Current criteria for disease detection rely on CT, Tc99m-MDP bone scan, and sometimes pelvic or whole-body magnetic resonance imaging. Given the increased sensitivity and specificity in detecting disease compared with these CI modalities, it is expected that PSMA PET will lead to earlier detection of recurrent disease both at the primary and distant sites. As such, the disease and thus the eligibility status of certain patients considered for, or enrolled in, existing clinical trials may change—eg, from biochemical recurrence only to established metastatic disease. This scenario is termed stage migration since the disease stage advances because of lesions found on PSMA PET but not on CI. In another scenario, response assessment may change from stable disease (on the basis of a lack of change on CI) to progressive disease (on the basis of PSMA PET detecting sites of disease that could not be detected by CI), resulting in potentially premature abandonment of efficacious treatment. Stage migration on the basis of more accurate determination of the presence or absence of metastatic disease will lead to reclassifying patients previously considered low risk to higher risk. However, better characterization of disease state by PSMA PET imaging may lead to more accurate study populations in clinical trials. The number of patients determined to be nonmetastatic by PSMA PET required to show clinical benefit from a therapeutic intervention will increase, as will the follow-up time and cost.

Following stage migration at baseline, the sample size of ongoing clinical trials in the (now truly) lower-risk space may need to increase, with longer follow-up and associated higher costs, to observe a sufficient number of events and maintain adequate power when using CI-defined end points (eg, radiographic progression-free survival [rPFS] and metastasis-free survival [MFS]). Developing PSMA PET databases to assess the magnitude of the impact on trial outcomes is important to both inform and optimize the study design of future clinical trials. Alternatively, risk groups could be defined using previously validated models, with patients then substratified by PSMA PET imaging findings.

Radiographic progression events on the basis of PSMA PET will likely occur earlier than those on the basis of CI. Trial outcome data may not be comparable with those from the pre-PSMA PET era. It is thus conceivable that end points may be reached earlier, potentially reducing the sample size, increasing the power, or decreasing the needed follow-up time. By contrast, certain patients with locally advanced prostate cancer and at high risk but with negative CI may undergo PSMA-based upstaging. If this occurred, the time to enroll lower-risk (PSMA-negative) patients into trials, the needed sample size, and the required follow-up time would have to increase to observe a statistically meaningful number of events. Trials testing PSMA PET as a valid measure for response will require baseline scans at the time of enrollment and a definition of criteria indicating progression (analogous to Prostate Cancer Working Group [PCWG] criteria for the declaration of progressive disease using bone scan).⁹ A major concern for both patients and physicians is whether earlier progression events as detected by PSMA PET will be clinically meaningful—ie, true disease progression versus early detection of an imaging marker that does not necessarily translate into worse overall outcome and survival. Although the detection of new (true-positive) lesions by PSMA PET is probably a manifestation of progression, future studies will need to define appropriate criteria for progression and validate whether this also justifies a change in current therapy. In the extreme, premature acceptance of PSMA PET-determined progression as a clinically meaningful surrogate marker (ie, biologically and quantitatively related to an accepted outcome parameter) could lead to abandoning therapy or commencing new or layered therapy without evidence of survival or clinical benefit. This would be analogous to trials from the 1990s that removed men from study for PSA progression only and without radiographic progression. For instance, metastasis-directed therapy is currently under study in several trials, including therapy to PSMA-positive sites as shown in ORIOLE.¹⁰ Regardless of the potential clinical benefit of these approaches, such data will not be comparable to those from past clinical trials with systemic therapy. Also, in randomized, unblinded trials, some patients assigned to the control arm may be classified as exhibiting progressive disease, potentially introducing bias favoring the treatment arm. Finally, the best frequency of follow-up PSMA scans also needs to be defined. As these data are still being collected, many experts consider it premature to use PSMA PET for routine clinical decision making until the results of prospective studies become available. Of note, the PSMA signal is modulated by androgens and the androgen receptor status (recently reviewed).¹¹ Short-term androgen deprivation therapy leads to increased PSMA expression, and this has two implications. First, it is not clear whether short-term androgen deprivation should be administered routinely before therapy to boost the PSMA signal to derive an accurate baseline assessment of the extent of disease. Second, a minimum time between baseline and follow-up (a 3 months' time window has been proposed)¹² may be necessary to assess therapy response, avoiding any short-term flare phenomenon. It is not clear whether these concerns similarly apply to modern androgen receptor targeting therapies. Measurement of testosterone levels on the day of imaging will provide information on whether castrate levels of testosterone were achieved.

MFS, as defined by PCWG2,¹³ is a commonly used end point in prostate cancer clinical trials, serving as a validated surrogate for overall survival (OS) in high-risk disease when using current CI.^14,15 MFS may not retain its status or strength of surrogacy when using PSMA PET; this question must be addressed in future phase III clinical trials. For instance, although the PSMA imaging signal correlates with PSA level, the PSA level per se and changes in PSA level are not currently accepted surrogate markers for patient outcome. Investigators have proposed a definition for what constitutes metastatic disease on PSMA-PET as well as PSMA response criteria.¹² For instance, on the basis of expert opinion and depending on clinical scenario, one proposed definition of progression includes two new lesions on a PSMA scan or one new lesion in a location consistent with prostate cancer spread, regardless of RECIST or PCWG criteria (eg, pelvic or retroperitoneal node even if < 1 cm; new focal PSMA uptake in bone marrow after exclusion of reasons for false positives, such as post-traumatic or degenerative change).

Whether rPFS as assessed by PSMA PET is an acceptable surrogate for OS needs to be studied. In determining the clinical validity of PSMA-rPFS for assessing OS, both trial-level and patient-level utility should be assessed. Addressing this question should be a priority for current and future prostate cancer clinical trials.

Cost and power considerations are similar as outlined for the use of the rPFS end point discussed above, if PSMA-positive lesions prove to be clinically significant. Importantly, consistency in monitoring and the use of PSMA PET in both arms of clinical trials is necessary to avoid any potential bias.

Considering the many ways in which PSMA PET can affect disease staging and outcome measures (Table 1), experts discussed several approaches regarding how to integrate PSMA PET into ongoing clinical trials that use CI as an outcome measure, including:

1. Discourage the use of PSMA PET for disease monitoring since the long-term implications of scan findings are currently unclear.

2. Perform PSMA PET at baseline and follow-up, but consider scan findings only as outcome measures for rPFS if confirmed by biopsy; however, this may still alter the traditional end point definition as long as PSMA PET–positive lesions do not meet traditional CI criteria (eg, < 1 cm PSMA-positive lymph node).

3. Observe PSMA PET findings until they become positive on the basis of CI criteria, acknowledging that changes in PSMA PET are currently not validated as meaningful end points in clinical trials.

No consensus was reached. Regardless of the final approach taken, the handling of potential PSMA PET results should be specified in protocol amendments to ensure uniform patient management and outcome measurements. Of note, each approach may result in additional psychologic burden to patients.

The clinical trial cost per patient may increase substantially if and when PSMA PET is made mandatory for study entry and evaluation of outcome. The trial (rather than the patient) should bear this cost. By contrast, the earlier ascertainment of disease outcomes may, in fact, decrease overall study duration and costs. Widespread access to PSMA PET across the country needs to be ensured using different means, including increasing the availability of PSMA outside of major academic hospitals. However, this is likely to occur with the current approval of Pylarify and the imminent approval of other PSMA PET imaging agents. As additional PSMA PET agents are approved, study subjects may need to be stratified if the diagnostic performance of these agents is significantly different.

As the availability and use of PSMA PET increase, and as both ongoing and future clinical trials are affected, discussing these issues with the appropriate review division at the FDA in the near future will be important. These discussions will be particularly critical if the trial under discussion is designed to support regulatory approval of an investigational product.

Investigators need to ensure that diverse populations are enabled to participate in clinical trials and have access to PSMA PET imaging throughout the country, regardless of ability to pay, geographic location, and/or sociodemographic status. Unequal or disparate access to PSMA PET by sociodemographic group must be studied, so that conclusions about the utility of this novel imaging technique can be applied appropriately. Patient-focused communication about PSMA PET must be developed and is essential to (1) ensure that patients understand how their staging and subsequent treatment and risks and benefits may change on the basis of PSMA PET results, (2) reduce stress about lesions identified on PSMA PET but not on CI, (3) adequately explain the potential reasons for and prevalence of false-positive PSMA PET results to patients, and (4) ensure patient understanding that PSMA PET remains an investigational test in clinical trials until its utility in various settings is fully validated. The costs of adding PSMA PET to existing and future clinical trials must be considered during trial design and amendments.

Ideally, PSMA PET may be performed as PET-CT in conjunction with a full-dose CT (including IV contrast, if and as prespecified by the trial), rather than as a separate, additional test. If this is not feasible, the CT of the PET-CT may be obtained with moderate dose for anatomical localization. Image acquisition protocols need to be standardized.

Since PSMA PET is a new technique, reader training must be implemented to reduce the incidence of false positives.^16,17 Adoption of defined criteria, such as those proposed in the European Association of Nuclear Medicine standardized reporting guidelines for PSMA,¹⁶ may be helpful in the interpretation of PSMA scans and in the standardized reporting of imaging findings. In addition to recording a prespecified number of target lesions on PSMA scans and CI, determination of involved regions and organ systems, as well as assessment of volumes of disease on PSMA PET and CI, may provide valuable and potentially predictive and/or prognostic information.

In conclusion, regardless of clinical state, PSMA PET has high sensitivity for prostate cancer lesions, frequently showing metastatic disease earlier and more extensively than CI does. The clinical implications of such findings and the potential role of PSMA PET as a predictive or prognostic marker in prostate cancer need to be investigated in future clinical trials. Until then, PSMA imaging findings, per se, should not affect patient management and trial outcome; however, retrospective subset analysis may potentially provide meaningful information.