The need for improved brain lesion segmentation techniques for children with cerebral palsy: A review
Alex M Pagnozzi1, Yaniv Gal2, Roslyn N Boyd3
1CSIRO Digital Productivity and Services Flagship, The Australian e-Health Research Centre, Brisbane, Australia; The University of Queensland, School of Medicine, St. Lucia, Brisbane, Australia.
Insights
Automated brain lesion segmentation in MRI can improve diagnosis and treatment for children with cerebral palsy (CP). This review explores techniques to quantify brain injury, enabling personalized therapies for better lifelong function.
Area of Science:
- Neuroimaging
- Medical Image Analysis
- Pediatric Neurology
Background:
- Cerebral palsy (CP) involves permanent motor disorders from developing brain disturbances.
- Accurate diagnosis and prognosis in CP are challenging due to heterogeneous brain injury and anatomical distortions.
- Current MRI analysis for CP is qualitative, under-utilizing data and hindering personalized treatment optimization.
Purpose of the Study:
- To review available brain injury segmentation approaches applicable to CP MRIs.
- To identify challenges and suggest future research directions for automated segmentation in CP.
- To highlight the potential of automated segmentation for improving CP patient outcomes.
Main Methods:
- Review of brain injury segmentation techniques, including detection of cortical malformations, white/grey matter lesions, and ventricular enlargement.
- Discussion of the strengths and weaknesses of existing segmentation algorithms for CP.
- Identification of adaptive, spatially consistent algorithms and cortical shape parameters for future research.
Main Results:
- Existing segmentation algorithms require modification for reliability in CP due to unique technical challenges.
- Atlas-based priors are ineffective in regions with substantial malformations.
- Adaptive algorithms with fast initialization and cortical shape analysis show promise.
Conclusions:
- Automated brain lesion segmentation is crucial for valid and reproducible quantification of injury in CP.
- Advanced segmentation techniques can elucidate the relationship between imaging features and patient outcomes.
- This approach has the potential to enable better tailoring of therapies for individual CP patients.
Abstract:
Cerebral palsy (CP) describes a group of permanent disorders of posture and movement caused by disturbances in the developing brain. Accurate diagnosis and prognosis, in terms of motor type and severity, is difficult to obtain due to the heterogeneous appearance of brain injury and large anatomical distortions commonly observed in children with CP. There is a need to optimise treatment strategies for individual patients in order to lead to lifelong improvements in function and capabilities. Magnetic resonance imaging (MRI) is critical to non-invasively visualizing brain lesions, and is currently used to assist the diagnosis and qualitative classification in CP patients. Although such qualitative approaches under-utilise available data, the quantification of MRIs is not automated and therefore not widely performed in clinical assessment. Automated brain lesion segmentation techniques are necessary to provide valid and reproducible quantifications of injury. Such techniques have been used to study other neurological disorders, however the technical challenges unique to CP mean that existing algorithms require modification to be sufficiently reliable, and therefore have not been widely applied to MRIs of children with CP. In this paper, we present a review of a subset of available brain injury segmentation approaches that could be applied to CP, including the detection of cortical malformations, white and grey matter lesions and ventricular enlargement. Following a discussion of strengths and weaknesses, we suggest areas of future research in applying segmentation techniques to the MRI of children with CP. Specifically, we identify atlas-based priors to be ineffective in regions of substantial malformations, instead propose relying on adaptive, spatially consistent algorithms, with fast initialisation mechanisms to provide additional robustness to injury. We also identify several cortical shape parameters that could be used to identify cortical injury, and shape modelling approaches to identify anatomical injury. The benefits of automatic segmentation in CP is important as it has the potential to elucidate the underlying relationship between image derived features and patient outcome, enabling better tailoring of therapy to individual patients.
