Pulmonary tuberculosis (PTB), caused by drug-sensitive strains of Mycobacterium tuberculosis (DS-TB), is treated with multiple drugs for 6 months (1). This long duration of treatment with multiple drugs has a cost implication and a risk of non-adherence predisposing to the development of drug resistance. In recent times, many trials have been done to guide the design of shorter treatment regimens for PTB (2–5). A reliable biomarker of cure is a crucial tool for the determination of sterilizing cure of tuberculous disease. None of the currently utilized biomarkers is sufficiently reliable in predicting cure. The current gold-standard technique for the assessment of successful treatment of tuberculous disease is sputum culture. Sputum culture may be negative despite persisting bacilli. A reason for negative culture in the presence of live bacilli may be due to the slow-growing nature of these bacilli that makes them non-culturable (6). Therefore, other biomarkers that can reliably predict sterilizing cure of tuberculosis are needed for the rational design of a shorter treatment course of ATT.

Metabolic imaging with fluorine-18 fluorodeoxyglucose (FDG) using hybrid positron emission tomography/computed tomography (PET/CT) scanners has been evaluated for their utility as a non-invasive tool for therapy response assessment in patients with tuberculosis. Findings on FDG PET/CT obtained at baseline can reliably discriminate between responders vs. non-responders to anti-tuberculous therapy (ATT) (7, 8). Similarly, FDG PET/CT obtained as early as 1 month into ATT can predict successful treatment outcome (9). More importantly, our group and others have recently shown that patients with residual metabolic activity (RMA) in their lungs on FDG PET/CT obtained at treatment completion have a higher risk of TB relapse (10, 11). Conversely, patients who achieved complete metabolic response to ATT at end-of-treatment (EOT) FDG PET/CT have a durable cure with no risk of TB relapse (11). These studies support the utility of FDG PET/CT as a useful non-invasive tool for treatment response assessment in patients treated for tuberculosis.

Despite its excellent performance as a tool for treatment response assessment, the application of FDG PET/CT is unlikely to become routine in clinical practice due to cost and availability issues. The availability and utilization of molecular imaging techniques are still very limited in developing countries where most patients with tuberculosis live (12, 13). Computed tomography (CT) is a more widely available imaging tool in TB-endemic regions of the world. Certain features on CT, including lung consolidation, lung nodules in the typical tree-in-bud pattern, pleural effusion, and others, are known to suggest active PTB (14–16). Determining the strength of the association between these CT features suggestive of active TB and the RMA on FDG PET is crucial for guiding the application of CT, a more widely available imaging technique, for treatment response assessment of patients treated for PTB. Therefore, in this study, we aimed to determine the association between RMA on FDG PET and CT features of active disease in patients treated with a standard regimen of anti-tuberculous medication.

The standard-of-care for the assessment of treatment success after ATT for pulmonary tuberculosis is sputum microscopy and culture. FDG PET/CT studies have reported residual or persisting metabolic activity in the lungs of patients declared cured based on the current standard of care (10, 11). We and others have previously shown that the presence of RMA in the lungs of patients declared cured according to the current standard of care is associated with early relapse of tuberculous disease (10, 11). These findings from previous studies suggest that FDG PET/CT may be a better modality for treatment response assessment than the current standard of care. Cost and availability are important issues that may limit the widespread application of FDG PET/CT for response assessment after ATT in routine care of patients with tuberculous disease. CT is a more readily available and cheaper modality in clinical practice. Therefore, in this current study, we sought to evaluate the strength of the associations of the different lung changes seen in patients who completed ATT and declared cured according to the current standard of care and the presence of RMA on FDG PET. We classified these CT changes into those known to be associated with active tuberculous disease and those associated with inactive lung changes induced by previous tuberculous disease. Out of the five CT features we classified as associated with active tuberculous disease, four showed a significant association with the presence of RMA on EOT FDG PET. The presence of pleural effusion was the only CT feature that did not show a significant association with the presence of RMA. This negative finding was understanding considering that pleural effusion was seen in only two patients in our study population. We also demonstrated a significant association between the combination of these five CT features of active tuberculous disease and the presence of RMA. The CT features we used as indicators of active tuberculous disease in our study have been previously reported as findings seen in patients starting ATT after the diagnosis of pulmonary tuberculosis (14–16). They have not been robustly assessed in patients who have completed ATT. Therefore, our findings are novel and may suggest that these known CT features of active tuberculous disease may be applied for treatment response assessment in lieu of FDG PET/CT in circumstances where the latter modality is not available or inaccessible.

The five CT features we classified as suggestive of inactive lung changes induced by successfully treated pulmonary tuberculosis disease, when applied in combination, showed no association with the presence of RMA on FDG PET, suggesting that the combination of these findings may help identify patients who have achieved a sterilizing cure. The individual CT features of inactive disease were not great in the assessment of their association with RMA. Two of the five CT features classified as suggestive of inactive disease showed significant associations with the presence of RMA. This finding is best understood on the background of the known heterogeneity of TB lesions where some lesions improved over time on treatment while some other lesions worsened or new lesions appearing in a patient on treatment (10, 18). For this reason, it is not uncommon to encounter, in the same patients, foci of inactive lung changes and other foci of active tuberculous disease. This heterogeneity in response to ATT that may be encountered in tuberculous disease is another strength of imaging for assessing the total-body burden of the disease and the response of each focus of disease to treatment. To address the heterogeneity of lung changes post treatment of pulmonary tuberculosis, a composite score derived from a combination of CT features may be better than considering each individual CT feature on their own. We showed that combined CT features of inactive disease showed a stronger strength of association with absence RMA on FDG PET than the individual CT feature of inactive disease on their own. Yen et al. in a cohort of 50 patients, derived a composite score for assessing lung changes seen in CT in the post tuberculosis setting (19). We did not apply the classification system utilized by Yen et al. because their study population is different from us. While we recruited patients who completed standard course of ATT and required no further treatment according to the current standard of care, Yen et al. studied patients with tuberculosis-induced destroyed lung syndrome for which the patients underwent surgical lung resection.

The greatest strength of FDG PET/CT in the management of patients with tuberculous disease is in its excellent negative predictive value, where a complete metabolic response to ATT is strongly associated with a sterilizing cure (11). On the contrary, not all patients with RMA eventually experience disease relapse on follow-up (11). RMA is primarily due to the persistence of dormant or slowly replicating but non-culturable bacilli that are difficult to eradicate with chemotherapy (6, 10, 20). The presence of these bacilli within tuberculous lesions stimulates host immune response giving rise to the FDG uptake in the lesions. Disease relapse due to the persisting bacilli occurs from the imbalance in the interplay between the host immune system and the bacterial virulence factors. A balance between these two factors (host and bacterial) ensures that the bacilli remain dormant without causing clinically overt disease. Another explanation for RMA that may be seen after the completion of ATT is the host inflammatory response induced by the lipid components of the dead bacterial cell membrane (21). FDG PET/CT was acquired within 2 weeks of completing ATT in our study. For this reason, a possibility exists that some of the patients with RMA in our study may have inflammation induced by dead bacilli rather than dormant or slowly replicating live bacilli. No FDG metric exists currently to differentiate between metabolic activity due to dead bacilli vs. activity due to live bacilli capable of causing disease relapse in the future. The discrimination between these two is of potential clinical importance and should be explored in future studies.

People living with HIV infection are at a higher risk of tuberculous disease, 10% risk per year in HIV-infected people compared with 10–15% lifetime risk in HIV-uninfected people (22, 23). This explains the prevalence of HIV infection seen in our study cohort, with two-thirds of them living with HIV infection at the time of this study. HIV is known to predispose to other inflammatory and infectious conditions, all of which may be confounders in the assessment of residual tuberculous lesion on FDG PET/CT after ATT (24). However, the presence of HIV was not significantly associated with the presence of RMA in our study cohort. This is not surprising considering that all HIV-infected patients included in our study were on stable antiretroviral treatment. This is in line with our experience where we have found the presence of HIV infection to not impact negatively on the accuracy of the quantification of the whole-body burden of disease in oncology patients (25–27). This observation is valid for HIV-infected patients on stable antiretroviral therapy. Interestingly, among HIV-infected patients, a CD4 T cell count of >200 cells/μL and a suppressed HIV viremia were significantly associated with the presence of RMA. These findings suggest that immune recovery evident by a high CD4 T cell count level is needed to mount a robust host immune response responsible for the RMA seen on EOT FDG PET. By extension, we speculate that advanced HIV infection may reduce RMA incidence regardless of the presence of persisting live bacilli.

The strengths of our study include its prospective design and the inclusion of a relatively large study population. At the same time, the study has got some important limitations. We did not acquire an FDG PET/CT scan at the commencement of ATT. Some of the lesions we classified as due to tuberculosis, especially those classified as suggestive of inactive post-tuberculous changes, could have predated the index tuberculous disease. We categorized cavitary lung lesions as suggestive of inactive tuberculous disease. In fact, lung cavity develops during the clinical stage of the disease. This may explain why the presence of cavitary lung lesion showed a significant association with the presence of RMA on EOT FDG PET in our analysis. Cavitary lesions are non-resolving so that they remain after a successful treatment regardless of when they develop. On this premise, we decided to classify cavitary lung lesions as representative of inactive lung changes induced by tuberculous disease rather than suggestive of active tuberculous disease. The classification of CT features as representing active disease was based on findings in the published literature (14–16). Unlike RMA on EOT FDG PET that has been robustly evaluated for its association with residual active tuberculosis, no such comprehensive assessment of the CT features characterized as representing active tuberculosis has been done in longitudinal studies. Our study did not assess the long-term outcome of patients with the CT features of active disease, representing a potential limitation that must be borne in mind when applying our study findings in clinical practice. Our study included mostly patients who completed a standard course of ATT for drug-sensitive TB. Drug-resistant tuberculosis presents a problem of larger magnitude in routine clinical practice as it is more difficult to treat, and the length of treatment is highly variable. It is unknown if the results from this study will be applicable in patients with drug-resistant tuberculosis.

We showed that CT features such as lung nodules, lung consolidation, micronodules in the tree-in-bud pattern, and thoracic lymph nodes, either alone or in combination, have a significant association with residual metabolic activity on end-of-treatment FDG PET/CT in patients who completed a standard course of anti-tuberculous treatment. This suggests that FDG PET/CT may be substituted with CT in assessing the treatment outcome of patients with pulmonary tuberculosis due to its limited availability and high cost. HIV infection does not confound the interpretation of FDG PET/CT obtained from treatment response assessment of pulmonary tuberculosis.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by Faculty of Health Sciences Research Ethics Committee, University of Pretoria. The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

CH, JE, MH, BE, and MS designed the study. IL, KM, IM, NM, and GP collected and analyzed the data. MM, MN, BF, and MS supervised the project and supervision. IL write the original draft of the manuscript. All authors contributed to the article and approved the submitted version.

The study received funding from CRDF Global for the project titled: The Clinical Research Unit (CRU) for the Advancement of Tuberculosis (TB) Biomarker-Targeted Interventions.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.