Quantitative depth-based [18F]FMCH-avid lesion profiling in prostate cancer treatment

Besides prognostic clinical factors, baseline risk assessment in personalized cancer research would benefit from quantitative disease characterization to inform therapy planning. Texture analysis of [18F]FMCH PET/CT imaging is paving the way for such purposes but its potential is still braked by radiomic feature limitation, such as redundancy and lack of standardization. In this work, we provide a method for a robust assessment of intratumor heterogeneity in patients affected by prostate cancer, through a depth-based ranking quantifying the level of centrality/outlyingness of the lesion with respect to peers. We interpret the results in terms of clinical information of lesions.