6533b7d9fe1ef96bd126cd48
RESEARCH PRODUCT
Three-Dimensional Reconstruction of the Bony Nasolacrimal Canal by Automated Segmentation of Computed Tomography Images.
Nicolas Toledano-fernandezJose M. Ramirez-sebastianFederico Saenz-francesMaria Urbasos-pascualLucía Jáñez-garcíaLuis Jáñez-escaladasubject
MaleModels AnatomicCritical Care and Emergency Medicinelcsh:MedicineComputed tomographyPolynomialsDiagnostic RadiologyNormality test0302 clinical medicineMedicine and Health SciencesSegmentationDegree of a polynomiallcsh:ScienceTomographyMusculoskeletal SystemTrauma MedicineMathematicsMultidisciplinaryNasolacrimal ductmedicine.diagnostic_testRadiology and ImagingAnatomyMiddle Agedmedicine.anatomical_structureSurgery Computer-AssistedPhysical SciencesNasolacrimal canalFemaleAnatomyResearch ArticleAdultComputer and Information SciencesImaging TechniquesTrauma SurgeryAutomated segmentationNeuroimagingSurgical and Invasive Medical ProceduresResearch and Analysis MethodsBone and BonesComputer Software03 medical and health sciencesImaging Three-DimensionalDiagnostic MedicinemedicineHumansSkeletonAgedMorphometrySkulllcsh:RBiology and Life SciencesComputing MethodsComputed Axial TomographyCross-Sectional StudiesAlgebra030221 ophthalmology & optometrylcsh:QTomography X-Ray ComputedNasolacrimal DuctMathematics030217 neurology & neurosurgeryNeuroscienceBiomedical engineeringVolume (compression)description
Objective To apply a fully automated method to quantify the 3D structure of the bony nasolacrimal canal (NLC) from CT scans whereby the size and main morphometric characteristics of the canal can be determined. Design Cross-sectional study. Subjects 36 eyes of 18 healthy individuals. Methods Using software designed to detect the boundaries of the NLC on CT images, 36 NLC reconstructions were prepared. These reconstructions were then used to calculate NLC volume. The NLC axis in each case was determined according to a polygonal model and to 2nd, 3rd and 4th degree polynomials. From these models, NLC sectional areas and length were determined. For each variable, descriptive statistics and normality tests (Kolmogorov-Smirnov and Shapiro-Wilk) were established. Main Outcome Measures Time for segmentation, NLC volume, axis, sectional areas and length. Results Mean processing time was around 30 seconds for segmenting each canal. All the variables generated were normally distributed. Measurements obtained using the four models polygonal, 2nd, 3rd and 4th degree polynomial, respectively, were: mean canal length 14.74, 14.3, 14.80, and 15.03 mm; mean sectional area 15.15, 11.77, 11.43, and 11.56 mm2; minimum sectional area 8.69, 7.62, 7.40, and 7.19 mm2; and mean depth of minimum sectional area (craniocaudal) 7.85, 7.71, 8.19, and 8.08 mm. Conclusion The method proposed automatically reconstructs the NLC on CT scans. Using these reconstructions, morphometric measurements can be calculated from NLC axis estimates based on polygonal and 2nd, 3rd and 4th polynomial models.
| year | journal | country | edition | language |
|---|---|---|---|---|
| 2016-01-01 | PLoS ONE |