1. IntroductionAlong with systemic therapies, radiation therapy (RT) plays an integral role in treating advanced lung cancer. However, the incidence of RT-induced cardiac complications in this patient population is as high as 33% [1,2,3]. Recent studies suggest that radiation-induced cardiac damage can manifest within two years of completing RT for lung cancer [1,4,5,6,7,8,9] and that mortality correlates with mean heart dose [4] or with the percentage of the heart receiving 5 Gy [10], 30 Gy and/or 50 Gy [8]. A national multicenter lung cancer trial and other single-institution studies have found correlations between early death and radiation dose to the heart [11]. Interestingly, many trials have not been able to correlate early deaths with clinically evident cardiotoxicity, suggesting that early non-cancer deaths may be partially due to subclinical cardiac dysfunction.The current paradigm for cardiac function assessment and management of cardiovascular disease relies primarily upon the assessment of global cardiac function, e.g., left ventricle ejection fraction (LVEF) [12,13]. However, LVEF alone is limited in its diagnostic and prognostic ability, as ventricular remodeling can compensate for regional cardiac dysfunction to maintain global function and cardiac output in acute and subacute phases post-RT before LVEF ultimately declines [14,15,16]. Thus, LVEF may be stable or even increases after radiation exposure due to compensatory mechanisms [16]. Therefore, identifying new markers capable of early detection of subclinical cardiac dysfunction after cardiac RT exposure is essential for risk stratification and prompt initiation of cardioprotective heart therapy to avoid non-reversible late cardiac damage [17].Magnetic resonance imaging (MRI) tagging [18] allows for quantitative assessment of myocardial contractility on a regional basis. As regional changes in myocardial contractility frequently occur before cardiac damage [14,15], regional cardiac dysfunction is potentially reversible, and myocardial strain-derived parameters can be early indicators of cardiotoxicity. This has been demonstrated in studies where MRI-generated strain parameters detected abnormal myocardial function in cancer patients treated with chemotherapy, despite normal systolic function [19,20,21,22,23,24,25]. Recent results from our preclinical work identified heterogeneous contractility patterns in different LV regions with reduced myocardial strain post-RT [15,16].

In this study, we investigated the capability of advanced MRI-generated parameters of heart mechanics for characterizing the myocardial contractility patterns post-RT and for early detection of RT-induced cardiotoxicity in a pre-clinical rat model of radiation-induced cardiotoxicity. We present the contractility index (ContractiX) derived from regional strain parameters. We hypothesize that ContractiX is a sensitive marker for early detection of RT-induced cardiotoxicity. We also examined the effect of RT on other MRI-generated heart contractility parameters, including ventricular torsion, diastolic strain rate, and mechanical dyssynchrony.

4. Discussion

In this study, we investigated the effect of localized heart irradiation on myocardial contractility patterns by studying different MRI-generated parameters that measure different aspects of systolic and diastolic functions, as well as mechanical dyssynchrony. The results demonstrated the sensitivity of the newly developed parameter, ContractiX, for early detection of subclinical cardiac dysfunction post-RT despite normal LVEF.

The results from this study illustrate the multifaceted effect of RT on myocardial contractility. Many reports, both clinically and pre-clinically, measure RT-induced cardiotoxicity utilizing LVEF as a measure of cardiac function [15,16,26,31,32,33,34]. Despite its value, LVEF mainly reflects global cardiac function, typically affected at a later time point post-RT. However, parameters of regional cardiac function can be measured for early detection of RT-induced cardiotoxicity. From this perspective, regional myocardial strain can be used to calculate the percentage of normally contracting myocardium (ContractiX), which serves as an early detector of the RT effect on heart function. The increase in LVEF post-RT could be attributed to undergoing ventricular remodeling and hypertrophy to maintain cardiac output in the face of RT injury.Our results demonstrate the value of ContractiX as an early marker of subclinical cardiac dysfunction post-RT despite normal global function (Figure 6). While normal LVEF was maintained post-RT, ContractiX showed a continuous decrease in values from the sham-treated rats to 8 weeks post-RT rats and then to 10 weeks post-RT rats. ContractiX parameters calculated from circumferential strain, both circumferential and longitudinal strains, or all three strain components demonstrated significant reductions between 8 weeks and 10 weeks post-RT (Figure 4), which presents ContractiX as a sensitive marker of cardiac dysfunction progression post-RT, especially since all ContractiX parameters demonstrated greater and more significant reductions post-RT compared to strain reductions (Figure 5). When ContractiX was examined versus LVEF on a scatter plot, the results identified a nonlinear inverse relationship between the two parameters (Figure 6), where ContractiX spanned a large range of values between 25% and 100% among different rat groups despite normal LVEF (≥59%) in all rats. It is possible that ContractiX may hold promise for risk stratification of cancer patients receiving higher incidental heart radiation exposure as part of their cancer treatment, for example, in lung and esophageal cancer patients. Ultimately, this could lead to prompt initiation of cardioprotective therapy before global heart function is affected to avoid clinical cardiac dysfunction and subsequent heart failure.

There are three unique features of the introduced ContractiX parameter: (1) it can be calculated based on different strain components or combinations of these strain values; (2) it is agnostic to the strain acquisition sequence and analysis technique; and (3) the threshold values are determined utilizing the mean and standard deviations of strain measurements from normal subjects in the population being investigated for accurate differentiation between normal and pathological cases. The last feature is important as normal strain ranges could be different in different studied groups, species or even based on the strain acquisition and analysis techniques.

Our results also demonstrate the effect of RT on other parameters of heart mechanics. For example, ventricular torsion showed a small, non-significant decrease post-RT. While this parameter did not distinguish early changes post-RT well in this study, it may be useful to compare individuals in other situations, e.g., with different doses or time points post-RT. The study showed that RT affects not only systolic function but also diastolic function [25,33]. This was represented by the reduced early-diastolic strain rate, which was significant for all strain components at 8 weeks and 10 weeks post-RT. Diastolic cardiac dysfunction could be due to tissue stiffening secondary to changes in myocardial tissue composition post-RT, e.g., the presence of interstitial and perivascular fibrosis, myocardial vacuolization, and/or necrosis, as previously illustrated in preclinical models [15,26] and shown to be an important mechanism of radiation-induced heart dysfunction [35]. Our study also demonstrates that RT affects not only contractility magnitude but also contractility timing. The results show that while different myocardial segments contract at almost the same time point (time-to-peak strain) in the healthy heart, this mechanical synchrony is affected by radiation such that the variation of time-to-peak strain between different myocardial segments increases at 8 weeks and 10 weeks post-RT (Figure 7c). The induced mechanical dyssynchrony constitutes an additional factor contributing to the weaker contractility performance post-RT. It should be noted that while ventricular torsion, diastolic strain rate, and mechanical dyssynchrony showed altered measurements post-radiation, they were not as sensitive as the proposed ContratiX parameter. Specifically, all ContractiX parameters (Figure 4) showed statistically significant reduction not only between sham and RT rats but also between 8 weeks post-RT and 10 weeks post-RT, which was not the case for other parameters shown in Figure 7. For example, ventricular torsion (Figure 7a) showed very close values in different rat groups without any significant differences between the groups. Furthermore, while diastolic strain rate (Figure 7b) and mechanical dyssynchrony (Figure 7c) showed significant differences between the sham and RT rats, they showed limited capability for differentiating between the 8 week post-RT and 10 week post-RT. Therefore, the additive value of the proposed ContractiX parameter is its higher sensitivity for detecting slight changes in heart contractility during different time points in the study, which may not be feasible using other cardiac measures and allows for a prompt intervention to avoid advancement to heart failure.

Our study has some limitations. The first limitation is the small number of animals used in the study. The main purpose of this study was to introduce the new measure, ContractiX, and provide a proof-of-concept about its sensitivity for detecting subclinical cardiac dysfunction, which was demonstrated by the results despite the small number of animals. However, future studies on a larger cohort are warranted to confirm these results with sufficient statistical power.

Another limitation of our study is that we did not include another group of salt-resistant rats. The choice of salt-sensitive rats for this study was based on their cardiovascular profile, which is more vulnerable to cardiac injuries [16,26,36] (radiation in this study) compared to the more immune salt-resistant rats, which makes it appropriate for the investigation in this study about comparing different MRI measures of cardiac function. However, studying the effect of rat strain (comparing salt-sensitive to salt-resistant rats) on the performance of the developed ContractiX parameter is warranted in future research.

Another limitation of the current study is the unavailability of perfusion or late gadolinium enhancement (LGE) MRI data. However, the focus of this study was to characterize the cardiac function and myocardial contractility patterns shortly after RT and before global heart function is affected, where ischemia development and myocardial infarction are not expected to occur during the acute phase post-RT. Nevertheless, larger studies with longer follow-up times are warranted to confirm the results from this study and evaluate late RT effects on the heart. Although the focus of this study was on situations where the heart receives large radiation doses based on the tumor’s proximity to the heart (such as in lung and esophageal cancers), the presented imaging biomarkers may be valuable in other cancer types that incur incidental heart irradiation, e.g., in breast cancer or lymphoma, and they should be evaluated in these types of exposures in the future.

There is great potential for the translational value of this study, especially in cases of significant radiation to the heart when the tumor’s location is close to the heart [8,37]. The implementation of the developed technique in clinical studies would allow for earlier detection of subclinical cardiac dysfunction that may not be detected using conventional cardiac MRI measures. Such capability would be valuable for risk stratification, prognosis, monitoring, and management of cancer patients. Based on the demonstrated sensitivity of ContractiX, it can be used for detailed assessment of baseline cardiac function even before treatment starts. Based on the information obtained about ‘weak’ heart regions that show borderline contractility, treatment planning could be optimized to minimize the radiation dose delivered to these regions. Follow-up scans would also be valuable for identifying patients at risk in whom cardioprotective therapy could be promptly initiated to avoid progression into heart failure. Another advantage of the proposed technique is that it does not require using a gadolinium contrast agent, which is important from a safety perspective, especially in patients with compromised kidney function where it can cause nephrogenic systemic fibrosis (NSF) and because of gadolinium deposition in other organs, including the brain [38]. Furthermore, as the imaging biomarkers presented in this study provide information about regional heart function, they could be valuable in future studies to investigate the relative radiation sensitivity of different cardiac substructures or myocardial segments, which may ultimately lead to improved RT planning and better outcomes. Moreover, a variety of cancer-related pharmacologic treatments result in cardiomyopathy, which is frequently preceded by alterations in myocardial strain. Thus, the use of ContractiX could be tested after other cancer treatments for its utility in identifying early subclinical cardiac dysfunction and treatment management.Finally, it should be noted that cardiac MRI provides a plethora of techniques for the comprehensive evaluation of heart health, including assessment of systolic function, diastolic function, hemodynamics, and tissue characterization [39]. Although different techniques reflect different aspects of the heart condition, ContractiX allows for early detection of subclinical cardiac dysfunction during ventricular remodeling when other parameters, e.g., ejection fraction, are still within the normal range. While a comprehensive evaluation of the heart function is always desirable, it may not be feasible in lung cancer patients who are at the advanced cancer stage and have limited capability of enduring a long comprehensive cardiac MRI exam that requires several breath-holds. In this case, an optimized short cardiac MRI exam that includes a few sequences for evaluating heart contractility and measuring ContractiX would be desirable and sufficient for risk stratification and prognosis in this patient population.