Prognostic impact of advanced lung cancer in ammation index (ALI) on patients with adenocarcinoma of esophagogastric junction (AEG) after radical resection

Qianchao Liao Shantou University Medical College Zifeng Yang Guangdong Academy of Medical Science Xu Hu A liated Hospital of Guangdong Medical University Chengbin Zheng Jieyang People’s Hospital Huolun Feng Guangdong Academy of Medical Science Zhenru Deng Guangdong Academy of Medical Science Jiabin Zheng Guangdong Academy of Medical Science Junjiang Wang Guangdong Academy of Medical Science Weixian Hu Guangdong Academy of Medical Science Yong Li (  liyong@gdph.org.cn ) Guangdong Academy of Medical Science


Introduction
In recent decades, the incidence of the adenocarcinomas of esophagogastric junction (AEG) has increased worldwide [1][2][3]. Based on the Siewert classi cation [4], AEG is de ned as a tumor with an epicenter crossing 5cm above or below the esophagogastric junction. Despite multimodality treatment strategies, the prognosis of AEG, especially advanced stage AEG, remains poor [5,6].
The TNM classi cation, related biomarkers and clinicopathological factors are essential indications affecting the prognosis of AEG [7][8][9][10]. An improved and individualized prognostic model could further help to stratify the risk of these patients and better guide their clinical management.
Recently, the advanced lung cancer in ammatory index (ALI) was reported as an independent marker of poor outcomes in patient with advanced non-small-cell lung cancer [21]. ALI combines BMI, albumin level and NLR into its evaluation. Compared with NLR which mainly re ects patients' in ammatory status, BMI and serum albumin are suggested as biomarker for nutritional condition [22,23]. Hence, ALI could have a better prospect as a conventional clinical biomarker for simultaneously evaluating both two aspects. The signi cance of ALI in nasopharyngeal cancer, large B cell lymphomas and esophageal cancer have been studied [24][25][26], however to our knowledge, the clinical value of ALI on purely AEG patients had not yet been established.
In this study, we aim to assess the clinical signi cance of ALI in patients with AEG after radical resection and developed a model for easy clinical implementation to better guide their management. We present the following article in accordance with the STROBE reporting checklist.

Patients
The records of patients with AEG who underwent radical resection at Guangdong Provincial Peoples' Hospital (Guangzhou, China), from January 2008 to December 2018, were assessed. Inclusion criteria included: 1. a pathological diagnosis of adenocarcinoma; 2. underwent radical surgical resection (R0); 3. the tumor epicenter was located within 5cm above or below the esophagogastric junction and tumor invaded the EG junction; 4. absence of distant metastasis; 5. had completed clinical and follow-up data.
Patients' clinicopathologic data were extracted from the hospitals inpatient management system and included age, sex, body mass index (BMI), Siewert classi cation, T stage, lymph node metastasis, lymphovascular invasion (LVI), perineural invasion (PNI), differentiation, tumor size and ALI. The Siewert subtype classi cation was de ned based on the patient's preoperative gastroscopy, computed tomography, digestive tract radiography and postoperative specimen pathological report. Regular postoperative follow-up was performed. Patients were observed at 3-month intervals for 2 years after surgery, and every 6 months for the 3rd to 5th year by reviewing hospital records, yia outpatient clinics and telephone interviews. Pathological staging was based on the 8th edition of the AJCC TNM classi cation.
All patients provided signed informed consent for using their data anonymously for research purposes. The study was performed in accordance with the standards of the Declaration of Helsinki.

Laboratory test
The patients' weight and height were routinely measured after admission. Blood samples were taken within 1 week of surgery. ALI was calculated using the following formula: ALI = BMI × serum albumin (Alb: g/dL)/ (absolute neutrophil count/ absolute lymphocyte count [NLR]), BMI was calculated using the formula: weight (kg)/ (height [m]) 2 [21].

Statistical analyses
Continuous variables are expressed as the mean± SD, and categorical variables were reported as frequencies with percentages. For comparing the correlation between ALI and other clinicopathological factors, the Student's t-test or Mann-Whitney U test was used for continuous variables, whereas the χ2 test or Fisher's test was used to compare categorical data. Patients' information about ALI, tumor size and overall survival (OS) time were added to the X-tile software (version 3.6.1) [27] for estimating the optimal cutoff value respectively.
Kaplan-Meier method was used to construct the survival curves and were compared by log-rank test. The Cox proportional hazards method was used to compare the HR for death and recurrence. OS was de ned as the time between the date of surgery until death from any reasons or last contact. Disease-free survival (DFS) was de ned as the time between the date of surgery until tumor recurrence or death from any reasons or the last contact. The clinicopathological variables with a p 0.10 in univariate analysis were selected for multivariate analysis by Cox regression models. Independent variables from multivariate analyses for OS with a p 0.05 were used to developed a nomogram to predict the 1-, 2-, and 3-year OS survival rates. A C-index value in the range of 0.5-1.0 between predicted and actual outcomes was used to evaluate the predictive ability and discriminative ability of the nomogram model. Calibration plots were used to assess the tting degree of the nomogram and the relative operating characteristic (ROC) curve with AUC was used to evaluate the discriminative and predictive ability. The statistical analysis was performed using the SPSS (version 22, Chicago, IL, USA), R (www.R-project.org, version 3.6.3) and GraphPad Prism (version 8.2.1) software.
The determined cutoff value using the X-tile software for ALI was 43.1. Patients with ALI lower than 43.1 were classi ed into a low-ALI group (n = 90 [61.2%]) while those with ALI higher than 43.1 into a high-ALI group (n = 57 [38.8%]). The correlation between the clinicopathological factors and ALI is shown in Table   1. More patients with low ALI had larger tumor size (p=0.011). Patients in the high-ALI group had higher preoperative BMI (p 0.001), albumin level (p 0.001), and NLR (p 0.001).

Construction of the nomogram for OS
Poor differentiation, lymphovascular invasion, tumor size, and low-ALI were the independent prognostic factors for OS used to construct the nomogram (Fig. 2). The C-index for the OS predicting nomogram of patients with AEG was 0.699 (95%CI 0.636-0.762). The dotted line in the calibration plot represented the predictive value of the nomogram, while the colorful line represented the actual 1-, 2-and 3-year OS rates. The more cohesive the colorful line to the dotted line, the more precise the nomogram's predictive capability. As shown in Fig. 3A, the colorful lines uctuated above and below the dotted line, indicating a reliable predictive value of the nomogram. Besides, the AUC for the 1-, 2-, and 3-year OS were 0.736, 0.712, and 0.697 respectively, indicating acceptable discrimination of this nomogram (Fig. 3B).

Subgroup survival analyses of ALI in different stage patients
Next, we evaluated whether ALI could be a predictor for OS or DFS in different stage AEG. We found that ALI is also independent prognostic predictor for stage III/IVA AEG patients. For these patients, the median OS time was 33 months in the low-ALI group and 87 months in the high-ALI group. 1,3,5-year overall survival rate was 81.4%, 47.7% and 31.0% respectively for low-ALI group, while 92.3%, 68.6% and 58.1% for high-ALI group (Fig. 1C). The median disease-free survival time was 27 months in the low-ALI group and 45 months in the high-ALI group. 1,3,5-year disease-free survival rate was 71.2%, 35.3%, 33.2% for low-ALI group and 79.5%, 53.5%, 44.1% for high-ALI group (Fig. 1D).  (Table 3).

Discussion
In recent years, there was rising incidence of AEG in western countries and similar trend also found in East Asia, probably for the sake of obesity, gastroesophageal re ux disease and eradication of Helicobacter pylori [1,28,29]. AEG had demonstrated different biological behavior and pathological features compared with either gastric or esophageal cancer [30]. It has attracted more attention and has gradually become a distinct tumor category from esophageal and gastric cancer. Thus, we believed deeper understanding of AEG could bene t to better management of these patients.
There was accumulating evidence indicating that in ammatory response and nutritional status playing important role in malignancies [11,12,[31][32][33][34]. Several studies had evaluated the in ammatory indexes, including NLR, PLR, GPS, as prognostic predictors in upper gastrointestinal cancers [35,36]. In another hand, poor nutritional status, for example, decrease of preoperative albumin level or low BMI value, had demonstrated as indicators for poor prognosis [22,23,37]. According to previous study, we thought ALI could be a potential surrogate combining both two dimensions. ALI's prognostic impact had been evaluated in several cancers [21,24,26] and it also had been showed to be independent predictor in gastric and esophageal cancer. However, in Yin's et al [38] study tumor mostly located at gastric body or pylorus and Tan's et al [25] study mostly included squamous cell carcinoma of esophagus. Since then, the prgnostic value of ALI for patients with AEG remain unclear.
The major nding of this study is the promising prognostic impact of ALI in patients with AEG after radical resection. Compared with the high-ALI group, patients in the low-ALI group had a worse prognosis.
Multivariate analyses showed that ALI was an independent risk factor for both OS and DFS. In line with the previous studies in several other malignancies, our study indicated ALI as a feasible predictor for both OS and DFS in patients with AEG after radical resection. Even though several nomogram models had been used for the prediction of prognosis of AEG [39,40], this is the rst to incorporate ALI into the survival model of AEG patients after radical resection. The C-index of our propoesd nomogram was 0.699, showing acceptable discrimninatory ability.
In this study, besides ALI, poor differentiation, lymphovascular invasion and tumor size were also recognized as independent prognostic factors for AEG. Poor differentiation already been incorporated into the AJCC classi cation for esophageal cancer as histologic Grade 3. Poor differentiation was also found to be a predictor for deeper tumor invasion and lymph nodes metastasis in gastric cancer [41,42]. Even though not well established in AEG, we believed that the prognostic value of differentiation status deserves more investigation. Lymphovascular invasion, the presence of malignant cell within endotheliallined space, is correlated with the ability of the cancer to metastasize [43]. In line with a previous study, it is also considered as a prognostic factor for AEG [44]. Tumor size had been used as a staging method in several solid tumors like lung and breast cancer. Enlarged tumor mass was correlated with increased risk of other adverse pathological features [45]. In our study, more patients in low-ALI group had enlarged tumor size (≥3.3) (p=0.003). This might be related to a larger tumor mass inducing more intense in ammatory response and consuming more host nutrition. Moreover, as tumor size increased, it might invade beyond the esophagogastric junction, leading to an increased risk of mediastinal metastasis, a unique feature for patients with AEG.
Another important nding of this study was the prognostic impact of ALI in patients with stage III/IVA AEG.

Consent for publication
Written informed consent for publication was obtained from all participants.

Availability of data and materials
The datasets used or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.