Automatic classification of prostate MR series type using image content and metadata
Deepa Krishnaswamy, Bálint Kovács, Stefan Denner, Steve Pieper, David Clunie, Christopher P. Bridge, Tina Kapur, Klaus H. Maier-Hein, Andrey Fedorov
TL;DR
The paper addresses the challenge of automatically classifying prostate MRI sequence types when free-text metadata is unreliable. It proposes a CNN that jointly processes image content and scanner-derived DICOM metadata to categorize four sequence types (T2W, DWI, ADC, DCE) using publicly available IDC datasets, with 4-fold cross-validation and ensembling. The key finding is that combining image data with metadata yields higher accuracy than using either source alone, although generalization to external collections is limited by cross-collection differences and DWI/ADC confusion driven by is4D and intensity similarity. This work highlights the importance of metadata in robust MRI data curation and provides publicly available code to facilitate reproducibility and broader adoption in AI-powered radiology workflows.
Abstract
With the wealth of medical image data, efficient curation is essential. Assigning the sequence type to magnetic resonance images is necessary for scientific studies and artificial intelligence-based analysis. However, incomplete or missing metadata prevents effective automation. We therefore propose a deep-learning method for classification of prostate cancer scanning sequences based on a combination of image data and DICOM metadata. We demonstrate superior results compared to metadata or image data alone, and make our code publicly available at https://github.com/deepakri201/DICOMScanClassification.