After his godmother and his business partner’s mother both died of breast cancer, Rubens Fernando Mendrone made a radical change: he quit his job at IBM to become a high-tech innovator and social entrepreneur.
Both women felt fine, so they had put off going for a mammogram. By the time doctors discovered their tumours, the disease had spread. Both died within six months of receiving their diagnosis.
A software engineer and artificial intelligence specialist, Mendrone set out to help other women avoid a similar fate. His solution is Linda, an app on a hand-held smartphone-sized device that uses body heat to detect suspicious patterns in breast tissue that could indicate a tumour.
Early evidence suggests it provides similar results to a mammogram without the hassle and discomfort, at a fraction of the cost. It is small enough to be available in doctors’ offices and it gives instant results.
Linda Lifetech was founded in Brazil, where more than 20,000 women used the app in the past year, but is moving its headquarters to an office on Toronto’s waterfront as it makes plans to enter the North American market.
“We truly believe that Linda is our legacy,” he said. “I think every woman can benefit.”
Mendrone’s app is a novel approach to an old problem: how to find malignant tumours before they pose a threat to a patient’s life.
Regular screening programs for lung, colorectal, cervical and breast cancers have already helped to reduce mortality, but participation is often below target levels. But now, a number of new technologies are emerging that are set to revolutionize how screening takes place, making it easier and more convenient than ever to get tested.
In Canada, screening for cervical, colorectal and breast cancer is recommended for people who are at average risk of developing those diseases. Screening for prostate cancer is also available on the advice of a doctor. However, in most cases, participation is still low.
Provincial health agencies want to see 60 percent of Canadians older than 50 screened for colorectal cancer, but no province has reached that goal.
There are many reasons for low screening rates, said Elizabeth Holmes, senior manager of health policy at the Canadian Cancer Society. They include not having a regular health care provider, being unaware that screening is available, or being unable to travel long distances for it.
In the case of colorectal cancer, the “ick factor” can put people off. In recent years, home test kits that involve collecting a stool sample and sending it for analysis have replaced colonoscopies for some people who aren’t considered at high risk, but uptake is still low. In New Brunswick, residents receive the kit for their 50th birthday, but just one-third return the test.
“It’s really easy to do,” said Holmes. “But some of these tests are quite intimate. They’re things we don’t talk about.”
Numerous organizations are searching for ways to boost participation. The Canadian Cancer Society has partnered with MaRS Discovery District to challenge researchers to design innovative solutions that would improve screening in underserved populations.
The competition, which has a $150,000 prize fund, is taking applications and will announce winners in April.
Tests that are easy to use, yet look for multiple diseases, form another promising avenue. California-based biotechnology company GRAIL has developed a blood test that can detect more than 50 types of cancer with one sample. Researchers say it holds promise for pancreatic and ovarian cancers, which have few early symptoms and no screening methods.
Meanwhile, a growing body of evidence shows that breath can be used to detect a variety of illnesses including cancers, irritable bowel syndrome and multiple sclerosis. New Brunswick based Breathe Biomedical is developing the first breath analytics technology that uses artificial intelligence and machine learning to spot early signs of lung cancer.
The test, which involves breathing into a machine about the size of a microwave, takes around half an hour. It collects volatile organic compounds found in the breath, and the sample is then analyzed in a laboratory using 200 different frequencies of infrared light and machine-learning software that looks for biomarkers of disease.
The technology is still in development, but Breathe BioMedical chief executive Stephen Graham says it could eventually be used in pharmacies and doctor’s offices across Canada. It could also be deployed to remote northern nursing stations and other underserved communities.
“It really levels the playing field,” Graham said.
Early results from Breathe’s prototypes found the test could detect early stage lung cancer with 86 percent accuracy, early stage breast cancer with 81.4 percent accuracy and COVID-19 pneumonia with 82 percent accuracy. The technology could also be applied to other illnesses, such as Alzheimer’s, Parkinson’s, a range of neurological disorders, tuberculosis and malaria.
Graham says the breath test would not replace current screening technologies, such as mammograms or CT scans, but could become part of an annual screening test.
He says a significant challenge is convincing the medical research community to accept breath analysis based on artificial intelligence as a reliable tool. “AI in medicine can be seen as a black box,” he said. The company is working with researchers at several Canadian universities to help prove the concept.
Mendrone also believes his technology will improve access to screening for many women. In Brazil, just 11 percent of eligible women had a mammogram last year. While these tests are more widely available in North America, his technology, Mendrone says, will be useful for millions of women who can’t easily access a mammogram or who don’t qualify for screening.
If Health Canada approves the technology, Mendrone aims to place a Linda scanner in every family doctor’s office and hopes it will eventually become a standard part of a woman’s annual physical.
The scanner uses infrared sensors to produce a heat map of the patient’s breasts. The image is taken just as a photograph would be, with no touching or radiation involved. Instead, a sensor looks for hot spots that suggest there is increased blood flow required to feed a tumour. Using machine learning, it indicates if there is anything suspicious that warrants further investigation.
The first peer-reviewed study of Linda’s results is pending publication. Mendrone says the study found his technology is just as effective as a mammogram in picking up early stage tumours. Its main use, however, will be to help doctors make decisions about which patients require further screening.
Mendrone said Brazilian women were reluctant to use the scanner at first as they thought it would take a regular photograph of their breasts. “But once they met Linda and saw how the images are, they said, ‘Oh, my God! I need to do this. This should be in my neighbourhood.’”
Learn more about the Colorectal Cancer Early Detection Challenge at a free webinar on February 1.
Photo credit: Linda Lifetech