Outcomes of patients undergoing lung resection for drug-resistant TB and the prognostic significance of pre-operative positron emission tomography/computed tomography (PET/CT) in predicting treatment failure

BACKGROUND: Surgery remains an adjunctive treatment for drug-resistant tuberculosis (DR-TB) treatment failure despite the use of bedaquiline. However, there are few data about the role of surgery when combined with newer drugs. There are no outcome data from TB endemic countries, and the prognostic significance of pre-operative PET-CT remains unknown. METHODS: We performed a prospective observational study of 57 DR-TB patients referred for surgery at Groote Schuur Hospital between 2010 and 2016. PET-CT was performed if there was nodal disease or disease outside the area of planned resection but did not influence treatment decisions. 24-month treatment success post-surgery (cure or treatment completion), including all-cause mortality, was determined. FINDINGS: 35/57 (61.4%) patients (median age 40 years; 26% HIV-infected) underwent surgery and 22/57 (38.6%) did not (11 patients were deemed unsuitable due to bilateral cavitary disease and 11 patients declined surgery). Treatment failure was significantly lower in those who underwent surgery compared to those eligible but declined surgery [15/35 (43%) versus 11/11 (100%); relative risk 0.57 (0.42–0.76); p < 0.01). In patients treated with surgery, a post-operative regimen containing bedaquiline was associated with a lower odds of treatment failure [OR (95%CI) 0.06 (0.00–0.48); p = 0.007]. Pre-operative PET-CT (n = 25) did not predict treatment outcome. INTERPRETATION: Resectional surgery for DR-TB combined with chemotherapy was associated with significantly better outcomes than chemotherapy alone. A post-operative bedaquiline-containing regimen was associated with improved outcome; however, this finding may have been confounded by higher use of bedaquiline and less loss to follow-up in the surgical group. However, PET-CT had no prognostic value. These data inform clinical practice in TB-endemic settings. FUNDING: This work was supported by the South African 10.13039/501100001322MRC (RFA-EMU-02-2017) and the 10.13039/501100001713EDCTP (TMA-2015SF-1043 & TMA- 1051-TESAII).