Cardiac MRI in the Assessment of Cardiac Injury and Toxicity From Cancer Chemotherapy

Abstract

HomeCirculation: Cardiovascular ImagingVol. 6, No. 6Cardiac MRI in the Assessment of Cardiac Injury and Toxicity From Cancer Chemotherapy Free AccessResearch ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissionsDownload Articles + Supplements ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toSupplemental MaterialFree AccessResearch ArticlePDF/EPUBCardiac MRI in the Assessment of Cardiac Injury and Toxicity From Cancer ChemotherapyA Systematic Review Paaladinesh Thavendiranathan, MD, MSc, Bernd J. Wintersperger, MD, Scott D. Flamm, MD, MBA and Thomas H. Marwick, MBBS, PhD, MPH Paaladinesh ThavendiranathanPaaladinesh Thavendiranathan From Department of Medicine, Division of Cardiology (P.T.) and Department of Medical Imaging (P.T., B.J.W.), Peter Munk Cardiac Center, University Health Network, University of Toronto, Toronto, Canada; Cardiovascular Imaging, Heart and Vascular Institute, Cleveland Clinic, OH (S.D.F.); and Menzies Research Institute, University of Tasmania, Tasmania, Australia (T.H.M.). Search for more papers by this author , Bernd J. WinterspergerBernd J. Wintersperger From Department of Medicine, Division of Cardiology (P.T.) and Department of Medical Imaging (P.T., B.J.W.), Peter Munk Cardiac Center, University Health Network, University of Toronto, Toronto, Canada; Cardiovascular Imaging, Heart and Vascular Institute, Cleveland Clinic, OH (S.D.F.); and Menzies Research Institute, University of Tasmania, Tasmania, Australia (T.H.M.). Search for more papers by this author , Scott D. FlammScott D. Flamm From Department of Medicine, Division of Cardiology (P.T.) and Department of Medical Imaging (P.T., B.J.W.), Peter Munk Cardiac Center, University Health Network, University of Toronto, Toronto, Canada; Cardiovascular Imaging, Heart and Vascular Institute, Cleveland Clinic, OH (S.D.F.); and Menzies Research Institute, University of Tasmania, Tasmania, Australia (T.H.M.). Search for more papers by this author and Thomas H. MarwickThomas H. Marwick From Department of Medicine, Division of Cardiology (P.T.) and Department of Medical Imaging (P.T., B.J.W.), Peter Munk Cardiac Center, University Health Network, University of Toronto, Toronto, Canada; Cardiovascular Imaging, Heart and Vascular Institute, Cleveland Clinic, OH (S.D.F.); and Menzies Research Institute, University of Tasmania, Tasmania, Australia (T.H.M.). Search for more papers by this author Originally published1 Nov 2013https://doi.org/10.1161/CIRCIMAGING.113.000899Circulation: Cardiovascular Imaging. 2013;6:1080–1091During the past 30 years, there has been a significant decrease in cancer mortality rates, predominantly attributable to improvements in treatment options.1 However, survivors are at increased risk of premature cardiac disease,2 both because of the overlap in risk factors for cancer and cardiovascular disease3 and the cardiotoxic effects of cancer chemotherapy. Two chemotherapeutic agent classes that are commonly associated with cardiotoxicity are the anthracyclines and tyrosine kinase inhibitors, both of which can cause left ventricular (LV) dysfunction and heart failure (HF).4,5Mechanisms of cardiac injury from cancer therapy have been summarized elsewhere.4,6 Briefly, anthracycline cardiotoxicity has been attributed to reactive oxygen species formation, transcriptional changes in intracellular adenosine triphosphate production in cardiac myocytes, and, more recently, through interaction with cardiac topoisomerase IIβ.4,6 Trastuzumab cardiotoxicity seems to be because of inhibition of cardiomyocyte human epidermal growth factor receptor 2, resulting in ATP depletion and contractile dysfunction.4 Other proposed mechanisms include immune-mediated destruction of cardiomyocytes.4,6 At the tissue level, early anthracycline toxicity has been associated with myocardial inflammation,7–9 vacuolization,9–12 and cell swelling/edema.11,13 These changes seem to occur before myocardial functional abnormalities.11,13 Later stages of toxicity are associated with myocardial fibrosis.14,15 Unfortunately, the use of myocardial biopsy is not feasible for diagnostic purposes in this setting. However, once HF manifests, the 2-year mortality can be as high as 60%.16 This emphasizes the importance of early recognition of cardiac injury and institution of cardioprotective therapy in an effort to prevent development of HF and allow uninterrupted completion of cancer therapy.17 Thus, the diagnosis is dependent on either direct evidence of myocardial damage or functional disturbance. Either signal may be identified by imaging or biomarker measurements.A commonly used definition of cardiotoxicity is a >5% reduction in LV ejection fraction (EF) to <55% with HF symptoms, or a >10% asymptomatic reduction to <55%.18 Cardiovascular imaging, therefore, has an important role in the identification of cardiotoxicity and, more importantly, in the detection of early cardiac injury using newer techniques.19 Although no established definition of early cardiac injury exists, in this article it is defined as the presence of myocardial changes without significant reduction in LVEF. However, despite multiple candidate imaging modalities, the best modality for these purposes is unknown.Multiple gated acquisition scans are commonly used to monitor cardiotoxicity using LVEF. Despite being a reproducible technique, its lack of sensitivity to small changes in LVEF limits its use in the detection of early cardiac injury.20 Echocardiography is also commonly used to monitor cardiotoxicity, although it too is relatively insensitive to small changes in LVEF.21 Its most promising application, however, is in the use of myocardial strain and strain rate imaging to detect early cardiac injury.22 Nonetheless, the challenges of acoustic window with echocardiography will remain.Biomarkers may also play a role in the detection of cardiotoxicity. In particular, troponin I release with anthracycline or trastuzumab therapy has been shown to be a marker of subsequent LV dysfunction.23,24 However, the need for multiple repeated measurements23 and the poor recovery in LV function in patients with proven troponin release limit its use in routine practice.24Cardiac MRI (CMR) is considered the reference standard for measurement of ventricular volumes and function,25,26 making it ideally suited to assess cardiotoxicity.6 However, more uniquely the ability for myocardial tissue characterization provides a noninvasive biopsy of the underlying histopathological myocardial changes, potentially allowing for recognition of early cardiac injury.13,27 Additional strengths include lack of ionizing radiation and absence of acoustic window limitations. In this systematic review of published works, recent abstracts, and ongoing clinical trials, we summarize the existing data on the use of CMR for the following clinically important scenarios: (1) detection of early cardiac injury, (2) identification of cardiotoxicity during or <1 year of treatment, (3) detection of late consequence of therapy (>1 year post-treatment), and (4) monitoring response to cardioprotective therapy.Literature ReviewThe search method adhered to the preferred reporting items for systematic reviews and meta-analysis (PRISMA) statement for reporting systematic reviews.28 A MEDLINE (1970 to June 2013) search was performed using the OVID search engine using the key words neoplasms, antineoplastic agents, cardiac toxicity, CMR, and their variations. All citations were screened for inclusion by using a hierarchical approach of assessing the title, abstract, and article (Figure 1). References of all selected articles and reviews were screened to identify additional studies. Any study of ≥10 patients that used CMR as the primary method to detect cardiotoxicity was included. No language or humans-only restrictions were imposed. Conference proceedings from Society of Cardiovascular Magnetic Resonance, European Society of CMR, American College of Cardiology, American Heart Association and, European Society of Cardiology Meetings were screened for relevant abstracts between 2009 and 2013. Any relevant studies from these meetings are referred to as preliminary studies in this review. The World Health Organization International Clinical Trials registry (clinicaltrials.org) was searched with above mentioned key words to identify relevant ongoing trials. Preliminary studies and ongoing trials are summarized in the Appendix Tables I and II in the online-only Data Supplement and the references for the preliminary studies are only provided in the Supplement. All relevant data were extracted using a standard data form, and primary authors were contacted when necessary to clarify study components.Download figureDownload PowerPointFigure 1. Literature search flow diagram.Detection of Early Cardiac InjuryVentricular Volumes, Systolic Function, and MassPrediction of subsequent cardiac toxicity15,20 based on small early changes in LV volumes, mass, or function has been difficult with echocardiography or multiple gated acquisition. In contrast, the high accuracy and precision of CMR may allow identification of early cardiac injury by showing subtle changes that do not meet current criteria for cardiotoxicity (Table 1).19,29 A preliminary study by Smith et al of 28 anthracycline-treated breast cancer patients illustrated that a significant increase in LV mass on day 3 of therapy (likely because of edema) predicted subsequent drop in LVEF at 1 year (Appendix Table I in the online-only Data Supplement). Another preliminary study by Grover et al of 36 patients with breast cancer demonstrated a significant increase in LV end-systolic volume by CMR as the earliest indicator of cardiac injury that occurred even before a rise in troponin or fall in global longitudinal strain.Table 1. Published CMR Studies Using LV Function, Mass, Strain, or Early Relative Gadolinium EnhancementAuthorMethodCancernAge*/SexTreatmentImaging TimingFindingsPrognosisLunning et al30Strain (feature tracking)Non-Hodgkin lymphoma10 (50% cardiotoxicity)Median (range) age 59 (55–74), 40% womenDoxo (300 mg/m2)Baseline, 3 mo post-DoxoSignificant reduction in circumferential strain post-therapy. Trend toward GLS reduction…Drafts et al29LV volumes, EF and circumferential (circ) strain (SPAMM)Breast and heme53Mean±SE, 50±2 y, 58% womenAnth (Doxo equivalent 50–375 mg/m2), radiotherapy 0%Baseline, 1, 3, and 6 moBy 6 mo significant increase in LVESV (48±3 to 54±3), reduction in LVEF (58±1% to 53±1%) and circ strain (−17.7±0.4% to −15.1±0.4%)…Ylänen et al31LVEFMultiple pediatric6214.6±3.2 y, 55% womenAnth (80–419 mg/m2), radiotherapy 11%Median, 7.8 y post-therapyLVEF and RVEF <55% in 79% and 80%, respectively. Significant dilation of LV and RV in 15% to 100% of patients. LV mass normal…Armstrong et al19LVEF and massMultiple pediatric114 (5% with previous CM)Median (range) age, 39 (22–53) y, 59% womenRadiotherapy 32% ± Anth (0 to >350 mg/m2)Median 27.8 y postcancer diagnosisLVEF <50% in 14%. LVEF and LV mass 2 SD below normal in 32% and 48% of patients…Neilan et al32LVMIMultiple91, all with known Anth CM, (82% stage C HF)43±18 y, 41% womenAnth (276±18 mg/m2), radiotherapy 33%Median 88 mo post-therapyInverse relationship between LVMI and Anth dose (r=−0.67)For 27 mo f/u, CV events† more common with lower LVMI (HR, 0.87), LVMI of <57 gm/m2—sens 100%, spec 85% to predict CV eventsChaosuwannakit et al33LVESV, LVEFBreast, lymphoma, or leukemia40 with cancer, 13 controls52±11 y, 70% womenAnth, trastuzumab (in combination or alone), radiotherapy 0%Baseline and 4 mo after treatment initiationSignificant increase in LVESV (49.7±18.2 to 54.9±12.5 mL) and LVEF (58.6±6.3 to 53.9±6.4%) at 4 mo…Oberholzer et al34LV and RV functionMultiple28Mean 16.4 y, 53% womenAnth 202 (81–462) mg/m20–20 y post-therapy (average 3 y)Reduction in LVEF in 15/28 to <55%…Wassmuth et al35Early gadolinium enhancement imaging to measure relative enhancement (RE)Multiple22Mean age 43 y, 77% womenDoxo (67±25 mg/m2) or epirubicin (76±19 mg/m2), radiotherapy 9%Pretherapy 3 d and 28 d postLV EF drop 67.8±1.4% to 58.9±1.9% at 28 d. Increase in RE to ≥5 on d 3 associated with drop in LVEF on d 28 and persistent EF reduction at 6 mo (subgroup)…Anth indicates anthracycline; CM, cardiomyopathy; CMR, cardiac MRI; CV, cardiovascular; Doxo, doxorubicin; EF, ejection fraction; GLS, global longitudinal strain; heme, hematologic; HF, heart failure; HR, hazard ratio; LV, left ventricular; LVESV, LV end-systolic volume; LVMI, LV mass index; LVEF, LV ejection fraction; RV, right ventricular; RVEF, RV ejection fraction; SPAMM, spatial modulation of magnetization.*Mean±SD unless otherwise stated.†Cardiovascular deaths, appropriate implantable cardioverter defibrillator therapy, admission for decompensated HF.An ongoing observational study is examining whether early change in CMR LVEF and volumes between baseline and 3 months in 150 patients receiving anthracycline therapy (with or without trastuzumab) could predict the degree of LVEF drop by 24 months (Appendix Table II in the online-only Data Supplement).36Myocardial StrainEarly reduction in myocardial strain or strain rate using echocardiography has been shown to predict subsequent cardiotoxicity.15,22 Similarly, CMR-based techniques such as spatial modulation of magnetization have shown a decline of mid-wall circumferential strain at 1 month of therapy in 53 asymptomatic patients treated with anthracyclines and have remained reduced at 6 months (Table 1).29 However, whether these early changes predicted subsequent cardiotoxicity was not reported. An ongoing trial in 50 patients with lymphoma with CMR will provide additional data on the use of myocardial strain to identify early injury and predict subsequent cardiotoxicity.37Myocardial InflammationMyocardial inflammation detected with CMR based on early gadolinium enhancement using T1-weighted fast-spin echo sequences has been used as a marker of myocardial injury in patients with myocarditis (Figure 2).38 Using this method in 22 asymptomatic patients receiving anthracyclines for solid or hematologic malignancies, a ratio of signal intensity (SI) differences between myocardium and skeletal tissue between pre- and postcontrast acquisition ≥5 on day 3 was associated with a reduction in LVEF at 28 days (Table 1).35 Also, there was a modest (r=−0.57) correlation between increased SI at day 3 and lower LVEF at 6 months.35 These changes may be attributable to either an increase in extracellular distribution volume or enhanced water exchange between the compartments because of early myocardial injury.Download figureDownload PowerPointFigure 2. Early gadolinium enhancement in a patient with breast cancer receiving adjuvant trastuzumab therapy referred with concern of cardiotoxicity. Left ventricular ejection fraction by cardiac MRI was 54%. Signal intensity measurements (mean/SD) in the myocardium and skeletal muscle precontrast (left) and postcontrast (right) illustrated. Calculated early gadolinium enhancement ratio was 16.3 (abnormal), suggestive of capillary leak/inflammation.These findings seem to be confirmed by a preliminary study by Kotwinski et al of 51 patients with breast cancer treated with anthracyclines (with or without trastuzumab; Appendix Table I in the online-only Data Supplement). Among the 20% with cardiotoxicity, a higher increase in early gadolinium enhancement was present between baseline and day 3 after first cycle of therapy and had an area under the curve of 0.75 to discriminate subsequent cardiotoxicity. High-sensitive troponin T was not elevated in any of the patients after the first cycle. An ongoing trial will perform early gadolinium enhancement imaging before, at completion of anthracyclines (3 months), and at 1 year in patients with lymphoma and will likely provide additional information about the value of early gadolinium enhancement in predicting cardiotoxicity.37Myocardial EdemaT2-weighted imaging or T2/T1 mapping techniques (Figure 3) can identify myocardial edema (Table 2).39 Using a rat model of early anthracycline toxicity, T1 and T2 values were elevated in explanted hearts of doxorubicin-treated animals in comparison to controls even in the absence of ventricular dysfunction or histopathologic evidence of myocardial fibrosis or necrosis.13 By direct measurement, there was a good correlation between myocardial water content and CMR-measured T1 and T2 values. A small human study of anthracycline-treated patients showed myocardial edema in 2 of 7 patients at the end of therapy.34 However, the prognostic significance of this finding was not presented. In another rat model of early doxorubicin toxicity, a treatment-related increase in SI >2.9 on postgadolinium images compared with pretreatment was associated with subsequent deterioration of LV function or mortality (Table 2).27 None of the treated animals without events or the control group had an increase in SI. All animals with events had histopathologic evidence of myocellular vacuolization attributed to intracellular edema without myocardial fibrosis or necrosis.Table 2. Published CMR Studies of Edema ImagingAuthorMethodCancernAge/SexTreatmentImaging TimingFindingsPrognosisLightfoot et al27Postcontrast, myocardial Gd-SINo malignancy40 ratsAge N/A, 100% maleControl (n=7), weekly doxo 1.5 mg/kg (n=19) and 2.5 mg/kg (n=14)Pre, 2, 4, 7, and 10 wk post doxo initiationIn animals experiencing adverse cardiovascular events* (16/33 treated animals), Gd-SI increased compared with baseline at time of eventAn increase in Gd-SI >2.9 at midterm from baseline predicted subsequent CV events*Oberholzer et al34TIRMMultiple7N/AAnthImmediately post anth2/7 had myocardial edema…Cottin et al13T1 and T2 relaxation times of ex vivo hearts (0.47-T magnet)No malignancy23 ratsAge N/A, 100% maleControl (n=10), low dose doxo 1 mg/kg intraperitoneal for 10 d (n=13; early injury model)7 d post completion of treatmentIncreased LV T1 and T2 values with doxorubicin. Correlations between T1 and T2 values and percent water content r=0.46 to 0.75. No change in LV performance…Thompson et al40T1 and T2 relaxation times of ex vivo hearts (0.47-T magnet)No malignancy22 rats4 wk, 100% maleControl (n=11), doxo (n=11) 1 mg/kg 7 wk, then 2 mg/kg 8–9 wk subcutaneous(chronic model)At 16, 17, 18, or 19 wkIncreased LV T1 values with doxo 651±30 vs control 623±20 ms and higher in more advanced disease. No difference in T2 values 45.8±2.6 vs 47.5±3.6 ms. Myocardial water content not increased…Anth indicates anthracycline; Doxo, doxorubicin; Gd-SI; gadolinium signal intensity; LV, left ventricular; N/A, not available; and TIRM, turbo inversion recovery magnitude sequence.*Deterioration of LV ejection fraction or unanticipated death.Download figureDownload PowerPointFigure 3. T2 maps at the basal, mid, and apical slices (A–C) and T1 map in the basal slice (D; in a patient with breast cancer referred for cardiotoxicity—asymptomatic—during adjuvant trastuzumab therapy). Cardiac MRI left ventricular ejection fraction was 46%, and T2 maps illustrated patchy areas of elevated T2 values (basal anterior septum, anterior and inferior wall, mid-anterior septum, apical septum, and lateral wall) and T1 maps illustrated increased regional native T1 values (in the basal anterior and inferior wall). Mean/SD T2 and T1 values per segment are provided.Three preliminary human studies (Appendix Table I in the online-only Data Supplement) have assessed the use of myocardial edema imaging during cancer therapy. Using T2-weighted sequences (short tau inversion recovery) in 28 patients with breast cancer, Smith et al illustrated that a significant increase in SI on day 3 of therapy was seen among patients who had a drop in LVEF at 1 year. Two other studies (n=29 and 21, respectively) by Grover et al using short tau inversion recovery imaging in patients with breast cancer receiving anthracycline and trastuzumab showed an increase in myocardial SI compared with skeletal muscle of >1.9 at 3 months after start of therapy in 58% and 52% of patients, respectively. The latter of the 2 studies identified cardiotoxicity in 43% of patients at 12 months. However, neither study reported whether the increase in SI at 3 months predicted subsequent cardiotoxicity.An ongoing trial of T1 and T2 mapping in 50 patients with lymphoma receiving anthracyclines will compare T2 values at the completion of chemotherapy (≈3 months) between those who have preserved and reduced LVEF at 1 year.37 This study will also examine the relationship of T1 and T2 values to various biomarkers and myocardial mechanics and diastolic parameters by echocardiography. Another ongoing study will also assess the ability of T1 signal at 3 months to predict LV function changes at 24 months in anthracycline-treated patients.36Detection of Cardiotoxicity During or Early After Cancer TherapyVentricular Volumes, Systolic Function, Mass, and StrainIn a recent study of patients with hematologic or breast malignancy treated with anthracyclines, CMR was performed before, at 1, 3, and 6 months into treatment.29 By 6 months, the LV end-systolic volume increased significantly with a concomitant drop in LVEF (Table 1). These changes were seen as early as 1 month into therapy. By 6 months, 26% of patients who had EF >50% at baseline experienced cardiotoxicity (EF <50%). In a previous publication, the same group demonstrated a significant increase in LV end-systolic volume and a decrease in LVEF between a baseline and 4-month CMR in 40 patients with hematologic or breast malignancy treated with anthracyclines (Table 1).33 Likewise, another small study of multiple malignancies detected a significant reduction in LVEF for the whole group by day 28 (Table 1) with an EF <55% in 27% of anthracycline- treated patients (Table 1).35 These findings are consistent among CMR-based studies, with other preliminary work by Grover et al (Appendix Table I in the online-only Data Supplement) emphasizing increase in LV volumes and reduction of LVEF and right ventricular EF as early as 1 month into therapy and persisting up to 12 months with cardiotoxicity identified in in a subgroup. An ongoing observational study is examining serial CMR versus multiple gated acquisition–based LVEF for the detection of cardiotoxicity, its association with biomarkers, and prognosis in 50 trastuzumab-treated women with breast cancer.41 Currently, only 1 study of 10 adult patients receiving anthracyclines for hematologic malignancy illustrated reduced circumferential strain at 3 months after therapy using a feature-tracking algorithm (Figure 4) but in the context of reduced LVEF.30Download figureDownload PowerPointFigure 4. Myocardial strain analysis using a feature-tracking algorithm in the same patient as in Figure 2. Peak systolic global longitudinal strain was mildly reduced, whereas circumferential strain was in the normal range. Yellow arrows (left) represent velocity vectors; curves represent strain measurements in each of the myocardial segments (6 segments) and a global curve (black).Late Gadolinium EnhancementLate gadolinium enhancement (LGE) imaging allows detection of myocardial fibrosis and scar and is widely used with good diagnostic and prognostic value in cardiovascular diseases.42 However, the limited data on the incidence, pattern, and prognostic significance of LGE in patients receiving cancer therapy are conflicting (Figure 5). Two retrospective and 1 prospective studies from the same group have illustrated the presence of LGE in the context of established cardiomyopathy in patients with breast cancer during5,43 and at end of therapy15 with anthracyclines and trastuzamab (Table 3). The pattern of LGE was subepicardial or myocarditis-like, and the incidence ranged between 94% and 100%. More recently, in 10 patients with non-Hodgkin lymphoma, new or progressive midmyocardial LGE was seen in 30% of patients 3 months after completion of therapy, some in the context of preserved LVEF.30 The potential prognostic value of LGE was shown in 2 retrospective studies5,43 where ≈40% of patients with LGE during therapy either had no improvement or further decline in LVEF at 6 months despite stopping trastuzumab and treatment with angiotensin converting enzyme (ACE) inhibitors and β-blockers. Unfortunately, neither of these studies adjusted for other covariates.5,43 Likewise, a preliminary report by Jordan et al (Appendix Table I in the online-only Data Supplement) of 51 patients with breast or hematologic malignancy receiving anthracyclines and decreasing LVEF described a diffuse pattern of signal enhancement on LGE images in all patients at 6 months into therapy reflecting either diffuse fibrosis or edema.Table 3. Published CMR Studies of Late Gadolinium EnhancementAuthorCancernAge*/SexTreatmentImaging TimingFindingsPrognosisLunning et al30Non-Hodgkin lymphoma10 (50% had Ctox)Median (range), 59 (55–74) y, 40% womenDoxo (300 mg/m2)Baseline, 3 mo postdoxo30% with 1 new or progressive segment of LGE…Drafts et al29Breast and heme5350±2 y, 58% womenAnth Rx, Doxo equivalent dose 50–375 mg/m2, radiotherapy 0%Baseline, 1, 3, and 6 moNo LGE…Neilan et al32Multiple91 with Anth CM, 82% stage C HF43±18 y, 41% womenAnth (276±18 mg/m2), radiotherapy 33%Median 88 mo post-RxLGE only seen in 6% (mean EF 36±8%)LGE not predictive of MACE at a median follow-up of 27 mo†Lawley et al44Breast cancer (HER2+)2564±7 y, 100% womenAnth (80%) with adjuvant trastuzumab, radiotherapy 48%Mean 20±8 mo post-trastuzumab completionLGE in 8% (basal inferior lateral wall midmyocardial segments (all with LVEF >50%)…Fallah-Rad et al15Breast cancer (HER2+)42 (10/42 had Ctox)42±9 y, 100% womenAnth, adjuvant trastuzumab, radiotherapy 98%Baseline and 12 moAll with Ctox had subepicardial linear LGE in lateral LV wall (18±4% of total LV mass). No LGE in patients without Ctox…Wadhwa et al5Breast cancer (HER2+)152 patients, 36 with Ctox52±10 y, 100% womenAnth, adjuvant trastuzumab, radiotherapy 69.1%.At diagnosis of toxicity and 3 and 6 mo later34/36 with Ctox had subepicardial LGE of 75% thickness of lateral wallDespite stopping therapy and starting HF therapy, 16/36 had no improvement or further decline in LVEFFallah-Rad et al43Breast cancer (HER2+)10 patients with Ctox (out of 160)40±8 y, 100% womenAnth, adjuvant trastuzumab, radiotherapy100%After toxicity detected by MUGA or echoSubepicardial LGE of lateral LV wall in all (mean EF 29±4% by CMR)With stopping trastuzumab 60% had improvement in LVEF within 6 mo, 40% did not (despite ACE and β-blocker)Anth indicates anthracycline; CM, cardiomyopathy; CMR, cardiac MRI; Ctox, cardiotoxicity; Doxo, doxorubicin; EF, ejection fraction; HF, heart failure; LGE, late gadolinium enhancement; LV, left ventricular; LVEF, LV ejection fraction; MACE, major adverse cardiovascular events; and MUGA, multiple gated acquisition.*Mean±SD unless otherwise stated.†Cardiovascular deaths, appropriate implantable cardioverter defibrillator therapy, admission for decompensated HF.Download figureDownload PowerPointFigure 5. Delayed enhancement imaging in the same patient as in Figure 3, illustrating a focal area of midmyocardial enhancement (arrows) involving the basal inferior septum seen in both short-axis (A) and 4-chamber (B) views. This finding is confirmed by lower postcontrast T1 values (mean/SD; C) in the segment with scar/fibrosis (277 ms) in comparison to the remote segment (350 ms).In contrast, a recent study of anthracycline-treated patients did not identify LGE in any patient with imaging performed at baseline, 1, 3, and 6 months, despite a significant drop in LVEF over the 6-month period (Table 3).29 An ongoing clinical trial will compare LGE with later LVEF changes in 90 trastuzumab-treated women with and without LV dysfunction at 6 months into therapy.45Arterial StiffnessIn patients with cardiovascular risk factors and advanced age, increased thoracic aortic stiffness has been associated with adverse cardiovascular events.46 Two studies (Table 4) have examined the impact of anthracycline therapy on aortic stiffness measured by CMR.29,33 In patients receiving anthracyclines for breast or hematologic malignancies, a significant increase in pulse wave velocity and decrease in ascending aortic distensibility was seen between baseline and 4 and 6 months.29,33 In both studies, the whole population had a reduction in LVEF. These studies demonstrate that the early consequences of anthracycline therapy extend beyond the cardiac system.36Table 4. Published CMR Studies of Vascular ChangesAuthorMethodCancernAge/SexTreatmentImaging TimingFindingsDrafts et al29Thoracic aortic pulsed wave velocity (PWV) using phase-contrast CMRBreast lymphoma or leukemia5350±2 y, 58% womenAnth, Doxo equivalent range of 50–375 mg/m2Baseline, 1, 3, and 6 moAt 6 mo in comparison to baseline significant increase in PWV from 6.7±0.5 to 10.1±1.0 m/sChaosuwannakit et al33Thoracic aortic PWV and distensibility using phase-contrast CMRBreast lymphoma or leukemia40 with cancer, 13 controls52±11 y, 70% womenAnth, cyclophosphamide, trastuzumab (in combination or alone)Baseline and 4 mo after treatment initiationSignificant increase in PWV (6.9±2.3 to 13.5±4.7 ms) and decrease in distensibility (4.1±1.6 to 1.9±1.2 mm Hg) at 4 mo in patients receiving therapyAnth indicates anthracycline; CMR, cardiac MRI; and Doxo, doxorubicin.Currently, the same group is assessing multiple CMR markers including arterial stiffness at baseline, 3, and 24 months of therapy in adults receiving anthracyclines alone or in combination with paclitaxel or trastuzumab.36 Another ongoing study will be assessing aortic stiffness changes in 25 patients with cancer being treated with anthracyclines.47 In addition, this study will measure aortic stiffness in 60 childhood cancer survivors 1 to 15 years postanthracycline therapy.47Detection of L