Virtual Prostate Biopsy with Prostate-specific Membrane Antigen and Magnetic Resonance Imaging: Closer to Reality in a Subgroup of Prostate Cancer Patients?

Prostate-specific membrane antigen (PSMA) is highly expressed on most prostate cancer (PC) cells, and this evidence has led to development of several PSMA ligands for positron emission tomography (PET) imaging. The nuclear medicine community has come a long way since the first in-human applications of 68Ga-PSMA-11, which date back to 2012, and at present the use of PSMA PET is suggested by several international guidelines for investigating PC in different clinical settings [1]. In particular, PSMA PET is recommended for detection of recurrence in patients with biochemical recurrence, either after prostatectomy and radiation therapy or with prostate-specific antigen persistence after radical treatment [1]. PSMA PET is also indicated for staging of high-risk PC at presentation and for identifying candidates for Lu-PSMA treatment.

There are far fewer papers regarding the use of PSMA PET for detection of PC [2], [3]. Technically, sensitivity may be limited owing to the spatial resolution of PET (around 5 mm) and very small focal PC lesions will be missed by PSMA PET. Furthermore, it is known that approximately 5% of PCs do not overexpress PSMA, and these may be missed. For these reasons, PSMA PET is not even mentioned in guidelines on the diagnostic workflow for patients with suspected PC, for which multiparametric magnetic resonance imaging (mpMRI) is recommended instead [1]. However, these arguments ignore the probability of PSMA overexpression in more aggressive PCs, leading to easy detection even for smaller foci, and the fact that lower expression in low-grade malignancies (ISUP1) could be a distinct advantage when trying to identify only those malignancies that need intervention [4].

The paper by Heetman and colleagues [5] in this issue of European Urology Open Science provides a change in perspective, suggesting the use of PSMA PET combined with MRI to avoid unnecessary biopsies. The rationale is very simple: both MRI (Prostate Imaging-Reporting and Data System [PI-RADS] score of 4 or 5) and a high maximum standardized uptake value (SUVmax) on PSMA PET have shown good specificity in detecting clinically significant PC. Combining the two methods using smart thresholds (high PI-RADS score and high SUVmax) allows such high specificity for the presence of clinically relevant PC that the need for histological confirmation may become obsolete. In summary, the combination of PI-RADS ≥4 with SUVmax ≥8 identified PC with a specificity of 98%, and the authors speculate that a nomogram combining predictive factors could be developed to identify patients for whom biopsy could be avoided.

Of course, the study is preliminary with a number of limitations, such as being single-center, retrospective, and small. In addition, biopsy will remain indispensable in certain patients with a risk of PC by providing important additional information (eg, Gleason score, specific tumor characteristics, DNA testing) for management decisions [6]. Nonetheless, it is very important to consider the hypothesis of whether histological confirmation of suspected lesions is still needed if mpMRI and PSMA PET, known to be the most accurate imaging techniques, are combined in a “virtual biopsy” [7]. Emmett et al. [8] previously confirmed the high negative predictive value and sensitivity of PSMA PET and mpMRI for identification of clinically significant PC (csPC) and showed that consideration of the intraprostatic PSMA uptake pattern improves the diagnostic accuracy for csPC [9]. Similarly, the PRIMARY trial found very high specificity (100%) for the combination of high PI-RADS score (4 or 5) and high PSMA SUVmax (>9). This current trial also confirms the strong association between grade group on histopathology and high PSMA SUVmax. It is important to note that early work demonstrated that the use of PSMA PET SUVmax alone has limited ability to identify which prostate sextants are tumor-positive and have a Gleason score of 7–10 [10]. In addition, recent data from the PEDAL study indicate that fixed SUVmax cutoffs are potentially not the way forward for optimizing identification of csPC via PSMA PET, and should be investigated in conjunction with other characteristics such as the patterns of tracer distribution, volume, and background/signal ratio [11]. Taken together, the evidence suggests that the first likely subgroup of patients who could safely avoid prostate biopsy are those with negative PSMA PET and negative MRI. A second possible group are the patients for whom a definitive therapy has already been determined, such as radical prostatectomy. In these patients, biopsy before surgical excision could be unnecessary and this potential warrants evaluation in future trials. Using these strategies, the morbidity and discomfort of patients [12] exposed to biopsies can be reduced. Furthermore, there are favorable financial implications for society if modern imaging technologies succeed in replacing biopsies in a subgroup of patients with PC.

From a philosophical point of view, it can be added that with virtual biopsy the concept of diagnosis itself changes from foundationalism to coherentism. Foundationalism holds that knowledge (in this case, diagnostic judgment) should ultimately be based on evidence that cannot be justified further. Assuming that biopsy and histopathology are the gold standard fits the foundationalist agenda. Coherentism is the view that knowledge requires convergence of many items of evidence, none of which is taken as the bedrock [13]. Acknowledged limitations of traditional biopsies in terms of reliability and partiality, together with the good prospects of imaging and artificial intelligence (AI)-supported biopsies, speak in favor of a conceptual shift from foundationalism to coherentism.

In conclusion, the study by Heetman and colleagues [5], although preliminary, brings to light a very interesting prospect. Prostate biopsy is an invasive procedure associated with patient morbidity and a financial burden on society. So far, the emphasis has been on reducing unnecessary biopsies via risk stratification for patients with low incidence of PC; however, if the chance of PC is close to 100% and treatment with radical prostatectomy is planned, a biopsy for histological confirmation could be seen as redundant. This pathway for diagnosing PC without histological proof could be an option for reducing the 1 million prostate biopsies performed in Europe each year [4]. Therapy without histological confirmation is already in practice for other malignancies such as renal cell carcinoma, for which histological proof is only used for radiologically indeterminate renal masses [4]. Lastly, there are potential further opportunities to improve on the method for imaging-based virtual biopsies using AI and radiomics [7].

: Stefano Fanti reports financial support (honoraria/grants/other) outside the submitted work from AAA, Amgen, Astellas, AstraZeneca, Bayer, BED, GE Healthcare, Janssen, Merck, Sofie Biosciences, and Telix. Ken Herrmann reports personal fees from Bayer, SIRTEX, Adacap, Curium, Endocyte, IPSEN, Siemens Healthineers, GE Healthcare, Amgen, Novartis, ymabs, Aktis Oncology, Theragnostics, Pharma15, Debiopharm, AstraZeneca, and Janssen, personal fees and other from Sofie Biosciences, nonfinancial support from ABX, and grants and personal fees from BTG, all outside the submitted work. The remaining authors have nothing to disclose.