DTI_PANORAMA

About

Creators:

NataliaAlves

MountainHill

Contact email:

han.liu@siemens-healthineers.com

Image Version:

6c176cf6-72a0-427f-a003-7c7b357568d1

Last updated:

Dec. 16, 2024, 4:09 p.m.

Associated publication:

Liu H, Gao R, Grbic S. AI-assisted Early Detection of Pancreatic Ductal Adenocarcinoma on Contrast-enhanced CT. arXiv. Published online March 18, 2025.

Interfaces

This algorithm implements all of the following input-output combinations:

Inputs Outputs

	Inputs	Outputs
1	Venous phase CT scan Description A contrast-enhanced CT scan in the venous phase and axial reconstruction Kind Image Read from `/input/images/venous-ct/<uuid>.mha` or `/input/images/venous-ct/<uuid>.tif` Venous phase CT scan Clinical Information Pancreatic CT Description Clinical information in JSON format, includes "age", "sex", "study_date" and "scanner". In formats, respectively, "042Y", "M", "YYYY-MM-DD HH:MM:SS", and "TOSHIBA". All values can be null. Kind Anything Read from `/input/clinical-information-pancreatic-ct.json` Example { "key": "value", "None": null } Clinical Information Pancreatic CT	PDAC Likelihood Description This indicates the likelihood of pancreatic ductal adenocarcinoma (PDAC) occurring, presented as a value between 0.0 and 1.0. Kind Float Write to `/output/pdac-likelihood.json` Example 42.0 PDAC Likelihood PDAC Detection Map Description Single-class, detection map of pancreatic ductal adenocarcinoma (PDAC) in 3D, where each voxel represents a floating point in range [0,1] Kind Heat Map Write to `/output/images/pdac-detection-map/<uuid>.mha` or `/output/images/pdac-detection-map/<uuid>.tif` PDAC Detection Map

Challenge Performance

Date	Challenge	Phase	Rank
Oct. 3, 2024	PANORAMA	Development Phase - Sanity Check	1
Jan. 23, 2025	PANORAMA	Development Phase - Tuning	1
Feb. 25, 2025	PANORAMA	Testing Phase	1

Model Facts

Summary

Pancreatic ductal adenocarcinoma (PDAC) is one of the most common and aggressive types of pancreatic cancer. However, due to the lack of early and disease-specific symptoms, most patients with PDAC are diagnosed at an advanced disease stage. Consequently, early PDAC detection is crucial for improving patients’ quality of life and expanding treatment options. In this work, we develop a coarse-to-fine approach to detect PDAC on contrast-enhanced CT scans. First, we localize and crop the region of interest from the low-resolution images, and then segment the PDAC-related structures at a finer scale. Additionally, we develop two strategies to further boost the detection performance, including (1) a data splitting strategy for model ensembling, and (2) a customized post-processing function. We participated in the PANORAMA challenge1 and ranked 1st place for PDAC detection with an AUROC of 0.9263 and an AP of 0.7243. Our code and models are publicly available at https://github.com/han-liu/PDAC_detection.

Mechanism

Please refer to our technical report: AI-assisted Early Detection of Pancreatic Ductal Adenocarcinoma on Contrast-enhanced CT (https://arxiv.org/pdf/2503.10068)

Validation and Performance

We ranked 1st place in the testing phase of the PANORAMA challenge, with an AUROC of 0.9263 and an AP of 0.7243. More details can be found in our technical report.

Uses and Directions

This algorithm was developed for research purposes only.

Warnings

Left empty by the Algorithm Editors

Common Error Messages

If you have any questions, feel free to open an issue at our GitHub page: https://github.com/han-liu/PDAC_Detection

Information on this algorithm has been provided by the Algorithm Editors, following the Model Facts labels guidelines from Sendak, M.P., Gao, M., Brajer, N. et al. Presenting machine learning model information to clinical end users with model facts labels. npj Digit. Med. 3, 41 (2020). 10.1038/s41746-020-0253-3