Abstract
PURPOSE: Radiation-induced morphea (RIM) is a very rare but devastating side effect of breast radiation therapy, characterized by progressive skin induration, pain, and discoloration, with no effective treatments currently available. This is a proof-of-concept study that aims to identify radiomic features from pretreatment magnetic resonance imaging (MRI) scans associated with the development of RIM in patients with breast cancer undergoing radiation therapy. METHODS AND MATERIALS: This is a retrospective analysis of a single institutional registry of patients who received diagnosis of RIM following breast radiation therapy from 2008 to 2022. Clinical and histopathological data were reviewed. Pretreatment MRI scans of these patients and matched controls were analyzed. Radiomic features were extracted from whole breast and fibroglandular tissue regions of interest. A total of 528 radiomic features were compared between patients who developed RIM and those who did not, using the Wilcoxon rank-sum test to identify statistically significant differences. RESULTS: We evaluated 10 patients who received clinical diagnosis of RIM, with a mean age of 63 years (range, 44-75 years). Among these, 7 patients had biopsy-proven RIM. Both clinical and histologic findings were correlated with radiomic analyses. Forty percent of the patients had a history of autoimmune disorders, including hypothyroidism, Graves' disease, systemic sclerosis, and systemic lupus erythematosus. Radiomic analysis identified 11 significant features, primarily related to tissue structure and texture. Nine of these features were from the contralateral breast, and 2 were from the ipsilateral breast. CONCLUSIONS: This is a pilot study on a small sample that demonstrates that radiomic features extracted from pretreatment MRI scans can serve as potential predictors for the development of RIM in patients with breast cancer. The integration of clinical and histopathological data with radiomic analysis highlights the distinct changes in breast tissue architecture that precede RIM onset. These findings pave the way for the early identification of patients at risk, allowing for more personalized surveillance and management strategies.