The future of computing imaging in healthcare?

The future of computing imaging in Healthcare? With Minerva, a software using IA based Automatic processing of medical images to help in the diagnosis of different pathologies, it would be possible.

?What is Minerva?

Minerva is a software, developed by?Lincbiotech (www.lincbiotech.com), oriented to the automatic processing of medical images to help in the diagnosis of different pathologies. Minerva allows the clinical user to work with much more complete information than that obtained by hospital systems.

To do this, Minerva uses different computer vision algorithms, digital image processing and artificial intelligence. Minerva uses the most advanced technology in these fields to achieve high quality results.

?What type of imaging studies can Minerva analyse?

Currently Minerva has a complete and continuously expanding pack focused on stroke (Minerva Stroke). Minerva Stroke allows for the extraction of biomarkers of interest from magnetic resonance imaging (MRI) and computerized tomography (CT) studies. More specifically Minerva can compute the following biomarkers:

-Generation of perfusion maps and calculation of infarct volume measurements, penumbra and mismatch. This makes it possible to visualize the degree of extension of a cerebral infarction and what the patient's condition is in terms of the severity of the infarction.

Example of visual results generated on CT perfusion imaging (Minerva Stroke):

-?????????Computation of MIP projections of the cerebral vasculature in angiography imaging and computation of collateral circulation between hemispheres. This provides additional information to assess how affected is the cerebral circulatory system of the affected hemisphere and helps in the visualization of possible blood flow blockages.

Example of the results obtained on angiography (Minerva Stroke):

-Detection and characterization of thrombi in brain CT images without contrast. The system automatically detects the presence of any thrombus causing brain stroke and provides different measurements of their morphological characteristics. Additionally, if the study includes angiographies, these are used for a more exhaustive evaluation of the permeability characteristics of the thrombus. This analysis provides vital information for choosing the optimal thrombectomy technique for subsequent removal of the thrombus.

Example of the results obtained on thrombus characterization (Minerva Stroke):

Additionally, Minerva has a module under development (Minerva Spine) for the characterization of the state of the spine in both CT and MR images. Currently Minerva Spine already includes a system for the analysis of CT images of the spine to characterize the state of the lumbar, thoracic and cervical vertebrae. Soon Minerva Spine will allow the analysis of MR studies to characterize disc degradation and lumbar stenosis among other biomarkers.

Example of the results of ?vertebral analysis (Minerva Spine):

?How does Minerva work?

Minerva has two modes of use: web platform and hospital/clinic integration. In both cases, Minerva executes its algorithms in the cloud after anonymizing the sensitive data of any file received by the system through an automatically integrated local system, so that the personal data of the patient, specialist or institution never enter the servers of Minerva, allowing secure anonymization.

The web platform allows the user to access a complete interface through a personal account where they can see the cases already processed and the results. Any user can upload their images in DICOM format to be processed automatically and view the results. The user can upload several studies and several images and does not need any type of manual selection of them, the system can automatically catalog each type of image and by itself execute all the algorithms for which there are images available. Likewise, the user can download pdf and dicom reports of the results, as well as files of the images generated in both dicom and nifti format alongside the computed biomarkers in an excel compatible file. This environment is suitable for both diagnostic and clinical research units.

The hospital and clinic integration system is designed to work directly with radiologists and specialists in their natural environment. The integration ensures that whenever a new study is available in the PACS system requiring analysis, it is automatically sent to the servers where Minerva executes its algorithms and once the process is finished, it sends back the dicom and pdf report of the results to the center. The Minerva system works through an advanced API that allows connection to any type of hospital system and is adaptable according to the needs of each clinic or hospital. This environment is specifically oriented to diagnostic aid processes in a real clinical context.

References

Scientific publications

Minerva has already been used to carry out both quantitative and qualitative studies and has scientific publications supportting its quality in terms of the results offered. The following publications contain comparative information against other systems and its use in public databases:

-Pérez-Pelegrí, M., Biarnés, C., Thió-Henestrosa, S., Remollo, S., Gimeno, A., Cuba, V., ... & Puig, J. (2021). Higher agreement in endovascular treatment decision-making than in parametric quantifications among automated CT perfusion software packages in acute ischemic stroke.?Journal of X-Ray Science and Technology, (Preprint), 1-12.?

-Pérez-Pelegrí, M., Puig, J., & Ruiz-Constantino, J. S. (2021, November). Acute Infarct Volume Prediction Based on CT Perfusion Metrics Derived from an Automated Software Package using Machine Learning Models. In?2021 International Conference on e-Health and Bioengineering (EHB)?(pp. 1-4). IEEE.

More information available:

https://lincbiotech.com/minerva-stroke-advanced-neuroimaging-software-analysis/??

or contact to Manuel Pérez Pelegrí, Ph.D. , R&D head of biomedical imaging at Lincbiotech