CT Data Curation for Liver Patients: Phase Recognition in Dynamic Contrast-Enhanced CT

TL;DR

This paper presents an automated data curation tool using a deep learning model to accurately identify phases in liver CT scans from real-world hospital datasets, facilitating large-scale medical imaging data collection.

Contribution

The work introduces a novel 3D SE architecture with an aggregated loss function for effective phase recognition in heterogeneous CT datasets, improving data harvesting accuracy.

Findings

Achieved a mean F1 score of 0.977 in phase classification.

Successfully harvested up to 92.7% of studies from noisy labels.

Outperformed existing models and standard approaches in accuracy.

Abstract

As the demand for more descriptive machine learning models grows within medical imaging, bottlenecks due to data paucity will exacerbate. Thus, collecting enough large-scale data will require automated tools to harvest data/label pairs from messy and real-world datasets, such as hospital PACS. This is the focus of our work, where we present a principled data curation tool to extract multi-phase CT liver studies and identify each scan's phase from a real-world and heterogenous hospital PACS dataset. Emulating a typical deployment scenario, we first obtain a set of noisy labels from our institutional partners that are text mined using simple rules from DICOM tags. We train a deep learning system, using a customized and streamlined 3D SE architecture, to identify non-contrast, arterial, venous, and delay phase dynamic CT liver scans, filtering out anything else, including other types of liver contrast studies. To exploit as much training data as possible, we also introduce an aggregated cross entropy loss that can learn from scans only identified as "contrast". Extensive experiments on a dataset of 43K scans of 7680 patient imaging studies demonstrate that our 3DSE architecture, armed with our aggregated loss, can achieve a mean F1 of 0.977 and can correctly harvest up to 92.7% of studies, which significantly outperforms the text-mined and standard-loss approach, and also outperforms other, and more complex, model architectures.