Texture analysis of PET images for predicting response to induction chemotherapy for oral squamous cell carcinoma

Abstract

We evaluated whether texture analysis of positron emission tomography (PET) scans could predict the response to induction chemotherapy (ICT) in patients with oral squamous cell carcinoma (OSCC). Our retrospective cohort study included 21 patients with OSCC who underwent ICT. We extracted 31 textural indices from pretreatment PET scans and performed receiver operating characteristic curve (ROC) analysis to evaluate these indices’ efficacy in predicting ICT response. The area under the ROC curve was statistically significant for one texture index: short-zone emphasis (SZE; p = 0.041).

Keywords

• Oral cancer
• Chemotherapy
• Response
• Textural indices