Submission ID 118525
| Issue/Objective | Nigeria has the highest Tuberculosis (TB) burden in Africa with an estimated incidence of 219/100,000 and accounts for 6.3% of the gap of unidentified TB cases globally. The approach to finding people living with TB has been through symptom screening for presumptive TB using the World Health Organization Four Symptom Screening (W4SS) and linking them to diagnosis. However, some clients do not report symptoms accurately. To address this, KNCV Nigeria TB LON project introduced chest Xray TB screening in communities using portable digital Xray (PDX) fitted with Artificial intelligence (AI). |
|---|---|
| Methodology/Approach | Nine PDX platforms were deployed across 8 project implementing states to screen for TB in identified TB hotspot and hard to reach Communities. These PDX platforms were fitted with AI- Computer Aided Detection for TB (CAD4TB). The CAD4TB generates scores from 1-100% based on abnormality patterns on the lungs. Presumptive TB were identified as those with CAD4TB score of ≥50. Following a developed algorithm all identified presumptive TB were tested with GeneXpert or TB LAMP. Those not able to produce sputum had their Xray films reviewed by a radiologist using a web based real time reporting platform known as XMAP. The TB yield from Chest Xray screening was compared to symptom screening. |
| Results | A total of 516,048 clients were screened from January - December 2024 across hotspot communities. A total of 55,583 presumptive TB were identified, 4062 cases diagnosed (TB yield of 7%) an average Number Needed to Test (NNT) of 14 and Number Needed to Screen (NNS) of 127. Disaggregating by screening method, a total of 247,093 clients were screened with W4SS, of those 26,414 presumptive TB were identified 1,327 TB cases diagnosed (5% TB yield) NNT was 20 and NNS as 186. While 268,955 clients were screened with Chest X ray with CAD4TB, 29,169 identified presumptive TB were evaluated, 2,735 TB were diagnosed (TB yield of 9%) NNT was 11 and NNS was 95 |
| Discussion/Conclusion | There was a higher TB yield and better TB cascade efficiency with chest Xray with CAD4TB compared to W4SS only. We recommend the use of AI enabled digital Xray to improve TB yield in communities. |
| Presenters and affiliations | Ogoamaka Chukwuogo KNCV Nigeria Bethrand Odume KNCV Nigeria Chidubem Ogbudebe KNCV Nigeria Nkiru Nwokoye KNCV Nigeria Sani Useni KNCV Nigeria Emperor Ubochioma NTBLCP |
