A new study highlights the potential predictive power of genome sequencing.
International medical humanitarian organization Médecins Sans Frontières/Doctors Without Borders (MSF) gives the world a failing grade when it comes to the global response to tuberculosis.
Sure, MSF doctors and volunteers are not schoolteachers, but they believe there are important—and, all too frequently, unheeded—lessons in the ongoing fight against the deadly disease. In fact, the data, courtesy of the World Health Organization (WHO), is stark: There were more than 1.6 million deaths attributable to tuberculosis and more than 10 million new cases of the disease (in developing countries) in 2017 alone. Remarkably, tuberculosis remains under-diagnosed, with more than one-third of those infected not aware that they have the disease, according to WHO.
And worse: Drug-resistant tuberculosis remains a significant challenge, according to MSF, with uptake of the WHO’s recommended treatment—bedaquiline—sluggish at best. The WHO estimates that more than 500,000 tuberculosis cases annually involve drug-resistant forms of the disease.
However, good news may be on the horizon, at least based on research published on October 11 in the New England Journal of Medicine (NEJM), in which members of the so-called CRyPTIC Consortium and the 100,000 Genomes Project suggest that DNA sequencing can be used to “accurately predict profiles of susceptibility” to anti-tuberculosis drugs. A related commentary appears with the article.
“The goal of this work is to develop whole-genome sequencing as an alternative to more time-consuming methods to diagnose drug-resistant tuberculosis,” explained Megan Murray, MD, DPH, professor, department of epidemiology, Harvard TH Chan School of Public Health, who was among several US-based researchers involved in the project in an interview with Contagion®. “[Our] study shows that existing sequencing technologies can provide results that are as accurate as phenotypic [approaches], at least for first-line tuberculosis drugs.”
Indeed, the international team of researchers, led by Timothy Walker, a specialist registrar in Infectious Diseases and Microbiology and DPhil student in Clinical Medicine at Oxford in the United Kingdom, obtained whole-genome sequences and associated phenotypes of resistance or susceptibility for the tuberculosis drugs isoniazid, rifampin, ethambutol, and pyrazinamide for disease isolates from 16 countries in 6 continents. For each tuberculosis isolate, they identified mutations associated with drug resistance and susceptibility for 9 genes and used them to compile complete susceptibility profiles.
In all, the team analyzed more than 10,000 tuberculosis isolates, and they were able to correctly predict resistance to isoniazid, rifampin, ethambutol, and pyrazinamide with 97.1%, 97.5%, 94.6%, and 91.3% sensitivity, respectively. Furthermore, susceptibility to these drugs was correctly predicted with 99.0%, 98.8%, 93.6%, and 96.8% specificity, respectively.
“We have known for a long time that predicting resistance to the drugs isoniazid and rifampicin is highly accurate from bacterial sequence data and there are several commercially available point-of-care tests that have been highly adopted globally that do just that,” noted Dr. Murray’s colleague and co-author Maha R. Farhat, MD, CM, assistant professor, Biomedical Informatics, Harvard Medical School. “[I]t’s not really surprising that whole-genome sequencing can improve the performance beyond what these tests can achieve and extend the range of drugs for which prediction is possible because it can capture the DNA pattern across most of the tuberculosis genome.”
Although the work of the CRyPTIC Consortium and the 100,000 Genomes Project researchers highlights the power of genome sequencing in identifying optimal treatment for tuberculosis in general—and drug-resistant tuberculosis specifically—it is not without its limitations. The authors were quick to acknowledge that the technology is resource-intensive and expensive, and that it requires a preparatory bacterial culture currently unavailable at point of care. At present, it can take several days, if not weeks, to obtain susceptibility data in a clinical setting.
“It is still not clear if whole-genome sequencing can be performed rapidly enough and at scale in resource-limited settings where it is needed the most,” Dr. Farhat said. “Also, the data presented in this study show clearly that the performance of the prediction from whole-genome sequencing declines quite a bit in settings with higher rates of antibiotic resistance which arguably is where they are needed the most.”
However, she and her colleagues remain sanguine about the potential for the approach at improving tuberculosis treatment globally. Countries such as the United Kingdom are already successfully using it in the diagnosis of drug-resistant tuberculosis, and clinicians are able to get results in a matter of days, according to Dr. Murray. Walker told Contagion® that he hopes the team’s findings will “help build the case for working towards a portable diagnostic test that helps tailor therapies based on genomic/genetic information derived from individual patients’ tuberculosis samples, wherever they are.”
“I suspect that in countries with easy access to these technologies, we will see much more rapid diagnosis of [multidrug-resistant] tuberculosis, which means that patients can be started on effective treatment much more quickly than previously,” Dr. Murray added. “This will not only improve patient outcomes but, we hope, reduce the time during which patient can infect others.”
Maybe then the world can finally earn a passing grade in the fight against tuberculosis.
Brian P. Dunleavy is a medical writer and editor based in New York. His work has appeared in numerous health care-related publications. He is the former editor of Infectious Disease Special Edition.