Imaging for cardiac dysfunction secondary to cancer treatment

\\(Background:\\) Cancer treatment-related cardiac dysfunction (CTRCD) and heart failure were first described in the published literature over 50 years ago. Once it was determined that CTRCD was frequently irreversible with a high mortality burden, cardiologists and oncologists worked together to utilize cardiac imaging to tailor anti-cancer treatments to reduce cardiac dysfunction risk. However, the guidelines established at that time predated several modern techniques for imaging the heart, and many cardiovascular treatments with demonstrated efficacy in reducing heart failure mortality were not available when traditional strategies for managing cardiac risk in cancer patients were established. Currently the morbidity and mortality of cancer treatment-related cardiac dysfunction remains high, and there is now a strong evidence base supporting the concept of increased long-term risk of cardiac complications in the cancer survivor community. Cardio-oncologists today now have a number of cardiac imaging techniques to choose from for the detection of CTRCD, including nuclear multi-plane acquisition scans (MUGAs), echocardiography and cardiac MRI (CMR). Several techniques themselves have recently introduced additional imaging features that allow for more thorough and accurate interrogation of cardiac function. There is currently a limited amount of high quality published scientific evidence evaluating the utility of these imaging techniques in the cancer population, which is reflected by the lack of a consensus among practitioners about which of these imaging modalities may be optimal. Given the significant differences in clinical information provided and risks of each technique, it is possible that one technique could more appropriate than the alternatives. While novel cardiac imaging techniques such as assessment of diastolic function and myocardial mechanics by CMR offer new insights into mechanisms of cardiac dysfunction, their utility in the cancer population is yet untested. \\(Aims:\\) The work in this thesis is divided into four distinct sections. The first section is concerned with recognizing CTRCD secondary to either chemotherapy or radiotherapy. The first section aims to answer the following: 1) What is the frequency and optimal imaging timing of cardiac dysfunction secondary to chemotherapy? 2) What are the most frequent cardiac imaging abnormalities in patients treated with chest directed radiotherapy? 3) What is the quantifiable increased risk of heart failure secondary to radiotherapy? 4) Who should be tested ‚Äö- in particular, what is the impact of traditional heart failure risk factors on echocardiographic measurements of cardiac function? The second section is concerned with identifying the strengths and limitations of the two most frequently utilized cardiac imaging techniques in Australia (MUGA and echocardiography) with the intention of determining whether one technique is likely to be superior to the other. The second section aims to answer the following: 5) how do MUGA and echocardiography compare in frequency of utilization, agreement in measurements and prediction of cardiac dysfunction in the cancer population? 6) Given the importance of accurate and reproducible LVEF measurements in the cardio-oncology population, can automation in echocardiography be utilized to improve patient outcomes? Section 3 investigates the use of novel imaging techniques to provide additional information regarding pathophysiology in addition to traditional cardiac imaging techniques. Cardiac MRI (CMR) is not practical for routine cardiac imaging due to issues of cost and access but may be a reasonable adjunct to either echocardiography or MUGA to further guide treatment by identifying mechanism and phenotypes of cardiac dysfunction. The third section aims to answer the following: 7) Does the observed decline in LVEF as measured by CMR in breast cancer patients reflect preload reduction or reduced myocardial contractility? 8) What does CMR tell us about the phenotypic diversity of cardiac dysfunction secondary to breast cancer therapy? The final section recognizes that cardiac imaging decisions do not occur based solely on clinical characteristics, and issues of financial burden and quality of life must be clarified before society can invest in widespread uptake. The fourth section asks the question: 9) What is the most cost-effective approach for preventing CTRCD? \\(Methods:\\) These questions were addressed using a number of study designs. First, a narrative review of peer reviewed published research describes the current state of cardiac dysfunction secondary to chemotherapy. Second, a systematic review was performed to quantify cardiac structural and functional changes secondary to chest-directed radiotherapy. Third, a meta-analysis was conducted to describe and quantify the increased risk of heart failure secondary to chest-directed-radiotherapy as treatment for cancer. Fourth, the role of traditional heart failure risk factors was tested in a cohort study of adult survivors of childhood cancer, in which the effect sizes of clinical and echocardiographic changes were sought. Fifth, a prospective cohort study of 35 consecutive cancer patients judged by their treating oncologist to be at risk of CTRCD were recruited from a single regional Australian hospital. They underwent simultaneous oncologist-guided serial cardiac imaging with nuclear multi-plane acquisition (MUGA) scans and echocardiography performed by researchers at 3-month intervals over 12 months. Sixth, a systematic review of published literature regarding automation in echocardiography, limited to seven common echocardiographic measurement techniques, was performed and reported. In addition, strength of scientific evidence for each automated technique was assessed, and recommendations for implementation in echo labs is provided. Seventh, the effects of preload reduction (reduction in end-diastolic volume) or reduced myocardial contractility (increase in end-systolic volume) were evaluated in a short-term cohort of HER2+ve early breast (EBC) cancer patients satisfying criteria for CTRCD in serial cardiac MRI examinations and global longitudinal strain measurements by echocardiography. Eighth, a longitudinal cohort design was used to identify different phenotypes of cardiac response to potentially cardiotoxic chemotherapy in HER2+ve EBC patients. Ninth, a Markov model was used to assess the cost-effectiveness of different cardioprotective strategies. \\(Results:\\) First, CTRCD is a relatively frequent complication of contemporary anti-cancer treatments and sequential cardiac imaging is advised for high-risk patients. Second, mediastinal radiotherapy is associated with long-term heart failure due to left ventricular systolic dysfunction, with limited information suggesting significant contribution from right ventricular systolic dysfunction. Despite previous reports in the literature, long-term diastolic left ventricular dysfunction was infrequent. Third, meta-analysis of available quality publications revealed that mediastinal radiotherapy approximately doubled the risk of heart failure (HR 1.83 [95%CI 1.09 to 3.08], p=0.022), but with significant heterogeneity between studies (I\\(^2\\) 88.5%). Metaregression demonstrated that 80% of heterogeneity could be explained by differing age at time of radiotherapy and length of follow-up. Fourth, traditional heart failure risk factors were demonstrated to increase risk of long-term cardiac dysfunction on echocardiographic imaging after median of 22.6 years of follow-up. Reduced 3D-LVEF was significantly associated with hypertension (OR 1.81 [95% 1.25 to 2.62]), reduced global longitudinal strain was associated with insulin resistance (OR 1.72 [95%CI 1.30 to 2.28]) and obesity (OR 1.58 [95%CI 1.19 to 2.11]) and diastolic dysfunction was associated with hypertension (OR 1.41 [95% 1.03 ‚Äö- 1.94]), insulin resistance (OR 1.44 [95%CI 1.08 to 1.92]) and obesity (OR 1.92 [95%CI 1.42 to 2.58]). These findings suggest that childhood cancer survivors with traditional heart failure risk factors may benefit from long-term monitoring of cardiac function. Fifth, it was determined; i) there was poor correlation between LVEF measurements by MUGA and echocardiographic techniques; ii) echocardiography found that 28% of patients had a significant decline in LVEF, whereas only 3% of patients were found to have LVEF decline by MUGA, and iii) relative change in GLS at 2 months was predictive of significant later LVEF decline, with receiver-operating curve demonstrating AUC 0.74 (95%CI 0.55 ‚Äö- 0.94). These results suggest that MUGA and echocardiography have large limits of agreement for LVEF measurements and that echocardiography may detect cardiac dysfunction earlier than MUGA. Sixth, current echocardiographic LVEF measurements suffer from low reproducibility, which reduces confidence in echocardiography-guided CTRCD diagnoses. One potential strategy for improving echocardiographic LVEF reproducibility is by increased utilization of automated measurement techniques. Seventh, a cohort of 83 patients with early HER2 +vet breast cancer was recruited, of which 15 patients (13.4%)developed CTRCD as defined by LVEF drop measured by CMR. Of these 12 cases (80%) were due to isolated increase in LVESV, 2 cases (13.3%) were due to concurrent increase in LVESV and decrement in LVEDV and 1 case (7.7%) was due to non-significant volume changes in both LVEDV and LVESV. No patients developed CTRCD due to isolated LVEDV reduction. These results suggest that preload reduction does not significantly contribute towards observed reduction in LVEF in breast cancer patients diagnosed with CTRCD. Eighth, of a cohort of 66 EBC patients, 13 (19.7%) developed systolic dysfunction, 13 (19.7%) developed diastolic dysfunction with 4 patients overlapping, and 44 (66.7%) had a phenotype characterized by absence of overt systolic and diastolic dys...