Risk Stratification for Oropharyngeal Squamous Cell Carcinoma Using Texture Analysis on CT - A Step Beyond HPV Status

Area of Science:

• Oncology
• Radiology
• Medical Imaging

Background:

• Human papillomavirus-associated oropharyngeal squamous cell carcinoma (OPSCC) is a growing public health concern.
• While generally having a better prognosis, OPSCC patients exhibit varied responses to treatment, necessitating improved prognostic tools.
• Existing radiomics models for OPSCC prognostication primarily focus on the primary tumor, leaving a gap for cases where the primary tumor is not visualized.

Purpose of the Study:

• To investigate the utility of radiomics from metastatic lymphadenopathy in developing a risk assessment model for OPSCC.
• To compare the prognostic efficacy of models derived from metastatic lymphadenopathy with those from the primary tumor.
• To determine if incorporating human papillomavirus (HPV) status enhances the prognostic accuracy of these radiomics models.

Main Methods:

• Retrospective analysis of 80 stage III-IV OPSCC patients diagnosed between February 2009 and October 2015.
• Extraction of texture features from pre-treatment CT scans of both primary tumors and metastatic lymphadenopathy via manual segmentation.
• Development of multivariate models using extracted features to predict treatment response.

Main Results:

• Texture analysis of both primary OPSCC and metastatic lymphadenopathy demonstrated the ability to develop predictive models for treatment outcomes.
• The developed models exhibited high predictive capability, with Area Under the Curve (AUC) values ranging from 0.78 to 0.85.
• These findings suggest that radiomics of metastatic lymphadenopathy is a viable method for prognostication in OPSCC.

Conclusions:

• Radiomics analysis of metastatic lymphadenopathy shows promise for prognostication in OPSCC, comparable to primary tumor analysis.
• This preliminary research provides a foundation for multi-center trials to refine treatment strategies.
• Optimizing treatment based on radiomics and HPV status could significantly improve patient quality of life in the era of personalized medicine.