A web-based system for the automated analysis of MRI in neuro-oncology

Magnetic Resonance Imaging (MRI) exams provide imaging data of high importance to various medical fields including that of neuro-oncology [1]. Beyond the technological aspect of this imaging technique, the assessment of this data is crucial and specialists are necessary. Recent trends aim at developing software systems for the automation of certain tasks [2]. The advances of artificial intelligence and in specific that of neural networks allowed for the automation of processes like that of segmentation based [3] on modern mathematical instruments, their software implementations and deployment in suitable hardware. Due to their complexity such systems are hard to be accessed by groups of specialists not belonging to the original research team and affiliated institutes. Even at premature stages such systems could provide useful insights to the broader medical community. This is especially the case when cross-facility experiments are necessary and novel approaches to research are implemented like virtual trials [ref]. This short manuscript communicates the design and development of a web-based system [Fig.1 and Fig.2] that allows for easier access to advanced methods of automated MRI assessment [4] while it allows for data management and feedback collection for further research. The system permits the inclusion of multiple imaging methods including systems that provide quantitative information [5] for the automatically segmented data. Last but not least, the outcome of this project suggests that such systems can be prototyped fast and render important radiomic advances [6] available to a larger audience. While funds allow it a publicly available beta is in www.rmn.ai