Epidemiology and Survival of Esophageal Cancer Patients in an American Cohort

Objectives This study seeks to delineate trends in esophageal cancer patients in an American cohort and, in particular, examine the impact of race and histology on survival. Methods The association between over 50 variables between histology and race subgroups was evaluated. Survival was calculated using Kaplan-Meier curves and a multivariable Cox regression analysis (MVA) was performed. Results Poorer survival was noted in black vs. white (193 ± 65 days vs. 254 ± 39, 95% CI 205-295, p=0.07) and squamous cell cancer (SCC) vs. adenocarcinoma (AC) (233 ± 24 days vs. 303 ± 48, 95% CI 197-339, p=0.01) patients. In patients with resectable cancer, blacks had poorer survival than whites (253 ± 46 days vs. 538 ± 202, 95% CI 269-603, p=0.03), and SCC had poorer survival than AC (333 ± 58 vs. 638 ± 152 days, 95% CI 306-634, p=0.006). A higher percentage of white patients received surgery compared to black patients (36% vs. 8%, p=0.08). MVA revealed that only surgery was an independent predictor of mortality (p=0.001). Conclusion Black race and SCC were associated with poorer survival. On MVA, surgery was an independent predictor of mortality. Clinicians should be aggressive in offering potentially curative procedures to patients and eliminating socioeconomic barriers.


Introduction
Esophageal cancer (EC) is one of the most common cancers diagnosed, with an estimated global incidence of 455,800 [1]. In the United States, it is estimated that there will be 17,290 new cases and 15,850 deaths in 2018 [2]. In addition to the rise in incidence over the past few decades [3], there appears to be a concurrent change in histological type, with studies reporting a 463% increase in the incidence of adenocarcinoma (AC) in Americans in 2000-2004 as compared to 1975-1979 [4]. Adenocarcinoma has replaced squamous cell carcinoma (SCC) as the most common type of esophageal carcinoma, in part due to an increase in risk factors for AC such as obesity and gastroesophageal reflux disease (GERD) [5]. The tumor biology of EC is still not fully understood, and patients with SCC have been shown to have significantly poorer survival than patients with AC [6]. Efforts at elucidating the underlying mechanisms as well as developing biomarkers for early detection and screening have proved daunting [7]. Moreover, there seems to be a racial disparity in survival between white and black patients with esophageal carcinoma for reasons that are still not fully clear, with the age-adjusted incidence and death for blacks almost twice that for whites [8].
This study uniquely analyzes symptoms and risk factors at presentation as well as various demographic, pathological, clinical, and laboratory variables. In addition to delineating trends in the presentation and treatment of esophageal cancer patients in an American cohort, it seeks to closely examine the impact of race and histology on survival in our veteran patient population.

Materials And Methods
The variables assessed in the study are shown in Table 1

Demographics and epidemiology
A total of 122 patients with esophageal cancer were identified. Their features are presented in Table 2.  The mean age was 64 ± 0.9 years and the mean body mass index (BMI) at diagnosis was 26 ± 0.7. The cohort was predominantly male (97.5%) and white (75.4%). AC was found in 61.5% of the patients and 33.6% had SCC. The majority of the cancers were in the lower esophagus (73.8%) as compared to the upper esophagus (26.2%).
The patients presented mostly with advanced stage carcinoma with almost half (41%) presenting with metastatic disease. The vast majority of patients had a documented history of tobacco (82.8%) and alcohol (73.8%) use. A significant number also had gastroesophageal reflux disease (GERD) (43.4%), whereas only 28.7% were on a proton pump inhibitor. The most common complaint at presentation was weight loss (55.7%) followed by dysphagia to solids (40.2%). In terms of interventions, 41.8% patients received a gastric or jejunal feeding tube, 21.3% received esophageal stents, 15.6% patients had an esophagectomy, and 66.4% patients received some form of chemotherapy or radiation.

Survival
On a Kaplan-Meier analysis, overall median survival for the entire cohort was 253 ± 28 days.

Stage
As expected, there was an inverse relationship between stage of disease at presentation and survival with a median survival of 1447 ± 651 days, 402 ± 137 days, 292 ± 53 days, and 134 ± 35 days for Stage I, II, III, and IV patients, respectively (95% CI 199-307, p<0.0001).

Resectability
Patient survival was stratified by resectability. Stage IV patients were considered to have nonresectable tumors (NR). Patients that had potentially resectable disease (defined as Stage I-III) were grouped into patients that received an esophagectomy (R+SURG), endoscopic mucosal resection (R+EMR), and patients that did not receive surgical/endoscopic intervention (R-). Patients with an unknown stage were not included in the analysis. Patients with NR cancer had poorer survival compared to patients with resectable cancer (stage I-III) (134 ± 45 vs. 470 ± 84 days, 95% CI 193-343, p<0.0001).
Among the group that was potentially resectable (stages I-III), the median survival for Rpatients compared with patients who received any kind of surgical/endoscopic intervention (R+Surg and R+EMR) was 364 ± 81 days vs. 1044 ± 278 days (95% CI 193-343, p<0.0001).
This data is represented in Figure 1. 2018

By Race
A univariate analysis was conducted to elucidate any differences in any of the measured variables between white and black patients ( Table 3).  .008) proved to be statistically significant between whites and blacks.

By Histology
A univariate analysis was conducted to elucidate any differences in any of the measured variables between patients with differing histological subtypes (  There was no significant difference in interventions such as chemotherapy, radiation, enteral tube feeding, or esophageal stenting between AC and SCC patients (see Table 4).
Of the patients in our cohort who had a history of achalasia, all presented had SCC. More patients with SCC complained of dysphagia to solids (p=0.02) and hoarseness (p=0.008) whereas more patients with AC presented complaining of fatigue (p=0.03). Patients with AC had a higher BMI at presentation compared with SCC (27.7+ 0.9 vs. 23.0 + 0.9, p=0.002).

Multivariable analysis
All variables with a univariate analysis p-value < 0.1 obtained on a comparison between the histology and race subgroups were included in a multivariable Cox regression analysis to elucidate predictors of mortality (  Of the variables included, only surgery (p=0.001) was statistically significant.

Discussion
Although there was a higher percentage of SCC in blacks as well as a poorer survival rate for SCC compared with AC, the multivariable analysis revealed that histology and race did not play a role in overall survival. The receipt of an esophagectomy was an independent predictor of survival.
Esophagectomy for locally advanced esophageal cancer is currently the standard of care, and neoadjuvant chemoradiation has been shown in several studies to confer a survival advantage [8][9][10]. Definitive chemoradiation alone has generally not been recommended, although a recent Cochrane meta-analysis suggested that survival outcomes for definitive chemoradiation may be equal to surgery in SCC patients [11]. Several studies have suggested that surgical resection for esophageal cancer is underused [12], likely due to a combination of factors, including limited access to high-volume centers and patient reluctance to undergo surgery.
In our study, on a Kaplan-Meier analysis, there was a statistically and clinically apparent survival benefit of receiving an esophagectomy (p<0.0001). However, in our cohort, there was a total of 52% patients that had potentially resectable tumors (R+ surgery/EMR), but less than half of them (41%) underwent resection. When analyzing causes for not receiving surgery in potentially resectable patients (R-) ( Table 6), 16% of the patients were found to have cancer that advanced after neoadjuvant treatment that precluded them from surgery. Almost one-third of the patients (11/38, 29%) refused surgery and were treated solely with chemoradiation despite having a good (ECOG 0-1) functional status. Given the significant survival advantage in patients who received surgery, it seems prudent to be aggressive in offering surgery to patients early and highlighting the survival benefits clearly as well as identifying and eliminating potential socioeconomic barriers.

Reason n=38 % (/38)
Tumors not amenable to surgery due to local invasion after neoadjuvant 6 16% Tumor too locally advanced at diagnosis 3 8% Later found to have metastases and plans for surgery were aborted 3 8% Lost to follow-up 8 21% Refused surgery/treated with chemoradiation despite good ECOG 11 29% Poor performance status 4 11% Previous gastrointestinal surgery that precluded esophagectomy 3 8% Although randomized prospective trials are lacking, the use of endoscopic methods instead of surgery for the treatment of early mucosal esophageal cancer have become ubiquitous, with EMR for T1a the most common approach in the United States and increasing in frequency for T1b tumors [13]. Observational studies have demonstrated equal survival rate at one, three, and five years for patients with T1a treated with either endoscopic resection or esophagectomy, although endoscopic therapy is associated with a slightly higher recurrence rate [14]. Similarly, when comparing patients with Stage I cancer who received endoscopic treatment vs. surgery in our study, there was no statistically significant difference in survival noted although this may have been secondary to the limited population size of this subgroup.
Several population-based studies have reported a poorer survival of black patients compared to whites [15][16][17]. In earlier studies, this was sometimes proposed to be secondary to the increased incidence of SCC (which has poorer survival) in blacks.
However, several epidemiological studies have demonstrated that there may be other factors at play. One study demonstrated worse five-year survival rates for blacks compared to whites (37% vs. 60%, p<0.001). However, on a multivariate analysis and controlling for histology among other factors, the relationship between race and survival was not significant when surgery was taken into consideration [15]. Similarly, a subsequent large analysis of the surveillance, epidemiology, and end results (SEER) database confirmed the disparity in survival and incidence in SCC between black and white patients. However, on a multivariate analysis, the survival disparity disappeared after adjusting for the receipt of esophagectomy [16].
Conversely, in a recent study, Taoili et al. confirmed that blacks had a lower rate of esophagectomy; however, the differences in survival persisted even when adjusting for cancerdirected surgery. Interestingly, independent of stage, white and black patients who did not have surgery experienced similar survival, although among patients who did undergo surgery, there was still a poorer overall survival for blacks compared to whites [17]. Possible explanations for this disparity in esophagectomy rates among blacks may be poorer health care delivery, less access to specialized surgeons, poor access to high-volume centers, or secondary to patient refusal. It is also possible that a proportion of these cancers may be less resectable due to the preponderance of SCC in these patients, and thus the location of cancers in the mid-esophagus [15][16][17][18].
In our study, there was a significant difference in survival between white and black patients with potentially resectable cancers (Stage I-III), with whites surviving twice as long as blacks (median survival 538 ± 202 vs. 253 ± 46 days, p=0.03, 95% CI 269-603). When comparing overall survival rates between histology groups, we noted differences in survival, with AC patients having superior survival than SCC patients (303 ± 48 days vs. 233 ± 24, 95% CI 197-339, p=0.01). AC has been shown to have superior survival to SCC in the literature [19][20]. In one study that evaluated survival between AC and SCC patients who underwent curative resection, the fiveyear survival rate was 42.3% vs, 30.3%, p<0.01. This disparity in survival was consistent even in patients who had a complete macroscopic and microscopic resection as well as negative lymph nodes (R0N0) [21]. This poorer survival may be secondary to the higher rate of occult micrometastases [21], the histology-specific aggressiveness of cancer, or other confounding factors and exposure relating to the patient population, including race.
In our study, almost 80% of blacks presented with SCC compared with 20% of whites (p<0.0001). The higher incidence of SCC in blacks compared to whites has been documented in large population studies [7], but the causes are not fully apparent. Proposed etiologies have involved the lower socioeconomic class (SEC), lower intake of fruits and vegetables, and increased rates of alcohol and tobacco usage [22][23]. However, it is not clear why these factors, while definite risk factors for SCC, result in a higher incidence of SCC in blacks compared with whites. It may be secondary to increased amounts of tobacco and alcohol consumed compared to whites, something hard to accurately quantify in large studies. Furthermore, while lower SEC has been reported to be a risk factor in many studies [22,[24][25], it is not exactly clear what the underlying exposures causing these associations are, as low SEC can be considered a surrogate for various lifestyle and environmental exposures, including poor nutrition, poor access to health care, and poor housing [22].
Underlying genetic polymorphisms and race-specific genetic susceptibility may prove to play a significant role in the pathogenesis of SCC. Many recent studies have found various single nucleotide polymorphisms (SNPs) in micro RNA (miRNA) sequences in SCC patients compared with controls [26][27]. These mutations, under the potential effects of factors like alcohol and tobacco, may lead to an alteration of miRNA expression and contribute to carcinogenesis.
Others have reported polymorphic mutations in genes coding for metabolic enzymes, Deoxyribonucleic acid (DNA) repair enzymes and cytokines, which may contribute toward SCC susceptibility [28][29][30]. While the majority of these studies were done investigating the particularly high incidence of SCC in China and the so-called "esophageal cancer belt," it may help to explain the increased incidence of SCC in blacks compared to whites.
This study benefits from several strengths. To our knowledge, this is the only cohort of esophageal cancer patients in the United States where such detailed variables, such as potential risk factors and symptoms at presentation, were included, in addition to various laboratory clinical and pathological data across all stages and treatment modalities. In addition, this study's detailed survival analysis presented in days and stratified by various variables gives providers a better idea of realistic median survival as compared to the five-year survival rates presented elsewhere in the literature. All patients in this cohort were veterans and the VA North Texas Health Care System (VANTHCS) was their primary hospital. All follow-up and treatment were conducted there. This fact, along with the comprehensive and centralized computerized health records, allowed us to collect accurate data endpoints and calculate reliable survival times.
There were a few weaknesses in this study. First, the cohort was of a relatively small sample size and the findings must be interpreted in light of the inherent bias in a retrospective study with a sample of this size. Our population was predominantly male, and of white or black race, and findings cannot be generalizable to other races or women. In addition, it is possible that our patient population had a higher incidence of baseline risks factors (e.g. smoking), which may affect our findings and make them not generalizable to other patient populations.

Conclusions
In conclusion, there was a significant difference in survival between AC and SCC patients and between White and Black patients with resectable cancer. However, histology and race did not play a role in survival when other factors were controlled in a multivariate analysis. Surgery was found to be an independent predictor of mortality. Given the improved survival benefit of surgery as well as its underuse, clinicians should be more aggressive in offering potentially curative esophagectomies to patients.

Additional Information Disclosures
Human subjects: Consent was obtained by all participants in this study. IRB VANTHCS Dallas issued approval 003230. The Institutional Review Board (IRB) met on 10/02/2017 to review the application for continued approval of your study, referenced above. The use of a consent form was waived in accordance with 45 CFR 46.116(d). 2. Your study was approved by the IRB for a 12 Months continuation. Your study will be subject to continuing review on or before 10/01/2018. 3. For any change to approved research projects require prior review and approval before implementing such changes. 4. Please note that although this study was issued waiver of informed consent, in accordance with federal regulatory law, you must inform the IRB of deviations from the approved protocol or any loss of, or compromise to PHI associated with this study. Animal subjects: All authors have confirmed that this study did not involve animal subjects or tissue. Conflicts of interest: In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.