Hematologic safety of 177Lu-PSMA-617 radioligand therapy in patients with metastatic castration-resistant prostate cancer

Background Myelosuppression is a potential dose-limiting factor in radioligand therapy (RLT). This study aims to investigate occurrence, severity and reversibility of hematotoxic adverse events in patients undergoing RLT with 177Lu-PSMA-617 for metastatic castration-resistant prostate cancer (mCRPC). The contribution of pretreatment risk factors and cumulative treatment activity is taken into account specifically. Methods RLT was performed in 140 patients receiving a total of 497 cycles. A mean activity of 6.9 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\pm$$\end{document}± 1.3 GBq 177Lu-PSMA-617 per cycle was administered, and mean cumulative activity was 24.6 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\pm$$\end{document}± 15.9 GBq. Hematological parameters were measured at baseline, prior to each treatment course, 2 to 4 weeks thereafter and throughout follow-up. Toxicity was graded based on Common Terminology Criteria for Adverse Events v5.0. Results Significant (grade ≥ 3) hematologic adverse events occurred in 13 (9.3%) patients, with anemia in 10 (7.1%), leukopenia in 5 (3.6%) and thrombocytopenia in 6 (4.3%). Hematotoxicity was reversible to grade ≤ 2 through a median follow-up of 8 (IQR 9) months in all but two patients who died from disease progression within less than 3 months after RLT. Myelosuppression was significantly more frequent in patients with pre-existing grade 2 cytopenia (OR: 3.50, 95%CI 1.08–11.32, p = 0.04) or high bone tumor burden (disseminated or diffuse based on PROMISE miTNM, OR: 5.08, 95%CI 1.08–23.86, p = 0.04). Previous taxane-based chemotherapy was associated with an increased incidence of significant hematotoxicity (OR: 4.62, 95%CI 1.23–17.28, p = 0.02), while treatment with 223Ra-dichloride, cumulative RLT treatment activity and activity per cycle were not significantly correlated (p = 0.93, 0.33, 0.29). Conclusion Hematologic adverse events after RLT have an acceptable overall incidence and are frequently reversible. High bone tumor burden, previous taxane-based chemotherapy and pretreatment grade 2 cytopenia may be considered as risk factors for developing clinically relevant myelosuppression, whereas cumulative RLT activity and previous 223Ra-dichloride treatment show no significant contribution to incidence rates. Supplementary Information The online version contains supplementary material available at 10.1186/s13550-021-00805-7.

Hematological decline is a frequent occurrence in patients with progressive mCRPC and considered a risk factor for poor outcome. Based on evidence derived from peptide receptor radionuclide therapy (PRRT) in neuroendocrine neoplasias, the risk of myelosuppression has been taken into account as dose-limiting factor also in RLT. While descriptive assessment of myelotoxic events has been included in a number of prospective and retrospective trials [9][10][11][12][13][14][15][16][17][18], their association with potential predisposing factors remains to be elucidated.
Pretreatment factors implicated in the risk of myelosuppression during radionuclide therapy may include preexisting hematologic impairment, previous myelotoxic therapies and bone tumor burden [19]. Irradiation to the bone marrow during RLT can further add to deterioration of hematopoietic function [20]. However, the impact of RLT-specific variables, including administered treatment activity and cumulative activity, has so far not been investigated.
The aim of this study was to examine incidence, severity and reversibility of myelosuppression in patients undergoing RLT with 177 Lu-PSMA-617 in a sizable and heterogenous cohort. Predisposing factors, including previous therapies, disease burden, as well as administered activity per cycle and treatment course, were then analyzed regarding their contribution to new onset hematologic adverse events.

Patients
A total of 140 patients were treated with 177 Lu-PSMA-617 in this retrospective single-center series. Production and administration of 177 Lu-PSMA-617 were performed in accordance with legal regulations set out in the German Drug Registration and Administration Act (AMG § 13 2b). Inclusion criteria for RLT mandated that patients have histologically proven, non-resectable, metastatic prostate cancer with disease progression under standard treatment. Indications were confirmed by an interdisciplinary team including board-certified nuclear medicine physicians, urologists, radiation oncologists, pathologists and oncologists. Sufficient PSMA expression in target lesions was defined as an uptake exceeding the liver uptake on 68 Ga-PSMA-11 PET/CT imaging, i.e., scores 2 and 3 according to EANM standardized reporting guidelines v1.0 [21]. An estimated glomerular filtration rate (eGFR [based on the Chronic Kidney Disease Epidemiology Collaboration equation]) of > 30 mL/ min/1.73 m 2 , hemoglobin ≥ 8.0 g/dL, white blood cells (WBC) ≥ 2.00 × 10 9 /L and platelets ≥ 75 × 10 9 /L were required for treatment initiation. Extent of bone tumor burden on PET/CT imaging was classified based on the Prostate Cancer Molecular Imaging Standardized Evaluation (PROMISE) initiative as well as previous reports and categorized into 1) uni-/oligo-/multifocal (1-20 lesions) or 2) disseminated/diffuse to obtain sufficient group sizes for subsequent analysis [22,23].

Administration
PSMA-617 was obtained from ABX GmbH (Radeberg, Germany), and radiolabeling with 177 LuCl 3 was carried out as described in detail before [10,16]. Quality control was overseen by experienced radiochemists and physicians with respective training in the field. 177 Lu-PSMA-617 was administered by slow intravenous injection over 30-60 s. Infusion of 1000 mL of saline was initiated 30 min before application at a continuous rate of 300 mL/h. With the intention to limit the uptake to the parotid and submandibular, icepacks were locally applied 30 min before therapy and continued for 1 h [24]. All therapies were performed as in-patient procedures at our nuclear medicine therapy ward. As mandated by radiation protection legislation, patients remained hospitalized for a minimum 48 h; median hospitalization was 3 (range 2-5) days per cycle.

Toxicity assessment
Repeat blood tests of hematological parameters (hemoglobin, white blood cells and platelets) were undertaken at baseline, prior to each therapy cycle, 2-4 weeks after each cycle and in 6-12-week intervals throughout followup. Severity of hematologic adverse events was graded based on Common Terminology Criteria for Adverse Events (CTCAE), version 5.0. Grade ≥ 3 toxicities were termed significant.

Statistical analysis
Results are presented as median with interquartile range (IQR) and mean ± standard deviation for continuous variables. Categorical variables are reported as frequencies with respective percentages. The paired Student's t-test was used to compare intraindividual changes in hematologic parameters. Logistic regression analyses were undertaken to explore risk factors relevant for hematological decline. Analysis was carried out per patient (patient-based) and per cycle (cycle-based). Significant hematologic toxicity was defined as an increase in toxicity to grade 3 or higher during the course of RLT and transformed into a dichotomized variable. First, logistic regression analysis was performed for each categorical risk factor. Odds ratios (OR) and 95% confidence intervals (CI) were calculated. Association of hematologic toxicity with continuous baseline variables and administered activity was analyzed using nonparametric rank correlation (Spearman's correlation coefficient denoted with r s ). Statistical analyses were performed with SPSS (version 27.0, IBM, Armonk, NY), and GraphPad Prism (version 9.0.1, GraphPad Software, San Diego, CA) was used to plot graphs. All tests were two-sided with p-values < 0.05 denominating statistical significance.

Results
One hundred forty consecutive patients with mCRPC (median age 72 [IQR 67-78] years) met the eligibility criteria for RLT and underwent treatment at our institution. Patient characteristics at baseline are summarized in Table 1. Upon treatment initiation, 109 (78%) patients had low-grade anemia (85 grade 1, 24 grade 2), 13 (9%) leukopenia (10 grade 1, 3 grade 2) and 15 (11%) thrombocytopenia (13 grade 1, 2 grade 2). Two patients with hemoglobin levels slightly below the inclusion threshold (both 7.7 g/dL) were treated after individual consent and lack of therapeutic alternatives. Patients received a total of 497 cycles of 177 Lu-PSMA-617 with a mean treatment activity of 6.9 ± 1.3 GBq given in a median of 3 (IQR 2-5) treatment cycles. RLT cycles were administered at intervals of 4-8 weeks, reaching a mean cumulative activity of 24.6 ± 15.9 GBq. The median follow-up period was 8 (IQR 4-13) months from the start of treatment.

Course of patients with significant toxicity
Three out of 13 patients with grade ≥ 3 hematologic toxicity spontaneously recovered to lower levels (grade ≤ 2) within 4 to 6 weeks. Nine (69%) patients with significant myelosuppression received transfusion therapy, eight of which were transfused with packed red blood cells and two received platelet concentrates (Table 3). Four (31%) patients could receive additional cycles of RLT either after spontaneous recovery or blood transfusion. Cytopenia was successfully managed in 10 patients. Two patients who experienced significant disease progression following their last cycle died briefly thereafter; one patient was lost to further follow-up. Of the two aforementioned study patients with grade 3 anemia upon treatment initiation, one spontaneously recovered to grade 2 after responding to RLT and one received packed red blood cells throughout the course of RLT and remained at stable grade 2 hemoglobin levels prior to discontinuing RLT due to disease progression after two cycles.

Analysis of predisposing factors for hematologic adverse events
Baseline parameters significantly associated with occurrence of grade  (Fig. 1B). Of laboratory values assessed at baseline, alkaline phosphatase (ALP) was correlated with grade ≥ 3 toxicities (r s = 0.23, p = 0.01) and a moderate inverse correlation of eGFR with the occurrence of grade ≥ 3 thrombopenia was observed (r s = -0.19, p = 0.03). Treatment activity per cycle and administered cumulative activity preceding hematotoxic adverse events showed no significant association with incidence of ≥ grade 3 myelosuppression (p = 0.29, 0.32) (Fig. 2B, C, Table 4).

Discussion
In  Available taxane-based chemotherapeutic agents for progressive mCRPC bear a risk hematotoxicity, especially to white blood cells. The phase 3 TAX 327 trial for docetaxel yielded grade ≥ 3 neutropenia in 32%; the TROPIC trial reported grades ≥ 3 leukopenia in 68% of patients receiving cabazitaxel [3,4,25]. Myelosuppression is also a known side effect in radionuclide therapy [26]. Toxic effects to hematopoietic cells are mediated by both blood-driven recirculating ß-irradiation and scatter radiation from bone metastases. Long-standing experience from peptide receptor radionuclide therapy (PRRT) in neuroendocrine neoplasias (NEN) with 177 Lu-labeled DOTA 0 -Tyr 3 -octreotate ( 177 Lu-DOTA-TATE) yielded moderate grade ≥ 3 hematotoxicity rates in the range of 8 to 11.3% [26][27][28][29]. Beyond the emergence of subacute toxicity, myelodysplastic syndrome (MDS) may develop as a rare, but severe long-term sequel after PRRT in 1-2% of all patients treated [26,30]. Apart from effects attributable to diverging biokinetics, it may be hypothesized that MDS is less likely observed after 177 Lu-PSMA-617 due to shorter survival of patients with mCRPC as compared to NEN.
Initial radioimmunological approaches targeting an extracellular PSMA epitope were limited by high rates of myelotoxicity related to the longer plasma half-life inherent to circulating antibodies [31]. In a phase 2 study with the 177 Lu-labeled monoclonal antibody J591 conducted by Tagawa et al., 47% of all patients developed grade 4 thrombocytopenia necessitating aggressive management and transfusion therapy in 30% of all patients enrolled [32,33]. Following the advent of the small-molecule ligands 177 Lu-PSMA-617 and 177 Lu-PSMA-I&T multiple studies, predominantly within compassionate use programs have included assessment of hematologic adverse events in heterogeneous mCRPC cohorts. Reported overall incidence rates are summarized in Table 5 [9]. It has to be acknowledged that comparison of white blood cell toxicity in our study was impeded by the fact that differential blood counts for neutrophils and lymphocytes were not available for analysis in all our patients. Recently, Barber et al. contributed a comparative retrospective study using both 177 Lu-PSMA-617 and 177 Lu-PSMA I&T in 83 patients previously treated with taxane-based chemotherapy and 84 taxane-naïve controls. Grade ≥ 3 anemia, leukopenia and thrombocytopenia occurred in 8% vs. 1%, 2 vs. 0% and 4% vs. 1% of all study patients [35]. The latter findings indicate an adverse impact of previous taxane-based chemotherapy on subsequent hematotoxcity during RLT, as described in our cohort. In the most recent randomized, multicentric phase 2 trial by Hofman et al. (TheraP, ANZUP 1603) previous treatment with docetaxel was an inclusion criterion. Here, grade ≥ 3 anemia, leukopenia and thrombocytopenia occurred in 8%, 1% and 11% of 98 patients receiving 177 Lu-PSMA-617, as compared to 8%, 1% and 0% in the standard-of-care arm (n = 85) treated with cabazitaxel [44]. We report new onset grade ≥ 3 anemia, leukopenia and thrombopenia in 7% (10/140), 4% (5/140) and 4% (6/140) of patients, respectively. Despite delimiting pre-existing cytopenia from therapy-emergent toxicity, differentiation of hematologic decline due to disease progression from true therapy-emergent toxicity remains challenging due to frequently overlapping phenomena. For a conservative estimate, we considered all new onset grade ≥ 3 toxicities in our analysis, regardless of disease progression being a likely contributing factor in a number of cases. Overall, our results appear well in line with data from the foregoing retrospective and prospective studies, taking into account the significant portion of patients with extensive tumor burden and baseline low-grade myelosuppression in the examined cohort.
Our study points toward an influence of predisposing factors on emergence of grade ≥ 3 hematologic adverse events, including taxane-based chemotherapy and initial grade 2 cytopenia. This may be explained by DNA damage conferred by cytotoxic agents [45]. In addition, sequential failure on multiple systemic treatments preceding RLT puts patients at higher odds of developing hematologic decline through disease progression over time. In their post hoc hematologic safety analysis of the ALSYMPCA trial, Vogelzang et al. report both previous taxane-based chemotherapy with docetaxel and baseline cytopenia (anemia and thrombocytopenia) to be associated with grade ≥ 2 thrombocytopenia in mCRPC patients undergoing 223 Ra-dichloride [19]. Interestingly, the placebo arm also contained relevant rates of new onset toxicity with grade ≥ 3 anemia, neutropenia and thrombocytopenia in 14%, 1% and 3%, underlining the notion that the natural course of mCRPC itself is linked to significant deterioration of bone marrow reserve. In further accordance with our observations in RLT, increased tumor burden (defined as ≥ 6 metastases) was also a predictive factor for hematotoxicity in mCRPC patients receiving 223 Ra-dichloride.
In our cohort, a slight trend toward grade ≥ 3 hematologic toxicities, especially thrombocytopenia, was observed with decreasing eGFR values. This effect has been described also in PRRT and attributed to decreased plasma clearance of recirculating radionuclides in chronic kidney disease [28,46].
Cumulative activity and individual treatment activity play a distinct role in defining appropriate regimens for RLT, and various RLT-schemes have been put forth. Rathke et al. clustered 40 patients into treatment groups receiving 4, 6, 7.4 or 9.3 GBq of 177 Lu-PSMA-617 reporting comparable safety and efficacy, while pointing out a lower mean platelet count in the 10 patients having received 9.3 GBq [39]. Our treatment routine allowed for individual dose adaptation and yielded no correlation between higher treatment activities or high cumulative activities with increased rates of hematologic adverse events. Potential bias must be considered when interpreting the bivariate association of treatment activity and hematotoxic events since myelosuppression was one reason for individual dose de-escalation.
A major limitation to the conducted analysis is undoubtably its retrospective design. The presented patient population is highly heterogenous and may differ from previously reported series, taking into account that both a fraction of patients omitting prior chemotherapy after interdisciplinary counseling and a  considerable number of patients with wide-spread bone tumor burden were included in our analysis. Prospective phase 3 data are much anticipated, with results from the VISION trial expected in near future [47].

Conclusions
Our findings suggest that repeated cycles of RLT with 177 Lu-PSMA-617 can be carried out at acceptable rates of myelosuppression with cytopenia being most frequently reversible, especially in earlier phases of disease progression. High bone tumor burden, previous taxane-based chemotherapy and initial hematologic decline are possible risk factors for developing significant new onset hematologic adverse events. Administered activity per cycle and cumulative activity had in turn no significant impact. These results call for further refining individualized treatment based on given risk factors for hematologic toxicity.