Anthracyclines are chemotherapy agents widely used to treat different types of cancer and are well recognised to induce cardiotoxicity¹. As cancer survival rates have improved the longer term consequences of chemotherapy-induced cardiotoxicity is increasingly recognised as a public health issue. Chemotherapy-induced cardiotoxicity can occur acutely following infusion of the agent or in the months and years after treatment. Significant cardiac dysfunction has been defined in cancer trials as a decrease in left ventricular ejection fraction (LVEF) of at least 10% by which time there will already be significant and possibly irreversible myocardial damage². There is therefore an unmet clinical need to identify techniques capable of providing an earlier warning of cardiotoxicity prior to irreversible cardiac damage. This would allow oncologists to alter patient treatment to reduce this risk.

Phosphorous MR Spectroscopy (³¹P-MRS) detects “high energy” phosphate metabolites in-vivo and allows quantification of the energy state of the heart, usually through the phosphocreatine (PCr) to adenosine triphosphate (ATP) concentration ratio. ³¹P-MRS has previously been used to study heart conditions³. Here we present preliminary results from an on-going study investigating whether ³¹P-MRS can detect changes in cardiac energetics in patients undergoing chemotherapy treatment for breast cancer.

9 patients recently diagnosed with breast cancer and due to receive 6 cycles of chemotherapy (FEC-T, EC-T or FEC-80) underwent MRI scanning of their heart prior to chemotherapy. Due to illness of 2 patients, only 7 were scanned during chemotherapy (between cycles 3 and 4). So far 5 of these patients have been scanned following completion of their treatment (~2-3 weeks following cycle 6). Patients were positioned supine and head first in the 3T Siemens Verio (Siemens Healthcare, Erlangen) between anterior and posterior parts of an 8-element cardiac ³¹P receive array coil (Rapid Biomedical, Germany). ³¹P MR spectra were acquired using a protocol⁴ that included a 3D UTE-CSI pulse sequence (with TR/TE = 1000/~0.6ms, FOV=(350mm)³, a 22x22x10 CSI matrix, acquisition weighting with 2 averages at k=0, and WSVD coil combination). This was applied without ECG gating and had a total acquisition time of 28 minutes. Patient specific B1-maps were generated, allowing calculation of the excitation voltage required to deliver a 30° pulse to the mid-septum of the myocardium during the spectral acquisition⁴. The spectrum from the mid-septum voxel was fitted using a custom Matlab implementation of AMARES which estimates the PCr/ATP ratio, corrected for saturation effects and blood contamination⁵. A stack of short-axis cardiac images were also acquired at each time point using a TrueFISP sequence (TR/TE = 85.8/1.45ms, flip angle = 50°, FOV=400x338, matrix = 256x205, GRAPPA = 3, slices = 2, slice thickness = 8mm, gap = 2mm). Imaging signal was received using the scanner’s body coil. Left ventricular ejection fractions (LVEFs) were calculated for each time point (QMass ,Medis). An experienced operator measured LVEF twice at each time point to provide a mean LVEF and standard deviation.An example ³¹P MR spectrum is shown in Figure 1. PCr/ATP ratios, LVEFs and troponin levels at pre-, mid- and post- chemotherapy are shown in Figure 2 and are plotted in Figures 3-5 respectively. The troponin levels of all patients increased between pre- and mid-chemotherapy. There was a further increase between mid- and post- chemotherapy in the 5 patients who have completed treatment. Although troponin levels are used in clinical management of cardiac infarction, their relevance in chemotherapy patients is not yet known. Therefore this needs to be further evaluated. It is not expected that all patients undergoing chemotherapy will experience cardiotoxicity. Only patient 3 experienced a greater than 10% decrease in LVEF between pre- and mid-chemotherapy. This patient reported extreme tiredness and nausea resulting in delayed treatment. Subsequent chemotherapy cycles were delivered at reduced dose and no further change in LVEF was detected between mid- and post-chemotherapy. Using an equivalent ³¹P MRS protocol to that applied here, Tyler at al⁶ reported an intra-subject variability in PCr/ATP of 20% in healthy male volunteers scanned twice (PCr/ATP = 2.07±0.38 and 2.14±0.46). Here the mean pre-chemotherapy PCr/ATP was 2.17±0.59 and mean patient age was 52.4±11.3 years. Patients 2, 3, 4 and 6 experienced a greater than 20% decrease in PCr/ATP (24.7%, 57.1%, 27.3% and 20.1% respectively) between pre- and mid-chemotherapy. Between mid- and post-chemotherapy only patient 2 experienced a further PCr/ATP decrease (of 17.5%). PCr/ATP ratios of age-matched healthy female subjects will be determined as part of this on-going study and repeatability will be further assessed.These preliminary results illustrate the potential of ³¹P MRS to detect changes in heart function of patients undergoing chemotherapy. As this study continues it will determine whether changes in high energy phosphates of the heart precede changes in LVEF in this population.