Visual Abstract
Abstract
177Lu-PSMA-617 is an effective and novel treatment in metastatic castration-resistant prostate cancer (mCRPC). Our ability to assess response rates and therefore efficacy may be improved using predictive tools. This study investigated the predictive value of serial 177Lu-PSMA-617 SPECT/CT (177Lu SPECT) imaging in monitoring treatment response. Methods: Fifty-six men with progressive mCRPC previously treated with chemotherapy and novel androgen signaling inhibitor were enrolled into the LuPIN trial and received up to 6 doses of 177Lu-PSMA-617 and a radiation sensitizer (3-(4-hydroxyphenyl)-2H-1-benzopyran-7-ol [NOX66]). 68Ga-PSMA-11 and 18F-FDG PET/CT were performed at study entry and exit, and 177Lu SPECT from vertex to mid thighs was performed 24 h after each treatment. SPECT quantitative analysis was undertaken at cycles 1 (baseline) and 3 (week 12) of treatment. Results: Thirty-two of the 56 men had analyzable serial 177Lu SPECT imaging at both cycle 1 and cycle 3. In this subgroup, median prostate-specific antigen (PSA) progression-free survival (PFS) was 6.3 mo (95% CI, 5–10 mo) and median overall survival was 12.3 mo (95% CI, 12–24 mo). The PSA 50% response rate was 63% (20/32). 177Lu SPECT total tumor volume (SPECT TTV) was reduced in 68% (22/32; median, −0.20 m3 [95% CI, −1.4 to −0.001]) and increased in 31% (10/32; median, 0.36 [95% CI, 0.1–1.4]). Any increase in SPECT TTV was associated with shorter PSA PFS (hazard ratio, 4.1 [95% CI, 1.5–11.2]; P = 0.006). An increase of 30% or more in SPECT TTV was also associated with a shorter PSA PFS (hazard ratio, 3.3 [95% CI, 1.3–8.6]; P =0.02). Tumoral SUVmax was reduced in 91% (29/32) and SUVmean in 84% (27/32); neither was associated with PSA PFS or overall survival outcomes. PSA progression by week 12 was also associated with a shorter PSA PFS (hazard ratio, 26.5 [95% CI, 5.4–131]). In the patients with SPECT TTV progression at week 12, 50% (5/10) had no concurrent PSA progression (median PSA PFS, 4.5 mo [95% CI, 2.8–5.6 mo]), and 5 of 10 men had both PSA and SPECT TTV progression at week 12 (median PSA PFS, 2.8 mo [95% CI, 1.8–3.7 mo]). Conclusion: Increasing SPECT TTV on quantitative 177Lu SPECT predicts a short PFS and may play a future role as an imaging response biomarker.
- metastatic prostate cancer
- SPECT
- lutetium-PSMA
- response biomarker
Although treatment resistance and short response duration remain common, 177Lu-PSMA-617 is an effective therapy in metastatic castration-resistant prostate cancer (mCRPC) (1–4). Accurate monitoring of response to 177Lu-PSMA-617 may improve patient outcomes by enabling treatment escalation, change in treatment, or a treatment holiday, dependent on imaging results. Interim and serial prostate-specific membrane antigen (PSMA) PET has recently been shown to be predictive of progression-free survival (PFS) with PSMA-targeted radionuclide therapy (5). Quantitative 177Lu-PSMA-617 SPECT/CT (177Lu SPECT) imaging after each 177Lu-PSMA-617 dose may also be valuable in response monitoring in addition to providing dosimetric information. This LuPIN trial substudy aimed to determine whether quantitative parameters on serial 177Lu SPECT imaging 24 h after 177Lu-PSMA-617 therapy were predictive of treatment response and PFS.
MATERIALS AND METHODS
The LuPIN trial is a prospective single-center phase I/II dose escalation and expansion trial of combination 177Lu-PSMA-617 and 3-(4-hydroxyphenyl)-2H-1-benzopyran-7-ol (NOX66) for men with mCRPC previously treated with at least 1 line of taxane chemotherapy and androgen signaling inhibitor. The clinical results have been previously published (6,7). The St. Vincent’s Hospital institutional review board approved the study protocol (HREC/17/SVH/19 and ACTRN12618001073291), and all patients provided informed written consent.
Screening
Men with progressive mCRPC, based on either conventional imaging (CT and bone scanning) or a rising prostate-specific antigen (PSA) level according to Prostate Cancer Working Group 3 criteria (8), were eligible for screening. Prior treatment with at least 1 line of taxane chemotherapy (docetaxel or cabazitaxel) and an androgen signaling inhibitor (abiraterone or enzalutamide) was required for inclusion. Men underwent screening with 18F-FDG and 68Ga-HBEDD-PSMA-11 PET/CT, bone scanning, and CT of the chest, abdomen, and pelvis. Molecular screening criteria were based on SUVmax rather than physiologic activity (liver or parotid). Men were eligible if they had an SUVmax of more than 15 on PSMA PET at 1 or more sites, an SUVmax of more than 10 at all measurable sites, and no 18F-FDG avidity without corresponding PSMA uptake.
Study Treatment
Men received up to 6 doses of 177Lu-PSMA-617 at 6-wk intervals, with 3 dose-escalated cohorts of NOX66 (400, 800, 1,200 mg). NOX66 suppositories were administered as a radiosensitizer on days 1–10 after each 177Lu-PSMA-617 injection. All cohorts were administered 7.5 GBq of 177Lu-PSMA-617 on day 1 via slow intravenous injection. The PSMA-617 precursor (AAA Novartis) was radiolabeled to no-carrier-added 177Lu-chloride according to the manufacturer’s instructions. Quality control tests for radionuclide and radiochemical purity were performed using high-pressure liquid chromatography and thin-layer chromatography. Blood was prospectively collected before assessment of adverse events and biochemical responses. The patients were treated on trial until they were no longer clinically benefiting from treatment.
Imaging Procedures and Analysis
68Ga-PSMA and 18F-FDG PET/CT scans were obtained at baseline and trial exit (after completing 6 cycles or when treatment was ceased), using the imaging acquisition and analysis parameters previously published (7). 177Lu SPECT (vertex to mid thighs) was performed 24 h after 177Lu-PSMA-617 injection using a Discovery 670 system (GE Healthcare) with the following parameters: medium-energy collimators, 3 bed positions, 60 projections over 360° with an acquisition time of 10 s per frame, 128 × 128 matrix, and 4.42 × 4.42 mm pixel size. An energy window centered on 208 keV ± 10% with a 165 keV ± 6.5% scatter window was used. An unenhanced low-dose CT scan was obtained immediately afterward using the following parameters: pitch of 1, tube voltage of 120 kV, automatic mAs control (reference mAs, 90), slice thickness of 3.7 mm, matrix of 512 × 512, and field of view of 40 cm. The SPECT projection images were reconstructed with an iterative ordered-subset estimation-maximum algorithm that used 4 iterations and 10 subsets using SPECTRA Quant (MIM Software, Inc.). No pre- or postreconstruction filters were applied. CT-based attenuation correction, dual-energy-window scatter correction, collimator-based resolution recovery, and quantitative conversion to SUV were performed during the reconstruction. The conversion from counts to units of activity was based on a cylinder phantom with known activity.
Quantitative Analysis
177Lu SPECT and 68Ga-PSMA PET/CT were analyzed semiquantitatively by a nuclear medicine physician using MIM (LesionID; MIM Software Inc.) software and a standardized semiautomated workflow to delineate regions of interest with a minimum SUV cutoff of 3. All lesions identified quantitatively were manually reviewed and physiologic activity removed. Whole-body quantitation was used to derive total tumor volume (TTV), SUVmax, and SUVmean for both 68Ga-PSMA PET and 177Lu SPECT (9).
Statistical Analysis
We measured PSA decline from baseline (≥50%) at any time point, PSA PFS as defined by Prostate Cancer Working Group 3 criteria, and overall survival (8,10). The Kaplan–Meier method was used to characterize time-to-event endpoints and estimate medians (presented with 95% CIs). We correlated changes in TTV, PSMA intensity, clinical parameters, and biochemical parameters with time-to-event outcomes, using univariate Cox proportional-hazards regression models (11,12). Continuous variables included increase in TTV, SUVmax, and SUVmean. P values below 5% were considered significant. We compared 68Ga-PSMA PET TTV with cycle 1 SPECT TTV using scatterplots and Pearson correlation coefficients. Reproducibility testing of SPECT TTV and PSMA SUVmax was undertaken using repeatability statistics calculated from a hierarchic linear mixed model that accounted for variance in score at the patient level (13). A 95% CI for the repeatability statistics was derived via bootstrapping. Analyses were performed using R (version 4.0.5).
RESULTS
Baseline Patient Characteristics
Of the men enrolled in LuPIN, 57% (32/56) had 177Lu SPECT imaging suitable for analysis, 30% (17/56) had incomplete SPECT data precluding analysis, and 13% (7/56) did not reach cycle 3 of treatment. Baseline characteristics are summarized in Table 1. In this LuPIN substudy, 53% (17/32) completed 6 cycles of treatment whereas 47% (15/32) completed between 3 and 5 cycles, and 63% (20/32) achieved at least a 50% PSA decline from baseline at any time point. At the time of analysis, 84% (27/32) were deceased. There was no difference in either PSA PFS or overall survival based on NOX66 dose. Overall, median overall survival was 12.3 mo (95% CI, 11.7–23.6 mo). Median PSA PFS was 6.3 mo (95% CI, 5.1–9.8 mo).
Baseline Patient Characteristics
177Lu SPECT Quantitation
SPECT quantitation measures at baseline and week 12, including SPECT TTV, SUVmax, and SUVmean, are summarized in Table 2. SPECT TTV was reduced between baseline and week 12 in 68% (22/32; median, −0.20 m3 [95% CI, −1.4 to −0.001]) and increased in 31% (10/32; median, 0.36 [95% CI, 0.1–1.4]). A 30% increase in SPECT TTV by week 12 was identified in 19% (6/32). SUVmax was reduced between baseline and week 12 in 91% (29/32; median, −28.9 [95% CI, −195 to +42]), and SUVmean was reduced in 84% (27/32; median, −2.6 [95% CI, −12 to +10]).
Summary of 177Lu SPECT Quantitation at Cycles 1 and 3
Correlation with Patient Outcomes
An increase in SPECT TTV between baseline and week 12 was associated with a significantly worse PSA PFS (hazard ratio, 4.1 [95% CI, 1.5–11.2]; P = 0.006). Median PSA PFS in those with an increase in SPECT TTV was 4.5 mo (95% CI, 2.8–5.6 mo), compared with 7.1 mo (95% CI, 6.3–10.7 mo) for those with no increase in SPECT TTV. A SPECT TTV increase of at least 30% was also associated with a shorter PSA PFS (hazard ratio, 3.3 [95% CI, 1.3–8.6], P = 0.02) (Fig. 1). Increased SUVmax or SUVmean between baseline and week 12 was not associated with PSA PFS or overall survival (Table 3). By week 12, 25% (8/32) of patients demonstrated PSA progression. PSA progression at week 12 was associated with significantly worse PSA PFS (hazard ratio, 26.5 [95% CI, 5.4–131]; P < 0.001). Patients with PSA progression at week 12 had a median PSA PFS of 3.5 mo (95% CI, 1.1–4.5 mo), versus 7.9 mo (95% CI, 6.3–10.7 mo) in those without PSA progression. In the 10 patients with SPECT TTV progression at week 12, 50% (5/10) had no concurrent PSA progression (median PSA PFS, 4.5 mo [95% CI, 2.8–5.6 mo]), and 5 of 10 men had both PSA and SPECT TTV progression at week 12 (median PSA PFS, 2.8 mo [95% CI, 1.8–3.7 mo]) (Fig. 2).
Kaplan–Meier curve for PSA PFS stratified by any increase in SPECT TTV at cycle 3 (A) or >30% increase in SPECT TTV at cycle 3 (B). TV = tumor volume.
Univariable Analysis of Clinical and Imaging Markers and Association with PSA PFS and Overall Survival
(A and B) Maximum-intensity projection and quantitation of 177Lu SPECT at cycle 1 (A) and cycle 3 (B) for patient with reduction in SPECT TTV and PSA and PSA PFS of 22 mo. (C and D) Maximum-intensity projection and quantitation of 177Lu SPECT at cycle 1 (C) and cycle 3 (D) for patient with increase in SPECT TTV > 30% but no increase in PSA and PSA PFS of 5 mo.
Reproducibility
TTV was compared between 68Ga PSMA-11 PET/CT at screening (PSMA PET TTV) and the baseline 177Lu SPECT (median time between scans, 15 d [range, 6–56 d]). There was a strong correlation between PSMA PET TTV and cycle 1 SPECT TTV (R = 0.87 [95% CI, 0.74–0.93], P < 0.001) (Fig. 3). Mean TTV was similar between PSMA PET and SPECT (PET TTV, 925 ± 856 cm3; SPECT TTV, 949 ± 852 cm3).
Scatterplot of log(PSMA TTV) at baseline vs. log(SPECT TTV) at cycle 1.
The repeatability of 177Lu SPECT quantitative analysis was assessed in all 32 patients. There was no evidence of a systematic difference between test and retest for SUVmax, SUVmean, or TTV. The repeatability estimate was 0.99 for SUVmax (95% CI, 0.97–0.99), 0.90 for SUVmean (95% CI, 0.81–0.95), and 0.99 for TTV (95% CI, 0.98–0.99) (Table 4).
Repeatability of 177Lu SPECT Quantitation Measures
DISCUSSION
This study found that quantified changes in SPECT TTV between baseline and 12-wk 177Lu-PSMA-617 predict PFS in men treated on a prospective PSMA-targeted therapy trial. To our knowledge, this was the first study to evaluate SPECT parameters for response biomarker capability, a potentially valuable development that uses a readily available tool to potentially enhance personalized treatment by directly assessing treatment response. 177Lu-PSMA-617 has proven an effective therapy for mCRPC, with randomized trials demonstrating both improved overall survival and improved radiographic PFS compared with the standard of care (3), as well as increased PSA 50% response rates and improved patient-reported outcomes compared with cabazitaxel (2). However, responses can be heterogeneous, and PFS remains relatively short (2,3). Combination trials with 177Lu-PSMA-617 are under way to investigate whether combining 177Lu-PSMA-617 with other agents may deepen and prolong responses (NCT04419402, NCT03658447, and NCT03874884) (14). Predictive and interim response biomarkers, both imaging and genomic, will be critical in personalizing treatments to optimize longer-term responses to PSMA-targeted radionuclide therapy (5,7). Although there are limitations in spatial resolution with 177Lu SPECT, its elegant potential as a response biomarker warrants further evaluation. Molecular imaging as an interim response biomarker has been particularly successful in optimizing treatment responses with 12-wk 18F-FDG PET in lymphoma (15–19). More recently, Gafita et al. proposed a 12-wk interim PSMA PET scan on the basis of its predictive value for early disease progression in a multicenter 177Lu-PSMA-617 therapy trial (RECIP 1.0) (5). Being able to identify treatment-resistant phenotypes early could allow either intensification with the addition of synergistic drugs or a change in treatment, thereby maximizing opportunities for treatment response in individuals and avoiding the clinical and financial costs of continuing futile treatment. Generally, this has been done in mCRPC by monitoring serum PSA response (20). Similarly to PSMA, there is heterogeneity of PSA expression, meaning it is not an accurate measure of disease volume in a significant proportion of men with mCRPC (21). In this study, both SPECT TTV and PSA progression were predictive of PFS. However, 21% of those in this study with no PSA progression had SPECT TTV progression. Gafita et al. had a similar finding, with 14% demonstrating PSMA PET progression before PSA progression (5). Larger numbers are required to determine whether use of SPECT TTV in combination with PSA can more effectively identify disease progression, but this study provides strong preliminary evidence, with 177Lu SPECT identifying progression in a subset of patients who had not yet experienced a PSA rise. These results raise the question of whether 177Lu SPECT or PSMA PET should be the preferred interim imaging response biomarker for 177Lu-PSMA-617 therapy.
Posttherapy imaging both with planar imaging and with 177Lu SPECT after radionuclide therapy has traditionally been used for dosimetric calculations to determine the dose to nontarget organs and tumor (22–25). Because of its significantly lower spatial resolution and inability to detect small lesions relative to PET imaging, 177Lu SPECT has not been considered for treatment response. In our direct comparison of quantitative findings between 68Ga-PSMA PET and 177Lu SPECT within 2 wk, TTV was very similar between 177Lu SPECT and PET, with a high correlation between the 2 modalities, although theoretically 177Lu SPECT will underestimate small-volume disease (26). However, evaluating disease progression requiring treatment change or intensification should not depend on identifying small-volume disease. A lesion that is below the spatial resolution for detection on 177Lu SPECT will become visible as its size increases. Although the findings from this trial confirm that 177Lu SPECT has potential for identifying clinically significant disease progression, the opposite may be a more difficult issue. 177Lu SPECT may struggle to confirm complete resolution of all sites of disease in men with exceptional responses to 177Lu-PSMA-617 therapy. Confirmation of an exceptional response may indeed require the spatial resolution of PSMA PET, and further research is required to more precisely define the limitations of 177Lu SPECT and appropriate minimal volume changes required to identify progression.
This study relied on quantitation of 177Lu SPECT data, rather than visual assessment, to determine an increase in TTV. It is becoming clear that assessment of treatment response using PSMA-based imaging for PSMA-targeted therapy must focus on changes in volume rather than measures of intensity (27,28). Accurate visual assessment of changes in tumor volume can be difficult, especially in the presence of large-volume metastatic bone disease. We found that the repeatability of tumor volume on 177Lu SPECT was high and comparable to PSMA PET/CT (29). However, quantitation remains outside routine reporting guidelines and is time-intensive. Further work is needed both to evaluate the accuracy of quantitation over visual assessment and to streamline quantitation to be more user-friendly for integration into routine clinical practice (9).
Changes in SPECT TTV were predictive of progression-free but not overall survival in this study. Although this finding may be due to the small patient cohort, it may also be due to the many patients with progressive disease on 12-wk posttherapy SPECT who were taken off trial and changed to an agent that proved effective. Further work is required to evaluate the benefit of 177Lu SPECT as a prognostic biomarker.
There were several limitations to this study. First, the patient numbers were small, and a larger cohort is needed to validate these findings. This was a single-center study, and 177Lu SPECT quantitative measures can vary significantly between centers and systems (30). Further evaluation is required to harmonize image acquisition and reconstruction across centers and imaging systems for results to be reproducible. Finally, appropriate volume cutoffs for a significant increase in SPECT TTV need to be defined. This will require trials with larger patient numbers and outcome data. However, this study provided a strong foundation on which to build further work.
CONCLUSION
Increasing SPECT TTV on quantitative 177Lu SPECT predicted a short PFS and identified progression in some men who had yet to demonstrate PSA progression. This tool shows promise as an imaging response biomarker.
DISCLOSURE
This investigator-initiated study was sponsored by St. Vincent’s Hospital, Sydney, and supported by a Cancer Institute NSW prostate translational research grant. Noxopharm Limited provided funding for drug and PET scans, and AAA/Novartis provided the PSMA-617 ligand. Louise Emmett has an advisory role with Noxopharm Limited, received grant funding support through St. Vincent’s Clinic Foundation, and received trials support from Novartis and Astellas. Anthony Joshua has an advisory role with Noxopharm Limited and received institutional funding from Novartis. Peter Wilson and Remy Niman are salaried employees of MIM Software, Inc. Sarennya Pathmanandavel received funding through the Cancer Institute NSW and St. Vincent’s Clinic Foundation. No other potential conflict of interest relevant to this article was reported.
KEY POINTS
QUESTION: Do the SPECT images acquired 24 h after 177Lu PSMA-617 therapy provide predictive information on patient outcomes?
PERTINENT FINDINGS: A change in SPECT TTV between dose 1 and dose 3 177Lu PSMA-617 is predictive of PFS.
IMPLICATIONS FOR PATIENT CARE: SPECT images early in treatment have the potential to predict the response to therapy, potentially allowing adjustment of treatment combinations or changes in therapy to improve patient outcomes.
ACKNOWLEDGMENTS
We thank the patients, as well as the clinical trials teams at the Department of Theranostics and Nuclear Medicine and the Kinghorn Cancer Centre, St. Vincent’s Hospital, for their support.
Footnotes
Published online Aug. 25, 2022.
- © 2023 by the Society of Nuclear Medicine and Molecular Imaging.
REFERENCES
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25.
- 26.
- 27.
- 28.
- 29.
- 30.
- Received for publication May 10, 2022.
- Revision received July 22, 2022.