Molecular imaging and biochemical response assessment after a single cycle of [225Ac]Ac-PSMA-617/[177Lu]Lu-PSMA-617 tandem therapy in mCRPC patients who have progressed on [177Lu]Lu-PSMA-617 monotherapy

Rationale: Despite the promising results of prostate-specific membrane antigen (PSMA)-targeted 177Lu radioligand therapy in metastatic castration-resistant prostate carcinoma (mCRPC), some patients do not respond and other patients with initially good response develop resistance to this treatment. In this study, we investigated molecular imaging and biochemical responses after a single cycle of [225Ac]Ac-PSMA-617/[177Lu]Lu-PSMA-617 tandem therapy in patients who had progressed on [177Lu]Lu-PSMA-617 monotherapy. Methods: Seventeen patients with mCRPC were included in a retrospective, monocenter study. Molecular imaging-based response was assessed by modified PERCIST criteria using the whole-body total lesion PSMA (TLP) and molecular tumour volume (MTV) derived from [68Ga]Ga-PSMA-11 PET/CT. Biochemical response was evaluated according to PCWG3 criteria using the prostate-specific antigen (PSA) serum value. Concordance and correlation statistics as well as survival analyses were performed. Results: Based on the molecular imaging-based response assessment, 5 (29.4%) patients showed partial remission and 7 (41.2%) had stable disease. The remaining 5 (29.4%) patients had further progression, four with an increase in TLP/MTV of >30% and one with stable TLP/MTV but appearance of new metastases. Based on the biochemical response assessment, 5 (29.4%), 8 (47.1%), and 4 (23.5%) patients showed partial remission, stable disease, and progressive disease, respectively. A comparison of the response assessment methods showed a concordance of 100% (17/17) between TLP and MTV and 70.6% (12/17) between TLP/MTV and PSA. Patients with partial remission, independently assessed by each method, had better overall survival (OS) than patients with either stable or progressive disease. The difference in OS was statistically significant for the molecular imaging response assessment (median OS not reached vs. 8.3 m, p = 0.044), but not for the biochemical response assessment (median OS 18.1 m vs. 9.4 m, p = 0.468). Conclusion: Based on both assessment methods, [225Ac]Ac-PSMA-617/[177Lu]Lu-PSMA-617 tandem therapy is an effective treatment for the highly challenging cohort of patients with mCRPC who have progressed on [177Lu]Lu-PSMA-617 monotherapy. Molecular imaging response and biochemical PSA response were mostly concordant, though a considerable number of cases (29.4%) were discordant. Molecular imaging response reflecting the change in total viable tumour burden appears to be superior to PSA change in estimating survival outcome after tandem therapy.


Introduction
Prostate carcinoma is the second most frequent malignancy in men [1], and a considerable number of patients ultimately progress to a metastatic castration-resistant stage (mCRPC) [2,3]. Apart from chemotherapy with taxanes (docetaxel, cabazitaxel) and next-generation androgen receptor signal inhibitors (abiraterone, enzalutamide) [4][5][6], radioligand therapy targeting prostate-specific membrane antigen (PSMA) is a promising therapy option, currently applied after failure of chemotherapy and hormone therapy [7]. PSMA radioligand therapy (PSMA-RLT) using PSMA ligands labelled with the beta emitter lutetium-177 (e.g. [ 177 Lu]Lu-PSMA-617) has shown encouraging results in various retrospective studies [8][9][10][11] and in a few prospective clinical trials on patients with mCRPC [12,13]. Some patients, however, do not respond to PSMA-RLT with beta emitters and other patients with initially good response develop resistance [14][15][16]. This cohort of patients has limited remaining therapeutic options. Recently, PSMA-RLT with alpha emitters such as actinium-225 e.g. in form of [ 225 Ac]Ac-PSMA-617 monotherapy [17][18][19][20] or [ 225 Ac]Ac-PSMA-617/[ 177 Lu]Lu-PSMA-617 tandem therapy, a concept first introduced by our group [21], has yielded impressive results. PSMA-RLT using alpha emitters may also be effective in patients who have progressed on [ 177 Lu]Lu-PSMA-617 monotherapy due to the higher radiobiological impact of alpha radiation (20 times higher weighting factor than beta radiation) [22][23][24]. However, the higher radiobiological impact of alpha radiation also affects 'organs-at-risk', especially the salivary glands, which may lead to severe xerostomia and impair patients' quality of life. Tandem therapy combining alpha and beta emitters with adapted doses, instead of monotherapy with alpha emitter, may reduce the occurrence of these substantial adverse effects. There is little data, however, about the efficacy of alpha emitters in these highly challenging patients. Besides biochemical response assessment, molecular imaging-based response assessment could be of high importance and interest but has not been reported. In this study, we investigated the efficacy of a single cycle of [ 225 Ac]Ac-PSMA-617/[ 177 Lu]Lu-PSMA-617 tandem therapy in patients who had progressed on [ 177 Lu]Lu-PSMA-617 monotherapy. We assessed molecular imaging response by [ 68 Ga]Ga-PSMA-11 positron emission tomography (PET)/computed tomography (CT) of total tumour burden and biochemical response by serum prostate-specific antigen (PSA) values. Both response assessment methods were evaluated as potential predictors of outcome for this tandem therapy approach.

Study design
We performed a retrospective monocenter study of mCRPC patients treated with one cycle of [ 225 Ac]Ac-PSMA-617/[ 177  tandem therapy in order to assess response by molecular imaging. The study design is depicted schematically in Figure 1. Patients (n=6) who had progressed on [ 177 Lu]Lu-PSMA-617 RLT but with known 'mismatch findings', meaning intense glucose metabolism on [ 18 F]FDG PET/CT with missing or low PSMA expression, were excluded from the tandem therapy. In addition, patients with abnormal hematological blood parameters (hemoglobin < 8 mg/dL, platelets < 75,000 /µL, leukocytes < 3000 /µL), renal insufficiency (estimated glomerular filtration rate (eGFR) < 45 mL/min), or poor overall conditions (Eastern Cooperative Oncology Group (ECOG) performance status ≥ 3) did not receive tandem therapy.

Patients and ethics
Seventeen mCRPC patients were included in this retrospective study. All patients received several pre-treatments. Detailed information about the pre-treatments and patient characteristics are presented in Table 1. The patients received an average of 5 ± 2 cycles (range: 2-9 cycles) of [ 177 [14,25]. Both radioligands were administered simultaneously during an inpatient stay in accordance with German radiation protection regulations. The mean administered activity of [ 225 Ac]Ac-PSMA-617 and [ 177 Lu]Lu-PSMA-617 was 4 ± 2 MBq (range: 1.8-6.9 MBq), corresponding to 44 ± 18 kBq/kg body weight (BW) (range: 19-74 kBq/kg BW), and 6 ± 1 GBq (range: 3.8-8.2 GBq), respectively. Administered activities of 225 Ac and 177 Lu were primarily chosen with regard to total tumour burden, sites of metastases, and patient condition including ECOG performance status, organ function, and degree of existing xerostomia before tandem therapy. For example, patients with high tumour burden or patients who never responded to [ 177 Lu]Lu-PSMA-617 monotherapy received higher 225 Ac activities with high 225 Ac/ 177 Lu ratios (up to a ratio of 1 MBq 225 Ac to 1 GBq 177 Lu). Patients with low tumour burden or unfavorable patient conditions received lower total amounts of 225 Ac and/or lower 225 Ac/ 177 Lu ratios. In order to prevent side effects, each patient received intravenous hydration (1000 mL 0.9% NaCl) 30 min before to 120 min after injection of the radioligand and cooling of the salivary glands. [ 68 Ga]Ga-PSMA-11 was administered intravenously, followed by a 500 mL infusion of 0.9% NaCl. No diuretics were applied. Before administering the radioactivity, blood tests including PSA serum value, creatinine, hemoglobin, leukocytes, and platelets were performed. The mean time between injection and PET acquisition was ~60 min, in accordance with standard procedures for prostate cancer imaging [26]. The PET/CT scans were performed using a Biograph 40 mCT PET/CT (Siemens Medical Solutions, Knoxville, TN, USA) accredited by European Association of Nuclear Medicine (EANM) Research Ltd. PET acquisition was performed from vertex to mid-femur with an acquisition time of 3 min per bed position, which covers a 21.4 cm extended field-of-view (TrueV). The PET datasets were reconstructed using an iterative 3-dimensional ordered subset expectation maximization algorithm (3 iterations; 24 subsets) with gaussian filtering and a slice thickness of 5 mm. Random correction, decay correction, scatter correction, and attenuation correction were applied. CT was performed with a low-dose technique using an X-ray tube voltage of 120 keV and tube current modulated by CARE Dose4D with a maximal tube current-time product of 30 mAs.

Response assessment
Molecular imaging-based response was assessed by calculating the whole-body total lesion PSMA (TLP) and molecular tumour volume (MTV) from [ 68 Ga]Ga-PSMA-11 PET/CT collected before and after [ 225 Ac]Ac-PSMA-617/[ 177 Lu]Lu-PSMA-617 tandem therapy by applying a semi-automatic tumour segmentation using Syngo.Via (Enterprise VB 30, Siemens, Erlangen, Germany). For tumour segmentation, a standard uptake value (SUV) threshold of 3.0 was applied, in accordance with Ferdinandus et al. [27], except for liver metastases (in 2/17 patients), where an SUV threshold of 1.5 × SUVmean of the healthy liver tissue was applied. Physiological uptake in healthy organs was manually excluded. A representative example of the semi-automatic tumour segmentation is presented in Figure 2. TLP was calculated in analogy to total lesion glycolysis (TLG), which is an established parameter in [ 18 F]FDG PET/CT for therapy monitoring [28], as the summed products of volume and uptake (SUV mean ) of all lesions. TLP values are expressed as mL × SUV to distinguish them from MTV values, which are presented in mL. Between PET scans, androgen deprivation therapy (ADT) and therapy with next-generation androgen receptor signal inhibitors such as enzalutamide or abiraterone were continued unchanged to avoid altering PSMA expression [29] (ADT in 17/17, enzalutamide in 13/17, and abiraterone in 1/17 patients). As in other studies [30,31], the PET response criteria in solid tumours (PERCIST) version 1.0 [32] were slightly modified: partial remission was defined as a decrease in TLP or MTV of >30%, progression as an increase of >30% or development of new metastases, and stable disease as a change between ±30%.
Biochemical response was evaluated by the serum PSA value measured on the same day that the PET/CT scans were collected using the Prostate Cancer Working Group 3 criteria (PCWG3) [33]: partial remission was defined as a decrease in PSA of >50%, progression as an increase in PSA of >25%, and stable disease as a change between -50% and +25%. However, these changes were not confirmed by a second measurement as is recommended by the PCWG3.

Toxicity
Hematotoxicity, renal toxicity, and xerostomia were assessed after one cycle of [ 225 Ac]Ac-PSMA-617/[ 177 Lu]Lu-PSMA-617 tandem therapy. Toxicity was recorded using the Common Terminology Criteria for Adverse Events version 5.0 (CTCAE) on blood values of hemoglobin, leucocytes, and platelets, and eGFR. CTCAE grades of xerostomia were assessed based on patient reporting via a questionnaire.

Statistics and survival analysis
Descriptive and concordance analyses were performed on the molecular imaging (TLP and MTV) and biochemical (PSA) response assessments. In addition, several parameters such as patient characteristics, imaging parameters, pretherapeutic parameters, and therapeutic parameters were tested for correlation with changes in TLP, MTV, and PSA. Spearman's correlation was applied for this purpose using Prism 8 (GraphPad Software, San Diego, USA). A p-value of <0.05 was regarded as statistically significant. Overall survival (OS) was defined as the interval from the start of [ 225 Ac]Ac-PSMA-617/ [ 177 Lu]Lu-PSMA-617 tandem therapy to the occurrence of any of the following: (1) death from any cause, (2) commencement of a different treatment such as chemotherapy, or (3) the last study visit. Patients were independently categorized by molecular imaging and biochemical response assessments into two groups: (a) patients with partial remission and (b) patients with stable or progressive disease. The cut-off follow-up date was October 30 th , 2020. Median follow-up was determined, and OS was analyzed using the Kaplan-Meier method.

Molecular imaging response
TLP and MTV responses after one cycle of tandem therapy were similar for each patient. 5/17 patients (29.4%) showed partial remission with a decrease in TLP/MTV of >30%. None of these patients developed new metastases. 8/17 patients showed a change in TLP and MTV between -30% and +30%. However, one of these patients (5.9%) developed new metastases and was classified as having progressive disease. Thus, stable disease was found in 41.2% of patients (7/17). The remaining 4 patients (23.5%) showed disease progression with an increase in TLP/MTV of >30% and three of these patients developed new metastases. The changes in TLP and MTV are visualized as waterfall plots in Figure 3A-B.  Figure 3C.

Concordance analyses
Comparison of the response assessment methods showed a concordance of 100% (17/17) between TLP and MTV and 70.6% (12/17) between TLP/MTV and PSA ( Table 3). Concordant examples of partial remission, stable disease, and progression are shown in Figure 4. Of the discordant patients, two patients had stable disease by molecular imaging but one had partial remission and one had progressive disease by PSA, one patient had partial remission by molecular imaging and stable disease by PSA, and two patients had progressive disease according to molecular imaging but stable disease by PSA ( Table 3). A discordant example is presented in Figure 5.

Correlation with pre-therapeutic parameters
Multiple pre-therapeutic parameters were tested for correlation with changes in TLP, MTV, and PSA.
All pre-therapeutic parameters tested with their corresponding r-and p-values are summarized in Table 4. No significant correlation was noted for any of the parameters tested.

Toxicity
[ 225 Ac]Ac-PSMA-617 and [ 177 Lu]Lu-PSMA-617 were well-tolerated without any serious acute adverse events. No relevant change in eGFR was observed after the single cycle of tandem therapy. One patient had grade 3 thrombocytopenia but no other grade 3/4 hematotoxicities were observed. One patient experienced mild (grade 1) xerostomia attributed to the tandem therapy. In 4 patients with known mild xerostomia prior to tandem therapy, no change in xerostomia grade was recorded. Detailed CTCAE prior to and after tandem therapy are summarized in Table 5. In patients with stable disease or partial remission by both assessment methods, no relevant weight loss and no decrease in overall ECOG performance status was observed. Rather, reduction in pain, as assessed by the visual analogue scale, was noted in 3 patients.

Discussion
This is the first study investigating molecular imaging and biochemical response assessments after a single cycle of [ 225 Ac]Ac-PSMA-617/[ 177 Lu]Lu-PSMA-617 tandem therapy in patients with mCRPC (n = 17) who had progressed on [ 177 Lu]Lu-PSMA-617 monotherapy. In this highly challenging patient cohort, 29.4% (5/17) of patients showed partial remission by either response assessment method after the single cycle of tandem therapy. Our study shows that a PSMA-targeted tandem therapy with alpha and beta emitters can be a successful treatment option for patients with mCRPC and resistance to monotherapy with beta emitters. The results are also in agreement with our previous study [21] in a different cohort of patients who had an insufficient response to monotherapy with [ 177 Lu]Lu-PSMA-617. In the present study, the patients had progressive disease by both [ 68 Ga]Ga-PSMA-11 PET/CT and serum PSA (PSA increase of >25% since last cycle), whereas patients in our previous study were only insufficient responders (defined by a PSA increase or decrease of <50%). These different inclusion criteria may explain why the PSA-based response rate after a median follow-up time of 2 m in this study is lower than that of the previous study (29.4% vs. 50%). All patients were heavily pre-treated with at least one line taxane (if not contraindicated) and next-generation androgen receptor signal inhibitors (enzalutamide and abiraterone), and showed resistance to [ 177 Lu]Lu-PSMA-617 monotherapy. These patients, accordingly, had limited therapeutic options remaining.
Two different methods, molecular imaging assessment using [ 68 Ga]Ga-PSMA-11 PET/CT to calculate TLP and MTV and biochemical assessment of serum PSA, were used to evaluate response. Comparing the TLP and MTV values before tandem therapy, a high correlation (r = 0.99 with p < 0.01) was observed as both reflect the total tumour burden. In assessing response, TLP and MTV provided a concordance of 100%, which is in line with the results of Hartrampf et al. [34]. No correlation between pre-therapeutic TLP or MTV and PSA was noted in our cohort (r = 0.03 with p = 0.90 and r = 0.11 with p = 0.67, respectively). Therefore, molecular imaging and PSA can be considered independent response assessment methods. For most patients, molecular imaging response was in accordance with biochemical response, which is currently the standard response method used in the clinic. A concordance of 70.6% (12/17) between molecular imaging and biochemical methods was noted in our study. Similar concordances of 65-87% have been observed in other studies with different cohorts of patients with prostate cancer [35][36][37]. Since 29.4% of patients had discordant results, a survival analysis was performed to evaluate which of these two methods better estimates the therapy outcome. Patients with partial remission, independently assessed by each method, had better OS than patients with either stable or progressive disease. The difference in OS was statistically significant for the molecular imaging response assessment (median OS not reached vs. 8.3 m, p = 0.044), but not for the biochemical response assessment (median OS 18.1 m vs. 9.4 m, p = 0.468). Accordingly, molecular imaging response can be considered a prognostic value for outcome of this tandem therapy. Despite the very small number of patients, this preliminary data suggests the need for treatment monitoring by [ 68 Ga]Ga-PSMA-11 PET/CT in patients with mCRPC undergoing tandem RLT. Based on our results, the total tumour burden by [ 68 Ga]Ga-PSMA-11 PET/CT (measured as either TLP or MTV) is an important parameter for monitoring PSMA-targeted tandem RLT and may be used to assess other treatments in patients with mCRPC. To confirm our findings, future studies in larger patient cohorts are recommended. No significant correlation was noted between several pre-therapeutic parameters and changes in TLP, MTV, or PSA ( Table  4). No patient, therefore, should be excluded from tandem RLT based on any of these parameters. The statistical analysis may be biased by the small number of patients and further studies are needed.
One cycle of [ 225 Ac]Ac-PSMA-617/[ 177 Lu]Lu-PSMA-617 tandem therapy was safe and well-tolerated. The patient, who experienced grade 3 thrombocytopenia had extensive progression with disseminated bone metastases (the patient is depicted in Figure 4C). This thrombocytopenia seemed to be more related to progression of disease than to the tandem therapy. No other grade 3/4 hematotoxicities were observed. Only one patient experienced mild xerostomia attributed to the cycle of tandem therapy. This low rate of toxicity, especially of xerostomia, may be related to the lower administered activity of 225 Ac (mean: 44 kBq/kg BW) in comparison to other studies applying [ 225 Ac]Ac-PSMA-617 as monotherapy [19,20,22]. For example, Kratochwil et al. recommended 100 kBq/kg BW in [ 225 Ac]Ac-PSMA-617 monotherapy to balance response and side effects [17]. Besides the expected low toxicity profile of the tandem therapy in comparison to 225 Ac monotherapy, we also assume, as discussed in our previous study [21], that 177 Lu may still have a contributing anti-tumour effect when augmented with low activities of 225 Ac in patients somewhat resistant to 177 Lu monotherapy. Alpha radiation even in lower quantities might be able to overcome beta resistance. From our clinical experience, this seems to be true to some extent and was observed in 4/9 patients that were re-challenged [ 177 Lu]Lu-PSMA-617 therapy after tandem therapy (1 patient with partial remission, 3 patients with stable disease); however, this hypothesis remains completely unproven.
Our results provide preliminary evidence and a rational starting point for further studies. Ideally, prospective studies should be conducted to establish the tandem approach in mCRPC management. A dose-finding study that includes evaluation of the absolute activities of 225 Ac and 177 Lu as well as the 225 Ac/ 177 Lu ratio may be one of the next steps to improve response without impairing quality of life. In addition, evaluation of tandem therapy is also conceivable in patients without resistance to 177 Lu monotherapy.

Limitations
The promising results reported herein should be considered in the light of some limitations. The study suffers from its retrospective nature and the limited number of patients. A further limitation is that the time for follow-up varied, which could impact the TLP, MTV, and PSA changes. The majority of patients received their follow-up ([ 68 Ga]Ga-PSMA-11 PET/CT and serum PSA on the same day) about 4 to 8 weeks after the tandem therapy. In a few patients, this time window could not be met due to the patient's condition or organizational reasons. In these few outliers, however, the assessment methods showed concordant results and did not impact the survival analyses. Another major limitation of this study is that no fixed activity protocol was used. The 225 Ac and 177 Lu activities were individually chosen based on tumour and patient characteristics. Furthermore, the limited availability of 225 Ac influenced the determination of the administered activities, resulting in a variable 225 Ac/ 177 Lu ratio. In addition, it has to be pointed out that there are several methods for calculating TLP and MTV [27,35,36]. Even though percentage-based thresholding (41% or 50%) is recommended by the EANM for TLG in [ 18 F]FDG-PET/CT [28], this method is known to be adequate solely for non-heterogenous uptake distributions. In order to not underestimate the lesion volume in case of heterogeneous PSMA expression, which is often observed in disseminated and confluent disease after many therapies, we applied the method published by Ferdinandus et al. [27] with a fixed SUV threshold of 3.0. For calculating liver metastases, we applied a threshold of 1.5 × SUVmean of the healthy liver, which appeared to be appropriate compared to visual findings.

Conclusion
[ 225 Ac]Ac-PSMA-617/[ 177 Lu]Lu-PSMA-617 tandem therapy is an effective treatment for the highly challenging cohort of patients with mCRPC who have progressed on [ 177 Lu]Lu-PSMA-617 monotherapy. Molecular imaging response (based on the [ 68 Ga]Ga-PSMA-11 PET/CT-derived parameters TLP and MTV) and biochemical serum PSA response were mostly concordant, though a considerable number of cases (29.4%) were discordant. Molecular imaging response reflecting the change in viable total tumour burden appears to be superior to PSA change in estimating survival outcome after tandem therapy. Larger, and ideally prospective, studies are recommended to confirm and expand these findings. prostate-specific membrane antigen; PSMA-RLT: PSMA-targeted radioligand therapy; RLT: radioligand therapy; SD: stable disease; TLG: total lesion glycolysis; TLP: total lesion PSMA.