Parameters from novel non-Gaussian diffusion models for identifying the expression levels of prognostic factors in nasopharyngeal carcinoma

Motivation: The role of parameters derived from novel non-Gaussian diffusion models in identifying prognostic factors associated with nasopharyngeal cancer has not been reported. Goal(s): To explore the association between diffusion parameters and prognostic factors in NPC, including Ki-67, EGFR, and tumor-stroma ratio. Approach: Eighty-three NPC patients performed continuous-time random-walk (CTRW) and fractional order calculus (FROC) models. ROC analysis was conducted using the diffusion parameters to evaluate accuracy in identifying prognostic factors. Results: Diffusion parameters from CTRW and FROC models can distinguish EGFR and TSR expression, but not Ki-67. Impact: Non-Gaussian diffusion parameters can non-invasively assess prognostic factors in NPC and help doctors better manage NPC patients.