Oncologic outcomes of patients with resectable primary gastric squamous cell carcinoma: a systematic review

Introduction

Primary gastric squamous cell carcinoma (PGSCC) is an uncommon type of gastric malignancy. Although the true prevalence of this disease cannot be accurately measured because of its rarity, it is roughly estimated to be around 0.04-0.5% of all gastric neoplasms [1-3]. Pure PGSCC closely resembles squamous cell carcinomas (SCCs) encountered in other organs and can be distinctively identified by the criteria established by Boswell and Helwig [4], i.e., the formation of keratinizing cell masses with pearl formation, the mosaic pattern of cell arrangement, the presence of intercellular bridges and the identification of keratin within neoplastic cells. In 2011, the Japanese Gastric Cancer Association published their updated diagnostic criteria, which postulate that: (i) PGSCC should originate entirely from the gastric mucosa; and (ii) all tumor cells must be squamous neoplastic cells, without a glandular component.

The exact pathogenetic mechanism leading to the development of SCC in the stomach is yet to be determined. Boswell and Helwig [4] hypothesized that gastric SCC arises from metaplastic foci thought to exist as a response to injurious insults, such as corrosive agent ingestion or peptic ulcer disease, or conceivably from chronic exposure to tobacco and alcohol [5,6]. Although this explanation is the most plausible one, it has also been suggested that ectopic squamous cell nests encountered in the stomach may also arise de novo from gastric pluripotent stem cells, which undergo malignant transformation and thus give rise to PGSCCs [7]. Attempts at identifying a causal link between human papilloma virus infection and squamous cell gastric cancer have been unsuccessful to this date [8,9].

Taking into account the histologic resemblance of PGSCC to other types of SCC, a similarly aggressive tumor behavior is to be expected. Reports suggest that the prognosis of this particular subtype of gastric cancer is poor [10,11], with expected survival rates of 17-50%, depending on stage and surgical resectability [3,11]. Although surgery combined with chemotherapy and/or radiation therapy is pursued in the treatment of patients with PGSCC, it should be emphasized that reliable long-term data on its effectiveness are still lacking. The aim of the present systematic review was to identify potential prognostic factors related to the oncologic survival of patients with resectable PGSCC and to gauge their long-term survival rates.

Materials and methods

Results

The initial database search yielded 991 results in total, of which 24 met the inclusion/exclusion criteria described above. The PRISMA flowchart of study selection is depicted in Fig. 1. The entire study pool was comprised of 23 case reports and 1 retrospective cohort study, totaling 38 patients with PSGCC [1,2,19-28,9,29-32,11,13-18] (the details of the included studies are available as Supplementary Table 1). Included studies were published between 2004 and 2019.

Figure 1 PRISMA flowchart of study selection

Overall, 38 patients (36 male, 2 female), with mean age 61.2 years, were identified. The majority of evaluated tumors were located in the gastric body (n=13, 34.3%), followed by the cardia (n=10, 26.3%), the antrum (n=8, 21%), and the fundus (7, 18.4%). With regard to the endoscopic appearance of the PSGCC, most lesions were either polypoid (Borrmann type I, n=14, 40%) or fungating (Borrmann type II, n=13, 37.3%), with a minority being ulcerated (Borrmann type III, n=7, 20%) or diffuse (Borrmann type IV, n=1 (2.7%). Preoperative staging data were available for a total of 22 patients, all of whom underwent preoperative endoscopy and computed tomography imaging. Additional imaging with positron emission tomography (PET) was performed in 4 (19.3%) patients and endoscopic ultrasound (EUS) in a single patient (4.5% of the entire patient pool). Most tumors were high stage at the time of diagnosis, with the T4 stage predominating in the patient pool (n=15, 50%), along with a high percentage of lymphatic spread (N positive tumors, n=15, 47%). Only one patient received neoadjuvant chemotherapy, while 20 patients (60%) received postoperative adjuvant chemotherapy, 3 underwent radiation therapy (10%) and 12 patients (40%) did not receive any adjuvant treatment. Patients were followed-up for an average of 30.7±14 months. During the follow up, 11 cases of death and 9 cases of disease recurrence occurred. Table 1 summarizes the above findings. The adjuvant and neoadjuvant regimens utilized are detailed in Supplementary Table 2.

Table 1 Characteristics of the pooled patient cohort

Survival analysis

The cumulative 3- and 5-year survival probabilities were estimated to be 62.2% and 51.9%, respectively. Regarding disease-free survival, the 3-year probability for being disease free was calculated to be 30.8%. The Kaplan-Meier survival curves for overall and disease-free survivals are depicted in Fig. 2 and 3. In total, 21 patients remained alive at the end of their follow up, with 8 of them (38%) being disease-free.

Figure 2 Kaplan-Meier curves of (A) overall survival and (B) disease-free survival

Figure 3 Kaplan-Meier curves of overall survival and disease-free survival stratified by tumor stage (panels A and B) and by lymph node status (panels C and D)

During univariate analysis of factors prognostic for overall survival, only the T4-stage and lymphatic spread were statistically significant predictors of the risk of death from PGSCC (P=0.019 and 0.05, respectively). When introduced into a multivariate model, however, neither of these variables exhibited statistical significance (Table 2).

Table 2 Univariate and multivariate analyses of prognostic factors for overall survival

Discussion

The paucity of existing literature on PGSCC is a major impediment in the process of understanding this disease and hinders our ability to formulate effective evidence-guided treatment plans. To overcome this obstacle, the present review is the first attempt to summarize all existing literature regarding the outcomes following surgery for PGSCC.

According to our results, PGSCC shows a greater preponderance for male patients, with an estimated 18:1 male: female ratio, and a mean patient age of 61.2 years. These findings suggest that PGSCC is a disease of the 7^th decade of life that exhibits an incidence pattern resembling that of other squamous neoplasms of the aerodigestive tract. Although the present patient sample is small, it does suggest that PGSCC shares, at least partly, the same risk factors as esophageal SCC.

The study by Akce et al [3] is, to this date, the largest report on patients with PGSCC, having accumulated 836 patients through national database searching. Their findings with regard to PGSCC demographics are in line with those obtained by the present systematic review, further revealing that the vast majority of gastric SCC cases (73.4%) are not amenable to surgery, and in fact the investigators observed that squamous histology is independently associated with poor patient survival.

Whatever the case may be, it must be acknowledged that PGSCC grows indolently and presents with nonspecific clinical symptoms, not unlike those associated with other gastric malignancies. This is in fact supported by the results of the systematic review, which demonstrated that most PGSCCs are large, averaging 7.1 cm in size, and present at a locally advanced stage (T3-T4, node positive) in their vast majority. Endoscopy is unable to distinguish between PGSCC and gastric adenocarcinoma, because of the macroscopic resemblance between the 2 types of lesion (Borrmann type I being the most common presentation for PGSCC); however, the histopathologic positive staining for p63 and cytokeratins 5, 6, 7 and 20 is unequivocally associated with PGSCC [1].

The treatment paradigm for this rare disease is the same as for other gastric malignancies. Nonetheless, it is uncertain whether oncologic outcomes in cases of localized disease are the same as in localized, resectable adenocarcinomas. Given the paucity of relevant literature on PGSCCs, we constructed a patient cohort from isolated cases to extrapolate survival data following surgical resection of these lesions. Kaplan-Meier curve analysis demonstrated 3- and 5-year overall survival to be approximately 62.2% and 51.9%, respectively, with even lower 3-year disease-free survival at 38% following curative intent surgery with or without adjuvant therapy. Akce et al [3], in their study, quote inferior 5-year overall survival rates of 14.7%, in a patient population that was largely managed non-operatively, mainly because of the presence of metastatic disease at presentation. These observations suggest that PSGCC amenable to surgical resection is likely to be associated with enhanced survival and metastatic disease is possibly an important factor influencing long-term oncologic outcomes.

Furthermore, T4 and node-positive stages were found to be significant negative predictors of survival in univariate but not in multivariate analysis. It is possible that this may be due to a type II statistical error related to the small accrued patient sample, although it is also possible that advanced T-stage correlates strongly with lymphatic spread and therefore cannot be assessed independently. Notwithstanding this limitation, survival following resection of non-metastatic PGSCC appears to be roughly equal to that following resection of gastric adenocarcinoma [33], and therefore surgery should be aggressively pursued whenever disease clearance is technically feasible, irrespective of T and N stage, in the effort to optimize oncologic outcomes.

As far as adjuvant therapy is concerned, no standardized treatment regimens have been established as yet, with most authors reporting the use of 5-FU or cisplatin-based drug combinations [14,16,27]. In our analysis, the use of adjuvant therapy was not associated with prolonged survival, nevertheless, given the lack of data, many factors were left unaccounted for, most importantly the performance status and comorbidities of included patients. In addition, radiotherapy was scarcely used: only in 2 patients of the entire cohort [20,23], for whom no follow-up data are available, and therefore no assumptions on its usefulness in the setting of PGSCC can be made. Taking into consideration the radiosensitivity of other squamous neoplasms of the digestive tract, we speculate that radiation therapy, in addition to chemotherapy, may be more effective in improving disease-free and overall survivals.

It must be emphasized that the present review inherently bears significant limitations that preclude the generalizability of the obtained results. More significantly, the small patient sample drawn from isolated case reports or small case series is thus highly subject to publication and indication biases. Moreover, a large number of patient and disease-related confounding factors were missing from the extracted reports (most importantly, patient comorbidities and overall status, lack of standardized adjuvant treatment protocols and limited follow-up length), and this, when taken together with the possibility of selection bias, reduces the likelihood that the presented outcomes are an accurate cross-sectional portrayal of PGSCC. Instead, they are a reflection of all existing literature, aiming to provide a rough assessment of survival following curative-intent surgical management of PGSCC.

In conclusion, PGSCC is commonly diagnosed in advanced stages, yet survival appears to be on a par with that observed for the most commonly encountered gastric adenocarcinomas. Surgery, when applicable, is the mainstay of treatment and, when coupled with adjuvant therapy, may lead to survival rates of up to 50%. More frequent and thorough reporting of the outcomes of PGSCC patients is still required to elucidate the role of chemoradiotherapy in these particular neoplasms.