Hematologic dysfunction and myeloid neoplasm risk in patients treated with lutetium-177 prostate-specific antigen membrane therapy

In 2022, the US Food and Drug Administration approved ¹⁷⁷Lu-PSMA-617 (¹⁷⁷Lu-PSMA), a radiotheranostic, for metastatic castration-resistant prostate cancer (mCRPC), following the landmark phase III VISION trial that demonstrated improved overall survival (OS) and radiographic progression-free survival (PFS) compared to best standard of care.¹ While ¹⁷⁷Lu-PSMA has been shown to improve quality of life,² hematologic toxicity is prevalent.^{1,3 177}Lu-PSMA targets cells with PSMA expression leading to β-particle emission and single-stranded DNA breaks. We previously identified therapy-related clonal cytopenia of undetermined significance (tCCUS) and therapy-related myeloid neoplasms (tMN) in 4% of patients with metastatic NET (neuroendocrine tumor) who received ¹⁷⁷Lu-DOTATATE,⁴ a radiotheranostic agent targeting the somatostatin receptor (SSTR). Among prostate cancer patients, the risk of tMN is elevated, with hazard ratios ranging from 2.0-6.5, influenced by treatment exposures.^5,6 However, to date, no studies have evaluated changes to the bone marrow in patients treated with ¹⁷⁷Lu-PSMA. While the current indication for ¹⁷⁷Lu-PSMA is in patients with mCRPC who have exhausted multiple prior lines of treatments (and whose median OS is approximately 15.3 months), ongoing efforts to advance ¹⁷⁷Lu-PSMA to earlier lines of therapy⁷ underscore the need for a more in-depth understanding of potential hematologic risks. Here, we evaluated hematologic dysfunction and tCCUS/ tMN prevalence in mCRPC patients treated with ¹⁷⁷Lu-PSMA. After Mayo Clinic Institutional Review Board approval, we identified adult patients treated at Mayo Clinic (enterprise-wide), with >1 dose of ¹⁷⁷Lu-PSMA between January 1, 2022, and December 31, 2023. Patients with <3-month follow-up were excluded from analyses. Cytopenias were defined per NCI CTCAE v5.0 and MN diagnoses was defined per the revised 4^th-edition World Health Organization criteria (2016).⁸ We evaluated patients who had a bone marrow biopsy (BMB) post-¹⁷⁷Lu-PSMA, focusing on those with assessable BM morphology, cytogenetics and molecular genetics. Standard descriptive statistical methods were used, specific tests are noted in Figure legends. Of 405 patients who received >1 dose of ¹⁷⁷Lu-PSMA, 42% (N=172) had follow-up with laboratories at least 3 months post-177Lu-PSMA, before subsequent cytotoxic cancer-directed therapy (Online Supplementary Figure S1). The median age at first ¹⁷⁷Lu-PSMA was 72.8 years (range, 45.1-93.8). Prior to receiving ¹⁷⁷Lu-PSMA, 92.4% (N=159) of patients had a history of chemotherapy (chemo), with 88.9% (N=153) and 23.2% (N=40) receiving docetaxel and/or carboplatin, respectively. Prior radiation therapy (RT) was present in 86.6% (N=149), with 69.8% (N=120) receiving salvage RT, and 12.8% (N=22) receiving palliative RT to sites of metastases only. Additionally, 7.0% (N=12) had a history of ²²³Radium-dichloride (Xofigo) treatment prior to ¹⁷⁷Lu-PSMA, and 13.3% (N=23) had prior treatment with PARPi (inhibitors). Patients received a median of six doses (range, 2-9) of ¹⁷⁷Lu-PSMA, amounting to a cumulative median exposure of 1,167 mci (range, 193-2867).

At baseline (<1 month prior to 1^st dose of ¹⁷⁷Lu-PSMA), the median hemoglobin (HGB) was 12.1 g/dL (range, 7.9-17.9) and anemia was present in 42.0% (N=68, grade 1-34.6% [N=56], grade2-6.8% [N=11], grade 3-0.6% [N=1], grade 4 - none (Online Supplementary Table S1; Figure 1). Baseline anemia was not associated with age, prior chemo, or prior RT, but was associated with prior PARPi (P=0.011) and Radium-²²³ therapy (P=0.013) (Online Supplementary Table S2). Among those who were not on subsequent cancer-directed therapy at 3 (N=145), 6 (N=111), and 12 (N=44) months post-¹⁷⁷LuPSMA, any grade of anemia was present in 69.7% (N=101; P<0.001), 75.7% (N=84; P<0.001), and 72.7% (N=32; P=0.003), with grade 3 anemia present in 8.3% (N=12), 10.8% (N=12), and 13.6% (N=6), respectively, and none with grade 4 anemia (Figure 1). Of those with HGB values available at all follow-up time points (N=29), persistent anemia was present in 58.6% (N=17) and those who were anemic at baseline were more likely to be anemic at 12-month follow-up (P=0.003; Online Supplementary Table S1; Figure 1). While the median mean corpuscular volume (MCV) was 95.0 fL (range, 74.8-116.7) at baseline and 98.8 fL (range, 80.4-124.4) at 12-months (P=0.004), there were no changes in red cell distribution width (RDW) with time (Online Supplementary Table S1). Thrombocytopenia was present in 10% of patients at baseline (grade 1-7.5% [N=12], grade 2-1.9% [N=3], grade 3-0.6% [N=1], grade 4 - none) with a median platelet (PLT) count of 208x10⁹/L (range, 28-387). Baseline PLT count was not associated with age, prior chemo, RT, or Radium-²²³ use, but was associated with prior PARPi use (P=0.044; Online Supplementary Table S1). At 3 (N=141), 6 (N=108), and 12 (N=41) months post-treatment, among those not on subsequent cancer-directed therapy, thrombocytopenia was present in 36.9% (N=52; P<0.001), 32.4% (N=35; P<0.001) and 24.4% (N=10; P=0.182), respectively. At 3 (N=116), 6 (N=90), and 12 (N=44) months, no patients had grade 4 thrombocytopenia, while grade 3 thrombocytopenia was present in 2.8% (N=4), 7.4% (N=8), and 0% (N=0), respectively. In patients assessable at all three follow-up time points (N=25), BMB, next-generation sequencing (NGS) and cytogenetics were performed in 10.5% (N=18 of 172) of patients, of which 83.3% (N=15) had prior history of chemo (22.2% - platinum), 100% (N=15) had prior RT (salvage - 53.3%, N=8, palliative-46.7%, N=7), and 5.5% (N=1) each had prior Radium-²²³ or PARPi (Table 1). BMB was performed 4.5 years (yrs) (range, 1.3-9.4), 7.9 yrs (range, 0.1-53.1) and 0.9 yrs (range, 0.4-4.7) from first chemo, RT, and ¹⁷⁷Lu-PSMA, respectively. Diagnoses included: tMDS in 50.0% (N=9), tCCUS in 38.9% (N=7), and metastatic carcinoma in 11.1% (N=2). Diagnosis of tMDS or tCCUS was not associated with prior treatments (P=0.83), older age (P=0.52), or total mci exposure (P=0.69).

Figure 1.Hematologic dysfunction in metastatic prostate cancer patients treated with ^*177Lu-PSMA, despite not being on concurrent cytotoxic cancer-directed therapies (other than androgen deprivation therapy). (A) Percent of patients with anemia, thrombocytopenia and neutropenia at baseline, 3, 6, and 12 months after last dose of ¹⁷⁷Lu-PSMA. Pairwise comparisons for each time point compared to baseline were assessed using paired t test for continuous variables and McNemar tests for binary variables. P<0.05 were considered significant. *P<0.05; **P<0.01; ***P<0.001. (B) Median laboratory values at baseline (0-1 month prior to first dose, N=162), 3 (N=145), 6 (N=113), and 12 (N=45) months after last dose. Pairwise comparisons for each time point compared to baseline were assessed using paired t test for continuous variables and McNemar tests for binary variables. P<0.05 were considered significant. *P<0.05; **P<0.01; ***P<0.001. (C) Oncoplot of pathogenic variants among 16 (of 18) assessable patients. Two patients had no variants identified. ANC: absolute neutrophil count; WBC: white blood cell count; HGB: hemoglobin; MCV: mean corpuscular volume; tCCUS: therapy-related clonal cytopenia of undetermined significance; tMDS: therapy-related myeloid neoplasm.

Table 1.Demographic and clinical characteristics of prostate cancer patients undergoing bone marrow biopsy with comprehensive molecular testing after ¹⁷⁷Lu-PSMA Therapy.

NGS revealed 34 somatic pathogenic variants (PV) in 18 patients, with a median of two PV (range, 0-5) per patient, and median variant allele fraction (VAF) of 8.5% (range, 2-82). The most common PV was PPM1D, with 16 PV identified among nine patients (60.0%), followed by TP53, with five PV identified among four patients (22.2%, 4 presumed biallaleic; Figure 1). PV in PPM1D and TP53 were mutually exclusive (P=0.023), and PV in TP53 correlated with complex cytogenetics (P=0.004). Specific PV or VAF were not associated with age (P=0.15), total mci exposure (P=0.23), or prior treatment (P=0.57). Abnormal karyotype was present in 38.9% (N=7), and five patients (27.8%) had a complex karyotype. The most common karyotypic abnormalities included deletion 5q (N=3, 16.7%) and monosomy 7 (N=3, 16.7%). Abnormal karyotype and was not associated with age (P=0.63), total mci exposure (P=0.80) or prior treatments (P=0.66). PV in PPM1D were predominantly found in patients with a normal karyotype (P=0.01), whereas PV in TP53 were found among those with abnormal karyotype (P=0.004). With a median follow-up of 1.6 yrs (range, 0.5-5.9) from first ¹⁷⁷Lu-PSMA, 72.2% (N=13) of those who had a BMB with molecular testing (N=18) were deceased, with 53.8% (N=7) due to progression of disease (Table 1).

To our knowledge, this is the first series to detail tCCUS/tMN in mCRPC patients following ¹⁷⁷Lu-PSMA treatment. Limitations of this retrospective analysis include the confounding effects of prior exposure to cytotoxic chemo and RT, which was present in 83.3% and 100.0%, respectively, making it challenging to determine whether ¹⁷⁷Lu-PSMA was directly or solely responsible. Additionally, the lack of comprehensive hematologic assessments in these patients, which may relfect providers’ perceptions of limited life expectancy or provider difficulty in distinguishing BM disease from treatment toxicity, may underestimate the prevalence of tCCUS/tMN in this population.

Within the VISION trial, 23.4% of patients experienced grade >3 BM suppression and drug-related adverse events leading to death occurred in 0.9%, with etiologies including pancytopenia and BM failure; without tMN diagnosis.¹ Among patients in our cohort who underwent BMB for cytopenias, tCCUS/tMN was identified in 88.9%, with DNA damage response/repair (DDR) pathway mutations present in 72.2% (50.0% with PPM1D, 22.2% with TP53). It is likely that prior exposures to chemo and RT, followed by ¹⁷⁷Lu-PSMA, positively selects hematopoietic progenitor cells with intrinsic resistance mechanisms to apoptosis, in this case, somatic DDR mutations. Over time, these mutant clones result in hematologic dysfunction and tCCUS/tMN. While genomic instability in TP53 mutant clonal hematopoiesis is a well documented pathway towards MN progression, mechanisms of MN progression with PPM1D mutations remain to be ascertained.^9-13 We also highlight the short latency for tCCUS/ tMN, akin to reports in ovarian cancer patients receiving platinum agents and PARPi.¹⁴ While the median survival of mCRPC patients receiving ¹⁷⁷Lu-PSMA is 15.3 months, current attempts to advance ¹⁷⁷Lu-PSMA in earlier stages of disease raise important long-term hematologic safety questions.^1,15-18 These findings underscore the need for further research into mechanisms of clonal evolution and hematologic surveillance strategies in this population.

Footnotes

• Received October 24, 2024
• Accepted March 26, 2025

Correspondence

^#DSC and MMP contributed equally as senior authors

Y. KUSNE - KusneYael@mayo.edu

Disclosures

MMP has received research funding from Stemline Pharmaceuticals, Kura Oncology, Epigenetix, Solu Therapeutics and Polaris Pharmaceuticals. OS is a consultant for ArtBio, Amgen, Astellas, AstraZeneca, Bayer, Clarity, EMD Serono, Fusion, Isotopen, Janssen, MacroGenics, Novartis, Pfizer, Point Biopharma, Ratio, Sanofi, Telix, and TeneoBio; and receives research support from Advanced Acceleratory Applications, Amgen, AstraZeneca, Bayer, In Vitae, Janssen, Lantheus, Merck, Sanofi, and Point Biopharma.

Contributions