Visual Abstract
Abstract
177Lu-PSMA is an effective treatment in metastatic castration-resistant prostate cancer (mCRPC). Our ability to assess response rates and adjust treatment may be improved using predictive tools. This study aimed to evaluate change in 177Lu-PSMA SPECT quantitative parameters to monitor treatment response. Methods: One hundred twenty-seven men with progressive mCRPC previously treated with androgen-signaling inhibition (99%) and chemotherapy (71%) received a median of 3 (interquartile range [IQR], 2–5) 8-GBq (IQR, 8–8.5 GBq) doses of 177Lu-PSMA-I&T. Imaging included 68Ga-PSMA-11 PET/CT (SUVmax > 15 at a single site and > 10 at all sites > 2 cm), diagnostic CT, and 177Lu SPECT/CT from vertex to mid thigh (24 h after treatment). 177Lu SPECT/CT quantitative analysis was undertaken at cycles 1 (baseline) and 2 (week 6) of treatment. Clinical and biochemical results were assessed to evaluate prostate-specific antigen (PSA) progression-free survival (PFS) and overall survival (OS). Results: A PSA reduction of more than 50% was seen in 58% (74/127). The median PSA PFS was 6.1 mo (95% CI, 5.5–6.7), and OS was 16.8 mo (95% CI, 13.5–20.1). At the time of analysis, 41% (52/127) were deceased. At baseline and week 6, 76% (96/127) had analyzable serial 177Lu SPECT/CT imaging. SPECT total tumor volume (TTV) was reduced between baseline and week 6 in 74% (71/96; median, −193; IQR, −486 to −41). Any increase in SPECT TTV between baseline and week 6 was associated with significantly shorter PSA PFS (hazard ratio, 2.5; 95% CI, 1.5–4.2; P = 0.0008) but not OS. Median PSA PFS in those with an increase in SPECT TTV was 3.7 mo (95% CI, 2.8–6.8), compared with 6.7 mo (95% CI, 5.8–10.6) in those with no increase in SPECT TTV. An increase in SPECT TTV greater than 20% was also associated with PSA PFS (hazard ratio, 1.9; 95% CI, 1.2–3.0; P = 0.008) but less significantly than any change in SPECT TTV. There was a significant difference in PSA PFS between patients with both increased PSA and SPECT TTV and patients with reduced SPECT TTV and PSA (median, 2.8 vs. 9.0 mo; P < 0.0001). Conclusion: Increasing PSMA SPECT TTV on quantitative 177Lu SPECT/CT predicts short progression-free survival and may play a future role as an imaging response biomarker, identifying when to cease or intensify 177Lu-PSMA therapy.
In metastatic castration-resistant prostate cancer (mCRPC), 177Lu-PSMA is an effective therapy, although treatment resistance and short response duration remain common (1–4). The ability to monitor early responses to 177Lu-PSMA therapy may improve patient outcomes by enabling treatment escalation, change in treatment, or a treatment “holiday,” dependent on imaging results. Interim and serial 68Ga-HBEDD-PSMA-11 PET/CT (68Ga-PSMA) have recently been found predictive of progression-free survival (PFS) with prostate-specific membrane antigen (PSMA)–targeted radionuclide therapy (5). Quantitative 177Lu SPECT/CT imaging after each 177Lu-PSMA dose may also be valuable in response monitoring in addition to providing dosimetric information. Increased tumor volume on 177Lu SPECT/CT at dose 3 (week 12) has been shown to predict early disease progression with 177Lu-PSMA therapy (6). This study aimed to determine if quantitative parameters on the week 6 177Lu SPECT imaging 24 h after 177Lu-PSMA therapy predicted treatment response and PFS.
MATERIALS AND METHODS
Men with mCRPC treated clinically with 177Lu-PSMA-I&T were enrolled into a retrospective registry. Enrolled men were treated at a single center with 177Lu-PSMA-I&T at 6-wk intervals until disease progression or treatment was changed. Information on prior treatment, years since diagnosis, and biochemical and hematologic parameters was collected. Patients were followed up after treatment to document prostate-specific antigen progression-free survival (PSA PFS) and overall survival (OS). 177Lu SPECT/CT 24 h after each treatment was undertaken for quantification of disease burden. The institutional review board of St. Vincent’s Hospital approved this retrospective study (Human Research Ethics Committee approval 2022/ETH00924) and waived the requirement to obtain informed consent.
Screening
Men underwent screening 68Ga-PSMA PET/CT; bone scan; and CT of the chest, abdomen, and pelvis. They were eligible if they had an SUVmax greater than 15 on 68Ga-PSMA PET at one or more sites, an SUVmax greater then 10 at all measurable sites of disease not impacted by partial voluming, and no sites of soft-tissue disease on contrast-enhanced CT without corresponding 68Ga-PSMA uptake. Patients with an Eastern Cooperative Oncology Group (ECOG) performance status of 1–3 were eligible for treatment. A minimum estimated glomerular filtration rate of 35 mL/min, hemoglobin of more than 70 g/L, and platelets of more than 70 (109/unit) were required.
Treatment
Men received treatment with 177Lu-PSMA-I&T at 6-wk intervals (median number of doses, 3; interquartile range [IQR], 2–5) between May 2018 and April 2022. A median of 8 GBq (IQR, 7.5–8 GBq) of 177Lu-PSMA-I&T was administered via slow intravenous injection. Dexamethasone, 8 mg orally, was administered on day 1 for each dose to minimize pain flare and vomiting (7,8). Blood work was routinely performed at 3-wk intervals to assess toxicity, adverse events, and biochemical response. Prior treatments and the date of diagnosis were documented, as was PSA PFS and OS. The clinical protocol included 6 doses of 177Lu-PSMA at 6-week intervals, with a multidisciplinary clinical decision dictating the duration of treatment based on evidence of disease progression or exceptional response. Patients were followed up after cessation of treatment to confirm clinical outcomes.
177Lu-PSMA-I&T PSMA-I&T precursor (ABX/Huayi) in sodium acetate buffer was added to non–carrier-added 177Lu-LuCl3 according to the institutional production protocol. Radiochemical purity was determined using high-pressure liquid chromatography and thin-layer chromatography.
Imaging Procedures and Analysis
Screening 68Ga-PSMA PET/CT was performed on all patients before consideration for treatment. All treated patients had 177Lu SPECT/CT (vertex to mid thighs) acquired 24 h after 177Lu-PSMA-I&T injection, with the 24-h time point determined on the basis of the TheraP trial protocol (8). SPECT imaging was undertaken with a Discovery 670 system (GE Healthcare) and a Tandem NM/CT 870 DR (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. For quantitation, the required SPECT calibration was performed on both cameras with a cylindric phantom (for determination of sensitivity factor and conversion from counts to units of activity), and CT attenuation correction was performed using a CIRS CT-data–to–electron-density phantom by MIM Software Inc. For images acquired on the Discovery 670, SPECT projection images were reconstructed with an iterative ordered-subset expectation-maximum algorithm with 4 iterations and 10 subsets using SPECTRA Quant (MIM Software, Inc.). No pre- or post-reconstruction 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. Images acquired on the Tandem NM/CT 870 DR were processed using the same reconstructive parameters on a GE Healthcare SmartConsole for quantitation. Diagnostic contrast CT chest, abdomen and pelvis was undertaken at each treatment cycle to assess for non–PSMA-avid visceral disease progression.
Quantitative Analysis
177Lu SPECT/CT images were analyzed semiquantitatively using MIM (LesionID; MIM Software Inc.) software and a standardized semiautomated workflow to delineate regions of interest with a minimum SUVmax cutoff of 3 and lesion size of at least 0.5 mm. 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 (9).
Statistical Analysis
We measured PSA decline from baseline (≥50% [PSA50]) at any time point; PSA PFS, as defined by Prostate Cancer Working Group 3 (PCWG3) criteria (with PSA progression defined as a PSA rise of ≥ 2); and OS (10,11). The Kaplan–Meier method was used to characterize time-to-event endpoints and to estimate medians (presented with 95% CIs). We correlated changes in SPECT TTV, SPECT PSMA intensity, and clinical and biochemical parameters with time-to-event outcomes, using univariable and multivariable Cox proportional-hazards regression models (12,13). Variables included increase in SPECT TTV, SUVmax, SUVmean, PSA, and radiographic progression. P values below 5% were considered significant. Analyses were performed using R (version 4.0.5).
RESULTS
Patient Characteristics
Between May 2019 and April 2022, 127 men underwent 177Lu-PSMA-I&T therapy. All had mCRPC, 99% (126/127) had received prior androgen-signaling inhibitor (ASI) treatment, and 70% (89/127) had received prior docetaxel. The mean age was 75 y (IQR, 70–80 y) (Table 1). Patients received a median of 3 doses up to a maximum of 10 (IQR, 2–5). Seven percent (9/127) had received prior 177Lu-PSMA-617 on trial. Fifty-eight percent (74/127) had a PSA reduction of more than 50% (PSA50). Median PSA PFS was 6.1 mo (IQR, 4.9–8.4 mo; 95% CI, 5.5–6.7 mo), and median OS was 16.8 mo (IQR, 6.3–14.9 mo; 95% CI, 13.5–20.1 mo). At the time of analysis, 41% (52/127) were deceased.
Patient Characteristics (n = 127)
SPECT Quantitation
Seventy-six percent (96/127) of men had analyzable 177Lu SPECT/CT data at baseline and week 6. 177Lu SPECT/CT quantitation measures at baseline and week 6, including SPECT TTV, SUVmax, and SUVmean, are summarized in Table 2. On the baseline 177Lu SPECT/CT, median SUVmean was 8.8 (IQR, 7–12), median SUVmax was 60 (IQR, 35–88), and median SPECT TTV was 411 cm3 (IQR, 128–1,169 cm3). SPECT TTV was reduced between baseline and week 6 in 74% (71/96; median, −193 cm3; IQR, −486 to −41 cm3) and increased in 25% (24/96; median, 103 cm3; IQR, 42–196 cm3). SUVmax was increased in 24% (23/96), and SUVmean was increased in 22% (21/96).
Baseline Predictive Biomarkers
Patient Outcomes
177Lu SPECT/CT
At baseline, SPECT TTV (dose 1) was not significantly associated with PSA PFS or OS (Table 2). SUVmean measured on dose 1 177Lu SPECT/CT was significantly associated with PSA PFS as a continuous variable (hazard ratio [HR], 0.90; 95% CI, 0.85–0.96; P = 0.0009) but not with OS (HR, 0.94; 95% CI, 0.9–1.0). When stratified by an SUVmean of more than 7 on dose 1 177Lu SPECT/CT, patients with an SUVmean of more than 7 had a median PSA PFS of 6.8 mo (95% CI, 6–9 mo), versus 3.0 mo (95% CI, 2.6–6 mo) in those with an SUVmean of less than 7. An SUVmean of more than 7 was significantly associated with longer PSA PFS (HR, 2.7; 95% CI, 1.6–4.4; P < 0.001) and OS (HR, 2.1; 95% CI, 1–4; P = 0.03).
Any increase in SPECT TTV between baseline and week 6 was associated with significantly shorter PSA PFS (HR, 2.5; 95% CI, 1.5–4.2; P = 0.0008) but not OS. Median PSA PFS in those with an increase in SPECT TTV was 3.7 mo (95% CI, 2.8–6.8 mo), compared with 6.7 mo (95% CI, 5.8–10.6 mo) in those with no increase in SPECT TTV (Fig. 1). An increase in SPECT TTV greater than 20% was also associated with PSA PFS (HR, 1.9; 95% CI, 1.2–3.0; P = 0.008), but less significantly than any change in SPECT TTV. Increases in SUVmax and SUVmean were both associated with PSA PFS (respectively: HR of 1.8 and 95% CI of 1.0–3.1 [P = 0.04] and HR of 2.1 and 95% CI of 1.2–3.8 [P = 0.01]) but not OS (Table 3).
(A and B) Reduction between SPECT TTV from baseline (A) to week 6 (B) in patient with PSA PFS at 14 mo. (C and D) Patient with increased SPECT TTV at 6 wk (C) and PSA PFS at 2.0 mo (D).
Univariable Analysis of Response Biomarkers (Baseline to 6-Week SPECT)
Biochemical
Twenty-three percent (22/96) of patients demonstrated a rise in PSA by week 6. A PSA rise was associated with significantly shorter PSA PFS (HR, 4.0; 95% CI, 2.3–6.9; P < 0.0001) and worse OS (HR, 2.4; 95% CI, 1.1–5.1; P = 0.02). A baseline lactate dehydrogenase level of more than 1.5 times the upper limit of normal was associated with worse PSA PFS (HR, 2.3; 95% CI, 1.3–4.1; P = 0.006), and an alkaline phosphatase level of more than 1.5 times the upper limit of normal was associated with both worse PSA PFS and worse OS (respectively: HR of 1.8 and 95% CI of 1.1–3.0 [P = 0.03] and HR of 2.2 and 95% CI of 1.1–4.5 [P = 0.04]). Baseline hemoglobin was associated with neither PSA PFS nor OS (Table 2).
Combination Biomarkers
Of the 25 patients with SPECT TTV progression at week 6, 44% (11/25) had no concurrent PSA progression (median PSA PFS, 3.7 mo; 95% CI, 2.8–4.7), and 14 men had both PSA and SPECT TTV progression at week 6 (median PSA PFS, 2.8 mo; 95% CI, 1.4–4.7) (Fig. 2). PSA PFS in patients with increased PSA and SPECT TTV differed significantly from that in patients with reduced SPECT TTV and PSA (median, 2.8 mo vs. 9.0 mo; P < 0.0001).
Kaplan–Meier curves for PSA PFS in patients with any increase vs. reduced SPECT TTV (A), rise in PSA at 6 wk vs. reduced PSA at 6 wk (B), and combination of reduced SPECT TTV with reduced PSA, rising PSA with reduced SPECT TTV, rising SPECT TTV with reduced PSA, and rising PSA with rising SPECT TTV (C).
Seventy-one men had reduced SPECT TTV by week 6. Eleven percent (8/71) of men had a reduction in SPECT TTV and PSA progression (median PSA PFS, 2.7 mo; 95% CI, 1.4–4.8). Of these, 2 of 8 had new PSMA-negative hepatic lesions identified on diagnostic CT, and 2 of 8 had new small-volume lesions identified on SPECT, despite a drop in SPECT TTV.
SPECT Multivariable Analysis
Both a baseline SPECT SUVmean of more than 7 and a change in SPECT TTV were found to be independently predictive for PSA PFS, whereas a change in SUVmean or SUVmax was not (Table 4).
Multivariable Analysis of SPECT Parameter for PSA PFS and OS
DISCUSSION
This study demonstrated that a change in tumor volume on 177Lu SPECT/CT between baseline and 6 wk of 177Lu-PSMA-617 therapy is predictive of short PFS. Furthermore, the combination of increased SPECT TTV and a PSA rise at 6 wk identified a subgroup of men at high risk of a poor response to 177Lu SPECT/CT therapy who may benefit from either a change in therapy or the addition of combination treatments that may have a synergistic benefit in conjunction with 177Lu-PSMA. Identifying effective response biomarker combinations such as early PSA rise and SPECT TTV, which provide strong information as early as 6 wk into treatment, is a big step toward being able to tailor treatment strategies to individual patients, thereby improving outcomes.
177Lu-PSMA has proven an effective therapy for mCRPC, with randomized trials demonstrating improved OS and high PSA response rates compared with standard-of-care therapies (2,3). However, responses can be heterogeneous, and a proportion of men with suitable 68Ga-PSMA PET screening reults may have limited treatment responses. At the same time, combination trials with 177Lu-PSMA are under way to investigate whether combining 177Lu-PSMA with other agents may deepen and prolong responses (NCT04419402, NCT03658447, NCT03874884) (7,14). Imaging biochemical and clinical interim response biomarkers will be critical in personalizing treatments to optimize longer-term responses to PSMA-targeted radionuclide therapy and, conversely, in stopping treatment early to mitigate opportunity cost if other treatment options are available.
The response evaluation criteria in PSMA PET/CT (RECIP) have recently been proposed for response assessment using a 12-wk PSMA PET scan (5). RECIP uses PET/CT quantification to derive volume and intensity scores applying a combination of a 20% increase in PSMA PET TTV and new lesions to determine disease progression. A recent analysis of the LUPIN trial has found that 177Lu SPECT/CT at week 12 is also predictive of treatment response (6). This study has confirmed this predictive value, also finding that 177Lu SPECT/CT predicts PFS earlier in treatment (6 wk) without the need for an additional 68Ga-PSMA PET scan. This finding has significant cost and availability advantages worldwide, with few countries having approved 68Ga-PSMA PET/CT for response assessment. Additionally, the current analysis found that only 4 of 96 patients had new lesions not identified by an increase in SPECT TTV, 2 of 4 of these being new PSMA-negative hepatic lesions not evident on PSMA SPECT/CT. More work is required to determine whether the presence of new lesions on PSMA imaging should be a requirement in classifying disease progression or whether increased tumor volume in conjunction with biochemical parameters is sufficient.
Currently accepted response biomarkers in mCRPC include a sequential rise in PSA (15) and RECIST/PCWG3 criteria progression on diagnostic CT and bone scan (16). Heterogeneity of PSA expression in mCRPC may limit its predictive value in a proportion of men (17). This possibility is reflected by the fact that 25% of the men in this study who had an increase in SPECT TTV did not have a concurrent PSA rise at 6 wk. This percentage is similar to the 21% identified in a previous SPECT TTV study and the 14% demonstrating 68Ga-PSMA PET progression before PSA progression by Gafita et al. (5,6). RECIST progression requires serial imaging to determine progression using PCWG3 criteria, limiting its value as an early response biomarker for treatment adjustment. Ongoing evaluation is required to determine whether response evaluation criteria should be modified to include 177Lu SPECT/CT, although diagnostic CT will remain important in identifying PSMA-negative progression.
Screening PSMA and 18F-FDG PET parameters have demonstrated strong prognostic value for 177Lu-PSMA therapy (18,19). In the TheraP trial, patients with a high PSMA SUVmean of more than 10 have an excellent PSA50 response rate (18). SUVmean is an indirect measure of PSMA expression heterogeneity, which impacts treatment effectiveness but cannot assess radiation sensitivity. However, change in SPECT TTV may provide individual information on radiation sensitivity, measuring tumor volume reduction after treatment. This study found that both the screening SUVmean (SPECT) and the change in SPECT TTV were independently predictive of PSA PFS, raising the possibility of developing effective imaging response nomograms as early as 6 wk into treatment.
Previous work has identified baseline PSMA and biochemical parameters that predict early treatment failure for patients undergoing 177Lu-PSMA therapy (13). A novel component of our study is the evaluation of baseline PSMA SPECT parameters in addition to the change in SPECT between baseline and after 6 wk of therapy. This study demonstrates that baseline PSMA predictive biomarkers can be derived from SPECT in addition to baseline PET imaging. This ability may be valuable when PSMA PET is not widely available.
This study relied on quantitation of SPECT data rather than visual assessment. This reliance on quantitation for effective predictive and response biomarkers in molecular imaging is increasing, including the highly valuable SUVmean on screening 68Ga-PSMA PET for 177Lu-PSMA therapy (18–20). However, quantitation is not the standard of care and is time-intensive. Further work is needed to streamline quantitation for widespread adoption into routine practice (9).
There were several limitations to this study, a single-center retrospective analysis of a clinical treatment program. Although PSA and survival data were rigorously collected, obtaining routine RECIST/PCWG3 bone scan criteria for radiographic progression was not possible. Additionally, 177Lu SPECT/CT quantitative measures can vary significantly between centers and systems, and these findings require validation in other clinical databases and quantitative programs (21). This study evaluated only the first 2 SPECT data time points quantitatively, and examination of subsequent time points could provide more comprehensive information on the value of 177Lu SPECT/CT. Finally, further research with larger patient numbers and outcome data is necessary to better define appropriate volume cutoffs for a significant increase in SPECT TTV that should be used to identify disease progression.
CONCLUSION
Increasing PSMA-SPECT-TTV on quantitative 177Lu SPECT/CT predicts short PFS and may play a future role as an imaging response biomarker, identifying when to cease or intensify 177Lu-PSMA therapy.
DISCLOSURE
This investigator-initiated study was sponsored by St. Vincent’s Hospital Sydney and supported by a Cancer Institute NSW prostate translational research grant. Sarennya Pathmanandavel received funding through the Cancer Institute New South Wales and St. Vincent’s Clinic Foundation. Louise Emmett has an advisory role with Clarity Pharmaceuticals; receives trial support from Novartis and Astellas; and receives grant funding support from St. Vincent’s Clinic Foundation. Anthony Joshua has an advisory role with and receives institutional funding from Novartis. Peter Wilson and Remy Niman are salaried employees of MIM Software, Inc. No other potential conflict of interest relevant to this article was reported.
KEY POINTS
QUESTION: Is there value in SPECT imaging of patients after 177Lu-PSMA therapy?
PERTINENT FINDINGS: A change in tumor volume on 177Lu SPECT/CT at week 6 is predictive of short PFS and may have potential as a response biomarker.
IMPLICATIONS FOR PATIENT CARE: 177Lu SPECT/CT has potential as an imaging response biomarker and may assist in the management of men with mCRPC undergoing 177Lu-PSMA therapy.
ACKNOWLEDGMENTS
We thank the patients and the clinical staff at the Department of Theranostics and Nuclear Medicine for their support.
Footnotes
Published online Sep. 8, 2022.
- © 2023 by the Society of Nuclear Medicine and Molecular Imaging.
REFERENCES
- Received for publication July 16, 2022.
- Revision received September 1, 2022.