Prognostic value of SMAD4 in resectable pancreatic cancer

This meta-analysis was conducted according to the Preferred Reporting Items for Systematics Reviews and Meta-Analyses guidelines [12]. PubMed, Embase, and Web of Science were comprehensively searched up to December 31, 2020. Studies were selected using the following search terms: “SMAD4 or DPC4,” “pancreatic cancer,” “resection,” “recurrence,” and “overall survival.” The articles chosen through this search process and the references reviewed were also manually searched for additional studies for inclusion.

The inclusion criteria were: 1) to study patients with PC, 2) to provide OS or recurrence data to evaluate the role of SMAD4 gene expression in the prognosis of PC patients, 3) to provide the hazard ratio (HR) with 95% confidence interval (CI), or to calculate these statistics in the data, 4) to divide SMAD4 gene expression into “high” and “low,” or “positive” and “negative”, 5) publication of the full text in English. The exclusion criteria were: 1) data from review, animal, or cell line studies; 2) no information on survival; and 3) publication in any language other than English. In cases of multiple publications on the same population, only the most recent or the most complete study was included in the analysis.

Two investigators independently reviewed each eligible article and extracted information from all publications meeting the inclusion criteria. The following information was collected from each study: the first author‘s name, publication year, country of study, number of patients, average or median age of patients, histology, disease stage, follow-up time, detection method, antibody used and its dilution, positive cut-off value, and OS data or recurrence data. OS was defined as the interval between initial diagnosis and death. Disease recurrence was divided into local recurrence and distant recurrence. The inconsistency in the research process was solved through debates and consultations. Each eligible study was evaluated using the Newcastle-Ottawa Quality Assessment Scale [13]. The scale comprises eight methodological items under three main dimensions: selection (4 items, 1 per item), comparability (1 item, up to 2 points), and outcome (3 items, 1 per item). The item scores are added to quantitatively compare the research quality. A higher score indicates higher quality. Score inconsistency was solved by discussion and negotiation.

The HR and 95% CI were used to estimate the effect of SMAD4 gene expression deficiency on the prognosis of patients with PC. In multivariate analysis, the Cox regression model was used to calculate HRs (95% CIs). If HR (95% CI) was not directly provided in the literature, the authors were contacted to obtain more information, or this factor was calculated by the method provided by Tierney [14]. The homogeneity test of the effect was performed by χ² test, and p ≤ 0.05 indicated that the difference was statistically significant. When the homogeneity assumption was not rejected, the set effect of the results was estimated by the fixed effect model. When the reverse held, the random effect model was also calculated. A sensitivity analysis was performed by deleting one study at a time to estimate the stability of the results. Publication bias was assessed using the Begg's and Egger ‘s test. All analyses were performed using Stata 12.0.

We searched articles from PubMed, Web of Science, and Embase, and found 46 potential related articles. After careful reading of titles, abstracts, tables, figures, and key data, 12 eligible articles [15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26] were included for meta-analysis (Figure 1).

Fig. 1

The characteristics of these studies are shown in Table 1 and Table 2. The expression of SMAD4 in all studies was evaluated by immunohistochemistry. The cutoff values of SMAD4 staining negative and positive were different in different studies. In 9 studies, positive results were determined by visible staining, while in other studies, different scores were used as thresholds. The included studies were of high quality.

Nine studies [15, 16, 17, 18, 19, 20, 21, 22, 23] with a total of 1676 patients were qualified to estimate the correlation between SMAD4 gene expression deficiency and OS after resection. Realizing the possible confounding factors and significant heterogeneity between studies (I² = 68.9 %, P = 0.001), adjusted by age, tumor size, differentiation, stage, lymph node status, and grade, the combined HR (95% CI) of the loss of SMAD4 gene expression for OS was 1.38 (0.98–1.81) (Figure 2).

Fig. 2

Five studies [15, 17, 21, 22, 23] comparing the loss of SMAD4 gene expression and recurrence were eligible in this meta-analysis. Regarding local recurrence, there were 5 relevant studies with a total of 664 patients. The heterogeneity was not statistically significant, with I 2=81.8% and P=0. The pooled OR (95% CI) of the loss of SMAD4 expression for local recurrence was 0.97 (0.52–1.80, P=0.914, Figure 3A). Moreover, distant recurrence was mentioned in 4 studies [17, 21, 22, 23] with a total of 558 patients. There was no significant heterogeneity with I 2=17.2% and P=0.305. The loss of SMAD4 gene expression was associated with a distant recurrence pattern (OR=1.36, 95% Cl [1.08 – 1.70], P=0.008, Figure 3B).

Fig. 3

For the loss of SMAD4 gene expression and OS, we assessed publication bias using Begg ‘s funnel plot and Egger ‘s test (Figure 4). For our analysis, no funnel plot asymmetry was found. Moreover, neither Begg ‘s test nor Egger ‘s test showed significant funnel plot asymmetry (Begg ‘s test P=1.0, Egger ‘s test P=0.987). As for SMAD4 gene expression loss and recurrence, due to less included studies, we did not conduct publication bias assessment.

Fig. 4

We used the leave-one-out method for sensitivity analysis, excluding one study each time to determine whether there was a single study affecting the results (Figure 5). But the results are stable and not significantly affected by any single study in our analysis.

Fig. 5