1. Background
While gastric cancer is the fourth most common cancer globally, Iran is one of the several South Central Asian countries where gastric cancer is the most common cause of cancer-related death (1). Gastric cancer is also the first and third most common cancer in males and females in Iran, respectively (2).
Based on anatomic location, gastric cancers are classified into two groups: Cardia and non-cardia. Several studies have shown that the survival rate of adenocarcinomas in the cardia and upper third of the stomach is noticeably low (3-5).
Unlike anatomic classification, a lot of classification systems have been proposed and used for gastric cancer based on histology. However, the prognostic value of all of these histopathologic classifications is still controversial due to the high heterogeneity among these tumors and the lack of reproducible histopathological criteria (6-8).
Among these classifications, the Laurén classification was established in 1965 and since then has been the most popular system, but its value in the prognostic area is still controversial (9). The World Health Organization (WHO) classification, established 12 years ago, is another popular classification system that includes all types of gastric tumors. Some of the major variants in the latest WHO classification (2019) are tubular adenocarcinoma, parietal cell carcinoma, adenocarcinoma with mixed subtypes, papillary adenocarcinoma, NOS, micropapillary carcinoma, mucoepidermoid carcinoma, mucinous adenocarcinoma, signet ring cell carcinoma, poorly cohesive carcinoma, medullary carcinoma with lymphoid stroma, hepatoid adenocarcinoma, and Paneth cell carcinoma.
Along with these classifications, some other histologic features such as perineural and lymphovascular invasions have been identified by many studies as poor prognostic factors (10-13). However, the clinical stage is still the most important determinant of prognosis and the TNM staging system is the means that is generally applied as the primary method for assessing the extent of disease and correlates with the prognosis of patients (14). This system is also a great help in choosing the most proper management for patients with gastric cancer. In this study we used the 8th edition of this system for our staging purposes.
Although most cases of gastric cancers in Iran are diagnosed at advanced stages (15), the 5-year survival rate is estimated for 25% of the patients (16); however, early detection and accurate diagnosis of gastric cancer can reduce the mortality rate, advanced-stage diseases, need for chemotherapy and/or chemoradiation, and surgical resection of the tumor and lymph nodes (17, 18).
2. Objectives
With regard to the high prevalence of this cancer in Iran and mentioning the importance of an accurate histopathologic evaluation, classification and staging for choosing the most proper management, the purpose of this study was to investigate the correlation of the histologic type of gastric cancer with some important variables. Most of these variables have been identified to have prognostic value including perineural invasion, lymphovascular invasion, surgical margin tumoral involvement, pT, pN, and tumor location and size.
3. Methods
We performed a retrospective cross-sectional study on 100 selected patients with gastric carcinoma referred to Shohadaye Tajrish Hospital from 2017 to 2022.
University's research ethics committee approved this academic work. Informed consent was acquired from all patients.
Inclusion criteria included the presence of primary epithelial neoplasms confirmed by surgical resection specimen and no history of chemoradiotherapy or other synchronous malignancy. Those who refused to allow researchers to use their clinical data or those with incomplete information were excluded from the study.
The clinicopathological data of all patients were obtained from the hospital information system. Variables included demographic data (age and sex), preferred surgical approach, histological classification, primary tumor location, tumor size, tumor grade, surgical margins, lymphovascular invasion, perineural invasion, T stage (AJCC 8th edition), N stage, and lymph node metastasis. The International Union Against Cancer clinicopathological staging (TNM) was applied for tumor evaluation. World Health Organization classification (2019) was used for histological typing and differentiation grading. Histology was exclusively used as the gold standard for the T stage. The gold standard for assessment of lymph node involvement was the histologic examination of tissue obtained in patients undergoing either lymphadenectomy or a peri gastric lymph node biopsy.
By calculating the median tumor size and using it as the standard, we categorized tumors into 2 groups: Small size (≤ 6) and large size (> 6) (19).
Statistical analysis was performed using IBM SPSS Statistics (version 21; IBM Corporation). A chi-square test was used to investigate correlations between the pathological and clinical findings. Data were considered significant if P < 0.05.
4. Results
A total of 100 patients were enrolled in this study. The patients had a mean ± SD age of 62.4 ± 13.44 years old (range 28 - 84 years) and were predominantly men (n = 66, 66 %).
The histopathological findings of the specimens are summarized in Table 1. Tubular carcinomas were most commonly found histologic type (n = 45, 45%); followed by mixed tubular & papillary, poorly cohesive, mucinous 25 (n = 25, 25.0%), poorly cohesive carcinomas (signet-ring cell carcinoma & poorly cohesive carcinoma NOS) (n = 24, 24%), and papillary carcinoma (n = 6, 6%). Other histologic types were not found among the evaluated tumors.
Table 1.
Histopathological Findings
| Tumor Histologic Type | No. (%) |
|---|---|
| Tubular | 45 (45.0) |
| Mixed tubular & papillary or poorly cohesive or mucinous | 25 (25.0) |
| Poorly cohesive (signet-ring cell carcinoma & poorly cohesive carcinoma NOS) | 24 (24.0) |
| Papillary | 6 (6.0) |
4.1. Histologic Type Correlation with Demographics and Significant Features
4.1.1. Age and Sex
There was no statistically significant correlation between the sex of patients and tumor histologic type (Table 2). The average age of patients with poorly cohesive carcinoma (signet-ring cell carcinoma & poorly cohesive carcinoma NOS) was lower than other types. The highest age was related to patients with tubular tumors (P-value = 0.007).
Table 2.
Age and Sex Distribution of Different Histologic Types of Gastric Cancer
| Variables | Total | Tumor Histologic Type | P-Value | |||
|---|---|---|---|---|---|---|
| Tubular | Mixed Tubular & Papillary or Poorly Cohesive or Mucinous | Poorly Cohesive | Papillary | |||
| Gender, No. (%) | 0.559 a | |||||
| M | 66 (66.0) | 31 (68.9) | 18 (72.0) | 13 (54.2) | 4 (66.7) | |
| F | 34 (34.0) | 14 (31.1) | 7 (28.0) | 11 (45.8) | 2 (33.3) | |
| Age | 62.4 ± 13.44 | 66.62 ± 11.83 | 62.68 ± 13.02 | 55.5 ± 14.28 | 57.17 ± 13.04 | 0.007 b |
a Based on chi-square
b Based on ANOVA
4.1.2. Lymphovascular/Perineural Invasion
Lymphovascular invasion was detected in 67 (67%) patients and showed a significant correlation with tumor type (P-value < 0.001) (Table 3). Most of the cases with lymphovascular invasion were patients with tubular type of carcinoma while none of the papillary cases showed this type of invasion.
Perineural invasion was found in 57 (57%) cases and significantly correlated with tumor histologic type (P-value = 0.024) (Tubular Tubular and poorly cohesive carcinomas had similar numbers of cases with perineural invasion, while papillary carcinoma again had the least number of perineural invasions.
Table 3.
Histologic Types of Gastric Cancer and Some Proposed Prognostic Variables a
| Variables | Total | Tumor Histologic Type | P-Value | |||
|---|---|---|---|---|---|---|
| Tubular | Mixed Tubular & Papillary or Poorly Cohesive or Mucinous | Poorly Cohesive | Papillary | |||
| Margin | 0.012 b | |||||
| All free | 88 (88.0) | 44 (97.8) | 22 (88.0) | 16 (66.7) | 6 (100.0) | |
| Radial involve | 7 (7.0) | 1 (2.2) | 1 (4.0) | 5 (20.8) | 0 (0.0) | |
| Proximal involve | 5 (5.0) | 0 (0.0) | 2 (8.0) | 3 (12.5) | 0 (0.0) | |
| LV | < 0.001 b | |||||
| - | 33 (33.0) | 19 (42.2) | 5 (20.0) | 3 (12.5) | 6 (100.0) | |
| + | 67 (67.0) | 26 (57.8) | 20 (80.0) | 21 (87.5) | 0 (0.0) | |
| PNI | 0.024 b | |||||
| - | 43 (43.0) | 26 (57.8) | 9 (36.0) | 5 (20.8) | 3 (50.0) | |
| + | 57 (57.0) | 19 (42.2) | 16 (64.0) | 19 (79.2) | 3 (50.0) | |
| PT | 0.049 b | |||||
| t1a | 7 (7.0) | 3 (6.7) | 2 (8.0) | 0 (0.0) | 2 (33.3) | |
| t2 | 8 (8.0) | 4 (8.9) | 2 (8.0) | 2 (8.3) | 0 (0.0) | |
| t3 | 37 (37.0) | 15 (33.3) | 10 (40.0) | 8 (33.3) | 4 (66.7) | |
| t4 | 15 (15.0) | 10 (22.2) | 3 (12.0) | 2 (8.3) | 0 (0.0) | |
| t4a | 30 (30.0) | 13 (28.9) | 8 (32.0) | 9 (37.5) | 0 (0.0) | |
| t4b | 3 (3.0) | 0 (0.0) | 0 (0.0) | 3 (12.5) | 0 (0.0) | |
| PN | < 0.001 b | |||||
| n0 | 18 (18.0) | 11 (24.4) | 3 (12.0) | 1 (4.2) | 3 (50.0) | |
| n1 | 18 (18.0) | 14 (31.1) | 1 (4.0) | 0 (0.0) | 3 (50.0) | |
| n2 | 39 (39.0) | 17 (37.8) | 9 (36.0) | 13 (54.2) | 0 (0.0) | |
| n3a | 17 (17.0) | 0 (0.0) | 10 (40.0) | 7 (29.2) | 0 (0.0) | |
| n3b | 8 (8.0) | 3 (6.7) | 2 (8.0) | 3 (12.5) | 0 (0.0) | |
a Values are expressed as No. (%)
b Based on Fisher exact test.
4.1.3. Surgical Margin Tumoral Involvement
Microscopic evaluation of surgical margins revealed that the margins of the majority of specimens (n = 88, 88.0%) were free of tumoral lesions, but it also had a significant correlation with the type of tumor (P-value = 0.012) (Table 3).
4.1.4. TNM Staging
The T stage of the majority of patients was T3 (n = 37, 37.0%) and T4a (n = 30, 30.0%) and the N stage showed a significant correlation with the type of tumor (P-value = < 0.001) (Table 3).
4.1.5. Tumor Location
The relationship between tumor type and location is summarized in Table 4. Tumors in the antrum region were more of the tubular type (P-value = 0.033) while tumors of the body region were more of the papillary type (P-value = 0.013).
Table 4.
Tumor Location in Relation to Histological Type of Gastric Cancer a
| Tumor Location | Total b | Tumor Histologic Type | P-Value b | |||
|---|---|---|---|---|---|---|
| Tubular | Mixed Tubular & Papillary or Poorly Cohesive or Mucinous | Poorly Cohesive | Papillary | |||
| Cardia | ||||||
| No | 72 (72.0) | 36 (80.0) | 15 (60.0) | 18 (75.0) | 3 (50.0) | 0.19 |
| Yes | 28 (28.0) | 9 (20.0) | 10 (40.0) | 6 (25.0) | 3 (50.0) | |
| Lesser curvature | ||||||
| No | 59 (59.0) | 25 (55.6) | 18 (72.0) | 13 (54.2) | 3 (50.0) | 0.494 |
| Yes | 41 (41.0) | 20 (44.4) | 7 (28.0) | 11 (45.8) | 3 (50.0) | |
| Antrum | ||||||
| No | 80 (80.0) | 31 (68.9) | 20 (80.0) | 23 (95.8) | 6 (100.0) | 0.033 |
| Yes | 20 (20.0) | 14 (31.1) | 5 (20.0) | 1 (4.2) | 0 (0.0) | |
| Whole gastric wall | ||||||
| No | 88 (88.0) | 42 (93.3) | 21 (84.0) | 19 (79.2) | 6 (100.0) | 0.242 |
| Yes | 12 (12.0) | 3 (6.7) | 4 (16.0) | 5 (20.8) | 0 (0.0) | |
| Pylorus | ||||||
| No | 92 (92.0) | 40 (88.9) | 23 (92.0) | 23 (95.8) | 6 (100.0) | 0.661 |
| Yes | 8 (8.0) | 5 (11.1) | 2 (8.0) | 1 (4.2) | 0 (0.0) | |
| Body | ||||||
| No | 89 (89.0) | 41 (91.1) | 22 (88.0) | 23 (95.8) | 3 (50.0) | 0.013 |
| Yes | 11 (11.0) | 4 (8.9) | 3 (12.0) | 1 (4.2) | 3 (50.0) | |
| Fundus | ||||||
| No | 96 (96.0) | 43 (95.6) | 25 (100.0) | 22 (91.7) | 6 (100.0) | 0.477 |
| Yes | 4 (4.0) | 2 (4.4) | 0 (0.0) | 2 (8.3) | 0 (0.0) | |
a Values are expressed as No. (%)
b Based on Fisher Exact test and chi-square.
4.1.6. Tumor Size
Table 5 shows the relationship between histologic type and tumor dimension.
Table 5.
Tumor Size and Histologic Type of Gastric Cancer
| Tumor Size | Total | Tumor Histologic Type | P-Value | |||
|---|---|---|---|---|---|---|
| Tubular | Mixed Tubular & Papillary or Poorly Cohesive or Mucinous | Poorly Cohesive | Papillary | |||
| Max size, No. (%) | 7.59 ± 6.25 | 5.76 ± 4.63 | 9.05 ± 7.4 | 10.62 ± 6.72 | 3.08 ± 2.27 | 0.002 a |
| ≤ 6 | 57 (57.0) | 30 (66.7) | 14 (56.0) | 8 (33.3) | 5 (83.3) | 0.031 b |
| > 6 | 43 (43.0) | 15 (33.3) | 11 (44.0) | 16 (66.7) | 1 (16.7) | |
a Based on ANOVA.
b Based on Fisher exact test.
4.1.7. Type of Surgery
No significant correlation was found between tumor histologic type and type of surgery (P-value = 0.434) (Table 6).
Table 6.
Type of Surgery and Histologic Type of Gastric Cancer a
| Type of Surgery | Total | Tumor Histologic Type | P-Value b | |||
|---|---|---|---|---|---|---|
| Tubular | Mixed Tubular & Papillary or Poorly Cohesive or Mucinous | Poorly Cohesive | Papillary | |||
| Total gastrectomy | 37 (37.0) | 16 (35.6) | 7 (28.0) | 10 (41.7) | 4 (66.7) | 0.434 |
| Subtotal gastrectomy | 13 (13.0) | 6 (13.3) | 6 (24.0) | 1 (4.2) | 0 (0.0) | |
| Gastrectomy | 19 (19.0) | 9 (20.0) | 5 (20.0) | 4 (16.7) | 1 (16.7) | |
| Partial gastrectomy | 20 (20.0) | 9 (20.0) | 3 (12.0) | 7 (29.2) | 1 (16.7) | |
| Resection | 7 (7.0) | 4 (8.9) | 1 (4.0) | 2 (8.3) | 0 (0.0) | |
| Distal gastrectomy | 4 (4.0) | 1 (2.2) | 3 (12.0) | 0 (0.0) | 0.(0.0) | |
a Values are expressed as No. (%).
b Based on Fisher exact test.
5. Discussion
Gastric cancer is the most common cause of cancer death in Iran (1) and although the overall incidence of gastric cancer has declined over past decades, the cardia group in anatomic classification has shown an increase (20). These statistics illustrate the importance of further investigations in this area, especially in countries like Iran.
In our study males (n = 66, 66.0%) were more affected by this cancer than females (n = 34, 34.0%). This result ties well with previous studies where the male-to-female ratio was 2: 1 (21).
Different histologic types have been shown to have different presentations, for example mixed carcinoma had great size (22) and correlated with nodal involvement (23, 24) In our investigation, we examined tumor size among different tumor types in two categories: Small and large size (Table 4). While both groups showed statistically significant correlation with different histologic types, most of the poorly cohesive tumors had large size, whereas other histologic types were mostly in the small-size group. We also examined nodal involvement according to the updated AJCC/UICC TNM staging system (Table 3). It has to be noted that the prognosis prediction abilities of the 8th and 7th editions of the AJCC/UICC TNM staging system for gastric cancer were analyzed in a recent study and proved that the 8th edition of the AJCC/UICC TNM staging system is superior to the previous edition (25). The majority of the poorly cohesive tumors which as mentioned above were large, had pN2 stage in this study. Lymph node involvement is considered a sensitive prognostic factor in patients with gastric cancer. Accordingly, a higher lymph node ratio that is defined as metastatic nodes to the total number of examined lymph nodes was found to be the predictor of a worse prognosis (26). In assessing the depth of tumor invasion according to the considered staging system, we found that most of evaluated tumors had the pT3 stage. A similar pattern of results was obtained in another study by Zhu et al. (27). They showed that most of their cases (418 of all 932 cases) had pT3 stage. They also compared histologic type and pT stage in pure and mixed histologic types and found significant relationships similar to our study. However, contrary to our findings, they did not show a significant correlation between tumor location and histologic type (P-value = 0.96).
It should be mentioned that one of the most important determinants in treatment of gastric cancer is histologic type (27-29). In histopathologic evaluation we also evaluated two factors proposed as negative prognostic factors, lymphovascular invasion and perineural invasion. Dobrițoiu et al. (30) showed that these factors are more prevalent among mixed type gastric carcinoma. Contrary to their findings, the incidence of lymphovascular and perineural invasions in our study were highest among poorly cohesive groups (87.5% and 79.2%, respectively).
Others have shown that surgical margin tumoral involvement correlates with a poorer prognosis (31). In evaluation of this negative prognostic factor, we found a significant difference between different histologic types (P-value = 0.012). The highest incidence of positive surgical margins (radial and proximal totally) was among poorly cohesive groups, while all the surgical margins in the papillary group were free of tumor.
The main limitation of this study was the lack of follow up of the patients for assessing survival rate.
5.1. Conclusions
Overall, our results demonstrate a correlation between tumor histology and our selected studied variables including perineural invasion, lymphovascular invasion, tumoral involvement of surgical margin, pT, pN tumor location, and size.
