From AI to Immunotherapy, Scientists Are Innovating to Eliminate TB


Apr 11, 2019


With only 17% of the world’s population, India is home to roughly a quarter of the world’s TB cases — a disproportionate burden. But only some patients are being treated; while 2.7 lakh tuberculosis cases were recorded in India in 2017, it is estimated that another nearly 1 lakh TB cases were never recorded, nor received formal treatment.

Where there is great burden, there is also opportunity for innovation. As part of the government’s efforts to eliminate TB by 2025, it launched the Tata Trusts-seeded TB Quest, a fund to foster new pharmaceuticals and technologies to help in the fight against TB. But what are these innovations? What are researchers trying to fix?

Improving TB diagnosis

One of the biggest hurdles in TB control is timely diagnosis. Tuberculosis is very difficult to diagnose by symptoms alone; its main signs — fatigue, coughing, chest pain and fever — are associated with any number of infections and conditions. A chest x-ray is usually needed to put doctors on the right track — but so are specialist radiologists who can interpret the x-ray results (interpreting lung x-rays is particularly difficult). Access to and follow through in getting x-rays are low; so are the numbers of trained radiologists: India has roughly one radiologist for every 1 lakh people, compared to Europe’s 1 to 10,000.

“On one hand, you want to screen people actively [with x-rays]; on the other hand, you don’t have enough trained professionals to do that,” says Dr. Pooja Rao, the co-founder of Qure.ai, a Mumbai-based start up trying to solve that gap by developing an artificial intelligence that can detect abnormal chest x-rays and refer patients for confirmation tests.

Called qXR, the screening program was developed through deep learning — that is, by running 2.2 million examples of x-rays and radiologist reports of both healthy lungs and lungs with TB through the algorithm until it developed pattern recognition and thus, the ability to spot an abnormal lung scan.

“Computers have gotten very good at understanding and interpreting images,” says Dr. Rao, who is also Qure’s head of R&D. She points to facial recognition software, and even a standard phone’s ability to auto-sort photos by type.

A validation study at Massachusetts General Hospital in Boston, U.S., found the AI was comparable to a radiologist in its ability to detect abnormalities.

Today, qXR is used in India, as well as 11 other countries, in private practices as well as public-private partnerships; it is used in vehicles that are part of proactive screening efforts, as well as in institutions where patients come seeking treatment. “You get results at the point of care,” Dr. Rao says. “That saves a lot of time and transmission.”

The next step for qXR, Dr. Rao says, is to see if it can predict tuberculosis, specifically, from x-rays with the same accuracy as a diagnostic test. If so, it could be a game-changer. Current diagnosis methods suffer from one of two problems: they are either time-intensive, like a culture test, which can take up to eight weeks to confirm TB, or they are costly, like a GeneXpert test, which screens for tuberculosis DNA in the blood. (While government subsidies make GeneXpert tests more affordable, they don’t apply in private settings, which is where most TB patients in India get diagnosed.)

But even after diagnosis, there are still challenges to eliminating TB. The biggest of these, in Dr. Rao’s opinion, aside from the problems Qure is attempting to address, is the spread of TB strains that are resistant to the current methods of antibiotic treatment.

Improving TB treatment

Mycobacterium tuberculosis, the bacteria that causes TB, is highly treatable with antibiotics. However, it does require dedication: a typical course of treatment usually involves a cocktail of several different antibiotics, which may or may not cause unpleasant side effects, and which must be taken strictly according to schedule, across a regimen lasting six to nine months. The strict adherence required over a long period of time often leads patients to lapse on their treatment as soon as their symptoms diminish, which far from curing TB, leads to the development of antibiotic resistant strains.

These versions of the disease are much, much more difficult to treat. They’re also increasingly common in India, rising by perhaps as much as 40% in four years.

Some scientists are on the hunt for new antibiotics that TB strains have not yet been exposed to (and thus have not had a chance to develop resistance to). Last year, researchers at Penn State University, in the U.S., published the results of a study finding that an antibiotic compound, known as kanglemycin A, was effective against some drug-resistant TB strains. Kanglemycin A, a naturally occurring antibiotic found in soil is, loosely speaking, a cousin to rifampicin, a drug considered one of the first-line medications in a typical TB treatment cocktail. But kanglemycin A is different enough to allow it to be effective against strains that have developed resistance to rifampicin’s effects. While not yet developed into or tested as a medication for human use, its discovery is still “a really exciting finding,” said Nikolay Zenkin, one of study’s leaders, giving scientists a promising place to start in developing new drugs.

Other researchers are looking for alternatives, or at least supplements, to antibiotics, in the fight against drug-resistant TB. Earlier this year, scientists in Germany announced they had identified a protein in human immune cells, which when acted upon through immunotherapy, could inhibit the bacteria’s destructive effects. The efforts are not aimed at killing the bacteria, but rather at using corticosteroids to inhibit the effects of the bacteria on the protein, and thereby reduce damage to lung tissue. While corticosteroids have been used to supplement tuberculosis treatment for decades, the team hopes to use their finding to develop more targeted treatments that use the body’s natural defenses against the bacteria.

“If we can support antibiotic treatment with immunotherapy, the duration of treatment would be shortened, which in turn would reduce secondary complications,” explains Jan Rybniker, of the University Hospital, Cologne, and the German Center for Infection Research (DZIF).

Yet another line of research focuses on using the bacteria against itself.
All bacteria produce byproducts that are toxic to themselves; in certain environments and in certain bacteria, these toxins are, essentially, a suicide mechanism; in other environments and in certain other kinds, bacteria also have mechanisms that simultaneously protect them from these toxins. Earlier this year, scientists at the Pharmacology and Structural Biology (IPBS–CNRS/UPS) and the European Molecular Biology Laboratory (EMBL) announced they had found one such “suicide toxin” in Mycobacterium tuberculosis, along with its “antidote toxin.” If the latter could be blocked, the theory goes, the bacteria would essentially kill itself.

At this point, however, it’s just a theory; among other things, there is no clarity on how a toxin so lethal to the bacteria could affect the patient. Unlike Qure’s qXR, which has already undergone testing and is available for use, none of these treatment developments are close to application yet. But they do each represent potential routes of curbing antibiotic resistance as well as the spread of the disease.

Still, there remains one gap that no scientist can fill: proactive seeking of diagnosis and spreading awareness. Stigma around the disease, and misconceptions that equate tuberculosis with poverty, keep people from seeking care, curbing transmission, completing treatment and discussing the disease openly.

“What humans can do is say, ‘Yes, this might be TB,'” Dr. Rao says, and seek diagnosis and treatment. “Even in India, where we have the highest TB burden in the world … a lot of TB is hidden from view. So, you have a perception that TB is a thing of the past.”

“But it’s not really a niche,” she says. “It’s quite widespread.”

Curbing TB, then, may be as much about the innovation of public mindsets, as it is about the innovation of care.


Written By Liesl Goecker

Liesl Goecker is The Swaddle’s managing editor.


Leave a Comment

Your email address will not be published. Required fields *.

The latest in health, gender & culture in India -- and why it matters. Delivered to your inbox weekly.