HFA-ICOS cardio-oncology baseline risk assessment for multiple myeloma therapies

Clinical background

Proteasome inhibitors and immunomodulatory drugs form the backbone of modern myeloma treatment, but both drug classes carry a substantial risk of cardiovascular complications: heart failure, arrhythmia, thromboembolism and ischaemic events. Carfilzomib has a particularly unfavourable safety profile, with a higher rate of cardiotoxicity than bortezomib [2]. At the same time, myeloma patients are often elderly with a high burden of comorbidity in the form of cardiovascular disease, renal failure and amyloidosis, which makes individual risk assessment before starting treatment complex.

The HFA-ICOS baseline cardiovascular risk assessment for myeloma therapies was developed to meet a concrete need: the absence of a standardised tool for cardiovascular risk stratification before starting cardiotoxic cancer treatment [1]. The instrument is not intended to determine whether cancer treatment should be given, but to guide how intensively cardiac function should be monitored during treatment and whether referral to cardio-oncology is warranted. The 2022 ESC cardio-oncology guidelines have incorporated the proforma into their recommendations for baseline assessment [4].

Applying the HFA-ICOS baseline cardiovascular risk assessment

The instrument is a hierarchical risk classification, not a summed score in the traditional sense. The risk factors are divided into three levels: very high risk, high risk and medium risk. The medium-risk factors are weighted 1 or 2 points as specified. The overall risk category is determined by the following algorithm:

$$\text{Risk} = \begin{cases} \text{Very high} & \text{if any very high risk criterion is present} \ \text{High} & \text{if any high risk criterion is present, or the medium-risk score is} \geq 5 \ \text{Moderate} & \text{if the medium-risk score} = 2\text{--}4 \ \text{Low} & \text{if the medium-risk score} = 0 \text{ or } 1 \end{cases}$$

Very high risk (any single criterion classifies the patient as very high risk regardless of the other factors):

• Existing heart failure or cardiomyopathy
• Previous proteasome inhibitor-related cardiotoxicity
• Venous thrombosis (DVT or pulmonary embolism)
• Cardiac amyloidosis
• Arterial vascular disease (ischaemic heart disease, previous PCI/CABG, stable angina, TIA, stroke or peripheral arterial disease)
• Previous cardiovascular toxicity from immunomodulatory drugs

High risk (any criterion classifies the patient as high risk if no very high risk criterion is present):

• LVEF <50% at baseline
• Raised BNP or NT-proBNP at baseline
• Age ≥75 years
• Previous anthracycline exposure

Medium risk factors (weighted 1 or 2 points; the total determines whether the patient falls into low, moderate or high risk if no very high or high risk criterion is present):

The proforma was not derived from a prospective cohort by statistical modelling, but produced through an expert consensus process at a workshop organised by the Cardio-Oncology Study Group of the Heart Failure Association (HFA) of the ESC in collaboration with the International Cardio-Oncology Society (ICOS) [1]. The risk factors were selected on the basis of available evidence that each parameter contributes to future cardiovascular risk and warrants inclusion. The proforma applies to patients due to be treated with proteasome inhibitors (bortezomib, carfilzomib) or immunomodulatory drugs (thalidomide, lenalidomide, pomalidomide) for myeloma or primary amyloidosis.

Interpretation in practice

The risk category translates into concrete management according to the design of the proforma [1, 4]:

A central message of the source document is that identifying high or very high risk should not lead to effective cancer treatment being withheld, but to cardiovascular risk being optimised and monitoring intensified [1]. A decision to switch to a less cardiotoxic alternative should be made only after multidisciplinary discussion between haematologist and cardiologist, weighing treatment efficacy against safety.

Validation and performance

Since the proforma rests on expert consensus and not on a derived cohort, there is no original discrimination or calibration to compare against. Two external validation studies have, however, been published.

Ho et al. (2026) evaluated the proforma in a retrospective cohort of 419 myeloma patients (mean age 68 years, 52.5% men) treated with proteasome inhibitors or immunomodulatory drugs in Hong Kong between 2012 and 2023 [3]. Patients were classified as low risk (n=184), moderate risk (n=150) and combined high/very high risk (n=85). Over a median follow-up of 35 months, the 3-year cumulative incidence of admission for heart failure was 4.3% in the low-risk group, 11.1% in the moderate-risk group and 28.3% in the high/very high-risk group (log-rank p < 0.001). The c-statistic for the proforma was 0.706 (95% CI 0.63 to 0.78). In a multivariable Fine-Gray model, previous heart failure (SHR 3.96), atrial fibrillation (SHR 2.30) and chronic kidney disease (SHR 4.34) were the strongest individual predictors. Compared with low risk, the moderate-risk group had an SHR of 2.29 and the high/very high-risk group an SHR of 6.39 for admission with heart failure.

Pons-Fuster et al. (2026) validated the proforma in a smaller single-centre retrospective cohort of 98 patients with myeloma or primary amyloidosis treated with proteasome inhibitors in Spain between 2019 and 2024 [2]. Cardiovascular events occurred in 22.5%. The original HFA-ICOS proforma showed moderate discrimination with an AUC of 0.66 and a sensitivity of only 50%, with frequent over-classification of risk. Carfilzomib exposure (HR 4.68, 95% CI 1.47 to 14.90) and a raised NT-proBNP before cycle 2 (>300 pg/mL; HR 3.13, 95% CI 1.10 to 8.93) were independent predictors. An exploratory modified model including these parameters and adjusting the age threshold to ≥65 years improved discrimination to an AUC of 0.72 and sensitivity to 90.9%, but this model is hypothesis-generating and requires external validation.

In summary, the larger cohort [3] shows that the proforma has an acceptable overall ability to stratify risk (c-statistic about 0.71), with a clear gradient of outcomes across the risk categories. The smaller cohort [2] indicates that sensitivity may be limited and that over-classification of risk is a problem, particularly with the mild NT-proBNP elevations that are common in myeloma patients with renal impairment or advanced age.

Limitations

The proforma applies only to patients due to be treated with proteasome inhibitors or immunomodulatory drugs for myeloma or related plasma cell disorders. It should not be applied to other cancer treatments; HFA-ICOS has separate proformas for anthracyclines, HER2-targeted therapies, VEGF inhibitors, kinase inhibitors in CML, RAF/MEK inhibitors and androgen deprivation therapy [1].

The instrument is consensus-based, not statistically derived, meaning that the scoring and weights rest on expert judgement rather than on empirically derived odds ratios. This carries a risk of both over- and under-classification, as the external validation has also shown [2].

Some specific pitfalls:

• Baseline NT-proBNP is a two-point medium-risk factor and a high risk factor if markedly raised, but mild elevations are common in myeloma because of renal impairment, age and amyloidosis, which can lead to over-classification of risk [2]. Pons-Fuster et al. found that a higher threshold (>300 pg/mL) and measurement before cycle 2 predicted cardiovascular events better than the baseline value [2].
• Cardiac amyloidosis classifies the patient as very high risk, but amyloidosis may be subclinical and requires specific investigation (CMR, bone marrow biopsy, SAP scintigraphy) for confirmation. Suspected but unconfirmed amyloidosis should be handled with caution.
• Carfilzomib is not a separate variable in the proforma, despite carrying a considerably higher risk of cardiotoxicity than bortezomib [2]. The proforma treats all proteasome inhibitors alike, which may underestimate the risk in patients due to receive carfilzomib.
• Previous anthracycline exposure is a high risk factor, but the proforma takes no account of the cumulative dose, which is decisive for the actual risk.
• The proforma is intended for baseline assessment and does not capture dynamic changes during treatment. A patient classified as low risk at baseline may develop cardiotoxicity during treatment and should be reassessed if new symptoms or biomarker changes appear.

References

1. 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
2. Pons-Fuster E, Riquelme-Perez A, Shumbar V, et al. Validation and Exploratory Refinement of the HFA-ICOS Score for Cardiovascular Risk in Proteasome Inhibitor-Treated Multiple Myeloma: Single-Center Retrospective Study. Cancers. 2026;18(12):1924. PMID: 42352458
3. Ho MH, Leung C, Leung CY, et al. The clinical utility of the HFA-ICOS risk proforma for predicting heart failure hospitalization in a real-world cohort of multiple myeloma patients. BMC Cardiovasc Disord. 2026;26:493. PMID: 42035004
4. 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