Chest X-rays and sputum scores flag who may fail shortened all-oral MDR-TB therapy

Shortened, all-oral regimens have transformed the treatment of multidrug-resistant or rifampicin-resistant tuberculosis, replacing months of painful injections and toxic drugs with tablets that patients can actually tolerate. Yet even as these World Health Organization-endorsed regimens improve outcomes, a stubborn minority of patients still die, fail treatment, or relapse. A new prospective cohort study from Lesotho, published in Nature Communications, offers a practical answer to a question that has haunted the field: can clinicians tell, on day one, which patients are at highest risk?

The research, conducted under the STEM-TB observational study and registered as NCT05871489, followed 238 adults with bacteriologically confirmed pulmonary MDR/RR-TB who were treated with a shortened all-oral regimen. The cohort was strikingly representative of the epidemic in southern Africa: the median age was 41 years, only 29 percent of participants were female, and 62 percent were living with HIV. Rather than examining drug exposures or resistance patterns, the investigators focused on two features that are available in almost every clinic that diagnoses tuberculosis: the appearance of the chest radiograph and the grade of the baseline sputum smear.

The microbiologic logic of the study rests on a single central metric, time-to-culture-conversion. When a patient’s sputum cultures flip from positive to negative, it signals that the drugs are killing the bacilli faster than they can replicate, and delayed conversion is a well-established warning sign of eventual failure. Among the 148 participants whose baseline cultures were positive, an impressive 142, or 95.9 percent, converted within four months. That figure confirms that the shortened all-oral regimen is highly effective at clearing culturable bacilli for the vast majority of patients, even in a population with a heavy burden of HIV.

But the aggregate success concealed meaningful variation. Four baseline features were consistently associated with delayed culture conversion and lower odds of treatment success: fibrosis, cavitation, bilateral disease, and a high sputum smear grade. Each of these markers reflects a different dimension of disease severity. Cavitation indicates that the infection has destroyed lung tissue and created air-filled spaces where bacilli can proliferate behind a physical barrier to drug penetration. Fibrosis signals chronic scarring that both harbors persistent organisms and impairs the local immune response. Bilateral disease means the bacilli have spread across both lungs, and a high smear grade simply quantifies the enormous bacillary load the patient carries at diagnosis.

The statistical findings were most compelling when these features appeared in combination. Cavitation together with a high smear grade was negatively associated with favorable final post-treatment outcomes, with an unadjusted odds ratio of 0.29 and a 95 percent confidence interval of 0.11 to 0.78; after adjustment, the odds ratio was 0.39 with a confidence interval of 0.14 to 1.14. Even more striking was severe fibrocavitary disease, defined as multi-lobar fibrosis with at least one cavity measuring five centimeters or greater, which carried an unadjusted odds ratio of 0.20 and an adjusted odds ratio of 0.26. In plain terms, patients whose lungs showed this combination of extensive scarring and large cavities, and whose sputum teemed with bacilli, faced roughly a quarter to a fifth the odds of a good final outcome compared with patients whose disease was less advanced.

One of the most technically interesting aspects of the analysis was the way effect modification by HIV status reshaped the picture. Bilateral disease was more strongly associated with unfavorable outcomes among participants without HIV, while fibrosis and high smear grade showed stronger associations among those living with HIV. This pattern makes biological sense. In people with intact cellular immunity, the spread of disease across both lungs may reflect an uncontained inflammatory process, whereas in immunocompromised patients, the chronic scarring and high bacillary burden may dominate the clinical trajectory because the immune system cannot compensate for the damage already done.

Overall, the study recorded 27 unfavorable outcomes among the 238 participants: 20 deaths, 3 treatment failures, and 4 recurrences. That translates to roughly 11 percent of the cohort experiencing an unfavorable result, a figure that is low by the standards of drug-resistant tuberculosis but far from zero. The distribution of those outcomes matters as much as the count. Deaths outnumbered failures and recurrences combined, underscoring that the residual risk in shortened all-oral therapy is not merely a matter of the drugs failing to clear the infection but of patients dying during or after treatment from causes that may include the advanced disease itself.

The practical implications reach beyond Lesotho. Because chest radiographs and sputum smears are inexpensive and universally available, the findings suggest a risk-stratification tool that requires no new laboratory infrastructure. A clinician who sees bilateral disease, cavitation, extensive fibrosis, and a high smear grade on the day of diagnosis can identify a patient who may need closer monitoring, more frequent culture follow-up, or a longer or augmented regimen. Conversely, patients whose radiographs show limited disease and whose smears are low-grade can be treated with confidence in the shortened regimen, avoiding unnecessary prolongation and its attendant costs and toxicities.

The study also highlights the value of prospective cohort designs in tuberculosis research. Rather than pooling retrospective data or relying on trial populations that may exclude the sickest patients, the investigators enrolled consecutive adults treated in routine program settings through Partners In Health Lesotho and the Lesotho Ministry of Health’s National Tuberculosis and Leprosy Programme. The cohort’s high proportion of people living with HIV reflects the reality of the epidemic in the region, and the results therefore carry weight for programs deciding how to deploy shortened all-oral regimens in high-burden settings.

For the field at large, the message is that the era of one-size-fits-all drug-resistant tuberculosis treatment may be giving way to risk-adapted strategies. The shortened all-oral regimen remains the backbone of care, and the overwhelming majority of patients, even those with positive baseline cultures, achieve culture conversion within four months. But the combination of radiographic and microbiologic severity at diagnosis now provides a scientifically grounded way to identify the minority who need something more. As programs scale up these regimens across high-burden countries, the humble chest X-ray and sputum smear may prove to be the most powerful prognostic tools available, guiding clinicians toward the patients whose disease demands vigilance from the very first tablet.

Subject of Research: Tuberculosis

Article Title: Baseline radiographic and microbiologic predictors of treatment outcomes in a shortened all-oral regimen for MDR/RR-TB

Article References: Shen, E. Q. L., Mpinda, S., Maama, L., Tamirat, M., Sebakeng, P., Oyewusi, L., LaHood, A., Trevisi, L., Romo, M., Osso, E., Rich, M. L., Mitnick, C. D., Seung, K. J., Franke, M. F., & Mikanda, K. K. (2026). Baseline radiographic and microbiologic predictors of treatment outcomes in a shortened all-oral regimen for MDR/RR-TB. Nature Communications. https://doi.org/10.1038/s41467-026-78700-9

Image Credits: AI Generated

DOI: 10.1038/s41467-026-78700-9

Keywords: Baseline, radiographic, microbiologic, predictors, treatment, outcomes, shortened, all-oral, regimen, RR-TB, scientific research