The use of 3D image analysis in the diagnosis and treatment planning of corrective surgery for mandibular asymmetry patients

Student thesis: Doctoral ThesisDoctor of Philosophy


Background: mandibular asymmetry poses a challenge in craniofacial diagnosis and treatment planning. Variability in etiology and presentations necessiates accurate assymetry diagnosis and quantification for treatment planning and follow-up. Conventional cephalometric radiographs have inherent superimpositions, magnifications and distortions, limiting their clinical value.
Aims: To develop and trial novel diagnostic Cone Beam CT image analysis procedures, to enhance treatment by quanitification and localization of mandibular assymetries. Development and trials employed simulation and patient datasets, including end-user clinicians' perception of simulation versus conventional surgical outcomes.
Methods: Twenty assymetric patients' CBCT volumes were segmented, constructing surface models, allowing 3D image analyses, using voxe3l-based co-registration. Mirroring techniques computed left to right side dissimilarities and validation studies initially compared known introduced assymetries. The SPHARM-PDM toolbox was then applied to a cohort of 60 patiens with varying degrees of assymetry to assess detection limits and functionality, including a Craniofacial Microsomia dataset. Finally, an algorith was developed to compute virtual mandibular surgical outcomes and a multi-centre study compared both real and virtual outcomes for 20 patients.
Results: SPHARM-PDM proved a valid technology to quantify degree and direction of mandibular assymetries with probabilities of (0.99-1) and (0.84-1) respectively, that true assymetry was within 0.5 mm (translation) or 5 (rotations). Whether right or left mandible was mirrored, no statistical output differences were identified, demonstrating consistency. The functional limit is identified as no anatomical correspondence e.g. where a complete structural loss or external augmentation has arisen.
Encouragingly, simulated outcomes based on virtual templates were better than actual outcomes, conventionally planned, in particular in difficult cases.
Conclusion: SHARM-PDM is useful for diagnosis and quantification of mandibular assymetries and virtual templates in treatment planning offer potentials to increase predictability and optimize craniofacial interventions.
Date of Award2014
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorRichard Cook (Supervisor) & Frederic Festy (Supervisor)

Cite this