Comparing treatment and outcomes in advanced esophageal, gastroesophageal junction, and gastric adenocarcinomas: a population-based study

Background: Treatment of advanced or metastatic esophageal adenocarcinoma (EAC) follows the guidelines for gastroesophageal junction adenocarcinoma (GEJC) and gastric adenocarcinoma (GAC), but patients with EAC are often excluded from clinical studies of GEJC/GAC. Objectives: Here we describe treatment and survival of patients with advanced EAC, GEJC, and GAC to provide population-based evidence on distinctions and similarities between these populations. Design: Retrospective cohort study of patients with unresectable advanced (cT4b) or metastatic (cM1) EAC, GEJC, or GAC (2015–2020) were selected from the Netherlands Cancer Registry. Methods: Overall survival (OS) was assessed using Kaplan–Meier methods, log-rank tests, and multivariable Cox regression. Results: In all, 7391 patients were included (EAC: n = 3346, GEJC: n = 1246, and GAC: n = 2798). Patients with EAC were more often males and more often had ⩾2 metastatic locations. First-line systemic therapy was received by 42%, 47%, and 36% of patients with EAC, GEJC, and GAC, respectively. Median OS was 5.0, 5.1, and 4.0 months for all patients with EAC, GEJC, and GAC, respectively (p < 0.001). Median OS from start of first-line therapy of patients with human epidermal growth factor receptor 2 (HER2)-negative adenocarcinomas was 7.6, 7.8, and 7.5 months (p = 0.12) and of patients with HER2-positive carcinoma receiving first-line trastuzumab-containing therapy was 11.0, 13.3, and 9.5 months (p = 0.37) in EAC, GEJC, and GAC, respectively. After multivariable adjustment, no difference in OS for patients with EAC, GEJC, and GAC was observed. Conclusion: Despite differences in clinical characteristics and treatment strategies, survival between patients with advanced EAC, GEJC, and GAC was similar. We advocate that EAC patients should not be excluded from clinical trials for patients with molecularly similar GEJC/GAC.


Introduction
Esophageal adenocarcinoma (EAC) is the most common subtype of esophageal cancer in Western countries and accounts for around two-thirds of cases. 1 In gastric and gastroesophageal junction cancers, 95% of cases are adenocarcinoma. 2 The main risk factors for EAC are Barrett's esophagus, gastroesophageal reflux disease, and obesity, the latter two are also risk factors for gastroesophageal junction adenocarcinoma (GEJC), whereas the main risk factor for developing gastric adenocarcinoma (GAC) is a Helicobacter pylori infection. 3,4 Gastric cancer can be classified into four distinct genomic subtypes. 5 One such subtype is characterized by an intestinal histology and exhibits chromosomal instability (CIN). 5 On a molecular level, EAC strongly resembles the intestinal CIN subtype as opposed to esophageal squamous cell carcinoma which is molecularly distinct from the gastric cancer genomic subtypes. 6 EAC, GEJC, and GAC share molecular similarities and could therefore be considered the same disease entity. 6 According to the current European Society for Medical Oncology (ESMO) guidelines for gastric cancer, first-line palliative systemic treatment consists of a fluoropyrimidine and platinum doublet, with the addition of trastuzumab in patients with a human epidermal growth factor receptor 2 (HER2)-positive tumor. [7][8][9][10] Recently, the addition of nivolumab to first-line chemotherapy showed an improved survival compared to chemotherapy alone in patients with advanced EAC, GEJC, or GAC, particularly for programmed death-ligand 1 combined positive score ⩾5. 11 In second line, paclitaxel in combination with ramucirumab is considered standard of care. 12 However, evidence for the efficacy systemic therapy in patients with EAC is limited, as these patients were excluded from pivotal clinical trials for GEJC/GAC (Supplemental Table 1). 8,[11][12][13][14][15][16] Although ESMO guidelines for palliative systemic therapy advise to treat patients with EAC mainly according to the guidelines of GEJC/ GAC, this may be hampered by formal registration constraints. 7,17 For example, the EMA registration of ramucirumab is limited to GEJC/GAC due to exclusion of patients with EAC from the pivotal RAINBOW trial. 12 In this population-based study, we compared treatment strategies, survival, and time to treatment failure (TTF) of patients with advanced EAC, GEJC, and GAC to provide populationbased evidence on distinctions and similarities of these populations.

Study population
Patients with an adenocarcinoma of the esophagus (C15.0-C15.9), gastroesophageal junction/cardia (C16.0), or stomach (C16.1-C16.9) diagnosed with unresectable advanced (cT 4b cN all cM 0 ) or synchronous metastatic disease (cT all cN all cM 1 ) in 2015-2020 were selected from the Netherlands Cancer Registry (NCR) (Supplemental Figure 1). Tumors were staged using the International Union Against Cancer TNM classification 7th edition (2015-2016) and 8th edition (2017-2020). 18,19 The NCR is a nationwide population-based cancer registry that covers the Dutch population and is based on notification of all newly diagnosed malignancies by the national automated pathology archive.  20 In short, TTF was assessed until the first progression that resulted in the end of first-line treatment. Patients were censored on date of the last hospital visit if no progression was registered.

Statistical analyses
Patient, tumor, and treatment characteristics were compared between patients with EAC, GEJC, and GAC using chi-squared test, Fisher's exact test, or ANOVA where appropriate. To evaluate OS and TTF, Kaplan-Meier methods and log-rank tests were used. OS was also evaluated using univariable and multivariable Cox proportional hazard models. The multivariable models were adjusted for available clinical characteristics. The proportional hazards assumptions for these variables were tested with time-dependent covariates as a function of the survival time. If these interaction terms were significant, the Schoenfeld residual plots were graphically inspected and if the residuals of the covariates changed over time, the covariates were deemed non-proportional and the Cox model was stratified for these covariates instead of adjusted. p Values of <0.05 were considered statistically significant. All analyses were conducted using SAS® version 9.4 (SAS Institute, Cary, NC, USA).

Discussion
In this nationwide population-based cohort study of patients with advanced EAC, GEJC, and GAC, we found that despite differences in patient and tumor characteristics as well as treatment, survival of patients with EAC was similar to patients with GEJC and GAC. This observation is consistent with response to cancer treatment being influenced more by the genetic profile of gastroesophageal tumors than by their anatomical location.
Recent research efforts in gastroesophageal cancers have led to the conclusion that on a molecular level EAC clusters with adenocarcinomas of the gastroesophageal junction and the CIN subtype of gastric cancer. 5,6 The CIN subtype echoes the intestinal tumor type on a histological level. 5 In our analysis of patients with an intestinal tumor type, also no survival differences between EAC, GEJC, and GAC were observed. In the curative setting, the four distinct genomic subtypes of gastric cancer, including the CIN subtype, have previously been associated with different prognostic outcomes. 22 In the palliative setting, the association of the subtypes of gastric cancer with prognostic outcomes is not yet know and in the current study the association of the four genomic subtypes of gastric cancer was unavailable due to the irrelevance in daily clinical practice.
Our results are in line with a previous pooled analysis study including patients from four randomized controlled trials reported no difference in survival, objective tumor response, and toxic effects between advanced EAC, GEJC, and GAC. 23 In addition, a retrospective observational study in the United States including patients with HER2-negative cancer receiving first-line treatment no difference in survival was reported between unresectable advanced gastric or gastroesophageal junction cancer versus EAC. 24 Survival in our study among patients with performance status of 0-1 receiving CapOx/ FOLFOX for a HER2-negative tumor (mOS: 8.3-8.8 months) was lower as compared to the recent phase III trial, CheckMate 649, which reported a median OS of 11.6 months for all randomly assigned patients receiving chemotherapy alone. 11 This difference in survival is probably the result of strict exclusion criteria besides performance status of 0-1, such as the presence of other (auto-immune) diseases and/or comorbidities.
Treatment strategies differ between patients with EAC, GEJC, and GAC. In our study, we observed that patients with GAC (7%) more often underwent palliative resection as compared to EAC (2%) or GEJC (3%). Indeed, palliative gastrectomy is regarded as a treatment option to relieve symptoms, whereas an esophagectomy is not. 7,17,25,26 The percentage of patients receiving first-line systemic therapy (without resection or radiation) was lowest in GAC (36%) compared to EAC (42%) and GEJC (47%). This could be due to several reasons: patients with GAC more often received palliative resection (with or without perioperative therapy), they were slightly older, more often female, more often had a performance status of >1, and may therefore have refrained from palliative systemic therapy.
In contrast to the ESMO guidelines primarily recommending first-line doublet therapy consisting of a fluoropyrimidine and platinum, we found a high number of EAC patients with HER2negative or unknown HER2 status who received first-line carboplatin and paclitaxel as compared to GEJC and GAC. 7,17 In a retrospective observational study from the United States, administration of first-line carboplatin and paclitaxel was approximately 2-5% and in a population-based study in Canada, the use of carboplatin/cisplatin and paclitaxel was 2%. 27,28 The high proportion of patients receiving first-line carboplatin and paclitaxel appears to be clinical practice in a    29,30 Another reason could be the usage of carboplatin Less than a third of patients who received first-line treatment received second-line treatment. We did not observe OS differences between patients with EAC, GEJC, and GAC receiving second-line treatment. Previous studies reported a survival benefit of second-line treatment with paclitaxel and ramucirumab compared to taxane monotherapy. 12,21 In patients with EAC, second-line paclitaxel and ramucirumab usage was almost 30% lower as compared to patients with GEJC or GAC. Patients with EAC were not eligible for the pivotal clinical trial of paclitaxel and ramucirumab in second line (RAINBOW) and consequently this regimen is not formally registered for EAC. 12 Therefore, carboplatin and paclitaxel can be chosen in first line when paclitaxel and ramucirumab is not considered an option in second line. However, our study shows that in clinical practice patients with EAC are tested for HER2 and if positive receive trastuzumab, and second-line paclitaxel and ramucirumab. Due to the changing treatment landscape in esophagogastric cancer, further studies should investigate whether the effectiveness of novel therapies is also similar for EAC, GEJC, or GAC.
The strength of our study is the use of population-based data, which reflects treatment and outcomes including frail patients, elderly and patients with comorbidities. Our study also has limitations. First, not all patients with HER2positive carcinoma received a trastuzumab-containing regimen, these patients (n = 80) were excluded from the survival analysis. It was previously reported in a population-based study from the Netherlands that 23% of patients with HER2positive cancers did not receive a trastuzumab containing regimen which could be partly due to the lack of registration for true EAC tumors but is potentially related to cardiac comorbidity. 32 Second, missing data on HER2 status and other variables, for example, performance status, could have resulted in suboptimal adjustment of multivariable models. Third, for patients with a HER2positive tumor, it was unknown if the result of the IHC staining patterns was 2+/3+ which could also have contributed to suboptimal adjustment. Lastly, misclassification of the primary tumor location could have occurred as information on tumor margins was unavailable. Classification of the primary tumor was based on the diagnosis in the clinical practice.
In conclusion, we have shown that despite differences in patient and tumor characteristics, survival is comparable between patients with advanced EAC, GEJC, and GAC. We advocate that patients with EAC should not be considered a separate entity and should be included in trials encompassing molecularly similar gastroesophageal junction and gastric cancers to be able to have similar benefit from novel treatment options.

Declarations
Ethics approval and consent to participate According to the Central Committee on Research involving Human Subjects, this type of study does not require approval from an ethics committee in the Netherlands. This study was approved by the Privacy Review Board of the NCR and the scientific committee of the Dutch Upper GI Cancer Group. Based on current Dutch legislation, it is not necessary to retrieve informed consent from patients for registration into the NCR. The privacy review board of the NCR reviews all data requests for studies with data of the NCR regarding privacy issues.

Consent for publication
Not applicable.