Back in 2010, Delft Imaging already provided an Odelca DR X-ray system to support screening in Tanzania.
In 2020, Delft Imaging partnered different organizations within Tanzania to support research on the utilization of CAD4TB, the artificial intelligence software that supports the automated detection of TB-related abnormalities on chest X-rays.
In 2021, the utilization of CAD4TB was further expanded with 4 CAD4TB solutions to support the National Tuberculosis Program of Tanzania in the screening and triage of TB in the country.
Shortly after, in 2022, Delft Imaging delivered another 17 CAD4TB solutions to Tanzania in support of the National TB Program. The CAD4TB solutions were installed in different clinics and hospitals across the country. The project was delivered with the necessary installation and training services, in collaboration with the local in-country partner of Delft Imaging.
Early 2023, Delft Imaging delivered a Delft Ultra, an ultra-portable X-ray system to Tanzania, together with the CAD4TB software and relevant training and installation services.
Making a difference
Already in 2014, a study was published on the utilization of CAD4TB in Tanzania. CAD4Tb was evaluated on chest radiographs of patients with symptoms suggestive of pulmonary tuberculosis enrolled in two cohort studies in Tanzania. CAD4TB was found to be significantly more accurate for the discrimination of smear-positive cases against non TB patients than for smear-negative cases. It also showed that CAD4TB significantly outperformed the clinical officer, but did not reach the accuracy of the expert reader. Note that the study used a significantly older version of CAD4TB than what is currently available.
In 2015, another study was published using CAD4TB looking at CAD4TB as a tool to help screen for pulmonary tuberculosis in a Tanzanian prison. The study showed that CAD4TB reliably evaluates chest X-rays from a mostly asymptomatic prison population, with a diagnostic performance inferior to that of expert readers but comparable to local readers. Note that the study used a significantly older version of CAD4TB than what is currently available.