Impact of Radiation on Cardiovascular Outcomes in Older Resectable Esophageal Cancer Patients with Medicare

Abstract

Objectives:

Preoperative radiotherapy improves outcomes for operable esophageal cancer patients, though the proximity of the heart to the esophagus puts patients at risk of radiation-induced cardiovascular disease. This study characterizes the impact of radiotherapy and different radiation techniques on cardiovascular morbidity among a cohort of esophageal cancer patients.

Methods:

We identified 1,125 patients aged 65 and older diagnosed between 2000 and 2011 with esophageal cancer who received surgery alone, or surgery preceded by either preoperative chemotherapy or preoperative chemoradiation from the SEER-Medicare database. We used Medicare claims to identify severe perioperative and late cardiovascular events. Multivariable logistic regression and Fine-Gray models were used to determine the effect of pre-surgery treatment on the risk of perioperative and late cardiovascular disease.

Results:

Preoperative chemotherapy or chemoradiation did not significantly increase the risk of perioperative cardiovascular complications compared to surgery alone. Patients treated with preoperative chemoradiation had a 36% increased risk of having a late cardiovascular event compared to patients treated with surgery alone (subdistribution hazard ratio [SDHR] 1.36; p=0.035). There was no significant increase in late cardiovascular events among patients treated with preoperative chemotherapy (SDHR 1.18; p=0.40). Among patients treated with preoperative chemoradiation, those receiving intensity modulated radiotherapy (IMRT) had a 68% decreased risk of having a late cardiovascular event compared to patients receiving conventional radiation (SDHR 0.32; p=0.007).

Conclusions:

This study demonstrates an increased risk of cardiovascular complications among operative esophageal cancer patients treated with preoperative chemoradiation, though these risks might be reduced with more cardioprotective radiation techniques such as IMRT.

Introduction

Radiation therapy represents a central component in the treatment of operable esophageal cancer¹. Among patients with localized or local-regional disease, clinical trials demonstrate that the use of preoperative radiation delivered concurrently with chemotherapy followed by surgery confers a survival benefit compared with surgery alone^2,3. The CROSS study in particular found a median survival of 49 months for patients treated with preoperative chemoradiation followed by surgery compared to 24 months for patients treated with surgery alone³.

The location of the esophagus with respect to the heart puts the heart at risk of receiving incidental radiation during treatment for esophageal cancer. With other cancer types, including breast cancer, lung cancer, and Hodgkin’s lymphoma, large clinical studies demonstrate a clear link between thoracic radiation and adverse cardiovascular outcomes^4-6. Preclinical studies have suggested that radiation-mediated cardiovascular disease results from microvascular and macrovascular inflammation and damage, ultimately leading to myocyte injury and eventual fibrosis within the myocardium, pericardium, and vasculature⁷. In the largely avascular cardiac valves, radiation-mediated injury remains less well understood, but is thought to be related to elevated systemic pressure given that the left-sided valves are most commonly affected⁸. Multiple small institutional studies have demonstrated a potential link between radiation and cardiovascular disease in esophageal cancer patients treated with radiotherapy^9-13. Additionally, a study evaluating patients within the Surveillance Epidemiology and End Results (SEER) cancer registry found increased cardiovascular mortality among esophageal cancer patients receiving radiation¹⁴. While these older studies demonstrate increased risks associated with radiation, the delivery of radiation in esophageal cancer has evolved over the past several decades¹⁵. Modern conformal treatment techniques, such as intensity modulated radiotherapy (IMRT), offer the ability to decrease radiation doses to the heart¹⁶. The purpose of this large population-based study was to characterize adverse cardiovascular events associated with radiation among operable patients with esophageal cancer. Furthermore, this study will analyze whether IMRT mitigates risks of radiation-induced heart disease.

Materials and Methods

Data Source

We identified esophageal cancer patients from the SEER-Medicare linked database. The National Cancer Institute oversees the SEER program which collects data on incident cancers diagnosed across the United States covering approximately 28% of the US population. Medicare provides federally funded health insurance for people over the age of 65. The SEER-Medicare linkage provides Medicare claims data for Medicare beneficiaries within the SEER database. Medicare claims indirectly capture information about cancer treatments and long-term risks of adverse events associated with treatment. The Institutional Review Board at the University of California San Diego found this study to be exempt from review.

Study Population

Our initial query of the SEER-Medicare database identified 10,353 patients with histologically confirmed non-metastatic esophageal cancer diagnosed between 2000 and 2011. We included only patients with squamous cell carcinoma or adenocarcinoma, and excluded patients with atypical histologies (Supplemental Table 1) given the rarity and different management/prognosis of non-squamous cell carcinoma and non-adenocarcinoma histology¹⁷. To help maintain focus on adverse outcomes associated with esophageal cancer, we excluded subjects with a diagnosis of more than one cancer. Esophageal surgery portends additional cardiopulmonary risk on top of that already conferred by chemotherapy and radiation; therefore, patients who did not undergo surgery as part of treatment for their esophageal cancer were excluded from the study. To ensure data completeness, patients were required to have continuous Medicare claims data from 1 year prior to diagnosis (to calculate pre-existing comorbidity) through death or the end of follow-up. This required patients to have continuous Medicare Part A (inpatient/hospitalization services) and B (outpatient/medical services) coverage over this period. We excluded subjects with Part C coverage, which largely includes managed care organizations¹⁸, because managed care organizations do not consistently submit detailed claims data. Additional selection criteria are described below, and the final study cohort included 1,125 subjects (see Figure 1 for selection schema).

Figure 1.

Patient selection.

Study Covariates

Patient demographic and tumor characteristics including histology, stage, and year of diagnosis were extracted from SEER. We used the Deyo adaptation of the Charlson comorbidity index to calculate pre-existing comorbidity from inpatient and outpatient Medicare claims during the year prior to diagnosis¹⁹. This study separately incorporated pre-existing cardiovascular disease as a study variable (described below), therefore we modified the Charlson score by removing cardiovascular disease to avoid double counting. We identified chemotherapy, radiation, and esophageal surgery using previously established methods^20-22 that rely on Medicare claims and International Classification of Diseases 9^th edition (ICD-9) Procedure and Diagnosis codes, as well as Healthcare Common Procedure Coding System (HCPCS) codes (see Supplemental Table 2 for codes). We identified individual fractions of radiation from radiation treatment claims, and considered only definitive courses of radiation, defined as 20 or more fractions of radiation. We assumed that a break between fractions of radiation of more than 30 days indicated a separate (and subsequent) course of radiation. We identified patients who received IMRT through radiation planning and treatment claims specific to IMRT. We defined concurrent chemoradiation as any administration of chemotherapy during or within a 2 week window surrounding the course of radiation. We excluded the few patients who received brachytherapy, or proton therapy. Additionally, we excluded the small number of patients who received preoperative radiation alone without concurrent chemotherapy (insufficient numbers for analysis), as well as patients receiving postoperative radiotherapy (see Figure 1).

Cardiovascular events

The primary study endpoint included clinically relevant cardiovascular events potentially associated with radiation ascertained with ICD-9 and HCPCS codes (Supplemental Table 2). This composite endpoint included hospitalization associated with any of the following: acute myocardial infarction, percutaneous coronary intervention (PCI), coronary artery bypass grafting (CABG), valve surgery, pericardial disease, or congestive heart failure or cardiomyopathy. We did not include cardiac arrhythmias in the primary endpoint because of the significant incidence of these events following esophageal surgery^23-25. We classified cardiovascular endpoints as preexisting, perioperative, and late based on timing with respect to cancer diagnosis and surgery. Preexisting cardiovascular events occurred during the year prior to cancer diagnosis. Perioperative cardiovascular events occurred between the date of esophageal surgery and 30 days after discharge from the hospital. Late cardiovascular events occurred more than 30 days after discharge.

Statistical analysis

We characterized the study cohort with median follow-up time, calculated from the date of surgery through death or last follow-up, and the median and 5-year survival, estimated with Kaplan-Meier analyses. We categorized patients into three groups based on the type of preoperative treatment received: preoperative chemotherapy alone, preoperative chemoradiation, or no preoperative treatment. Among these three treatment groups, we assessed for differences in baseline patient and tumor characteristics with Fisher’s exact tests. We used multivariable models to determine whether the risks of perioperative and postoperative cardiovascular events varied by preoperative treatment. Multivariable models incorporated potential confounding factors including patient age, race, sex, marital status, year of diagnosis, tumor histology, stage (localized versus regional), Charlson comorbidity score, preexisting cardiovascular disease, and the receipt of adjuvant chemotherapy. The impact of preoperative treatment on perioperative cardiovascular events was assessed with a multivariable logistic regression. The impact of preoperative treatment on late cardiovascular events was assessed with a multivariable Fine-Gray competing risk regression model^26-30, considering death as a competing event and censoring at last follow-up. Esophageal cancer has a high mortality³¹, and the risk of death differs by treatment³. Using a Fine-Gray regression approach has the advantage of accounting for these competing events³². To determine whether the risk of late cardiovascular events varied by patient characteristics, we performed an additional analysis introducing interaction terms with treatment and select variables into the multivariable model. A significant interaction would indicate that the impact of treatment on the risk of cardiovascular disease varied by that patient characteristic. The 5-year incidence of late cardiovascular events was assessed with cumulative incidence analyses. Due to the small number of cardiac deaths (N=58), we did not assess cardiovascular mortality in this analysis. Statistical analyses were conducted with SAS version 9.4 (SAS Institute Inc., Cary, NC) with p-values < 0.05 considered significant.

Results

Among the 1,125 esophageal cancer patients in this study the median follow-up time was 1.5 years (interquartile range [IQR]: 0.5-4.2 years), and for the patients alive at the end of the study (N=355; 32%), the median follow-up time was 4.7 years (IQR: 2.3-7.8 years). The median overall survival for the entire cohort measured from the date of surgery was 1.9 years, and the 5-year overall survival was 33%. When considering preoperative treatment, we found that 613 patients (54%) received surgery alone, 137 (12%) received preoperative chemotherapy, and 375 (33%) received preoperative chemoradiation. Table 1 demonstrates patient characteristics stratified by treatment group. Compared to patients receiving surgery alone, patients receiving chemoradiation were younger, more likely married, diagnosed at a later time period, were more likely to have regional disease on presentation, and were more likely to have received chemotherapy after surgery.

Table 1.

Patient characteristics.

		Treatment group – N (%)
Characteristic	N	Surgery alone	Preoperative chemotherapy	Preoperative chemoradiation	p-value
Total	1,125	613	137	375
Age					<0.001
65-74	750	360 (58.7)	92 (67.2)	298 (79.5)
75+	375	253 (41.3)	45 (32.9)	77 (20.5)
Race					0.46
White	1,034	562 (91.7)	123 (89.8)	349 (93.1)
Other	91	51 (8.3)	14 (10.2)	26 (6.9)
Marital status					<0.001
Married	773	392 (64)	98 (71.5)	283 (75.5)
Other	352	221 (36.1)	39 (28.5)	92 (24.5)
Sex					0.37
Male	868	463 (75.5)	108 (78.8)	297 (79.2)
Female	257	150 (24.5)	29 (21.2)	78 (20.8)
Year of diagnosis					<0.001
2000-2003	365	225 (36.7)	50 (36.5)	90 (24)
2004-2007	402	216 (35.2)	46 (33.6)	140 (37.3)
2008-2011	358	172 (28.1)	41 (29.9)	145 (38.7)
Histology					0.99
Squamous cell carcinoma	283	153 (25)	35 (25.6)	95 (25.3)
Adenocarcinoma	842	460 (75)	102 (74.5)	280 (74.7)
Stage					<0.001
Localized	482	325 (53)	50 (36.5)	107 (28.5)
Regional	643	288 (47)	87 (63.5)	268 (71.5)
Charlson comorbidity index					0.98
0	680	370 (60.4)	84 (61.3)	226 (60.3)
1+	445	243 (39.6)	53 (38.7)	149 (39.7)
Pre-existing cardiovascular disease					0.24
No	1,070	577 (94.1)	>126 (92.0)	362 (96.5)
Yes	55	36 (5.9)	<11 (9.0)	13 (3.5)
Chemotherapy after surgery					<0.001
No	891	545 (88.9)	91 (66.4)	255 (68)
Yes	234	68 (11.1)	46 (33.6)	120 (32)

In the perioperative period, 302 patients (27%) experienced a cardiovascular event, with the majority of these classified as congestive heart failure (N=259; 23%) or acute myocardial infarction (N=61; 5.4%) (Table 2). On multivariable analysis, we found that the risk of perioperative cardiovascular events did not correlate with the type of preoperative treatment. Compared to the surgery alone group, there was no increased risk of developing a perioperative cardiovascular event among those receiving preoperative chemotherapy (odds ratio [OR] 0.83, 95% confidence interval [CI] 0.54-1.27; p=0.50), or among those receiving preoperative chemoradiation (OR 0.92, 95% CI 0.67-1.26; p=0.97). Table 3 demonstrates the full results of the multivariable model.

Table 2.

Perioperative and late cardiovascular events.

	5-year cumulative incidence (95% confidence interval)
Late Cardiovascular Endpoint	All patients	Surgery alone	Preoperative chemotherapy	Preoperative chemoradiotherapy
Any cardiovascular event	25.7 (22.8-28.7)	23.4 (19.7-27.4)	23.9 (16.5-32.0)	30.0 (24.8-35.4)
Acute myocardial infarction	4.9 (3.6-6.6)	4.9 (3.1-7.2)	3.3 (1.1-7.7)	5.8 (3.4-9.1)
PCI or CABG	3.3 (2.2-4.7)	3.3 (1.9-5.2)	1.9 (0.4-6.1)	4.0 (2.0-7.0)
CHF or cardiomyopathy	21.4 (18.7-24.2)	19.5 (16.0-23.3)	21.4 (14.4-29.4)	24.4 (19.6-29.6)
Pericardial disease	5.5 (4.1-7.1)	3.4 (2.0-5.4)	4.9 (2.0-9.9)	9.2 (6.2-12.9)
Heart valve procedure	0.3 (0.1-1.0)	0.4 (0.1-1.5)	0	0.3 (0.03-1.7)
	Number of events (%)
Perioperative Endpoint	All patients	Surgery alone	Preoperative chemotherapy	Preoperative chemoradiotherapy
Any cardiovascular event	302 (26.8)	167 (27.2)	40 (29.2)	95 (25.3)

Abbreviations: CABG = Coronary artery bypass graft; PCI = percutaneous coronary intervention; CHF = congestive heart failure.

Table 3.

Multivariable logistic and Fine-Gray competing-risks regression analysis of perioperative and late cardiovascular events, respectively.

	Perioperative Cardiovascular Event		Late Cardiovascular Event
Characteristic	Odds Ratio (95% CI)	P value	Subdistribution Hazard Ratio (95% CI)	P value
Preoperative Treatment
None (Surgery Only)	1 [Reference]	-	1 [Reference]	-
Chemotherapy	0.83 (0.54 – 1.27)	0.50	1.18 (0.80 – 1.74)	0.40
Chemoradiotherapy	0.92 (0.67 – 1.26)	0.97	1.36 (1.02 – 1.80	0.035
Age Group
65-74	1 [Reference]	-	1 [Reference]	-
≥75	0.61 (0.46 – 0.82)	<0.001	1.05 (0.80 – 1.37)	0.73
Race
White	1 [Reference]	-	1 [Reference]	-
Other	0.86 (0.52 – 1.42)	0.55	0.93 (0.56 – 1.55)	0.78
Marital Status
Married	1 [Reference]	-	1 [Reference]	-
Other	0.77 (0.57 – 1.05)	0.10	1.15 (0.86 – 1.54)	0.34
Sex
Male	1 [Reference]	-	1 [Reference]	-
Female	1.21 (0.85 – 1.74)	0.29	0.84 (0.58 – 1.22)	0.36
Year of Diagnosis
2000-2003	1 [Reference]	-	1 [Reference]	-
2004-2007	1.05 (0.76 – 1.46)	0.65	1.16 (0.88 – 1.54)	0.30
2008-2011	1.26 (0.89 – 1.78)	0.18	0.84 (0.60 – 1.17)	0.29
Histology
Squamous Cell Carcinoma	1 [Reference]	-	1 [Reference]	-
Adenocarcinoma	1.35 (0.96 – 1.91)	0.09	1.03 (0.73 – 1.45)	0.87
Stage
Localized	1 [Reference]	-	1 [Reference]	-
Regional	1.00 (0.75 – 1.33)	0.99	0.96 (0.74 – 1.25)	0.79
Charlson Comorbidity Index
0	1 [Reference]	-	1 [Reference]	-
≥1	0.54 (0.41 – 0.72)	<0.001	1.30 (1.01 – 1.68)	0.04
Pre-existing Cardiovascular Disease
No	1 [Reference]	-	1 [Reference]	-
Yes	0.33 (0.18 – 0.58)	<0.001	1.45 (0.85 – 2.49)	0.26
Postoperative Chemotherapy
No	-	-	1 [Reference]	-
Yes	-	-	0.84 (0.61 – 1.14)	0.26

A total of 259 patients experienced a late cardiovascular event, including 128 surgery alone patients, 34 preoperative chemotherapy patients, and 97 preoperative chemoradiotherapy patients. Of the 55 patients with preexisting cardiovascular disease, 37 (67.3%) developed a perioperative and/or late cardiovascular event. The median time to development from surgery of any late cardiovascular event for the entire cohort was 0.8 years (IQR: 0.2-2.7 years). The median time to development of specific late cardiovascular events from surgery for the entire cohort is shown in Supplemental Table 3. The 5-year cumulative incidence of late cardiovascular events among all patients was 25.7%; among surgery alone patients was 23.4%; among preoperative chemotherapy patients was 23.9%; and among preoperative chemoradiotherapy patients was 30.0%, with most of the events consisting of congestive heart failure, pericardial disease, or myocardial infarction (Figure 2, Table 2). The five-year cumulative incidence of late cardiovascular events was significantly higher in the preoperative chemoradiotherapy group compared with the surgery alone group (p=0.02). On multivariable Fine-Gray analysis, patients treated with preoperative chemoradiation had a 36% increased risk of developing a late cardiovascular event compared to those undergoing surgery alone (subdistribution hazard ratio [SDHR] 1.36, 95% CI 1.02-1.80; p=0.035). We did not find a significant increase in late cardiovascular events risk among those receiving preoperative chemotherapy (SDHR 1.18, 95% CI 0.80-1.75; p=0.40) compared to those undergoing surgery alone. Additionally, the receipt of postoperative chemotherapy was not significantly associated with the development of a late cardiovascular event (SDHR 0.84, CI 0.61-1.14, p=0.26). Table 3 demonstrates the results of the full multivariable model.

Figure 2.

Cumulative incidences of late cardiovascular events between treatment groups. Gray’s p-values for comparisons of cumulative incidences between groups: surgery alone versus preoperative chemotherapy, p=0.64, surgery alone versus preoperative chemoradiation, p=0.02; preoperative chemotherapy versus preoperative chemoradiation, p=0.33. ***Exact number cannot be reported per the SEER-Medicare data use agreement.

The impact of preoperative chemoradiotherapy on late cardiovascular outcomes did not differ according to patient age, race, sex, marital status, year of diagnosis, tumor histology, stage (localized versus regional), Charlson comorbidity score, preexisting cardiovascular disease, or receipt of postoperative chemotherapy (all interaction p-values >0.05).

Among the cohort of patients receiving chemoradiotherapy, 71 (19%) received IMRT. Compared to patients who received conventional radiation therapy, those who received IMRT were significantly more likely to be female (29.6% versus 18.8%, p=0.04) and receive chemotherapy after surgery (43.7 versus 29.3%, p=0.02). Additionally, patients diagnosed later in the study period were significantly more likely to receive IMRT compared to conventional radiation (p<0.001). The receipt of IMRT was not significantly associated with any other cofactors, including race, age, and preexisting cardiovascular disease. The 5-year cumulative incidence of late cardiovascular events was significantly lower (p=0.008) for patients receiving IMRT (12.1% [95% CI 5.1-22.2%]) compared to patients receiving conventional radiation (33.2% [95% CI 27.3-39.2%]) (Figure 3). On multivariable analysis, patients receiving IMRT had a 68% decreased risk of having a late cardiovascular event compared to patients receiving conventional radiation (SDHR 0.32, 95% CI 0.14-0.73; p=0.007).

Figure 3.

Impact of radiation modality on postoperative cardiovascular events. The cumulative incidence of late cardiovascular events was significantly lower in patients who received IMRT compared with those who received conventional radiation (p=0.008).

Discussion

This study demonstrates a link between the receipt of radiation and long-term postoperative cardiovascular outcomes among operable esophageal cancer patients receiving curative intent treatment. Our key finding of a 36% increased relative risk of adverse, clinically relevant cardiovascular events supports the existing literature evaluating the risk of radiation-induced cardiotoxicity in esophageal cancer. A series of single institution studies with sample sizes ranging from 15 to 123 have described a similar association between radiation in esophageal cancer and a variety of cardiovascular endpoints. This includes studies identifying a link between radiation and increased myocardial perfusion defects⁹, decreased ejection fraction^10,11, and clinically significant cardiac toxicity^12,13. Similarly, a study evaluating patients within SEER found a 2.8% absolute increased risk of death from cardiovascular disease among esophageal cancer patients treated with radiation¹⁴. Although our study, in conjunction with existing research, highlights the potential for increased cardiovascular morbidity and mortality, clinical trial data support a cancer-specific survival advantage associated with preoperative radiation in esophageal cancer^3,33. Therefore, esophageal cancer patients are still likely to benefit from preoperative radiation despite the increased risk of cardiovascular morbidity and mortality.

The timing of cardiovascular events after radiation deserves further discussion. Imaging research demonstrates that myocardial perfusion defects can occur as early as 6 months after treatment^34-36. Furthermore, research in breast cancer demonstrates that the long-term cardiovascular effects of radiation stretch over a patient’s lifetime^4,37. The probability of long-term survival with esophageal cancer is lower than breast cancer, though given the trend towards improved survival over time with esophageal cancer³⁸, one could argue that the clinical importance of radiation-induced cardiovascular disease among esophageal cancer survivors will increase in the future.

Another key finding of this study relates to the observed reduction in late cardiovascular events among patients receiving IMRT. Research evaluating radiation dosimetry demonstrates that the more conformal radiation dose distributions with IMRT have the potential to reduce radiation doses to the heart compared to conventional forms of radiation^16,39-42. Similarly, other retrospective analyses have found potential reductions in cardiovascular toxicity associated with IMRT compared to older radiation techniques^43,44. Furthermore, a cancer registry study found that IMRT was associated with a decreased risk of cardiac mortality compared to conventional radiation⁴⁵. The potential benefits of more conformal radiation modalities raise the question about the possible utility of proton therapy in esophageal cancer, which can further reduce radiation doses to the heart⁴⁶. A recent early report of a phase II randomized trial comparing proton therapy to IMRT demonstrated reduced toxicity among the proton therapy patients, though did not specifically report on cardiovascular toxicity⁴⁷. An ongoing phase III randomized trial with a co-primary endpoint of overall survival and adverse cardiopulmonary events will help define the potential benefit of proton therapy in esophageal cancer (NCT03801876).

An important consideration of this study relates to the potential contribution of chemotherapy to the risks of adverse cardiovascular events. Combined chemotherapy and radiation represents the standard approach in the definitive management of non-metastatic esophageal cancer¹. The typical chemotherapy doublets used concurrently with radiation in esophageal cancer include 5-fluorouracil and cisplatin, or carboplatin and paclitaxel¹. 5-fluorouracil can cause coronary spasms or ischemia, and paclitaxel can cause bradycardia⁴⁸. Furthermore, all of these chemotherapy agents have the capacity for radiosensitization⁴⁹, which could influence the adverse impact of radiation on the heart. Whether different concurrent chemotherapy agents delivered with radiation lead to different risks of cardiotoxicity remains an important but unanswered question.

This study has limitations worth mentioning. The SEER-Medicare linked database lacks specific details of radiation including the radiation dose and target, both of which could impact the amount and anatomic distribution of radiation received by the heart, which could influence the patient-specific risks of cardiovascular disease. Along these lines, we cannot assess the dose to adjacent organs, including the lungs, therefore cannot thoroughly assess the interplay between radiation to the lungs and heart. Given that radiation for esophageal cancer can lead to both cardiac and pulmonary toxicity⁵⁰, this interplay between radiation to the lungs and heart deserves further evaluation. Tumor location within the esophagus (i.e., upper thoracic versus lower esophageal) likely impacts the level of cardiac radiation exposure and the subsequent risk of a postoperative cardiac event; however, this information is not available through the SEER-Medicare database and thus could not be incorporated in our analysis. Furthermore, this study is subject to additional limitations inherent to using claims data. Patients were required to have one year of continuous Medicare Part A and B coverage prior to diagnosis. While this is a standard approach when working with claims data^51-53, this may have failed to capture preexisting cardiovascular comorbidities occurring prior to this time period. Moreover, the Medicare coverage inclusion requirement in our study was necessary to reduce the risk of missing data, though this eligibility requirement may have excluded individuals within lower socioeconomic groups from our analysis^54-56. The SEER-Medicare database also lacks information on important cardiovascular risk factors, including smoking, obesity, exercise, and family history, any of which could mitigate or moderate the risk of cardiovascular morbidity and mortality. This study included Medicare beneficiaries over the age of 65, therefore the findings in this study may not generalize to a younger population. Finally, with the increased knowledge and awareness surrounding radiation-induced heart disease, one could hypothesize that radiation oncologists treating patients today might prioritize cardiac sparing in their radiation plans, which could decrease the risks of radiation-induced heart disease for patients treated in current practice. However, prioritizing cardiac sparing often comes at the expense of increased lung dose, and practicing radiation oncologists must carefully balance the risk of cardiac toxicity with that of pulmonary toxicity.

Despite these limitations, this population-based study demonstrates that preoperative chemoradiation increases the risk of cardiovascular complications among esophageal cancer patients treated with curative intent. Furthermore, this study demonstrates the capacity of more advanced radiation techniques such as IMRT to reduce these risks, which provides a rationale to use cardioprotective radiation techniques. While additional research incorporating radiation dose is needed to quantify patient-specific risks, the findings of this study can help patients and radiation oncologists better understand the risks of radiation in esophageal cancer.