With New Technology Comes Great Responsibility: Prostate-Specific Membrane Antigen Imaging in Recurrent Prostate Cancer

Prostate-specific antigen (PSA) is both a blessing and a curse for prostate cancer. It is a blessing because it makes prostate cancer one of the few cancers with a reliable serum biomarker, but it is also a curse because it can drive overdiagnosis and overtreatment in newly diagnosed disease and also because it creates anxieties for men with recurrent disease.¹ In patients previously treated with radiation or surgery for localized prostate cancer, 30%-40% will experience a rise in PSA, an initial signal of recurrent disease.^2,3 With an estimated 1.4 million men diagnosed each year worldwide, this could mean approximately 500,000 will be detected with recurrent disease.⁴ Given the ability of PSA to detect even low levels of recurrent disease, this often creates a vexing scenario where men have biochemical evidence of disease for years, but no anatomic evidence of disease on conventional imaging with computed tomography (CT) or Technetium-99m (Tc99) bone scans. The recent advances in positron emission tomography (PET) technology, specifically prostate-specific membrane antigen (PSMA) PET, are now able to detect anatomic disease earlier in the course of recurrence. Despite the initial exuberance over the potential of PSMA PET imaging, it is possible that the ubiquitous use of this new technology, like PSA before it decades ago, will lead to substantial overtreatment and stage migration in patients with recurrent prostate cancer, who might otherwise have lived many years without clinically meaningful disease.

The PET agents, ⁶⁸Ga-PSMA-11 and ¹⁸F-DCFPyL, bind to PSMA and have been approved by the US Food and Drug Administration since 2020 and 2021, respectively, for patients with newly diagnosed high-risk disease or patients with recurrent disease.⁵ The sensitivity and specificity of this new technology cannot be denied. For patients with recurrent disease, the ¹⁸F-DCFPyL PSMA scan was found to have a sensitivity of 95.8% and a positive predictive value of 81.9%.⁶ In a study of 90 patients with biochemically recurrent disease (median PSA = 2.5 ng/mL), 77.8% had positive findings on ¹⁸F-DCFPyL PSMA scan, including 50% of patients with a PSA of 0.5 ng/mL or less.⁷

The ability to stage primary prostate cancer more accurately will be a critical new tool for determining if surgery or radiation with androgen deprivation therapy (ADT) is appropriate when evaluating the curability of a given patient. PSMA PET will be used to limit morbidity of futile surgical resections in patients with known metastatic disease or perhaps select patients for whom a clinical trial may be most appropriate.⁸ Studies need to be done to evaluate how systemic ADT affects findings on a PSMA scan in patients undergoing radiation to their primary tumor within their prostate. In addition, new imaging strategies, such as PSMA PET, may increase the efficacy of salvage options for these patients by identifying those who are the best candidates for such interventions. There will be hope among patients and physicians alike that the identification of oligometastatic findings on a PSMA scan may be successfully treated with tailored focal radiation. This optimism should be tempered by the fact that ¹⁸F-DCFPyL PSMA likely does not detect all the disease present and may be even less sensitive for bone (micro)metastases than the other PET agents.⁹ Furthermore, data from prospective studies evaluating PET-directed therapy suggest that the majority of patients with limited disease are not cured with intervention.¹⁰ That said, a noteworthy progression-free survival benefit has been demonstrated when PSMA scan–directed stereotactic ablative radiation is used in patients with limited disease (1-3 lesions).¹¹ A separate study demonstrated that in patients with more than a solitary finding on PSMA scan, the benefits of PET-directed therapy (in terms of a biochemical response) may be limited to a median of 3-6 months.¹² More prospective investigation in PET-directed therapy is required to better understand the durability of response and benefits on the basis of the extent of disease seen on PSMA imaging.

The greater challenge and need for discipline will come from the large number of patients who have recurrent disease after definitive surgery or radiation and who have exhausted their salvage options. For decades, these patients were considered to have biochemically recurrent prostate cancer because despite a rising PSA, their conventional (CT and Tc99) scans were negative and biochemical detection was the only basis for their disease status. It is important to acknowledge that this is not a distinct biologic condition but rather an artifact of a gap in detection technology. Physicians had the ability to detect circulating biomarker–based evidence of recurrent disease (ie, PSA) but did not have the radiographic capability of visualizing the small amounts of disease that accounted for the rise in PSA. The generally accurate presumption was that these patients had microscopic amounts of metastatic disease, but the amounts were below the detection limits of conventional CT and Tc99 bone scans. Over the decades, investigators attempted novel imaging agents such as ProstaScint and ¹⁸F-sodium fluoride PET to narrow this gap, but both suffered from sensitivity and specificity issues, respectively, limiting their wide acceptance.^13,14 Additional PET platforms such as ¹¹C-choline were also used, but there has been a lack of randomized or prospective data.¹⁵ PSMA-based PET imaging is clearly superior to its predecessors; however, this added ability to detect smaller amounts of disease is not necessarily a game changer.⁶ This imaging strategy is indeed revolutionary and will change how we see prostate cancer. The question is, should it change how we treat recurrent prostate cancer?

The advent of PSMA imaging will challenge how practitioners manage recurrent prostate cancer. Up to this point, no clear standard of care existed for these patients. Practitioners would frequently monitor patients for years until the patients had a PSA that reached a value of concern to the practitioner or the patient or both. Others would adhere to known data in the field and resist treatment until patients had metastatic disease manifest on CT or Tc99 bone scans, where randomized data clearly demonstrate the efficacy of androgen deprivation–based combinations.^16-18 But now with the introduction of PSMA PET, previously subclinical disease can be identified and anatomically localized. These findings will exist in a data void without any clear guidance from prospective trials informing the management of PSMA-detectable recurrent prostate cancer with negative CT and Tc99 scans, especially if metastasis number is more than just a limited number of lesions (eg, 1-3). These imaging findings may increase anxiety in patients, especially if they are labeled as metastatic with all the connotations that term engenders. This anxiety combined with a plethora of treatment options for metastatic castration-sensitive prostate cancer (mCSPC) will likely lead to a strong temptation to initiate therapy with androgen deprivation, in combination with an antiandrogen (eg, enzalutamide and abiraterone) and/or docetaxel.^16-19

But will aggressive early therapy improve outcomes in patients with positive PSMA imaging? Of course, our hope is it will. However, the reality is none of the extant trials done in mCSPC have used PSMA PET imaging as an eligibility criterion. In fact, men with negative CT and Tc99 scans were ineligible for all the trials in mCSPC.^16-19 Therefore, what initially seems like a reasonable extrapolation of mCSPC data quickly becomes a presumption of benefit in a population with less advanced disease. Furthermore, the majority of patients with mCSPC in these trials had de novo disease, that is, disease that was aggressive enough to evolve despite screening or over a short period of time. Recurrent disease that evolves perhaps many years after an initial diagnosis and treatment (with surgery or radiation) is likely to be more indolent and thus may not require immediate therapy upon detection on PSMA PET imaging. Indeed, some mCSPC trials specifically excluded patients with recurrent disease, whereas others that demonstrated clear efficacy in patients with de novo mCSPC had confidence intervals that cross 1.0 in subgroups of patients with previous local treatment.^16-19 Moreover, large randomized studies have been done using immediate or deferred ADT in patients with biochemically recurrent disease in the traditional sense (rising PSA, with a negative CT and Tc99 scan). These studies failed to demonstrate the efficacy of ADT to improve survival.²⁰ An important question that is being evaluated in prospective trials is, can PET-directed therapy in limited recurrence be combined with systemic therapy to achieve durable responses or even cures? Obviously, with ADT, time to PSA progression could be improved but perhaps at great cost to quality of life for men who may have 5-10 years until radiographic progression on CT or Tc99 scans, so patient selection will be important.^21-23

What is missing in this equation is knowledge of the natural history of PSMA PET–positive disease. Although extrapolations from existing mCSPC data are possible, there is great risk in these assumptions. There are, however, well-established natural history data on patients with biochemically recurrent prostate cancer which comes from Johns Hopkins where the general practice for decades was to follow patients over time with PSA recurrent disease until they developed overt metastasis on CT or Tc99 scan. These data clearly demonstrate that the PSA doubling time can be used to identify patients at risk for metastatic progression on the basis of conventional imaging. For patients with a PSA doubling time that is rapid (< 3 months), metastatic disease on conventional imaging is expected (50% at 1 year), but in patients where the PSA doubling time is beyond 3-6 months, the time line is very different.²⁴ In an analysis of 450 patients with a median follow-up of 8 years, patients with a PSA doubling time of 3-9 months had a median time to metastatic progression on conventional imaging of 4 years, and for patients with a PSA doubling time beyond 9 months, it was over 13 years. For added context to demonstrate the relative incongruence of the indolent nature of the biochemically recurrent population, most studies of patients with mCSPC have a median overall survival of 4-6 years.^16-18 Thus, treating metastatic disease on the basis of PSMA PET findings could cause a lead time bias of perhaps 5-10 years or more. The patient might not actually benefit from such intervention but would incur the long-term effects of ADT combinations and other therapies.²³

It is also noteworthy that patients can frequently have PSMA PET imaging findings with PSAs < 0.5 (when a doubling time is difficult to ascertain) and in patients with PSA doubling times as long as 20 months.^7,25 These data highlight a potential fundamental flaw in using a lone PSMA scan to determine the need for aggressive, prolonged treatment of disease in the absence of a broader clinical context.

We recently encountered a 63-year-old patient who is illustrative of the potential stability of PSMA scans over time. He was initially treated with definitive radiation in 2011. By 2016, he had biochemically recurrent disease and he enrolled on an immunotherapy trial for 6 months with no PSA decline after which he was monitored. By 2019, his PSA was 29 ng/mL, with testosterone levels generally in the normal range (175-442 ng/dL with lower limit of normal = 181 ng/mL) and a negative CT and Tc99 bone scan. An ¹⁸F-DCFPyL PSMA PET scan was performed as part of a clinical trial (ClinicalTrials.gov identifier: NCT03173924). This scan demonstrated multiple sites of abnormal focal uptake including in unusual sites (spleen, liver, and mesentery; Fig 1). Despite these findings and his elevated PSA, his PSA doubling time (more than 17 months) predicted near-term stability, so he was not treated with androgen deprivation–based therapy. Twenty-one months after his initial PSMA scan, his PSA was 49. A repeat PSMA scan demonstrated minor increases in the number of foci with mild increases in standardized uptake value while his conventional imaging (CT and Tc99 scans) remained negative. Tissue acquired from his PSMA-positive local recurrence was consistent with prostate adenocarcinoma.

FIG 1.

Indolent recurrent prostate cancer remains negative on CT and Tc99 scans despite numerous and anatomically atypical ¹⁸F-DCFPyL PET/CT findings. A 63-year-old male with a history of Gleason 8 (4 + 4) prostate cancer, status after neoadjuvant androgen-deprivation therapy followed by radiation treatment with brachytherapy boost in 2011. (A) Baseline ¹⁸F-DCFPyL PET/CT (PSA = 29 ng/mL, normal testosterone) in 2019, including maximal intensity projection, and axial PET/CT images show a large DCFPyL-avid intraprostatic lesion, abnormal focal uptake at the left seminal vesicle, with additional subcentimeter DCFPyL-avid presacral node, and multiple foci in the mesentery and serosal surface of the colon, liver, and spleen (arrows). (B) At the 21-month follow-up, with a PSA of 49 ng/mL, ¹⁸F-DCFPyL PET/CT imaging shows some new DCFPyL-avid mesenteric implants but otherwise similar to the previous scan. CT and Tc99 bone scans have remained negative throughout this clinical course. CT, computed tomography; PET, positron emission tomography; PSA, prostate-specific antigen; Tc99, Technetium-99m.

The lack of progression of disease on Tc99 and CT scans in this case with numerous findings in atypical locations is not surprising on the basis of PSA doubling time. A doubling time beyond 15 months would not be expected to be associated with metastatic disease for many years.²⁴ Furthermore, while the anatomic sites appear atypical, similar previously subclinical findings have been reported on PSMA imaging.^26,27 Although this case may represent an extreme example, the conservative decision of not to treat with ADT and antiandrogens or chemotherapy allowed this patient to maintain his quality of life and did not result in metastatic progression on CT and Tc99 scans.

Thus, as this case illustrates, several key questions remain to define the role for PSMA imaging in recurrent prostate cancer. Does therapy (metastasis-directed or otherwise) improve symptomatic progression and/or survival and at what cost to quality of life? Several trials using this new imaging technology are underway to initiate therapy earlier. However, the necessary corollary is understanding the natural history of disease progression on PSMA imaging in men who wish to defer treatment and its associated toxicity. Does PSA doubling time indeed have the same predictive value it did in the pre-PSMA era? To answer this question, there are plans to conduct serial PSMA imaging on patients with recurrent disease at the National Cancer Institute (Bethesda, MD) as part of an upcoming prospective trial. Until these types of trials with treatment and/or serial imaging without treatment have been completed, providers will have to weigh what is known about the often-indolent clinical course of recurrent prostate cancer against the optimism-inspired drive to deploy treatments that have established benefit in quite a different clinical population (ie, mCSPC). Now that PSMA-based imaging technology will unveil metastatic disease earlier, what will it reveal about our capability to be judicious in how we treat recurrent prostate cancer?

AUTHOR CONTRIBUTIONS

Conception and design: Ravi A. Madan, Peter L. Choyke

Provision of study materials or patients: Ravi A. Madan, Peter L. Choyke

Collection and assembly of data: All authors

Data analysis and interpretation: All authors

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

With New Technology Comes Great Responsibility: Prostate-Specific Membrane Antigen Imaging in Recurrent Prostate Cancer

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).