Myeloid Toxicity of Radionuclide Cancer Therapy

Introduction

Theranostic nuclear oncology is founded upon the principle of integration of diagnostic and therapeutic radionuclide ligands, which target a tumor-specific biomarker to deliver personalized molecular radiation therapy of metastatic cancer. ¹ It epitomizes the precision oncology paradigm, in that the tumor-specific target molecule is defined by imaging with positron or γ-emitting radionuclides, and the same ligand, labeled with α- or β-emitting tumoricidal radiation, precisely targets the previously identified avid metastases.

Clinical application of theranostic clinical oncology is increasing exponentially, for diagnosis, staging, therapy, monitoring of tumor response, and prognostication in cancer care. ^{2 –4} However, acceptance of radionuclide theranostics within the oncology community, and incorporation into routine oncology practice, either when used alone, or in combination with chemotherapy–immunotherapy, has been slow. Reasons why oncologists seem reluctant to embrace theranostic radionuclide therapy (TRT) in patients with advanced cancers are multifactorial and include paucity of randomized controlled trial data, nonreimbursement, availability logistics, regulatory issues, and concerns related to radiation toxicity. Myelotoxicity is a common, but manageable, adverse event consequent on radionuclide therapy. This commentary evaluates the risk–benefit ratio in theranostic treatment of thyroid cancer, lymphoma, neuroendocrine neoplasm, and prostate cancer and elucidates mechanisms of hematological radiation toxicity. The potential for minimization of myelotoxicity is explored.

Radioiodine Therapy of Differentiated Thyroid Cancer

Radioiodine (RAI) for targeted therapy of differentiated thyroid cancer (DTC) represents the oldest and most well-studied radioisotope in medical use. ^5,6 Initial surgical management followed by RAI remains the mainstay of therapy and achieves a cure in most DTC patients. ⁷ The efficacy of RAI for both low and high prescribed activity in DTC is well established. Long-term follow-up (median 10 years) of patients from randomized studies has repeatedly demonstrated an overall survival (OS) of 90%–95% and a low risk of recurrence. ^{8 –12} The 75-year experience with RAI for DTC has informed practice in other cancer subtypes and generated controversy regarding the risk of secondary hematologic malignancies.

A SEER registry study of over 100,000 patients, recently reported a relatively high risk of early development of acute leukemia (AL) following RAI when compared with thyroidectomy alone (AL; hazard ratio, 1.79; 95% confidence interval [CI], 1.13–2.82; p = 0.01). ¹³ This increased risk was reported in both low- and high-risk disease. ¹³ Under close scrutiny however, it is apparent that these findings, and the absolute conclusion regarding the significantly increased hematologic risk of RAI, are flawed. The actual data presented a negligible difference in incidence rate of secondary hematologic malignancies (SHM) of ∼8 per 100,000 patient years, which was not adjusted for relevant patient factors such a disease stage. ¹⁴ It is also notable that relevant factors relating to the pathobiology of SHM, especially myelodysplastic syndrome/acute leukemia (MDS/AL), were not addressed in the SEER data. ^{15 –21} The absence of dose exposure data, the number of patients receiving supraphysiologic thyroxine, staging, baseline blood results, concurrent therapies associated with comorbidities and, crucially, the lack of diagnostic data relating to AL diagnosis such as bone marrow karyotyping consistent with secondary disease, is a significant limitation of the study. Finally, both the earlier-than-anticipated incidence of AL following therapy, and conflicting reduction in risk of other SHM subtypes, for example myeloma, require explanation. ^{12,14 –16} The findings and recommendations of the retrospective SEER study, based upon incomplete data, were refuted by the nuclear medicine community. ¹⁶

However, these data, and the engendered controversy, highlight two important points regarding TRT of cancer. First, there is continued prescient concern from the oncology community regarding the long-term myelotoxic sequelae of radionuclide therapy, and attendant anxiety of both clinicians and patients. Second, there is a limited understanding of the true pathobiology and natural history of secondary MDS/AL following TRT of cancer.

Iodine-131-Rituximab Radioimmunotherapy of Non-Hodgkin's Lymphoma

Most trials of ¹³¹I-rituximab radioimmunotherapy (¹³¹I-rtiuximab RIT) for lymphoma have focused on the management of low-grade non-Hodgkin's lymphoma (NHL), in particular, follicular lymphoma (FL). The authors' own 10-year single-center experience of ¹³¹I-rituximab RIT in over 100 patients with relapsed low-grade NHL (predominantly FL), reported a progression-free survival (PFS) of 32 months in patients achieving a complete remission based on ¹⁸ F-fluorodeoxyglucose positron-emission tomography/computed tomography ( ¹⁸ F-FDG PET/CT). ²² The 8-year OS was 48%, demonstrating achievement of durable responses. All patients had prospective individualized dosimetry of 2 Gy to red marrow. Myelotoxicity acute grade 3/4 hematologic toxicity was observed in up to 10% of patients and MDS was observed in 4% (6 patients) with 1 progressing to AL. ²² All therapy was completed over 4 weeks, all patients being treated as outpatients with a single therapy dose of ¹³¹I-rituximab RIT. Thus ¹³¹I-rituximab RIT provided a cost-effective and safe treatment for relapsed/refractory indolent NHL with a negligible long-term risk of SHM. ²²

The ¹³¹I-rituximab RIT protocol was subsequently combined with 1 year of standard maintenance rituximab for chemotherapy-free induction treatment of FL in the Phase II INITIAL study (Australian Clinical Trial Registry Notification No. 12607000153415). ²³ The long-term update of this study (median follow-up 6 years) reported an estimated OS of 84% and PFS of 76%. No significant long-term or persistent grade ≥3 hematologic toxicity was observed. No instance of MDS/AL was recorded. There was no treatment-related morbidity or mortality reported. ²⁴

While not directly comparable, these ¹³¹I-rituximab RIT results are favorable when considered alongside outcomes reported from alternative induction regimen trials in FL. At a median follow-up of 34.5 months, the phase III GALLIUM study (comparing obinutuzumab vs. rituximab combined with standard chemotherapy induction followed by 2 years of maintenance) reported an estimated 3-year OS of 94% versus 92.1%; and PFS 80% versus 73.3% for Obinutuzumab and rituximab, respectively. Unsurprisingly, the use of combination chemotherapy and 2-year maintenance monoclonal antibody resulted in higher rates of grade ≥3 hematologic toxicity, up to 20% during the maintenance phase, and up to 9% during the extended follow-up period. ²⁵ Two cases (0.3%) of MDS/AL were also reported. ²⁵

Regarding alternative chemotherapy-free induction, the phase III RELEVANCE study compared rituximab plus oral lenalidomide with standard rituximab (RL) chemotherapy, both with 2 years of rituximab maintenance. At a median follow-up of 37.9 months, there was an estimated 3-year OS of 94% for both arms; and PFS of 77% versus 78%. ²⁶ Hematologic toxicity analysis reported a 32% incidence of grade ≥3 neutropenia for rituximab–lenalidomide and 50% for rituximab chemotherapy. Four cases of SHM were reported in the lenalidomide arm (0.8%) and two (0.4%) in the chemotherapy arm. Crucially, the trial failed to meet its endpoint with respect to demonstrating superiority of RL-based induction over standard of care. ²⁶

Existing data support the long-term efficacy and safety of induction ¹³¹I-rituximab RIT for FL and highlight the importance of patient selection and timing with respect to ameliorating the long-term risks of SHM. The risk of MDS/AL following alkylating chemotherapy is well established. ²⁷ Typically manifesting 5–10 years postexposure, the double-stranded deoxyribonucleic acid (DNA) damage resulting from therapy leads to aberrant clonal hematopoiesis (CH) and expansion over time, eventually reaching a critical limit associated with clinical sequelae. ²⁸ Therefore, the modestly increased long-term incidence of SHM following ¹³¹I-rituximab RIT for relapsed disease is to be anticipated. ²² There is, however, concern for potentially accelerating and increasing the risk of secondary MDS/AL with sequential exposure to DNA-damaging agents. The reduced intervals for marrow recovery, such as those encountered with ¹³¹I-rituximab RIT consolidation following induction chemoimmunotherapy of FL, heighten this risk.

The Southwest Oncology Group (SWOG) has conducted two pivotal studies evaluating the role of combination chemoimmunotherapy and anti-CD20-conjugated RIT in FL. The phase II SWOG S0801 trial evaluated the role of standard rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone (RCHOP) induction, followed by ¹³¹I-tositumomab RIT consolidation and maintenance rituximab. Despite an encouraging 3-year PFS of 90%, a significant proportion of patients failed to tolerate 4 years of maintenance therapy. Acute and long-term toxicity was predominantly hematological with up to 57% of patients experiencing grade ≥3 hematologic toxicity and 5% developing secondary myeloid disorders, including MDS/AL within a relatively short time frame. ²⁹ The phase III SWOG S0016 study compared RCHOP with ¹³¹I-tositumomab-RIT-CHOP in previously untreated FL. The 10-year update reported durable remissions, with RCHOP patients achieving a significantly superior PFS of 56% versus 42% for ¹³¹I-tositumomab RIT arm (p = 0.01). However, there was no significant difference in the OS. Long-term hematologic toxicity was highlighted, occurring in 5 (1.8%) patients who were treated with RCHOP and 13 (4.9%) who received the ¹³¹I-tositumomab-RIT CHOP developed myeloid malignancies (AL or MDS; p = 0.058). The estimated 10-year cumulative incidence of death from SHM was 3.2% (95% CI, 1.5%–6.1%) in the RCHOP arm and 7.1% (95% CI, 4.3%–10.9%) in the ¹³¹I-tositumomab-RIT CHOP arm. ³⁰ Both studies failed to demonstrate superiority of RIT consolidation following R-chemo for FL and have again highlighted the risk of hematologic toxicity following combined alkylator and radionuclide exposure.

Prior purine analog therapy is also associated with a significantly increased risk of secondary MDS/AL following RIT salvage with both ⁹⁰Y and ¹³¹I. At a median follow-up of 5 years, an estimated 10-year risk of MDS/AL of 29% was reported in patients with prior fludarabine exposure, versus 13% in those without (p = 0 · 012). ³¹ The median time to MDS/AL was 41 months with a median survival of 1.6 months. This increased incidence of MDS/AL following RIT consolidation in patients with prior purine analogs confirmed findings from previous studies of chemotherapy–RIT combinations, which reported a 14% crude cumulative 11-year incidence of MDS/AL. ³²

Durable remissions can be achieved with front-line RIT for FL, but careful patient selection with respect to prior therapies and timing of RIT is crucial to ameliorate SHM. Combination chemotherapy–RIT, including RIT consolidation, confers an increased 5- to 10-year risk of MDS/AL without significant improvement in efficacy compared with single modality treatment with either. This increased incidence of MDS/AL may be attributable to the different mechanisms of complex double-stranded DNA damage to myeloid progenitors arising from alkylator, purine analog, and radiation exposure. Critically, trials to date have not routinely performed in-depth genomic analysis, to either interrogate individual baseline risk or monitor for evidence of MDS/AL clonal expansion. Therefore, little is known about the natural history/pathogenesis of SHM in the setting of RIT, or indeed, TRT of cancer. The availability of pharmacogenomic platforms and validation of next-generation sequencing (NGS) for detection of MDS/AL-related circulating myeloid clones means that such analysis is now possible.

Lutetium-177-Octreotate Peptide Receptor Radionuclide Therapy of Neuroendocrine Tumors

Studies in neuroendocrine tumors (NETs) have focused on outcomes of salvage peptide receptor radionuclide therapy (PRRT) following progression/relapse after prior somatostatin–analog treatment, chemotherapy, and radiotherapy. Their literature review of long-term outcomes of over 2000 patients reported a 10% incidence of short-term grade ≥3 hematologic toxicity, most often manifesting as self-limiting thrombocytopenia during cycle one. ³³ With respect to SHM, MDS/AL was a rare event, occurring in 1.4% of patients at a median duration of follow-up of up to 62 months. Pathologic findings in these patients confirmed complex chromosomal abnormalities consistent with secondary MDS/AL following chemotherapy/radiation exposure. ³³ Factors associated with myelotoxicity included age >70, impaired renal function, baseline cytopenias (a likely indicator of inherent genetic damage), prior number of therapies, prior alkylator exposure, and prior radiotherapy. ³³

These results supported the findings of an earlier review of the long-term outcomes of 807 patients. However, in that study, the incidence of self-limiting acute grade ≥3 hematologic toxicity was more prevalent in patients receiving yttrium-90-based PRRT. This did not, however, translate to an increased risk of SHM associated with ⁹⁰Y exposure. At a median follow-up of 30 months (range 1–180 months), MDS occurred in 2.35% of patients. Risk factors for MDS included previous chemotherapy, acute platelet toxicity (possible indication of bone marrow sensitivity), and bone marrow factors such as chemotherapy with myelotoxic agents, tumor invasion of bone marrow, radiotherapy to marrow fields, immunosuppressive agents, previous radionuclide therapies, previous bone marrow toxicity, previous cytopenias, and prior myeloproliferative diseases. Additional analysis showed that these identified risk factors were of limited predictive value (<30%), consistent with the existence of unidentified individual genetic susceptibilities for SHM. ³⁴

No significant association between the conventional risk factors and subsequent development of persistent hematologic dysfunction (PHD), defined as diagnosis of SHM or otherwise unexplained cytopenia for >6 months, was reported in the Dutch study of ¹⁷⁷Lu-DOTATATE PRRT of NET. ³⁵ The 2.9% incidence of SHM in that study was similar to that noted in previous reports and in registry data. However, given that 90% of patients were prior chemotherapy naive, the median time-to-PHD of 41 months (range 15–84), raises the probability of inherent drivers of irreversible myeloid injury. ³⁵

The minimal incidence of long-term hematology toxicity, coupled with favorable survival outcomes, has led to the incorporation of PRRT into international NET treatment guidelines. ³⁶ Acknowledging the potential concern for both acute and long-term hematologic toxicity, these guidelines recommend careful patient selection based on evidence of adequate baseline marrow and renal reserve. Furthermore, based on existing evidence, there has been a shift away from the use of high-energy β ⁹⁰Y, toward ¹⁷⁷Lu-based therapy. There has also been interest in enhancing efficacy with chemotherapy–PRRT combinations for advanced and relapsed/refractory NET patients. The combination of PRRT with alkylating agents such as capecitabine (CAP) or temozolomide (TEM), is associated with an increased incidence of both acute and long-term hematologic toxicity. ^{37 –40} However, true long-term data are sparse. Our recent update of NET patients treated with ¹⁷⁷Lu-octreotate-PRRT CAPTEM (n = 37) highlights this increased risk. At a median follow-up of 7 years (range 1–10), the authors observed a 16% incidence of persistent hematologic toxicity, defined as sustained grade ≥3 hematologic toxicity beyond 36 months of follow-up, and 8% incidence of MDS/AL with a median time-to-event of 46 and 34 months, respectively. ⁴¹ Based on historical experience of radionuclide–chemotherapy combinations for cancer, this early evidence does suggest a potential for increased toxicity with chemotherapy–PRRT combinations, and clinical application in heavily pretreated patients should be very cautious.

Radium-223 Dichloride Therapy of Metastatic Castrate-Resistant Prostate Cancer

The pivotal phase III ALSYMPCA trial established efficacy of targeted α emitter therapy of metastatic castrate-resistant prostate cancer (mCRPC) using single-agent radium-223 dichloride (²²³Ra). At the time of initial reporting, ²²³Ra therapy was associated with a modest but significant improvement in OS (median, 14.9 months vs. 11.3 months; hazard ratio, 0.70; 95% CI, 0.58–0.83; p < 0.001), ⁴² delayed time-to-skeletal-events, ⁴³ and minimal acute toxicity (predominantly low-grade anemia). ⁴² At a median follow-up of 13 months (range 0–36), updated safety analysis did not report any significant long-term cytopenias, and no cases of MDS/AL were observed. However, 1 ²²³Ra patient did develop aplastic anemia 16 months postfinal injection (<1%). Although true long-term follow-up is lacking, given that the cohort included heavily pretreated patients (>50% with prior docetaxel) and those with a high burden of baseline skeletal disease (>40% with a superscan), ⁴⁴ these results established the therapeutic utility and sustained hematologic safety of radionuclide therapy of mCRPC. In doing so, they provided a platform for future studies of targeted radioligand therapy (RLT), especially radioimmunoconjugates of antiprostate-specific membrane antigen (PSMA) monoclonal antibodies.

Lutetium-177/Actinium-225 Prostate-Specific Membrane Antigen Radionuclide Therapy of Metastatic Castration-Resistant Prostate Cancer

Results of the TheraP phase II randomized study of ¹⁷⁷Lu-PSMA-RLT versus cabazitaxel have confirmed the superior efficacy profile of RLT versus standard chemotherapy mCRPC. At a median follow-up of 18 months, PSA responses were more frequent among men in the ¹⁷⁷Lu-PSMA-RLT group than in the cabazitaxel group (66% vs. 37%) by intention to treat (95% CI, 16–42; p < 0 · 0001). Overall, RLT was associated with an increased incidence of acute cytopenias (all grades), in particular advanced-grade thrombocytopenia and leukopenia (consistent with previously published preliminary data ⁴⁵ ). Strikingly, there was a significant difference in incidence of acute hematologic toxicity, with a reported incidence of febrile neutropenia, as a medical emergency, of 4% versus 13% for ¹⁷⁷Lu-PSMA-RLT and cabazitaxel, respectively. ⁴⁶ This efficacy data are consistent with reported results from numerous preceding retrospective studies. ^{47 –50} The phase III VISION study of best standard of care with or without ¹⁷⁷Lu-PSMA-RLT has recently been reported to have met its primary endpoints, significantly improving OS and radiographic PFS in mCRPC. The safety profile is yet to be detailed but is reported to be consistent with previously published data. ⁵¹

With respect to risk factors for acute hematologic toxicity of ¹⁷⁷Lu-PSMA-RLT for mCRPC, at median follow-up of 12 months, our recently published experience of 100 heavily pretreated patients failed to identify any statistically significant association with the development of grade ≥3 hematologic toxicity and risk factors of age >70, prior/concurrent therapy, presence of metastases, and number of cycles of therapy completed. ⁵⁰ The duration of follow-up is limited, nevertheless, given the grave prognosis faced by mCRPC patients (5-year OS of ∼30% ⁵² ), SHM is unlikely to be of significant clinical concern in heavily pretreated patients.

Despite the efficacy of ¹⁷⁷Lu-PSMA-RLT, the majority of patients do experience progressive disease with a requirement for additional repeat cycles or further salvage therapy, due to radioresistance of tumors to β-emitters. ⁵³ In this context, ²²⁵Ac-PSMA-RLT has been studied as an alternative salvage option, following failure of ¹⁷⁷Lu-PSMA-RLT. ^{53 –56} Owing to higher linear energy transfer and short-tissue penetration, targeted α therapy can induce cell damage even in β-emitter-resistant tumors, with the added advantage of reduced marrow penetration and consequently less hematologic toxicity. ^55,56

Preliminary retrospective results report complete response rates for PSA of 70%–87%, and for ⁶⁸ Ga-PSMA PET/CT of 13%–30%. ^55,56 Acute hematologic toxicity is mostly self-limiting, however, there is an increased incidence of grade ≥3 toxicity in patients presenting with widespread skeletal metastases on baseline ⁶⁸ Ga-PSMA-PET/CT. ^54,56 Despite this promising activity, therapy-associated irreversible hematologic and salivary gland toxicity are of concern. ⁵⁴

Reports of ²²⁵Ac-PSMA-RLT salvage following ¹⁷⁷Lu-PSMA-RLT failure have demonstrated a PSA response in >50% of patients, however, survival benefit appears to be modest with median PSA-PFS of 3.5 months (95% CI, 1.8–11.2) and median OS of 7.7 months (95% CI, 4.5–12.1). ⁵³ Acute hematologic toxicity remains a concern with incidence of grade ≥3 anemia and thrombocytopenia observed in 35% and 19% of patients, including irreversible cases. ⁵³ Given that the primary aim of therapy is to improve and maintain patient quality of life, further research is required to establish the optimal timing and dosing schedule of ²²⁵Ac-PSMA-RLT for mCRPC, particularly in heavily pretreated patients.

Discussion

Conventional risk factors, such as age >70, prior chemotherapy with alkylating agents or purine analogs, prior radiation/radioisotope exposure, heavy bone marrow involvement by tumor and evidence of compromised bone marrow reserve must be considered when planning radionuclide therapy. However, their limited predictive for SHM highlights the unmet need for reliable personalized risk stratification before radionuclide therapy of advanced cancer. This is especially important considering the global expansion of TRT applications and consideration of novel combinations. ²

MDS/AL is one of the most common hematologic malignancies in adults, with advanced age being the predominant risk factor. ^{57 –59} It is now understood that this clonal disease is driven by recurrent, acquired somatic genetic mutations. Initiating mutations in hematopoietic stem cells drive CH, which, when accompanied by an accumulation of cooperating driver DNA mutations, result in progression to MDS and secondary AL (Fig. 1). We now have genomic tools to help us understand the evolution of MDS/AL following radionuclide exposure and to prospectively screen/monitor patients and individually tailor TRT to minimize the risk of SHM. The potential implications of such pharmacogenomic analysis in TRT of cancer has already been explored with genetic profiling identifying germline and tumor specific mutations predictive of disease risk and response to radiation therapy. ⁶⁰ Targets for potential tumor-specific PRRT combinations have also been identified, providing proof of concept for the prospective application of genomic analytic techniques to improve therapeutic efficacy. ⁶⁰ Thus far, predominantly applied in the setting of pathway/checkpoint inhibitor development and selection, cross-fertilization of pharmacogenomics to theranostic oncology will facilitate customization of TRT to patient and tumor-specific genetic variants predictive of both response and toxicity.

OPEN IN VIEWER

FIG. 1.

Pathogenesis of secondary hematologic malignancies following radionuclide cancer therapy. CCUS, clonal cytopenia of uncertain significance; CHIP, clonal hematopoiesis of indeterminate potential; MDS/AL, myelodysplastic syndrome/acute leukemia.

Bone marrow karyotyping for chromosomal abnormalities remains the most conventional and widely available diagnostic genetic test for MDS/AL. ⁶¹ Patients exposed to alkylating agents or radiation injury develop abnormalities involving chromosomes 5 (−5/del [5q]) and 7 (−7/del [7q]), and trilineage dysplasia with an insidious onset at a median of 4–7 years. ⁵⁷ Patients developing secondary MDS/AL following TRT predominantly do so after prior exposure to alkylating/purine analog agents. However, monitoring for cytogenetic abnormalities with serial invasive bone marrow biopsies is impractical.

With the evolution of sequencing technologies and progressive reduction in cost, chromosomal studies are being superseded by radiogenomics. ^{62 –65} An evolving field of study, radiogenomics utilizes whole-genome sequencing data to identify signatures of mutation combinations, which predict risk of therapy-related toxicity. Research has primarily focused on analysis of single nucleotide polymorphisms (SNPs), which represent a major source of genetic variation, and may be used as a predictive tool for the development of toxicity following exposure to radiation. ⁶⁵ The majority of studies have reported on the use of candidate gene and genome-wide association studies (GWAS), which can identify an array of SNP variants, using large discovery and replication cohorts to identify associations. ⁶² These studies applied to patients treated with radiotherapy for prostate, breast, and lung cancer have successfully identified and validated a number of SNPs associated with a significantly increased risk of adverse clinical outcomes. ^{66 –70} To date, no SNPs have been linked to an increased risk of acute or long-term hematologic toxicity following radiation exposure. ⁶²

The potential of radiogenomic analysis for risk stratification for individual patients before TRT must be balanced against a number of challenges, including requirements for detailed dosimetry data, long-term follow-up, in addition to conventional baseline assessment. Other considerations include variability of tools/methods, the impact of other epigenetic and “panomic” factors, and large sample size requirements. ⁷¹ Despite these limitations, GWAS and candidate gene studies are a validated method for identifying SNPs predictive of radiation-related toxicity and may be appropriately applied to prospective assessment of TRT-related toxicities.

In addition, NGS of circulating nucleated cells has also clarified the age dependence of CH as being driven by a continued cumulative risk of acquired mutations with subsequent clonal expansion, leading to an increased detectability in older age, as opposed to the process being primarily driven by specific age-dependent DNA mutations. ⁷² In the healthy preretirement age population, CH accounts for ≤10% of circulating nucleated cells, whereas they are found in up to 15%, 20%, and 30% of adults in their mid 70's, 90’s, and >100 years of age, respectively. ^{73 –77} Although the nature and significance of mutations vary, >90% of patients with MDS are known to harbor disease-associated driver mutations. These findings also explain why age >70 appears to be an independent risk for SHM following TRT. Commonly mutated genes include DNMT3A, TET2, ASXL1, TP53, RUNX1, and those that are components of the 3′ RNA splicing machinery (e.g., SF3B1, U2AF1, SRSF2, and ZRSR2). ^{58,78 –82} Analysis has led to the identification of higher-risk patients with clonal hematopoiesis of indeterminate potential (CHIP), and clonal cytopenia of uncertain significance (CCUS), who do not meet diagnostic criteria for MDS, but harbor associated pathogenic mutations. ^83,84

Decreased OS in those ≥70 years of age, greater risk of hematologic malignancy, and an increased risk of cardiovascular complications are associated with CHIP. ⁷⁴ The risk of transformation/progression to a hematologic malignancy in CHIP-positive patients is estimated to be 0.5%–1% per year. ⁸⁵ Patients with CCUS are at a significantly increased risk for development of MDS/AL when compared with patients with idiopathic cytopenias without evidence of CH, with a 5-year cumulative risk of progression of 82% for CCUS versus 9% for CHIP, and a 10-year cumulative risk of progression of 95% versus 9%. ⁷⁶ These findings highlight the dramatic impact and early clinical sequelae of CH. ⁸³ The risk of development of MDS/AML in patients with CH varies considerably depending on the aberrant gene, with mutations of TP53 and U2AF1 associated with a high risk. Mutations of DNMT3A, TET2, JAK2, SRSF2, IGH1, and IDH2 are associated with an increased risk ^{86 –88} and a large clonal burden. ⁸⁹ It is therefore conceivable that patients exhibiting the harboring features of CCUS/CHIP would be at increased risk of SHM following radiation exposure, thus enhancing the limited predictive value seen with conventional risk factors. Although use of CH testing for the screening, diagnosis, and management of MDS/AL remains ancillary, international guidelines have been developed for standardization of reporting, allowing for incorporation of genomic testing into prospective studies. ⁸⁴

Given our limited understanding of the mechanisms of secondary MDS/AL, consequent upon TRT of advanced cancer, performance of minimally invasive genomic analysis can clearly play a crucial role in the screening/monitoring of prospective patients. The theranostic oncology community now have an opportunity to embrace genomic analysis in prospective studies of TRT to elucidate the mechanism of secondary hematologic toxicity following radionuclide therapy by prior individual risk assessment to minimize the occurrence of SHM.

Conclusions

The immediate benefit of TRT of advanced cancer must be balanced against the long-term risk of MDS/AL. The risk of long-term hematologic toxicity in patient populations treated in early phase studies is apparent, and is likely to increase with the advent of novel combination chemo-TRT regimens. Integration of novel genomic analysis to predict individual risk and characterize the true pathogenesis of long-term SHM will enable appropriate selection of patients and tailor TRT to minimize toxicity while optimizing efficiency in the control of metastatic cancer.