How Accurately does PSMA Inhibitor 18F-DCFPyL-PET-CT Image Prostate Cancer?

In this issue of Clinical Cancer Research, Morris and colleagues (1) report the performance of a novel PSMA-binding agent, 18F-DCFPyL, in locating extra-prostatic foci of prostate cancer in men with biochemical recurrence (BCR) following intent to cure therapy–radical prostatectomy or radiotherapy.

Thirty to forty percent of men who undergo curative treatment of primary prostate cancer have biochemical evidence of persistent/recurrent cancer based upon posttherapy rising serum PSA (BCR). A challenge is to identify the source of cancer using ever evolving imaging methods. CT and ^99mTc-labeled bisphosphonates imaging bone metastasis–related sclerosis form the core of conventional imaging evaluation for metastatic disease. MRI has further improved our ability to detect locally recurrent prostate cancer due to its inherent high soft-tissue contrast. Given the higher resolution of PET and poor performance of ¹⁸F-fluorodeoxyglucose, the workhorse PET tracer for imaging most other cancers, additional PET agents have been developed. These include ¹¹C-choline, ¹¹C-acetate, and ¹⁸F-fluciclovine that target proliferation and metabolic pathways upregulated in prostate cancer. Two of these are FDA-approved for imaging prostate cancer in the BCR setting (2). All of these are still limited in their ability to characterize the full extent of BCR disease.

Glutamate carboxypeptidase II, commonly known as PSMA, has long been regarded as a marker of prostate cancer cells (3). Because of what has been regarded as specificity for prostate cancer, PSMA was used to develop prostate cancer-specific molecular imaging agents. PSMA-targeted imaging approaches would, in theory, provide more specific information about the presence and extent of prostate cancer and potentially improve upon the relatively poor sensitivity of conventional imaging approaches that did not target prostate cancer biomarkers. The first PSMA agent to be approved for clinical use was the antibody-based agent ¹¹¹In capromab pendetide (Prostascint). Unfortunately, this agent suffered from poor sensitivity since the antibody targeted an intracellular epitope, thus limiting access of the target to the imaging agent.

Subsequent imaging agent development has focused on the extracellular component of the PSMA molecule. Radiolabeled peptides that bind the PSMA molecule have confirmed the original hypothesis that PSMA can serve as a target for more accurate and precise localization of prostate cancer (4, 5). ⁶⁸Ga-PSMA-11 PET, which has been used extensively internationally, has finally been approved in the United States. The authors now present performance data that will support likely approval of ¹⁸F-DCFPyL as the second clinically available PSMA imaging agent in the United States. The disease detection rate, reflected as the percent of positive ¹⁸F-DCFPyL PET scans, was similar to that reported for ⁶⁸Ga-PSMA-11. Furthermore, by allowing only patients with negative prior imaging (including CT, MRI, bone scan, or PET imaging with ¹⁸F-fluciclovine or ¹¹C-choline), the investigators demonstrated the added clinical benefit of ¹⁸F-DCFPyL PET information. Interestingly, the study focused on demonstrating a high colocalization rate (CLR) as the primary endpoint of the study, at the FDA’s recommendation. As discussed by the authors, the value of negative ¹⁸F-DCFPyL PET scans remains to be determined, though negative ¹⁸F-DCFPyL PET scans will likely predict a more favorable prognosis (Fig. 1).

Can the performance of 18F-DCFPyL-PET/CT be improved? Not all prostate cancer is detected. As reported by Morris and colleagues (1), detection rates in men with BCR vary with serum PSA level–from 36% of men with PSA < 0.5 ng/mL to 97% with PSA > 5 ng/mL. Can apparent false negative rates be attributed to the size of tumor cells aggregates, which may be too small to be detected with current imaging? Furthermore, expression of PSMA by prostate cancer cells is known to be heterogeneous (2). Conversely, PSMA is expressed by cells other than prostate cancer, including nonneoplastic prostate glands, vascular endothelia of many solid tumors, proximal renal tubular cells, and salivary glands. Few IHC studies confirming that positive scans reflect prostate cancer have been done. Confirmation of detection often relies on indirect data. Finally, when prostate cancer is detected outside of the prostate using more sensitive methods, will treatment change the course of the patient’s cancer? Is there lead time bias (the Will Rogers phenomenon), where the 5-year outcome by stage can be improved just by upstaging higher risk, lower stage cancers using a more sensitive staging modality.

With the imminent availability of two PSMA PET imaging tracers in the United States, a major question that will be faced by imaging centers is which agent to use and when. Because of the more favorable imaging characteristics of F-18 versus Ga-68, ¹⁸F-DCFPyL may have a slightly higher detection rate on a per-lesion basis, particularly for small lesions. Whether this will lead to a clear difference in detection rate in clinical practice remains to be seen because the data to date suggest equivalent detection rates for both tracers. Financial and workflow considerations will likely inform this decision at each imaging center. To this point, F-18 production can be more readily scaled for wide use in such a common disease as prostate cancer than can Ga-68. As to when PSMA PET imaging should be ordered, the data from this study and others demonstrate reasonably high detection rates at PSA levels down to 0.5 ng/mL, but lower detection rates when PSA is < 0.5 ng/mL. Understandably, patients will likely want to know where their cancer recurrence is as soon as the PSA starts rising; however, whether that warrants an immediate move to PSMA PET imaging remains to be determined.

The wide availability of PSMA PET imaging in the United States will also further expand the opportunities to both understand prostate cancer biology and how PSMA contributes to the biology of other cancers in ways that were simply not possible without this technology. With the ever-increasing treatment options available to patients with prostate cancer, assessing treatment response will almost certainly be the next growth area for PSMA PET imaging applications. Such application will provide further insight into how molecularly targeted or immunotherapy-based treatment modalities modify PSMA expression and determine the optimal use of PSMA PET imaging for informing treatment decisions and assessing response to treatment.

Finally, PSMA PET imaging will pave the way for selecting patients for radiopharmaceutical-based therapies. By replacing the positron-emitting radioactive isotope label with isotopes ideally suited for therapy, such as beta emitter Lu-177 and alpha emitter Ac-225, PSMA radiopharmaceutical therapies can effectively treat prostate cancer sites identified by PSMA PET imaging. The success of this approach has already been demonstrated in somatostatin receptor-expressing neuroendocrine tumors, with Ga-68 dotatate PET for detection and Lu-177 dotatate for therapy. Lu-177–based PSMA therapies are actively being developed in the United States and widely used internationally. Success has been reported in limited studies with Ac-225 PSMA.

In conclusion, the expansion of PSMA PET imaging availability in the United States is an exciting development for the entire prostate cancer field, especially for patients with BCR disease who will now have access to this technology. We hope to see drastic improvements in treatment outcomes as the improved accuracy of prostate cancer tumor burden characterization provided by PSMA PET is incorporated into clinical decision making for patients with prostate cancer.

D.L. Chen reports a patent for WO 2015103526 A1 (Mach RH, Chu W, Zhou D, Michel LS, Chen DL. Radiolabeled tracers for poly(adp-ribose)polymerase-1 (PARP-1), methods and uses thereof) issued and licensed to Trevarx; and reports an honorarium from Telix Pharmaceuticals for serving on an expert panel. No disclosures were reported by the other author.

L.D. True acknowledges funding from the Pacific Northwest Cancer SPORE 2P50CA097186. D. Chen acknowledges funding from the Institute of Prostate Cancer Research, University of Washington.