Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

CHRONIC LYMPHOCYTIC LEUKEMIA

The clinical utility of polygenic risk scores for chronic lymphocytic leukemia

To the Editor:

We read with interest the article by Kleintsern et al., reporting polygenic risk scores (PRSs) for monoclonal B-cell lymphocytosis (MBL) in Caucasians and for chronic lymphocytic leukaemia (CLL) in African Americans [1]. These data undoubtedly provide further evidence of the polygenic nature of the heritable risk to CLL and its precursor lesion MBL. In their report, which builds on earlier work, they assert that the PRS is a strong predictor of CLL risk, and importantly that the results of their study will aid the identification of individuals at high risk [2]. We believe that assertions regarding the clinical utility of a PRS for CLL need to be tempered. For any PRS to be clinically useful, it firstly needs to provide sufficient risk discrimination and secondly, it must be relevant to strategies for prevention and early detection.

The risk discrimination provided by PRS is a consequence of the PRS-frequency distributions of those with CLL compared to those without CLL. This can be quantified by: (i) comparison of the relative risk (RR) for those at the top and bottom tails of the PRS distribution (Fig. 1), (ii) comparison for specified PRS cut-offs of the proportion of CLL with ‘positive’ PRS (‘detection rate’ (DR)) versus the proportion of unaffected falling within the same PRS score range (the ‘false-positive rate’ (FPR) (1–specificity)) (Fig. 2) [3]. For CLL, Kleinstern et al. report a RR of 11.7 between the top 20% and bottom 20% of the population PRS, providing an area under the curve (AUC) of 0.72. This translates to a FPR of 5% for a DR of 22%, or, to achieve a DR of 50%, tolerance of a FPR of 19% is required (Fig. 1 and Fig. 2). For MBL, a RR of 5 between the top 20% and bottom 20% of the population is associated with an AUC of 0.65, affording a DR of 50% for a FPR of 28% (Fig. 2). Advocates of PRSs tend to focus its performance on the tails of the distribution, diverting attention from the fact that for 90% of individuals their PRS lies relatively close (<2 standard deviations) to the mean. Modification by these PRSs against baseline risk results in small absolute differences in risk. Furthermore, clinical decision-making is generally driven by absolute risk rather than RR, and for those in the top 20% of PRS, lifetime risk for CLL will be only be elevated from 0.44 to 1.54% (3.5-fold) [1, 4]. Two possible avenues have widely been suggested to improve the predictive capability of PRSs. Firstly, it is argued that the discriminatory value of PRS will improve if additional disease-associated SNPs are identified. Aside from the fact that this is a questionable assertion, the prospects of developing large enough GWAS to harvest 80% of the heritable risk of CLL (requiring a sample size of at least 80,000), seems problematic given that the largest GWAS of CLL so far conducted has been based on only 6200 cases [5, 6]. Secondly, the predictive value of PRS could theoretically be boosted when combined with non-genetic risk factors. However, as the authors point out, there are no known non-genetic factors aside from age, sex and ancestry that have been shown to influence CLL risk. A family history (FH) of CLL is a risk factor but it is correlated with PRS, and models incorporating both FH and PRS result in modest improvements in discrimination [2]. While epidemiology is a dynamic field, it would seem naive to predict imminent discovery of modifiable risk factors for CLL to complement PRSs.

Fig. 1: Overlapping relative frequency distributions of polygenic risk scores in chronic lymphocytic leukaemia.
figure1

The detection rate for a false-positive rate of 5% is 22%.

Fig. 2: Receiver operator plots of polygenic risk scores for monoclonal B-cell lymphocytosis (MBL) and chronic lymphocytic leukaemia (CLL).
figure2

Area under the curve (AUC) calculated using the risk-screening converter (https://www.medicalscreeningsociety.com/rsc.asp) and the relative risk estimates between the top and bottom 20% of the population PRS for CLL (orange) and MBL (blue) from Kleinstern et al. [1, 3]. The AUC provides an estimate of the probability that a randomly selected subject with the condition has a test result indicating greater suspicion than that of a randomly chosen individual without the condition. The solid line represents a receiver operator curve based on polygenic risk scores from known risk SNPs. An AUC of 0.5 (dashed line) indicates a classifier that does not differ from the null (i.e. a classifier that does not provide any useful information in discriminating cases from controls).

Along with concerns regarding adequate risk discrimination, PRS testing does not currently offer any direct benefit in the prevention or early detection of CLL. Since no modifiable risk factors for CLL exist, the opportunity to reduce an individual’s risk through lifestyle changes is not an option. While chemoprevention of CLL might appear an attractive proposition, no agents are currently licenced for this indication. Moreover, the risk-benefit profile of a therapy informs the level of disease risk at which administration is justified; thus the value of PRS stratification in determining administration is contingent on the performance of PRS in discriminating those at sufficiently elevated risk. The administration of a chemopreventative agent based on PRS-stratified groups has not yet been trialled and would require careful consideration and includes the selection of an appropriate end point such as overall survival (OS). Over the past decades, treatment of early-stage CLL with chlorambucil, fludarabine and combination chemoimmunotherapy has been trialled with no evidence of an OS benefit [7, 8]. With the advent of novel agents such as Bruton’s tyrosine kinase and B-cell lymphoma 2 inhibitors, treatment of early-stage CLL, and particularly of those at high-risk of disease progression, is being revisited [9]. Whilst data from these trials are currently too immature to draw conclusions regarding OS, an increase in adverse events of clinical interest (including atrial fibrillation, bleeding and hypertension) was associated with Ibrutinib therapy when compared to a placebo in the CLL12 trial [9].

It is important to note that in the absence of counselling and appropriate implementation, communicating DNA based disease risk assessments has the potential to incur harm [10]. Currently, when an individual is diagnosed with CLL, one third of patients will never require treatment in their lifetime and an additional third of patients will have an indolent phase followed by disease progression [7]. Given that the PRS implemented by Kleinstern et al. was developed to predict CLL incidence and not morbidity or mortality, the use of PRS in screening for CLL may well raise the issue of over diagnosis.

The GWAS of CLL that have been conducted have undoubtedly informed our understanding of CLL aetiology. However, given the moderate RRs provided by PRS and the lack of efficacious early interventions, the clinical utility of PRS for CLL is currently questionable.

References

  1. 1.

    Kleinstern G, Weinberg JB, Parikh SA, Braggio E, Achenbach SJ, Robinson DP, et al. Polygenic risk score and risk of monoclonal B-cell lymphocytosis in caucasians and risk of chronic lymphocytic leukemia (CLL) in African Americans. Leukemia. 2021. https://doi.org/10.1038/s41375-021-01344-9.

  2. 2.

    Kleinstern G, Camp NJ, Goldin LR, Vachon CM, Vajdic CM, de Sanjose S, et al. Association of polygenic risk score with the risk of chronic lymphocytic leukemia and monoclonal B-cell lymphocytosis. Blood. 2018;131:2541–51.

    CAS  Article  Google Scholar 

  3. 3.

    Wald NJ, Old R. The illusion of polygenic disease risk prediction. Genet Med. 2019;21:1705–7.

    Article  Google Scholar 

  4. 4.

    Sud A, Chattopadhyay S, Thomsen H, Sundquist K, Sundquist J, Houlston RS, et al. Analysis of 153 115 patients with hematological malignancies refines the spectrum of familial risk. Blood. 2019;134:960–9.

    CAS  Article  Google Scholar 

  5. 5.

    Law PJ, Berndt SI, Speedy HE, Camp NJ, Sava GP, Skibola CF, et al. Genome-wide association analysis implicates dysregulation of immunity genes in chronic lymphocytic leukaemia. Nat Commun. 2017;8:14175. 2017/02/06

    CAS  Article  Google Scholar 

  6. 6.

    Janssens ACJW, Aulchenko YS, Elefante S, Borsboom GJJM, Steyerberg EW, van Duijn CM. Predictive testing for complex diseases using multiple genes: Fact or fiction? Genet Med. 2006;8:395–400. 2006/07/01

    Article  Google Scholar 

  7. 7.

    Dighiero G, Maloum K, Desablens B, Cazin B, Navarro M, Leblay R, et al. Chlorambucil in indolent chronic lymphocytic leukemia. N. Engl J Med. 1998;338:1506–14.

    CAS  Article  Google Scholar 

  8. 8.

    Herling CD, Cymbalista F, Groß-Ophoff-Müller C, Bahlo J, Robrecht S, Langerbeins P, et al. Early treatment with FCR versus watch and wait in patients with stage Binet A high-risk chronic lymphocytic leukemia (CLL): a randomized phase 3 trial. Leukemia. 2020;34:2038–50. 2020/08/01

    CAS  Article  Google Scholar 

  9. 9.

    Langerbeins P, Bahlo J, Rhein C, Gerwin H, Cramer P, Fürstenau M, et al. Ibrutinib versus placebo in patients with asymptomatic, treatment-naïve early stage CLL: primary endpoint results of the phase 3 double-blind randomized CLL12 trial. Hematological Oncol. 2019;37:38–40.

    Article  Google Scholar 

  10. 10.

    Hudson K, Javitt G, Burke W, Byers P, Committee ASI. ASHG Statement on Direct-to-Consumer Genetic Testing in the United States. Am J Hum Genet. 2007;81:635–7.

    Article  Google Scholar 

Download references

Acknowledgements

RSH acknowledge grant support from Cancer Research UK (C1298/A8362) and the Wellcome Trust (214388). AS is in receipt of a National Institute for Health Research (NIHR) Academic Clinical Lectureship, an Academy of Medical Sciences Starter Grant for Clinical Lecturers and funding from the Royal Marsden Biomedical Research Centre. This is a summary of independent research supported by the NIHR Biomedical Research Centre at the Royal Marsden NHS Foundation Trust and the Institute of Cancer Research. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health.

Author information

Affiliations

Authors

Contributions

AS, PJL, and RSH researched, reviewed, drafted and edited the manuscript. AS generated the figures.

Corresponding author

Correspondence to Amit Sud.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Sud, A., Law, P.J. & Houlston, R.S. The clinical utility of polygenic risk scores for chronic lymphocytic leukemia. Leukemia 35, 3608–3610 (2021). https://doi.org/10.1038/s41375-021-01429-5

Download citation

Search

Quick links