Ruperez Maria, Busang Jacob, Mureithi Linda, Shanaube Kwame, Klinkenberg Eveline, Gachie Thomas, Burnett James M, Kosloff Barry, de Haas Petra, Hayes Richard, Fidler Sarah, Schaap Ab, Floyd Sian, Ayles Helen
Clinical Research Department, London School of Hygiene & Tropical Medicine (LSHTM), London, United Kingdom.
Health Systems Trust, Cape Town, South Africa.
PLOS Glob Public Health. 2025 Jun 20;5(6):e0003787. doi: 10.1371/journal.pgph.0003787. eCollection 2025.
WHO recommends computer-aided detection (CAD) in chest X-ray (CXR) for systematic screening of TB. Increased detection of individuals with high CAD score but without bacteriologically confirmed TB can be expected, requiring guidance on their clinical management. We followed participants of a TB prevalence survey (TBPS) in Zambia and South Africa with a high CAD score but no bacteriologically confirmed TB over a median time of 9 months and assessed their clinical outcomes. At the TBPS participants with TB-suggestive symptoms or a CAD score ≥40 submitted two sputum samples for Xpert-Ultra testing, and, an additional sample was collected the next day for liquid culture and Xpert-ultra testing. Participants with a CAD score ≥70 and no bacteriologically confirmed TB were eligible for follow-up. At follow-up visit participants were asked about TB symptoms and treatment, underwent a repeat CXR with CAD, and those with either TB-suggestive symptoms or a CAD score ≥70 at follow-up submitted a sputum sample for Xpert-Ultra testing. A composite "clinical" outcome was defined based on changes in CAD-score and TB-suggestive symptoms between the TBPS and the follow-up. Of the 254 eligible TBPS participants 162 (65%) completed follow-up. Most of the participants self-reported previous TB (65% 105/162), were from Zambia (79%, 128/162,) and male (70%, 97/162). Overall, 43% (70/162) participants progressed clinically/remained radiologically abnormal and 6% (10/162) developed TB between the TBPS and the follow-up, with an overall TB incidence rate of 7% per year (95% CI: 3.8-13.3). Patients with high CAD score but no bacteriological confirmation may have had a past TB or other pulmonary lesions identified in the CXR, which may need to be investigated. Also, these participants may be at risk of progressing to TB over time and could benefit from a follow-up visit and from repeated assessment of symptoms and CXR.
世界卫生组织建议在胸部X线检查（CXR）中采用计算机辅助检测（CAD）对结核病进行系统筛查。预计CAD高分但未得到细菌学确诊的结核病患者的检出率将会增加，这就需要对其临床管理提供指导。我们对赞比亚和南非结核病患病率调查（TBPS）中CAD高分但未得到细菌学确诊的结核病患者进行了随访，中位随访时间为9个月，并评估了他们的临床结局。在TBPS中，有结核病疑似症状或CAD评分≥40的参与者提交两份痰标本进行Xpert-Ultra检测，第二天再采集一份标本进行液体培养和Xpert-Ultra检测。CAD评分≥70且未得到细菌学确诊的结核病患者符合随访条件。在随访时，询问参与者结核病症状和治疗情况，进行重复的带有CAD的CXR检查，随访时出现结核病疑似症状或CAD评分≥70的参与者提交一份痰标本进行Xpert-Ultra检测。根据TBPS和随访之间CAD评分和结核病疑似症状的变化定义了一个综合“临床”结局。在254名符合条件的TBPS参与者中，162名（65%）完成了随访。大多数参与者自我报告既往有结核病（65%，105/162），来自赞比亚（79%，128/162），男性（70%，97/162）。总体而言，43%（70/162）的参与者临床病情进展/影像学仍异常，6%（10/162）在TBPS和随访之间患结核病，总体结核病发病率为每年7%（95%CI：3.8-13.3）。CAD高分但未得到细菌学确诊的患者可能既往有结核病或CXR检查发现有其他肺部病变，这可能需要进行调查。此外，这些参与者可能随着时间推移有进展为结核病的风险，可能会从随访以及症状和CXR的重复评估中获益。
