-RADS score were used for analysis. If prostate cancer was identified, the highest Gleason score was reported. Many patients had a component of benign hyperplasia (BPH) in the transitional zone (not separately reported). Significant BPH is recognized as a possible factor accounting for a higher PSA score, as is the presence of prostatitis. In Figure 1, the overall characteristics of the entire study group are summarized. Note the associated trend towards higher average PSA values, higher age, higher PSA density and higher likelihood of identifying cancer as the maximal PI-RADS scores increase.
We then further analyzed the specific subgroup of patients with a maximal MRI finding of PI-RADS 3 as shown in Figure 2 and Figure 3.
A total of 34% of the patients with PI-RADS 3 were found to have prostate cancer ≥ Gleason score 3 + 3) with 15.6% demonstrating Grade Group 2 (Gleason grade 3 + 4) or higher on MRI-guided fusion biopsy. Patients having a prior diagnosis of prostate cancer in the past on active surveillance and not found to have cancer currently (n = 2) were excluded. Twelve out of the 32 patients had new findings of cancer: one patient demonstrated Gleason 4 + 4 adenocarcinoma, two patients demonstrated Gleason 4 + 3 adenocarcinoma, two patients demonstrated Gleason 3 + 4 adenocarcinoma, and seven patients demonstrated Gleason 3 + 3 adenocarcinoma (one in active surveillance). Of patients who eventually underwent surgical treatment, concordance of the final pathology with the preoperative biopsy results was high, although one patient with initial Gleason 3 + 3 was upgraded to Gleason 3 + 4. Six patients identified with cancer
![]()
Figure 1. Overall characteristics of study group.
![]()
Figure 2. Characteristics of maximum score PI-RADS 3 patients.
![]()
Figure 3. Management decision as of 2022 for PI-RADS 3 Patients with cancer.
had concurrent findings of prostatitis (see Figure 2), yet their average PSA was lower than the group, suggesting that this alone did not prompt the biopsy. Patients with cancer in the PI-RADS 3 group had a lower PSAD average than those without cancer (0.137 (ng/ml/cc) vs 0.21). Four of the twelve cancer patients had a family history of prostate cancer, whereas an additional two patients had a family history of female breast cancer. Gleason 4 with cribriform morphology was not initially reported during the early years of this study but was included in the results after pathologic re-review, as it generally conveys a poorer prognosis [6]. As regards treatment, as of this writing in 2022 six patients are in active surveillance protocols, three elected to be treated with external beam radiation therapy, and three proceeded with robotic assisted laparoscopic prostatectomy. As seen in Figure 2, we noted a high correlation with positive biopsy findings in the areas identified as PI-RADS 3 (10 of 12).
4. Discussion
This project demonstrates our experience in a large community hospital with the management of patients with a maximal mpMRI PI-RADS score of 3. This is a common finding on MRI, and often presents a challenge for the treating physician and patient as to whether to proceed with a biopsy. Although our sample size is relatively small, we believe that this is the result of progressive stratification during the workup process for PSA elevation using mpMRI. Our group of 7 Urologists in a single large community hospital required 4 years to find 32 cases of PI-RADS 3 patients that were biopsied. In many communities based on prior studies, the finding of a PI-RADS 3 result typically reduces the likelihood of the offer of a biopsy. We believe that this study had to be done in a retrospective fashion. We did not feel that it would be ethical to perform a prospective study where a group of men with PI-RADS 3 would receive a biopsy and then be compared to a group that did not. As expected, there appeared to be a significant likelihood of finding cancer in patients with mpMRI PI-RADS 4 or 5, and almost no patients were reported as PI-RADS 1 or 2. We noted a positive correlation of an increased pre-biopsy average PSA of patients with a finding of cancer. Of the remaining far smaller group that were found to have PI-RADS 3, and elected to proceed with biopsy, we identified a 34% chance of finding any prostate cancer in this group, with 5 of the 12 biopsies containing elements of Gleason grade 4 (Grade Group 2 or higher) (including one with Gleason pattern 4 cribriform morphology). Radiologists occasionally report a “predicted PSA” on mpMRI reports, based on the measured volume of the prostate, assuming that larger glands will likely produce more PSA. (A PSA density (PSAD = PSA/prostate volume) of 0.12 ng/ml/cc is used to predict the expected PSA score). PSAD has been studied as an independent predictor of the likelihood of finding cancer on biopsy, with higher values conferring a higher chance of cancer [7] [8] [9]. We noted the average PSAD in PI-RADS categories 3, 4 and 5 were higher than the “predicted” PSA, confirming this association. Increasing age also predicted a higher chance of finding cancer.
In the interpretation of prostate MRI, the possibility of inter-observer variation between Radiologists must also be considered. A recent meta-analysis attempted to address this issue, finding good inter-observer agreement in interpretation of PI-RADS scores of 4 or 5, as compared to lesser, but still generally acceptable agreement in the interpretation of PI-RADS 3 [10]. We noted a strong correlation of positive cancer findings in PI-RADS 3 lesions, suggesting that our experience parallels that of centers of excellence.
The racial composition of our cohort paralleled the surrounding towns. An ethnic estimation of the overall geographic study area consists of 54.3% White (non-Hispanic), 25.6% Asian, and 18.1% Hispanic. Black or African Americans make up a relatively small percentage. In our PI-RADS 3 group, there were 23 White patients (72%, of whom 9 had cancer), 5 Hispanic men (16%, of whom 3 had cancer), and 4 Asian men, of whom none had cancer. No Black men in the entire study group had a PI-RADS 3 score: however, all 7 Black men in the PI-RADS 4 group were found to have cancer, and the average PSA was 11.7, which was higher than the overall study population for PI-RADS 4 patients. It is unclear whether this suggests a selection bias for Black men being referred for Urologic consultation until their average PSA is higher than the rest of the population or other unidentified factors. One Black man had a PIRADS score of 5 and was also found to have cancer on biopsy.
For clinicians, several challenges exist. While many organizations advocate some form of screening for prostate cancer (at least with a PSA test) for patients between the approximate ages of 50 and 70 - 75, the best methods for this are debated. Another challenge is deciding which cancers are “clinically insignificant” (i.e. unlikely to cause harm to the individual during their lifespan) and which are “clinically significant” (i.e. likely to progress and produce morbidity or mortality during their lifetime). It is this latter group that should be offered treatment for cancer. Many men are now in active surveillance protocols, in which definitive treatment for the prostate cancer is not immediately undertaken unless evidence of clinical progression is noted. Routine follow up biopsies or follow up pelvic MRI) are frequently used to evaluate for the possibility of progression, in which case definitive treatment may be started. Prior studies have suggested that cancers identified on biopsies done for patients with a maximum score of PI-RADS 3 are associated with a low likelihood of clinically significant prostate cancer, defined as Gleason score 3 + 3 (Grade Group 1) [11]. Recently, however, several groups have questioned this, suggesting that clinically important prostate cancers may be missed with such an approach, particularly if other risk factors are present (older age, smaller median prostate size (which would correlate with a higher PSAD, such as found in our study) [12] [13]. These studies suggest that the size of the MRI-identified lesion did not have a positive correlation with the likelihood of finding cancer. These authors have called into question the frequent practice of not performing a biopsy for PI-RADS 3 lesions, finding that a higher than previously reported incidence of clinically significant prostate cancer may in fact be present. Findings of clinically significant prostate cancer (defined here as Grade group ≥ 2) may range as high as 27% for a first biopsy [10] [11]. Our data finds a significant percentage of all-grade prostate cancers (34%) in this group, with 5 of the 12 positive biopsies in the PI-RADS 3 group demonstrating a Gleason score of ≥6 (i.e. ≥Grade Group 2). This equates to the finding of “clinically significant” prostate cancer in 15.6% of all PI-RADS 3 findings on MRI.
The use of MRI will likely continue to increase, both at academic and community hospitals. This will lead to more findings of PI-RADS 3 cases, challenging Urologists who need to educate patients on the implications.
The newer literature has begun to suggest that the widespread practice of not offering a biopsy to a man with a PI-RADS 3 lesion may be short sighted, and that there are patients in this group that harbor clinically significant prostate cancer-thus excluding them from diagnosis simply based on an “equivocal” PI-RADS score. This may be an incorrect approach without including other factors that may favor a biopsy—PSA density is one such example [14]. We believe that our study is the only one of its kind that employed MRI-fusion guided biopsies for 100% of the study group, increasing the likelihood that the significant lesions identified on imaging were in fact the tissue that was sampled.
Finally, we note that most of the Urology practiced around the world is performed by community Urologists, and our large community hospital experience should add to the world’s experience with this challenging diagnosis and management.
5. Conclusion
Our findings, as well as the evolving literature, suggest that only performing a prostate biopsy on an MRI PI-RADS finding of 4 or 5 may miss some men who have clinically significant (i.e. Grade Group 2 or higher) prostate cancer. An individualized approach is highly advised, utilizing as many additional decision making factors as possible.
Funding
No author has any financial relationships to disclose.
This study was not funded by any individual, grant, or other funding agency, including private or governmental.