Cordoba team demonstrates workload reduction in 2D and 3D, published in Radiology

With the European launch of Transpara powered by Fusion AI, the platform has attracted a lot of interest from around the world. To date it is doing well in the USA since FDA accreditation last year and is now installed in over 20 countries including Sweden, Germany, France, Spain and UK.

Transpara has been subject to rigorous clinical testing at the Hospital Universitario Reina Sofia in Cordoba, Spain, under the leadership of Dr Alvarez Benito. The results of the work entitled, “Using autonomous AI to reduce the workload of breast cancer screening with breast tomosynthesis: a retrospective validation,” were initially presented at the RSNA 2020 and are now published in the high impact Radiology journal https://pubs.rsna.org/doi/10.1148/radiol.2021203555.

The study confirmed that Transpara can confidently identify very likely normal 2D and 3D mammograms in screening, which could be used to safely reduce workload without affecting screening performance.

Dr Alvarez Benito comments, “With the retrospective analysis of exams from a consecutively collected screening cohort of 15,986 women, we clearly showed that by using Transpara we can significantly reduce workload by up to 70%, without compromising the sensitivity in either 2D or 3D screening.

This saving is so important as there is a global shortage of radiologists, with workload also on the increase due to the introduction of new imaging techniques such as 3D breast tomosynthesis.” According to Dr Alvarez Benito, “In our first study, we used Transpara version 1.6.0. The results were so compelling that we felt the need to bring this deep learning technology to women in our breast screening setting.

We are now prospectively implementing Transpara in our screening program and we are very excited to see the results from our prospective trial. Even better, we have observed that the latest version of Transpara has significantly improved the detection performance of the system, and a higher sensitivity is achieved at high specificity.”

The latest version of the product, now powered by FusionAI, has reached a new high-performance level: FusionAI accuracy has improved by up to 28%, bringing a further leap in performance which sets a new benchmark in evidence-based AI decision support.