Preoperative Inammatory Response to Esophageal Cancer Is More Reliable Prognostic Factor in Patients Received Esophagectomy

Background: Inammation and nutrition are closely related to the progression of gastrointestinal malignancies. We aimed to explore the potential of preoperative inammation-based or nutrition-based biomarkers as predictors of survival in patients with resectable esophageal squamous cell carcinoma (ESCC) using multivariate Cox analysis. Methods: We included 122 patients with resectable ESCC (stages I–IV) in the study. We assessed the inammation-based modied Glasgow prognostic score (mGPS), nutrition-based modied controlling nutritional status (mCONUT) score, CRP(C-reactive protein),serum albumin, lymphocyte counts, and total cholesterol. The relationships of these biomarkers with overall survival (OS) and recurrence-free survival (RFS) were evaluated. Three Cox model were performed for single parameters(CRP, albumin, lymphocyte, total cholesterol), for mCONUT and mGPS,and for clinicopathological factors. Results: The cut-off values for CRP, albumin, and mCONUT measured using receiver operating characteristic (ROC) curves were 0.3, 3.5, and 4, respectively. Patients with high mGPS and high mCONUT scores were signicantly associated with shorter OS and RFS (p < 0.05).Multivariate Cox analysis showed that mGPS,pStage,tumor location were independent prognostic factors both FRS and OS. Also, Cox analysis for single parameters showed that preoperative CRP, lymphocyte counts were independent prognostic factors for RFS and albumin was prognostic factor for OS. Conclusions: Preoperative inammation-based mGPS is most reliable independent prognostic factor in patients with resectable ESCC. Suppression of preoperative inammation can be improved nutritional status and may improve the prognosis in these patients.


Introduction
Esophageal cancer is the ninth most common cancer worldwide, with 604100 new cases the standard treatment for patients with locoregionally con ned esophageal carcinoma. Despite curative surgery, a high recurrence rate is observed in patients with esophageal cancer (1)(2)(3). Perioperative chemotherapy and chemoradiotherapy have been developed to improve the survival of patients undergoing esophagectomy (4). Approximately 57% -85% of patients with esophageal cancer have over 10% unintentional weight loss before treatment, which is associated with poor survival (5,6).
An acute phase response (APR) leads to both in ammatory and neoplastic conditions (7), following the increased production of proin ammatory cytokines including IL-6 and TNF-α from tumor cells. These cytokines act on the liver to increase hepatic synthesis of the acute phase protein, C-reactive protein (CRP). Elevated CRP levels are associated with bodyweight loss and poor prognosis in gastrointestinal malignancies (8,9).The modi ed Glasgow prognostic score (mGPS), an in ammation-based biomarker consisted of serum CRP and albumin levels, and it has been validated as an independent prognostic factor for various malignancies (10). On the other hand, the controlling nutritional status (CONUT) score is nutrition-based biomarker consisted of serum albumin, total cholesterol, and total lymphocyte counts, but not included in ammatory markers (11). It has been associated with postoperative complications and prognosis in several cancers (12)(13)(14)(15). As prognostic markers associated with resectable esophageal cancer, the mGPS score was controversial for predicting survival, regardless of different cut-off values for each marker, while the CONUT score is a favorable predictor of survival, (16,17). It is not clear whether preoperative these immunological, nutritional factors, in ammatory responses as a single marker, or their comprehensive markers determine the prognosis of patients with resectable esophageal cancer.
In the present study, we retrospectively reviewed patients who underwent curative surgery for esophageal squamous cell carcinoma (ESCC) in our department and attempted to identify the highest priority prognostic factors for ESCC patients after esophagectomy by evaluating these preoperative parameters using multivariate analysis.

Patients And Methods
We retrospectively reviewed 122 consecutive patients with biopsy-proven invasive ESCC who underwent esophagectomy with two or three-eld lymphadenectomies at the Kawasaki Medical School Hospital between January 2011 and July 2020. Preoperative tumor stages were de ned by upper GI ,endoscopy with biopsy, enhanced computed tomography, and positron emission tomography-computed tomography (PET-CT). Tumor stages were determined based on the eighth edition of the UICC classi cation of malignant tumors (18). Patients with R1/2 resections were excluded, and only those with R0 resections were evaluated. Neoadjuvant therapy consisting of 5-FU/cisplatin or 5FU/cisplatin/docetaxel was administered to seven cStageII/III patients after 2017. Seven patients of cStageII/III were received neoadjuvant therapy and esophagectomy was performed 3-7 weeks later after completion of chemotherapy. Three surgical procedures were performed. Ivor Lewis esophagectomy with two-or threeeld lymph node dissections was performed before 2013. Esophageal reconstruction was performed using a gastric tube in the retrosternal or posterior mediastinal roots. However, cases after 2013 utilized thoracoscopic and laparoscopic-assisted approaches. Trans-hiatal esophagectomy with middle, lower mediastinal, and abdominal lymphadenectomy were also performed in the relatively early stages of lower esophageal cancer. Postoperative complications were evaluated according to the Clavien-Dindo classi cation grade 2 or higher (19). This study was approved by the Institutional Review Board of Kawasaki Medical School (Authorization No: 3382-2). Data were collected from the medical records at Kawasaki Medical School Hospital, and details of this retrospective study are published on the hospital home page. All patients were followed up regularly until July 2020 or until death. Informed consent was obtained from all the patients.
Nutritional, immunological, and in ammatory response assessment Peripheral blood samples were collected for routine laboratory measurements of albumin and CRP within 1-4 weeks before surgery. While CRP was measured using anti-human CRP mouse monoclonal antibody sensitive latex (SEKISUI, Tokyo, Japan), serum albumin was measured using bromocresol purple (SEKISUI, Tokyo, Japan). Cholesterol was measured using the HMMPS method (FUJIFILM, Tokyo, Japan). The mGPS was determined as previously described based on the CRP and serum albumin levels (20). A time-dependent receiver operating characteristic (ROC) curve was calculated to evaluate the sensitivity and speci city of CRP and albumin in predicting 3-year recurrence-free survival (RFS). The Youden's index was estimated to determine the optimal cut-off for CRP and serum albumin ( Fig. 1A and 1B). Based on these analyses, the optimal cut-off values for CRP and albumin were 0.3 mg/dl and 3.5 g/dl, respectively. The mGPS for esophageal cancer was calculated as follows: elevated CRP (> 0. 3 mg/dl) with hypoalbuminemia (< 3.5 g/dl) was allocated a score of 2 (mGPS 2), while the presence of only one of these abnormalities was allocated a score of 1 (mGPS1). The absence of both these abnormalities was allocated a score of 0 (mGPS0). The mCONUT was calculated as per previously published criteria (11). As shown in Table 1, minimal changes were observed in serum albumin values.
The mCONUT score, body mass index (BMI,kg/m 2 ) were evaluated. ROC curves estimated the cut-off values for these parameters for predicting 3-year RFS.

Statistical analysis
The chi-square or Fisher's exact t-test was used for categorical variables. The Mann-Whitney U test was used for continuous variables. When comparing three or more groups, the Holm method was used.
Overall survival (OS) was de ned as the time between surgery and the patient's death or when the nal information on vital status was available. RFS was de ned as the time between surgery and cancer recurrence, death, or when the nal information on vital status was available. The Kaplan-Meier method estimated survivals, and groups were compared using the two-sided log-rank test. Univariate and multivariate analyses for OS and RFS were performed using the Cox proportional hazard regression model, and survival with a 95% con dence interval (CI) was determined. Results were considered signi cant at p < 0.05. Three Cox models were used for nutritional ,in ammatory and clinicopathological parameters, factors in the model were selected based on the following criteria: Age and gender was basically included, and (i) the number of explanatory variables were approximately 1/10 the number of event occurrences, (ii) factors dependent on each other were not entered, (iii) useful independent prognostic factors were selected from previously published data.
Statistical analyses were performed using JMP (version 14; SAS, Tokyo, Japan) and R version 3.

Preoperative mGPS values for predicting OS and RFS after esophagectomy
Kaplan Meier survival curves for DFS and OS in the different mGPS groups were analyzed. A high mGPS score was signi cantly associated with an unfavorable OS (p = 0.00357) and RFS (p = 0.00173).

Preoperative mCONUT values for predicting OS and RFS after esophagectomy
There were no patients in the severe mCONUT group, and therefore, we compared the three groups,  (Table 4). Cox proportional hazard model for single nutritional and in ammatory covariates showed preoperative CRP and Lymphocyte counts were independent prognostic factors for DFS, and preoperative albumin level was independent prognostic factor for OS(Table5). As the comprehensive biomarkers,Cox analysis showed that only mGPS emerged as independent prognostic factors for OS and RFS(Table6). Cox analysis for clinicopathological covariates and mGPS showed that pStages, tumor locations, and mGPS were independent prognostic factors for both RFS and OS(Table7).

Discussion
This is the rst multivariate analysis of whether the preoperative in ammatory-based biomarker (mGPS) or the nutritional-based biomarker( mCONUT) can be most reliable prognostic factors in patients with resectable ESCC. We have demonstrated that the mGPS score was an independent prognostic factor for both OS and RFS in Cox analysis. Namely, in ammatory-based comprehensive biomarker represented by CRP and albumin predict postoperative cancer recurrence and survival. Another Cox analysis showed that preoperative CRP predicted the cancer recurrence. As the cut off value of CRP level in GPS, O'Gorman et al. set it at 1 mg/dL (21). In the present study, using the ROC curve to estimate RFS, we set the cut-off value at 0.3 mg/dL, which was consistent with the Japanese standard CRP value before 2019 (22). As per the patient care and health information from Mayo Clinic, the normal CRP level is < 10 mg/L, and > 10 mg/L is a sign of serious infection, trauma, or chronic disease (23). Obesity is also associated with elevated CRP levels, probably due to the abundance of proin ammatory mediators such as TNF-α and IL-6 in adipose tissues (24).The standard CRP values in Japan were inconsistent with those in other countries. Therefore, the standard CRP value of ≥ 1 mg/dl is not suitable for Japanese patients with a mean BMI of < 25 kg/m 2 (25). A meta-analysis showed that preoperative high CRP levels were associated with poor OS in esophageal SCC patients received surgery(26), but there was little reports association between preoperative CRP and RFS regarding esophagectomy. In the present study, preoperative high CRP level can predict recurrence of esophageal cancer after esophagectomy, this might be the result of the preoperative high CRP status persisting during the perioperative period and continued to activation of cancer cells (27). Preoperative low serum albumin level as an indicator of nutritional status was associated with poor OS in Cox analysis. It was possibly that preoperative albumin level also affected the postoperative nutritional status, and it affected the response to treatment and increased the risk of chemotherapy induced toxicity at the time of recurrences(28).

The GPS score was rst introduced by Forrest et al. in 2003 as a prognostic marker for inoperable non-
small cell lung cancer [20]. There are seven retrospective reports before 2020 on the association between mGPS scores and esophageal surgery, all from Asian countries, including Japan and China (29)(30)(31)(32)(33)(34)(35). The mGPS scores for prognostic factors in esophageal cancer patients after esophagectomy are controversial, might be the different cut-off value of CRP. In the present study, we found that the mCONUT score could not be a prognostic factor for OS and RFS in Cox analysis. The parameter common to both the mCONUT score and mGPS is serum albumin. The major difference between mGPS and mCOUNT is whether the CRP value is included. As mentioned above, we showed that preoperative CRP level was important biomarker for predicting cancer recurrence after esophagectomy, and albumin level was independent prognostic factor for OS in Cox analysis. In general, CRP and albumin might be linked together, and hypoalbuminemia re ects the presence of systemic in ammation (36, 37). This phenomenon was also recognized in this study (Figure4). Scot et al. reported an increase in the systemic in ammatory response associated with weight loss, reduced albumin concentration, reduced performance status, and reduced survival in patients with inoperable non-small-cell lung cancer (37). Cancer-induced cachexia, a guideline for the oncologist, states that proin ammatory cytokines such as IL-1,6 and TNFα are released as a local response to the tumor cells. These cytokines act on the liver and increase hepatic synthesis of acute-phase proteins such as CRP and brinogen, inducing muscle proteolysis, fat lipolysis, and cancer-related cachexia(38). Thus, pretreatment positive CRP levels lead to worse outcomes in cancer patients. As a nutritional intervention for cancer-induced cachexia, positive CRP levels may prevent the improvement of nutritional status. Nutritional interventions for patients with cancer are controversial. Recent ESPEN(European Society for Clinical Nutrition and Metabolism) guidelines showed that preoperative nutritional intervention for patients of severe malnutrition received major surgery was effective to reduced postoperative surgical site infection. But there was no description of the relationship between nutritional intervention and prognosis (39). On the other hand, Gullett et al. showed that nutritional intervention does not recommend in patients undergoing major cancer surgery for routine use (40). They also classi ed the wasting syndromes into three categories(starvation, cachexia and sarcopenia) in cancer patients and reported that nutritional intervention might be effect in these patients with starvation (38). Based on these ndings, it seems reasonable that patients with improved mGPS score during NAC(Neoadjuvant chemotherapy) had a better prognosis, although the number of cases was small, as reported by Otowa et al. (31). To improve survival in resectable esophageal cancer patients, preoperative chemotherapy or chemoradiotherapy that results in reduced CRP levels may be effective and improve the patient's nutritional status. Unfortunately, we could not evaluated the NAC patients because of small number of patients in this study.
This study has some limitations. First, it was a retrospective study and included data from a single institution. Large retrospective cohort studies might be necessary for the future to accurately evaluate the role of preoperative in ammatory, nutritional, and immunological factors as predictors of survival in patients with resectable esophageal cancer.
In conclusion, this study suggests that preoperative mGPS score is an independent prognostic factor in patients with resectable esophageal cancer. CRP and albumin levels are simple and easy to measure in clinical practice. More effective preoperative therapies and nutritional interventions which lead to negative CRP levels might be necessary to improve survival in these patients.

6, Availability of data and materials
The datasets generated and/or analysed during the current study are not publicly available because permission of our hospital and university were not given, but are available from the corresponding author on reasonable request.
Due to technical limitations, tables 1 to 7 xlsx are only available as a download in the Supplemental Files section. Figure 1 Receiver operating characteristics (ROC) curves Shown are the ROC curves for (A) CRP (mg/dl) with a cut-off value of 0.3mg/dl; sensitivity, 0.467; speci city, 0.812, and AUC, 0.643, and (B) albumin (g/dl) with a cut off value of 3.5g/dl; sensitivity, 0.292; speci city, 0.905; and AUC, 0.589. Kaplan Meier survival curves in ESCC patients after esophagectomy based on mCONUT scores.

Figures
(A) The ve-and three-year OS were 66% and 61.3%, respectively, in the low mCONUT group and 43.8%,43.8% in the high mCONUT group. The low mCONUT group showed signi cantly better OS than the high mCONUT group (p < 0.0349).
(B) The low mCONUT group showed signi cantly better RFS than the high mCONUT group (p < 0.0351). The ve-and three-year RFS were 63.2% and 60.6%, respectively, for the low mCONUT group and 40.8% Preoperative serum albumin level and in ammatory response.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. Table1.xlsx