Background Automatic tumor segmentation based on Convolutional Neural Networks (CNNs) has shown to be a valuable tool in treatment planning and clinical decision making. We investigate the influence of 7 MRI input channels of a CNN with respect to the segmentation performance of head&neck cancer. Methods Head&neck cancer patients underwent multi-parametric MRI including T2w, pre- and post-contrast T1w, T2*, perfusion (k trans , v e ) and diffusion (ADC) measurements at 3 time points before and during radiochemotherapy. The 7 different MRI contrasts (input channels) and manually defined gross tumor volumes (primary tumor and lymph node metastases) were used to train CNNs for lesion segmentation. A reference CNN with all input channels was compared to individually trained CNNs where one of the input channels was left out to identify which MRI contrast contributes the most to the tumor segmentation task. A statistical analysis was employed to account for random fluctuations in the segmentation performance. Results The CNN segmentation performance scored up to a Dice similarity coefficient (DSC) of 0.65. The network trained without T2* data generally yielded the worst results, with ΔDSC GTV-T = 5.7% for primary tumor and ΔDSC GTV-Ln = 5.8% for lymph node metastases compared to the network containing all input channels. Overall, the ADC input channel showed the least impact on segmentation performance, with ΔDSC GTV-T = 2.4% for primary tumor and ΔDSC GTV-Ln = 2.2% respectively. Conclusions We developed a method to reduce overall scan times in MRI protocols by prioritizing those sequences that add most unique information for the task of automatic tumor segmentation. The optimized CNNs could be used to aid in the definition of the GTVs in radiotherapy planning, and the faster imaging protocols will reduce patient scan times which can increase patient compliance. Trial registration The trial was registered retrospectively at the German Register for Clinical Studies (DRKS) under register number DRKS00003830 on August 20th, 2015.

Purpose: To dynamically minimize radiofrequency (RF)-induced heating of an active catheter through an automatic change of the termination impedance. Methods: A prototype wireless module was designed that modifies the input impedance of an active catheter to keep the temperature rise during MRI below a threshold, ΔT max . The wireless module (MR safety watchdog; MRsWD) measures the local temperature at the catheter tip using either a built-in thermistor or external data from a fiber-optical thermometer. It automatically changes the catheter input impedance until the temperature rise during MRI is minimized. If ΔT max is exceeded, RF transmission is blocked by a feedback system. Results: The thermistor and fiber-optical thermometer provided consistent temperature data in a phantom experiment. During MRI, the MRsWD was able to reduce the maximum temperature rise by 25% when operated in real-time feedback mode. Conclusion: This study demonstrates the technical feasibility of an MRsWD as an alternative or complementary approach to reduce RF-induced heating of active interventional devices. The automatic MRsWD can reduce heating using direct temperature measurements at the tip of the catheter. Given that temperature measurements are intrinsically slow, for a clinical implementation, a faster feedback parameter would be required such as the RF currents along the catheter or scattered electric fields at the tip. K E Y W O R D S active catheter, active implantable medical devices, bluetooth low energy, interventional MRI, MR safety, radio frequency induced heating | 1049 ÖZEN Et al.

X-ray fluoroscopy is the gold standard for coronary diagnostics and intervention. Magnetic resonance imaging is a radiation-free alternative to x-ray with excellent soft tissue contrast in arbitrary slice orientation. Here, we assessed real-time MRI-guided coronary interventions from femoral access using newly designed MRI technologies. Six Goettingen minipigs were used to investigate coronary intervention using real-time MRI. Catheters were custom-designed and equipped with an active receive tip-coil to improve visibility and navigation capabilities. Using modified standard clinical 5 F catheters, intubation of the left coronary ostium was successful in all animals. For the purpose of MR-guided coronary interventions, a custom-designed 8 F catheter was used. In spite of the large catheter size, and therefore limited steerability, intubation of the left coronary ostium was successful in 3 of 6 animals within seconds. Thereafter, real-time guided implantation of a non-metallic vascular scaffold into coronary arteries was possible. This study demonstrates that real-time MRI-guided coronary catheterization and intervention via femoral access is possible without the use of any contrast agents or radiation, including placement of non-metallic vascular scaffolds into coronary arteries. Further development, especially in catheter and guidewire technology, will be required to drive forward routine MR-guided coronary interventions as an alternative to x-ray fluoroscopy.