Diagnosis of cancer: ColonFlag software is 2 times more accurate than tests

Israel's Medial company offers less invasive solutions, based on more data for drug-resistant patients. It is ColonFlag - a predictive software that is twice as accurate as the traditional stool test method.

ColonFlag compares new blood tests with previous diagnosis of patients, as well as Medial proprietary database of 20 million anonymous tests that lasted for three decades and three continents, to assess the possibility. the ability of patients to have cancer.

Israel's second-largest health protection organization is currently using the software, and Medial (works on a combination of 'health' and 'algorithm') is working with Kaiser Permanente and two hospitals. America's leading, to develop other applications for databases and analytical tools.

Picture 1 of Diagnosis of cancer: ColonFlag software is 2 times more accurate than tests
Medial has released a software that predicts colon cancer twice as accurate as traditional stool testing.(Artwork).

'Our algorithms can automatically scan all patient parameters and detect microscopic changes over time, to find the corresponding models for the end result' , Nir Kalkstein, Co-founder and director of Medial technology said, especially about clinical symptoms. The database allows his team to 'find similar symptoms in the past and then identify from data correlations that can lead to end results'. With 45 employees, Medial has become the first company in the application of cheap, easy tests that are highly effective.

Kalkstein has experience with both the application of administrative procedures and big data applications. He performed military service in the Israel Defense Force (IDF) at a high-end cybersecurity research unit, after which he founded a company called Final (short for financial algorithms). on the day of his demobilization, ie in 2001. Although he did not know anything about finance, he excelled in programming to predict stock market activities, based on market reactions. Past school with similar events.'We just look at data and not use any economic model,' Kalkstein said.

According to the Jerusalem Post, Kalkstein eventually became a billionaire, but he still didn't see enough. By 2009, he said, he decided to "invest his time and resources in things that have a positive impact on people's lives ." With some of the students from the university and IDF, he started with Medial at a garage near his home in the Tel Aviv suburb, recruiting Ori Geva technology expert as chief executive, and consultant Ofer Ariely was president.

Although they lack the knowledge of health care, they learned quickly. In 2011, Medial held an informal competition with intensive care doctors at Rabin Medical Center, Israel's largest hospital, to predict which cancer patients would survive. Data scientists have won over medical doctors. Varda Shalev, a leading doctor, who also runs research and development for Maccabi health services, currently uses ColonFlag, said: 'Considering many aspects at the same time is very difficult, but With data and technology, this is easy. '

According to a Kaiser study published last year in the Journal of Digestive Diseases, the software is ten times more likely to identify patients at risk of cancer from 6 months to 1. year, compared to the doctor's diagnosis, while those cases are still under observation.

Currently, Kalkstein's group is focusing primarily on R&D. Medial raised $ 40 million from Hara Ventures investment fund of Hong Kong billionaire Li Ka-shing, and the company said its next product, expected in the second quarter of 2018, will be related to diabetes. .

Further: the wearable device can predict epilepsy, enabling patients with this disease to find the safest place before getting sick. This device is also based on data analysis to predict symptoms including heart failure, acute kidney damage, and infection. Kalkstein said: 'There is a lot of work to do, but I am willing to devote all my resources to investing in this industry; because no industry has great potential like this. '