Abstract
Aims
We aimed to determine the prevalence of right ventricular (RV) systolic dysfunction on cardiovascular magnetic resonance imaging (CMR) and its impact on long-term adverse outcomes in a large cohort of cancer survivors treated with anthracycline-based chemotherapy.
Methods and results
Consecutive cancer survivors treated with anthracyclines who underwent clinical CMR for suspected anthracycline-related cardiomyopathy were studied. The primary endpoint was a composite of all-cause death or major adverse cardiac events (MACE): heart failure hospitalization, heart transplantation, ventricular assist device implantation, resuscitated cardiac arrest, or life-threatening ventricular arrhythmia. The secondary endpoints were all-cause death, and cardiac death or MACE. Among 249 survivors who underwent CMR at a median of 2.9 years after cancer treatment, RV systolic dysfunction was present in 54 (21.7%). Of these, 50 (92.6%) had an abnormal left ventricular ejection fraction (LVEF). At a median follow-up time after the CMR of 2.7 years, 105 survivors experienced the primary endpoint. On Kaplan–Meier analyses, the cumulative incidence of the primary endpoint was significantly higher in survivors with abnormal RVEF compared with those with normal RVEF (P = 0.002). However, on Cox multivariable analyses, RVEF was not associated with the primary endpoint (HR 1.04 per 5% decrease; 95% CI 0.93–1.17; P = 0.46) after adjustment for non-imaging variables and LVEF. RVEF was also not associated with the secondary endpoints.
Conclusion
Among anthracycline-treated cancer survivors undergoing CMR for suspected cardiotoxicity, RV systolic dysfunction was present in one in five cases, accompanied by LV systolic dysfunction in nearly all cases, and was not independently associated with long-term outcomes.
Keywords: magnetic resonance imaging, cardiomyopathy, anthracyclines, cardiotoxicity, right ventricular function
Introduction
Anthracyclines are used in the treatment of a variety of solid organ tumours and haematologic malignancies, including breast cancer, lymphoma, leukaemia, and sarcoma. Anthracyclines have been linked to an increased risk of cardiovascular disease—particularly cardiomyopathy and heart failure—and mortality.1,2 The most commonly described imaging abnormality associated with cardiotoxicity is left ventricular (LV) systolic dysfunction.
Structural and functional abnormalities of the right ventricle (RV) are prevalent in various cardiomyopathies. RV dysfunction is increasingly being recognized to have incremental prognostic implications over clinical risk factors and LV dysfunction.3,4 Small studies have described RV dysfunction in cancer patients during anthracycline treatment5–8 and in survivors after anthracycline treatment.8–10 However, whether RV dysfunction has long-term prognostic value in cancer survivors has not been studied. This knowledge has the potential to guide the surveillance of cancer survivors after anthracycline treatment and to contribute to improved outcomes. We hypothesized that RV dysfunction is independently associated with long-term adverse outcomes in anthracycline-treated cancer survivors.
Accordingly, the objective of this study was to determine the prevalence of RV systolic dysfunction on cardiovascular magnetic resonance imaging (CMR), its determinants, and its impact on long-term adverse outcomes in a large cohort of anthracycline-treated cancer survivors with suspected cardiotoxicity.
Methods
Study design and patient selection
We conducted a retrospective cohort study of consecutive adults treated with anthracyclines for breast cancer, lymphoma, leukaemia, or sarcoma who were referred for clinical CMR at the University of Minnesota, Minneapolis, MN, USA for suspected cardiotoxicity based on symptoms or other testing. Survivors with late gadolinium enhancement (LGE) on CMR were excluded since they had cardiomyopathies related to causes other than their cancer treatment, as we have shown recently.11 Detailed demographic and clinical data were collected by review of the electronic medical record. The study was approved by University of Minnesota’s Institutional Review Board with a waiver of signed informed consent.
CMR protocol
CMRs were performed using 1.5 T scanners (Siemens Sonata, Avanto, or Aera, Siemens, Malvern, PA, USA) with phased-array coil systems. A typical CMR protocol was as follows: first, localizers were acquired to identify the cardiac position. Next, cine CMR images were acquired in the short-axis (every 10 mm to cover the entire LV from the mitral valve plane through the apex) and three long-axis views (two-, three-, and four-chamber) using a steady-state free precession sequence. Standard LGE CMR imaging was performed 10–15 min after administration of gadolinium contrast (0.15 mmol/kg), using a two-dimensional segmented inversion-recovery gradient-echo sequence in identical views as cine CMR imaging. Typical inversion delay times were 280–360 ms. The same CMR protocol was used during the entire study period.
CMR analyses
CMRs were analysed by a study investigator with expertise in CMR, blinded to all other information. EFs were determined by quantitative analysis according to standard recommendations.12 LV and RV volumes were quantified by planimetry of the end-diastolic and -systolic endocardial borders on a stack of short-axis cine CMR images acquired from the base to the apex, which were used to calculate the respective end-diastolic and -systolic volumes. Stroke volumes were calculated by subtracting the end-systolic volumes from the end-diastolic volumes. EFs were calculated by dividing the stroke volume by the end-diastolic volumes. The presence of LGE was determined visually. Normal ranges from a comprehensive systematic review of CMR studies reporting normal values were used to classify LVEF and RVEF as normal or abnormal.13 Precession (Heart Imaging Technologies, Durham, NC, USA) was used for analyses.
Clinical follow-up and outcomes
Follow-up data were collected through a review of electronic medical records blinded to CMR data. The pre-specified primary endpoint was a composite of all-cause death or major adverse cardiac events (MACE) after the CMR: heart failure hospitalization, heart transplantation, ventricular assist device implantation, resuscitated cardiac arrest, or life-threatening ventricular arrhythmia. When a recipient experienced more than one event, the first event was chosen for outcome analyses. The secondary endpoints were all-cause death and cardiac death or MACE. Mortality status and death dates were cross-checked with data from the Minnesota Department of Health’s Office of Vital Records. The cause of death was adjudicated by two investigators after a review of the electronic medical records and death certificates, with a third opinion in cases of disagreement.
Statistical analysis
Normally distributed continuous variables were expressed as mean ± standard deviation (SD), and non-normally distributed continuous variables were presented as medians with interquartile range (IQR). Categorical variables were expressed as counts with percentages. Comparison between groups was performed with a two-sample Student’s t-test for continuous, normal variables, and Mann–Whitney rank sum test for continuous, non-normal data. Pearson χ2 tests were used to compare discrete data between groups; in those cases where the expected cell count was <5, Fisher exact test was used. Pearson correlation analysis was used to examine relationships between LVEF and RVEF. Logistic regression analysis was used to identify correlates of RV systolic dysfunction. Kaplan–Meier analyses and unadjusted and adjusted Cox proportional hazards regression analyses were used to evaluate the relationships between clinical variables and the primary endpoint of all-cause death or MACE, and the secondary endpoint of all-cause death. For the secondary endpoint of cardiac death or MACE, non-cardiac death was considered a competing risk; accordingly, cumulative incidence functions were estimated, and Fine–Gray subdistribution hazard modelling was performed. The assumption of proportional hazards was assessed by plotting the scaled Schoenfeld residuals for each independent variable against time; these correlations were found to be non-significant for all variables included in the multivariable models. All eligible variables were tested for collinearity using the variance inflation factor measure before being included in the final multivariable models. All multivariable models showed acceptably low multicollinearity. All tests were two-tailed. A P-value of <0.05 was used to denote statistical significance. Analyses were performed using RStudio, version 1.4.1103 (RStudio, Inc.).
Results
Study cohort
Three hundred and fifteen consecutive survivors with breast cancer, lymphoma, leukaemia, or sarcoma had CMRs after receiving anthracyclines. Of these, 21 were excluded because they received anthracyclines after the CMR, 19 were excluded because they received trastuzumab after the CMR, and 26 were excluded because most had LGE that was suggestive of a cardiomyopathy unrelated to anthracyclines. Of the 26, 16 had ischaemic cardiomyopathy, 3 had myocarditis, 2 had cardiac sarcoidosis, 1 had lymphoma infiltration, 1 had hypertensive cardiomyopathy, and the cause of the LGE was unclear in 3. Thus, 249 cancer survivors formed the study cohort (Figure 1).
Figure 1.
Study patients. Flow chart of study patients.
Patient characteristics
Patient characteristics at the time of the CMR are provided in Table 1. The mean age of the cohort was 55 years and 40% were men. The most prevalent cancers were lymphoma and breast cancer. The CMRs were done at a median of 2.9 years after their cancer treatment. In addition to anthracyclines, 38% received radiation therapy to the chest and 9% received trastuzumab.
Table 1.
Patient characteristics at the time of the CMR
| All survivors (n = 249) | Abnormal RVEF (n = 54) | Normal RVEF (n = 195) | |
|---|---|---|---|
| Age, mean (SD), years | 54.5 (15.3) | 53.5 (16.9) | 54.8 (14.9) |
| Women, n (%) | 149 (59.8) | 26 (48.1) | 74 (37.9) |
| Body mass index, median (IQR), kg/m2 | 28.1 (24.4, 32.1) | 29.5 (25.3, 32.1) | 28.0 (24.2, 32.0) |
| Cancer type | |||
| Lymphoma, n (%) | 92 (36.9) | 26 (48.1) | 66 (33.8) |
| Breast cancer, n (%) | 85 (34.1) | 11 (20.4) | 74 (37.9) |
| Leukaemia, n (%) | 54 (21.7) | 12 (22.2) | 42 (21.5) |
| Sarcoma, n (%) | 18 (7.2) | 5 (9.3) | 13 (6.7) |
| Cancer treatment | |||
| Time since cancer treatment, median (IQR), years | 2.9 (0.9, 9.2) | 3.4 (0.8, 10.1) | 2.8 (0.9, 8.9) |
| Cumulative anthracycline dose, median (IQR), mg/m2 | 240 (200, 300) | 245 (207, 300) | 240 (186, 300) |
| Radiation therapy involving the chest, n (%) | 95 (38.2) | 16 (29.6) | 79 (40.5) |
| Trastuzumab, n (%) | 23 (9.2) | 3 (5.6) | 20 (10.3) |
| Comorbidities | |||
| Hypertension, n (%) | 109 (43.8) | 28 (51.9) | 81 (41.5) |
| Dyslipidaemia, n (%) | 125 (50.2) | 22 (40.7) | 103 (52.8) |
| Diabetes mellitus, n (%) | 43 (17.3) | 9 (16.7) | 34 (17.4) |
| Former tobacco use, n (%) | 94 (37.8) | 17 (31.5) | 77 (39.5) |
| Current tobacco use, n (%) | 15 (6.0) | 2 (3.7) | 13 (6.7) |
| Coronary artery disease, n (%) | 19 (7.6) | 4 (7.4) | 15 (7.7) |
| Heart failure, n (%) | 68 (27.3) | 24 (44.4) | 44 (22.6) |
| Cerebrovascular disease, n (%) | 7 (2.8) | 3 (5.6) | 4 (2.1) |
| Peripheral vascular disease, n (%) | 6 (2.4) | 2 (3.7) | 4 (2.1) |
| Chronic obstructive pulmonary disease, n (%) | 18 (7.2) | 2 (3.7) | 16 (8.2) |
| Obstructive sleep apnoea, n (%) | 35 (14.1) | 6 (11.1) | 29 (14.9) |
| Cardiac symptoms | |||
| NYHA functional class (I/II/III/IV), n (%) | 31/22/14/1 (12.4/8.8/5.6/0.4) | 9/8/6/1 (16.7/14.8/11.1/1.9) | 22/14/8/0 (11.3/7.2/4.1/0.0) |
| Cardiac medications | |||
| Aspirin, n (%) | 71 (28.5) | 15 (27.8) | 56 (28.7) |
| Beta-blockers, n (%) | 99 (39.8) | 21 (38.9) | 78 (40.0) |
| ACE-inhibitor/ARB, n (%) | 89 (35.7) | 20 (37.0) | 69 (35.4) |
| Statin, n (%) | 67 (26.9) | 11 (20.4) | 56 (28.7) |
| Diuretic, n (%) | 48 (19.3) | 17 (31.5) | 31 (15.9) |
| Mineralocorticoid receptor antagonist, n (%) | 12 (4.8) | 4 (7.4) | 8 (4.1) |
ACE, angiotensin-converting enzyme, ARB, angiotensin receptor blocker; CMR, cardiovascular magnetic resonance; IQR, interquartile range; NYHA, New York Heart Association; RVEF, right ventricular ejection fraction; SD, standard deviation.
Cardiac risk factors were common (44% hypertension, 50% dyslipidaemia, 17% diabetes mellitus, and 38% former tobacco use). At the time of the CMR, 27% had been diagnosed with heart failure.
CMR findings
The CMR findings are listed in Table 2. The CMRs were done at a median of 2.9 years since their cancer treatment. The median LVEF was 51.2% and the median RVEF was 57.7%. The prevalence of LV dysfunction defined as an LV ejection fraction (LVEF) <57%13 was 69.9%.
Table 2.
CMR markers
| All survivors (n = 249) | Abnormal RVEF (n = 54) | Normal RVEF (n = 195) | |
|---|---|---|---|
| LVEDVI, median (IQR), mL/m2 | 64.2 (53.6, 80.0) | 65.9 (47.7, 81.8) | 64.2 (55.4, 77.2) |
| LVESVI, median (IQR), mL/m2 | 30.0 (23.7, 42.3) | 36.5 (26.2, 55.9) | 29.3 (23.3, 40.1) |
| LVEF, median (IQR), % | 51.2 (42.4, 57.5) | 40.9 (31.3, 49.3) | 53.6 (45.9, 58.9) |
| Abnormal LVEF, n (%) | 174 (69.9) | 50 (92.6) | 124 (63.6) |
| RVEDVI, median (IQR), mL/m2 | 56.7 (47.0, 67.2) | 58.5 (45.8, 70.4) | 56.4 (47.1, 65.4) |
| RVESVI, median (IQR), mL/m2 | 22.7 (18.2, 30.8) | 34.1 (25.5, 43.5) | 21.1 (16.4, 26.5) |
| RVEF, median (IQR), % | 57.7 (51.1, 65.5) | 45.5 (39.3, 47.5) | 60.6 (55.9, 66.7) |
CMR, cardiovascular magnetic resonance; IQR, interquartile range; LVEDVI, left ventricular end-diastolic volume index; LVEF, left ventricular ejection fraction; LVESVI, left ventricular end-systolic volume index; RVEDVI, right ventricular end-diastolic volume index; RVEF, right ventricular ejection fraction; RVESVI, right ventricular end-systolic volume index.
Prevalence of RV systolic dysfunction and relation with LV systolic dysfunction
RV systolic dysfunction, defined as an RV ejection fraction (RVEF) <51%,13 was present in 54 (21.7%) of cancer survivors. In these survivors, the median RVEF was 45.5%. Of the 54 survivors with RV systolic dysfunction, 50 (92.6%) also had concomitant LV systolic dysfunction.
Among the 174 survivors with LV dysfunction, the prevalence of RV systolic dysfunction was 28.7%. The prevalence of severe (≤35%) systolic dysfunction was lower with the RV compared with the LV (3.6% vs. 12.9%; P < 0.001) (Figure 2 Top). There was a moderate, positive correlation between RVEF and LVEF (r = 0.50; P < 0.001) (Figure 2 Bottom).
Figure 2.
Comparison of RVEF and LVEF. (Top) Distribution of cancer survivors by LVEF and RVEF categories. The number of survivors in each LVEF category is noted in green and the number in each RVEF category is noted in orange. (Bottom) Scatterplot showing the correlation between LVEF and RVEF.
Correlates of RV systolic dysfunction
On unadjusted logistic regression analyses, correlates of RV systolic dysfunction were heart failure at the time of CMR and lower LVEF (Table 3). On adjusted analyses, the correlates were male sex [odds ratio (OR) 2.19, 95% confidence interval (CI) 1.05–4.59; P = 0.037] and lower LVEF (OR 1.54, 95% CI 1.29–1.83; P < 0.001) (Table 3).
Table 3.
Correlates of RV dysfunction
| Univariable | Multivariablea | |||
|---|---|---|---|---|
| OR (95% CI) | P-value | OR (95% CI) | P-value | |
| Age (per 10-year increase) | 0.95 (0.78–1.15) | 0.58 | 0.92 (0.74–1.15) | 0.46 |
| Male sex | 1.52 (0.83–2.79) | 0.18 | 2.19 (1.05–4.59) | 0.037 |
| Body mass index (per 1 kg/m2 increase) | 1.01 (0.96–1.06) | 0.68 | – | – |
| Hypertension | 1.52 (0.83–2.78) | 0.18 | – | – |
| Diabetes mellitus | 0.95 (0.42–2.12) | 0.90 | – | – |
| Coronary artery disease | 0.96 (0.31–3.02) | 0.94 | – | – |
| Chronic obstructive pulmonary disease | 0.43 (0.10–1.93) | 0.27 | – | – |
| Obstructive sleep apnea | 0.72 (0.28–1.82) | 0.48 | – | – |
| Time between anthracycline treatment and CMR (per 1-year increase) | 1.00 (0.96–1.05) | 0.85 | – | – |
| Cumulative anthracycline dose (per 10 mg/m2 increase) | 1.02 (0.99–1.05) | 0.21 | 1.01 (0.98–1.04) | 0.64 |
| Chest radiation therapy | 0.62 (0.32–1.18) | 0.15 | – | - |
| Trastuzumab | 0.51 (0.15–1.80) | 0.30 | – | – |
| Heart failure diagnosed before the CMR | 2.75 (1.46–5.17) | 0.002 | 1.34 (0.56–3.19) | 0.51 |
| LVEF (per 5% decrease) | 1.53 (1.32–1.77) | <0.001 | 1.54 (1.29–1.83) | <0.001 |
The multivariable model included age, male sex, cumulative anthracycline dose, heart failure diagnosed before the CMR, and LVEF. Bold indicates values with statistical significance (p < 0.05).
CMR, cardiovascular magnetic resonance; LVEF, left ventricular ejection fraction; RV, right ventricular.
Clinical follow-up and outcomes
The median follow-up time after the CMR was 2.7 years (IQR 1.0–5.9 years). No survivors were lost to follow-up. During follow-up, 105 survivors experienced the composite endpoint of all-cause death or MACE. There were 89 deaths, 41 heart failure hospitalizations, 2 heart transplantations, 2 left ventricular assist device implantations, and 1 resuscitated asystolic cardiac arrest. There were no instances of life-threatening ventricular arrhythmia. Of the 89 deaths, the cause of death was adjudicated as cardiac in 12 (13.5%), non-cardiac in 74 (83.1%), and unclear in 3 (3.4%).
Associations of RV systolic dysfunction with all-cause death or MACE
On Kaplan–Meier analyses, the cumulative incidence of all-cause death or MACE was significantly higher in survivors with abnormal RVEF compared with those with normal RVEF (P = 0.002) (Figure 3A).
Figure 3.
Survival curves for abnormal and normal RVEF. (A) Primary endpoint of all-cause death or MACE. Kaplan–Meier estimates of the cumulative incidence are represented in orange for abnormal RVEF and green for normal RVEF. (B) Secondary endpoint of all-cause death. Kaplan–Meier estimates of the cumulative incidence are represented in orange for abnormal RVEF and green for normal RVEF. (C) Secondary endpoint of cardiac death or MACE. Cumulative incidence function estimates are represented in orange for abnormal RVEF and green for normal RVEF, solid lines for cardiac death or MACE, and dashed lines for non-cardiac death.
On Cox proportional hazards regression analyses (Table 4), on univariable analyses, RVEF was associated with all-cause death or MACE. On multivariable analysis, RVEF was independently associated with the composite endpoint after adjustment for non-imaging risk factors with a hazard ratio (HR) of 1.11 for every 5% decrease (95% CI 1.01–1.22; P = 0.035). However, when LVEF was added to the multivariable model, RVEF was no longer associated with the composite endpoint (HR 1.04 for every 5% decrease, 95% CI 0.93–1.17; P = 0.46). LVEF was significantly associated with the composite endpoint with a HR of 1.14 for every 5% decrease (95% CI 1.03–1.25; P = 0.009).
Table 4.
Association of CMR markers with all-cause death or MACE
| Univariable | Multivariable 1a | Multivariable 2b | ||||
|---|---|---|---|---|---|---|
| HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value | |
| Lymphoma (relative to breast cancer) | 1.75 (1.03–2.99) | 0.039 | 1.61 (0.85–3.04) | 0.14 | 1.70 (0.90–3.22) | 0.10 |
| Leukaemia (relative to breast cancer) | 4.23 (2.47–7.24) | <0.001 | 2.65 (1.24–5.69) | 0.012 | 2.59 (1.21–5.56) | 0.014 |
| Sarcoma (relative to breast cancer) | 2.29 (1.05–5.01) | 0.038 | 2.45 (0.97–6.20) | 0.058 | 2.62 (1.02–6.72) | 0.044 |
| Age (per 10-year increase) | 1.12 (0.98–1.29) | 0.089 | 1.08 (0.93–1.25) | 0.33 | 1.09 (0.93–1.26) | 0.28 |
| Time between anthracycline treatment and CMR (per 1-year increase) | 0.95 (0.92–0.99) | 0.006 | 0.96 (0.93–0.99) | 0.016 | 0.96 (0.92–0.99) | 0.011 |
| Cumulative anthracycline dose (per 10 mg/m2 increase) | 0.96 (0.94–0.98) | <0.001 | 0.98 (0.96–1.01) | 0.13 | 0.98 (0.96–1.00) | 0.043 |
| Chest radiation therapy | 0.46 (0.29–0.71) | <0.001 | 1.11 (0.62–1.99) | 0.73 | 1.15 (0.64–2.06) | 0.65 |
| Trastuzumab | 0.23 (0.07–0.73) | 0.012 | 0.33 (0.09–1.13) | 0.077 | 0.36 (0.10–1.26) | 0.11 |
| Heart failure diagnosed before the CMR | 1.89 (1.27–2.81) | 0.002 | 1.74 (1.13–2.68) | 0.012 | 1.36 (0.84–2.20) | 0.21 |
| LVEDVI (per 10 mL/m2 increase) | 1.05 (0.95–1.15) | 0.35 | – | – | – | – |
| LVEF (per 5% decrease) | 1.17 (1.09–1.27) | <0.001 | – | - | 1.14 (1.03–1.25) | 0.009 |
| RVEDVI (per 10 mL/m2 increase) | 1.00 (0.89–1.13) | 0.99 | – | - | – | – |
| RVEF (per 5% decrease) | 1.16 (1.06–1.27) | 0.001 | 1.11 (1.01–1.22) | 0.035 | 1.04 (0.93–1.17) | 0.46 |
Multivariable 1 model included non-imaging variables (cancer type, age, time between anthracycline treatment and CMR, cumulative anthracycline dose, chest radiation therapy, trastuzumab, and heart failure diagnosed before the CMR), and RVEF.
Multivariable 2 model included non-imaging variables, LVEF, and RVEF. Bold indicates values with statistical significance (p < 0.05).
CI, confidence interval; CMR, cardiovascular magnetic resonance; HR, hazard ratio; LVEDVI, left ventricular end-diastolic volume index; LVEF, left ventricular ejection fraction; MACE, major adverse cardiac events; RVEDVI, right ventricular end-diastolic volume index; RVEF, right ventricular ejection fraction.
Associations of RV systolic dysfunction with all-cause death
On Kaplan–Meier analyses, the cumulative incidence of all-cause death was not significantly different in abnormal RVEF compared with those with normal RVEF (P = 0.11) (Figure 3B).
On Cox proportional hazards regression analyses (Table 5), RVEF was not associated with all-cause death on either univariable or multivariable analyses.
Table 5.
Association of CMR markers with all-cause death
| Univariable | Multivariable 3a | Multivariable 4b | ||||
|---|---|---|---|---|---|---|
| HR (95% CI) | P-value | HR (95% CI) | P-value | HR (95% CI) | P-value | |
| Lymphoma (relative to breast cancer) | 1.71 (0.93–3.15) | 0.085 | 1.20 (0.60–2.40) | 0.61 | 1.22 (0.61–2.44) | 0.58 |
| Leukaemia (relative to breast cancer) | 5.17 (2.86–9.37) | <0.001 | 2.07 (0.96–4.50) | 0.065 | 2.03 (0.94–4.41) | 0.072 |
| Sarcoma (relative to breast cancer) | 2.72 (1.16–6.36) | 0.021 | 1.93 (0.75–4.99) | 0.17 | 1.94 (0.75–5.02) | 0.17 |
| Age (per 10-year increase) | 1.06 (0.92–1.22) | 0.44 | ||||
| Time between anthracycline treatment and CMR (per 1-year increase) | 0.94 (0.90–0.97) | 0.001 | 0.94 (0.91–0.98) | 0.005 | 0.94 (0.91–0.98) | 0.006 |
| Cumulative anthracycline dose (per 10 mg/m2 increase) | 0.95 (0.93–0.97) | <0.001 | 0.97 (0.95–0.99) | 0.012 | 0.97 (0.95–0.99) | 0.009 |
| Chest radiation therapy | 0.35 (0.21–0.59) | <0.001 | 0.83 (0.43–1.58) | 0.56 | 0.83 (0.43–1.58) | 0.57 |
| Trastuzumab | 0.09 (0.01–0.67) | 0.019 | 0.12 (0.02–0.95) | 0.044 | 0.13 (0.02–0.98) | 0.047 |
| Heart failure diagnosed before the CMR | 1.16 (0.74–1.84) | 0.53 | 1.12 (0.69–1.82) | 0.65 | 1.03 (0.59–1.77) | 0.93 |
| LVEDVI (per 10 mL/m2 increase) | 0.96 (0.86–1.07) | 0.44 | ||||
| LVEF (per 5% decrease) | 0.97 (1.06–1.15) | 0.19 | 1.04 (0.93–1.16) | 0.49 | ||
| RVEDVI (per 10 mL/m2 increase) | 1.00 (0.87–1.14) | 0.97 | ||||
| RVEF (per 5% decrease) | 1.10 (1.00–1.21) | 0.058 | 1.09 (0.98–1.21) | 0.12 | 1.07 (0.96–1.20) | 0.24 |
Multivariable 3 model included non-imaging variables (cancer type, age, time between anthracycline treatment and CMR, cumulative anthracycline dose, chest radiation therapy, trastuzumab, and heart failure diagnosed before the CMR), and RVEF.
Multivariable 4 model included non-imaging variables, LVEF, and RVEF. Bold indicates values with statistical significance (p < 0.05).
CI, confidence interval; CMR, cardiovascular magnetic resonance; HR, hazard ratio; LVEDVI, left ventricular end-diastolic volume index; LVEF, left ventricular ejection fraction; MACE, major adverse cardiac events; RVEDVI, right ventricular end-diastolic volume index; RVEF, right ventricular ejection fraction.
Associations of RV systolic dysfunction with cardiac death or MACE
Cumulative incidences of cardiac death or MACE and non-cardiac death are shown in Figure 3C. The cumulative incidence of cardiac death or MACE was higher in survivors with abnormal RVEF compared with those with normal RVEF (P < 0.001). The cumulative incidence of non-cardiac death was not different between survivors with and without abnormal RVEF (P = 0.89).
On Fine–Gray subdistribution hazard modelling (Table 6), on univariable analyses, RVEF was associated with cardiac death or MACE. On multivariable analysis, RVEF was independently associated with cardiac death or MACE after adjustment for non-imaging risk factors with a subdistribution hazard ratio (sHR) of 1.36 for every 5% decrease (95% CI 1.18–1.56; P < 0.001). However, when LVEF was added to the multivariable model, RVEF was no longer associated with cardiac death or MACE (sHR 1.16 for every 5% decrease; 95% CI 0.97–1.40; P = 0.10). LVEF was significantly associated with cardiac death or MACE with a sHR of 1.34 for every 5% decrease (95% CI 1.16–1.56; P < 0.001).
Table 6.
Association of CMR markers with cardiac death or MACE
| Univariable | Multivariable 5a | Multivariable 6b | ||||
|---|---|---|---|---|---|---|
| sHR (95% CI) | P-value | sHR (95% CI) | P-value | sHR (95% CI) | P-value | |
| Lymphoma (relative to breast cancer) | 1.21 (0.62–2.38) | 0.59 | 1.09 (0.55–2.15) | 0.80 | 1.06 (0.52–2.15) | 0.88 |
| Leukaemia (relative to breast cancer) | 1.38 (0.63–3.01) | 0.42 | 1.00 (0.43–2.32) | 1.00 | 0.86 (0.36–2.05) | 0.73 |
| Sarcoma (relative to breast cancer) | 0.30 (0.04–2.16) | 0.23 | 0.21 (0.03–1.43) | 0.11 | 0.21 (0.03–1.31) | 0.095 |
| Time between anthracycline treatment and CMR (per 1-year increase) | 1.01 (0.98–1.05) | 0.44 | 1.03 (0.99–1.08) | 0.14 | 1.02 (0.98–1.07) | 0.35 |
| Cumulative anthracycline dose (per 10 mg/m2 increase) | 0.99 (0.96–1.01) | 0.23 | 1.03 (0.99–1.08) | 0.066 | 0.96 (0.94–0.99) | 0.016 |
| LVEF (per 5% decrease) | 1.40 (1.25–1.58) | <0.001 | 1.34 (1.16–1.56) | <0.001 | ||
| RVEF (per 5% decrease) | 1.31 (1.14–1.50) | <0.001 | 1.36 (1.18–1.56) | <0.001 | 1.16 (0.97–1.40) | 0.10 |
Multivariable 5 model included non-imaging variables (cancer type, time between anthracycline treatment and CMR, and cumulative anthracycline dose), and RVEF.
Multivariable 6 model included non-imaging variables, LVEF, and RVEF. Bold indicates values with statistical significance (p < 0.05).
CI, confidence interval; CMR, cardiovascular magnetic resonance; LVEF, left ventricular ejection fraction; MACE, major adverse cardiac events; RVEF, right ventricular ejection fraction; sHR, sub-distribution hazard ratio.
Discussion
In a large cohort of 249 cancer survivors who had CMR for suspected cardiotoxicity at a median of 2.9 years after anthracycline treatment, we performed the first large systematic investigation into the prognostic value of RV systolic function. The prevalence of RV systolic dysfunction was 22% compared with a prevalence of 70% for LV systolic dysfunction. Among those with RV systolic dysfunction, 93% also had concomitant LV systolic dysfunction. Determinants of RV systolic dysfunction were male sex and LVEF. Decreased RVEF was associated with the primary endpoint of all-cause death or MACE at a median follow-up of 2.7 years after the CMR, independent of non-imaging variables, but importantly, not independent of LVEF. Decreased RVEF was also not independently associated with the secondary endpoints of all-cause death and cardiac death or MACE.
Our study has several strengths. First, we assessed RV systolic function using CMR, widely considered as the reference standard imaging technique for the assessment of ventricular size, function, and structure.14 For the assessment of the RV, CMR is superior to echocardiography. Echocardiography is limited for the assessment of the RV due to its shape—the RV has a highly variable shape, a relatively thin free wall, and heavy trabeculations—and its location—the RV is in the near field of parasternal echocardiographic windows, and in apical views, it may be obscured by the ribs, the sternum, or the lungs.15 Second, we studied hard outcomes of death or MACE, rather than cardiotoxicity defined by reductions in LVEF as is done in most studies of anthracycline-treated cancer patients. LVEF-based definitions of cardiotoxicity may be confounded by measurement variability, particularly with echocardiography,16,17 the possibility of stress-induced cardiomyopathy18 being mistaken for cardiotoxicity, and end-diastolic volume-mediated reductions in EF.19 Finally, we studied a large number of cancer survivors with suspected cardiotoxicity who had a substantial number of adverse outcomes during follow-up. This allowed for robust adjustments of potential confounders in the multivariable models.
While there have been no prior studies of the long-term prognostic implications of RV systolic dysfunction, a few studies have examined its prevalence in anthracycline-treated cancer survivors. In a study of 62 long-term survivors of childhood cancer studied using CMR at a median of 7.8 years after anthracycline therapy, Ylanen et al.10 described LVEF <55% in 79% and RVEF <55% in 81%. These high rates of LV and RV systolic dysfunction have not been described in subsequent papers. In a study of 246 survivors of childhood lymphoma and acute lymphoblastic leukaemia (77% treated with anthracyclines) examined by echocardiography at a mean of 21.7 years after diagnosis, Christiansen et al.9 found RV systolic dysfunction in 30% of the survivors. Consistent with our findings, RV systolic dysfunction was associated with cardiotoxic cancer treatment and LV systolic dysfunction. Contradictory to our findings, LV systolic dysfunction was present in only 11% of survivors, a third of the prevalence of RV systolic dysfunction. In a third study of 274 adult lymphoma survivors (all treated with anthracyclines) who had echocardiography at a mean of 13 years after lymphoma diagnosis, Murbraech et al.20 described LV systolic dysfunction in 30.8% and RV systolic dysfunction in 6.2%. Of the 17 survivors with RV systolic dysfunction, 15 (88.2%) also had LV systolic dysfunction. These findings of a lower prevalence of RV systolic dysfunction than LV systolic dysfunction, and the presence of LV systolic dysfunction in nearly all patients with RV systolic dysfunction are similar to ours.
We found that only 22% of all survivors and 29% of survivors with LV systolic dysfunction had RV systolic dysfunction. This raises the possibility that the RV is less susceptible than the LV to direct permanent injury from anthracyclines. While it is presumed that the RV may be more sensitive to the toxic effects of anthracycline than the LV due to its thinner walls and fewer myofibrils, the RV has also been shown to display remarkable resiliency in the face of various insults, which is related to unique anatomic, physiologic, and genetic factors that differentiate it from the LV.21 The LV and the RV have been shown to have markedly different propensities for injury in animal studies of chronic anthracycline cardiotoxicity.22–24 In rabbits with chronic anthracycline cardiotoxicity, Lenčová-Popelová et al.24 described profound molecular remodelling of myocytes, non-myocyte cells, and extracellular matrix in the LV, with distinctly weaker or completely absent changes in the RV. While RV systolic dysfunction was infrequent in survivors with LV systolic dysfunction in our study, the converse was not true—93% of survivors with RV systolic dysfunction had LV systolic dysfunction. This suggests two possible causes for RV systolic dysfunction in anthracycline-treated cancer survivors (i) direct toxicity from anthracyclines, manifesting less often in the RV than in the LV and (ii) a consequence of LV systolic dysfunction manifesting through multiple mechanisms including increased afterload from secondary pulmonary hypertension, and ventricular interdependence from septal dysfunction.3 Future studies should investigate the relative contributions of these two causes to RV systolic dysfunction in anthracycline-treated cancer survivors. They should also examine the prognostic implications of RV dysfunction in all-comer cancer survivors treated with anthracyclines.
Limitations
Our cohort consisted of survivors referred clinically for a CMR for suspected cardiotoxicity. Thus, our results may not be generalizable to all-comer cancer survivors treated with anthracyclines. RV-focused cine CMR images (placing the basal short-axis slice immediately on the myocardial side of the RV)25 were not acquired. Discerning the cause of death may be difficult,26 particularly in cancer survivors with cardiac disease. For instance, it is challenging to distinguish between cardiac and cancer causes of death in a patient who died of advanced cancer after their anthracycline-based cancer treatment was prematurely stopped due to cardiotoxicity. Accordingly, we used all-cause death in our primary endpoint. All-cause mortality as an outcome is objective and free of bias that could arise from the adjudication of the cause of death.26 The composite endpoint of all-cause death or MACE is well accepted in heart failure trials27 and studies of cancer treatment-related cardiotoxicity.2 Finally, a recent large study of cancer survivors demonstrated that severe cardiotoxicity defined as heart failure or an asymptomatic decrease in LVEF to <40% was associated with a 10-fold increase in all-cause death compared with survivors without or with milder forms of cardiotoxicity.2 We, therefore, believe that all-cause death or MACE is a valid and relevant primary endpoint for our study.
Conclusions
Among anthracycline-treated cancer survivors undergoing CMR for suspected cardiotoxicity, RV systolic dysfunction was present in 22% and was accompanied by LV systolic dysfunction in 93% of cases. While RV systolic dysfunction had prognostic value over non-imaging variables, it was not associated with the primary endpoint of all-cause death or MACE, at a median follow-up of 2.7 years after the CMR, independent of LVEF. Decreased RVEF was also not independently associated with the secondary endpoints of all-cause death, and cardiac death or MACE.
Contributor Information
Sanya Chhikara, Cardiovascular Division, Department of Medicine, University of Minnesota Medical School, 420 Delaware Street SE, MMC 508, Minneapolis, MN 55455, USA.
Matthew Hooks, Cardiovascular Division, Department of Medicine, University of Minnesota Medical School, 420 Delaware Street SE, MMC 508, Minneapolis, MN 55455, USA.
Pal Satyajit Singh Athwal, Cardiovascular Division, Department of Medicine, University of Minnesota Medical School, 420 Delaware Street SE, MMC 508, Minneapolis, MN 55455, USA.
Andrew Hughes, Cardiovascular Division, Department of Medicine, University of Minnesota Medical School, 420 Delaware Street SE, MMC 508, Minneapolis, MN 55455, USA.
Mohamed F Ismail, Cardiovascular Division, Department of Medicine, University of Minnesota Medical School, 420 Delaware Street SE, MMC 508, Minneapolis, MN 55455, USA.
Stephanie Joppa, Cardiovascular Division, Department of Medicine, University of Minnesota Medical School, 420 Delaware Street SE, MMC 508, Minneapolis, MN 55455, USA.
Pratik S Velangi, Cardiovascular Division, Department of Medicine, University of Minnesota Medical School, 420 Delaware Street SE, MMC 508, Minneapolis, MN 55455, USA.
Prabhjot S Nijjar, Cardiovascular Division, Department of Medicine, University of Minnesota Medical School, 420 Delaware Street SE, MMC 508, Minneapolis, MN 55455, USA.
Anne H Blaes, Division of Hematology, Oncology and Transplantation, Department of Medicine, University of Minnesota Medical School, 420 Delaware Street SE, MMC 508, Minneapolis, MN 55455, USA.
Chetan Shenoy, Cardiovascular Division, Department of Medicine, University of Minnesota Medical School, 420 Delaware Street SE, MMC 508, Minneapolis, MN 55455, USA.
Acknowledgements
This work was supported by National Institutes of Health grant K23HL132011, a University of Minnesota Clinical and Translational Science Institute KL2 Scholars Career Development Program Award (National Institutes of Health grant KL2TR000113-05), and a University of Minnesota Clinical and Translational Science Institute K-R01 Transition to Independence Grant (supported by the National Institutes of Health grant UL1TR002494) to Chetan Shenoy.
Data Availability
The data underlying this article cannot be shared publicly due to privacy reasons. De-identified data will be shared on reasonable request to the corresponding author.
References
- 1. Bowles EJ, Wellman R, Feigelson HS, Onitilo AA, Freedman AN, Delate T, et al. ; Pharmacovigilance Study Team . Risk of heart failure in breast cancer patients after anthracycline and trastuzumab treatment: a retrospective cohort study. J Natl Cancer Inst 2012;104:1293–305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Lopez-Sendon J, Alvarez-Ortega C, Zamora Aunon P, Buno Soto A, Lyon AR, Farmakis Det al. Classification, prevalence, and outcomes of anticancer therapy-induced cardiotoxicity: the CARDIOTOX registry. Eur Heart J 2020;41:1720–9. [DOI] [PubMed] [Google Scholar]
- 3. Sanz J, Sanchez-Quintana D, Bossone E, Bogaard HJ, Naeije R.. Anatomy, function, and dysfunction of the right ventricle: JACC state-of-the-art review. J Am Coll Cardiol 2019;73:1463–82. [DOI] [PubMed] [Google Scholar]
- 4. Purmah Y, Lei LY, Dykstra S, Mikami Y, Cornhill A, Satriano Aet al. Right ventricular ejection fraction for the prediction of major adverse cardiovascular and heart failure-related events: a cardiac MRI based study of 7131 patients with known or suspected cardiovascular disease. Circ Cardiovasc Imaging 2021;14:e011337. [DOI] [PubMed] [Google Scholar]
- 5. Boczar KE, Aseyev O, Sulpher J, Johnson C, Burwash IG, Turek Met al. Right heart function deteriorates in breast cancer patients undergoing anthracycline-based chemotherapy. Echo Res Pract 2016;3:79–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Abdar Esfahani M, Mokarian F, Karimipanah M.. Alterations in the echocardiographic variables of the right ventricle in asymptomatic patients with breast cancer during anthracycline chemotherapy. Postgrad Med J 2017;93:271–4. [DOI] [PubMed] [Google Scholar]
- 7. Tanindi A, Demirci U, Tacoy G, Buyukberber S, Alsancak Y, Coskun Uet al. Assessment of right ventricular functions during cancer chemotherapy. Eur J Echocardiogr 2011;12:834–40. [DOI] [PubMed] [Google Scholar]
- 8. Grover S, Leong DP, Chakrabarty A, Joerg L, Kotasek D, Cheong Ket al. Left and right ventricular effects of anthracycline and trastuzumab chemotherapy: a prospective study using novel cardiac imaging and biochemical markers. Int J Cardiol 2013;168:5465–7. [DOI] [PubMed] [Google Scholar]
- 9. Christiansen JR, Massey R, Dalen H, Kanellopoulos A, Hamre H, Ruud Eet al. Right ventricular function in long-term adult survivors of childhood lymphoma and acute lymphoblastic leukaemia. Eur Heart J Cardiovasc Imaging 2016;17:735–41. [DOI] [PubMed] [Google Scholar]
- 10. Ylanen K, Poutanen T, Savikurki-Heikkila P, Rinta-Kiikka I, Eerola A, Vettenranta K.. Cardiac magnetic resonance imaging in the evaluation of the late effects of anthracyclines among long-term survivors of childhood cancer. J Am Coll Cardiol 2013;61:1539–47. [DOI] [PubMed] [Google Scholar]
- 11. Modi K, Joppa S, Chen KA, Athwal PSS, Okasha O, Velangi PSet al. Myocardial damage assessed by late gadolinium enhancement on cardiovascular magnetic resonance imaging in cancer patients treated with anthracyclines and/or trastuzumab. Eur Heart J Cardiovasc Imaging 2021;22:427–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Schulz-Menger J, Bluemke DA, Bremerich J, Flamm SD, Fogel MA, Friedrich MGet al. Standardized image interpretation and post processing in cardiovascular magnetic resonance: society for Cardiovascular Magnetic Resonance (SCMR) board of trustees task force on standardized post processing. J Cardiovasc Magn Reson 2013;15:35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Kawel-Boehm N, Maceira A, Valsangiacomo-Buechel ER, Vogel-Claussen J, Turkbey EB, Williams Ret al. Normal values for cardiovascular magnetic resonance in adults and children. J Cardiovasc Magn Reson 2015;17:29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Huang H, Nijjar PS, Misialek JR, Blaes A, Derrico NP, Kazmirczak Fet al. Accuracy of left ventricular ejection fraction by contemporary multiple gated acquisition scanning in patients with cancer: comparison with cardiovascular magnetic resonance. J Cardiovasc Magn Reson 2017;19:34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Ostenfeld E, Flachskampf FA.. Assessment of right ventricular volumes and ejection fraction by echocardiography: from geometric approximations to realistic shapes. Echo Res Pract 2015;2:R1–R11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Thavendiranathan P, Grant AD, Negishi T, Plana JC, Popovic ZB, Marwick TH.. Reproducibility of echocardiographic techniques for sequential assessment of left ventricular ejection fraction and volumes: application to patients undergoing cancer chemotherapy. J Am Coll Cardiol 2013;61:77–84. [DOI] [PubMed] [Google Scholar]
- 17. Lambert J, Lamacie M, Thampinathan B, Altaha MA, Esmaeilzadeh M, Nolan Met al. Variability in echocardiography and MRI for detection of cancer therapy cardiotoxicity. Heart 2020;106:817–23. [DOI] [PubMed] [Google Scholar]
- 18. Cammann VL, Sarcon A, Ding KJ, Seifert B, Kato K, Di Vece Det al. Clinical features and outcomes of patients with malignancy and takotsubo syndrome: observations from the international takotsubo registry. J Am Heart Assoc 2019;8:e010881. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Melendez GC, Sukpraphrute B, D'Agostino RB Jr, Jordan JH, Klepin HD, Ellis Let al. Frequency of left ventricular end-diastolic volume-mediated declines in ejection fraction in patients receiving potentially cardiotoxic cancer treatment. Am J Cardiol 2017;119:1637–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Murbraech K, Holte E, Broch K, Smeland KB, Holte H, Rosner Aet al. Impaired right ventricular function in long-term lymphoma survivors. J Am Soc Echocardiogr 2016;29:528–36. [DOI] [PubMed] [Google Scholar]
- 21. Brittain EL, Hemnes AR, Keebler M, Lawson M, Byrd BF 3rd, Disalvo T.. Right ventricular plasticity and functional imaging. Pulm Circ 2012;2:309–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Van Vleet JF, Ferrans VJ.. Clinical and pathologic features of chronic adriamycin toxicosis in rabbits. Am J Vet Res 1980;41:1462–9. [PubMed] [Google Scholar]
- 23. Van Vleet JF, Ferrans VJ, Weirich WE.. Cardiac disease induced by chronic adriamycin administration in dogs and an evaluation of vitamin E and selenium as cardioprotectants. Am J Pathol 1980;99:13–42. [PMC free article] [PubMed] [Google Scholar]
- 24. Lencova-Popelova O, Jirkovsky E, Mazurova Y, Lenco J, Adamcova M, Simunek Tet al. Molecular remodeling of left and right ventricular myocardium in chronic anthracycline cardiotoxicity and post-treatment follow up. PLoS One 2014;9:e96055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Kramer CM, Barkhausen J, Bucciarelli-Ducci C, Flamm SD, Kim RJ, Nagel E.. Standardized cardiovascular magnetic resonance imaging (CMR) protocols: 2020 update. J Cardiovasc Magn Reson 2020;22:17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Lauer MS, Blackstone EH, Young JB, Topol EJ.. Cause of death in clinical research: time for a reassessment? J Am Coll Cardiol 1999;34:618–20. [DOI] [PubMed] [Google Scholar]
- 27. Anker SD, Schroeder S, Atar D, Bax JJ, Ceconi C, Cowie MRet al. Traditional and new composite endpoints in heart failure clinical trials: facilitating comprehensive efficacy assessments and improving trial efficiency. Eur J Heart Fail 2016;18:482–9. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data underlying this article cannot be shared publicly due to privacy reasons. De-identified data will be shared on reasonable request to the corresponding author.