Multimodal Dynamic MRI for Structural and Functional Assessment of Cancer

Abstract

In this work, a novel dynamic contrast-based magnetic resonance imaging (MRI) method, called ‘split dynamic MRI’, is presented, which show to improve the differential diagnosis of breast cancer, as compared to conventional MRI interpretations, and facilitate staging of primary rectal cancer and lymph node metastasis. Our results showed a diagnostic accuracy of 96% for distinguishing malignant from benign breast cancer, and 90% when differentiating rectal cancer patients with and without lymph node metastasis. The proposed method has a unique advantage over conventional dynamic MRI, as it provides comprehensive information about both structural and physiological characteristics in the tumour. Consequently, combining these information channels may provide a more extensive characterization of tumour aggressiveness, and thus identify patients in need of either intensified or reduced treatment (personalized treatment). Through mathematical simulations, this work show that the split dynamic MRI method provides reliable information of physiological properties in tumours, as compared to conventional dynamic MRI, without compromising international guidelines concerning image resolution for structural interpretation of breast cancer. Reliable identification of individual tumour aggressiveness may represent a major breakthrough in treatment stratification, as this may enable a more personalized treatment. However, conventional MRI findings are associated with substantial misinterpretation and poor diagnostic accuracy. In this context, our results suggest that the split dynamic MRI method can provide physicians with valuable information for clinical decision making in breast cancer and rectal cancer.