To improve the way lung cancer is diagnosed, a group of students and researchers from the Massachusetts Institute of Technology and Harvard Medical School have begun work on commercialising a breath-based technology. The idea, says Joseph Azzarelli, a team member, is to develop a “single-use device that patients would blow on and that would be scanned by a smartphone” to provide information assisting diagnosis.
The team is among many scientists and researchers working to find better means of diagnosing cancer, from biopsies, mammograms and colonoscopies to magnetic resonance imaging. The aims are to cut costs, avoid painful procedures and radiation exposure, raise accuracy and detect earlier.
“The science is still preliminary but others have shown that there’s a correlation between the molecules present in exhaled breath and the occurrence of lung cancer, and potentially other cancers,” says Mr Azzarelli.
One of the reasons for developing the breath-based test — which uses a card slightly larger than a postage stamp — is to reduce the cost and inconvenience of lung cancer testing. This would allow more people to be tested regularly.
“We’re building the card to be responsive to early-stage lung cancer, knowing those patients do much better in terms of treatment and chance of cure,” says Alexander Blair, another team member. The team won this year’s Accelerate and Launch prizes, two of three contests that make up the MIT $100K, an MIT-run entrepreneurship competition.
Nanotechnology, meanwhile, is behind the technique being developed by Miami-based start-up Entopsis. It uses a piece of glass designed to capture nanoscale molecules, proteins and cells from a body fluid (it is currently using urine) and bind them to its surface, creating a molecular profile of patients.
Rather than looking for predetermined biomarkers for cancer, the technique works by using cloud-based algorithms. Analysis compares the molecular profile of the patient’s sample with a database containing the molecular profiles of people with a range of cancers.
“If patients with breast cancer have a specific molecular profile and your sample matches that with high confidence, then you have breast cancer and the computer will tell you, and how certain it is of that answer,” says Obdulio Piloto, Entopsis chief executive.
“We’re training the computer to learn what human disease looks like.” The aim, he adds, is to have “one test that’s non-invasive and offers a low-cost diagnosis” for a wide range of conditions.
While some technologies for cancer diagnostics are in early development, much interest is focused on “liquid biopsy”, a means of diagnosis from DNA that has leaked from tumours into the bloodstream. It allows detection in organs such as the pancreas, which lie deep inside the body and cannot easily be screened.
People are conventionally tested after they present with clinical symptoms. “By that time, the disease has often spread beyond the pancreas,” says Anirban Maitra, co-director of the University of Texas MD Anderson Sheikh Ahmed bin Zayed al-Nahyan Center for Pancreatic Cancer Research. Liquid biopsies offer the chance of early-stage detection and by way of a far less costly and painful test. For those who have the cancer, liquid biopsies avoid the need for repeated invasive biopsies, which both cost thousands of dollars and run a higher risk of complications.
“If you can sample the bloodstream and extract the cancer-specific information”, says Dr Maitra, “you can get a pretty informative view into the disease itself without having to obtain a biopsy.”
While the tests are becoming less expensive, more sensitive and easy to use, this can come with its own problems.
“The impression is that if we could use more and more sensitive detection assays for mutations, the better our chances of getting an accurate read for a liquid biopsy,” says Dr Maitra. “But sometimes the pendulum can swing too much the other way.”
He cites recent research on ovarian cancer showing that some of the patients who did not have ovarian cancer had circulating mutant DNA in their blood. “We are finding the same in pancreatic cancer,” he says.
“As liquid biopsies are being promoted as a cancer-screening tool for the general population, we have to accept the fact that there is an element of over-diagnosis,” says Dr Maitra. “And we have to find ways to make sure we swing that pendulum back to the centre.”
Get alerts on Harvard Medical School when a new story is published