Automated Quality Control

INCREASINGLY, MEDICAL IMAGING is being used to quantify the evolution and response to therapy of diseases because good imaging performance indicates that image quality should be sufficient to meet the clinical requirement for the examination to have a successful diagnostic process. 

 This quantification benefits from the use of multimodal images and advanced image processing such as linear and nonlinear registration, image segmentation, and classification methodologies. Example applications include quantification of lesions, and non-lesional pathology in multiple sclerosis (MS) and Alzheimer’s disease. 

Often, post-acquisition image-processing tools have been demonstrated to be robust when used on high quality images from a single scanner. However, the introduction of artifacts associated with patient motion and the increased variability associated with multiple scanners can significantly degrade the accuracy and precision of results obtained. 

The method most often used to avoid the dissemination of error along image analysis pipelines is the use of a visual quality control (QC) verification step. This step is performed manually by experts. The downstream image assessment thus becomes susceptible to intra- and inter-evaluator variability, as well as human error related to the failure to identify deviations from approved sequence parameters. In addition, this approach rarely provides quantitative outputs for tracking or comparative analyses. ¹



The most important parts of an AQC for CT include: 

  • Patient Identity Verification
  • Positioning Verification 
  • CT Parameters Verification ( Slice thickness, CTDI) 
  • CT number Verification
  • SNR Verification
  • Uniformity and plane distance accuracy 
  • Contrast (spatial) resolution Verification 

The most important parts of an AQC for MRI include: 

  • l.Patient Identity Verification
  • MRI Parameters Verification
  • SNR Verification
  • Ghosting (SGR) Verification
  • IV Enhancement Verification 
IQBMI developed dedicated software for checking the quality of fMRI data. This feature is able to evaluate quality of images in terms of SNR, SFNR, SGR, Spike, motions, and other artifacts. Flagging the studies with poor image quality automatically is essential for an accurate diagnosis and IQBMI aim is adding this feature for bringing this advantage to you to ensure minimal errors in reporting procedure.