Diagnostic Performance of PSMA-Targeted PET Tracer Excels in Relapsed Prostate Cancer

Michael J. Morris, MD

Imaging with ¹⁸F-DCFPyL-PET/CT outperformed that of standard imaging modalities—such as bone scan, CT, MRI, and FDG PET—in patients with biochemically relapsed prostate cancer in the phase 3 CONDOR trial.

“The CONDOR study met its primary endpoint. Indeed, it well exceeded it. And it demonstrated excellent diagnostic performance of PyL PET in men with biochemically relapsed prostate cancer, even at very low PSA values,” said Michael J. Morris, MD, when presenting findings during the 2020 ASCO Virtual Scientific Program. “It clearly showed superiority to the standard imaging these men received as part of their local work-ups.”

PSMA-targeting PET has been suggested to be a superior method for imaging in prostate cancer compared with standard imaging modalities although it is not yet approved for use in the United States. ¹⁸F-DCFPyL is an investigational PSMA-targeted PET tracer. PyL is a lysine linked urea-based small molecule that targets the extracellular domain of PSMA. For imaging, 9 (±20%) mCi of trace is administered intravenously as a bolus injection and then PET imaging is completed 1 to 2 hours after.

The CONDOR trial is the second of 2 prospective, multicenter trials examining the diagnostic performance of ¹⁸F-DCFPyL-PET/CT. The first was the OSPREY trial, which examined the use of the PET tracer in patients with clinically localized disease preoperatively as well as in patients with locally recurrent or metastatic disease. This phase 2/3 study showed high specificity in detecting lymph node disease involvement and significant positive predictive value (PPV) in the detection of metastatic disease.¹

CONDOR explored the use of ¹⁸F-DCFPyL in patients with rising prostate-specific antigen (PSA) concentrations. Patients were eligible for the study if they had biochemical relapse in terms of a PSA of at least 0.2 ng/mL following radical prostatectomy or at least 2 ng/mL over the nadir following radiation therapy or cryotherapy, and negative or equivocal imaging per stand-of-care work-up. Those who were undergoing systemic therapy or who had received androgen deprivation therapy within 3 months were excluded from participating in the trial.²

The primary end point was the correct localization rate (CLR), which equates to PPV with anatomical lesion location matching, and was defined as the percentage of patients with a 1:1 correspondence between ¹⁸F-DCFPyL-PET/CT and the composite truth standard, which was matched with an anatomic atlas created for the study.

A secondary end point was the percentage of patients who had a change in their treatment plan as a result of the ¹⁸F-DCFPyL-PET scan, which was assessed with pre- and post-imaging questionnaires from the clinician, and disease detection rate was an exploratory end point.

"Because this patient population has neither available tissue nor informative standard imaging, the end points of this study warrant special scrutiny,” said Morris, who is clinical director of the Genitourinary Medical Oncology Service and Prostate Cancer Section Head in the Division of Solid Tumor Oncology at Memorial Sloan Kettering Cancer Center in New York.

For the sake of the trial, composite truth standard was defined—in order of importance—as either evaluable local histopathology findings from surgery or biopsy, informative conventional imaging, or confirmed PSA response (≥50% decline from baseline) in patients treated with radiotherapy only following imaging, based on feasibility. This was used to define true positive from false positive.

For a positive result, the study assumed that the lower bound of the 95% confidence interval of the CLR exceeded 20% from at least 2 of the 3 blinded, independent central readers—due to the low PSA values of the patient population, that 60% had positive ¹⁸F-DCFPyL-PET/CT findings, and that 30% of positive PSMA-PET scans are confirmed per the composite truth standard.

All patients who had a positive scan with at least 1 lesion detected on ¹⁸F-DCFPyL-PET/CT were required to undergo a confirmatory study to assess the composite truth standard. To further minimize bias of the readings, 2 adjudicators examined the composite truth standard and a statistician compared the outcomes for the final true/false positive verdict.

A total of 208 men were dosed with ¹⁸F-DCFPyL in the study. These patients had a median age of 68 years (range, 43-91) and were a median of 71 months (range, 3-356) from the time of their diagnosis. Nearly half (49.5%) of all patients had undergone radical prostatectomy, 35.6% underwent radical prostatectomy and radiation therapy, 14.9% only received radiotherapy, and 27.9% received at least 1 systemic therapy for their prostate cancer. Almost three-quarters of the men (73.6%) had a total Gleason score below 8.

The median PSA was 0.8 ng/mL (range, 0.17-98.45), and 31.2% of patients had a PSA of at least 2.0 ng/mL.

“The PSA values were representative of this group of men. As expected, the median PSA of this population was quite low at 0.8,” Morris commented. “And generally, these are the values at which most decisions regarding salvage therapy and therapeutic plans are made.”

Across the 3 readers, the rate of positive ¹⁸F-DCFPyL scans was 59.1% to 65.9%, and 11.5% to 15.9% were unevaluable due to the standard of truth not being submitted or suggesting a false negative finding.

The CLR, or the rate of true positives over true and false positives, ranged from 84.8% (95% CI, 77.8%-91.9%) to 85.6% (95% CI, 78.8%-92.3%) to 87.0% (95% CI, 80.4-93.6%).

“For every reader, the CLR was excellent. The lower bound of the 95% confidence interval was well in excess of the 20% benchmark,” Morris said.

A high CLR was found across all PSA levels with a median rate of 73.3% (95% CI, 51.0%-95.7%) for patients with a PSA below 0.5 ng/mL and a median rate of 96.4% (95% CI, 89.6%-100%) in those with a PSA of at least 5 ng/mL.

Even at the lowest PSA levels, ¹⁸F-DCFPyL was able to detect disease with a median detection rate of 36.2% (95% CI, 24.9%-47.6%) in patients with less than 0.5 ng/mL PSA. In patients with a PSA of at least 5 ng/mL, the detection rate was 96.7% (95% CI, 90.2%-100%).

A total of 63.9% of evaluable patients had a change in their management plan resulting from their ¹⁸F-DCFPyL-PET/CT scan, 78.6% of which were due to positive scans and 21.4% due to negative scans.

Twenty-one percent of the men had a change in the goal of their disease management from a noncurative approach to a curative treatment, 28% changed from receiving salvage local therapy to systemic therapy or added systemic therapy, 23.9% went from observation status to receiving therapy, and 4.4% changed from a planned treatment strategy to observation status alone.

“The results demonstrated that actionable information was furnished to clinicians in order to make clinically significant decisions,” Morris said, although optimized treatment patterns still need to be further defined.

No significant safety issues were associated with ¹⁸F-DCFPyL. Only 3 patients had at least 1 treatment-related adverse event, which were 1 case each of fatigue, hypersensitivity, and headache. Hypersensitivity was the only grade 3 event reported, which was observed in a patient with a significant allergic history. This safety profile was considered to be similar to that from the OSPREY trial.

“The CONDOR trial coupled with the OSPREY study has now established the performance characteristics of PyL across the full spectrum of prostate cancer,” Morris concluded.

References

1. Progenics reports results of phase 2/3 trial of PSMA PET imaging agent PyL for the detection of prostate cancer. News release. Progenic Pharmaceuticals Inc. October 5, 2018. Accessed May 30, 2020. https://bit.ly/2BgbeNg.
2. Morris MJ, Carroll PR, Saperstein L, et al. Impact of PSMA-targeted imaging with 18F-DCFPyL-PET/CT on clinical management of patients (pts) with biochemically recurrent (BCR) prostate cancer (PCa): Results from a phase III, prospective, multicenter study (CONDOR). J Clin Oncol. 2020;38(suppl):5501. doi:10.1200/JCO.2020.38.15_suppl.5501

<<< 2020 ASCO Virtual Scientific Program