Histological evaluation of tumors plays an important role in malignancy identification and, more recently, in stratification of treatment response and prognosis. Quantitative imaging methodology has been applied to cross-sectional imaging datasets with pathology correlation to model tumor composition and tissue classification. Several digital image analysis studies have demonstrated prognostic significance for tumor heterogeneity, proportional representation of complex stroma, and immune cell infiltration. Texture analysis has been used for classification of several tissue subtypes. We examine here a new approach that uses wavelet transformations of images, which can optimize spatial and frequency distributions for characterized intratumoral tissues. We have developed and evaluated a model for colon intratumoral tissue classification using wavelet scattering features, achieving 85.10% accuracy on test data with precision (85.19%), recall (85.10%), and F1-score (85.08%) across eight tissue types.