Background
Oral squamous cell carcinoma shows a high propensity for invasive growth and cervical lymph node metastasis [
1]. Every year, an estimated 263,000 new cases of oral squamous cell carcinoma (OSCC) are reported worldwide [
2]. Five-year survival rates of patients with OSCC have remained in the vicinity of 50% over the last few decades [
3,
4].
Solid malignant tumors tend to show rapid proliferation and growth in a hypoxic microenvironment [
5]. Hypoxia inducible factor-1 (HIF-1) is an important regulator of cellular response to hypoxia. It is a heterodimer that comprises of HIF-1α and HIF-1β subunits [
5‐
7], the former being its functional subunit [
8]. High expression levels of HIF-1α, a marker of tumor hypoxia, have been reported in patients with OSCC [
9]. Overexpression of HIF-1α has recently been linked with unfavorable prognosis in several types of malignant tumors [
10,
11]. However, definitive evidence of this association is yet to be obtained.
Though a previous meta-analysis revealed an association between HIF-1α expression and poor overall survival in patients with OSCC, the analysis was not comprehensive [
12]. The main prognostic factors for OSCC include tumor size, lymph node status, tumor stage, and differentiation. We performed a literature review to evaluate the association of HIF-1α expression with clinicopathological characteristics and overall survival in patients with OSCC.
Methods
Search strategy
Full-text original research articles relating to expression of HIF-1α in patients with OSCC were identified on a search of PubMed, Web of Science, and EMBASE databases on February 05, 2016. The following keywords were used: “hypoxia inducible factor 1α,” “HIF 1α,” “oral squamous cell carcinoma,” “OSCC,” “mouth neoplasm,” “oral cancer,” “oral carcinoma,” or “oral tumor” with all possible combinations.
The inclusion criteria were as follows: (1) Language of publication: English; (2) studies that investigated the correlation between HIF-1α and prognosis in patients with a histopathological diagnosis of squamous cell carcinoma; (3) HIF-1α protein expression determined by immunohistochemical examination of paraffin-embedded surgical specimens; and (4) adequate data for estimation of hazard ratio and overall survival reported.
Conference abstracts, reviews, letters, case reports, unpublished studies, duplicate publications, and experimental studies were excluded from the analysis.
Two researchers independently reviewed the eligible studies and performed data extraction using a standard format. Data pertaining to the following variables were extracted: name of first author, year of publication, country, sex distribution, sample size, definition of HIF-1α positive, clinicopathological features, and overall survival analysis. Any disagreement on the eligibility criteria was resolved by consensus among the authors.
Methodological assessment
The Newcastle-Ottawa quality assessment scale (NOS) was used for methodological assessment of the included studies [
13]. The maximum score was 9; studies with a total score of ≥6 points were deemed to be of a high quality. Two researchers assessed the scores independently; any disagreements in this respect were resolved by consensus.
Statistical analyses
Statistical analyses were performed using Stata (version 11.0; Stata Corporation, College Station, Texas, USA). Kaplan-Meier curves were plotted using Engauge Digitizer version 4.1 (free software downloaded from
http://sourceforge.net). Hazard ratio at 95% confidence interval was used to assess the relationship between HIF-1α and overall survival. An HR >1 correlated with poor prognosis. HRs were either estimated directly from the reported data using methods described by Parmar et al. [
14] or were derived from the Kaplan-Meier curve. Odds ratio at 95% CI were calculated to evaluate the association of HIF-1α expression with clinicopathological features.
Heterogeneity of the included studies was checked by chi-square-based Q statistical test. A value of I
2 < 50% or P value >0.1 was considered to be indicative of no significant heterogeneity, and a fixed-effects model was used for analysis. Otherwise, the random-effects model was used.
Sensitivity analysis was performed to evaluate the robustness of the individual studies. Begg’s funnel plot and Egger’s regression test were used to evaluate the effect of potential publication bias on the results of this analysis [
15]. All
P values were two-sided;
P < 0.05 was considered as statistically significant.
Discussion
Oral squamous cell carcinoma is a fatal disease; the typically poor prognosis is attributable to its invasive growth. Surgical resection is the mainstay of treatment. Nonetheless, survival rates of patients with OSCC have not shown improvement over the last few decades. More reliable therapeutic biomarkers are urgently needed for the treatment of OSCC. The results of our meta-analysis showed an association of HIF-1α expression levels with clinicopathological characteristics and overall survival of these patients.
HIF-1α degradation is inhibited under hypoxic conditions as well as under some conditions also under normoxic conditions [
31]. HIF-1α activates the transcription of downstream target genes which regulate many biological processes, including angiogenesis, cell proliferation, glucose metabolism, pH regulation, and migration [
12,
32]. Overexpression of HIF-1α has been shown to be associated with tumor cell growth in patients with cervical, endometrial, gastric, colorectal, pancreatic, ovarian, breast, and head/neck carcinomas [
33‐
37]. However, a consensus on this association is yet to be attained.
Zhu et al. reported a significant association of HIF-1α overexpression with tumor stage, histological differentiation, lymph node status, and poor prognosis [
21]. In contrast, Fillies et al. reported significantly lower OS in patients with low expression of HIF-1α [
27]. Therefore, we performed this meta-analysis to determine the clinical and prognostic significance of HIF-1α overexpression in patients with OSCC.
The results of the present meta-analysis showed that increased expression of HIF-1α protein, as detected with IHC, was significantly associated with larger tumor size (T3/T4), more advanced TNM stage (III/IV), and lymph node metastasis, but not with poor differentiation. This suggested an association of HIF-1α overexpression with the biological behavior of OSCC. High expression of HIF-1α was defined based on a combination of the percentage of stained cells and the intensity of staining. Heterogeneity in tumor size, lymph node status, and histology were eliminated. We found this method as being particularly useful to define HIF-1α expression in OSCC cells.
In the present meta-analysis, 14 eligible studies were included for the assessment of the association of HIF-1α and overall survival. The results showed a significant association between HIF-1α overexpression and poor overall survival. On subgroup analyses, a significant association between HIF-1α overexpression and poor prognosis was found only in Asian population (HR = 2.33, 95% CI = 1.72–3.15), without significant heterogeneity in this respect. Therefore, this association is subject to considerable geographical variability. It is pertinent to mention here that geographical variations have been reported in the context of gastric cancer and head-neck cancer [
32,
38]. This phenomenon may be attributable to genetic differences between different ethnic groups [
39]. Further studies are required to clarify this issue.
Some limitations of the present study need to be considered while interpreting our results. Firstly, only studies published in English were included in this analysis, which may have introduced an element of publication bias. Secondly, preoperative chemoradiation may have affected the prognosis of patients with OSCC. Data on preoperative treatment were available only for six studies, which may have contributed to the heterogeneity. Thirdly, the number of eligible studies was relatively small, which limits the statistical powder of our analysis. Finally, HRs were extracted from Kaplan-Meier curves in a few studies, which may not have been entirely accurate, thus contributing to the heterogeneity.
Acknowledgements
We would like to acknowledge and thank the native English-speaking biologists who provided medical writing services on behalf of Medjaden Bioscience Limited.