Silent tuberculosis hides in people who feel fine. In South Africa, doctors used digital chest scans and a computer tool to look for this hidden disease in over 1,300 adults. They wanted to know if these tools could catch the infection before symptoms appeared. The results were mixed and did not meet the strict standards set by the World Health Organization.
Human readers looking at the scans found the disease in only about half of the silent cases. The computer tool did slightly better but still missed a large number of people who had the infection. When people had symptoms, the tools worked better but still left many cases undetected. Specificity was high, meaning they rarely said someone had the disease when they did not.
The study team noted that neither the human readers nor the computer tool reached the target of catching 90 percent of cases while correctly identifying 80 percent of healthy people. This gap means that relying on these current methods alone might leave too many people untreated. The findings suggest that screening strategies need to change to find more silent cases without causing too many false alarms.