| PART ONE Three-Dimensional Reconstruction of Brain Surface with MR Imaging:Technical Consideration and Comparison.SubjectIn this part, we reconstructed brain surface structures in various ways and compare their advantages and disadvantages. Content including following: (1) To reconstruct brain surface three-dimensional image based on trace Diffusion-Weighted Images (DWI-3D), and evaluated its efficiency. (2) Three-dimensional reconstruction of brain surface anatomy: Comparison of the techniques based on 3D flash and diffusion-weighted imaging. (3) Three-dimensional reconstruction of brain surface anatomy with MRICro software.Materials and Methods1. Patients: 50 normal cases (male 24, female 26, 11～84 years, average 55 years old) with Diffusion-Weighted Images(DWI-3D), 55 years old and elder cases are 23 persons (11～55 years old, average 39.6 years), younger than 55 years old cases are 27 persons (58~84 years old , average 68.2 years). 30 normal cases (male 17, female 13, 21～66 years old, average 45 years) underwent both DWI and 3D flash. 33 normal adult volunteers (male 17, female 16, 21～39 years old , average 25.6 years) underwent 3D flash MR scan. MR images were utilized to reconstruct brain surface structure with MRICro software.2. MR Scan Protocol: The following parameters were used to obtain MR images. DWI: TR/TE: 5000ms/118ms, field of view (FOV) 220mm×220mm, matrix 128×128, slice thickness 5mm, no gap, parallel to orbitomeatal line, 25 slices were acquired, 4 averages, diffusion-weighted coefficient was b=1000mm~2/s. 3D flash: TR/TE: 30ms/5ms, field of view (FOV) 220～240 mm×220～240 mm, matrix 256×256×196, slice thickness 1.2mm, gaps between slices were 20%, contiguous sagittal images were acquired.3. Images Reconstruction: DWI data were 3D reconstructed with PC workstation using RT 3D software (volume rendering) and observed in various orientations. 3D flash images were also 3D reconstructed with PC workstation using RT 3D software (volume rendering), and then eliminated signals from the scalp and skull and showed brain surface anatomy. MRICro 3D reconstruction was done with PC and Windows2000 system.4. Images Evaluation and Analysis: All of the images were evaluated and analyzed by a group of neuro-radiologists. SPSS 10.0 for Window and t-test was used for statistical analysis.Results1. Three-Dimensional Reconstruction of Brain Surface Image Based on Trace Diffusion-Weighted Images1.1 Analysis of bilateral hemisphere of 50 cases showed: Precentral gyrus, postcentral gyrus, superior parietal lobule, superior frontal gyrus, precentral sulcus, central sulcus, postcentral sulcus, intraparietal sulcus and superior frontal sulcus were best shown among all structures, their scores were 2.61～2.77; Supramarginal gyrus, middle frontal gyrus, inferior frontal gyrus and lateral sulcus were clear shown in the majority of hemisphere with average scores of 2.0～2.49; Angular gyrus, inferior frontal sulcus and superior temporal gyrus were not demonstrated satisfactorily and average scores were 1.67～1.89; Middle temporal gyrus, inferior temporal gyrus, superior temporal sulcus and inferior temporal sulcus were shown bad with average scores of 0.87～1.26.1.2 There was no obviously difference between the bilateral hemispheres in 50 cases.1.3 obviously difference was found between two groups with different ages (≤55years old vs. >55 years old) on the scores of superior frontal gyrus, middle frontal gyrus, superior frontal sulcus, postcentral gyrus, Precentral gyrus, inferior frontal sulcus, inferior frontal gyrus, Supramarginal gyrus, intraparietal sulcus and postcentral sulcus. Central sulcus, precentral sulcus and superior parietal lobule were well shown in both groups with no statistical difference. Inferior temporal gyrus, inferior temporal sulcus, middle temporal gyrus, superior temporal sulcus, angular gyrus and superior temporal gyrus were badly shown with no statistical difference between the two groups.2. Three-dimensional reconstruction of brain surface anatomy: Comparison of the techniques based on 3D flash and diffusion-weighted imaging.Main structures of brain surface were clearly displayed on three-dimensional images based on 3D flash sequence. Average scores were all above 2.50. Precentral gyrus, postcentral gyrus, superior parietal lobule, supramarginal gyrus, superior frontal gyrus, superior temporal gyrus, middle temporal gyrus, precentral sulcus, central sulcus, postcentral sulcus, intraparietal sulcus lateral sulcus and superior temporal sulcus were best displayed with average scores above 2.90.For images based on DWI, precentral gyrus, postcentral gyrus, superior parietal lobule, superior frontal gyrus, precentral sulcus, central sulcus, postcentral sulcus, intraparietal sulcus and superior frontal sulcus were best shown with average scores between 2.60～2.75, However, supramarginal gyrus, angular gyrus, middle frontal gyrus, inferior frontal gyrus, superior temporal gyrus, lateral sulcus, inferior frontal sulcus could not be well shown, with average scores between 1.67～2.48. Middle temporal gyrus, inferior temporal gyrus, superior temporal sulcus and inferior temporal sulcus can only get scores from 0.88 to 1.27.Scores of images based on 3D flash were much higher than that based on DWI with distinct differentiations. Differentiation in supramarginal gyrus, angular gyrus, middle frontal gyrus, inferior frontal gyrus, superior temporal gyrus, middle temporal gyrus, inferior temporal gyrus, lateral sulcus, inferior frontal sulcus, superior temporal sulcus and inferior temporal sulcus were markedly with P values below 0.001.3. Three-dimensional reconstruction of brain surface anatomy with MRICro software.Main structures of brain surface were clearly displayed on three-dimensional images reconstructed with MRICro software. Central sulcus, precentral sulcus, postcentral gyrus, postcentral sulcus, precentral gyrus, intraparietal sulcus and superior parietal lobule were shown best, average scores were above 2.80. Average scores of lateral sulcus, superior frontal gyrus, angular gyrus, supramarginal gyrus, superior frontal sulcus, inferior frontal sulcus, inferior frontal gyrus, middle frontal gyrus, superior temporal sulcus, superior temporal gyrus, middle temporal gyrus, inferior temporal sulcus were 2.52～2.77. Average score of inferior temporal gyrus was 2.39.ConclusionMR three-dimensional reconstruction of brain surface anatomy could be obtained with various methods. Each method has its advantage, as well as limitations. Our research could be regarded as a guild for selection on clinical experience and brain research.PART TWOBrain Lesions: Localization Efficiency Based on MR Brain Surface Three-Dimensional Reconstruction Images.SubjectMR brain surface three-dimensional images were obtained in 23 patients with brain lesions close to cortical. We try to evaluate the displaying effectiveness of these lesions and their relationships with surrounding surface structures based on 3D images and 2D orthogonal slices, and to discuss the value of three-dimensional images in localization of brain lesions.Materials and Methods1. Patients: 23 patients (male 15 cases, female 8 cases, 8～74 years old, average 38.7 years old, mean 41 years old ) were included. Tumor 15 cases, chronic brain infarction 5 cases, arteriovenous malformation 2 cases, schizencephaly 1 case.2. MR Scan Protocol: 3D flash sequence.3. Images Reconstruction: 3D flash images were 3D reconstructed with PC workstation using RT 3D software (volume rendering), and then eliminated signals from the scalp and skull and showed brain surface anatomy.4. Images Evaluation and Analysis: 3D reconstructed images and orthogonal 2D images of 23 patients were respectively evaluated by 2 experienced neuro-radiologists and 2 junior doctors. Locations and extensions of the lesions were recorded.5. Statistical analysis: Mann-Whitney's U Test and Chi-square analysis with SPSS 10.0 for Window.Results1. 3D reconstructed images in 21 cases (91.3%) clearly showed the lesions. Images in 2 patients were unsatisfied.2. Location from 4 doctors: According to 2D orthogonal images, there were 8 (6.78%) sub-lobes regarded by all 4 doctors and 51(43.22%) sub-lobes were localized by only one doctor. Total sub-lobes were 118. Whereas, there were 16(17.2%) sub-lobes regarded by all 4 doctors and 28(30.11%) sub-lobes were localized by only one doctor based on 3D reconstructed images. Total sub-lobes were 93.3. Location from 4 doctors showed: (1)Total regarded sub-lobes were reduced from 118 to 93; (2)Based on 3D images, sub-lobes regarded by all 4 doctors increased from 6.78% to 17.20%. Sub-lobes localized by only one doctor reduced from 43.22% to 30.11%. Distinct differentiation could be found after statistically analyzed. (3)There is no obvious differentiation between the twice localizing at the level of brain lobes. (4)There is a quite different located record among 4 doctors according to 3D and 2D images in the frontal lobe. (5)The sub-lobes recorded by all 4 doctors was merely 17.2% even though according to 3D reconstructed images.4. Localization records by 2 experienced neuro-radiologists and 2 junior doctors based on 2D and 3D images respectively: (1)For 2 experienced neuro-radiologists, concordance rate increased from 50% based on 2D images to 59.7% on 3D images. (2)For 2 junior doctors, concordance rate increased from 16.4% based on 2D images to 32.1% on 3D images. Distinct differentiation could be found after statistically analyzed. (3)Concordance rate of experienced neuro-radiologists (50%) was much higher than that of junior doctors (16.4%) according to 2D images. (4)Concordance rate of experienced neuro-radiologists (59.7%) was still much higher than that of junior doctors (32.1%) according to 3D images. But the gap decreased.Conclusions1. Three-dimensional reconstruction of brain surface could well show lesions and their relationships with surface structures.2. According to 3D images, doctors were easier to get a concordance localization opinion compare to 2D.3. Concordance rates were increased in both experienced neuro-radiologists and junior doctors based on 3D images. 3D images seem to be more helpful to junior doctors and clinical doctors with little tomographic anatomy knowledge.4. Even though based on 3D images, different localization opinions still exist among doctors. Concordance rate of 4 doctors was merely 17.2%. Even between 2 experienced neuro-radiologists, concordance rate was merely 59.7%.PART THREE Peeling off the skull and scalp: a simple way of mapping the superficial cerebral veins on the brain surfacePurpose: A simple technique was developed to display superficial cerebral veins (SCVs) along with brain surface structures in simulation images of craniotomy in 64 patients with brain tumors.Materials and methods: Sixty-four consecutive patients (32 women, 32 men; mean age, 45 years) with brain tumors were studied. All patients underwent MR examination with 3D contrast-enhanced MPRAGE sequence for the neuronavigation system, from which 3D simulation images of craniotomy were reconstructed. We cut the skull and scalp away piece by piece like the way of peeling an apple so that the peeled off window might correspond to the craniotomy planned at surgery. Image quality was subjectively scored from 0 (poor) to 3 (excellent) for both brain surface structures and SCVs respectively. Detectability of the SCVs was evaluated through comparing the images to intraoperative photographs available in 20 cases. Correlations between grades of image quality and degrees of the dural/meningeal enhancement and the effacement of sulci/ subarachnoid space were analyzed.Results: Details of the brain surface structures and the SCVs were generally well demonstrated with average scores of 2.81 for brain surface structures and 2.77 forthe SCVs. Grades of image quality were negatively correlated with degrees of bothdural/meningeal contrast enhancement and local sulci/ subarachnoid spaceeffacement. Detectability analysis of the SCVs showed that veins larger than 2 mmwere demonstrable in 100%, those from 1 to 2 mm in 90%, and those smaller than 1mm in 59.3%.Conclusion: With this simple technique, simulation images of craniotomy were easilyobtained, and would help neurosurgeons "preview" the brain surface structures andSCVs before operation.PART FOURAsymmetry of Central Sulcus in Normal Adults: Application of a New ApproachSubjectPresent a new method to study the asymmetry of central sulcus based on a set ofsoftware. We hope that this novel method could be used to study brain surfacestructures between different persons in future studies.Materials and Methods1. Patients: 33 normal adult volunteers (male 17, female 16, 21～39 years old, average 25.6 years old) were included. All volunteers were right handed.2. MR Scan Protocol: 3D flash scan sequence.3. Images Reconstruction: First, AFNI (Analysis of Functional Neurolmages) was used to convert MR data to .HEAD and .BRIK format files and normalized to the standard brain system (Talairach stereotaxic space). Then, these files were converted to .iso format by MRICro and strip signals from scalp and skull. Brain surface structures were displayed on PC monitor.4. Trace Central Sulcus: Spatial parameters of bilateral central sulcus of 33 volunteers were extract using MRICro software.5. Data and Statistical Analysis: With the help from Origin 6.0 Professional, we can draw projection curves of the bilateral central sulcus of 33 volunteers on XY (axial), XZ (coronal) and YZ (sagittal) coordinate system. Average curve of 33 volunteers was calculated and compared between left central sulcus and right central sulcus. Results1. XY (axial): The original point, curve morphologic and end point from right central sulcus is much variable than that of left central sulcus. The original point and end point of right central were closer to center compared with left with distinct differentiation. Three marked variability (SD peaks) along the superficial central sulcus could be seen bilaterally.2. XZ (coronal): On the coronal plane, curve morphology from 33 volunteers were quite coordination. The end point of left central culcus went little further than right.3. YZ (sagittal): Both of the central culcus were variation on sagittal plane. Right central sulcus was much obviously than left. Right central sulcus was slightly anterior than left, especially within the scope of 30～50 on Z axial.Conclusions1. It is feasible to combine AFNI,MRICro and Origin 6.0 Professional as a new way to study brain surface structures.2. Curves morphology from 33 volunteers was quite symmetric between bilateral hemispheres. Four features worth to be attention: First, The original point, curve morphologic and end point from right central sulcus is much variable than that of left central sulcus. Second, Three marked variability (SD peaks) along the superficial central sulcus could be seen bilaterally. Third, the original point and end point of right central were closer to center compared with left with distinct differentiation. Forth, Right central sulcus was slightly anterior than left, especially within the scope of 30～50 on Z axial.
|
PDF Full Text Request