Clinical Impact and Dynamics of Clonal Hematopoiesis with ¹⁷⁷Lu-PSMA-617 Therapy in Advanced Prostate Cancer

Visual Abstract

Abstract

The prevalence, clinical impact, and clonal dynamics of clonal hematopoiesis (CH) in patients receiving ¹⁷⁷Lu-PSMA-617 (LuPSMA) for metastatic castration-resistant prostate cancer (mCRPC) are unknown. Methods: Targeted next-generation sequencing of 21 genes recurrently mutated in CH was performed on DNA extracted from the peripheral blood of patients who received at least 4 cycles of LuPSMA for mCRPC at our institution between 2022 and 2023. Pathogenic somatic mutations with a variant allele fraction of at least 1% were identified using a standardized pipeline. Clinical outcomes pertaining to efficacy (overall survival [OS], measured from date of planned cycle 5) and hematologic toxicity of LuPSMA were collected from the electronic medical record. Results: Fifty patients treated with LuPSMA were eligible, with a median follow-up of 23 mo. At least 1 CH variant was detected in 33 patients (66%). The most common mutations were TET2 (n = 16), PPM1D (n = 15), and DNMT3A (n = 6). OS was similar in patients with or without CH (12-mo OS, 92% vs. 80%; hazard ratio, 0.89; 95% CI, 0.3–2.68). There was a trend toward greater hematologic toxicity in patients with CH, with a greater need for growth factor support (12% vs. 0%). In patients with serial samples available, the emergence of new clones or expansion of preexisting CH variants was detected in most patients, particularly with PPM1D and TP53-mutant clones. The key limitation was the small sample size and short follow-up. Conclusion: CH was highly prevalent and tended to lead to greater hematologic toxicity in patients with mCRPC receiving LuPSMA. Expansion or emergence of DNA damage repair CH clones was very common during and after LuPSMA therapy. Further study of the impact of CH on radiopharmaceutical therapy, particularly when used in earlier prostate cancer disease settings, is warranted.

Treatment with ¹⁷⁷Lu-PSMA-617 (LuPSMA) has proven effective for metastatic castration-resistant prostate cancer (mCRPC) both before and after chemotherapy. Hematologic toxicity with LuPSMA in mCRPC is not uncommon, with grade 3 or higher anemia, thrombocytopenia, or leukopenia seen in approximately 5% and 10% of participants in the PSMAfore (1) and VISION (2) trials receiving LuPSMA before and after chemotherapy, respectively. The incidence of longer-term hematologic toxicity, specifically therapy-related myeloid neoplasms (t-MNs; myelodysplasia and acute myeloid leukemia) has yet to be reported from these trials, likely because of the relatively short follow-up (median, <36 mo) in these studies and the competing risks of prostate cancer death in advanced mCRPC. A recent single-institution case series reported the development of a t-MN during or after LuPSMA in approximately 1% of patients with mCRPC (3). There is concern that the use of LuPSMA in early-stage prostate cancer, where survival may be several years and in which trials are ongoing, may uncover a higher rate of late hematologic toxicities, such as t-MNs.

Clonal hematopoiesis (CH) is characterized by the expansion of hematopoietic stem and progenitor cells driven by somatic mutations in leukemia driver genes (4). The prevalence of CH increases with age. Further, CH is associated with decreased overall survival (OS) because of an increased risk of inflammatory conditions, such as coronary heart disease and hematologic malignancies, particularly myelodysplasia and acute myeloid leukemia. CH is commonly found in patients with solid tumors, including those with prostate cancer, and the prevalence, clonal architecture, and spectrum of mutations are shaped by the type and duration of cytotoxic anticancer therapy (5). Patients exposed to chemotherapy, especially to multiple lines or for an extended duration, have higher rates of CH, with mutations predominantly appearing in genes involved in regulation of DNA damage response (DDR-CH) (e.g., TP53, PPM1D, CHEK2, SRCAP) (6,7). Expansion of these clones during and after exposure to cytotoxic therapy may contribute to the development of cytopenias and t-MNs (8). Despite the observation of cytopenias after LuPSMA therapy, published data on the prevalence of CH and clinical outcomes for patients receiving radiopharmaceutical therapies are limited, with 1 study suggesting a high proportion of patients with new CH mutations at the time of progression during or after treatment with LuPSMA (9).

On the basis of these considerations, we performed a single-institution analysis of the prevalence and clonal dynamics of CH in patients receiving standard-of-care LuPSMA in routine practice and evaluated the impact of CH on therapeutic efficacy and toxicity. Additionally, we compared the evolution and emergence of CH in patients treated with LuPSMA to patients receiving other prostate cancer therapies.

MATERIALS AND METHODS

Patient Cohort

This was a single-center study of patients with mCRPC who received 4 or more cycles of standard-of-care LuPSMA at our institution between June 2022 and October 2023 and who had consented for biobanking, with whole blood available at therapy initiation (i.e., day 1 of cycle 1) as well on the first days of cycle 3 and planned cycle 5 of LuPSMA therapy. Details of cohort selection are shown in Supplemental Figure 1, available at http://jnm.snmjournals.org. Since there were no data on CH in patients receiving LuPSMA at the time of study design, we evaluated CH in patients who had received at least 4 cycles of LuPSMA, reasoning that patients with CH after 4 cycles of treatment would have evidence of CH at earlier timepoints, which would permit us to study CH clonal dynamics before, during, and after LuPSMA therapy.

We additionally obtained paired pretherapy and posttherapy blood from patients with mCRPC who received standard-of-care taxane chemotherapy (docetaxel and cabazitaxel) and ²²³Ra at our institution between 2018 and 2023 and who had consented for biobanking. The study received approval from the Dana-Farber/Harvard Cancer Center Institutional Review Board (IRB 21-629 and 01-045).

CH Mutational Analysis

DNA was extracted from whole blood and analyzed using a targeted next-generation sequencing platform covering 21 genes recurrently mutated in CH. Details on the sequencing platform, performance, and targeted regions have been published (10). Deduplication, alignment, filtering, and variant calling was performed using a custom pipeline, as previously described, to identify mutations with a variant allele frequency (VAF) of 0.01 or greater (11,12). Cluster amplification and sequencing were performed on an Illumina NovoSeq to a mean deduplicated sequencing depth of 3109x. After initial alignment of raw reads, we called duplex consensus reads using fgbio and realign using bwa-mem. Consensus reads were required to have reads from both families and consensus reads with n’s in excess of 5% of bases were discarded. Reads 1 and 2 were soft-clipped from the 5′ end by 10 bases to reduce errors generated during end repair. Single nucleotide and small insertion and deletion calling were performed with samtools-0.1.18 mpileup and Varscan 2.2.3. Variants were excluded if they had fewer than 5 supporting alternate reads. Potential germline variants were excluded by filtering those with a VAF exceeding 0.35 that were not frameshift, nonsense, or known recurrent hotspot mutations.

Treatments and Outcomes

LuPSMA was administered per standard-of-care at a dose of 7.4 GBq every 6 wk for a maximum of 6 cycles, with dose reductions and delays on the basis of the clinical judgment of the treating medical oncologist and nuclear medicine physician. One patient received a seventh cycle of therapy a few months after completing cycle 6. Laboratory evaluations were conducted on day 1 of every cycle, and use of transfusions or growth factor support was determined by the clinical judgment of the treatment team. Clinical details pertaining to toxicity and efficacy were abstracted from the electronic medical record. PSA50 was defined as a reduction of at least 50% in PSA levels from cycle 1. OS was calculated from the date of (planned) cycle 5 of therapy.

Statistical Analysis

Patients’ CH status was determined by the presence or absence of CH mutations after 4 treatment cycles. Patient information and prior treatments were summarized for all patients and by CH status. The median follow-up time from cycle 1 for the overall cohort was estimated using the reverse Kaplan–Meier method. Hematologic parameters were compared between patients with and without CH using Wilcoxon rank-sum test in univariable analyses. ANCOVA was performed in multivariable models, adjusting for age at cycle 1 as a continuous covariate. Cox proportional hazards models were used to examine the association of CH status with OS, where OS was calculated from cycle 5. VAF was summarized by patient gene, with VAF and fold change calculated relative to values at cycle 1. Changes before and after therapy were visualized separately for patients treated with taxane chemotherapy and ²²³Ra. Statistical analyses were performed using SAS version 9.4 (SAS Institute Inc.).

RESULTS

Fifty patients who received at least 4 cycles of LuPSMA and had available biobanked blood were eligible for study inclusion, with a median age of 73 y (interquartile range [IQR], 52–87 y). The median follow-up time was 23 mo (IQR, 20–26 mo). All patients received at least 1 prior taxane chemotherapy and an androgen receptor pathway inhibitor. The median number of LuPSMA cycles received was 6 (range, 4–7) (Table 1). Overall, at least 1 CH variant was detected in 33 patients (66%), and patient characteristics at initiation of LuPSMA, stratified by the presence or absence of a CH variant, are shown in Supplemental Table 1. Patients with CH were generally older (median age, 74 y vs. 66 y). A similar proportion of patients with and without CH received prior pelvic radiotherapy (28 [85%] vs. 15 [88%], respectively) and more than 1 taxane chemotherapy (15 [46%] vs. 8 [47%], respectively). The 44 unique CH variants detected in the cohort are shown in Figure 1. The most common variants seen were PPM1D (n = 15, 30%), TET2 (n = 8, 16%), DNMT3A (n = 6, 12%), and TP53 (n = 4, 8%). All identified CH mutations in the 33 patients with CH are listed in Supplemental Table 2.

TABLE 1.

Patient Characteristics (n = 50)

Characteristic	Value
Age at cycle 1 of LuPSMA (y)*	73 (68–76)
Race (%)
White	44 (88)
Black	1 (2)
Hispanic or Latino	4 (8)
Asian	1 (2)
Site of metastasis
Bone	45 (90)
Lymph node	37 (74)
Liver	4 (8)
Lungs	12 (24)
Prior therapies received
1 ARPI	19 (38)
≥2 ARPIs	31 (62)
1 taxane	27 (54)
≥2 taxanes	23 (46)
223Ra	5 (10)
Pelvic radiotherapy	43 (86)
PSA at cycle 1 of LuPSMA (ng/mL)*	36 (6–156)
PSA50 response at cycle 5	40 (87)

^*Expressed as median and IQR.

ARPI = androgen receptor pathway inhibitor.

Qualitative data are expressed as number and percentage.

FIGURE 1.

Oncoprint of CH mutations detected in post-cycle 4 samples in 33 patients with at least 1 CH variant at this time point.

Impact of CH on Efficacy and Safety of LuPSMA

PSA50 rates up to the date of planned cycle 5 were similar between patients with (26/33, 87%) and without (14/17, 88%) CH. Landmark 12-mo OS after cycle 5 was also similar between patients with CH (2 patients excluded as they completed only 4 cycles) and without CH (92%; 95% CI, 72–98] vs. 80%, 95% CI, 50–93; hazard ratio, 0.89; 95% CI, 0.3–2.68) (Supplemental Fig. 2).

Overall, changes in hematologic parameters (hemoglobin, white blood cell count, absolute neutrophil count, platelets) between LuPSMA cycles 1–6 did not significantly differ between patients with and without CH (adjusted P ≥ 0.1) (Fig. 2; Supplemental Table 3). The median percentage change in platelet count from cycle 1 to cycle 6 was greater in patients with CH (−21.1%; IQR, −34.2% to −12.3% vs. −6.4%; IQR, −19.2% to −1.3%], P = 0.05), but this was not statistically different when adjusted for age (P = 0.13). Table 2 highlights the impact of CH on clinically significant hematologic toxicity during therapy. Although a similar proportion of patients with and without CH experienced a dose delay (6/33, 18%] vs. 4/17, 24%) or dose reduction (5/33, 15% vs. 2/17, 12%) during therapy, patients with CH had a numerically higher requirement for growth factor support (4 patients vs. 0 patients without CH), although the proportion of patients needing blood product transfusion was fairly similar (3 [9%] with CH vs. 1 [6%] without CH). No patient developed a secondary hematologic malignancy or t-MN during follow-up.

FIGURE 2.

Changes in hemoglobin levels (A), platelet count (B), white blood cell count (C), and absolute neutrophil count (D) during LuPSMA therapy in patients with and without CH (CH+ and CH−, respectively). Red boxplots denote CH+ patients, and blue boxplots denote CH− patients.

TABLE 2.

Summary of Hematologic Toxicity During LuPSMA Therapy

Variable	CH-negative (n = 17)	CH-positive (n = 33)
Dose delay	4 (24)	6 (18)
Dose reduction	2 (12)	5 (15)
Any blood product transfusion	1 (6)	3 (9)
Red blood cells	1 (6)	3 (9)
Platelets	0	2 (6)
Receipt of any growth factor	0	4 (12)
Erythropoietin	0	0
Romiplostim	0	3 (9)
Eltrombopag	0	1 (3)

Data expressed as number and percentage.

Evolution of CH During LuPSMA Therapy

Serial samples before, during, and after LuPSMA therapy were available in 9 patients with CH (Supplemental Table 4). Changes in the VAF of CH variants from these patients are shown in Figure 3. New CH clones (mutations in PPM1D, CHEK2, U2AF1, and TET2) were detected in 5 patients (56%) after initiation of LuPSMA. There was a general increase in the VAF of most CH variants present at the time of LuPSMA initiation, particularly in DDR-CH (TP53 and PPM1D). Fold changes in the VAF during therapy in 5 patients with CH (noted on their cycle 1 sample) are shown in Supplemental Figure 3, with significant increases in VAF (>3-fold) seen in PPM1D and TP53 mutant clones in 1 patient.

FIGURE 3.

Changes in VAF of CH clones before, during, and after LuPSMA therapy in 9 patients with serial samples available. Red asterisks denote new CH mutations. C1D1 = cycle 1 day 1.

To compare CH evolution during LuPSMA with other cytotoxic mCRPC therapies, we evaluated paired pretherapy and posttherapy samples from 21 patients with mCRPC receiving taxane chemotherapy (median age, 69 y; IQR, 51–88 y; median time between samples, 4.1 mo; range, 1.1–14.5 mo) and 22 patients receiving ²²³Ra (median age, 75 y; IQR, 61–91 y; median time between samples, 5.8 mo; range, 3.0–9.2 mo). Among these cohorts, 12 and 14 patients had CH detected in either their pretherapy or posttherapy sample, respectively (Supplemental Table 5). VAF changes between pretherapy and posttherapy samples are shown in Supplemental Figure 4. The emergence of new CH variants, most of which were TP53- and PPM1D-mutant clones, was observed in 5 patients (24%) after taxane chemotherapy and in 3 patients (14%) after treatment with ²²³Ra. In contrast with the LuPSMA cohort, the VAFs of CH variants appeared relatively stable in most of the samples of patients before and after therapy with either taxane chemotherapy or ²²³Ra.

DISCUSSION

In this cohort of 50 patients with heavily pretreated mCRPC receiving at least 4 cycles of LuPSMA in routine practice, we noted a high overall prevalence of CH. Although the presence of CH was not significantly associated with the clinical efficacy of LuPSMA, there was some evidence of a trend toward greater hematologic toxicity during therapy in patients with CH who completed at least 4 cycles of therapy, with a numerically higher requirement for growth factor support and greater reduction in platelet count during therapy. Furthermore, in an exploratory analysis of the small subset of patients with serial samples available, we observed the emergence of new DDR-CH and expansion of existing DDR-CH clones in most LuPSMA-exposed patients. To our knowledge, this is among the first studies evaluating the prevalence, clinical impact, and clonal evolution of CH during LuPSMA therapy. Our findings provide a potential mechanistic rationale for the development of t-MNs after LuPSMA and other radiopharmaceutical therapies. This is particularly notable at a time when LuPSMA is being used earlier in the treatment of prostate cancer, where survival may be several years and patients may be more likely to experience long-term hematologic toxicity, including t-MNs.

The relationship between CH and prostate cancer outcomes and therapies is being increasingly recognized. Given its association with age, CH is highly prevalent in patients with prostate cancer, with 24% of patients with mCRPC receiving first-line androgen receptor pathway inhibitor in a randomized phase 3 trial (Alliance A031201) harboring at least 1 CH clone at baseline (13). The higher prevalence of 66% in our study, with a mutational spectrum skewed toward DDR-CH genes, is consistent with other reports focusing on more heavily pretreated patients (8,9) and could be explained by exposure to taxane chemotherapy and pelvic radiotherapy, as well as differences in VAF cutoffs (1% vs. 2%) and the larger CH gene panel used in our study. Although larger studies evaluating a range of solid tumors have suggested poorer clinical and survival outcomes among patients with CH (14), our finding that patients with CH receiving LuPSMA had a similar OS to those without CH is consistent with results from other prostate cancer studies. CH did not impact outcomes in the Alliance A031201 trial (13), and no significant associations between CH and survival have been observed in institutional reports of patients with metastatic prostate cancer (15,16). One likely explanation for this is that the follow-up in all of these studies was relatively short, with significant competing risks from death from prostate cancer. It is possible that any adverse impact of CH on cancer-specific outcomes and OS would only be observed in clinical scenarios with longer anticipated survival times.

Our finding that patients with CH tended to experience more hematologic toxicity during LuPSMA therapy is consistent with CH predisposing patients to clonal cytopenias (4). Cytopenias, particularly thrombocytopenia, are a recognized adverse effect of LuPSMA, and we noted a potential trend toward a greater reduction in platelet count during therapy in patients with CH, although this was not statistically significant after adjusting for age (1,2). We also noted a numerically higher requirement for growth factor support in patients with CH, consistent with recent work showing that patients with CH were more predisposed to severe anemia after PSMA-targeted radiopharmaceutical therapy (17). It is also worth emphasizing that only patients who received at least 4 cycles of therapy were included in our cohort; it is very possible that patients with CH may have discontinued therapy because of toxicity before cycle 4, and our cohort selection would have missed this and may therefore underestimate the associations between CH and hematologic toxicity. Nevertheless, our observations suggest that knowledge of CH status in patients receiving LuPSMA may help guide closer hematologic monitoring and early initiation of growth factor support in this population.

Prior studies have reported an increase in CH clone size and emergence of new CH variants after cytotoxic and radiopharmaceutical therapy for solid tumors, particularly involving DDR-CH (i.e., PPM1D and TP53) (6,18). Mutations in these genes, especially TP53, are some of the most likely to promote progression to t-MNs (19–21). Data on the impact of therapies used in advanced prostate cancer on CH are relatively sparse, with a report involving 10 patients receiving a poly(adenosine diphosphate–ribose) polymerase inhibitor for either recurrent or metastatic prostate cancer showing a new CH clone (in PPM1D) developing in 1 patient and mixed changes occurring after therapy in preexisting CH clones (8). A recent report evaluated CH in blood samples taken at baseline and disease progression in patients receiving LuPSMA or cabazitaxel in the THERA-P trial, noting significantly higher odds of treatment-emergent clones (particularly in PPM1D) and more frequent expansion of CH clones between timepoints in the LuPSMA arm (9).

Our data build on these prior observations by confirming that LuPSMA is associated with significant expansion of preexisting CH clones or the development of new CH clones, primarily involving mutations in PPM1D and TP53. Furthermore, we noted that DDR-CH clonal expansion and emergence occurred to a greater degree in patients treated with LuPSMA compared with those receiving ²²³Ra or taxane chemotherapy, whereas CH expansion occurred least frequently in patients receiving taxane chemotherapy for mCRPC, consistent with prior reports on the limited impact of taxane chemotherapy on CH (6). Although these data are derived from relatively small sample sizes and subject to the inherent biases of a retrospective observational cohort, they highlight the potential impact of radiopharmaceuticals in driving DDR-CH clonal development and expansion. Although the clinical implications of these findings in heavily pretreated patients are limited because of their median OS of 1–2 y, this may have greater significance as LuPSMA is used earlier in the treatment of prostate cancer. The PSMA-DC (NCT05939414) (22) and PSMAddition (NCT04720157) trials (23) are evaluating LuPSMA in nonmetastatic and metastatic hormone-sensitive prostate cancer, respectively, where survival may be 5–10 y or greater (24,25), allowing ample time to observe transformation from CH to t-MN.

Key limitations to our work include a sample size of 50 patients treated at 1 institution, the limited number of patients with serial samples, and the relatively short follow-up and OS, which makes it difficult to evaluate the medium- and longer-term impact of CH. We reasoned that CH clones present at and beyond 4 cycles of LuPSMA therapy were likely present at therapy initiation (potentially at a lower VAF). Although this allowed us to evaluate temporal changes in CH clones during therapy, it is a limitation in interpreting the impact of CH on survival and toxicity, as there is an inherent immortal time bias, although we attempted to account for this by using a landmark analysis for OS after cycle 5 of therapy. Finally, the trends reported for hematologic toxicity in patients with CH and the emergence of DDR-CH and clonal expansion in the LuPSMA cohort compared with taxane chemotherapy or ²²³Ra therapy were not statistically tested, given the small sample sizes, and should be interpreted as exploratory in nature.

CONCLUSION

CH was highly prevalent in a cohort of patients with mCRPC treated with at least 4 cycles of LuPSMA after prior taxane chemotherapy, but did not appear to correlate with therapeutic efficacy. Patients with CH required growth factor support during therapy, unlike those without CH. In the subset of patients with serial samples available, most patients demonstrated new CH clones or the expansion of baseline clones, particularly in DDR-CH, providing evidence of selection of these CH genotypes by LuPSMA. These observations suggest that awareness of CH may inform clinical management of hematologic toxicity during LuPSMA therapy and invites further evaluation of the impact of CH on clinical outcomes, particularly the risk of t-MNs for individuals treated with radiopharmaceuticals earlier in the prostate cancer disease spectrum.

DISCLOSURE

This work was supported in part by a DFCI Medical Oncology Grant, the Jim & Jean Wood Cancer Research Fund, and an ASCO Career Development Award (Praful Ravi). Lachelle Weeks is supported by the Edward P. Evans Foundation and is the Timmerman Traverse Clinical Investigator for the Damon Runyon Cancer Research Foundation. Adam Sperling reports funding from the National Institute of Diabetes and Digestive and Kidney Diseases (R01DK14153) and an EvansMDS Foundation award and serves as a consultant for Janssen. Praful Ravi reports research funding paid to the institution from Novartis, Blue Earth Diagnostics, Convergent Therapeutics, AstraZeneca, and Lantheus and has relationships with Bayer (advisory board), Curium (advisory board), and Janssen (consultant). Wanling Xie serves on an advisory board for PCCTC. Heather Jacene reports research funding paid to the institution from Blue Earth Diagnostics and Lantheus, receives royalties from Cambridge Publishing, receives honoraria from Medscape and Ideology, and serves on the scientific advisory board of Luminance Biosciences. Lachelle Weeks serves as a consultant for and on the advisory boards of Vertex and Sobi. No other potential conflict of interest relevant to this article was reported.

KEY POINTS

QUESTION: What is the impact of CH on patients receiving LuPSMA for mCRPC, and what are the clonal dynamics during and after therapy?

PERTINENT FINDINGS: CH was highly prevalent in patients receiving LuPSMA for mCRPC. Although it did not impact therapeutic efficacy, there was some evidence for greater hematologic toxicity during LuPSMA therapy in patients who had CH. New or expansion of existing CH clones occurred in all patients receiving LuPSMA in whom serial samples were available, particularly in genes involved with the DDR-CH (PPM1D and TP53).

IMPLICATIONS FOR PATIENT CARE: Awareness of CH at initiation of LuPSMA may help guide closer hematologic monitoring during therapy. Clonal expansion during LuPSMA therapy may underpin the potential risk of t-MNs. Further studies, particularly in populations treated with LuPSMA in earlier stages of prostate cancer, are needed.