Related Experiment Videos
Assessment of rigid multi-modality image registration consistency using the multiple sub-volume registration (MSR)
C Ceylan1, U A van der Heide, G H Bol
1Department of Radiotherapy, University Medical Center Utrecht, The Netherlands.
Physics in Medicine and Biology
|May 7, 2005
Summary
Assessing medical image registration quality is challenging without a gold standard. The multiple sub-volume registration (MSR) method evaluates rigid registration consistency by analyzing sub-image deformations, showing high agreement with visual inspection.
Area of Science:
- Medical imaging analysis
- Image registration quality assessment
- Radiology and Nuclear Medicine
Background:
- Accurate registration of multi-modal medical images (CT, MRI, PET, SPECT) is crucial for clinical applications.
- Evaluating the quality of automatic image registration is difficult due to the lack of a universally accepted gold standard.
- Existing methods often rely on subjective visual inspection, which can be inconsistent.
Purpose of the Study:
- To introduce and validate a novel method, Multiple Sub-volume Registration (MSR), for assessing the consistency of rigid medical image registration.
- To provide an objective, quantitative measure for evaluating registration quality independent of the specific algorithm or similarity metric used.
- To demonstrate the MSR method's applicability across various imaging modalities and anatomical regions.
Main Methods:
- The Multiple Sub-volume Registration (MSR) method involves registering sub-images from one dataset to another, followed by a coarse non-rigid registration.
- Local deformations between sub-volume registrations and the overall full registration are analyzed to quantify registration consistency.
- The MSR method was tested on 15 multi-modal datasets (CT, MR, PET) from diverse anatomical sites (brain, head/neck, cervix, prostate, lung) using rigid registration with normalized mutual information.
Main Results:
- The MSR method successfully classified registration quality, achieving results that agreed with expert visual inspection across all tested cases.
- Statistical analysis (ANOVA) confirmed a significant correlation between MSR-based classification and visual assessment (p < 0.02).
- The MSR method proved effective for multi-modality image datasets and various anatomical locations.
Conclusions:
- The Multiple Sub-volume Registration (MSR) method offers a reliable and objective approach to assess the consistency of rigid medical image registration.
- This technique overcomes the limitations of subjective visual inspection and the absence of a gold standard.
- The MSR method is a versatile tool applicable to diverse imaging modalities and anatomical regions, enhancing the quality control of medical image registration.