This work aims at defining a validation approach to subjectively assess the realism of breast phantom images acquired with x-ray systems. Various breast phantoms were used to address specific questions. First, the realism of a breast neoplasm model was evaluated. Second, different materials for phantom production were evaluated for x-ray contrast, sharpness, inhomogeneity, and artifacts. Third, 3D printing techniques for manufacturing breast tissues were explored. Lastly, computational models were assessed by comparing manual and automatic segmentations of the breast tissues. Two distinct grading scales were introduced to aid radiologists in evaluating the phantoms. The first scale assesses computational phantoms based on the realism of glandular tissue distribution, morphology, and the preservation of fine parenchymal structures. The second scale is designed to support the validation of physical phantoms, focusing on the realism of the generated x-ray image. The scales have been applied to the selected cases and will be further used in studies involving the production of anthropomorphic phantoms.