Detection of abdominopelvic lymph nodes in multi-parametric MRI

Tejas Sudharshan Mathai, Thomas C. Shen, Daniel C. Elton, Sungwon Lee, Zhiyong Lu, Ronald M. Summers

Abstract: Highlights•We propose a pipeline for lymph node detection in multi-parametric MRI studies.•T2 fat suppressed (T2FS) and DWI series were linearly interpolated to preserve their traits during training.•We used a Dynamic Head (DyHead) neural network integrated with the complete IoU loss.•At test time, either T2FS can be used alone or blended with any available DWI series (after co-registration).•A 53.5% precision and ∼78% sensitivity at 4 FP/vol was seen with ≥10% improvement (p < 0.05) over prior works.

External IDs:doi:10.1016/j.compmedimag.2024.102363