Cardiac Imaging Methods for Chemotherapy-related Cardiotoxicity Screening and Related Radiation Exposure: Current Practice and Trends

Discussion

MUGA scan usage is declining and it is mostly replaced by Echocardiography. All three trends for declining MUGA usage and increasing Echo and CMR usage were seen consistently across all disease subtypes, types of chemotherapy, sex and age groups (Table IV). Although we cannot identify all the reasons for this decline in usage of MUGA scans, increased awareness of radiation risks is likely a notable contributor. The radiation exposure from a single MUGA scan is about 7.8 mSv (6). This is equivalent to 2.5 times the yearly background radiation in United States (3 mSv). For example, in the case of adjuvant trastuzumab use in a woman with breast cancer, CRC screening is recommended prior to initiation of therapy, every 3 months during therapy and every 6 months for two years after completion of therapy per the manufacturer prescribing information (requiring total of 9 scans) (9). The Canadian Trastuzumab Working Group recommends screening during therapy for a minimum of 5 scans, while stating no evidence exists for continued screening following completion of therapy (10). However, ongoing screening may be necessary in patients who received anthracyclines due to their late cardiac effects (1). If we assume a woman with breast cancer who undergoes trastuzumab-based adjuvant chemotherapy and receives 9 MUGA scans for CRC screening and one positron emission tomography-computed tomography (PET-CT) for staging, the radiation exposure is approximately 100 mSv (70 mSv for MUGA scans and 25-30 mSv for PET-CT) (6). Using BEIR VII estimates, exposure to 100 mSv of radiation in a 40-year-old has additional lifetime risk of 1,300 cases of cancer per 100,000 women (11). Typically, this patient (depending on the stage of the disease) may also receive chest wall radiation and, possibly, more CRC screening for the anthracycline part of the adjuvant regimen plus additional surveillance imaging during the disease course, thereby greatly increasing the radiation exposure and secondary cancer risk. Using other imaging methods, such as echo-based techniques instead of MUGA, could reduce the total radiation burden significantly.

The number and frequency of cardiac imaging studies and duration of follow-up required for CRC monitoring depends on the cardiotoxicity risk. The incidence of cardiotoxicity can vary depending on patients' characteristics, such as age, duration of survival after the inciting agent, presence of other co-morbidities like hypertension, cardiovascular disease, pre-existing left ventricular dysfunction and on treatment characteristics, such as dose of the drug, associated mediastinal radiation or usage of other cardiotoxic drugs. The majority of patients in our study underwent only 1 or 2 cardiac imaging studies during the 4 years (Table III). However, in patients who require frequent monitoring and a long monitoring period, such as in breast cancer patients who receive trastuzumab, the number of imaging studies was higher. Forty-six percent of patients who received trastuzumab and 32% of breast cancer patients had 5 or more studies over 4 years compared to only 18% of all patients in the study group (Table III). Also 55% of patients who received more than one cardiotoxic drug underwent 5 or more studies during the 4 years. As can be seen in Table IV, the breast cancer and trastuzumab subgroups saw the largest decline in MUGA scan usage over the four years and have the lowest relative usage of MUGA in 2014, likely reflecting the clinically appropriate drive to lower their total radiation burden.

Other factors contributing to these observed trends in imaging modality choice could be the more comprehensive anatomic and functional detail available using Echo and CMR, as well as their capability for early detection of cardiotoxicity. A decline in left ventricular ejection fraction may be a late manifestation of cardiotoxicity (12) and there has been an increasing focus on early detection of CRC using techniques, such as strain monitoring using Echo (13) and T2-weighted imaging for myocardial edema using CMR (14). There have been algorithms proposed to incorporate these techniques into the CRC screening protocols (15). It is essential to point out here that 2D Echo for serial LVEF monitoring has a high test-retest variability that is reported to be anywhere from 11% to as high as 17% (16). This raises the odds that a drop in LVEF of 10% detected using 2D Echo could be due to a sampling error (17). This can be improved by using a higher LVEF cut-off threshold by 2D Echo of 60% instead of 55% (18), by using contrast for better delineation of endocardial borders (interobserver variability of 9.6%) (19) or by using 3D Echo (interobserver variability of 4.9%)(20).

This study suffers certain limitations. The study population is from a single health care delivery system in the Minneapolis-St. Paul Metropolitan area and vicinities in the United States. Therefore, this practice may not be representative of the general population. While there could be other indications for cardiac imaging in cancer patients, we assumed that the majority of cardiac imaging studies in this population were performed for CRC screening. Despite these limitations, we report the current practice trends over 4 years in a system where there was broad availability and consistent access to all the three imaging modalities.