Quantitative diffusion magnetic resonance imaging for prediction of human papillomavirus status in head and neck squamous-cell carcinoma: A systematic review and meta-analysis

Neuroradiol J. 2019 Aug;32(4):232-240. doi: 10.1177/1971400919849808. Epub 2019 May 14.


Purpose: Head and neck squamous-cell carcinoma (HNSCC) related to human papillomavirus (HPV) infection represents a distinct biological and prognostic subtype compared to the HPV-negative form. Prior studies suggest a correlation between the apparent diffusion coefficient (ADC) values on diffusion-weighted imaging (DWI) of primary tumor lesion and HPV status in HNSCC. In this meta-analysis, we compared the average ADC of primary lesion between HPV-positive and HPV-negative HNSCC.

Methods: A comprehensive literature search of PubMed and Embase was performed. Studies comparing the average ADC on echo-planar DWI of primary tumor lesions between HPV-positive and HPV-negative HNSCC were included. The standardized mean difference was calculated using fixed- and random-effects models. Tau-squared estimates of total heterogeneity and Higgins inconsistency index (I2 test) were determined.

Results: A total of five studies, pooling data of 264 patients, were included for meta-analysis. Among these five studies, three had included oral cavity, hypopharyngeal, and/or laryngeal HNSCC in addition to oropharyngeal subsite. Primary lesions were comprised of 185 HPV-negative and 79 HPV-positive HNSCC. The meta-analysis showed lower average ADC values in HPV-positive HNSCC compared to the HPV-negative form, with a standardized mean difference of 0.961 (95% confidence interval 0.644-1.279; p < 0.0001). Since there was no significant heterogeneity in analysis (p = 0.3852), both random- and fixed-effects models resulted in the same estimates of overall effect.

Conclusions: HPV-positive HNSCC primary lesions have a lower average ADC compared to the HPV-negative form, highlighting the potential application of quantitative diffusion magnetic resonance imaging as a noninvasive imaging biomarker for prediction of HPV status.

Keywords: Apparent diffusion coefficient; head and neck tumors; human papillomavirus; meta-analyses; squamous-cell carcinoma.

Publication types

  • Meta-Analysis
  • Systematic Review

MeSH terms

  • Diffusion Magnetic Resonance Imaging / methods
  • Early Detection of Cancer / methods
  • Female
  • Head and Neck Neoplasms / virology*
  • Humans
  • Male
  • Middle Aged
  • Papillomaviridae
  • Papillomavirus Infections / diagnosis*
  • Squamous Cell Carcinoma of Head and Neck / virology*