Keywords: Breast cancer, MRI, Response prediction, Radiomics

"MRI-BASED RADIOMICS IN BREAST CANCER Optimization and Prediction"

Breast cancer is the most common type of cancer in women in the Netherlands. Currently, 1 in 7 women will develop breast cancer during their lifetime. The treatment of breast cancer, consisting of a combination of surgery, systemic therapy and radiotherapy has improved tremendously in recent decades. Survival with a 5-year term has increased from 52% to 88% over the past 60 years. Because of this improvement, the focus is increasingly on quality of life and preventing overtreatment. In this dissertation we investigated whether the application of MRI-based radiomics within the treatment of breast cancer can contribute to this. Radiomics is a technique that uses mathematical programs to process medical images into quantitative data, called radiomic features. The use of MRI-based radiomics in predicting tumor response to systemic therapy with the aim of adapting clinical treatment was investigated in this thesis. The use of MRI-based radiomics in predicting metastases in axillary lymph nodes was also examined. Both studies showed that the prediction models created are not (yet) contributing in a way that clinical treatments can be adapted. It has been found that the radiomics features are sensitive to different variations in the examined data. The effect of variation in the examined data on the radiomic features will first have to be investigated and demonstrated before new prediction models can be made.  

Click here for the full dissertation.

Click here for the live stream.