Format

Send to

Choose Destination
Acta Radiol. 2019 Feb 28:284185119825487. doi: 10.1177/0284185119825487. [Epub ahead of print]

Tumor heterogeneity in oral and oropharyngeal squamous cell carcinoma assessed by texture analysis of CT and conventional MRI: a potential marker of overall survival.

Author information

1
Department of Radiology, Shanghai Ninth People's Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, PR China.

Abstract

BACKGROUND:

The combined prognostic value of texture analysis on computed tomography (CT) and conventional magnetic resonance imaging (MRI) in oral and oropharyngeal squamous cell carcinoma (OOSCC) has rarely been reported.

PURPOSE:

To evaluate the ability of texture-based heterogeneity indices extracted from CT and conventional MRI to predict the overall survival (OS) of OOSCC patients.

MATERIAL AND METHODS:

Eighty-eight patients with OOSCC underwent contrast-enhanced CT, and T2-weighted (T2W) and contrast-enhanced T1-weighted (T1W) imaging. For each imaging protocol, four histogram (HISTO) and six gray-level co-occurrence matrix (GLCM) parameters were extracted. Intraclass correlation coefficients (ICCs) were used to evaluate variation in inter-observer delineation. The association of clinicopathologic factors and texture parameters with OS was analyzed using log-rank testing and Cox regression. The Harrell Concordance Index (C-index) was used to evaluate the prognostic performance of the predictors and Cox model.

RESULTS:

Higher inter-observer delineation agreement of texture analysis on contrast-enhanced T1W imaging was observed relative to that on CT and T2W imaging. T stage ( P = 0.026, hazard ratio [HR] = 2.35), GLCM_Entropy on CT ( P = 0.003, HR = 3.15), and HISTO_Energy ( P = 0.036, HR = 0.27) and HISTO_Skewness ( P = 0.026, HR = 0.41) on contrast-enhanced T1W imaging were independent predictors for the OS of OOSCC patients. The multivariate Cox model performed strongly in predicting OS, with a significantly higher C-index than the independent predictors separately (C-index: 0.751 vs. 0.600-0.627, P < 0.005 for all).

CONCLUSION:

Texture analysis using CT and conventional MRI could provide a valuable method for survival stratification of OOSCC patients.

KEYWORDS:

Head and neck cancer; computed tomography; heterogeneity; magnetic resonance imaging; prognosis; texture analysis

PMID:
30818979
DOI:
10.1177/0284185119825487

Supplemental Content

Full text links

Icon for Atypon
Loading ...
Support Center