FDG PET correlates weakly with HIF-1 α expression in solid tumors: a meta-analysis

Abstract

Background

Hypoxia-inducible factor (HIF)-1α plays a key role in hypoxic adaptation of tumor cells. Overexpression of HIF-1α is associated with tumor aggressiveness and worse prognosis in several malignancies. Presumably, expression of HIF-1a may be reflected by positron emission tomography with 2-deoxy-2 [fluorine-18] fluoro-D-glucose (18F-FDG PET). There are inconsistent data about relationships between FDG PET and HIF-1α.

Purpose

To provide evident data about associations between maximum standardized uptake value (SUV_max) and HIF-1α expression in solid tumors.

Material and Methods

MEDLINE, SCOPUS, and EMBASE databases were screened for relationships between SUV and HIF-1α up to August 2019. Overall, 21 studies with 1154 patients were identified. The following data were extracted from the literature: authors; year of publication; number of patients; and correlation coefficients.

Results

Correlation coefficients between SUV_max and HIF-1α were in the range of −0.51–0.71. The pooled correlation coefficient was 0.27 (95% confidence interval [CI] = 0.14–0.41). Furthermore, correlation coefficients for some tumor entities were calculated. For this sub-analysis, data for primary tumors with >2 reports were included. The calculated correlation coefficients in the analyzed subgroups were as follows: head and neck squamous cell carcinoma: ρ = 0.25 (95% CI = 0.07–0.42); non-small lung cell cancer: ρ = 0.27 (95% CI = −0.14–0.67); uterine cervical cancer: ρ = −0.09 (95% CI = −0.89–0.71); thymic tumors: ρ = 0.39 (95% CI = 0.04–0.58).

Conclusion

SUV_max of FDG PET correlated weakly with expression of HIF-1α both in overall sample and tumor subgroups. Therefore, FDG PET cannot be used for prediction of hypoxia in clinical practice.

Keywords

FDG PET hypoxia-inducible factor-1α HIF-1α

Introduction

Positron emission tomography (PET) with 2-deoxy-2 [fluorine-18] fluoro-D-glucose (18F-FDG PET) is well established for diagnosing and staging in different malignant tumors. Besides diagnostic potential, 18F-FDG PET parameters can also predict histologic and biologic characteristics such as tumor stage and grade (1 –3). For example, in breast cancer, invasive carcinomas have higher uptake than lower-grade tumors (1). Furthermore, PET can also be used for monitoring of treatment response (4). Finally, PET can predict prognosis in several tumors. So far, it has been shown that high metabolic activity determined by the standardized uptake value (SUV) was associated with poor survival in squamous cell carcinoma of the head and neck region (5).

These effects may be related to possible associations between PET parameters, particularly SUV_max, with underlying histopathology. In fact, some reports confirmed this hypothesis (6 –8). For example, Cochet et al. (8) suggested that SUV_max correlated strong with expression of proliferation marker Ki 67 in breast cancer.

Besides proliferation activity, other histopathological biomarkers play an important role in different malignancies. One of them is hypoxia-inducible factor (HIF)-1α. According to the literature, hypoxia-inducible factors, particularly HIF-1α, play a key role in the adaptation of tumor cells to hypoxic and nutrient-deprived conditions by upregulating the transcription of several pro-oncogenic genes (9). Overexpression of HIF-1α is significantly associated with increase of mortality risk and worse prognosis in several malignancies (10 –12).

Previously, some authors analyzed associations between PET parameters and expression of HIF-1α in several tumors with controversial results (13 –20). While some authors identified significant correlations between the parameters, others did not (13,18,19). Furthermore, most reports investigated small samples of patients.

The aim of the present meta-analysis was to provide evident data regarding possible relationships between SUV retrieved from FDG PET and HIF-1α.

Material and Methods

Data acquisition

The endpoint of the present meta-analysis was the assumption that SUV_max derived from FDG PET may have strong correlations with expression of HIF-1α.

For this analysis, MEDLINE, EMBASE, and SCOPUS databases were screened for associations between PET (SUV_max) and expression of HIF-1α up to July 2019. The strategy of data acquisition is shown in Fig. 1.

Fig. 1.

Flow chart of the data acquisition.

The following search words were used: PET (or) positron emission tomography (or) SUV (or) standardized uptake value AND HIF-1α (or) HIF1α (or) HIF 1α (or) HIF-1 alpha (or) HIF1 alpha (or) hypoxia-inducible factor (or) hypoxia inducible factor. Secondary references were also recruited. The primary search identified 873 records. After exclusion of duplicate articles (n = 739), abstracts of the remaining 134 articles were checked. Exclusion criteria were: review articles, experimental animals and in vitro studies, case reports, and non-English publications. Overall, 93 articles were excluded after checking of abstracts. Thereafter, full texts of the remaining 41 articles were retrieved and analyzed. Inclusion criteria were: original articles; studies investigating humans; reported correlation coefficients between SUV_max derived from FDG PET and expression of HIF-1α; in the English language. Articles that did not contain correlation coefficients between PET/SUV and HIF-1α were also excluded (n = 20). Therefore, 21 studies were included into the present analysis (13 –33). The following data were extracted from the literature: authors; year of publication; diagnosis; number of patients; and correlation coefficients. In all studies, Spearman’s correlation coefficient for correlations between SUV_max and HIF-1α was reported. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (PRISMA) was used for the research (34).

Meta-analysis

First, the methodological quality of the acquired 21 studies was checked by one observer (AS) using the Quality Assessment of Diagnostic Studies (QUADAS 2) instrument (35). Fig. 2 shows the results of QUADAS proving. Of the involved studies, 5 (24%) were prospective, 8 (38%) retrospective, and in 8 (38%) studies the design was unclear.

Fig. 2.

Methodological quality of the involved 21 studies according to the QUADAS 2 criteria.

Second, the meta-analysis was undertaken by using RevMan 5.3 (Computer program, version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014). Heterogeneity was calculated by means of the inconsistency index I² (36,37).

Third, DerSimonian and Laird random-effects models with inverse-variance weights were performed without any further correction (3).

Results

Overall, the acquired 21 studies comprised 1154 patients with different malignant tumors (Table 1). They were predominantly non-small cell lung cancer (NSCLC; 28.3%) and head and neck squamous cell carcinomas (HNSCC; 10.8%).

Table 1.

Included studies and primary tumors.

Tumors	Studies (n)	Patients
NSCLC	3	588 (51.0)
HNSCC	8	236 (20.5)
Thymic tumors	2	82 (7.1)
Uterine cervical cancer	2	78 (6.8)
Gastric cancer	1	50 (4.3)
Pulmonary NET	1	34 (2.9)
Endometrial cancer	1	29 (2.5)
Mesothelioma	1	21 (1.8)
Non-thymic mediastinal neoplasms	1	21 (1.8)
Pulmonary pleomorphic carcinoma	1	15 (1.3)
Total	21	1154

Values are given as n (%).

HNSCC, head and neck squamous cell carcinoma; NET, neuroendocrine tumor; NSCLC, non-small cell lung cancer.

The correlation coefficients between SUV_max and HIF-1α were in the range of −0.51–0.71 (Fig. 3). The pooled correlation coefficient was 0.27 (95% confidence interval [CI] = 0.14–0.41).

Fig. 3.

Forest plots of correlation coefficients between SUV_max derived from 18F-FDG PET and expression of HIF-1α in all involved studies/malignancies.

On the next step, correlation coefficients for every tumor entity were calculated. For this sub-analysis, data for primary tumors with more than two reports were included.

The calculated correlation coefficients in the analyzed subgroups were as follows: HNSCC: ρ = 0.25 (95% CI = 0.07–0.42); NSCLC: ρ = 0.27 (95% CI = −0.14–0.67); uterine cervical cancer: ρ = −0.09 (95% CI = −0.89–0.71); thymic tumors: ρ = 0.39 (95% CI = 0.04–0.58) (Fig. 4).

Fig. 4.

Forest plots of correlation coefficients between SUV_max derived from derived from 18F-FDG PET and expression of HIF-1α in different primary tumors.

Discussion

It is well-known that uncontrolled growth of cancer cells results in imbalance between oxygen demand and angiogenesis. This causes tumor hypoxia, which provokes activation of cellular adaptation mechanisms. HIF-1α plays a key role in hypoxic adaptation of tumor cells. HIF-1α is overexpressed in hypoxia and initializes a cascade including transcription of a network of genes that control several aspects of tumor biology (39 –41). So far, HIF-1α induces transcription of vascular endothelial growth factor (VEGF) to promote endothelial cell proliferation and blood vessel formation (40 –43). Furthermore, HIF-1α stimulates glucose metabolism of cancer cells via upregulation of glucose transporter (GLUT) 1 and 3 (40–43). Moreover, other relevant factors such as erythropoietin, inducible nitric oxide synthase, and different matrix metalloproteinases are also activated by HIF-1α (39 –42). Overall, these mechanisms result in aggressiveness, increase metastatic potential, and promote tumor progression (39 –42). In fact, previously, numerous studies identified a significant role of HIF-1α in several malignancies. It has been shown that overexpression of HIF-1α was associated with increased mortality and worse prognosis of HNSCC (10). In lung cancer, patients with positive HIF-1α expression in tumor tissues had lower overall survival rate than patients with negative HIF-1α expression (12). Similar results were reported for uterine cervical cancer (11). In addition, HIF-1α predicts resistance to anticancer therapy (9 –11). Therefore, visualization of tumor hypoxia may be of great importance. Some novel tracers were proposed to reflect tumor hypoxia HIF-1α. They showed promising preclinical and clinical results (44). However, they are not widely available and, therefore, cannot be used in daily clinical routine. In contrast, the FDG is available in clinics worldwide. Therefore, the possibility to predict on imaging, which is widely used in clinical setting, expression of HIF-1α is very important. In fact, if metabolic parameters of FDG PET are associated with expression of HIF-1α, FDG PET can be used as surrogate biomarker for tumoral hypoxia. Hence, tumor behavior can also be predicted. Theoretically, metabolic activity measured by FDG PET may reflect expression of HIF-1α. As mentioned above, HIF-1α induces upregulation of GLUT 1 and 3, which are main cellular mediators for FDG (40 –43). Presumably, parameters of FDG PET like SUV_max may well correlate with expression of HIF-1α. Some experimental investigations supported this hypothesis (45,46). Previously, only few clinical studies investigated this question. Some authors reported promising data about relationships between FDG PET and HIF-1α. However, because of the small number of participants, the reported results cannot apply as evident. The present meta-analysis is the first report based on a large sample including different tumors. Our results did not confirm the hypothesis about possible associations between SUV_max derived from FDG PET and expression of HIF-1α. A wide spectrum of correlation coefficients was identified and the pooled coefficient was 0.27. This finding indicates that FDG PET is only weakly associated with tumoral hypoxia. Furthermore, in the tumor subgroups, namely NSCLC, HNSCC, uterine cervical cancer, and thymic tumors SUV_max correlated weakly with HIF-1α. These findings are difficult to ascertain. Presumably, the HIF-1α processes only a small part of all glucose metabolisms. It is also possible that other hypoxia factors, such as hexokinase IX, may be stronger associated with glucose metabolism/FDG PET. Furthermore, the FDG uptake might be influenced by a complex interaction of several histopathological features such as intratumoral inflammatory cells, extracellular matrix, and other factors. These might be some of the reasons for the identified results in the present analysis.

The present meta-analysis identified some problems. To date, there are no reports analyzing associations between SUV_max and expression of HIF-1α in frequent and less frequent solid malignancies such as breast cancer, colonic cancer, esophageal carcinoma, different sarcomas, pancreatic cancer, hepatocellular carcinoma, and cholangiocellular carcinoma. Further investigations are needed.

Furthermore, we found significant heterogeneities among the studies investigating the same tumors. So far, in HNSCC, Grönroos et al. (14) observed an inverse correlation between SUV_max and expression of HIF-1α (−0.194), but in the study by Zhao et al. (33), the correlation coefficient was 0.577. Moreover, the same work group of Kaira et al. (17,23) identified different correlations between SUV_max and expression of HIF-1α in NSCLC, namely 0.71 and 0.11. This finding is difficult to ascertain. The identified variations of the published correlation coefficients might be related to different ratio of tumor subtypes, or different analysis methods of HIF-1α expression. Finally, for uterine cervical cancer and thymic tumors, the number of patients was small, namely 78 and 82, respectively. This fact relativizes the estimated correlation coefficients.

The present analysis also identified some methodological problems in the acquired studies. Most of the acquired reports were retrospective or the study design was not given. Furthermore, according to the QUADAS criteria, most included studies showed index test bias. Some reports also had patient selection bias. In addition, the included studies used different HIF-1α antibodies and different methods of estimation of staining intensity. Finally, only correlation coefficients between SUV_max and HIF-1α were included into the present meta-analysis. However, associations between these parameters may be of non-linear character. Therefore, tumors with high expression of HIF-1α may also have higher SUV values, but this relationship cannot be displayed by a correlation analysis. Moreover, real association between FDG uptake and HIF-1α is mainly based on the absence of evidence and this is not evidence of absence. Clearly, further prospective studies with well-defined inclusion and exclusion criteria and homogeneous methodology are needed.

In conclusion, the present meta-analysis based on the current literature showed that SUV_max of FDG PET correlated weakly with expression of HIF-1α in both the overall sample and tumor subgroups. Therefore, FDG PET cannot be used for prediction of hypoxia in clinical practice.

Authors' note

Alexey Surov is currently affiliated with Department of Radiology and Nuclear Medicine, Otto-von-Guericje University Magdeburg, Magdeburg, Germany.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Alexey Surov

Stefan A Schmidt

References

Kitajima

Miyoshi

Yamano

, et al. Prognostic value of FDG-PET and DWI in breast cancer. Ann Nucl Med 2018; 32:44–53.

Guo

Ren

Huang

, et al. Expression of Glut-1 in primary and recurrent head and neck squamous cell carcinomas, and compared with 2-[18F]fluoro-2-deoxy-D-glucose accumulation in positron emission tomography. Br J Oral Maxillofac Surg 2008; 46:180–186.

Miccò

Vargas

Burger

, et al. Combined pre-treatment MRI and 18F-FDG PET/CT parameters as prognostic biomarkers in patients with cervical cancer. Eur J Radiol 2014; 83:1169–1176.

Vera

Dubray

Palie

, et al. Monitoring tumour response during chemo-radiotherapy: a parametric method using FDG-PET/CT images in patients with oesophageal cancer. EJNMMI Res 2014; 4:12.

Abgral

Keromnes

Robin

, et al. Prognostic value of volumetric parameters measured by 18F-FDG PET/CT in patients with head and neck squamous cell carcinoma. Eur J Nucl Med Mol Imaging 2014; 41:659–667.

Surov A, Meyer HJ, Wienke A. Standardized uptake values derived from 18F-FDG PET may predict in lung cancer microvessel density and expression of KI 67, VEGF, and HIF-1αα but not expression of cyclin D1, PCNA, EGFR, PD L1 and p53. Contrast Media Mol Imaging 2018; Article ID 9257929. DOI: 10.1155/2018/9257929

Surov A, Meyer HJ, Wienke A. Can Imaging Parameters provide information regarding histopathology in head and neck squamous cell carcinoma? A Meta-Analysis. Transl Oncol 2018;11:498–503.

Cochet

Pigeonnat

Khoury

, et al. Evaluation of breast tumor blood flow with dynamic first-pass 18F-FDG PET/CT: comparison with angiogenesis markers and prognostic factors. J Nucl Med 2012; 53:512–520.

Albadari

Deng

The transcriptional factors HIF-1 and HIF-2 and their novel inhibitors in cancer therapy. Expert Opin Drug Discov 2019; 14:667–682.

10.

Gong

Zhang

Zhou

, et al. Prognostic Value of HIFs Expression in Head and Neck Cancer: A Systematic Review. PLoS One 2013; 8:e75094.

11.

Huang

Chen

Xiao

, et al. Overexpression of hypoxia-inducible factor-1α is a predictor of poor prognosis in cervical cancer: a clinicopathologic study and a meta-analysis. Int J Gynecol Cancer 2014; 24:1054–1064.

12.

Ren

Yang

, et al. The expression of hypoxia-inducible factor-1α and its clinical significance in lung cancer: a systematic review and meta-analysis. Swiss Med Wkly 2013; 143:w13855.

13.

Furukawa

Miyata

Kushitani

, et al. Association between [18F]-fluoro-2-deoxyglucose uptake and expressions of hypoxia-induced factor 1α and glucose transporter 1 in non-small cell lung cancer. Jpn J Clin Oncol 2015; 45:1154–1161.

14.

Grönroos

Lehtiö

Söderström

, et al. Hypoxia, blood flow and metabolism in squamous-cell carcinoma of the head and neck: correlations between multiple immunohistochemical parameters and PET. BMC Cancer 2014; 14:876.

15.

Han

Lee

Cho

, et al. Role of FDG-PET as a biological marker for predicting the hypoxic status of tongue cancer. Head Neck 2012; 34:1395–1402.

16.

Choi

Suh

, et al. Correlation between expression of biological markers and [F]fluorodeoxyglucose uptake in endometrial cancer. Oncol Res Treat 2014; 37:30–34.

17.

Kaira

Shimizu

Kitahara

, et al. 2-Deoxy-2-[fluorine-18] fluoro-d-glucose uptake on positron emission tomography is associated with programmed death ligand-1 expression in patients with pulmonary adenocarcinoma. Eur J Cancer 2018; 101:181–190.

18.

Kaira

Murakami

Endo

, et al. Biological correlation of ¹⁸F-FDG uptake on PET in pulmonary neuroendocrine tumors. Anticancer Res 2013; 33:4219–4228.

19.

Kaira

Endo

Abe

, et al. Biologic correlates of ¹⁸F-FDG uptake on PET in pulmonary pleomorphic carcinoma. Lung Cancer 2011; 71:144–150.

20.

Kaira

Endo

Abe

, et al. Biologic correlation of 2-[18F]-fluoro-2-deoxy-D-glucose uptake on positron emission tomography in thymic epithelial tumors. J Clin Oncol 2010; 28:3746–3753.

21.

Kaira

Serizawa

Koh

, et al. Relationship between 18F-FDG uptake on positron emission tomography and molecular biology in malignant pleural mesothelioma. Eur J Cancer 2012; 48:1244–1254.

22.

Kaira

Abe

Nakagawa

, et al. 18F-FDG uptake on PET in primary mediastinal non-thymic neoplasm: a clinicopathological study. Eur J Radiol 2012; 81:2423–2429.

23.

Kaira

Serizawa

Koh

, et al. Biological significance of 18F-FDG uptake on PET in patients with non-small-cell lung cancer. Lung Cancer 2014; 83:197–204.

24.

Miyawaki

Ikeda

Hijioka

, et al. SUVmax of FDG-PET correlates with the effects of neoadjuvant chemoradiotherapy for oral squamous cell carcinoma. Oncol Rep 2010; 23:1205–1212.

25.

Norikane

Yamamoto

Maeda

, et al. Correlation of (18)F-fluoromisonidazole PET findings with HIF-1α and p53 expressions in head and neck cancer: comparison with (18)F-FDG PET. Nucl Med Commun 2014; 35:30–35.

26.

Park

Suh

Kim

, et al. Correlation between biological marker expression and F-fluorodeoxyglucose uptake in cervical cancer measured by positron emission tomography. Onkologie 2013; 36:169–174.

27.

Shimomura

Sasahira

Yamanaka

, et al. [ 18F]fluoro-2-deoxyglucose-positron emission tomography for the assessment of histopathological response after preoperative chemoradiotherapy in advanced oral squamous cell carcinoma. Int J Clin Oncol 2015; 20:308–316.

28.

Surov A, Meyer HJ, Hölhn AK, et al. Associations between 18F-FDG-PET and complex histopathological parameters including tumor cell count and expression of KI 67, EGFR, VEGF, HIF-1α and p53 in head and neck squamous cell carcinoma. Mol Imaging Biol 2019;21:368–374.

29.

Surov A, Meyer HJ, Höhn AK, et al. Metabolo-volumetric parameters of 18F-FDG-PET can predict expression of EGFR and HIF 1alpha in uterine cervical cancer. Cancer Biomark 2019;24:135–140.

30.

Takebayashi

Izuishi

Yamamoto

, et al. [18F]Fluorodeoxyglucose accumulation as a biological marker of hypoxic status but not glucose transport ability in gastric cancer. J Exp Clin Cancer Res 2013; 32:34.

31.

Toba

Kondo

Sadohara

, et al. 18F-fluorodeoxyglucose positron emission tomography/computed tomography and the relationship between fluorodeoxyglucose uptake and the expression of hypoxia-inducible factor-1α, glucose transporter-1 and vascular endothelial growth factor in thymic epithelial tumours. Eur J Cardiothorac Surg 2013; 44:e105–112.

32.

Yamada

Uchida

Kwang-Lee

, et al. Correlation of metabolism/hypoxia markers and fluorodeoxyglucose uptake in oral squamous cell carcinomas. Oral Surg Oral Med Oral Pathol Oral Radiol 2012; 113:464–471.

33.

Zhao

Yang

Zhou

, et al. Fluorodeoxyglucose uptake in laryngeal carcinoma is associated with the expression of glucose transporter-1 and hypoxia-inducible-factor-1α and the phosphoinositide 3-kinase/protein kinase B pathway. Oncol Lett 2014; 7:984–990.

34.

Moher

Liberati

Tetzlaff

, et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009; 6:e1000097.

35.

Whiting

Rutjes

Westwood

, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011; 155:529–536.

36.

Leeflang

Deeks

Gatsonis

, et al. Systematic reviews of diagnostic test accuracy. Ann Intern Med 2008; 149:889–897.

37.

Zamora

Abraira

Muriel

, et al. Meta-DiSc: A software for meta-analysis of test accuracy data. BMC Med Res Methodol 2006; 6:31.

38.

DerSimonian

Laird

Meta-analysis in clinical trials. Control Clin Trials 1986; 7:177–188.

39.

Osinsky

Zavelevich

Vaupel

Tumor hypoxia and malignant progression. Exp Oncol 2009; 31:80–86.

40.

Ruan

Song

Ouyang

Role of hypoxia in the hallmarks of human cancer. J Cell Biochem 2009; 107:1053–1062.

41.

Denko

NC.

Hypoxia, HIF1 and glucose metabolism in the solid tumour. Nat Rev Cancer 2008; 8:705–713.

42.

Costa

Hypoxia-inducible factor-1 (HIF-1). Mol Pharmacol 2006; 70:1469–1480.

43.

Semenza

GL.

Defining the role of hypoxia-inducible factor 1 in cancer biology and therapeutics. Oncogene 2010; 29:625–634.

44.

Mees

Dierckx

Vangestel

, et al. Molecular imaging of hypoxia with radiolabelled agents. Eur J Nucl Med Mol Imaging 2009; 36:1674–1686.

45.

Burgman

Odonoghue

Humm

, et al. Hypoxia-induced increase in FDG uptake in MCF7 cells. J Nucl Med 2001; 42:170–175.

46.

Pugachev

Ruan

Carlin

, et al. Dependence of FDG uptake on tumor microenvironment. Int J Radiat Oncol Biol Phys 2005; 62:545–553.