Clinical background
BCR-ABL tyrosine kinase inhibitors have transformed chronic myeloid leukaemia (CML) from a fatal disease into a chronic condition with almost normal life expectancy, but treatment is lifelong and carries a non-negligible risk of cardiovascular toxicity. Nilotinib and ponatinib in particular are associated with arterial thrombosis, peripheral arterial disease and QT prolongation, while dasatinib can cause pleural effusions and pulmonary hypertension [1,2]. In the ENESTnd trial, the rate of cardiovascular events at 10 years of follow-up reached 16 and 24 per cent for nilotinib 300 mg and 400 mg twice daily respectively, compared with 3.6 per cent for imatinib [2].
The clinical challenge is to determine which patient can start treatment with minimal monitoring and which needs intensified cardiac follow-up or referral to cardio-oncology. The 2022 ESC cardio-oncology guidelines give a class I recommendation for baseline cardiovascular risk assessment before treatment with second- and third-generation BCR-ABL tyrosine kinase inhibitors [2,3]. The HFA-ICOS baseline assessment was developed to standardise this assessment and to provide a common terminology for communicating risk between haematologists and cardiologists.
Applying the HFA-ICOS baseline cardiovascular risk assessment
The instrument was published in 2020 as a position statement from the Heart Failure Association (HFA) of the ESC in collaboration with the International Cardio-Oncology Society (ICOS) [1]. Unlike statistically derived risk models, the HFA-ICOS proforma rests on expert consensus rather than on a formal derivation cohort. A working group assembled the evidence on risk factors for cardiovascular toxicity for each therapy class and then formulated the proformas through a consensus process [1].
The calculation combines three logical levels. Certain criteria immediately classify the patient as very high or high risk regardless of the score. The remaining factors are medium-risk factors weighted 1 or 2 points, which are summed to give a medium-risk score. The overall risk classification is:
where is the sum of the weighted medium-risk factors. Very high risk criteria include, among others, previous arterial thrombosis during TKI treatment, heart failure or left ventricular systolic dysfunction, pulmonary arterial hypertension, an LVEF below 50 per cent at baseline and a QTc above 480 ms. High risk criteria include arterial vascular disease, an abnormal ankle-brachial index, a 10-year cardiovascular risk above 20 per cent and age 75 years or older. Medium risk factors include age bands, hypertension, diabetes, hyperlipidaemia, smoking, obesity, chronic kidney disease, arrhythmia, venous thromboembolism and a prolonged QTc interval, with differentiated weighting [1].
Interpretation in practice
The risk class guides two decisions: whether referral to cardio-oncology is warranted, and how intensively cardiac function should be monitored during treatment [1].
| Risk class | Management |
|---|---|
| Low | Standard monitoring. ECG and blood pressure at routine visits. No specific cardiology referral needed. |
| Moderate | Optimise modifiable risk factors before starting. Consider baseline echocardiography. Follow-up with ECG and lipids every 3 to 6 months. |
| High | Referral to cardio-oncology or a cardiologist before starting. Echocardiography and biomarkers at baseline. A risk reduction strategy is discussed in a multidisciplinary team. Consider an alternative TKI with lower cardiovascular toxicity. |
| Very high | Mandatory multidisciplinary assessment. Detailed cardiac evaluation before starting. The choice of treatment is discussed explicitly in the light of cardiovascular risk. Consider a different TKI or intensive cardioprotection. |
It is important to emphasise, as the HFA-ICOS position statement does explicitly, that no patient should be denied effective cancer treatment on the basis of risk class alone, and that the decision should be made in multidisciplinary discussion [1].
Validation and performance
A central problem is that the HFA-ICOS proforma rests on expert consensus and not on a statistical derivation cohort. External validation is therefore particularly important, and for BCR-ABL tyrosine kinase inhibitors it is limited.
The only published validation specific to BCR-ABL TKIs to date is a single-centre study from Imperial College Healthcare NHS Trust in London comprising 229 CML patients treated with nilotinib between 2006 and 2021 [2]. The median age at the start of treatment was 49 years and half were men. The overall incidence of cardiovascular events was 20.9 per cent over a median treatment duration of 34.4 months. HFA-ICOS classified patients well: the incidence of cardiovascular events rose from 11.2 per cent at low risk to 28.2 per cent at moderate risk (HR 2.51, 95 per cent CI 1.17 to 5.66) and 32.4 per cent at high/very high risk (HR 3.57, 95 per cent CI 1.77 to 7.20). For ischaemic events the corresponding figures were 5.2, 17.9 and 21.6 per cent, with an HR of 3.9 (95 per cent CI 1.91 to 7.89) for high/very high risk [2]. The study was retrospective and based on one TKI (nilotinib); baseline QTc data were missing, which may have affected the risk classification.
A systematic review and meta-analysis from 2025 identified HFA-ICOS as the most frequently validated instrument for cancer therapy-related cardiac dysfunction, with 11 external validations [4]. Overall, however, the instrument showed suboptimal performance. In the largest validation cohort, which concerned anthracycline treatment (the CARDIOTOX registry, 1,066 patients), an AUC of 0.78 was reported for symptomatic or moderate to severe asymptomatic cardiac dysfunction at 12 months, but that study did not include BCR-ABL TKIs [4]. A pooled analysis of external validations, mainly in breast cancer patients treated with anti-HER2 therapies, showed a pooled C-statistic of 0.60 (95 per cent CI 0.52 to 0.68) [4]. Calibration was consistently unsatisfactory: the instrument underestimated risk, with observed event rates exceeding those predicted in every risk class. In the breast cancer cohort, observed rates were 12 per cent at low risk (predicted below 2 per cent), 15 per cent at moderate (predicted 2 to 9 per cent), 25 per cent at high (predicted 10 to 19 per cent) and 41 per cent at very high risk (predicted 20 per cent or more) [4]. The underestimation was most pronounced when mild forms of cardiac dysfunction were included as an outcome.
Limitations
The instrument's most important limitation is that it is consensus-based and not statistically derived. The point weights and thresholds rest on expert judgement, not on empirical optimisation against an outcome in a cohort. The only validation specific to BCR-ABL TKIs covers nilotinib alone and comes from a single centre with a retrospective design [2]. Ponatinib, which carries the highest cardiovascular toxicity among BCR-ABL TKIs, has no specific validation.
Baseline QTc data were missing in the nilotinib validation, meaning that one of the few variables that can score 2 points in the medium-risk score was not fully tested [2]. The systematic review found that almost all validations were judged to be at high risk of bias, and that calibration was consistently unsatisfactory with systematic underestimation of risk [4].
The instrument does not capture dynamic risk. A patient classified as low risk at baseline may move up one or two classes through new events during treatment, for example new-onset hypertension or arrhythmia. Risk should therefore be reassessed at every clinical contact, not only at the start of treatment.
The commonest misuse is to interpret a high risk class as a reason to withhold effective cancer treatment. HFA-ICOS is a tool for prioritising monitoring and risk reduction, not for contraindicating oncological treatment [1].
References
- Lyon AR, Dent S, Stanway S et al. Baseline cardiovascular risk assessment in cancer patients scheduled to receive cardiotoxic cancer therapies: a position statement and new risk assessment tools from the Cardio-Oncology Study Group of the Heart Failure Association of the ESC in collaboration with the International Cardio-Oncology Society. Eur J Heart Fail. 2020;22(11):1945-1960. PMID: 32463967
- Fernando F, Andres MS, Claudiani S et al. Cardiovascular events in CML patients treated with Nilotinib: validation of the HFA-ICOS baseline risk score. Cardiooncology. 2024;10(1):42. PMID: 39010172
- Lyon AR, López-Fernández T, Couch LS et al. 2022 ESC Guidelines on cardio-oncology developed in collaboration with the European Hematology Association (EHA), the European Society for Therapeutic Radiology and Oncology (ESTRO) and the International Cardio-Oncology Society (IC-OS). Eur Heart J. 2022;43(41):4229-4361. PMID: 36017568
- Gomes C, Geels J, Debray TPA et al. Risk prediction models for cancer therapy related cardiac dysfunction in patients with cancer and cancer survivors: systematic review and meta-analysis. BMJ. 2025;390:e084062. PMID: 40987514
