Blood test interpretation using AI

Based in Zurich, Switzerland, Smart Blood Analytics Swiss SA is a pioneering R&D company that has championed the use of advanced machine learning algorithms in AI models that can more accurately and efficiently predict diseases and medical conditions based on individual’s blood test results solely. The company released on the market two medical devices. First is SBAS Software which is intended for physician, and mySmartBlood, a mobile application intended for symptomatic end-users. We find out more following the company’s recent award for Most Innovative Machine  Learning Blood Analytics Tool, 2021.

For effective detection, diagnosis and treatment of more than 70 % of diseases, blood tests results are vital. Thanks to timely and accurate blood tests results, professionals and leading authorities of the global medical and healthcare sectors can ensure diagnoses and early intervention for patients that will facilitate better quality of care further down the line and efficiency in managing downstream costs.

However, the interpretation of blood test results can become unnecessarily complicated when multiple parameter values are out of the normal range – or indeed, even when all of those values are within the normal range.

Fortunately, Smart Blood Analytics Swiss SA (SBAS) offers a solution. The Zurich-based company was founded in 2016 by dr. Marko Notar, who still sits at the helm of SBAS and its work at the forefront of the world’s medical research sector. “The current health care system is increasingly dependent on reliable blood test results, the true power of which is frequently underestimated,” states dr. Notar on the SBAS website. “They carry way more information than we ever thought possible. Blood test results represent information and not the knowledge itself. Together they form the true power of Artificial Intelligence, which is efficiently combined in our products.”

A small but flexible and highly specialised organisation with globally renowned expertise in the interpretation of blood test results, SBAS is defined by its innovative approach to its science. This approach primarily centres on the belief in the power of artificial intelligence and its benefits to the future of medical research. Using advanced machine learning algorithms, SBAS is able to provide a broad portfolio of AI models that can accurately and efficiently predict the most probable diagnoses based solely on an individual’s blood test results from the field on internal medicine and neurology. As a result, SBAS is able to offer probable diagnoses and expanding differential diagnosis leading to more effective treatment and better overall outcomes.

This process has been perfected in the firm’s first flagship product, the SBAS Software. A cloud-based software service that is accessible through either an API or web-application, SBAS Software employs a two-step process.

The first step is to gather its patients’ blood test results that include complete blood count with differential and common parameters from biochemistry, haemostatic, enzyme, lipid, protein & amino acids, vitamins and hormone panels. This usually provides sufficient data for the SBAS Software models to predict the most probable disease, confirm a human-made diagnosis, or expand the differential diagnosis.

These ground-breaking models have been built on SBAS’s signature machine-learning algorithms to analyse hundreds of groups of medical conditions and diseases across the field of internal medicine, ranging from rheumatology to cardiology, toxicology, haematology to endocrinology, pulmonology and nephrology to gastroenterology and field of neurology. The SBAS Software is also able to recognise disease-related blood laboratory patterns that go beyond existing medical knowledge. As a result, SBAS’s solutions offer higher diagnostic accuracy compared to traditional quantitative interpretations based only on reference ranges.  

SBAS is then able to deliver a comprehensive report complete with a list of ten of the most likely diseases with the corresponding ranking factors and parameter evidence and an intuitive multilevel pie chart – a valuable asset in increasing the efficiency and improving the accuracy of a doctor’s diagnostics.

Consequently, SBAS’s pioneering solution is recognised internationally for its benefits in improving the diagnostic process as it enables doctors to make more accurate and more efficient decisions, resulting in better outcomes for patients and reduced costs for healthcare systems. As a result, SBAS’s team of professionals within the fields of medicine, machine learning, laboratory biomedicine and chemistry have since been called upon to take their solution even further.

As a result, SBAS has launched its research and development of an AI diagnostic test for Covid-19 infected patients that uses only a routine blood test results. Research around the viral infection that has been at the centre of worldwide attention for over a year and a half has found that symptomatic Covid-19 patients develop characteristic changes in their routine blood parameters. However, it is usually only a small fraction of information contained in routine blood test results that is actually extracted.

With machine learning, a diagnostic is able to better identify subtle patterns within the data extracted, making it a more ideal tool for recognising the indicators of Covid-19’s presence. Differentiating Covid-19 from 100 other viral and bacterial diseases, based on only 35 routine blood test results, SBAS’s COVID-19 diagnostic model achieves sensitivity of 83.5% and specificity of 99.1% (N=4434). Designed for adults presenting to emergency rooms and hospitals, SBAS’s COVID-19 diagnostic model is fast establishing itself as a highly-valued asset in the global pandemic that continues to dominate headlines over twenty months after it first entered public awareness.

Since SBAS Software is made to be used only by physicians, the SBAS team also felt the need to offer something to the end-users. They launched a mobile application for iOS and Android mobile devices called mySmartBlood, which operates by using the SBAS Software model. It is an in vitro diagnostic medical device for symptomatic users that provides a list of potential groups of diseases based solely on users’ blood test results.

The users simply download their Mandatory List of Parameters (MBPL), measure the required blood parameters in the nearest laboratory, enter the results in the mySmartBlood application, and create a report. The report can be viewed directly on the phone, or it can be downloaded as a PDF document that users can share with their physician. It’s a great tool that can guide its users to the right healthcare professional, enables physicians to make more accurate diagnostic decisions, and significantly shortens the path to the final diagnosis of the users.

Now, SBAS’s research is pushing the abilities of its AI-driven models even further, pursuing the answers to questions such as whether other serious medical conditions like brain tumours could be diagnosed through routine blood tests. Although many within the medical and healthcare sectors still show reluctance to embrace the power of AI in advancing scientific research within this sphere, SBAS is pioneering the move to its acceptance.

From demonstrating its algorithm’s alignment to the Seven Principles of Healthcare Ethics – those being autonomy or self-determination, beneficence, nonmaleficence, confidentiality, justice, role fidelity and veracity – to its publication and recognition on internationally acclaimed platforms including the Scientific Reports and the 38^th Annual J.P Morgan Healthcare Conference in San Francisco, SBAS is demonstrating its increasing and impressive credibility.

Now, in addition to its recent receipt of its CE mark for SBAS Software and mySmartBlood and ISO 13485 standard certificate stating that Smart Blood Analytics Swiss SA meets all EU safety and health requirements and is cleared for entry into the EU and EEA areas, SBAS is celebrating further success. The firm has won the award for Most Innovative Machine Learning Blood Analytics Tool, 2021 at the GHP – Global Excellence Awards, a testament to the advanced and pioneering technology developed by a team with unparalleled expertise and commitment to the progression of medical science.

For further information, please contact Ziga Osterc or visit https://www.smartbloodanalytics.com/en/home