Sage Journals: Discover world-class research

Abstract

Objective

This study aims to evaluate the impact of economic status on the cause-specific survival (CSS) of patients with esophageal adenocarcinoma (EAC).

Methods

Using propensity score matching (PSM), we analyzed data from 4676 patients with EAC, categorized into low household income (LHI) and high household income (HHI), sourced from the Surveillance, Epidemiology, and End Results database. Significant pre-PSM differences in age, gender, race, and tumor stage (T stage) were observed, necessitating PSM adjustment. The PSM incorporated these variables along with radiotherapy to balance the groups effectively. Kaplan–Meier curves and accelerated failure time (AFT) and random survival forest (RSF) models were used to assess the prognostic impacts.

Results

According to the pre-PSM data, a significant difference in CSS was observed between the LHI and HHI groups (median CSS = 1.25 years vs 1.50 years, P < .05). Post-PSM analysis maintained the result (P < .05). Using the AFT model, HHI was found to significantly enhance CSS (pre-PSM time ratio [TR] = 1.145, post-PSM TR = 1.146). Male gender (pre-PSM TR = 0.762, post-PSM TR = 0.819) and certain non-single marital statuses (such as divorced, pre- and post-TRs <1.000) are also prognostic risk factors for EAC (P values <.05). Additional significant prognostic risk factors for EAC included advanced American Joint Committee on Cancer stages (III and IV), higher tumor stages (T3 and T4), higher node stages (N1–N3), without surgery, and without chemotherapy, as identified using the AFT and/or RSF mothods. The results derived from the RSF model are consistent with those of the AFT model.

Conclusion

Higher economic status is positively related to EAC CSS. Beyond standard clinical treatments, interventions such as enhancing socioeconomic support and preventive aspects, including intensive health education and EAC screening, could play a crucial role in improving EAC prognosis.

Keywords

esophageal adenocarcinoma economic status cause-specific survival propensity score matching income

Introduction

As one of the most common cancers, esophageal cancer (EC) ranks 10th in incidence among all tumors worldwide.¹ The main treatments for this cancer include surgical resection, radiotherapy, and chemotherapy.^2-4 Despite current medical advancements, the prognosis for EC patients remains poor, with a 5-year survival rate of only 20%, slightly higher than that of pancreatic cancer (which has the lowest survival rate).⁵ Thus, EC is one of the most lethal cancers. Additionally, the vast majority of patients with EC are diagnosed at an advanced stage,⁶ further limiting the effectiveness of treatment. Esophageal adenocarcinoma (EAC), one of the two main subtypes of EC (the other being esophageal squamous cell carcinoma), has seen increasing incidence over the past decades, particularly in developed countries such as the United States.^7,8 Despite advancements in treatment, the prognosis for patients with EAC remains suboptimal, and the factors influencing this prognosis are not yet fully understood. Therefore, this study aims to identify the key prognostic factors affecting patients with EAC, which could inform the development of more targeted and effective treatment and management strategies for EC.

In addition to the clinical and pathological characteristics of cancer, sociological factors affecting cancer prognosis are receiving increasing research attention.^9,10 Among these factors, economic status is widely believed to have a significant impact on the prognosis of various cancers. As shown by many studies, the overall survival rate of low-income patients is significantly lower than that of high-income patients across various cancers due to potential differences between low-income and high-income patients’ access to diagnostic and treatment resources.¹¹ For example, a study of hepatocellular carcinoma revealed that patients from rural areas and low-income families were diagnosed more advanced tumor stages (T stages) at diagnosis and had significantly higher mortality rates.¹² Similarly, in patients with bone and soft tissue sarcomas, low income was significantly associated with lower 5-year overall survival rates.¹³ This trend is also evident in EAC, where previous studies have also shown that lower median household income is associated with a worse prognosis in early-stage (T1aN0M0) patients with EAC.¹⁴ However, early-stage patients represent only a small proportion of the EAC patient population,⁶ and the incidence of early-onset EC is declining.¹⁵ Therefore, it is crucial to investigate how socioeconomic factors affect the prognosis of the broader EAC population, rather than those at certain stages, to inform disease-specific interventions and aspects of prevention.

This study aims to evaluate the impact of economic status on the survival of patients with EAC in the United States using a large-scale retrospective cohort study based on the SEER (Surveillance, Epidemiology, and End Results) database,¹⁶ which may provide valuable references for the management of EAC.

Materials and Methods

Study Design and Ethical Approval

This study uses retrospective data from the SEER program¹⁶ and aims to evaluate the impact of economic status on the cause-specific survival (CSS) of patients with EAC in the United States. We use the propensity score matching (PSM)¹⁷ method to rigorously control for confounding factors, ensuring comparability between the high household income (HHI) and low household income (LHI) groups, thereby more accurately assessing the impact of economic status on the prognosis of patients with EAC. The data used in this study have been anonymized in compliance with SEER data usage requirements¹⁶; thus, ethical committee approval is not required. This study complies with the RECORD guidelines.¹⁸

Study Subjects and Data Collection Criteria

Data for this study were sourced from the SEER*Stat 8.4.3.¹⁹ The inclusion criteria were as follows: patients pathologically diagnosed with EAC between 2000 and 2020; aged 18 to 89 years; with clear information on gender, age, race, marital status, and American Joint Committee on Cancer (AJCC) staging; recorded information on surgery, radiotherapy, and chemotherapy; and clear CSS information and follow-up time. The exclusion criteria were as follows: CSS time <1 month (i.e., the interval from the diagnosis of EAC to death due to the disease is less than 1 month) and the presence of multiple lesions. The detailed process of patient selection is shown in Figure 1.

Figure 1.

Detailed process of patient selection in this study.

Variables and Endpoint

In this study, all independent variables were recorded as categorical variables. The variables collected in this study were demographic characteristics (economic status, gender, age, race, marital status), clinical characteristics (AJCC, tumor (T)-node (N)-metastasis (M) staging [T, N, M staging] information), treatment information (surgery, radiotherapy, chemotherapy, the time from diagnosis to first treatment), and follow-up information (follow-up time and CSS).

Specifically, economic status is defined as the median household income of the patient’s county in the past 12 months (adjusted to 2021 inflation dollars). Patients in the LHI group have a median household income of less than $65,000, while those in the HHI group have a median household income of at least $65,000. Gender, age, race, and marital status refer to the attributes of patients at the time of diagnosis. Race is categorized into four groups: White, Black, American Indian/Alaska Native, and Asian or Pacific Islander. Since White individuals constitute the majority of study subjects, race in this study is grouped as White and Others (American Indian/Alaska Native, Asian or Pacific Islander). Marital status includes the following categories: Single (never married), Widowed, Divorced, Unmarried or Domestic Partner, Married (including common law), and Separated.

For the AJCC stage and its T, N, M staging information, the 7th edition was used in this study. Information on whether the patient received the three treatment modalities—surgery, radiotherapy, and chemotherapy—was also collected for analysis. The primary endpoint of the study was CSS, defined as the interval from the diagnosis of EAC to death due to the disease. Follow-up time was converted into years in this study.

PSM Parameters

To reduce the sample differences between the LHI and HHI groups, this study used 1:1 matching and the “optimal” matching method. Matching variables included parameters that were unbalanced between the HHI and LHI groups before PSM, such as age. The matching effectiveness was evaluated using chi-square tests between the two groups, Love plots, and standardized mean differences (SMDs). An SMD less than 0.1 indicated good matching.

Statistical Analysis

All data analyses were performed using R 4.2.2 statistical software and its packages tableone, MatchIt, survival, survminer, randomForestSRC, and cobalt.^20-22 Chi-square tests were used to detect the balance between the LHI and HHI groups and to test for collinearity among different variables. Kaplan–Meier survival analysis was used to evaluate CSS differences between the LHI and HHI groups. accelerated failure time (AFT) and random survival forest (RSF) models were employed to explore factors affecting CSS in patients with EAC. The statistical significance level was set at P < .05.

Results

Baseline Characteristics and PSM

This study included 2082 patients with EAC in the LHI group and 2594 in the HHI group (Table 1). Before PSM, significant differences were observed in multiple baseline characteristics between the LHI and HHI groups. Age distribution showed that 54.4% of patients in the LHI group were under 65 years old, compared to 47.1% in the HHI group (P < .001, Table 1). In terms of gender distribution, 12.3% of the HHI group were female, higher than the 10.4% in the LHI group (P = .049, Table 1). Regarding racial distribution, 95.5% of patients in the LHI group were White, compared to 93.8% in the HHI group (P = .014, Table 1). Additionally, there were significant differences in the T stage between the two groups (P < .001, Table 1).

Table 1.

The Characteristics of Patients Between the Low Household Income (LHI) and High Household Income (HHI) Groups Based on Data Before Propensity Score Matching.

Characteristics		LHI Group (N = 2082)	HHI Group (N = 2594)	P Value
Age group (years)	<65	1133 (54.4%)	1223 (47.1%)	<.001
	≥65	949 (45.6%)	1371 (52.9%)
Gender	Female	216 (10.4%)	318 (12.3%)	.049
	Male	1866 (89.6%)	2276 (87.7%)
Race	Others	94 (4.5%)	161 (6.2%)	.014
	White	1988 (95.5%)	2433 (93.8%)
Marital status	Single (never married)	305 (14.6%)	422 (16.3%)	.639
	Widowed	141 (6.8%)	185 (7.1%)
	Divorced	244 (11.7%)	289 (11.1%)
	Unmarried or domestic partner	7 (0.3%)	9 (0.3%)
	Married (including common law)	1358 (65.2%)	1662 (64.1%)
	Separated	27 (1.3%)	27 (1%)
AJCC stage	I	218 (10.5%)	245 (9.4%)	.246
	II	469 (22.5%)	541 (20.9%)
	III	772 (37.1%)	1015 (39.1%)
	IV	623 (29.9%)	793 (30.6%)
Tumor stage	T1	474 (22.8%)	499 (19.2%)	<.001
	T2	281 (13.5%)	361 (13.9%)
	T3	1038 (49.9%)	1439 (55.5%)
	T4	289 (13.9%)	295 (11.4%)
Node stage	N0	671 (32.2%)	772 (29.8%)	.174
	N1	1010 (48.5%)	1312 (50.6%)
	N2	287 (13.8%)	384 (14.8%)
	N3	114 (5.5%)	126 (4.9%)
Metastasis stage	M0	1459 (70.1%)	1801 (69.4%)	.655
	M1	623 (29.9%)	793 (30.6%)
Surgery	No	1370 (65.8%)	1689 (65.1%)	.644
	Yes	712 (34.2%)	905 (34.9%)
Radiotherapy	No	319 (15.3%)	444 (17.1%)	.107
	Yes	1763 (84.7%)	2150 (82.9%)
Chemotherapy	No or unknown	168 (8.1%)	189 (7.3%)	.344
	Yes	1914 (91.9%)	2405 (92.7%)
Diagnosis to treat	<1 month	363 (17.4%)	456 (17.6%)	.928
	≥1 month	1719 (82.6%)	2138 (82.4%)

Notes: AJCC, American Joint Committee on Cancer.

After performing PSM for age, gender, race, and T stage, differences in radiotherapy between the LHI and HHI groups were found (P < .05, data not shown). Therefore, age, gender, race, T stage, and radiotherapy were ultimately included as variables in the PSM process. The post-PSM data indicate that the differences in baseline characteristics between the LHI and HHI groups were significantly reduced. After matching, there were no significant differences between the two groups in terms of age, gender, race, marital status, AJCC stage, T stage, N stage, and M stage (P > .05, Table 2). Moreover, all SMD values of the five factors (age, gender, race, T stage, and radiotherapy) were less than 0.1 (Supplemental material 1). These results suggest that PSM effectively balanced the baseline characteristics between the LHI and HHI groups, providing a more balanced sample for subsequent analysis. In the matched data, there is a significant collinear association between the AJCC stage and the M stage (P < .05). Consequently, the M stage variable will be excluded from subsequent analyses.

Table 2.

The Characteristics of Patients Based on Data After Propensity Score Matching.

Characteristics		Low Group (N = 2082)	High Group (N = 2082)	P Value
Age group (years)	<65	1133 (54.4%)	1106 (53.1%)	.419
	≥65	949 (45.6%)	976 (46.9%)
Gender	Female	216 (10.4%)	201 (9.7%)	.470
	Male	1866 (89.6%)	1881 (90.3%)
Race	Others	94 (4.5%)	89 (4.3%)	.762
	White	1988 (95.5%)	1993 (95.7%)
Marital status	Single (never married)	305 (14.6%)	355 (17.1%)	.209
	Widowed	141 (6.8%)	134 (6.4%)
	Divorced	244 (11.7%)	225 (10.8%)
	Unmarried or domestic partner	7 (0.3%)	8 (0.4%)
	Married (including common law)	1358 (65.2%)	1343 (64.5%)
	Separated	27 (1.3%)	17 (0.8%)
AJCC stage	I	218 (10.5%)	216 (10.4%)	.424
	II	469 (22.5%)	428 (20.6%)
	III	772 (37.1%)	782 (37.6%)
	IV	623 (29.9%)	656 (31.5%)
Tumor stage	T1	474 (22.8%)	469 (22.5%)	.956
	T2	281 (13.5%)	290 (13.9%)
	T3	1038 (49.9%)	1043 (50.1%)
	T4	289 (13.9%)	280 (13.4%)
Node stage	N0	671 (32.2%)	645 (31%)	.791
	N1	1010 (48.5%)	1039 (49.9%)
	N2	287 (13.8%)	289 (13.9%)
	N3	114 (5.5%)	109 (5.2%)
Metastasis stage	M0	1459 (70.1%)	1426 (68.5%)	.282
	M1	623 (29.9%)	656 (31.5%)
Surgery	No	1370 (65.8%)	1349 (64.8%)	.515
	Yes	712 (34.2%)	733 (35.2%)
Radiotherapy	No	319 (15.3%)	335 (16.1%)	.523
	Yes	1763 (84.7%)	1747 (83.9%)
Chemotherapy	No or unknown	168 (8.1%)	154 (7.4%)	.451
	Yes	1914 (91.9%)	1928 (92.6%)
Diagnosis to treat	<1 month	363 (17.4%)	369 (17.7%)	.839
	≥1 month	1719 (82.6%)	1713 (82.3%)

Notes: AJCC, American Joint Committee on Cancer.

Prognostic Differences Between the LHI and HHI Groups

To explore the prognostic differences between the LHI and HHI groups, Kaplan–Meier curves were plotted based on pre- and post-PSM data. According to the pre-PSM data, over a follow-up period exceeding 10 years, the LHI group exhibited shorter CSS at nearly every stage than the HHI group; moreover, a statistically significant difference in CSS was observed between the LHI (median CSS = 1.25 years) and HHI (median CSS = 1.50 years) groups (P < .05, Figure 2A). Similarly, post-PSM data showed that the LHI group had shorter CSS than the HHI group (median CSS: 1.25 years vs 1.50 years) (P < .05, Figure 2B). These results indicate that improving the economic conditions of patients with EAC may help enhance prognosis and increase survival rates.

Figure 2.

Kaplan–Meier curves before (A) and after (B) propensity score matching (PSM). CSS, cause-specific survival.

Factors Affecting the Prognosis of Patients with EAC

The study further calculated time ratios (TRs) based on the AFT model. Analysis of pre- and post-PSM data revealed that various social and demographic factors significantly impact the prognosis of patients with EAC. Household income level was a significant prognostic factor both before and after matching; patients in the HHI group had significantly longer survival times than those in the LHI group (pre-PSM: TR = 1.145, 95% CI: 1.070-1.227; post-PSM: TR = 1.146, 95% CI: 1.066-1.231; P values <.05, Table 3). Gender also emerged as an important factor, with male patients having significantly worse prognoses than female patients (pre-PSM: TR = 0.762, 95% CI: 0.682-0.852; post-PSM: TR = 0.819, 95% CI: 0.724-0.928; P values <.05, Table 3). In terms of marital status, divorced patients had poorer prognoses than single ones, with TRs below 1 before and after matching (P values <.05, Table 3).

Table 3.

Factors Affecting the Cause-Specific Survival (CSS) of Patients Based on Data After Propensity Score Matching (PSM).

Characteristics		Pre-PSM Time Ratio (95%CI)	Post-PSM Time Ratio (95%CI)
Household income	Low	Reference	Reference
	High	1.145 (1.070-1.227)^*	1.146 (1.066-1.231)^*
Age group	<65	Reference	Reference
	≥65	0.944 (0.878-1.015)	0.927 (0.858-1.001)
Gender	Female	Reference	Reference
	Male	0.762 (0.682-0.852)^*	0.819 (0.724-0.928)^*
Race	Others	Reference	Reference
	White	1.105 (0.954-1.280)	1.094 (0.921-1.301)
Marital status	Single (never married)	Reference	Reference
	Widowed	0.843 (0.719-0.989)^*	0.858 (0.724-1.018)
	Divorced	0.848 (0.744-0.966)^*	0.858 (0.747-0.985)^*
	Unmarried or domestic partner	0.615 (0.356-1.064)	0.587 (0.333-1.035)
	Married (including common law)	1.062 (0.964-1.170)	1.102 (0.996-1.220)
	Separated	0.960 (0.689-1.336)	0.962 (0.665-1.390)
AJCC stage	I	Reference	Reference
	II	0.864 (0.736-1.014)	0.864 (0.731-1.021)
	III	0.669 (0.554-0.809)^*	0.687 (0.565-0.836)^*
	IV	0.385 (0.323-0.460)^*	0.370 (0.308-0.444)^*
T Stage	T1	Reference	Reference
	T2	1.338 (1.178-1.520)^*	1.310 (1.148-1.496)^*
	T3	1.045 (0.943-1.158)	1.065 (0.956-1.187)
	T4	0.885 (0.781-1.004)	0.892 (0.784-1.014)
N stage	N0	Reference	Reference
	N1	1.022 (0.925-1.130)	1.010 (0.911-1.121)
	N2	0.869 (0.761-0.993)^*	0.894 (0.776-1.028)
	N3	0.726 (0.613-0.860)^*	0.702 (0.590-0.836)^*
Surgery	No	Reference	Reference
	Yes	2.757 (2.525-3.010)^*	2.769 (2.522-3.039)^*
Radiation	No	Reference	Reference
	Yes	0.996 (0.900-1.102)	0.922 (0.826-1.028)
Chemotherapy	No or unknown	Reference	Reference
	Yes	2.125 (1.879-2.403)^*	2.065 (1.814-2.351)^*
Diagnosis to treat	<1 month	Reference	Reference
	≥1 month	1.129 (1.033-1.233)^*	1.098 (1.000-1.205)

Notes: *The difference was statistically significant with a P value of <.05. Household income refers to the median household income.

Certain disease-related parameters also significantly influence the prognosis of patients with EAC. According to AJCC staging, patients with disease progression to stage III and IV had significantly worse prognoses, especially those at stage IV, who exhibited the lowest TR (pre-PSM: TR = 0.385, 95% CI: 0.323-0.460; post-PSM: TR = 0.370, 95% CI: 0.308-0.444; P values <.05, Table 3). Conversely, patients at stage T2 showed longer survival times than those at stage T1 (pre-PSM: TR = 1.338; post-PSM: TR = 1.310, 95% CI not including 1; P values <.05, Table 3), suggesting that in cases of smaller EAC (T1–T2 stages), prognosis is not always inversely proportional to tumor size.

Surgical treatment significantly improved survival rates, considerably extending survival times (pre-PSM: TR = 2.757, 95% CI: 2.525-3.010; post-PSM: TR = 2.769, 95% CI: 2.522-3.039; P values <.05, Table 3). Chemotherapy also emerged as a significant factor for improved prognosis, with TRs above 2 before and after matching (95% CI not including 1; P values <.05, Table 3). In contrast, radiotherapy showed no significant impact on prognosis. Pre-PSM data indicated that the interval from diagnosis to initial treatment affected prognosis, but post-PSM data showed that this variable did not significantly affect CSS (95% CI including 1; P values >.05, Table 3).

The above results suggest that eight factors—economic status, gender, marital status, AJCC stage, T stages, N stage, surgical treatment, and chemotherapy—can influence the prognoses of patients with EAC. An RSF model was used to verify the impact of these eight factors on the prognoses of patients with EAC in the United States. Pre-PSM data showed that patients with low income, male gender, certain non-single marital statuses (such as widowed, divorced, and unmarried or domestic partner), high AJCC stages (III and IV), high T stages (T3 and T4), high N stages (N1–N3), no surgical treatment, and no clear chemotherapy history had a higher relative frequency of mortality (Figure 3). The results derived from pre-PSM data were not only validated by post-PSM data (Figure 4) but also consistent with the AFT model results mentioned earlier.

Figure 3.

Random survival forests illustrating the impact of different variables on the relative mortality of patients with esophageal adenocarcinoma (based on pre-propensity score matching data). For marital status: SG, Single (never married); WD, Widowed; DD, Divorced; UD, Unmarried or Domestic Partner; MR, Married (including common law); SE, Separated.

Figure 4.

Random survival forests illustrating the impact of different variables on the relative mortality of patients with esophageal adenocarcinoma (based on post-propensity score matching data). For marital status: SG, Single (never married); WD, Widowed; DD, Divorced; UD, Unmarried or Domestic Partner; MR, Married (including common law); SE, Separated.

Discussion

The impact of economic status on cancer prognosis has attracted increasing research attention.¹¹ Numerous studies have highlighted the role of socioeconomic factors in determining survival outcomes across various cancers, such as soft tissue sarcoma¹³ and melanoma.²³ However, the effect of economic status on EAC prognosis remains underexplored. This study evaluated the impact of economic status on the CSS of 4676 patients with EAC diagnosed in the United States, revealing that patients in the LHI group had significantly poorer CSS than those in the HHI group. PSM was employed to reduce baseline characteristic differences between the two groups, resulting in a more balanced sample.¹⁷ Data both before and after PSM showed the same trend, highlighting the importance of economic status on CSS. Additionally, male gender, high AJCC stages (III and IV), high T stages (T3 and T4), high N stages (N1–N3), certain non-single marital statuses (such as widowed, divorced, and unmarried or domestic partner), absence of surgical treatment, and lack of a clear chemotherapy history were identified as risk factors for the prognosis of patients with EAC. These findings may provide valuable references for the clinical management of EAC.

Economic status is closely associated with the prognosis of various cancers, including EAC. Abdel-Rahman et al reported that patients with malignant melanoma of the skin and poor socioeconomic status had worse overall survival compared to those with better economic status.²³ They also also observed that lower socioeconomic status was associated with poorer melanoma-specific survival.²³ Similarly, in bone and soft tissue sarcoma, hepatocellular carcinoma, thyroid cancer, and limited-stage small cell lung cancer, low-income patients exhibited significantly lower survival rates than high-income patients.^10,12,13,24 A same trend was found in EAC: both studies by Kramer et al.²⁵ and Geng et al.,¹⁴ and our own research have revealed the positive impact of HHI on CSS in patients with EAC. Unlike studies by Geng and Kramer,^14,25 which focused on a limited subset of early-stage (T1aN0M0) or locally advanced (T4) patients, our analysis encompassed patients with EAC across all AJCC stage, thereby enhancing the generalizability of our findings that LHI is associated with poorer CSS in patients with EAC. Our study is notable for its larger sample size (n = 4676), longer time scale (up to 10 years), and the use of multiple statistical methods (Kaplan–Meier curves, AFT, and RSF) to corroborate the association between economic status and CSS.

The mechanisms that underpin the effect of lower socioeconomic status on EAC CSS are likely multifactorial. In certain cancers, late cancer diagnosis, differences in treatment methods, and greater physiological stress among LHI patients may contribute to the effect of economic status on prognosis.^10,12,13 According to EAC-related reports, Kramer et al discovered that in T4-stage patients with EAC, an LHI of less than $46,000 was associated with a higher overall survival risk²⁵; however, they did not further discuss the potential mechanisms of this finding. Further, Geng et al assumed that the worse prognosis in LHI patients might be associated with more aggressive EC phenotypes, lower follow-up monitoring rates, and limited health care resources in rural areas (those with LHI tend to live in rural areas).¹⁴ In our study, the disease severity (e.g., AJCC stages) and treatment modalities (surgery, radiotherapy, and chemotherapy) between the HHI and LHI groups were balanced after PSM, a robust method to reduce confounding factors.²⁶ Moreover, results from the AFT model showed that, even after adjusting for these covariates, economic status remained an independent prognostic factor for EAC CSS. Therefore, it is unlikely that disease severity and initial treatment modalities (at least for surgery, radiotherapy, and chemotherapy) are the primary reasons for the impact of socioeconomic status on CSS in EAC patients.

However, several factors, including treatment effectiveness, may influence the prognosis of EAC patients with lower economic status. One such factor is treatment adherence (e.g., continuity of chemotherapy), as low-income cancer individuals are more likely to exhibit poorer adherence due to the high costs of treatment.²⁷ Additionally, considering that the income levels in our study refer to the median household income of certain counties in the United States and that patients with EAC and LHI tend to reside in rural areas, the limited health care resources in and around these regions may affect the CSS of patients with EAC in the LHI group. Even among patients with a precancerous condition for EAC (i.e., Barrett’s esophagus), the inequity in health care access due to economic disparities has been documented.²⁸ Besides health care resources, poor health literacy may impact the prognoses of patients with EAC and lower economic status.^29,30 Furthermore, EC patients often experience negative emotions, such as anxiety and depression, during radiotherapy, which can lead to complications such as malnutrition and worsen their condition.³¹ In such cases, low-income individuals have fewer opportunities for specialized care and community support, potentially resulting in poorer quality of life and prognoses.^31,32 In summary, lower socioeconomic status may affect EAC CSS through multiple factors, including limited health care resources, unhealth habits, and psychological factors. Therefore, it is recommended to allocate more health care and educational resources to economically disadvantaged regions, as doing so may help intervene in and prevent EAC.

Other social factors are also significant prognostic indicators for patients with EAC. Our study found that male patients with EAC had worse prognoses than female patients. Previous research has attributed this to biological differences, treatment choices, and lifestyle differences between the sexes.^33,34 Certain non-single marital statuses (such as widowed, divorced, and unmarried or domestic partner) were identified as risk factors for poor EAC prognosis compared to being single. This situation may be related to differences in psychological support and lifestyle changes.³⁵ These findings underscore the importance of considering social factors into the clinical management of EAC.

Apart from social factors, the prognoses of patients with EAC are closely related to disease characteristics and treatment modalities. It is generally believed that higher stages of the same pathological type of cancer are associated with worse overall prognoses.^36,37 Our study found that this holds true for EAC, with patients at advanced AJCC stages (III and IV), high T stages (T3 and T4), and high N stages (N1–N3) experiencing poorer prognoses. Interestingly, the AFT and RSF methods used in this study consistently showed that patients with T2-stage EAC had better prognoses than those with T1-stage EAC. This suggests that for early-stage patients, the depth of tumor invasion (propria and more superficial) may not be fully correlated with the EAC prognosis and may be influenced by other factors, such as tumor location and differentiation.³⁸ This observation may provide some reference value for prognosis assessment of early-stage (T1-2) cancers. Additionally, the three major treatments (surgery, radiotherapy, and chemotherapy) are important factors in achieving favorable prognoses in EAC.^25,39,40 Our study also found that surgery and a clear chemotherapy history are important determinants of EAC prognosis, and radiotherapy is associated with relatively reduced mortality in patients with EAC. This highlights the importance of aggressive treatment in improving outcomes for patients with EAC.

This study emphasizes the importance of considering multiple factors in managing patients with EAC. However, there are certain limitations. First, as a retrospective study, it is subject to potential selection bias. Although PSM was used, it cannot eliminate all potential confounding factors. Second, the data were sourced from a single SEER database, which may limit the generalizability of the findings due to the lack validation from multicenter data. Lastly, the definition of income level in this study was relatively simplistic and did not account for more complex economic factors, such as household composition and the patients’ actual economic burden. As a result, this study could not analyze key aspects like equivalized income for each individual. Future research should expand sample sizes and incorporate multicenter data to address these limitations, providing more comprehensive intervention strategies to improve patient survival and quality of life.

Conclusion

Low-income patients have significantly poorer CSS than high-income patients and that economic status is an important independent factor for the prognoses of patients with EAC. Considering multiple social and demographic factors comprehensively is significant for improving the prognoses of patients with EAC.

Supplemental Material

Supplemental Material - The Impact of Economic Status on Cause-specific Survival in Patients With Esophageal Adenocarcinoma in the United States: A Retrospective Analysis

Supplemental Material for The Impact of Economic Status on Cause-Specific Survival in Patients with Esophageal Adenocarcinoma in the United States: A Retrospective Analysis by Xiu-Shun Xiang, Guo-Sheng Li, Jun Liu, Xiang Gao, Gui-Yu Feng, Jing-Xiao Li, Tao Huang, Jian-Ji Guo, Nuo Yang, and Hua-Fu Zhou in Cancer Control.

Footnotes

Acknowledgments

The public cohorts used in this study were obtained from the Surveillance, Epidemiology, and End Results (SEER) database ().

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was supported by Guangxi medical and health appropriate technology development and popularization application project (S2022070), Self-financed scientific research of Guangxi Zhuang Autonomous Region Health Commission (Z20211032), National Key Clinical Specialty Construction Project of China, Guangxi Medical and Health Key Discipline Construction Project, Guangxi Key Clinical Specialty Construction Project.

Ethical Statement

ORCID iD

Hua-Fu Zhou

Supplemental Material

Supplemental material for this article is available online.

Appendix

References

Sung

Ferlay

Siegel

, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71:209-249.

Wang

Shao

Wang

, et al. Efficacy and safety of neoadjuvant immunotherapy in surgically resectable esophageal cancer: a systematic review and meta-analysis. Int J Surg. 2022;104:106767.

Yang

Hong

. Advances in targeted therapy for esophageal cancer. Signal Transduct Targeted Ther. 2020;5:229.

Wang

Song

Dai

Zhang

. Comparative study of laparoscopic-assisted and open total gastrectomy for Siewert Types II and III adenocarcinoma of the esophagogastric junction. J Cell Physiol. 2019;234(7):11235-11239. doi:10.1002/jcp.27777. 30478913.

Siegel

Miller

Fuchs

Jemal

. Cancer statistics, 2021. CA Cancer J Clin. 2021;71:7-33.

Chen

Chuang

Lin

. The influence of socioeconomic status on esophageal cancer in taiwan: a population-based study. J Personalized Med. 2022;12:595.

Radkiewicz

Asplund

Lagergren

. Incidence trends and survival in early-onset esophagogastric adenocarcinoma: a Swedish population-based cohort study. Cancer Epidemiol Biomarkers Prev. 2023;24:OF1-OF8.

Lin

Wang

Fang

Zheng

. International trends in esophageal cancer incidence rates by histological subtype (1990-2012) and prediction of the rates to 2030. Esophagus. 2022;19:560-568.

Saadaat

Abdul-Ghafar

Haidary

Atta

Ali

. Esophageal carcinoma and associated risk factors: a case-control study in two tertiary care hospitals of kabul, Afghanistan. Cancer Manag Res. 2022;14:2445-2456.

10.

Huang

Wang

Mao

Chen

Dong

. Socioeconomic factors are associated with the prognosis of Thyroid Cancer. J Cancer. 2021;12:2507-2512.

11.

Rosskamp

Verbeeck

Sass

Gadeyne

Verdoodt

De Schutter

. Social inequalities in cancer survival in Belgium: a population-based cohort study. Cancer Epidemiol Biomarkers Prev. 2021;30:45-52.

12.

Wong

Kim

Ahmed

Singal

. Patients with hepatocellular carcinoma from more rural and lower-income households have more advanced tumor stage at diagnosis and significantly higher mortality. Cancer. 2021;127:45-55.

13.

Penumarthy

Goldsby

Shiboski

Wustrack

Murphy

Winestone

. Insurance impacts survival for children, adolescents, and young adults with bone and soft tissue sarcomas. Cancer Med. 2020;9:951-958.

14.

Geng

Gudur

Radlinski

, et al. Socioeconomic disparities affect outcomes in early-stage esophageal adenocarcinoma: a SEER analysis. Clin Gastroenterol Hepatol. 2023;21:2797-2806.

15.

Torrejon

Deshpande

Wei

Tullio

Kamath

. Proportion of early-onset gastric and esophagus cancers has changed over time with disproportionate impact on Black and hispanic patients. JCO Oncol Pract. 2022;18:e759-e769.

16.

Institute

. Surveillance, epidemiology, and end results (SEER) program. In: Nov 2023 Sub (1975-2021) - Linked to County Attributes - Time Dependent (1990-2022) Income/Rurality, 1969-2022 Counties. Bethesda: DCCPS, Surveillance Research Program: National Cancer Institute. www.seer.cancer.gov SEER*StatDatabase:Incidence-SEERResearchData,8Registries

17.

Kane

Fang

Galetta

, et al. Propensity score matching: a statistical method. Clin Spine Surg. 2020;33:120-122.

18.

Benchimol

Smeeth

Guttmann

, et al. The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement. PLoS Med. 2015;12:e1001885.

19.

Institute

. Surveillance research program. Bethesda: National Cancer Institute SEER*Stat software (seer.cancer.gov/seerstat).

20.

Kazuki Yoshida

. Tableone: create ‘table 1’ to describe baseline characteristics with or without propensity score weights. 2022. https://CRANR-projectorg/package=tableone

21.

Terry

Therneau

Atkinson

Cynthia

. A package for survival analysis in R. 2022. https://CRANR-projectorg/package=survival

22.

Alboukadel Kassambara

Biecek

Fabian

. survminer: drawing Survival Curves using’ggplot2. 2021. https://CRANR-projectorg/package=survminer

23.

Abdel-Rahman

. Prognostic impact of socioeconomic status among patients with malignant melanoma of the skin: a population-based study. J Dermatol Treat. 2020;31:571-575.

24.

Zhou

Shi

Chen

, et al. Association of race, socioeconomic factors, and treatment characteristics with overall survival in patients with limited-stage small cell lung cancer. JAMA Netw Open. 2021;4:e2032276.

25.

Kramer

Swanson

Fernando

, et al.

Locally advanced adenocarcinoma of the esophagus: is esophagectomy associated with improved overall survival?

J Am Coll Surg. 2024;239:68-75.

26.

Austin

. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res. 2011;46:399-424.

27.

Xiong

Brown

Horras

Yuan

. Impact of cost-related medication nonadherence on economic burdens, productivity loss, and functional abilities: management of cancer survivors in medicare. Front Pharmacol. 2021;12:706289.

28.

Fujiyoshi

Jugnundan

, et al. Impact of residing in below median household income districts on outcomes in patients with advanced barrett’s esophagus. J Can Assoc Gastroenterol. 2023;6:137-144.

29.

McCain

McManus

McQuaid

, et al. Alcohol intake, tobacco smoking, and esophageal adenocarcinoma survival: a molecular pathology epidemiology cohort study. Cancer Causes Control. 2020;31:1-11.

30.

Wang

Cao

Liu

, et al. The effect of socioeconomic status on health-care delay and treatment of esophageal cancer. J Transl Med. 2015;13:241.

31.

Luo

Chen

Chang

Jiang

Yin

. Analysis of the risk factors for negative emotions in patients with esophageal cancer during the peri-radiotherapy period and their effects on malnutrition. Int J Gen Med. 2023;16:6137-6150.

32.

Karunanayake

Rennie

Hagel

, et al. Access to specialist care in rural saskatchewan: the saskatchewan rural health study. Healthcare. 2015;3:84-99.

33.

Dijksterhuis

WPM

Kalff

Wagner

, et al. Gender differences in treatment allocation and survival of advanced gastroesophageal cancer: a population-based study. J Natl Cancer Inst. 2021;113:1551-1560.

34.

Argyrakopoulou

Dalamaga

Spyrou

Kokkinos

. Gender differences in obesity-related cancers. Curr Obes Rep. 2021;10:100-115.

35.

Wang

Chen

, et al. Effect of marital status on the survival of patients with adenocarcinoma of the esophagogastric junction: a population-based, propensity-matched study. Cancer Control. 2021;28:10732748211066309.

36.

Liu

, et al. Prediction of lymph node metastasis of lung squamous cell carcinoma by machine learning algorithm classifiers. J Cancer Res Therapeut. 2023;19:1533-1543.

37.

Scher

Elson

Kharofa

. Prognostic implications of advancing nodal stage in anal cancer: evaluating AJCC 8 staging changes through an NCDB analysis. Int J Colorectal Dis. 2020;35:2099-2104.

38.

Wang

Jiang

Xiao

Xue

Zhang

. Prognostic impact of tumor length in esophageal Cancer: a systematic review and Meta-analysis. BMC Cancer. 2021;21:988.

39.

Chen

Liu

Han

, et al. Does chemoradiotherapy benefit elderly patients with esophageal squamous cell cancer? A propensity-score matched analysis on multicenter data (3JECROG R-03A). BMC Cancer. 2020;20:36.

40.

Rahouma

Harrison

Kamel

, et al. Consequences of refusing surgery for esophageal cancer: a national cancer database analysis. Ann Thorac Surg. 2018;106:1476-1483.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB