We improve radiology workflow to reduce under-reporting and provide prioritizing patients using AI by auto detecting up to 14 brain lesions
Market
The global CT scanners market size is poised to grow by USD 1.79 billion during 2020-2024
Projected increase: CAGR of 6.5% by 2026..
23 million brain computed tomographies a year
52 million CT scans are performed annually, and over 51 million within the EEA
Key issues
Problem owner
- hospital managers are facing significant delays in image interpretations
- our solution improves radiology workflow by (not limited to) fast classification and prioritization
Workforce shortage
- data: 7% global data volume growth
- technical advancement
- aging of radiology personnel
Population aging
- 25% in 2030 > age 65
- more imaging diagnostics required
“
Based on the analysis of more than 246 108 brain CT scans, we developed a CE certified BrainScan.ai system
“
Dataset Q2'2021
0
| Providers | # of all examinations |
|---|---|
| Data source A | 35 432 |
| Data source B | 37 004 |
| Data source C | 7 613 |
| Data source D | 27 495 |
| Data source E | 31 000 |
| Data source F | 70 000 |
| Data source G | 37 564 |
| Total | 246 108 |
Dataset Q4'2021
0
| Providers | # of all examinations |
|---|---|
| Data source A | 71 291 |
| Data source B | 48 840 |
| Data source C | 14 864 |
| Data source D | 39 674 |
| Data source E | 32 705 |
| Data source F | 70 000 |
| Data source G | 37 564 |
| Total | 314 938 |
Solution
- unique range of pathologies - up to 26 classes
- triage - wide # of pathologies detected help prioritizing emergency patients
- prevent under-reporting
- complex brain diagnostic
- seamless integration powered by BrainScan cloud solution
“
BrainScan is a vertical solution that focus on the brain scans, but horizontal in the meaning of number of lesions detected
“
How does it work?
CT Scanning
The series of scans from the machine is sent as DICOM files to the PACS server. The script filters all brain CT scans. Anonymizes them and forward to the BrainScan Cloud.
Analysis
Algorithms powered by Artificial Intelligence automatically analyze scans finding potential pathological changes. The results are saved in a form of infographic and are visible by the doctors as the next serie in the study and saved in a DICOM file.
Results
The results of the analysis in a form of infographic (an additional series saved in the DICOM file) and in a structured text form (JSON file) are returned to the PACS server from which they were sent. The advantages of a form of infographic is that it is available in every DICOM viewers used by doctors.
The effectiveness of our algorithms
| # | Lesion |
ROC AUC |
Sensitivity |
Specificity |
Precision |
Accuracy |
NPV |
|---|---|---|---|---|---|---|---|
| Hemorrhage / Bleeding | |||||||
| 1 | Hemorrhage / Bleeding | 0.976 | 0.782 | 0.990 | 0.854 | 0.975 | 0.982 |
| Pericerebral hematoma | |||||||
| 2 | Pericerebral hematoma | 0.979 | 0.752 | 0.993 | 0.773 | 0.985 | 0.991 |
| Acute ischemic stroke | |||||||
| 3 | Acute ischemic stroke | 0.846 | 0.660 | 0.826 | 0.487 | 0.793 | 0.984 |
| Chronic ischemic stroke | |||||||
| 4 | Chronic ischemic stroke | 0.910 | 0.464 | 0.986 | 0.466 | 0.972 | 0.904 |
| Malacia | |||||||
| 5 | Malacia | 0.923 | 0.643 | 0.943 | 0.699 | 0.892 | 0.923 |
| Edema | |||||||
| 6 | Edema | 0.934 | 0.566 | 0.981 | 0.691 | 0.952 | 0.970 |
| Atrophy | |||||||
| 7 | Atrophy | 0.914 | 0.806 | 0.856 | 0.763 | 0.838 | 0.872 |
| Calcification | |||||||
| 8 | Calcification | 0.887 | 0.779 | 0.840 | 0.735 | 0.818 | 0.870 |
| Vascular malformation | |||||||
| 9 | Vascular malformation | 0.890 | 0.776 | 0.835 | 0.717 | 0.814 | 0.876 |
| Hypodense lesion / area | |||||||
| 10 | Hypodense lesion / area | 0.889 | 0.723 | 0.862 | 0.724 | 0.816 | 0.864 |
Y’2021
“
BrainScan CT - decision support system for the interpretation of head computed tomography based on machine learning is a medical product labeling the CE2274 mark.
BrainScan company develope a high quality product, as evidenced by received certificates confirming compliance with the requirements of the ISO13485: 2016 standard and the requirements of the MDD 93/42 / EEC Directive ».
“
Benefits:
- Shorter time between examination and scan description
- Early alert - detailed auto-detection of pathological changes
- Increased number of analysed CT scans in time unit, cause higher efficiency of radiologists.
- Downsed error level of incorrect diagnosis by omitting existing pathological changes in the CT scans.
- Empowering the image of the facility through the use of the latest technologies and scientific achievements.
- Double verification of the diagnosis
- Increased work efficiency, give human beings more time to analyze much more difficult cases
Partners
Awards & prizes
BrainScan Sp. z o.o. implements the project "BrainScan - a system for automatically supporting medical diagnostics, detection and location of potential pathological changes occurring in imaging of the head", co-financed under the Intelligent Development Operational Programme 2014-2020 (IDOP) measure 1.1 / sub-measure 1.1.1 co-financed by the European Regional Development Fund (ERDF). Application # POIR.01.01.01-00-0573 / 18.