Use Case:
QIPCM-MIRA as a radiomics support tool
QIPCM-MIRA as a radiomics support tool
There are clear requirements for standardization, quality assurance and control at each step in the radiomics process.
QIPCM, in collaboration with UHN’s Radiation Medicine Program, has validated tools and services for:
In order to facilitate the process of data curation, meta-data filtering and feature extraction, the MIRA platform, built by Dr. Tony Tadic, Dr. Nathan Becker and Tirth Patel at the Princess Margaret is used. MIRA is a web application and software framework for automating data science and integrating image processing pipelines (e.g. pyradiomics, deep learning modules) with clinical systems such as Raystation, Mosaiq and various DICOM PACS and databases.
Dr. Lillian Siu, Dr. Anna Spreafico, Dr. Scott Bratman and Dr. Benjamin Haibe-Kains at Princess Margaret Hospital have been working with QIPCM for both the INSPIRE (https://clinicaltrials.gov/ct2/show/NCT02644369) and ROMA (https://clinicaltrials.gov/ct2/show/NCT03759730) trials. INSPIRE is a single centre, nonrandomized, multi- cohort trial of pembrolizumab in subjects with advanced solid tumors. The goal of ROMA is to evaluate the role of microbiome as a biomarker in locally advanced oropharyngeal squamous cell carcinoma.
Requirement | QIPCM Solution |
---|---|
Image Anonymization and Transfer | Custom pipeline for de-identification and secure transfer from each institution to the QIPCM repository |
Image QA | Image QA by QIPCM including QA of the meta data to provide summary of adherence to acquisition protocols and anomalous cases. |
Centralized Storage with Image Access Logs | Images that pass QA are forwarded to centralized PACS |
Remote Connection | QIPCM Virtual Desktop Infrastructure (VMs) |
GPU for deep learning | QIPCM access to HPC4health compute cluster. |
Custom Feature Extraction | Feature extraction using the Pyradiomics software package. List of features to be extracted is customizable. |