The focus of this review will be on studies that have used CMR techniques to quantify MCT metrics in patients who have undergone RT. Though clearly important, a full review of chemotherapy-induced cardiotoxicity (without RT) and evaluation of cardiotoxicity metrics using other imaging modalities or biomarkers is beyond the scope of this article. We point the interested reader to numerous informative prior reviews [12], [37]–[39]. Searching PubMed, Google Scholar, ScienceDirect, and Google, with the following phrases: ‘radiotherapy and cardiotoxicity using MRI’ and ‘Using MRI to detect cardiotoxicity after radiotherapy’ resulted in ‘54’ and ‘11’, ‘9980’ and ‘12000’, ‘1128’ and ‘882’, ‘154000’ and ‘95300’ papers, respectively. Papers were sorted by relevance using automatic filtering and the first 1000 papers from each database was selected for initial screening. Next, papers were excluded based on the title and/or the abstract review leading to 1560 eligible papers for comprehensive review. Final exclusion was made based on lack of RT in the treatment plan, missing MRI-based metrics for cardiac function assessments, dismissing myocardial evaluation in MRI-based studies, and the use of non-human subjects for RT-induced MCT measurements. In total, 21 papers ranging from 2011-2022 were identified that specifically focused on RT-induced MCT evaluation in human studies using CMR.
Figure 1 shows the workflow for identification of studies via databases following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 flow diagram [40]. Ten studies were cross-sectional and eleven were longitudinal. Studies included patients who received chest radiation for a number of malignancies including: breast cancer, lung cancer, esophageal cancer, and Hodgkin’s and non-Hodgkin’s lymphoma. In cross-sectional reports, the sample size, mean dose to the heart, and time of comparison since RT varied between 5-80 patients (age 19-70 years old), 2.1-22.9 Gy, and 23.5 months to 24 years, respectively. In the longitudinal studies, sample size varied from 10 to 81 patients (age 8.5-69 years old), the mean heart dose ranged between 2-13.9 Gy, and the length of serial observation (after acquiring baseline pre-RT data) ranged from 0 to 24 months after RT completion. In this review article, findings will be summarized and grouped by types of MRI-derived metrics, followed by discussion and suggestions for future works. Metrics were assessed either globally (i.e., on the whole LV or LV myocardium) or regionally (i.e., at focal locations using an American Heart Association (AHA) model or over random segmental regions). Figures 2-3 show a detailed summary of longitudinal and cross-sectional studies including number of patients, type of dataset (i.e., type of cancer), amount of radiation to the heart, and the timing of MRI-based measurements.
II.A Left ventricular ejection fraction (LVEF)
LVEF is the most common screening tool to quantify global cardiac function and is defined as the stroke volume divided by the end diastolic volume. Normal LVEF ranges between 50-70% [41]. In terms of LVEF, cardiotoxicity has been defined as any LVEF decline to <50% or an LVEF decline >10%, ≥10%, or >15% from baseline to <55%, <50%, or ≥50%, respectively [42]. Benefits of this metric and defined cutoffs for cardiotoxicity are that LVEF is a widely accepted and clinically utilized metric of cardiac function assessable by multiple imaging modalities (though these modalities may result in slightly different values for the same person [43]). A primary drawback is that declines in LVEF may not be observed in early MCT [44], [45] since global cardiac function may be able to compensate for mild early or focal cardiac damage. Thus, reliance on LVEF alone may delay the diagnosis and early treatment of MCT.
II.A.1 LVEF Decrease
A decrease in LVEF was reported in a few studies. In 31 patients with Hodgkin’s disease who were treated with mediastinal RT, LVEF dropped below 55% in 23% of patients at 24 years post-RT treatment [46]. Also, long-term survivors of Hodgkin’s lymphoma (HL) and non-HL (NHL) who were treated with mediastinal RT in combination with chemotherapy (85% of the cohort) demonstrated lower LVEF compared to healthy controls (53±5% vs 60±5%, p<0.001) at a median of 20 years post-treatment [47]. Finally, in a small study with ten patients with different thoracic malignancies, LVEF dropped below 50% in only one patient mid-treatment and in two patients at the end of therapy (all patients’ LVEF were above 50% at baseline) [48]. No measurements were repeated beyond the end of RT.
II.A.2 LVEF Unchanged
LVEF stayed within normal limits in multiple studies. In 20 breast cancer patients who received 3D conformal RT (3DCRT), a median LVEF of 63% was reported at 8.3 years [45]. Similar findings were noted for 5 HL survivors who underwent chemotherapy and proton RT with LVEF of 60% at 5 years post treatment [49]. Over shorter follow-ups, LVEF did not change at 3 months post-RT in 51 breast cancer patients who were exposed to a low dose of radiation to the heart (2 Gy) [44], at 2- or 6-months in 21 lung cancer patients following 3DCRT (69-70%) [50], at 6 months in 11 chest tumor patients who underwent chemoradiotherapy [51], at either therapy completion or 13 months post-therapy in 66 breast cancer patients who were treated by chemoradiotherapy (60%) or only RT (62-64%) [52], and finally at 6- or 18-months in 19 and 24 survivors of esophageal cancer, respectively, who received chemoradiotherapy (60-65%) [53], [54]. Additionally, no change in LVEF was noted for 81 pediatric cancer patients with different malignancies who were treated by chemotherapy with/without irradiation at 1-year and 2-year follow up (range: 60-62%) [55]. Finally, in 49 left-sided breast cancer patients who were treated by intensity-modulated radiation therapy (IMRT), 3DCRT, with/without chemotherapy, LVEF was measured over a 24-month follow-up. No significant reduction of LVEF was noted at 6 months in the whole cohort (59.2- 61.2%, p=0.059); however, patients who underwent 3DCRT with/without chemotherapy demonstrated a temporary decrease of LVEF which was resolved at 12- and 24-months. Interestingly, for the IMRT group, the LVEF increased at 24-months follow up (60.1-63.6%, p=0.017) [56]. Table 1 shows a summary of LVEF results for RT-induced MCT studies using MRI.
Table1. LVEF changes in RT-induced MCT studies using MRI.
Study
|
Decrease
|
Increase
|
Constant
|
Machann et al., [46]
|
<55% in 23% of patients at 24 years post-RT
|
-
|
-
|
Van Der Velde et al., [47]
|
53±5% vs 60±5%, p< 0.001 at 20 years post-RT
|
-
|
-
|
Heggemann et al., [56]
|
In 3DCRT with/without chemotherapy group: temporary decrease at 6 months (59-62%, p=.031)
In 3DCRT without chemotherapy group:
temporary decrease at 6 months (59-63%, p=.005) resolved at 12 and 24 months:
(p=0.443)
|
For IMRT group: at 24 months follow-up (60.1-63.6%, p=0.017)
|
At 6 months: the whole cohort (59.2-61.2%, p=0.059)
|
Goyet et al., [55]
|
-
|
-
|
Baseline: 62±8%
One-year: 60±7%
Two-year: 61±6%
|
Traber et al., [48]
|
Half-therapy completion: (dropped below 50% in 1/10 patients) after RT completion: (dropped below 50% in 2/10 patients)
|
-
|
-
|
Bergom et al., [45]
|
|
-
|
Normal range (63% (52-77%)) at 8.3 years post-RT
|
Takagi et al., [54]
|
|
-
|
Normal range (baseline: 63±9%,
0.5 year follow up: 65±12%, 1.5 year follow up: 61±11%)
|
Bates et al., [49]
|
At 5 years: >5% LVEF drop in two patients with mean cardiac RT dose of ≥ 10 Gy and a total anthracycline dose of greater than 250 mg/m2
|
-
|
At 5 years: Median LVEF was 60% (52-61%)
|
Umezawa et al., [53]
|
-
|
-
|
Normal range (baseline: 60.4±8.9%, 6 months: 62.8±12.7%, and 1.5 years: 62±10.4%)
|
Vallabhaneni et al., [51]
|
|
|
No significant % changes in patients with higher radiation (-10.4±7.7%) or patients with minimal radiation (-8.9±8.6%) at 6 months post-RT
|
Lideståhl et al., [50]
|
|
|
Pre-RT: 69 (63-74) %,
2 months: 69.5 (65.5-74.8) %,
3 months: 70 (63-75) %
|
Speers et al., [44]
|
-
|
-
|
No significant changes of LVEF at 3 months post-RT
|
Tahir et al., [52]
|
-
|
-
|
In epirubicin-chemotherapy-based
followed by RT group:
unchanged (at baseline: 60±5%, therapy completion: 60±6% and after 13±2 months: 60±6%)
|
-
|
-
|
In left-sided RT only group:
unchanged (baseline: 62±5%,
therapy completion: 64±6%,
13±2 months: 62±5%)
|
II.A.3 Discussion on LVEF
Recent RT-induced MCT studies showed no significant changes of LVEF and/or relations between LVEF and dose mostly due to short follow-ups [51], [52], low dose of radiation to the heart [44], small sample size [49], and uncertainty in the exact cumulative dose for longer follow-up studies. However, studies with longer follow-up (20 years and more) showed a decrease in LVEF [46], [47]. Also, studies that were conducted on patients treated with older RT techniques such as anterior-mantle-field, in which overdosage is expected in anterior parts of the irradiated volume, showed declines in LVEF following RT [46]. These patients with LVEF drop were treated for HL and NHL and were exposed to large RT fields covering large volumes of the mediastinum and heart with simple RT techniques. More recent studies, including those for breast cancer, have used more advanced techniques and were therefore able to reduce cardiac dose. This may also contribute to the observed lack of changes in LVEF. Overall, these findings suggest that LVEF is a poor indicator for the reliable detection of early RT-induced MCT.
II.B Cardiac Chamber Dimensions
A limited number of studies investigated the dimensions and mass of different cardiac chambers to find possible early changes of subclinical myocardial dysfunction and associated dose-dependency.
II.B.2 Increase in Cardiac Chamber Dimensions
Increase of cardiac volumes and dose-dependency were noted in a few studies. In 3/11 patients with esophageal cancer, LV systolic volume increased at 3-5 months following neoadjuvant chemoradiotherapy [57]. An increase of LV end-systolic volume was also noted in 80 HL and NHL patients who were treated by mediastinal RT with/without chemotherapy at 20 years compared to healthy control (p=0.01) [47].
Dose-dependency response was found in two studies. In 81 pediatric cancer patients (with different malignancies) who were treated with chemotherapy and irradiation, increase in indexed LV end-diastolic volume (2.1±6.5 ml/m2, normalized to body surface area) at 2 years was correlated with the radiation dose to 98% of the heart volume (5.8±4.6 Gy, p<0.05) and LV (6.1±5.0 Gy, p<0.05) [55]. In addition, a higher LV mass index, a predictor of cardiovascular events, was shown to be correlated with LV mean dose (p=0.012), V10 (LV volume that received 10 Gy, p=0.027), and V25 (LV volume that received 25 Gy, p=0.016) at a mean of 8.3 years in 20 breast cancer patients who underwent anthracycline-based chemotherapy and 3DCRT [45].
II.B.1 Cardiac Chamber Dimensions Unchanged
The dimensions of cardiac chambers stayed within normal range in two studies with breast cancer populations. In the first study with 49 patients who underwent IMRT, 3DCRT with/without chemotherapy, LV and RV volumes did not change over 24 months follow-up [56]. In the second study with 51 patients who were treated by RT, RV ejection fraction did not change at 3 months even with a temporary significant drop at the end of therapy [44].
II.B.2 Decrease in Cardiac Chamber Dimensions
Decreases of cardiac chambers were shown in two studies. At 20-year follow-up of 80 HL/NHL survivors who were treated by mediastinal RT with/without chemotherapy, a drop in RV volumes and LV diastolic volume/mass were noted (P<0.06) compared to healthy controls [47]. Also, in 66 breast cancer patients who were treated by epirubicin-based chemotherapy followed by RT, right- and left-sided chamber sizes were significantly decreased at the end of therapy (p<0.05), while only RV systolic volume and RA diastolic volume dropped significantly at 13±2 months (p<0.05). For those who underwent RT treatment, RV and LV volumes were reduced at both therapy completion and 13±2 months (p<0.05) without affecting the atrial dimensions [52]
II.B.2. Discussion on cardiac chamber dimensions
Over short follow-ups, cardiac chambers dimensions did not change and/or were recovered after therapy completion [44], [56], potentially due to compensatory features of the heart. Changes in cardiac chamber dimension were mostly noted over longer follow-up times [47], when damages are less likely to be reversible, and in concurrent treatments (e.g., epirubicin-based chemotherapy followed by RT) [52], in which cardiotoxic drugs enhance the probability of cardiac dysfunction. Finally, a relationship between radiation dose and LV end-diastolic volume was noted in pediatric cancer patients [55], who are shown to be more susceptible to radiation damage compared to adult patients [20].
II.C T1/T2 Mapping
T1 and T2 relaxation times are tissue-specific time-constants that can provide useful information regarding myocardial abnormalities and possible associated pathologies (e.g., fibrosis, edema) [36].
II.C.1 T1 Mapping
T1, or longitudinal relaxation time, is a biological parameter that quantifies the time required for nuclei of hydrogen atoms to recover towards thermodynamic equilibrium along the main magnetic field. It can be measured either globally or locally, which allows for quantification of heterogenous T1 distribution. T1 values differ based on local molecular environment (e.g., location-specific constitutive properties, temperature, and pressure), as well as sex, age, and other parameters. Normal native T1 values of the myocardium acquired using a modified Look-Locker inversion recovery (MOLLI) method at 1.5T and 3T scanners were reported as 950±21 ms and 1052±23 ms, respectively [58]. Notably, these values can vary with magnetic field strength, vendor, scanner model, and physical location. Pathologies can also change the tissue properties (e.g., water content) and hence the T1 values. Myocardial abnormalities associated with changes in T1 value include myocardial fibrosis, edema, inflammation, infiltrative diseases, amyloidosis, and hemosiderosis [36].
II.C.1.1 T1 increase
Increase of T1 signal was reported in a few studies. In 24 survivors of esophageal cancer who were treated with chemoradiotherapy, increase of T1 value was noted at 0.5 years (1257±35 ms, P <0.01) and 1.5 years (1238±56 ms, p<0.024) compared to the baseline (1183±46 ms) at the basal septum (a highly irradiated area (43±4 Gy)). However, no correlation was found between regional radiation dose and percent change of T1 at the basal septum [54]. Over longer follow-up periods, a significant increase of T1 was found in 80 HL and NHL survivors at 20 years following mediastinal RT with/without chemotherapy compared to healthy controls (980±33 ms vs 964±25 ms, p=0.01) [47].
II.C.1.2 T1 unchanged
The majority of studies found T1 values within normal range. During RT of 10 patients with different thoracic malignancies, no T1 signal changes were noted at half-therapy (956±14 ms) and at therapy completion (968±72 ms), compared to baseline (966±39 ms) [48]. A similar cohort of patients in another study demonstrated no T1 changes at 6 months in either highly or minimally irradiated patients [51]. Over longer follow-up periods, T1 in 28 patients with chest tumors stayed around 1009 ms (p=0.054) with no dose-dependent response at 46.6 months following chemoradiotherapy [59]. T1 signal also remained unchanged among 40 esophageal cancer patients who underwent neoadjuvant chemoradiotherapy (959.2±34.7 ms) compared to control (949.9±28.4 ms) (p=0.4) at 67.6 months [60].
Fluctuation of T1 values over different follow-up times was noted in a single study of 66 patients with breast cancer who were treated by epirubicin-based chemotherapy followed by RT or with left-sided RT alone. T1 value increased at therapy completion in the group receiving epirubicin-based chemotherapy followed by RT and returned to baseline at 13±2 months, while the group receiving left-sided RT only demonstrated no significant changes in T1 at both follow-ups [52].
II.C.1.3 T1 decrease
Decrease of T1 signal was found in 51 patients with breast cancer post-RT treatment (immediately at the end of treatment, -20 ms, p=0.022 and at 3 months, -23 ms, p<0.001) [44]. A summary of T1 value findings using MRI for RT-induced MCT studies is shown in Table 2.
Table 2. T1 value changes in RT-induced MCT studies using MRI (no dose-dependency was reported).
Study
|
Decrease
|
Increase
|
Constant
|
Van Der Velde et al., [47]
|
|
Significant increase of native T1 values compared to healthy control at 20 years post-RT (980±33 ms vs 964±25 ms, p=0.01)
|
-
|
Traber et al., [48]
|
|
-
|
Normal range: (baseline (966±39 ms), half-time RT (956±14 ms), and after RT (968±72 ms))
|
Takagi et al., [54]
|
|
In the basal septum (highly radiated area, 43±4 Gy), native T1 values were higher at 0.5 year (1257±35 ms, p<0.01) and 1.5 year (1238±56 ms, p< 0.024) compared to the baseline (1183±46 ms)
|
At the apical lateral wall (nonradiated area 3± 4 Gy), no significant T1 differences were found at different time points
|
Speers et al., [44]
|
End of the treatment
(−20 ms, p=0.022) and three months post-treatment (−23 ms, p< .001)
|
-
|
|
Tahir et al., [52]
|
-
|
In epirubicin-chemotherapy-based followed by RT group:
baseline: 1244±29 ms,
therapy completion: 1293±34 ms, p<0.001)
|
In epirubicin-chemotherapy-based followed by RT group: changes returned to baseline at 13±2 months (1250±26 ms)
|
-
|
-
|
In left -sided RT only group: constant
(baseline: 1237±29 ms, at therapy completion: 1237±42 ms, and 13±2 months: 1239±39 ms)
|
Ricco et al., [59]
|
|
|
Mean T1: 1009 ms (range 933–1117 ms)
with no dose-dependency at 46.4 months post-RT (p=0.054)
|
Vallabhaneni et al.,[51]
|
|
|
No significant %T1 changes in
patients with higher radiation (-1.3±3.7%) or patients with minimal radiation (-3.7±2.0%) at 6 months post-RT
|
de Groot et al., [60]
|
|
|
No differences between neoadjuvant chemoradiotherapy and control (959.2±34.7 ms vs 949.9±28.4 ms, p=0.4) at 67.6 months
|
II.C.1.4 Discussion on T1 mapping
Different behaviors of T1 signal (e.g., increase, decrease, unchanged) were reported following cancer treatment. Increase of T1 was noted in longer follow-ups [47], regional analysis at high-dose regions [54], and with concurrent treatments [52]. Another explanation is the type of cancer. For example, in breast cancer patients, the mean heart dose is low; typically, only a small region of the LV is irradiated. Therefore, it is not surprising that there are no T1 changes [52]; whereas in patients with esophageal cancer, much larger volumes of the heart can be in the high-dose fields [54]. Despite increase of T1 value in high dose regions, no correlations were found between T1 increase and dose [47], [54]. Unchanged T1 values were mostly seen in heterogenous patient population [59], shorter follow-ups, [48] and non-irradiated regions [54]. Interestingly, one study reported a decrease of T1 at the end of treatment and three months post-treatment [44]. Given these findings, it might be premature to conclude the consistent progression of fibrosis and/or inflammation shortly after RT in all cases. Further studies are required to examine this matter more fully.
II.C.2 T2 Mapping
Similar to T1, T2 (or transverse relaxation time) is a biological parameter that is tissue-specific. Normal T2 values of myocardium acquired from steady-state free precession (SSFP) technique ranged between 52.18±3.4 ms and 45.1 ms at 1.5T and 3T scanners, respectively. T2 changes are mostly associated with change of water content in the tissue. The main pathology associated with longer T2 values is myocardial edema [36].
II.C.2.1 T2 unchanged
In RT-involved treatments, T2 values were reported to be within normal range. At 24 years follow up of 80 HL and NHL survivors who were treated with mediastinal RT with/without chemotherapy, T2 values did not differ from healthy controls (both 50 ms, p=0.13) [47]. In shorter follow-ups, T2 signal in 51 breast cancer patients did not change at therapy completion or 3 months post-RT [44]. Similar findings were noted in highly or minimally irradiated patients with chest tumors at 6 months post chemoradiotherapy [51].
In 66 breast cancer patients with two different treatments, T2 stayed unchanged in the left-sided RT group at the end of therapy and 13±2 months post-RT (46-47 ms); however, in patients who underwent epirubicin-based chemotherapy followed by RT , T2 increased at therapy completion (48 ms vs 45 ms, p<0.001) and returned to baseline at 13±2 months (46 ms) [52]. Table 3 shows T2 signal changes in RT-induced MCT reports using MRI.
Table 3. T2 signal changes in RT-induced MCT studies using MRI (No decrease or dose-dependency were reported).
Study
|
Increase
|
Constant
|
Van Der Velde et al., [47]
|
-
|
No differences were noted between cancer patients and healthy control at 20 years post-RT (50±3 ms vs 50±2 ms, p=0.13)
|
Speers et al., [44]
|
-
|
No significant changes of T2 signal over different time points (therapy completion and 3 months post-RT)
|
Vallabhaneni et al., [51]
|
|
No significant %T2 changes in patients with higher radiation (3.9±9.5%) or patients with minimal radiation (-3.4±12.1%) at 6-months following RT
|
Tahir et al., [52]
|
In epirubicin-chemotherapy-based followed by RT group: increased at therapy completion compared to baseline (48±3 ms vs 45±3 ms, p<0.001)
|
In epirubicin-chemotherapy-based followed by RT group: returned to baseline at 13±2 months (46±3 ms)
|
-
|
In left-sided RT only group: constant at baseline (46±3 ms), therapy completion (47±2 ms) and after 13 ± 2 months (46±3 ms)
|
II.C. Discussion on T2 mapping
T2 signal stayed unchanged over longer follow-ups [47] (with a higher likelihood of progression of fibrosis rather than edema) and over shorter follow-ups at low dose (e.g., 2 Gy) [44], except for concurrent treatments [52]. Even with concurrent treatments, the changes were resolved in a few months. These findings can demonstrate the low likelihood of myocardial edema development following RT [44], [47], [52]. Regional analysis of T2 mapping in longitudinal studies of patients with higher dose to the heart over both short and long-term follow ups are required for further evaluation.
II.D Extracellular volume fraction (ECV)
The cellular components of myocardium, including the interconnected cardiac muscles, are embedded in a complex three-dimensional extracellular space that accounts for the interstitial (or extracellular) component of the myocardium. One of the distinct features of myocardial pathologies (e.g., myocardial fibrosis, inflammation, edema) is the expansion of this extracellular space. Quantitative evaluation of extracellular expansion is now possible by acquisition of the hematocrit and pre- and post-T1 values of myocardium and blood pool (before vs after administration of a contrast agent (e.g., gadolinium)). Equation 1 shows the formula for ECV calculation.
Normal myocardial ECV values of 25±4% and 26±4% at 1.5T and 3T, respectively, have been reported [36], [58]. It should be noted that ECV values even in healthy volunteers may differ based on age, sex and type of scanner [36].
II.D.1 ECV increase
Increase of ECV and its dose-dependency have been reported in a few recent studies. Increase of ECV (32% vs 26%, P<0.01) was noted at 6 months post chemoradiotherapy in 24 patients with esophageal cancer in high dose regions (43±4 Gy) [54]. Segmental analysis of ECV (i.e., measured over focal regions) in 40 patients with esophageal cancer who received neoadjuvant chemoradiotherapy showed a linear relationship between mean dose per segment and ECV (a 0.136%-point increase of ECV for each Gy (p<0.001)) at 67.6 months [60]. Lastly, a substantial increase of ECV (45%) was also reported in an esophageal cancer patient at 8 years following chemoradiotherapy [61].
II.D.2 ECV unchanged
No significant change and/or relation between ECV and dose were also reported in a few studies. In 80 HL and NHL survivors who were treated by mediastinal RT with/without chemotherapy, ECV did not differ compared to healthy controls in patients with 20-year follow-up (28% vs 29%, p=0.24) [47]. Similarly, global ECV (i.e., in the whole myocardium) of 27% was measured in the myocardium of 20 breast cancer patients after 8.3 years of anthracycline-based chemotherapy and 3DCRT [45]. Over shorter time-points, ECV stayed unchanged at 2 years post-RT for 30 patients with various chest malignancies [62] and among 66 breast cancer patients who underwent epirubicin-based chemotherapy followed by RT (28% vs 29%) or left-sided RT treatment only (30% vs 30%) at 13±2 months follow-up [52]. A 6-months follow-up of 11 patients with various chest tumors also did not show significant ECV changes between highly or minimally irradiated patients [51]. Table 4 shows ECV measurements following RT treatment.
Table 4. ECV changes in RT-induced MCT studies using MRI (No decrease of ECV was reported).
Study
|
Increase
|
Constant
|
Dose dependent
|
Van Der Velde et al.,[47]
|
-
|
No differences with healthy control at 20 years post-RT (28±3% vs 29±3%, p=0.24)
|
-
|
Bergom et al., [45]
|
|
Mean global ECV of 27% (range: 23-34%) at 8.3 years post-RT
|
-
|
Takagi et al., [54]
|
In basal septum (high radiated area, 43±4 Gy) in 0.5 year follow up (26±3% vs 32±3%, p<0.01)
|
-
|
-
|
Tahir et al., [52]
|
-
|
In epirubicin-chemotherapy-based followed by RT group:
(Baseline: 28±2% and
13±2 months: 29±2%, p=0.52)
|
-
|
-
|
In left-seded RT only group:
(baseline: 30±3% and
13±2 months: 30±3%)
|
-
|
de Groot et al., [60]
|
(28.4±1.0% vs 24.0±0.9%; p<0.001)
|
-
|
Linear relation between mean dose per segment and ECV increase (a 0.136%-point increase of ECV for each Gy (p<0.001) at 67.6 months
|
Canada et al., [62]
|
|
no associations with dose: median 28% [26–31] at 2 years post-RT
|
-
|
Mukai-Yatagai et al., [61]
|
Increase of ECV (45%) at 8 years
|
-
|
-
|
Vallabhaneni et al., [51]
|
|
No significant % ECV changes in patients with higher radiation (-11.5±20.8%) or patients with minimal radiation (-8.1±2.9%) at 6-months following RT
|
|
II.D. Discussion on ECV
Global analysis of ECV [47] and short follow-up times (less than 2 years) [52], [62] did not show any changes post-RT. However, segmental analysis (i.e., by the American Heart Association (AHA) model) of ECV at high-dose regions over short follow-up times (0.5 year) [54] and global analysis over longer follow-ups in concurrent treatments (greater than 67.6 months) [60] demonstrated an increase of ECV signal. This might suggest that short follow-up increases in ECV are associated with focal and/or short-term inflammation caused by chemoradiotherapy, followed by fibrosis. Though, it should be noted that in concurrent treatments, such as the use of cisplatin (which increases the risk of late cardiovascular events) and 5-fluorouracil (which is associated with myocardial ischemia), it may be difficult to differentiate the effects on the interstitial myocardium (i.e., ECV changes) from RT alone versus systemic treatment [54]. ECV was also elevated at higher RT dose regions [54]. More importantly, one study found a linear relationship between ECV and segmental dose as a 0.136%-point increase of ECV for each Gy (P <.001) [60]. Further systematic analysis of ECV is necessary to determine its value for identifying fibrosis and/or inflammation associated with RT-induced MCT. It should also be noted that ECV, like T1/T2 but in lesser extent, is subject to scanner and patient-specific variabilities; therefore, longitudinal studies where each follow-up measurement is compared to its baseline value are more meaningful to report.
II.E Late Gadolinium enhancement (LGE)
LGE MRI is a standard non-invasive method for assessing ischemic and nonischemic myopathic processes with high spatial resolution [63]. Notably, LGE can detect increase of extracellular space that represents fibrous scar tissue [59]. No enhanced area (0% LGE signal/volume) is expected for healthy tissues, while increases in LGE are expected with increased degree of scaring.
II.E.1 LGE Increase
Enhanced LGE volumes were reported in a few studies. In a study of patients with Hodgkin’s disease survival of at least 20 years, 29% of 31 patients demonstrated late enhancement 24 years following mediastinal RT [46]. Similarly, LGE was noted at 20 years in 25% of 80 HL and NHL survivors who were treated by mediastinal RT with/without chemotherapy [47]. In another study of patients with various chest tumors, 9 of 28 patients showed LGE (2.3 ml (0.2-6.1)) at 46.4 months post-chemoradiotherapy [59]. Over shorter follow-ups, the prevalence of LGE was increased at 1.5 years following chemoradiotherapy compared to baseline (78% vs 7%, p<0.01) among 24 esophageal cancer patients [54]. Also, in 11 esophageal patients, LGE was noted in the subepicardial and mid-wall portion of the myocardium at 3-5 months after neoadjuvant chemoradiotherapy [57].
A few studies showed dose-dependent LGE response. In 12 out of 24 esophageal patients, LGE was detected in 15.38% of AHA segments receiving 40 Gy and in 21.2% of AHA segments receiving 60 Gy at 23.5 months post-RT [64]. In another study, a progressive increase of LGE signal was noted at >30 Gy dose in 68% (13/19) of esophageal cancer patients at 6 months and 1.5 year following post-chemoradiotherapy [53]. Also, in two years follow-up of children who were undergoing chemotherapy with/without RT for various malignancies, increase of LV myocardial scaring (0.4±1.5%, p<0.05) was correlated with the dose to 20% of LV volume (11.9±4.0 Gy, p<0.05) [55].
Lastly, one study was done on evaluation of RT-induced LA chamber enhancement at 3.1 years among 7 patients with lymphoma and esophageal cancer who were treated with RT. They found that there is a linear relationship between the LA mean dose (25.9 Gy) and its scar volume (2.5 cm3) (p=0.03) and between the ratio of LA scar-to-wall volume and dose (P<0.01) at 3.1 years [65].
II.E.2 LGE unchanged
On the other hand, no signs of LGE were found in 10 patients with different thoracic malignancies during RT treatment [48], 21 lung cancer patients at 2- and 6-months post-3DCRT [50], 66 patients with breast cancer (at 13 months) who were treated by either epirubicin-based chemotherapy followed by RT or left-sided RT only [52], 20 breast cancer patients at 8.3 years after anthracycline-based chemotherapy and 3DCRT [45], 40 esophageal cancer patients at 67.6 months past neoadjuvant chemoradiotherapy [60], or a 70-year-old esophageal cancer patient who was examined at 8 years post-chemoradiotherapy [61].
II.E.3 LGE decrease
Decreases in LGE signal were noted in low dose regions at 0.5 year (-0.2% change in 0-10 Gy regions) and 1.5 year (-0.8%, -3.2%, -1.9%, -4.4% changes corresponding to 0-10 Gy, 10-20 Gy, 20-30 Gy, 30-40 Gy of radiation, respectively) post chemoradiotherapy in 19 patients with esophageal cancer [53]. A summary of LGE findings in RT-induced MCT analyses is reported in Table 5.
Table 5. LGE changes at RT-induced MCT studies using MRI
Study
|
Decrease
|
Increase
|
Constant
|
Dose dependent
|
Machann et al., [46]
|
|
26% and 3% of the whole cohort demonstrated ischemic and cardiomyopathic late enhancement, respectively at 24 years post-RT
|
-
|
-
|
Van Der Velde et al., [47]
|
|
In 25% of patients at 20 years post-RT.
|
-
|
-
|
Traber et al., [48]
|
|
-
|
Only 2 patients (out of 10) demonstrated LGE at the baseline with no other LGE findings during RT
|
-
|
Bergom et al., [45]
|
|
-
|
No LGE was observed at 8.3 years.
|
-
|
Takagi et al., [54]
|
|
LGE changes were noticed at 1.5 year follow up among 79% of participants
(7% vs 78%, p<0.01)
|
No significant prevalence of LGE at 0.5 year follow up
(p=0.16)
|
-
|
Tahir et al., [52]
|
-
|
-
|
No LGE at all timepoints for both treatments
|
-
|
Umezawa et al., [64]
|
|
Higher LGE percentage in 60 Gy dose line (21.21%), and 40 Gy dose line (15.38%) in 12 of 24 patients who demonstrated LGE increased in 23.5 months
|
0% LGE percentage at segments out of the radiation field
|
|
Umezawa et al., [53]
|
At 6 months: -0.2% signal decrease corresponding to 0-10 Gy At 1.5 year: -0.8%,
-3.2%, -1.9%, -4.4% signal decrease corresponding to 0-10 Gy, 10-20 Gy,
20-30 Gy, 30-40 Gy, respectively.
|
At 6 months: 0.4%, 1.1%, 5.7%, 35.7%, and 38.1% signal increase corresponding to 10-20 Gy, 20-30 Gy, 30-40 Gy, 40-50 Gy, and 50-60 Gy, respectively. At 1.5 year: 17.5%, and 20.1% signal increase corresponding to 40-50 Gy, and 50-60 Gy, respectively.
|
|
A progressive increase of signal was noted at >30 Gy at both time points (6 months and 1.5 year).
|
Huang et al., [65]
|
|
|
|
1) A linear relation between LA scar-enhanced volume and mean dose with an average LA scar volume of 2.5 cm3 and average mean dose of 25.9 Gy
(R2 = 0.8514, p=0.03) at 3.1 years. 2) linear relation between radiation received by the cardiac tissue and the ratio of (LA scar/LA wall) at 3.1 years.
|
Burke et al., [57]
|
|
LGE was noted at the subepicardial and mid-wall after 3-5 months following therapy
|
|
|
Mukai-Yatagai et al., [61]
|
|
|
No signs of LGE at 8 years
|
|
Ricco et al., [59]
|
|
LGE noted in 9 of 28 patients of 2.3 ml (0.2-6.1), all within their left myocardium or septum at 46.4 months.
|
|
-
|
de Groot et al., [60]
|
|
|
No significant difference between neoadjuvant chemoradiotherapy and control group at 67.6 months post-RT
|
|
Lideståhl et al., [50]
|
|
|
No visual signs of LGE was detected at 2- and 6-month post-RT
|
|
de Goyet et al., [55]
|
|
At 2-year follow-up (4.4±2%) compared to baseline (4.0±1.6%), p<0.05
|
Between baseline (4.0±1.6%) and 1-year follow-up (4.3±2)
|
Increase of LV myocardial scarring (0.4±1.5%, p<0.05) at 2-year follow-up was correlated with the radiation dose received by 20% of volume of the left ventricle (11.9±4.0 Gy; r=0.85; p<0.05).
|
II.E. Discussion on LGE
Global changes of LGE were mostly noted over longer follow-ups (>20 years) [47], particularly in patients treated by older RT treatment techniques (e.g., mediastinal RT with anterior mantle-field technique) [46], and/or those who received higher dose to the heart (>22.9±4 Gy) [60]. Over shorter follow-ups (3-5 months), only those with segmental analysis showed an elevated LGE signal after RT [54], [57], [64]. Dose-dependent response of LGE was also noted in a few studies. A linear relationship was found between LA enhanced LGE volume and average received dose (25.9 Gy) [65]. Similarly, a progressive increase of LGE signal was found at >30 Gy (threshold) [53] with higher signal intensity changes at shorter follow-ups (6 months vs 1.5 year), suggesting that RT-induced inflammatory response may be present for 6 months and diminish in 1.5 years [53]. Finally, enhanced volumes on LGE MRI were noted in younger patients. Children, as a high risk group, showed a dose-dependent increase of LV myocardial scarring which could be due to myocardial hypoperfusion or an early sign of radiation damage [55].
Several studies did not find any signs of LGE, potentially due to small sample size (e.g., 1 patient) [61], low radiation dose to the heart (e.g., 2±2 Gy) [52], global (rather than regional) analysis [45], and non-ideal timing of follow-up (i.e., too early to detect acute inflammation/edema or too late to detect the formation of scar tissue as a late radiation effect (2-6 months follow-up)) [50]. Nevertheless, dose-dependency of LGE and ECV findings indicate the potential of CMR to detect early changes/correlations between MCT (possibly myocardial fibrosis) and RT dose.
II.F. Strain (circumferential, radial, longitudinal)
Cardiac function is a combination of contraction, twist, and expansion of myocardium in multiple axes. Regional and global quantification of myocardial deformation (e.g., via measurement of strain) in any of these axes may serve as a potential metric to detect sub-clinical myocardial changes and assess myocardial function. Strain can quantify spatial components of contractile function over multiple directions (e.g., radial, circumferential, and longitudinal). With regards to the LV, MRI-derived strain measurements in healthy subjects have reported pooled mean values of global longitudinal strain (GLS), global circumferential strain (GCS), and global radial strain (GRS) as -18.6% (-19.5% to -17.6%), -21.0% (−22.4% to -19.6%), and 38.7% (30.5% to 46.9%), respectively [66]. Notably, a relative drop of greater than 15% of GLS is known to be clinically significant and could be an early indicator of CVT [27], [67].
II.F.1 Regional (decrease)
In 22 breast cancer patients who underwent RT, decrease in the magnitude of circumferential, radial, and longitudinal strains were noted in segmental regions 6 weeks post-treatment, with a negative correlation between radial strain and maximum dose in segments 6 and 14 (p<0.05) of the 16 segment AHA model, as well as a negative correlation between the magnitude of longitudinal strain and dose in segment 6 (p<0.01) [68]. (Note that longitudinal and circumferential strain are typically negative when using diastole as the reference configuration.)
II.F.2 Global (decrease/constant)
Global measurements of strain were reported in a few studies. In longer follow-ups, significant decreases in the magnitude of GLS, GRS, and GCS were reported in 80 HL and NHL survivors 20 years following mediastinal RT with/without chemotherapy [47]. Similarly, a drop in GLS magnitude (-14.6%) was shown after 8.3 years among 16 of 20 breast cancer patients who were treated with anthracycline-based chemotherapy and 3DCRT [45]. Over shorter follow-ups, GLS and GCS were reduced in magnitude in breast cancer patients with concurrent treatment (epirubicin-based chemotherapy followed by RT) at both therapy completion and 13±2 months post therapy (-18% to -17% for both strains and timepoints) [52]. On the other hand, 6-months follow up of 11 patients with various chest tumors did not show any significant strain changes [51]. Also, no relation was found between global strain and dose in any of the studies. Table 6 summarizes strain changes in RT-induced MCT studies using MRI.
Table 6. Strain changes in RT-induced MCT studies using MRI (No increase of strain was reported).
Study
|
Decrease
|
Constant
|
Dose-dependent
|
Van Der Velde et al., [47]
|
GLS (-19.5±2.5 vs -20.6±2, p=0.01)
GCS (-17.9±2.5 vs -20.4±2.2, p<0.001)
GRS (69±15 vs 76±15, p=0.02) at 20 years post-RT.
|
-
|
|
Vallabhaneni et al., [51]
|
|
GLS: No significant % changes in patients with higher radiation (-15.2±15.2%) or patients with minimal radiation (-6.8±3.1%)
GCS: No significant % changes in patients with higher radiation
(-8.4±14.5%) or patients with minimal radiation (-4.0±6.9%)
at 6-months following RT.
|
|
Bergom et al., [45]
|
Abnormal lower absolute strain values in 16/20 patients -14.6% (-17.8% to -11.1%) compared to normal range -22.1% to -15.9% at 8.3 years post-RT
|
|
|
Tahir et al., [52]
|
In epirubicin-chemotherapy-based followed by RT group:
GLS (baseline: -18±2%,
therapy completion: -17±2%, p=0.01;
13±2 months: -17±2% p=0.01)
GCS (baseline: -18±2%,
therapy completion: -17±3%, p=0.03;
13±2 months: -17±3% p=0.01)
|
In epirubicin-chemotherapy-based followed by RT group: GRS changes were not significant (baseline: 36±7%; post-therapy completion: 34±8%; 13±2 months post-therapy completion: 34±6%)
|
|
-
|
In left-sided RT group:
GLS (baseline: -18±2%, therapy completion: -18±2%, 13±2 months post-therapy completion: -18±1%)
GCS (baseline: -18±2%, therapy completion: -18±2%, 13±2 months post-therapy completion: -19±3%)
GRS (baseline: 39±6%, therapy completion: 39±9%, 13±2 months post-therapy completion: 39±7%)
|
|
Tang et al., [68]
|
At 6 weeks vs pre-treatment:
3D circumferential: segment 7 -16.39 vs -19.17, segment8 -18.73 vs -20.67, segment13 -15.95 vs -18.48.
2D circumferential: segment 1 -19.13 vs -23.57, segment 14 -25.01 vs -28.29.
3D radial: segment 8 25.46 vs 35.32, segment 9 17.54 vs 27.24.
3D longitudinal: segment 7 -19.12 vs -21.44
|
|
Negative correlation:
3D radial at segment 6 and max dose
(-0.443, p=0.05),
2D radial at segment 14 and max dose
(-0.543, p=0.01),
Positive correlation:
3D longitudinal strain at segment 6 and max dose (0.669, p<0.01).
|
II.F. Discussion on strain
Similar to other global metrics (e.g., LVEF), no correlation was found between GLS/GCS/GRS and dose [45], [47], [52]. However, segmental strain analysis showed a potential relationship between local strain reduction and radiation dose [68]. Significant reduction of strain (without dose-dependency) was noted in longer follow-up studies [45], [47]. It is unclear if this reduction is due solely to time or whether concurrent treatments (e.g., anthracycline and RT exposure) and/or other cardiac risk factors play a significant role. Longitudinal studies over longer and shorter follow-up times in patients with single and concurrent treatments are needed to fully examine this matter. Lastly, it has been shown that strain can detect myocardial changes earlier than other global metrics (e.g., LVEF) [69]. For example, compensatory features of the heart can preserve LVEF by increasing cardiac torsion following a drop in circumferential strain due to myocardial fiber dysfunction [27], [70]. Therefore, it may be beneficial to utilize strain measurements over LVEF alone if one aims to measure global cardiac function following RT treatment.