The Diagnostic Performance of PSMA PET/CT with 18F-DCFPyL in Prostate Cancer Patients (OSPREY) - Christopher Wallis & Zachary Klaassen

June 1, 2021

In this UroToday Journal Club, Christopher Wallis and Zachary Klaassen discuss The Journal of Urology publication on the Phase 2/3 Prospective Multicenter Study of the Diagnostic Accuracy of Prostate-Specific Membrane Antigen PET/CT with 18F-DCFPyL in Prostate Cancer Patients (OSPREY). Prostate cancer staging is critical in multiple steps in the disease continuum, from the initial diagnosis to the detection of recurrence and progression of the disease. An alternative radiotracer in this space is 18 F-DCFPyL, which is a small molecule that binds to the extracellular domain of the PSMA intermembrane protein. Drs. Wallis and Klaassen review the study design,  objectives, and raise several discussion points of this data. The objective of the OSPREY study was to evaluate the diagnostic performance of this new approach in men with newly diagnosed, or known metastatic prostate cancer, compared to the histopathologic gold standard. 

Biographies:

Christopher J.D. Wallis, MD, Ph.D., Instructor in Urology, Vanderbilt University Medical Center, Nashville, Tennessee

Zachary Klaassen, MD, MSc, Urologic Oncologist, Assistant Professor Surgery/Urology at the Medical College of Georgia at Augusta University, Georgia Cancer Center


Read the Full Video Transcript

Christopher Wallis: Hello, and thank you for joining us for this UroToday Journal Club. Today, we are discussing a recently published study entitled, A Phase 2/3 Prospective Multicenter Study of the Diagnostic Accuracy of Prostate-Specific Membrane Antigen PET-CT with 18F-DCFPyL in Prostate Cancer Patients, also known as the OSPREY study.

I'm Chris Wallis, a fellow in Urologic Oncology at Vanderbilt, and with me today is Zach Klaassen, an Assistant Professor in the Division of Urology at the Medical College of Georgia.

This is the citation of this recently accepted paper in the Journal of Urology.

Prostate cancer staging is critical in multiple steps in the disease continuum, from initial diagnosis, to the detection of recurrence and progression of the disease.

Conventional imaging has historical use, CT, MRI, and bone scan. However, these approaches are suboptimal and lack sensitivity, particularly when PSA levels are low.

For the past few years, there have been alternate PET-CT approaches using 11 carbon choline and 18 fluorine-fluciclovine, which are FDA approved for men with suspected recurrent disease. However, these two, as a dimensional limited gene, lack sensitivity in men with low PSA's.

More recently, 68 Gallium-PSMA PET-CT was approved at selected sites in the US. However, this approach is limited by the need for on-site production and a short half-life.

An alternative radiotracer is 18 F-DCFPyL, which is a small molecule that binds to the extracellular domain of the PSMA intermembrane protein.

So the objective of the OSPREY study was to evaluate the diagnostic performance of this new approach in men with newly diagnosed, or known metastatic prostate cancer, compared to the histopathologic gold standard.

To do this, the authors accrued two separate cohorts. Cohort A comprised of adult men with a newly diagnosed high-risk prostate cancer, who were planned for radical prostatectomy and pelvic lymphadenectomy. And their definition of high-risk disease included those with T3a or greater, PSA of 20 or greater, and Gleason score of eight or greater.

Cohort B included men with radiographic evidence of local or distant recurrence based on conventional imaging, this is CTMI and bone scan, who were planned for biopsy.

Patients were excluded if they had prior ADT exposure.

The authors undertook this as a multicenter open-label phase II/III single-arm study. Baseline imaging, using conventional approaches, was used for all men, and this included bone scan, as well as chest CT, abdomen, and pelvis. And for men in whom contrast couldn't be given, a non-contrast chest CT and a pelvic MRI were used. Following this baseline imaging, patients then received a PET-CT scan using 18 F-DCFPyL, and PET images and non-contrast CT images were obtained at that point. Central imaging was then performed for both the PET-CT, which was interpreted by three independent blinded reviewers, as well as the baseline conventional imaging, which was interpreted by a separate reader, independent of the PET-CT results.

In terms of outcomes, the authors used the histopathologic gold standard, and this was assessed by local pathologists who were blinded to the imaging finding. For cohort A, that is men with high-risk disease, undergoing radical prostatectomy, histopathologic gold standard comprised the pelvic lymphadenectomy nodal packets from an extended template dissection. The number and size of positive lymph nodes were identified.

In cohort B, among men with the known recurrent disease based on conventional imaging, a biopsy of at least one lesion, based on these CT MRI or bone scan results, was used as the gold standard.

In cohort A, the co-primary outcomes were patient-level sensitivity and specificity of PET-CT scanning. Secondary outcomes included positive predictive value, negative predictive value, the detection of M1 disease, and the detection of disease within the prostate in the primary tumor. There was no size or SUV cut-off employed, for the PET-CT interpretation.

In cohort B, the primary outcomes are patient-level sensitivity and positive predictive value. And this was assessed by region and PSA level.

Safety endpoints were also assessed, including adverse events, which were followed from the time of PET-CT through to treatment or biopsy, depending on the cohort.

Analyses were performed only among patients who had a histologic gold standard.  In cohort A, the authors determined that they had an 80% power to detect for the specificity of 80% and sensitivity of 40%. In cohort B, no formal statistical testing was undertaken.

I will now pass it over to Dr. Klaassen, and he can take us through the results of the study.

Zachary Klaassen: Thanks Chris. So we can see here in this flow diagram for cohort A, there were 385 patients that were eligible. Ultimately, 268 in cohort A underwent a PET-CT scan, 16 did not undergo radical prostatectomy or pelvic lymph node dissection. And 252 underwent surgery and a planned pelvic lymph node dissection with pathology results available.

In Cohort B, there were 117 patients. Ultimately, 98 patients underwent image-guided biopsy and had pathology, and 93 patients had a viable pathology for prostate adenocarcinoma. This included 20 pelvic and 19 extra pelvic lymph node biopsies, 44 bone biopsies, and 10 visceral and soft tissue lesion biopsies.

So these are the demographic baseline characteristics. Looking at cohort A in the middle, and cohort B on the right, you can see here that the median age for cohort A was 65, whereas it was 68 for cohort B. The majority of these patients were not Hispanic or Latino, and this included 86.9% of patients in cohort A that were white and 86.3 in cohort B that were white.

In terms of the primary tumor T stage, we see that the most common for cohort A was T1c at 32.5%, followed by T3a at 20.9%. In terms of the T stage for cohort B, the most common T stage was T3a at 20.5%.

In terms of the regional lymph nodes, end-stage for cohort A, the most common was N0 at 58.2%, and for cohort B, was N0 at 39.3%.

For distant metastasis M stage, the most common M stage for cohort A was 80.6% for M0, and for cohort B, was also M0, at 58.1%.

This is a continuation of that demographics table, and you can see here that the total Gleason grade, most commonly for cohort A, was Gleason 8 at 44.8% for cohort A, as well as Gleason 7 for cohort B, at 33.3%.

The median PSA for cohort A was 9.7, and for cohort, B was 7.1. Of course, nobody in the court A group underwent prior prostatectomy, but 47.1% of patients in cohort B underwent prior prostatectomy.

This table looks at the DCFPyL-PET/CT diagnostic performance in the high-risk prostate cancer cohort, otherwise known as cohort A. You can see sensitivity here is in dark blue, specificity in light blue, positive predictive value in red, and negative predictive value in yellow. The take-home message from this slide is that for all pelvic lymph nodes and pelvic lymph nodes greater than five millimeters, the specificity positive predictive value and negative predictive value were all quite high, ranging from low 80% to 97.9%. However, you can see here, the sensitivity for all pelvic lymph nodes was low at 40.3%, which improved slightly to 60% if you consider pelvic lymph nodes greater than five millimeters.

This is a representative picture of a patient that had high-risk prostate cancer, essentially negative bone scintigraphy. But when you look at the images on the right after their DCFPyL-PET/CT scan, we see multiple areas of metastatic deposits.

This figure looks at the sensitivity and positive predictive value of DCFPyL-PET/CT in metastatic disease states, sensitivity is in blue, and positive predictive value is in red. And generally, for overall bone lymph node and visceral soft tissue metastatic sites, all of these had a very high sensitivity, over 95%, and a positive predictive value ranging from 81% to 90%.

Similarly when we are looking at the sensitivity and positive predictive value by anatomic region, the pelvis, with a sensitivity of 100%, and a positive predictive value of 77.8%, in the extra pelvic region, a sensitivity of 96.4%, and a positive predictive value of 83.1%.

This looks at the sensitivity and positive predictive value in cohort B across all PSA ranges. And certainly, you can see that in PSA greater than 20, excellent sensitivity and positive predictive value, which continues through to about the PSA of two to less than five, where we lose the positive predictive value down to about 69%. And then certainly as we get down to a PSA of 0.2 to less than one and less than 0.2, the sensitivity drops down to about 50% for the lowest PSA, and a positive predictive value of 67%.

So several discussion points from this trial; in cohort A, DCFPyL-PET/CT scan demonstrated improved diagnostic performance over conventional imaging modalities, with comparable sensitivity of about 40%, but a threefold higher positive predictive value for detecting pelvic lymph node metastases. Unfortunately, the sensitivity for DCFPyL-PET/CT did not meet the pre-specified endpoint in cohort A, and this may be explained by the no size threshold set for defining a positive lesion considered detectable for PET, but any lesion considered positive by microscopy on pathology, regardless of the size of the lesion.

We also saw that the clinical utility of DCFPyL-PET/CT was confirmed across multiple subcategories of recurrent or metastatic prostate cancer. And we saw that DCFPyL-PET/CT detected the presumptive metastatic disease in 57.6% of cohort B patients that had negative conventional imaging.

So in conclusion, DCFPyL-PET/CT scan demonstrated high positive predictive value, high negative predictive value, and specificity for pelvic lymph node involvement in men with newly diagnosed high-risk prostate cancer, despite a low sensitivity, when compared with histopathology. In men with post-therapy suspected recurrent or metastatic disease, DCFPyL-PET/CT demonstrated high sensitivity and positive predictive value in all sites of disease, and across all PSA ranges.

Finally, DCFPyL-PET/CT provided reliable information to help improve the staging of prostate cancer, compared to conventional imaging.

We thank you very much for your attention. We hope you enjoyed this UroToday Journal Club.
email news signup