The application can soon assist clinicians worldwide to produce faster and more accurate reports for improved patient outcomes.

Interpreting and reporting vascular ultrasound studies is commonly done through a manual, time-consuming, and error-prone process.

To report a vascular ultrasound scan, a clinician, usually a sonographer or radiologist, has to manually review and analyse 50 to 150 individual images consisting of various types of ultrasound images and doppler waveforms of each patient. The end result is a hand-written, paper-based template filled with drawings, numbers, and measurements, which can take as long as 20 minutes per patient in some complex cases.

This may soon be redundant after 10 Dec 2019, the day on which Singapore’s Health Sciences Authority (HAS) approved a device called Augmented Vascular Analysis (AVA) invented by See-Mode Technologies—a medtech startup based in Singapore and Australia.

AVA is an AI-based software for automated analysis and reporting of vascular ultrasound scans, one of the most common scans used for patients with cardiovascular diseases. AVA utilizes multiple deep learning models for image analysis, as well as text recognition and signal processing algorithms, and puts together a computer-generated report in less than one minute with a single click. This renders the application as an intelligent assistant to the clinician interpreting and reporting vascular ultrasound studies.

After this vascular ultrasound report is generated by the software, the control is handed over to the clinician who reviews the report and if necessary, makes adjustments before confirming the final report. With this methodology, See-Mode’s AI augments the clinical workflow, resulting in greater overall productivity, accuracy and improved patient outcomes.

Reducing the chance of human error

As See-Mode’s first product, AVA, has been approved as a Class B medical device. Said Dr. Sadaf Monajemi, co-founder of See-Mode Technologies: “During our collaboration with different hospitals, we observed cases where the mistakes in hand-written ultrasound worksheets could potentially result in human error. Given that these reports are produced by clinicians who acquire and review hundreds of images per day, human error is inevitable. Our intention with AVA is to give an assistive tool to clinicians to improve efficiency and minimize potential errors without making any changes to their established clinical workflow.”

With the HSA approval in hand, See-Mode is now pursuing regulatory approval for AVA in other regions, including Australia’s Therapeutics Good Administration, Europe’s Conformité Européene (CE) and USA’s Food and Drug Administration. Dr. Milad Mohammadzadeh, See-Mode’s co-founder, said: “While we work towards getting AVA to a larger user base, we are continuing on our mission of assisting doctors to predict and prevent stroke. See-Mode is building machine learning and computational models for extraction of stroke biomarkers from CT and MRI. We are collaborating with some of the most prominent stroke centers in the world to assess these models and introduce them into hospitals, aiming to help doctors improve patient outcomes.”

Seed-funding pays off

As a startup, See-Mode announced its seed funding round of US$1M in early 2019, with participation from SGInnovate and Cocoon Capital in Singapore, and Blackbird Ventures in Australia. Said Steve Leonard, Founding CEO, SGInnovate: “As people around the world grow older and live longer, there is an increasing need for new healthcare solutions. Globally, rising demands for greater healthcare infrastructure and resourcing are exceeding the ability of public and private systems to meet them. SGInnovate firmly believe AI, as a group of technologies, will be vital in helping medical professionals perform their jobs more effectively and efficiently.”

As one of the earliest investors in See-Mode, SGInnovate is happy to know that doctors and clinicians will be able to improve patient care using the AI systems being built by the See-Mode team.