Median Technologies has reported positive performance for its iBiopsy artificial intelligence (AI)-powered Computer-Aided Diagnosis (CADx) product for lung cancer screening (LCS) to characterise malignant and benign lung nodules.
The assessment follows a large-scale study that included a cohort of 1,696 patient cases from the National Lung Screening Trial (NLST) with a total of 15,608 lung nodules.
The training set was based on 1,224 patients, while the test set was based on 472 patients.
iBiopsy CADx demonstrated 95.2% sensitivity and 95.7% specificity for lung nodule characterisation, displaying an AUC of 0.991.
Median founder and CEO Fredrik Brag said: “We have achieved an outstanding performance with our innovative deep learning algorithm for lung nodule characterisation.
“This promising and powerful digital biomarker could have a profound impact on lung cancer screening and help to dramatically reduce the 1.8m lung cancer deaths per year by diagnosing lung cancer patients at a stage that could save their lives.
“Our plan in the coming months is to perform an additional large-scale study for an end-to-end lung cancer screening digital biomarker. Results of this study including both nodule detection (CADe) and characterisation in Q4, 2021.”
Several studies have shown that low-dose computed tomography (LDCT) LCS can reduce mortality rates among patients significantly.
However, screening efforts suffer from low rates of adoption primarily due to the challenges in characterising malignant from benign nodules.
Median Technologies expects that iBiopsy CADx can help in eliminating barriers and support widespread LCS adoption.
Earlier this year, Median signed a research collaboration agreement with the University of California San Diego, in the US, regarding iBiopsy.