Medical Imaging Assistant Non-Annotated MRI Frame Warning

The development of an Assistant helps the automation of cancer diagnosis. If a large dataset is available, it is possible to have recommendation methodologies that automatically can classify the set of medical images and lesions detection, giving a high probability to those cases where intervention is necessary. These methodologies will bring us a high impact to the clinical field. Such automation is crucial since it reduces the inspection performed by the radiologist, that is still rudimentary in current clinical setups. Besides having large datasets, the follow-up of the patient is crucial. This dataset means that the annotation of a given patient should be analyzed through time. On one hand, based on the literature we plan to improve our Assistant within a context of scaling our solution. We aim to understand how clinical institutions can use our system with impactful healthcare systems. On the other hand, the Magnetic Resonance Imaging (MRI) [1] technology is a type of scan that allows the visualization of several slices, framing layer-by-layer and a picture taken of each slice. This technology can help diagnose the lesions' progress across the set of slices. In our annotations dataset, we need to provide complete information regarding the existence of injuries in all frames that a lesion is presented. Therefore, we can not accept frames where there is evidence of lesion, but the clinician passed through it missing the lesion annotation. With this issue, our goal is to surpass the problem of misannotation a frame on a set of slices. The idea is, whenever the system starts to receive at, for instance, slice s = 121 and finish at the slice s = 134 it needs to guarantee that ALL slices between those are annotated. In the following image (uploaded), we can see if the slice s = 128 is not annotated, the system will trigger a warning when the clinician press the saving button. In short, there is the (high) probability of a clinician to jump between slices. For that reason, we need to guarantee that the system warns the clinician to fix that problem. The solution to it is, each time a clinician presses the saving button or closes (the behavior is the same as the last) the respective patient, we will warn the clinician. Budget = 30$ - 50$ [1]: [login to view URL]