Abstract
Background: The aim of this study was the analysis of the influence of prognostic factors on short- and long-term outcomes of gastric cancer resection. Patients and Methods: A database of 709 patients who had gastric cancer resection between 2007 and 2015 was compiled. Results: Total gastrectomy (TG) and subtotal proximal gastrectomy (SPG) significantly increased the risk of overall complications (p=0.0015 and 0.0173, respectively) and surgical complications (p=0.0141 and 0.0035, respectively). Moreover the resection of an additional organ was an independent prognostic factor of overall complications (p<0.0001), systemic complications (p=0.0503), surgical complications (p<0.0001) and relaparotomy (p=0.0259). T stage (p<0.0001), N stage (p<0.0001), M stage (p<0.0001) and radical resection (p<0.0001) significantly affected 5-year survival rates. Conclusion: Early diagnosis and radical resection was crucial in 5-year survival rates. However, the type of gastrectomy and the resection of an additional organ were the most important factors in short-term outcomes of treatment for such patients.
Gastric cancer is the second most common cancer of the gastrointestinal tract in the world. In 2020, the incidence of new cases of gastric cancer was estimated at 1,089,103, with a mortality rate of 768,793 people globally. Unlike East Asia, gastric cancer is a relatively rare neoplasm in North America and some highly developed countries in Western Europe. However, even there it is one of the most common causes of death from malignant neoplasms (1). In Poland, the number of deaths from gastric cancer reaches 5000 per year. In 2018, 3155 men and 1745 women died from this in our country (2).
Currently, the majority of publications concern aspects of multimodal therapy with pre- and perioperative chemotherapy, even in the presence of oligometastasis. Nevertheless, surgical resection is still the most effective treatment for such patients, and the principles of surgery of gastric cancer are usually well established (3-5).
Although data from a large number of articles reveals the impact of prognostic factors on the prognosis of patients with gastric cancer resection, survival is significantly improved by the effectiveness of surgical treatment measured by short-term outcomes. Most reports involving prognostic factors for patients after gastric cancer resection focus on the 5-year survival rate. There are no in detail articles on the influence of prognostic factors on short-term outcomes of treatment such as overall complications, systemic complications, surgical complications, relaparotomy and perioperative mortality (6-30). There are many prognostic factors that could affect the short-term outcomes of gastric cancer resection including gender, age, location, histologic type, tumour staging, type of gastrectomy, number of retrieved lymph nodes or resection of an additional organ.
In this study we carried out univariate and multivariate analysis of prognostic factors which, in addition to long-term outcomes (5-year survival), also affected the short-term outcomes of gastric cancer resection. Detailed analysis of prognostic factors made it possible to obtain interesting conclusions about gastric cancer resection, which may significantly influence the optimization of surgical treatment outcomes.
Patients and Methods
Patient population. Between 2007 and 2015, 709 patients had gastric cancer resection at the First Department of General, Oncological, Gastrointestinal and Transplantation Surgery, Jagiellonian University Medical College in Krakow. Their database with clinicopathologic features and surgical characteristic was reviewed retrospectively (Table I).
Diagnosis and clinicopathological features. Routine preoperative diagnostics included gastroscopy during which samples were drawn for histopathological examination, chest X-ray, abdominal ultrasonography and, in selected cases, computed tomography. Gastric cancer was defined as a histologically verified primary adenocarcinoma located in the stomach. Patients with other gastric tumors such as gastric lymphoma, gastrointestinal stromal tumor or recurrence of gastric cancer were excluded. The site of the tumor was defined as the location of the main portion of the neoplasm in the upper, middle or lower third of the stomach. Sometimes at diagnosis, gastric cancer was so advanced that occupied two analyzed locations in the stomach (upper/middle/lower), in these cases the location was classified as ‘other’ (Table I); it was not possible to assume which location was primary. To correctly evaluate the influence of tumor location in the stomach on short- and long-term outcomes of gastric cancer resection in univariate analysis and the logistic regression model, location classified as other was not included. The Lauren classification was used for the histologic evaluation of the tumor (31). The gastric cancer staging system was evaluated by pathological examination of the surgical specimens in accordance with the eighth edition of the TNM staging system of the American Joint Committee on Cancer/Union for International Cancer Control (AJCC/UICC) (32).
Surgical characteristics and treatment. The types of gastrectomy, total/subtotal distal/subtotal proximal and reconstruction options were determined by the location of the tumor, a histopathological examination, and the stage of the disease. However, all resections included in the study were carried out by conventional laparotomy. Lymphadenectomy with the removal of perigastric lymph nodes, lymph nodes around the left gastric artery, common hepatic artery, celiac trunk, splenic artery, hepatoduodenal ligament or additionally paraaortic lymph nodes (D2 or D2+) was standard in cases of radical stomach resections. In some cases, segmental pancreatectomy, spleen and/or bowel resections were performed to obtain potential oncological radicality. Cases with unresectable procedures (gastrojejunostomy, by-pass surgery or explorative laparotomy) without gastric cancer resection were excluded from the study. Some patients with advanced gastric cancer (stage II or higher according to TNM classification) received combined therapy using different chemotherapeutic regimens with fluorouracil, cisplatin with fluorouracil, irinotecan with fluorouracil or etoposide, doxorubicin and cisplatin.
Follow-up. Perioperative mortality was defined as any death during the hospital stay after surgery. After discharge, patients had a followed-up every 3-6 months or shorter intervals whenever justifiable. The dates of death were verified using data obtained from the census registry office. All relevant short-term outcomes of treatment, overall complications, systemic complications, surgical complications, relaparotomy as well perioperative mortality were collected and entered into an electronic database (Table II). In some cases, the same patient manifested both systemic and surgical complications. The 5-year survival was understood as long-term outcomes of treatment after gastric cancer resection (Table II). To each of the parameters of short- (overall complications, systemic complications, surgical complications, relaparotomy, perioperative mortality) and long-term (5-year survival) outcomes of treatment was carried out the univariate and multivariate analysis of affecting prognostic factors. These prognostic factors are presented in Table I as clinicopathologic features, and surgical characteristics.
Statistical methods. To determine the relationship between prognostic factors and the outcomes of treatment after gastric cancer resection, we applied the χ2 test and a logistic regression model. The χ2 test was used in the univariate analysis. In view of the available database of the presence of short-term outcomes of treatment and 5-year survival rates, logistic regression models were used in multivariate analysis, which allowed the generation of a dependency model for a binary dependent variable and one or more predictors (prognostic factors). p<0.05 was considered statistically significant in two-tailed analysis. The calculations were performed with the statistical package STATISTICA v. 13 and StatsDirect v. 3.3.4.
Results
Prognostic factors: clinicopathological features and surgical characteristics. The study involved 479 (67.6%) males and 230 (32.4%) females. The mean age±SD (min-max) was 63.9±12.0 (range=22-89) years, and the incidence of patients older than 70 years was 33.7%. Most tumors were located in the upper third part of the stomach (31.9%). The rates of intestinal and diffuse type of Lauren classification were comparable at 43.6% and 43%, respectively. The proportion of patients with T stage, N stage, M stage according to the eighth edition of TNM (AJCC/UICC) are presented in Table I. The most common type of gastrectomy was total gastrectomy (64.9%) and the most common resected additional organ was the spleen (12.7%). The incidence of patients who had >15 lymph nodes removed was 84.1%. The proportion of patients who underwent radical resection was 59.4% (Table I).
Outcomes of treatment. Postoperative complications are presented in Table II. The overall percentage of complications was 19.6%, while the rates of systemic and surgical complications were 11.6% and 11.3%, respectively. The proportion of relaparotomy was 4.8%. The incidence of perioperative mortality was 3.1%. The 5-year survival rate was 39.1%. Median follow-up was 58.6 (39.5-103.6) months.
Univariate and multivariate analyses. Univariate analysis by the χ2 test revealed potential prognostic factors affecting the short-term outcomes of gastric cancer resection such as overall complications, systemic complications, surgical complications, relaparotomy and perioperative mortality (Table III). The same analysis was used to detect potential prognostic factors of 5-year survival for patients (Table IV). Subsequently, prognostic factors statistically significantly affecting short-term outcomes and 5-year survival in the univariate analysis were assessed with the logistic regression model (Tables V, VI, VII, VIII and IX). Finally, multivariate analysis showed that total gastrectomy [odds ratio (OR)=1.12, 95% confidence interval (Cl)=0.93-1.36, p=0.0015], and subtotal proximal gastrectomy (OR=1.28, 95% Cl=1.04-1.58, p=0.0173) were independent risk factors of overall complications (Table V). The same types of gastrectomy significantly increased the incidence of surgical complications (total gastrectomy: OR=1.19, 95% Cl 0.94-1.49, p=0.0141; subtotal proximal gastrectomy: OR=1.16; 95% Cl=0.91-1.47, p=0.0035) (Table VII). Resection of an additional organ significantly increased rates of overall complications (OR=2.56, 95% Cl=1.79-3.65, p<0.0001), systemic complications (OR=1.52, 95% Cl=0.99-.32, p=0.0503), surgical complications (OR=3.04; 95% Cl=2.07-4.46, p<0.0001), and relaparotomy (OR=1.9, 95% Cl=1.08-3.36, p=0.0259) (Tables V, VI, VII and VIII). The presence of distant metastases (M1 stage) was identified as an independent prognostic factor for a higher perioperative mortality proportion (OR=1.36; 95% Cl=0.39-4.66, p=0.0422) (Table VIII). The multivariate analysis showed that T stage (OR=0.52, 95% Cl=0.41-0.66, p<0.0001), N stage (OR=0.48, 95% Cl=0.39-0.6, p<0.0001), M stage (OR=0.01, 95% Cl=0.001-0.08, p<0.0001), location in the middle third part of the stomach (OR=1.57, 95% Cl=0.73-3.37, p=0,0239) and radical resection (OR=2.84, 95% Cl=2.01-10.95, p<0.0001) significantly affected 5-year survival (Table IX).
Discussion
Clinicopathological features Gender. Gastric cancer is twice as common in men as it is in women (33). Gender is reported as being not an independent prognostic factor of survival (6, 7). However, in a large database - which is presented in the study Li et al. - white female patients as well as male Asian patients showed an independent prognostic factor for better survival in gastric cancer. According to the same database, the prognosis among black female patients between 1973 and 2003 was better than that of men, whereas sex-related survival was not significant in the black race between 2004 and 2013 (8). In Poland, epidemiological data on gastric cancer has stabilized in the last few years (5, 34, 35). Multivariate analysis in our study showed that gender was not a significant prognostic factor that affected overall complications, systemic complications, surgical complications, relaparotomy as well perioperative mortality and 5-year survival.
Age. The incidence of gastric cancer increases with age and most often occurs in the 7th decade of life (33). It is usually acknowledged that in gastric cancer surgery older patients demonstrate increased risk of complications and a poorer prognosis (6, 7, 9-13). Although most articles on the impact of age on prognosis of patients with gastric cancer recognize old age as more than 60 years, the 7th decade of life is usually a period of relatively good health nowadays, therefore, in our study, the border age affecting the outcomes of treatment was fixed at 70 years (6, 7, 10, 11). In the present study, the logistic regression model revealed that even at an age >70, this was not an independent prognostic factor influencing the short- and long-term outcomes of treatment for patients with gastric cancer resections. Similarly, Kulig et al. did not find differences in complications and perioperative mortality rates, except for a higher incidence of cardiopulmonary complications in older patients undergoing gastric cancer resection (6.6% vs. 12.3%). The median survival of patients was not significantly longer in younger (30.8 months) vs. older (24.1 months) patients (p=0.056) (14). Another report from Poland stratified patients into four groups according to their age: 29-0 years (group I), 51-65 years (group II), 66-75 years (group III) and 76-92 years (group IV). The middle-aged patients (group II) had significantly better 3-year survival than either the youngest (group I) or the oldest patients (group IV) (15). Nakamura et al. reported that youth is a prognostic factor of better survival for early gastric cancer, however for advanced gastric cancer, younger patients are prone to more advanced stages of TNM and a poorer prognosis (36). Saito et al. concluded that elderly patients undergo less aggressive surgical resections such as extended lymph node excision or multi-organ resections and present a poorer prognosis (37).
Tumor location. Incidence of gastric cancer situated in the lower or middle third part of the stomach have been steadily going down. Incidence in the upper third part is stable or higher, which could be caused by an inappropriate diet, or obesity and reflux disease in patients with gastric cancer (17-1). However, the occurrence of this prognostic factor depends on geographical location and race, for example, in Asia the proportion of tumors in the lower third part of the stomach exceeds 50% (6).
Tan et al., Zhao et al., and Lee at al. indicated that the location in the upper third part of the stomach was not an independent prognostic factor of poor survival (6, 25, 26). In this study, the most common location was the upper third part of the stomach (31.9%). In multivariate analysis, location was not a prognostic factor that affected overall complications, systemic and surgical complications, relaparotomy as well perioperative mortality, but the location in the middle third part of the stomach had influence on improved 5-year survival rates (p=0.0239). Conclusions from most reports are consistent with the results of our study. Petrelli et al. reviewed 50 studies including 128,268 patients with gastric cancer, they indicated that a location in the upper third part of the stomach entailed increased risk of tumor-related death and was an important prognostic factor of poorer prognosis (22). Matsuda et al. and Yu et al. Also found the same, i.e. that patients who underwent resection for gastric cancer with proximal cancer had a poorer 5-year survival (23, 24).
Lauren types and tumor staging. In line with the results of our study, the proportion of intestinal type gastric cancers has decreased, while the incidence of diffuse type has recently increased (38-42). Tan et al. noticed that the histological type in gastric cancer patients does not effect prognosis (6). However, most reports have shown that patients with intestinal type gastric cancer demonstrate significantly better survival than diffuse type and mixed type (27-29). In our study, the Lauren type was not an independent prognostic factor that affected the short-term outcomes of treatment and the 5-year survival rates.
As we all know, early stages (according to TNM classification) have an influence on better prognosis after gastric cancer resection (6, 7, 10, 25, 26, 30). Although in studies from the Netherlands and France where stage IV remained stable over the last few years, the staging of TNM of gastric cancer at diagnosis significantly decreased over the last decades (38, 43-47). In the present study, early gastric cancer occurred in 9.9% of cases, while the proportion of patients with stage IV cancer compromised 40.3%. In multivariate analysis, T stage (p<0.0001), N stage (p<0.0001) and M stage (p<0.0001) influenced the 5-year survival and the presence of distant metastases (M1 stage) influenced perioperative mortality (p=0.0422). Nevertheless, TNM stages was not independent prognostic factors affecting short-term outcomes of treatment such as overall complications, systemic complications, surgical complications and relaparotomy.
Surgical characteristics
Type of resection. Some articles reported that the 5-year survival rates for patients with gastric cancer reveal no significant changes depend on type of gastrectomy at the same stage of the disease (48-50). Studies carried out by the Polish Gastric Cancer Study Group have shown that total gastrectomy improves the 5-year survival in patients with stage IIIA according to the TNM classification (51). In multivariate analysis from the present study, the type of gastrectomy (total, subtotal, proximal or subtotal distal) statistically significantly influenced the overall and surgical complications, but not the systemic complications, relaparotomy as well perioperative mortality and 5-year survival rate.
In Asia, the proportion of patients with gastric cancer who undergo radical resection exceeds 90% (6, 52). In this study, this type of resection occurred in 59.4% of cases. The fact is that opposite to palliative, radical resection is an independent prognostic factor of better survival which is documented in patients in our study (52, 53-55). An interesting conclusion from the present study was that, the type of resection (radical, palliative) was not an independent risk factor in the logistic regression model influencing the short-term outcomes of treatment such as overall complications, systemic complications, surgical complications, relaparotomy and perioperative mortality.
Number of retrieved lymph nodes. In the study of Tan et al., the percentage of patients with gastric cancer with the removal of >15 lymph nodes was raised from 38.6% to 81.7% over a 30-year period (6). Although some reports indicated higher rates of postoperative complications after extended lymphadenectomy (D2, D2+ or D3) compared to D1, it is a routine procedure for patients with potentially radical gastric cancer resection, as it raises the oncological radicality and results in better survival for such patients (56-8). Marubini et al. did not observe any differences in perioperative mortality depending on the extent of lymphadenectomy (58). In the present study, dissection of >15 lymph nodes was a significant prognostic factor of 5-year survival in univariate (p=0.003) but not in multivariate analysis (p=0.691). Interestingly, this was not an independent prognostic factor influencing overall, systemic and surgical complications, relaparotomy and perioperative mortality in the logistic regression model.
Resection of an additional organ. Combined resection of the neoplasm-related affected organs such as the spleen, pancreas or bowel, like an extended lymphadenectomy, may cause increased oncological radicality and better survival in patients of gastric cancer resection (52, 53, 55, 59). However, multiple organ resection is associated with relatively high rates of complications and perioperative mortality (53, 59, 60). Furthermore, Mita et al. indicated that extended multi-organ resection could be beneficial only if radical surgery is carried out in such patients (54). Wang et al. reported that patients with locally advanced gastric cancer (T4-TNM) extending to the bowel might benefit from radical resection with acceptable rates of postoperative complications (52). Opposite to Wang et al., in multivariate analysis from our study, the resection of an additional organ was an independent risk factor of overall complications (p<0.0001), systemic (p=0.0503) and surgical complications (p<0.0001) as well relaparotomy (p=0.0259), and did not significantly influence 5-year survival rates (p=0.5373).
Study limitations. A limitation was that this was a retrospective, single-centre research project. The prognostic factors which were chosen for the analysis, for which data in the patient database of the patients was the most complete. Nevertheless, an analysis of all prognostic factors affecting overall complications, systemic and surgical complications, relaparotomy, as well perioperative mortality and 5-year survival of gastric cancer resection exceeds the size of a discussion for a single scientific article.
In fact, conclusions concerning 5-year survival did not differ from the previous papers, regarding the effect of prognostic factors on outcomes of gastric cancer resection. That said, early diagnosis followed by a radical resection is crucial in gastric cancer patients. However, the study database was from recent years (the final follow-up date was December 18, 2020). Moreover, analysis of prognostic factors after gastric cancer resection apart from the 5-year survival details their influence in the short-term treatment outcomes, such as overall complications, systemic and surgical complications, relaparotomy as well as perioperative mortality, which is an added benefit of our study.
Conclusion
When possible, a subtotal distal gastrectomy should be the preferred type of gastric cancer resection, as it causes significantly less incidence of overall and surgical complications compared to total and subtotal proximal gastrectomy. Resection of an additional organ such as the spleen, pancreas and/or the bowel is not recommended, because of the significantly increased rates of overall complications, systemic and surgical, as well as relaparotomy. In patients with resection of an additional organ, gastric cancer was too advanced in order for this to make a difference in the 5-year survival. Furthermore, the presence of distant metastases significantly increased the proportion of periopertive mortality. A location in the middle third part of the stomach was demonstrated as an independent prognostic factor of improved 5-year survival. However, early recognition with radical resection is still crucial for 5-year survival of patients after gastric cancer resection.
Footnotes
This article is freely accessible online.
Authors’ Contributions
Piotr Kulig: concept of the study, collection and analysis of patients database, statistical analysis, database results analysis, preparing articles for discussion, writing the manuscript. Przemysław Nowakowski: collection and analysis of patients database, statistical analysis, database results analysis, preparing articles for the discussion. Marek Sierzęga: concept of the study, collection and analysis of patients database, database results analysis, critical review. Radosław Pach: collection and analysis of patients database, statistical analysis, database results analysis. Oliwia Majewska: collection and analysis of patients database, preparing articles for the discussion. Anna Markiewicz: collection and analysis of patients database, statistical analysis, database results analysis. Piotr Kołodziejczyk: concept of the study, database results analysis. Jan Kulig: concept of the study, database results analysis, critical review. Piotr Richter: concept of the study, database results analysis, critical review.
Conflicts of Interest
The Authors declare that they have no conflicts of interest.
- Received April 26, 2021.
- Revision received May 25, 2021.
- Accepted May 27, 2021.
- Copyright © 2021 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.