ScanAim.AI

Project summary

Project presentation

https://youtu.be/c2-3pEUQUNQ

For the first time in the world, a unique and innovative method for increasing the accuracy of artificial intelligence predictions has been proposed, through two-stage and repeated training of a number of neural networks. Based on the proposed method, a computer program is proposed that helps a radiologist, in automatic mode, identify patients with asymptomatic, undetected chronic diseases, initiating earlier diagnosis and preventive treatment. This program will allow achieving stability and accuracy of predictions, even for images with high noise levels, low contrast and blurred boundaries of the objects being searched. The proposed method can be used both for the study of other human organs and in non-medical applications, where poor-quality images are encountered.

Consumer segments:

The program is intended for medical institutions, enterprises that develop and manufacture X-ray medical diagnostic devices, as well as organizations that provide teleradiology services.

Problems and existing alternatives:

Diagnosis and assessment of a compression fracture or determination of the cardiothoracic index from chest X-rays requires morphometric studies, which is a very long and laborious work for a radiologist.

The cardiothoracic index is the main indicator of the presence of cardiomegaly. If the pathology is not detected in time, the risk of heart failure increases and 50% of patients die within 5 years.

Identification of vertebral compression fractures can provide early treatment, as about 66% of compression fractures go unnoticed or unreported in cases of osteoporosis, which affects one in three women over the age of 50 and one in five men worldwide, according to the latest data from the International Osteoporosis Foundation.

Early diagnosis of carotid atherocalcinosis plays an important role in preventing ischemic stroke, after which up to 84-87% of patients die or remain disabled.

Unique value:

The developed artificial intelligence algorithm made it possible to achieve stability and accuracy of neural network predictions even for X-ray images with high noise, low contrast, and blurred boundaries of the objects to be searched for.

What fundamentally distinguishes the developed algorithm from competitors' artificial intelligence algorithms developed for X-ray computed tomography (CT) or magnetic resonance imaging (MRI) images, which are characterized by high contrast and low noise, without superimposing images of human organs on each other.

Solution to the problem:

The proposed program based on artificial intelligence will allow:

- increase the number of patients with early detection of chronic diseases;

- reduce costs for the treatment of chronic diseases;

- expand the capabilities of radiologists in quantitative assessment of disease signs;

- increase the attractiveness of the medical institution through advanced and innovative capabilities.

Promotion channels:

Internet

medical exhibitions

Sources of income:

Making money can take three years

channels:

- selling the program to medical institutions;

- sale of the program as an OEM product to companies that develop and manufacture X-ray medical diagnostic equipment;

- sale of the program to organizations that provide teleradiology services.

Cost structure:

The team members' salaries are planned to be $12,600 for 6 months of work.

A senior research fellow who will develop machine vision algorithms and neural network ensembles will be paid a salary of $800 per month.

A senior software engineer who will write the program's code and also be involved in its testing will be paid a salary of $800 per month.

The SMM Manager/Designer, who will be responsible for the design of the application's user interface, as well as the marketing company, will be paid a salary of $500 per month.

To develop machine vision algorithms, train neural networks, and develop a user interface, you need a high-performance personal computer with large RAM. It is planned to spend $1,900 on this.

The ScanAim.AI program is a medical device and therefore must undergo certification. This will cost $3,000.

It is planned to spend $5,100 on administrative expenses, including rent, telephone/internet, website support, salaries for the director ($300 per month) and accountant ($200 per month).

It is planned to spend $2,400 on a marketing campaign (website development, advertising).

Key metrics:

The success of the product will be determined by the number of sales of the program per year

Hidden advantage:

The algorithm of the proposed program is very complex and was developed over 2 years. The development of the program requires knowledge in the field of medical physics, neural networks and, of course, programming skills in various computer languages.

Competitor analysis:

Market analysis methodology:

We studied information obtained from the Internet. A special place was taken by the article "How Zebra Medical Vision Developed Clinical AI Solutions" ( https://www.datarevenue.com/en-blog/ai-diagnostic-solutions-zebra-medical ). As well as materials on the website https://www.nanox.vision/ai

Market analysis:

The US healthcare AI market is projected to grow from $2.9 billion in 2022 to $51.3 billion by 2030, at a compound annual growth rate (CAGR) of 43.22%, driven by the need for more efficient healthcare services and potential cost savings. ( https://www.itnonline.com/content/lunit-announces-completion-volpara-acquisition-advance-ai-driven-cancer-care ).

The global teleradiology market was estimated at US$8.6 billion in 2023 and is expected to reach US$29.8 billion by 2033.

Development plan:

The program's artificial intelligence algorithm has now been developed, demonstrating the feasibility of the task. The grant funds will be used to create a commercial computer program. The commercial program will be created in three versions:

1) OEM version for developers and manufacturers of software for X-ray diagnostic devices;

2) version for medical institutions;

3) version for organizations that provide teleradiology services.

The program sales strategy will be as follows:

1) The OEM version will be sold, as a one-time purchase, to developers and manufacturers of X-ray diagnostic devices at a price of $15,000;

2) the version of the program for medical institutions will be sold as a license for one year at a price of $1200. If the license is extended for the next year, the price of the program will be reduced;

3) for organizations that provide teleradiology services, the price for processing one X-ray image will be $0.1.