In this review, gastric cancer accounted for 4.5% of all histopathologically-diagnosed malignancies seen during the studied period in our setting. These data are comparable with other African studies which reported the incidence of gastric cancer to range from 1.1% to 6.0% of all cancers [6, 7, 24, 25]. High figures of 16.3 and 15.1% of all malignancies for males and females, respectively, were reported by Kitinya et al. in northeastern Tanzania. Dietary, genetic factors and variation in the infectivity rate of H. pylori may be responsible for this regional variation [7, 9]. Our figure for gastric cancer in this study may actually be underestimated by the retrospective nature of the study. A better picture of the incidence of gastric cancer in this region requires a prospective comprehensive data collection.
In agreement with other studies [26, 27], the peak age incidence of gastric cancer in this study was found to be in the fifth decade of life, which is about a decade or two earlier compared to the findings in developed countries . It is possible that the earlier age occurrence of gastric cancer is related to the life expectancy in the country, rather than any special demographic feature of gastric cancer. Compared with findings in developed countries, gastric cancer in sub-Saharan Africa tends to be very aggressive with short periods of time between the onset of symptoms and diagnosis . It has been reported that the occurrence of gastric cancer at a young age is associated with a worse prognosis [13, 16].
The male predominance demonstrated in this study was in keeping with previous observations reported in studies done elsewhere [26–28]. The exact reason for this male preponderance is not known, although it is possible that estrogen may protect women against the development of this form of cancer .
A strong association with socio-economic status (SES) has been frequently observed, with individuals of lower SES having higher risk. SES is, of course, not a causal factor, but is a surrogate for many other factors, including sanitary and dietary conditions . As reported in other studies done in developing countries [6–8], the majority of patients in this study had low socioeconomic status with poor education and more than three-quarters of them were unemployed. This observation has an implication on accessibility to health care facilities and awareness of the disease.
The etiopathogenesis of gastric cancer in developing countries is of great interest. It is possibly multifactorial and associated with complex interactions. It is, however, very difficult to know the precise roles of the different factors, such as genetic, premalignant lesions, H. pylori infection and diet [9, 24, 30]. The association between chronic H. pylori infection and the development of gastric cancer remains controversial . Several studies have shown significant associations between H. pylori seropositivity and gastric cancer risk [9, 24, 30, 32, 33]. It is, however, not known why some individuals with H. pylori infection develop gastric cancer whereas others do not. The virulence factors of H. pylori have been investigated. There is increasing evidence to suggest that certain H. pylori, containing a gene called CagA, associated with cytotoxin expression, are more strongly associated with gastric cancer. Several studies have suggested that CagA positive H. pylori are more common in patients who develop gastric cancer [32, 33]. Gastric cancer is generally accepted as a multistep-progression disease from chronic gastritis, chronic atrophic gastritis, intestinal metaplasia, dysplasia and, subsequently, to cancer. Infection with H. pylori has been linked to gastric carcinogenesis . It is the main pathogenic factor in the development of chronic atrophic gastritis and intestinal metaplasia . Determination of H. pylori seroprevalence was not performed in this retrospective study, because tests for H. pylori status were not routinely performed in patients with gastric cancer during the study period and, therefore, it was difficult to establish the association between H. pylori infection and gastric cancer.
Prospective studies have demonstrated a significant dose-dependent relationship between smoking and gastric cancer risk [36, 37]. There is little support for an association between alcohol and gastric cancer . In this study, we could not determine the association among gastric cancer and smoking and alcohol.
In the present study, the majority of patients presented late with an advanced stage of cancer (stage III and IV), which is in keeping with other studies in developing countries [6–8]. If gastric cancer is diagnosed at an early stage, patients can have a highly favorable prognosis and avoid extended surgery, which may produce complications, especially in the elderly people. However, early symptoms of gastric cancer are non-specific and vague and, therefore, many people in our area who have dyspeptic symptoms are treated for peptic ulcers regardless of the cause of dyspepsia. Subsequently, some of these patients, whose cause of dyspepsia is cancer, are diagnosed with late-stage gastric cancer or one of its complications. Late presentation in our study may be attributed to lack of awareness of the disease, low standard of education, low socioeconomic status, lack of accessibility to health care facilities and lack of screening programs in this region. As gastric cancer appears to occur at a younger age in our population as compared to the Western world, patients over 40 years of age with vague dyspeptic disorders or a long history of epigastric pain should be recommended for esophagogastroduodenoscopy. If any suspicious lesion is observed, multifocal biopsies should be taken. Of course, in order to achieve this there is a need to have endoscopy equipment available at least at all regional referral hospitals and to train more surgeons and physicians in endoscopy.
This study showed a wide spectrum in the gross and histopathological features. The common anatomical site for gastric cancer in this study was gastric antrum, which is similar to studies done in developing countries [6–8], but at variant with what is obtained in developed countries where gastric cardia is becoming the most common site of gastric cancer . Grossly, according to the Borrmann classification system, the ulcerating type was the most common tumor in this study. Similar macroscopic appearance was reported by Cassell and Robinson . However, our findings did not match with those of Schindler et al., who found infiltrative lesion (linitis plastica) to be the most common type. The most common histopathological type of gastric cancer in this study was adenocarcinoma, accounting for 95.1% of cases, which is consistent with what is reported in the literature [6–8, 10, 20]. Other reported gastric neoplasms are gastric lymphoma, and gastrointestinal stromal types of gastric malignancies, which is similar to the worldwide experience . The differences in prognosis and treatment approaches require the need for distinction and differentiation of tumor types. More than half of the gastric adenocarcinomas in this study were the intestinal type, based on Lauren classification. Compared with the diffuse types, the intestinal type is known to be associated with a better prognosis . The overall five-year survival rate for patients with intestinal-type carcinoma in this study was higher than that of patients with diffuse carcinomas. Similar findings were also reported by Riberio et al., who found the five-year survival rate to be higher in patients with intestinal type than diffuse type cancer.
The development of endoscopic techniques has improved the proportion of gastric cancers detected at an early stage, particularly in Japan, which has the highest incidence of the disease and the most developed programs for screening . Upper gastrointestinal endoscopy, which is an important diagnostic tool in patients with gastric cancer, is not widely available and not easily affordable to most low socioeconomic class patients in Tanzania. Early detection of the tumor is usually possible when all dyspeptic patients had upper gastrointestinal endoscopy as may occur in developed countries, such as Japan and the United State of America .
Lymph node involvement is one of the most important prognostic factors in gastric cancer . Lymph node metastasis at the time of diagnosis in this study was recorded in 31.9% of cases. Lymph node involvement in our series was greater in signet ring cell than in other types of carcinoma. Similar observation was reported by Gürsan et al.. High lymph node metastasis in this study is attributed to the late presentation in the majority of patients and this confirmed the highly metastatic potential of gastric cancer. In the present study, distant metastasis was documented in 29.3% of cases and occurred mainly to the transverse colon, adnexia, peritoneum and the liver. A similar distant metastatic pattern was reported by Alatise et al. in Nigeria. Late presentation in our area in the majority of patients may also be responsible for the high distant metastatic rate.
The treatment of gastric cancer requires a multidisciplinary approach. Treatment modalities of gastric cancer include surgery combined with chemotherapy and radiotherapy given either as neo- or adjuvant therapy . Surgery is and, most probably, will remain the cornerstone of curative management of resectable gastric cancer; however, this benefit is limited to patients who present with early and, perhaps, localized disease . Complete resection of a gastric tumor with resection of adjacent lymph node is the only chance for a cure [46, 47]. However, most of the patients we see in our environment present late with advanced disease at the time of diagnosis, for which only palliative surgery is possible. In this study, only 2.2% of patients had gastric resection with curative intent, 22.4% had gastrectomy and 53.8% of patients underwent gastro-jejunostomy alone due to the advanced nature of the disease. Only a very small proportion of patients in this study had resectable lesions. The low resection rate of gastric cancer with “curative” intent in this study could be explained by the high proportion of patients with advanced gastric cancer at presentation. Our patients tended to present late as evidenced by the facts that there was a long interval between onset of symptoms and presentation. This is similar to findings in previous studies in other developing African and East European countries [47–49]. While surgical resection remains the cornerstone of gastric cancer treatment, the optimum extent of nodal resection remains controversial, with randomized studies failing to show that the D2 procedure improves survival when compared with D1 dissection . In the present study, D2 lymphadenoctomy was carried out on fit patients with locally advanced disease. However, for optimal postoperative recovery and functional outcome, less radical surgery (D1 lymphadenoctomy) was performed on high risk and very old patients and those with wide-spread metastatic disease. In a D2 resection all tumor and N2 lymph nodes are resected, while in a D1 resection only N1 lymph nodes are removed and in a D0 resection only the tumor is removed without the lymph nodes . The scope of lymphadenectomy should be individualized and decided based on an accurate preoperative and intraoperative assessment of the extent of disease and the patient’s fitness. Patients that are fit for surgery should have a D2 lymphadenectomy with preservation of the pancreas and spleen. For more advanced cases and high risk patients, a more limited D1 lymphadenectomy should be carried out. The high rate of D1 lymphadenoctomy in our series can be explained by the fact that the majority of patients in our environment present late with advanced disease at the time of diagnosis and only palliative surgery was possible.
Multimodal treatment involving chemotherapy or radiotherapy, in addition to surgery, is thought to be a promising strategy for improving loco-regional control of gastric cancer . The majority of gastric cancer patients in this study had stage III or IV disease at presentation and were, therefore, candidates for adjuvant therapy. In our series, the use of chemotherapy and radiotherapy was reported in only 22.1% and 5.1% of cases, respectively. Lack of accessibility to radiotherapy and other adjuvant therapies, such as monoclonal antibodies, angiogenic inhibitors and antisense agents, may be responsible for discouraging results and low use of adjuvant therapy in this study. Many studies reported the vital role played by adjuvant chemo radiotherapy [46–48].
In this study, the postoperative complication rate was 37.1%, which is a higher rate than that reported by other authors [51, 52]. Surgical wound infection was the most common complication but was usually superficial and easily controlled by local wound care. Additional operations were required in 15 (6.7%) patients who developed postoperative complications, such as anastomotic leak, intra-abdominal abscesses, peritonitis and intestinal obstruction. These required additional operations that are associated with prolonged hospital stay and the cost of treatment. Improvement of surgical technique is, therefore, crucial to lower the occurrence of these postoperative complications while their prompt recognition and treatment would reduce the attendant mortality. In addition, prior treatment of preoperative comorbidities is essential to the postoperative recovery of patients with gastric cancer. A higher complication rate in this study is attributed to late presentation and delayed definitive treatment. The presence of comorbidities and a poor level of fitness in our patients increase the risk of postoperative complications and lowers patients’ ability to survive major complications when they occur.
Despite considerable improvement in the surgical treatment of gastric cancer, recurrences still constitute the main cause of death in surgical patients [53–55]. Recent series showed overall recurrence rates of 22 to 50% after curative surgery, mostly within two years from the operation [53–56]. In our study, recurrence rate was reported to be 19.4% of cases, attributing this to the presence positive resection margins, the high stage of the tumor and presence of metastasis at the time of diagnosis.
The median duration of hospital stay in our study was 16 days, which is higher than that reported in other studies [51, 52]. This can be explained by the high rate of postoperative complications which required additional operations/care and, subsequently, prolonged hospital stay.
The postoperative mortality following gastric cancer surgery is 1.7 to 16% in Western countries [57–59]. Our overall mortality rate in the present study was 18.1%, a figure that is comparable to 18.6% reported by Johnson et al. in Ethiopia. In the present study, the mortality rate was significantly high in patients with preoperative co-morbidity, high stage and grade of the tumor and those who had metastases.
The prognosis of gastric cancer has remained poor in most developing countries where most patients are already in an advanced stage of the disease at the time of diagnosis, which has been proven both in the present study and in most studies done in developing countries [6, 7, 46–48]. However, when it is diagnosed and treated early, gastric cancer is curable as a five-year survival rate of over 90% has been achieved in Japan . In this study, the overall five-year survival rate of 6.9% is significantly low compared to the five-year survival rates of gastric cancer patients ranging from 68 to 92% reported in developed countries and Japan [13, 47]. The low overall five-year survival rate in the present study may be explained by the fact that most of our patients generally seek medical attention when the disease has reached an advanced stage. Therefore, diagnosis is made when the chance of a full cure is low. The follow-up of patients in this study was generally poor as more than two-third of patients were lost to follow-up by the end of five years.
The potential limitations of this study included the following: first, the fact that information about some patients was incomplete in view of the retrospective nature of the study might have introduced some bias in our findings. Second, we did not determine the association of H. pylori with gastric cancer because of lack of necessary facilities at the study center. Third, this study included patients who were evaluated and treated at a single institution, which may not reflect the whole population in this region, despite the fact that approximately 70% of oncology patients in northwestern Tanzania are managed at our center.
However, despite these limitations, the study has provided local data that can help health care providers in the management of patients with gastric cancer. The challenges identified in the management of gastric cancer in our setting need to be addressed in order to deliver optimal care for these patients.