Optimal patient selection for prostate-specific membrane antigen (PSMA) targeted radioligand therapy has become a hot topic since the Food and Drug Administration (FDA) regulatory approval of the first PSMA-targeted radioligand therapeutic agent, lutetium Lu 177 vipivotide tetraxetan (177Lu-PSMA-617; Pluvicto, Novartis) in March 2022.1 177Lu-PSMA-617 is indicated for the treatment of patients with PSMA-positive metastatic castration-resistant prostate cancer (mCRPC) in the post androgen receptor pathway inhibition, post taxane-based chemotherapy setting.
In the prescribing information for 177Lu-PSMA-617, the FDA recommends selection of patients for treatment using gallium Ga 68 gozetotide PSMA (68Ga-PSMA-11; Locametz, Novartis/Advanced Accelerator Applications) “or an approved PSMA-11 imaging agent based on PSMA expression in tumors.”2
In the phase 3 VISION trial, on which regulatory approval of 177Lu-PSMA-617 was based, 68Ga-PSMA-11 PET/CT imaging was used to determine eligibility in patients with progressive mCRPC previously treated with at least one androgen-receptor pathway inhibitor and one or two taxane regimens.3 Eligible patients were those with at least one PSMA-positive metastatic lesion and no PSMA-negative lesions that would be excluded by study protocol criteria. After screening, 831 patients were randomized in a 2:1 ratio to 177Lu-PSMA-617 (7.4 GBq every 6 weeks for 6 cycles) plus standard of care (SOC) or SOC alone. SOC was determined by the individual investigators, but radium Ra 223 dichloride, chemotherapy, immunotherapy, and investigational drugs could not be included.
At a median follow-up of 20.9 months for the primary analysis, median overall survival (OS) was 15.3 months on the 177Lu-PSMA-617 arm versus 11.3 months on the SOC alone arm, a significant 38% reduction in risk of death (HR 0.62; 95% CI, 0.52-0.74; P <.001).
Thomas Powles, MBBS, MD, (Barts Cancer Centre, London, UK) and Brian Rini, MD (Vanderbilt University, Nashville, TN) invited Phillip Kuo, MD, PhD, one of the VISION investigators, to discuss the selection criteria for treatment in the trial,4 and how quantitative imaging parameters in the study were associated with treatment outcomes, which was the focus of a substudy presented at the 2022 annual meeting of American Society of Clinical Oncology (ASCO).5 Dr Kuo is Professor of Medical Imaging, Biomedical Engineering, and Medicine at the University of Arizona, Tucson. He also serves as senior medical director for Invicro (Needham, MA), the contract research organization (CRO) charged with performing the VISION substudy presented at ASCO.
Dr Kuo remarked on the “impressive” speed at which PSMA PET/CT has been adopted into regular practice, but he emphasized that it was used differently in the VISION trial. Prior to VISION, he explained, PSMA PET radiopharmaceuticals were used to diagnose intermediate-to high-risk prostate cancer prior to surgery or biochemical recurrence. In 2017 when the trial began, he was asked to develop rules that would exclude patients least likely to benefit from PSMA-targeted radioligand therapy from the VISION trial. “In the VISION trial, we used PSMA PET imaging in a novel way by creating read rules that had never been tested in a phase 1 or a phase 2 trial, and we went directly into a phase 3 trial,” Dr Kuo claimed.
His team’s first decision was to target an exclusion rate by imaging of 10-20%, which was deemed neither too nor insufficiently restrictive. “Back in 2017, there were too few data to set a definitive exclusion rate, Dr Kuo explained.” Another “huge” decision, Dr Kuo recalled, was not to use 18F-fluorodeoxyglucose (18F-FDG) PET imaging in addition to PSMA PET.
“18F-FDG PET across all sorts of tumors, including prostate cancer, has great prognostic value and even in 2017 data was accumulating on the combination of the two,” he said. “But we decided for many reasons, but primarily just for practicality, that we were only going to use PSMA PET.” Consideration of the potential costs of additional scans was another reason, he noted.
“So we created a new definition, where lesion PSMA positivity was defined as ‘greater than liver’ on visual assessment of a 68Ga-PSMA-11 whole-body scan,” Dr Kuo explained. This excluded 4.9% of patients who had no PSMA-positive lesion in any organ system. The next step was to review the CT scans to identify lesions that fulfilled certain size criteria for evaluation. About 8.7% of patients who had lymph nodes >2.5 cm in short-axis anywhere in the body, solid organ/parenchymal metastasis(es) ≥1 cm in short axis, or bone metastasis with soft tissue component ≥1 cm in short axis were also excluded from the trial. Bone metastases themselves were not used to exclude patients. “We wound up excluding about 14% of patients, so we hit our target range,” Dr Kuo confirmed.
Dr Kuo defended this form of visual assessment as being very practical and a method that nuclear medicine physicians are very familiar with. He stressed that the rules were vetted across a PSMA data set. In addition, a reproducibility study mandated by the FDA with three new central readers applying the same VISION read criteria showed “moderate to substantial agreement.” All the readers agreed on 77% of cases.
“Importantly, we saw agreement for exclusion cases was lower than inclusion,” he noted, adding that, “it’s a lot easier to find a lesion that’s hot than one that’s cold on any kind of PET scan.” He also cited support from a study carried out by UCLA researchers that assessed outcomes in mCRPC patients treated with 177Lu-PSMA-617 but who would have been excluded from the VISION trial based on the PSMA PET/CT criteria.6 It found that these patients had significantly worse outcomes compared with those who would have been eligible for the trial.
Additional evidence that the team ”did a reasonable, maybe a very good job” on the patient selection criteria came from the results of a VISION substudy presented at ASCO by VISION co-investigator Andrew J. Armstrong, MD, Professor of Medicine, Professor in Surgery, and Professor in Pharmacology and Cancer Biology at Duke University, Durham, NC.5 The substudy, which should be regarded as a post hoc exploratory analysis, according to Dr Armstrong, assessed the association between pretreatment 18Ga-PSMA-11 PET/CT qualitative parameters and outcomes with 177Lu-PSMA-617 therapy, with the aim of identifying optimal PET criteria that could prognosticate outcomes in this patient population.
Imaging data meeting quality requirements were analyzed for 548 patients. PSMA expression was quantified by five whole-body PSMA PET parameters: mean standardized uptake value (SUVmean), SUVmax, tumor load (whole-body PSMA-positive tumor volume × SUVmean), tumor volume (whole body PSMA positive lesions calculated from the total volume of voxels with positive uptake), and the presence of PSMA-positive lesions by region.
The most important parameter to emerge, Dr Kuo observed, was SUVmean which was significantly associated with all efficacy endpoints of OS, rPFS, objective response rate (ORR), and prostate-specific antigen (PSA) 50. Higher whole-body SUVmean was associated with prolonged rPFS, as shown by quartile distribution. In the highest quartile, SUVmean of ≥10.2 was associated with the greatest rPFS of 14.1 months, with an incremental decrease with decrease in SUVmean to an rPFS of 5.8 months in the lowest SUVmean quartile (<6.0). “This still compared favorably with 3.4 months in the overall control group in the trial,” Dr Kuo commented.
The results for OS were “a little different,” in that there was no significant association with SUVmean overall, but by quartiles SUVmean was associated with improved OS, with a 21.4 month OS for the highest quartile (SUVmean ≥9.9), but “not quite the incremental change we saw in rPFS,” Dr Kuo noted. “I hypothesize that this is probably due to the fact that we’d already removed the 14% of patients who were least likely to respond, and that is why even the bottom quartile of SUVmean did better than the SOC arm,” he suggested. Drs Powles and Rini noted that SUVmax was not associated with clinical outcomes, which Dr Kuo said was “not surprising, because it’s really just the single hottest voxel in the volume of disease where there could be hundreds of voxels.”
Dr Rini wondered whether one or more of these parameters might enter the clinical world in the future. “If enough people ask for SUVmean we will be able to meet the demand, but right now, it’s not something that you can just ask your radiologist or nuclear medicine people to provide for you,” Dr Kuo conceded. However, for the future, “the whole multidisciplinary community needs to understand and then drive forward what, as in this case, PSMA PET imaging can do for personalized oncology,” he urged.
Linda Brookes, MSc is a freelance medical writer/editor based in New York and London.