Drafting a blueprint for designing successful clinical trials in clonal haematopoiesis
Abstract
‘As our understanding of the biology of clonal hematopoiesis expands, a pressing need in the field becomes the design and implementation of clinical trials to help mitigate the risk for progression to overt myeloid neoplasm. Effective clinical trial design will be informed by use of personalized genetic risk to determine eligibility, strategic endpoint selection, and identification of suitable interventions with a goldilocks balance of toxicity and reduced risk of progression. We will only reach this milestone through collaboration’.
Commentary on: Haque et al. A blueprint for pursuing therapeutic interventions and early phase clinical trials in clonal haematopoiesis. Br J Haematol 2025; 206:416-427.
Approximately a decade ago, clonal haematopoiesis (CH), caused by acquired mutations in genes associated with myeloid neoplasia (MN), was initially characterized as a common age-related condition associated with an increased risk for blood cancers such as myelodysplastic syndrome (MDS) and acute myeloid leukaemia (AML).1 In CH, genetically related haematopoietic stem and progenitor cells (HSPCs) have a fitness advantage compared to normal HSPCs due to somatic mutations and make an outsized contribution to blood cell production. Subtypes of CH, including CH of indeterminate potential (CHIP) and clonal cytopenia of uncertain significance (CCUS), have become clinically defined precursors of MN.2 CHIP and CCUS confer a variable, yet defined, risk of progression.3, 4 CH is also associated with an increased risk of age-related inflammatory diseases including ischaemic cardiovascular disease, liver disease, osteoporosis and chronic pulmonary disease.5 Early intervention in CH could theoretically mitigate both malignant and non-malignant outcomes.
In this issue of the British Journal of Haematology, Haque et al. review available literature and summarize expert discussions held during the October 2021 inaugural meeting on somatic mutations and predisease to develop a series of recommendations on approaching clinical trial design in CH.6 The authors consider thoughtful selection of clinical trial end-points, therapeutic strategies, patient selection and genetic risk stratification. In the spirit of the collaborative summit that inspired these recommendations, the authors highlight key areas of exploration that must be addressed through unified, multicentre collaboration as the field of early intervention in CH evolves.
CLINICAL TRIAL END-POINTS
Cancer prevention (MN-free survival) or improved overall survival (OS) may seem obvious end-points for CH clinical trials. However, the relative health of the CH population and increased risk of age-related non-malignant comorbidity underscore the need to define clinically significant surrogate end-points that can be measured over a shorter timeframe. To this end, the authors offer improvement in clinically significant cytopenias, reduction in inflammatory biomarkers and reduction in clone size (reduced variant allele fraction) as options. While haematological improvement may reduce the symptom burden in patients with CCUS, it remains to be seen if this will have any impact on survival or MN risk. Likewise, as Haque et al. point out that despite the benefit hypothesized to correspond with reduced serum levels of pro-inflammatory cytokines, more data are required to confirm the clinical significance of cytokine levels as a biomarker of efficacy. Patient-reported outcome measures seem fundamentally critical in this early stage of therapeutic clinical trial design for CH. Low-risk interventions with significant quality-of-life benefits may be acceptable, even in the absence of significant improvement in survival.
THERAPEUTIC STRATEGIES
- Is the benefit–risk ratio for early use of MDS-approved agents justified?
- How do we collaborate to build robust correlative science into each clinical trial that informs the design of future studies?
- Are there retrospective analyses that may be done on existing trial specimens in other diseases that may aid in the repurposing of approved drugs for preventing malignant or non-malignant outcomes in CH?
The authors rightly point out that there are no data to back routine alteration of standard-of-care treatment for solid malignancies in the context of therapy-related CH. However, carefully designed clinical trials testing how alternative solid tumour-directed therapies impact clonal outgrowth and ultimately therapy-related MN risk could be attractive and impactful. Beyond therapeutics, global collaboration on clinical studies prospectively investigating the impact of CH on outcomes related to cytotoxic therapy or cellular therapies will be essential to designing future trials to prevent therapy-related MN in these contexts.
PATIENT SELECTION
Selection of the appropriate CH population involves identifying individuals at the highest risk for the outcome of interest. For instance, a CH intervention trial designed to prevent cardiovascular disease outcomes may selectively enrol patients with demographic, clinical or laboratory features that are known risk factors for cardiovascular disease. Haque et al. propose that investigators should homogenize eligibility criteria across planned trials in an effort to minimize barriers to effective interventions for CH.6 One downside of homogenization of criteria is the clinical and genetic heterogeneity among patients with CH that may influence whether or not a patient is responsive to a particular experimental intervention. Variation in eligibility criteria at this early stage – including multiarm studies that enrol two or more distinct population groups – may help pinpoint some of the therapeutic significance of CH heterogeneity.
GENETIC RISK STRATIFICATION
Genetic risk assessment on a personalized basis may inform design of trials in that patients with high-risk precursor states may be more likely to experience benefits from intervention. The past few years have seen a growing appreciation for the distinct leukaemogenic potential of various genotypes of CH, including genotypes seen in real-world analyses.4 Risk of progression from CHIP/CCUS to MN can now be estimated using the CH risk score (CHRS), MN predict, or in CCUS, the clonal cytopenia risk score (CCRS).3, 7, 8 These risk models may help enable design of clinical trials which enrol specific risk groups. High-risk CH – as defined by the CHRS – also correlates with risk of cardiovascular morbidity, suggesting that risk-specific enrolment may also be possible for clinical trials with non-malignant end-points.9
Many large academic centres have established ‘CHIP clinics’ to investigate these and other questions, although answers may not emerge for many years. As suggested by Haque et al., unified data collection and standardized management algorithms may accelerate efforts.6 Haque et al. highlight essential elements that should be considered in all clinical trials for CH and provide a draft blueprint for trial design. We join in the call for collaborative meetings, multicentre collaboration and biospecimen banking. As we utilize these shared resources to improve our understanding of CH biology and outcomes, iterative revisions of this blueprint will bring us closer to more firm guidelines for the field.10
CONFLICT OF INTEREST STATEMENT
SAP served on the MDS advisory board and the AML advisory board for Bristol Myers Squibb. SAP served on the Acute Leukaemia advisory board for Syndax. LDW reports advisory board and consulting fees from Vertex and Sobi all unrelated to the present work.
