Longitudinal control and segmentation of indeterminate pulmonary nodules

Accurate and reproducible segmentation of indeterminate pulmonary nodules and primary lung tumors including estimation of volume based on medical images. It can play a pivotal role in the diagnosis, staging, and ultimately assessment of response to cancer therapy. 

It is complex and time-consuming task to manually keep track of multiple nodules over time in terms of volume, growth and appearance. This can cause radiologists to overlook measures of growth over time and the fact that the nodules themselves change morphologically.

In case that manual volume estimation etc. could be replaced by an algorithm we hypothesize that it is more likely that the patient achieves timely diagnosis and a possibility of rapid treatment. An algorithm will also reduce manual workload within the area of tools related to volume measures, automation, smart reporting, time-overview etc.

Triage of chest X-ray in normal and abnormal examinations Expediated reporting of chest CT and chest X-Ray

An artificial intelligence (AI) system that with high accuracy can differentiate normal from abnormal chest X-rays. The potential is enormous if radiologists could prioritize abnormal chest X-rays with critical findings and normal images were handled solely by the algorithm. In case critical findings were flagged by the algorithm it could potentially reduce the backlog of exams and bring urgently needed care to patients more quickly.

Many patients could benefit from CT of the chest especially in acute patients as not sufficient information is present in standard chest X-ray. But radiologist resources are to scarce to analyze a growing number of CT scans and the many images included. Algorithms focusing on this aspect is sorely needed.

Detection and triage of (acute) cerebral symptoms / trauma

Early recognition and differentiation of acute ischemic stroke from intracranial hemorrhage and stroke mimics and the identification of large vessel occlusion (LVO) are critical to the appropriate management of stroke patients.

One way of doing that is to automate some of the process and also help the radiologist prioritize the different exams. Especially when there is a lack of neuro-radiological competencies.

When correct detection from the beginning there is more space available at the stroke unit for the people who really need it and furthermore the algorithm can assist the radiologist when the capacity of radiologist is low.

Detection and triage of bone fractures

The amount of fractures are high and fractures must not be overseen due to lack of radiologist, reading technologists or lack of experience at the younger radiologists . AI can assist especially the youngest doctors by analyzing the images and reviewing cases.

Detection and triage of bone fractures by AI will most likely be less time consuming and might be able to assist even the emergency doctors and nurses with a first read in the emergency room.

Brain

Several diseases in the brain are affecting the ration between the gray and whiter matter and the ventricular system including, but not limited to, degenerative brain diseases like dementia (Alzheimer, vascular changes) autoimmune diseases like Multiple Sclerosis, Subarachnoid bleedings, Stroke, Meningitis, that all heavily rely on advanced CT and MRI imaging to detect and monitor the disease changes in the brain.

With the increasing amount of imaging possibilities and increase in data size, combined with the increasing drug treatment possibilities in i.e Alzheimer's disease imaging interpretation are becoming a bottle neck, which is why AI methods to detect the above mentioned diseases as well as technology to automatically segment gray and white matter volume and ventricular size that can be used to track the consequences of the diseases will be of great value to mitigate the mentioned bottle necks.