Measurements And Localization Of Brain Structures On CT Images

Cerebrovascular disease,can be classified into two kinds from its origin: hemorrhagic cerebrovascular disease and ischemic cerebrovascular disease.There are about 100 million people died from cerebrovascular disease in our country every year, and the fatality rate is about 45%.Substantial data shows that in the rescured survivors after suffering from cerebrovascular disease,about 50%～80%patients are left with some degrees of disabled symtoms.According to statistics report,the rescued survivors with treatment of cerebrovascular disease are easy to relapse,with about 20%～47%relapsing rate in 5 years.Although hemorrhagic cerebrovascular disease only accounts for 15%of cerebrovascular disease,it has higher prevalence rate,higher disability,and higher mortality rate,which has become an acute and severe disease in neurology.At present,the surgery time of cerebral hemorrhage advocate ultra-early(intracerebral hemorrhage time＜6 h) or early surgery(cerebral hemorrhage time＜12 h),therefore,fast and accurate diagnosis of bleeding and bleeding parts is required.Diagnosis of cerebral hemorrhage by clinical physicians in emergency department are viewing the CT images,and measuring blood chips only in accordance with CT bleeding rough terms.The shortcoming of this clinical diagnosis approach is its long diagnosis time and inaccurate blooding calculation,which undoubtedly have an impact on patient's surgery plan and surgery symptoms.With the development of modern computer science and technology,medical image processing and analysis has attracted more and more attention,and been widely used in clinical,doctor-assisted diagnosis and surgery to accurate and reliable diagnosis.Combination of medical image analysis algorithm and computer software, can automatically locate cerebral hemorrhage faster on CT images,and precisely measure the amount of bleeding.However,it is difficult to determine the anatomy name of bleeding region,which is difficult to acquire diagnostic conclusions,for the structure name of bleeding region can't not be located by computer.At present,many research work have been carried out to the study of localizing brain structure.For example,the localization of cerebral sulcus and gyrus under the brain surface,and the three-dimensional position at the internal structure of the brain,have a wide range of applications and achieved satisfactory therapeutic effect,but there is little research about computer-assisted positioning bleeding region on CT images to help clinical physician to diagnosis hemorrhagic cerebrovascular disease.In this study based on the anatomical images according to the clinical classification of cerebral hemorrhage (bleeding into the basal ganglia region,thalamus hemorrhage,lobar hemorrhage, brainstem hemorrhage,cerebellar hemorrhage and intraventricular hemorrhage),we aim to determine cerebral structures and their standard sizes to divide brain structures by measuring and analyzing the CT brain data from healthy adults(20～86-year-old people),so as to assist computerized automatic location of cerebral anatomy structures,and to achieve a rapid diagnosis of cerebral hemorrhage by computer.The purpose of this study include:1.measurement the height of quadrigeminal cistern on CT images,analysis its feasibility as a bio-mark to distinguish between the cerebrum and cerebellar;2.measurement the length and width of the pons and incline angles of internal capsule on CT images,and set standard diameters for the distinction between the brain stem and basal ganglia region;3.using the three-dimensional reconstruction software -3DMed,to reconstruct the lateral ventricle,the quadrigeminal cistern and skull,so as to assist the Medicine Anatomy Teaching.Methods:1.The height of quadrigeminal cistern of 20 sets CT brain data from 20 healthy adults,with thickness of 2.5mm and scaned by ears-canthus line(OML) were measured.Morphological description were done on the CT data acquired using Centricity DICOM Viewer software(GE viewer-chip software software) to check the beginning and ending scan calibration,and calculate the average height and minimum height of the quadrigeminal cistern.2.152 sets of CT brain data from 152 healthy adults(male 80,female 72,aged from 20 to 86),scanned by ears-canthus line(OML),were selected.Morphological descriptions were done on the CT data acquired.Divided by age:Ⅰgroup age＜60,Ⅱgroup age≥60,the length and width of pons were measured by MxLite View DICOM Viewer software(Philips viewer-chip software);By skull type,it was divided into long,square,round three groups,and incline angles of internal capsule were measured by MxLite View DICOM Viewer software.Using the statistical software SPSS13.0 to calculate the mean and standard deviation,and inter-group comparison were done by analysis of variance and t-tests,and also the standard values of the length and width of pons and incline angles of internal capsule were calculated.3.A set of CT brain data from 20 healthy adults with thickness of 2.5mm, scanned by ears-canthus line(OML) was selected.A segmented software -Ground truth editor was used to segment the lateral ventricle,the quadrigeminal cistern and skull.Three-dimensional reconstruction software -3DMed was used to reconstruct the segmented results to yield a three-dimensional visual model,which represents the relationship of spatial location among them.Results:1.The quadrigeminal cistern is located in the back of the midbrain,and shown as an arc of low-density shadow region between the midbrain and the vermis of cerebellum,with a minimal height of 13.20mm.Its average height is(15.38±1.27) mm.2.The fourth ventricle is located between the pons and the vermis of cerebellum. In front of the pons is the prepons cistern,and outside of it is the cistern of cerebello-pons angle cistern.The average length and width of pons are(29.98±1.91) mm,(23.93±1.57) mm for male,and(28.48±1.74) mm,(23.00±1.60) mm for female respectively.Factorial analysis showed that there is no significant differences in interaction effect,but the main effect of inter-group difference is significantly(F = 22.182,P = 0.000),the length of pons for male is longer than that of female. Two-sample t test results showed that there is significant difference between＜60-year-old male and＜60-year-old female(t = 4.388,P = 0.000),length of ports for＜60-year-old male is longer than that of female for＜60-year-old;and there is also significant difference between≥60-year-old male with the length of pons and≥60-year-old female(t = 2.541,P = 0.014),length of pons for≥60-year-old male is longer than that of female for≥60-year-old.The results calculated by two-way ANOVA showed that there is significant differences between gender groups(F = 25.106,P= 0.000),the length of pons for male is longer than that of female,and age groups showed no significant difference(F = 0.002,P = 0.968).About the measurements of the width of the pons,the results of factorial analysis showed that it is no significant differences in interaction effect,but the main effect of inter-group difference is significantly(F = 14.629,P = 0.000),the width of pons for male is wider than that of female.Two-sample t test results showed that there is significant difference between＜60-year-old male and＜60-year-old female(t = 2.290,P = 0.026), the width of pons for＜60-year-old male is wider than that for＜60-year-old female; there is significant difference between≥60-year-old male and≥60-year-old female (t= 2.945,P = 0.005),the width of pons for≥60-year-old male is wider than that for≥60-year-old female.The results calculated by two-way ANOVA showed that there is significant differences between gender groups(F = 12.543,P = 0.001),the width of ports for male is wider than that of female,and age groups showed no significant difference(F = 0.628,P = 0.429).3.The internal capsule is shown as a low-density region shaped "＞＜".The posterior limb of internal capsule is shown as a clear line dividing the thalamus which is located in the back-inside of it and the basal ganglia region which is located in front-outside of it.The incline angles of the internal capsules of the long,square,and round skulls were 36.40°±2.48°,39.09°±2.78°and 42.64°±2.02°respectively. Statistical analysis showed that the inter-cranial-type were significantly different (P＜0.001).4.Models of the lateral ventricle,the quadrigeminal cistern and the skull were built,and they have a clear spatial location relationship.All the reconstructed structures could be arbitrarily matched-displayed and jointly-displayed,in addition they could be rotated around any spatial axis.They can also be showed three-dimensional space at any location around the shaft on rotation of arbitrary angle.Conclusion:1.The quadrigeminal cistern measurement results showed that the minimum height＞10mm.The quadrigeminal cistern can be observed in the 10mm conventional CT scan.As long as the quadrigeminal cistern appears,we can determine that the structure of its back is the cerebellum.Therefore,the quadrigeminal cistern can be used for the division of cerebrum and cerebellum on 10mm CT data.2.Under the basis of the subarachnoid cisterns around the pons,the length and width measurements of the pons can localize the pons accurately.Therefore,the subarachnoid cisterns around the pons,the fourth ventricle and the length and width measurements of the pons can be the assisted to localize sign of pons and brain stem.3.The posterior limb of internal capsule located between the thalamus and the basal ganglia region,which showed up in CT "/ \"-shaped low-density region,with the thalamus and the basal ganglia region on both sides of a clear distinction.Under the basic localization of the low-density region of capsule,the measurements diameter of the incline angles of the internal capsules can distinguish them accurately. Therefore,the internal capsule can help to determine the division of the thalamus and the basal ganglia region.4.Reconstruction results of the lateral ventricle,the quadrigeminal cistern and the skull showed a very good display of their three-dimensional spatial relationship, and can be used for clinical teaching.Three-dimensional image reconstruction has the following advantages:Ⅰ.It can be observed by rotating along any axis at any angle;Ⅱ.Three-dimensional image reconstruction can be sectioned by any plane;Ⅲ.In the implementation of cross-sectional,sagittal cutting surface and coronal plane,we can conduct some interactive operation;Ⅳ.We can synchronize the camera operation, dynamic three-dimensional display,and export the process for pre-operative and clinical teaching.