Multimodal assessment of acute cardiac toxicity induced by thoracic radiotherapy in cancer patients. Study protocol

Despite its demonstrated effectiveness against cancer RT damages not only malignant cells but also normal cells within our bodies. Consequently, like any other cancer therapy it has side effects that can range from moderate to severe [5]. In particular, thoracic RT has well-documented side-effects upon the cardiovascular system [6], which are commonly known as “cardiotoxicity”. Early studies demonstrated that the exposure to ionizing radiation causes a dose-dependent cardiac damage that may result in a premature death of the patient due to ischemic events, cardiac valve failure/disorders or pericardial injury [7]. In recent years, the prevention, assessment and clinical management of cardiotoxicity in oncological patients has established the basis for a new discipline and an emerging field in oncological medicine called “cardio-oncology” [8]. As described above, radiation to the chest has the potential to cause damage to the heart in a dose-dependent manner. Subsequently, post-RT cardiac dysfunction can progress into heart failure via a complex mechanism that involves damage to the endothelium and microvascular injury [9, 10]. This protocol will assess the dose-dependent cardiotoxic effects of thoracic RT in cancer patients.

Unlike other cancer therapies that can also cause cardiotoxicity (including anthracyclines & chemotherapeutics), RT-derived cardiac damage is usually manifested after years or even decades from initial exposure [ [11]]. Indeed, studies demonstrate that exposure to radiation can damage the intimal layer of blood vessels also triggering atherogenesis, and over the years these can lead to ischemia. Concomitantly, the damage to small vessels can also affect cardiac contractility [12]. Unfortunately, these effects develop over a prolonged period of time and become clinically apparent only after the cardiac performance is compromised and the reduction in the patient’s heart function can be measured, difficulting its diagnosis. At present, the sensitivity and specificity of the tools employed to measure cardiac function are somewhat imperfect and sometimes over-estimate or under-estimate RT-derived cardiotoxicity.

For years the echocardiogram has been the standard tool to assess cardiotoxicity in cancer patients subjected to thoracic RT. By this method, cardiac function is estimated based on volumetric changes of the left ventricle during the cardiac cycle by Simpson’s method, commonly known as Left Ventricular Ejection Fraction (LVEF) a parameter that is affected by pre and post load conditions. Thus, a 10%-drop in LVEF has been traditionally considered as a parameter to define “cardiotoxicity”. However, as previously pointed out this is a rather “late” event in the progression to heart failure, and by the time these changes occur there are structural modifications on the myocardium often irreversible, or difficult to compensate at best. Recently, a variety of novel echocardiographic techniques have been developed; these are largely based on the transient deformation (shortening) of the myocardial fibers assessed with changes of the “speckle” pattern observed in an echocardiogram during a normal cardiac cycle. Although myocardial deformation during the cardiac cycle is a multidimensional process it can be simplified into three basic strains [13] as shown by Fig.1 (Modified from Bansal) for analysis: (a) longitudinal, (b) radial and (c) circumferential. -Longitudinal strain refers to the shortening of the myocardial fibers along its major axis (longitudinal). -Radial strain denotes the thickening of ventricle wall in reference to its radius -Circumferential strain relates to a reduction in the circumference of a heart cavity (ventricle or atrium) during the cardiac cycle. Figure 1 shows a simplified diagram of the components of myocardial deformation that can be assessed with Strain by Speckle-Tracking echocardiogram (hereafter, simply called SST). In summary, the use of the SST allows a more accurate and much earlier assessment of alterations in myocardial contractility compared to the traditionally used 10%-drop in LVEF and thus previous studies have assessed RT-induced acute myocardial damage by SST in multiple settings and has been suggested as an early indicator of subclinical cardiac dysfunction [14‐16]. Furthermore, SST alterations predict mortality among acute heart failure patients [17]. Hence, one of the main goals of this proposal is to evaluate SST patterns as early predictors for the development of cardiotoxicity and acute damage during or after thoracic RT treatment across different dosages.

Studies demonstrate that MRI provides a higher spatial resolution to accurately visualize myocardial lesions compared to other techniques such as scintigraphy [18]. A study in esophageal cancer patients that received high doses of RT (>60Gy) used MRI to demonstrate that RT induces cardiac fibrosis and edema [19] suggesting this is an effective technique to assess RT-induced cardiac damage. Therefore, our proposal will incorporate MRI analysis pre- and post-RT in cancer patients as an imaging assessment along with SST. MRI will include Cardiac Magnetic Resonance Cinema Imaging (Long axis balance, Balance 4 cameras, Short shaft full balance, Right ventricular balance), Anatomical imaging (Inversion recovery single shot balance, 3D short axis covering the entire heart and aorta with free respiratory trigger), Flow Imagining (2D outflow tract of the aorta and 2D pulmonary artery outflow tract) and Quantitative Imaging (T1 map short axis apical section, T1 map short axis medial section, T1 map short axis basal section, T2 map short axis apical section, T2 map short axis medial section, T2 map short axis basal section).

In addition to alteration in the SST patterns or MRI, a number of soluble factors have been postulated as markers for cardiac dysfunction and will be assessed as early predictors of cardiotoxicity:

Circulating Endothelial Cells (CECs) are a subpopulation of fully-differentiated mature endothelial cells that have detached from the vascular walls. This may occur under certain circumstances such as tumor angiogenesis, derived from smoking, by acute myocardial infarction [24] or hypertension [25]. Although CECs were described over 30 years ago, they were only recently recognized as a reliable marker for endothelial dysfunction [26], and studies have postulated their role on the pathogenesis of ischemic vascular disease [27]. In particular, apoptotic CECs are predictors of cardiac vasculopathy in heart transplanted patients [28] and are well established as indicators of vascular injury and endothelial damage. On the other hand, RT is known to produce endothelial vascular injury through a variety of mechanisms [29]. Hence, our proposal will assess changes in apoptotic CECs prior and after RT in cancer patients. We hypothesize that high-RT doses will increase apoptotic CECs. Our proposal will compare imaging by SST and MRI against soluble/biochemical markers as early predictors of cardiac dysfunction in RT-treated cancer patients.

As explained above cardiotoxic damage by RT is dose dependent. Evidently, the success of a RT treatment in eradicating a tumor is determined by the total radiation dose given measured in Gray (Gy), however, this is limited by the tolerance of the surrounding normal tissue [5]. Similarly, previous studies have demonstrated that cardiotoxicity effects are proportional to the radiation dose applied to the patient. In order to quantify the dose-dependent effects of thoracic RT we will stratify our patients into two categories according to the calculated mean heart dose received. Therefore, our proposal will stratify patients receiving low (<10Gy) and high (>10Gy) doses of RT (see Fig. 2 and the methodology section for further details).

In recent decades, novel more effective therapies against cancer have significantly prolonged patient survival. Consequently, the number of cancer survivors with “late” anti-cancer treatment derived complications is expected to rise in coming years. Among these complications, cardiotoxicity represents a significant risk for premature death [8]. It is estimated that approximately 50% of all oncological patients receive some type of RT during their treatment. As occurs with other therapies, thoracic RT is associated to a risk of developing cardiotoxicity in a percentage of patients. Unfortunately, current standard techniques only allow the detection of cardiac dysfunction at an advanced stage in the progression to heart failure (a 10% drop in LVEF). Here we propose the use of a combination of echocardiogram SST, MRI and serum factors associated cardiac/endothelial dysfunction as early predictors of late cardiotoxicity derived from RT treatment.

Estimates in the United States indicate that approximately 4 million cancer survivors die from cardiovascular disease derived complications. An earlier identification of patients at a higher risk for late thoracic RT-derived cardiotoxicity may help identify the best candidates for a risk adapted clinical follow-up in order to prevent long-term cardiac complications aiming to minimize the risk of death. Likewise, this will pave the way to testing new interventions to reduce cardiac damage, as it has been the case for anthracycline-induce cardiotoxicity. The SST echocardiogram is a relatively simple, low-cost, non-invasive assay that can be routinely performed by cardiologists in any clinical setting.

To evaluate and compare the changes in different modalities: Strain by Speckle Tracking (SST), MRI, soluble markers of cardiac dysfunction and apoptotic CECs as early predictors of cardiotoxicity in cancer patients receiving low or high mean heart doses of RT.

GOAL 1-To prospectively enroll cancer patients scheduled to receive high or low doses of thoracic RT with curative intent.

Patients: Inclusion criteria:

Patients: Exclusion criteria:

The cardiovascular risk score ACC/AHA 2013 for every patient will be calculated prior to RT (previously validated by Acevedo et al. [30] RT planning for every patient will be elaborated following international guidelines. Heart volume will be contoured according to the atlas provided by Duane et al. [31]. Whenever, possible the most advanced equipment will be used to deliver the planned dose to the targeted tumor, minimizing the dose to the heart or other organs that may be at risk. Strategies employed may include: Intensity modulated RT (IMRT), volumetric modulated arc therapy (VMAT), or image guided RT (IGRT). For dosimetry planning and restrictions for the heart volume we will follow the criteria provided by Marks et al. [32].

Accordingly, patients will be stratified based on the mean heart dose received into two groups: Low mean heart dose (<10Gy) and high mean heart dose (≥10 Gy) (see Fig. 2). Patients will be treated in an Electa Axesse or Varian Clinac 2100, and planning will be done using the Monaco 5.11 or Eclipse with Montecarlo and Pencil beam algorithm. For adjuvant breast cancers standard Tangent technique will be used with Field-in-field to minimize hot spots in the target. Usual cardiac blocks will be used to reduce heart dose. 40–60 Gy in 15–30 fractions will be delivered. An additional Oblique supraclavicular field will be added for regional lymph node RT if needed. 6–18 Mv X-rays will be used. For neoadjuvant esophagus cancer VMAT technique will be used. 41.4 Gy in 23 fractions will be delivered to the target. 6 Mv X-rays will be used. A Cone beam CT (CBCT) prior to each treatment will used to verify setup. For definitive treatment, in esophagus or lung cancer 60 Gy in 30 fractions will be delivered with same technique and CBCT protocol. The medical oncologist will prescribe concomitant chemotherapy if required.

We anticipate a total n of 60 patients for this study. Patients will be categorized according to the diagram shown by Fig. 2 and distributed equally into groups: Low dose RT (n = 30), and high dose RT (n = 30) Sample size was calculated assuming a 30% change by SST in the group of patients receiving high-dose RT, an alpha = 5% and a statistical power = 80% (beta = 0.2). This gives a sample size of 28 patients. Baseline status (pre-RT) will be compared to 1 week following RT (post-RT). An additional 10% of patients will be recruited (n = 30) in order to avoid data loss during follow up.

GOAL 2-To assess imaging and cardiac parameters using SST and MRI in cancer patients prior and after thoracic RT.

The SST echocardiogram for consented patients is performed following the standard procedure in a medical attention box at the Cancer Center. The procedure lasts approximately 20 min, and is performed by a trained cardiologist using a portable device. Parameters measured include heart contractility, ventricular/atrial ejected fractions. Obtained data are used to calculate longitudinal, radial and circumferential strains as described above (see Fig. 1). Combined data are also used to calculate global longitudinal strain (GLS) using the average peak strain from 3 apical projections of the left ventricle. SST measurements will be recorded for every patient prior to RT, and also 1 and 12 weeks after completion of RT treatment as described in Fig. 2. Importantly, concomitant use of chemotherapy in patients will also be recorded. For analyses, the % of change in every patient will be assessed comparing levels before and after RT performing a paired analysis.

High-resolution, non-contrast-enhanced cardiac nuclear magnetic resonance (CMR) scans will be performed at the Radiology and Diagnostic Imaging Unit of Red Salud UC-Christus (Santiago, Chile) on a Philips 3.0 T Ingenia Gyroscan MRI scanner equipped with Release 5.3 Clinical Packages (Philips Healthcare, Best, The Netherlands). Following safety questionnaire and signing of informed consent patients will be in supine position for research studies. Trained radiologists from the Radiology Department at the School of Medicine in the Pontificia Universidad Catolica de Chile will blindly review and inform the obtained MRI results. CMR scans will be performed at baseline (pre-RT), and after the administration of last fraction (post-RT) according to radiation therapy treatment plan.

The same imaging protocol will be used for pre and post RT measurements. The body coil will be used for signal transmission, and a 16-element anterior and 12- element posterior phased-array coil will be used for signal reception. Image acquisition will be performed as described [33], with: i) electrocardiography-gated, motion corrected, non-contrast-enhanced T1 mapping using the modified look-locker sequence (MOLLI) parameterized as follows: slice thickness 8.0 mm, field of view 300 × 400 mm, flip angle 50 degrees, minimum TI 120 ms, inversion-time increment 80 ms and sensitivity encoding (SENSE) factor 3. Freebreathing or breath-hold will be used according to patient’s characteristics. Additionally, ii) electrocardiography-gated, breath-hold-gated, steady-state free precession long-axis cine images in 2, 3, and 4-chamber views will be acquired parameterized as follows: slice thickness 8.0 mm, no gap, field of view 280 × 300 mm, 208 × 256 matrix, repetition time 2.9 ms, echo time 1.2 ms, flip angle 70 degrees, temporal resolution 0.05 indicating a normal distribution).

A logistic multivariate regression analysis will determine SST at 1-week post-RT, other independent variables will be included: age, disease details, soluble markers status comparing LD versus HD. Analyses will also assess interaction or confusion among outcome associated variables on bivariate analyses. OR for these tests will be reported using a confidence interval of 95% Diagnostic yield of MRI versus standard SST. A diagnostic yield analysis comparing MRI against SST will report area under the curve (AUC), sensitivity, specificity, positive and negative predictors, and likelihood ratios (all at 95% confidence intervals). Finally, all analyses and calculations will use STATA v.12 (College Station, TX, StataCorp LP).

We expect increases in serum biomarkers of cardiac dysfunction after RT, especially in those patients receiving higher doses of radiation. Regarding statistical analyses we expect significant changes by comparing pre-RT versus post-RT levels in each case.

Our 3-year work plan for this proposal is shown below and describes the main tasks for every specific goal of the project and the estimated time frame for publication of manuscripts and presentation at scientific meetings. The goals or specific tasks indicated are organized quarterly by every year of the proposal. (see Fig. 3). This protocol will follow the SPIRIT guidelines (See SPIRIT checklist in supplementary material).