2024 Informatics Tool Challenge Winners Announced
, by Jennifer K. Loukissas, M.P.P.
Three projects were funded through the 2024 DCEG Informatics Tool Challenge, two in support of the Pediatric Proton and Photon Therapy Comparison Cohort. The third will facilitate the development and evaluation of polygenic risk scores across diverse populations. Since its establishment in 2014, the competitive program has provided support for innovative approaches to epidemiological methods, data collection, analysis, and other research efforts using modern technology and informatics.
Innovative and Realistic Patient CT Scan Extension for Dosimetry and Epidemiology Studies
Sergio Morato Rafet (REB), Matthew Mille (REB), Choonsik Lee (REB), Jungwook Shin (REB), Monjoy Saha (ITEB), Keith Griffin (REB), Kai-ling Chen (CBIIT), Madhu Kanigicherla (CBIIT)
Standard computed tomography (CT) scans for radiotherapy planning typically cover only a small portion of the patient's body near the target treatment area and omit organs of interest for research into radiation late effects. This tool will integrate a deep learning image auto-segmentation process with a novel anatomy extension algorithm based on a library of real CT patient images to extend partial body anatomies. It is designed to provide realistic and complete patient anatomies for the Pediatric Proton and Photon Therapy Comparison Cohort, where medical records do not include CT images of the entire patient anatomy. The completeness of this data will allow researchers to estimate organ doses and explore the short- and long-term treatment-related adverse effects of proton and photon therapies.
Online Dashboard for Individualized Dose Assessment from Large-Scale Radiotherapy Treatment Data
Jungwook Shin (REB), Todd Gibson (REB), Cari Kitahara (REB), Matthew Mille (REB), Sergio Morato Rafet (REB), Choonsik Lee (REB), Srujan Boppana (CBIIT)
This web-based tool will integrate heterogeneous data collected from participants in the Pediatric Proton/Photon Therapy Comparison Cohort and stored in two separate databases: one with medical records and a second with radiation treatment information. Investigators will be able to link the databases for individualized organ dose assessment, providing the ability to evaluate short- and long-term treatment related adverse effects. Specifically, the tool will (1) monitor time-series data collection, (2) run combinational data searches across the two databases, (3) to perform mathematical operations for the combined data, and (4) display and download the search and/or operation results.
PRS-Flow: An Integrated Toolkit and Pipeline for Polygenic Risk Scores Development
Haoyu Zhang (BB), Mitchell Machiela (ITEB), Xiaoyu (Kevin) Wang (CGR), Bill Wheeler (IMS), Kai-ling Chen (CBIIT), Madhu Kanigicherla (CBIIT), Phyllip Cho (CBIIT)
This integrated toolkit and pipeline is structured and designed to enhance the development and evaluation of multi-ancestry polygenic risk scores (PRS) across global populations. The team will develop a command-line based software supplemented by a detailed website for data preprocessing. This platform will provide users with calculated linkage disequilibrium (LD) references from diverse populations and a consistent syntax for implementing multiple cutting-edge multi-ancestry PRS methods. Additionally, they will create an interactive web tool that enables users to retrieve missing SNPs through LD proxies, ensuring robust, reproducible results essential for advancing personalized medicine across diverse populations. PRS-Flow will not only address current limitations within the field but also enhance the accessibility and accuracy of PRS for researchers worldwide.