MRI-based methods for response assessment to neoadjuvant therapy in esophageal cancer

Esophageal cancer is known to have low 5-year overall survival rates. Furthermore, the degree of response is associated with patient prognosis, with the most favorable long-term prognosis for the 29% of patients reaching pathologic complete response. For the introduction of personalized treatment schemes accurate prediction or assessment of response is required in which, depending on response, patients could follow a different treatment path. This could include for example a wait-and-see approach with omission of surgery, intensification of neoadjuvant therapy or primary surgery. In this thesis a range of topics has been discussed with a main focus on using MR-based (functional) techniques to predict response to neoadjuvant treatment for patients with esophageal cancer. The studies presented in this thesis showed the potential of functional MRI in the prediction of response in patients with esophageal cancer undergoing neoadjuvant chemoradiotherapy. Furthermore, it was found that different modalities (e.g. DCE-MRI, DW-MRI and 18F-FDG PET/CT) could be of complementary value in predicting response, each visualizing different physiological processes in the tumor. Therefore, the use of multi-parametric multimodality models should be further researched in larger patient cohorts. In addition, a methodical study on fitting methods for IVIM-MRI was researched. This study showed that differences in trend in fitted parameters occurred between the different fitting approaches. It is expected that these differences influence the accuracy and precision in longitudinal studies for response assessment in which IVIM will be used. Therefore, it was concluded that it is of great importance to explicit report specifications of fitting approaches used in IVIM studies to be able to compare outcomes between studies. Finally, esophageal tumor motion was researched on different time scales. Large variations were found in tumor motion both between and within patients with esophageal cancer. The stochastic nature of the variations in motion demonstrated the importance of real-time tumor motion management during radiotherapy in order to safely reduce treatment margins.