Authors (including presenting author) :
Chan NW(1), Xiao L(1), Chan TY(2), Chan YL(2), Liu T(2), Tse HF(2), Lee KW(2), Sin KW(2), Lai CC(1), Leung J(1), Ngar DYK(1)
Affiliation :
(1)Medical Physics Unit, Department of Clinical Oncology, Tuen Mun Hospital, (2)Department of Radiology and Nuclear Medicine, Tuen Mun Hospital/Pok Oi Hospital/Tin Shui Wai Hospital
Introduction :
Imaging quality is critical to ensure accurate imaging interpretation and therefore optimal patient care. Measuring and monitoring the imaging performance and operation stabilities of equipment in radiology are essential for obtaining good quality imaging and accurate diagnosis. Routine quality control of the MR and CT imaging systems such as periodic image quality test and constancy test, detects ongoing errors in radiologic practice which helps to obtain information of system variability before they can adversely affect clinical images. However, it is a laborious process to analyze the voluminous data in the QC report for imaging system.
Objectives :
Our newly developed Python-coded program shows its superiority in data visualizing over the performance of conventional QC alone. Its algorithm design facilitates in processing massive datasets and integrating parameters from the disorganized QC reports. This intelligent program was applied to acquire a general assessment of the image quality parameters to ensure high quality images delivered. It also helps early detection of failures and image artifacts in the imaging systems. The quality of diagnostic images and the performance improvements in safety and efficiency are further enhanced when combining this Python program with routine QC satisfying accreditation and regulatory requirement.
Methodology :
This study was begun in December 2021 and captured the QC reports of two MR and two CT imaging systems over the past seven years in our cluster hospitals. PyCharm was used as a Python code editor in the study. Different open-source Python libraries such as Pandas, NumPy and Glob, were adopted in the Python script to perform task automation, data extraction and data visualization.
Result & Outcome :
The Python program illustrated the trends of variability for various useful QC data including center frequency, signal-to-noise ratio, artifact evaluation, image uniformity, geometric distortion, spatial resolution, homogeneity and noise. It provides an efficient and visualizable evaluation of routine QC to radiologic technologists which enables effective monitoring of the imaging performance and operation stabilities of the MR and CT imaging systems. The generalization of the developed program facilitates data interpretation and easy data retrieval for different clinical specialties in the hospital including radiologists, radiographers, engineering specialists, electrical technicians and medical physicists. This could serve as a common platform or even web-based platform which provides a mean for data discussion and future possibility of data comparison between different systems, imaging modalities or hospitals.
