Clinicopathological and prognostic features of early gastric cancer patients after surgery: a retrospective study


 Background: Despite the decline in the incidence of gastric cancer, the incidence of early gastric cancer has increased. Hence, understanding the clinicopathological and prognostic features of early gastric cancers could help us understand the development of gastric cancer and improve the prognosis of early gastric cancer. Methods: A total of 244 patients diagnosed with early gastric cancer after surgery at Xiangya Hospital Central South University were retrospectively analyzed. Results: General data showed that in patients with a mean age of 54.30±10.68 years (M:F = 1.6:1), the median tumor size was 2.203±1.245 cm. A total of 15.6% of patients had lymph node metastasis. By univariate analysis, the longest diameter of the tumor, T stage, total number of dissected lymph nodes, number of metastatic lymph nodes, metastatic-to-total dissected lymph node (LN) ratio, vascular invasion, NLRc, and MLRc were associated with disease-free survival; tumor size, invasive depths, vascular invasion, NLRc, MLRc, NWRc and LWRc were associated with lymph node metastasis. Additionally, the longest diameter of tumor and total number of dissected lymph nodes were independent factors for early gastric cancer patients; tumor size, invasive depths, vascular invasion and NLRc were independent risk factors for lymph node metastasis in EGC. Conclusion: The longest diameter of the tumor and total number of dissected LNs were independent prognostic factors for EGC patients. Additionally, the longest diameter of the tumor, tumor invasive depths, vascular invasion and NLRc were the independent risk factors for lymph node metastasis in EGC patients.


Introduction
In recent years, the incidence rates of gastric cancer (GC) have been declining steadily worldwide [1]. However, GC is still the fth most common malignant tumor and the third leading cause of cancer-related mortality globally, with an estimated 1,033,701 new cases and 782,654 deaths in 2018 [2,3]. This translates into a high fatality to new case rate of 75%, which is much higher than that of other prevalent solid cancers, including breast and prostate cancers [2]. The incidence of GC varies regionally, with approximately 70% occurring in Eastern Asia, Central and Eastern Europe and South America [4]. Approximately 50% of patients are diagnosed with advanced disease, with a 5-year cancer-speci c survival rate lower than 30% [5]. However, the prognosis of GC markedly depends on the stage at diagnosis. In Western countries, including Europe and the United States, the 5-year survival rate does not exceed 25% [6], but in Japan, the 5-year cancer-speci c survival rate is 90% with the diagnosis of early GC of up to 50% [7].
Similar to other common cancers, GC rarely occurs in children and young persons, with a peak incidence ranging from 55 to 80 years, and gastric cancer rates are twice as high in men as in women [8]. GC is divided into early gastric cancer (EGC) and advanced gastric cancer, with signi cant differences in recurrence and cancer-related survival rates. EGC is de ned as a carcinoma invading the mucosa and/or submucosa with or without lymph node metastases [9]. Despite a decline in the incidence of total gastric cancer and the aging of the population, the incidence of early gastric cancer has been steadily increasing since the late 1980s [10].
Recent studies on EGC data remain incomplete and con icting; with the growing incidence of early gastric cancer, it is essential to focus on the clinical and prognostic features of early gastric cancer, as they are controversial. We retrospectively examined patients with early gastric cancer in our center, aiming to achieve a better understanding of clinicopathologic features, the development of gastric cancer and prognostic factors in early gastric cancer and to improve the outcomes of gastric cancer.

Patient selection
Data of patients who were diagnosed with EGC from January 2013 to December 2018 after surgery at Xiangya Hospital Central South University were collected through a pathology information system. The eligibility criteria in this study included (1) patients who were histopathologically diagnosed with early gastric cancer (carcinoma invading the mucosa and/or submucosa) after surgery; (2) no other synchronous malignant neoplasia; (3) no perioperative or postoperative chemotherapy or radiotherapy applied.
This retrospective study was supported by the Medical Ethical Committee of Xiangya Hospital, Central South University, and due to the retrospective nature of the study, informed consent was waived.

Data of routine blood collection
All patients' preoperative and postoperative peripheral routine blood tests were drawn within 7 days before the operation and 7 days after the operation. iNLR (initial preoperative neutrophil-to-lymphocyte ratio), iMLR(initial preoperative monocyte-to-lymphocyte ratio), iNWR (initial preoperative neutrophil-towhite blood cell ratio), iLWR (initial preoperative lymphocyte-to-white blood cell ratio), iPLR (initial preoperative platelet-to-lymphocyte ratio), iMWR (initial preoperative monocyte-to-white blood cell ratio),

Follow-up data
Patients who underwent surgery were followed every 3 months for the rst 2 years, every 6 months between 2 and 5 years, and then every 12 months thereafter. We supplemented the follow-ups with our outpatient system or communicated with patients by telephone. We examined the postoperative EGC patients by chest scan, abdominal CT scan, magnetic resonance imaging (MRI) scan, abdominal ultrasound scan, endoscopy (including gastroscopy and colonoscopy) and positron emission tomography-computerized tomography (PET-CT) (if necessary) to evaluate tumor recurrence and distant metastasis. All patients were monitored until October 1, 2019, or death. Metastatic recurrence revealed by imaging and death were regarded as the endpoint events.

X-tile and statistical analyses
Patient characteristics between each group were compared by chi-square test. The values for the iNLR, iMLR, iPLR, iNWR, iLWR, iMWR, NLRc, MLRc, PLRc, NWRc, LWRc, MWRc, longest diameter of the tumor, total number of dissected lymph nodes (LNs), number of metastatic LNs, and ratio of metastatic-to-total dissected LNs were determined by receiver operating characteristic (ROC) curve, based on the results of the ROC curve, X-tile analyses were conducted to assess the optimal cutoff values based on the integral optic density (IOD). The disease-free survival (DFS) and overall survival (OS) were calculated by the Kaplan-Meier method, and differences between Kaplan-Meier curves were investigated by the log-rank test. Signi cant predictors for survival identi ed by univariate analysis were further assessed by multivariate analysis using a multivariate Cox proportional hazards regression model (forward stepwise method, conditional likelihood ratio). The signi cant predictors for lymph node metastasis by univariate analysis were further assessed by logistic regression. Statistical analyses were performed using SPSS V22.0 (SPSS Inc., USA).P value was considered statistically signi cant below the 5% level.

Results
Clinicopathologic features for EGC patients For our research, 244 patients were eventually enrolled as our study group (Fig. 1). As shown in Table 1, the EGC group comprised 149 males and 95 females, and the male-to-female ratio was approximately 1.6:1, with a mean age of 54.30 ± 10.68 years, ranging from 24 to 82 years. All patients underwent D2 lymphadenectomy and curative gastrectomy (68.9% underwent the open method and 31.3% underwent the laparoscopy method). Most EGC patients underwent Billroth II reconstruction (71.7%), the tumor was most commonly located in the antrum and pylorus (61.9%), and the average size of the longest diameter of the tumor was 22.03 ± 12.45 mm, ranging from 3 to 85 mm. The most common pathological features included moderate-poor differentiation (58.2%), T1b (50.4%), N0 (84.4%), non-perineurium invasion invasion(99.2%), nonvascular invasion (90.6%), HER2-negative immunohistochemistry (81.6%) and TMN stage Ia (76.6%). More details are shown in Table 1.    Fig. 2f-i). Moreover, the best cutoff values were 2.0 cm for the diameter of the tumor, 13 for the total number of dissected LNs, 1 for the number of metastatic LNs, and 6% for the ratio of metastatic-to-total dissected LNs determined by X-tile program (Fig. 3c-f).

Risk factors for lymph node(LN) metastasis in EGC patient
Since the LN metastasis plays a critical role for the selection of clinical treatment (EMR/ESD or surgery) in EGC patients, it is essential to identify the risk factors for LN metastasis in EGC patients. ROC curve was conducted to calculate the best cutoff points of each viriate for the LN metastasis. Results showed that the optimal cutoff points of NLRc, MLRc, NWRc, LWRc, and tumor maximal dimension were 5.29, 0.49, 0.70, 0.13 and 1.9, respectively(Table S1, Figure S1). Next, we investigated the risk factors for lymph node metastasis according to univariate analysis. As shown in Table 3

Prognostic features for EGC
We further investigated the relationship between the clinicopathologic features and prognosis of EGC patients according to univariate analysis. As shown in Table 5, we found the longest diameter of the tumor > 2 cm, T1b, total number of dissected LNs ≤ 13, number of metastatic LNs ≥ 1, ratio of metastaticto-total dissected LNs ≥ 6%, vascular invasion, NLRc > 7.80, MLRc > 1.40 were both associated with recurrence/metastasis and overall survival in EGC patients. Next, Cox proportional hazard models were performed to identify independent prognostic factors. Those identi ed as signi cant prognostic factors by univariate analysis were further assessed by a multivariate analysis. The multivariate analysis results showed that the longest diameter of the tumor (≤ 2/ 2 cm) (RR: 7.404, P = 0.001) and total number of dissected LNs (≤ 13/>13) (RR: 0.289, P = 0.015) were independent risk factors for recurrence/metastasis in EGC patients (Table 6), otherwise, longest diameter of tumor (≤ 2/ 2 cm) (RR: 4.109, P = 0.033) was the independent risk factors for overall survival in EGC patients( Table 7). The DFS and OS for the signi cant prognostic factors identi ed by the multivariate analysis is presented in Fig. 5.   OS: overall survival, EGC: early gastric cancer, CI: con dence interval, RR: relative risk, LN: lymph node, cm: centimetre, NLRc = postoperative neutrophil-to-lymphocyte ratio change, MLRc = postoperative monocyte-to-lymphocyte ratio change.

Discussion
Gastric cancer (GC) has been a critical health burden due to its high morbidity and mortality worldwide, with approximately 1,033,701 new cases and 782,654 deaths in 2018 [2]. Gastric cancer can be divided into early gastric cancer (EGC) and advanced gastric cancer, with signi cant differences in prognosis.
Moreover, despite the declining incidence of GC and due to the advancement of early diagnostic techniques and health examinations, the incidence of EGC seems to have steadily increased since the late 1980s [10][11]. Many studies have indicated clinicopathologic features, multiple genes and molecules in advanced gastric cancer; however, there are limited studies focusing on the clinicopathologic and prognostic features of EGC.
Our retrospective analysis of 244 patients with EGC found that the male-to-female ratio was 1.6:1, indicated that gastric cancer is a sex-related carcinoma regardless of whether early gastric cancer or overall gastric cancer patients were included, similar results were indicated by McGuire, S [8]. We further investigated the prognostic features in EGC and found no signi cant differences in age, sex, surgical methods, reconstruction methods, tumor location, tumor differentiation, perineurium invasion or immunohistochemistry score for HER2. As our previous study indicated that laparoscopic distal gastrectomy achieves the same degree of radicality and short-term prognosis as open distal gastrectomy [12], we now veri ed no signi cant difference between laparoscopic and open surgical methods in terms of long-term survival of GC patients. Additionally, we found that HER2 expression had no prognostic in uence in EGC. However, some retrospective studies have suggested that HER2 positivity is the second worst prognostic factor [13,14], whereas other studies have suggested that HER2 status has no relationship with short-and long-term survival according to univariate and multivariate analyses [15].
We found that the longest diameter of the tumor = 2 cm were the best cutoff points for the prognosis factor in our large-scale group of EGC patients. Furthermore, the longest diameter of the tumor (≤ 2/ 2 cm) and total number of dissected LNs(≤ 13/ 13) were independent risk factors for EGC patients according to multivariate analysis, with RRs of 7.404 (P = 0.001) and 0.289 (P = 0.015), respectively. Li Y reported that lesions of over 2 cm might be more likely to have lymph node metastasis in EGC [17]. Similar results have been veri ed by Li H et al. [18]. Our research is the rst to identify the relationship between the longest diameter of the tumor and disease-free survival rate by using X-tile software and ROC curves.
The radical dissection of lymph nodes is a highly effective procedure in gastric cancer to improve the prognosis of GC and limit LN metastasis. Our study found that LN metastasis was 13.9% in EGC, which is higher than the LN metastasis rate of 2.5-8.6% in Japan [19].  [20]; a similar suggestion was given by Korean and Japanese investigators [21,22]. Our research also showed that the number of dissected LNs lower than 13 was associated with a poor prognosis of EGC patients, and this might be because limited LN dissection might lead to residual cancer, increasing the risk of recurrence or metastasis.
Since lymph node metastasis remains a critical role in the therapeutic approach (EMR/ESD vs surgery) for EGC patients, it is important to identify the risk factors for EGC patients [23]. Our study demonstrated that longest diameter of the tumor (≤ 1.9/ 1.9 cm), T stage (Tis + T1a/T1b), vascular invasion (yes/no), and NLRc (≤ 5.29/ 5.29) were the independent risk factors for LN metastasis. Similarly, Japanese Gastric Cancer Association (JGCA) points that nonulcerated, well-differentiated and lesions diameter less than 2 cm are the indications for EMR/EDS approach in EGC patients [24]. Additionally, our study also found that the invasive depth of EGC(T stage) and vascular invasion were also associated with lymph node metastasis, Chu YN et al also found that submucosal invasion depth and LVI were the predictive factors for lymph node metastasis [25], hence, it is truely for the application of ultrasound gastroscope to identify the invasive depths in EGC patients, which could not only measured the lesion size, but also measured the invasive depths of EGC tumors [26]. In collaboration with previous studies, our research indicated that tumor size ≤ 1.9 cm, ultrasound gastroscope presents the non-submucosa invasion and non-vascular invasion might be the indications for ESD/EMR in EGC patients.

Conclusion
The longest diameter of the tumor and total number of dissected LNs were independent prognostic factors for EGC patients. Additionally, the longest diameter of the tumor, tumor invasive depths, vascular invasion and NLRc were the independent risk factors for lymph node metastasis in EGC patients.

Declarations
Ethics approval and consent to participate The study pro le of patients enrolled in our study.