Hybrid Rendering of Multidimensional Image Data

The most important rendering methods applied in medical imaging are surface and volume rendering techniques. Each approach has its own advantages and limitations: Fast surface-oriented methods are able to support real-time interaction and manipulation. The underlying representation, however, is dependent on intensive image processing to extract the object surfaces. In contrast, volume visualization is not necessarily based on extensive image processing and interpretation. No data reduction to geometric primitives, such as polygons, is required. Therefore, the process of volume rendering is currently not operating in real time. In order to provide the radiological diagnosis with additional information as well as to enable simulation and preoperative treatment planning we developed a new hybrid rendering method which combines the advantages of surface and volume presentation, and minimizes the limitations of these approaches. We developed a common data representation method for both techniques. A preprocessing module enables the construction of a data volume by interpolation as well as the calculation of object surfaces by semiautomatic image interpretation and surface construction. The hybrid rendering system is based on transparency and texture mapping features. It is embedded in a user-friendly open system which enables the support of new application fields such as virtual reality and stereolithography. The efficiency of our new method is described for 3-D subtraction angiography and the visualization of morpho-functional relationships.

Magnetic resonance-based motion analysis of the shoulder during elevation

Changes in shoulder motion patterns are relevant in various shoulder diseases, but no in vivo information exists about the relative positions in vivo of the shoulder girdle bones and the supraspinatus muscle in three-dimensional space. Thus, the objective of this study was to perform a motion analysis of these structures during passive arm elevation using open magnetic resonance imaging and three-dimensional image processing. Fourteen volunteers were examined in five positions of abduction (30 degrees-150 degrees) with an open magnetic resonance system. After segmentation and three-dimensional reconstruction, the axis of the supraspinatus, humerus, clavicle, and the plane of the glenoid were determined, and the relative movements were calculated. The ratio for glenohumeral to scapulothoracic motion was 1.5:1 at 60 degrees and 2.4:1 at 120 degrees abduction. At 30 degrees, the axis of the supraspinatus was nearly horizontal, and during abduction a continuous elevation (+123 degrees at 150 degrees abduction) was measured. In the transverse plane, the angle between the supraspinatus and the clavicle axes became larger during abduction because of an increasing retroversion of the clavicle. The study shows specific three-dimensional motion patterns for each bone of the shoulder girdle and the supraspinatus muscle during passive elevation. The technique and results can be used for future studies in patients with pathologic changes of shoulder girdle motion.

A new hybrid renderer for virtual bronchoscopy

PURPOSE

To improve the diagnosis of pathologic modified airways, a new hybrid visualization system has been developed and tested based on digital image analysis and synthesis of spiral CT as well as the visual simulation of bronchoscopy.

METHOD/MATERIALS

20 patients with pathologic modifications of the airways (tumors, lung transplantation) were examined with Spiral-CT. The shape of the airways and the lung tissue is defined by an automatic volume growing method and a following geometric reconstruction by the computation of geometric primitives. This is the basis of a multidimensional display system which visualizes volumes, surfaces and computation results simultaneously. The enable the intuitive and immersive inspection of the airways a virtual reality system, consisting of two graphic engines, a head mounted display system, data gloves and specialized software was integrated.

RESULTS

In 20 cases the extension of the pathologic modification of the airways could be visualized with the virtual bronchoscopy. The user interacts with and manipulates the 3D model of the airways in an intuitive and immersive way. In contrast to previously proposed virtual bronchoscopy systems the described method permits truly interactive navigation, detailed quantitation of anatomic structures and a "see through" the bronchial wall. The system enables a user oriented and fast inspection of the volumetric image data.

CONCLUSIONS

To support radiological diagnosis with additional information a virtual bronchoscopy was developed. It enables the immersive and intuitive interaction with 3D Spiral CTs by truly 3D navigation in the airways. The system was tested with 20 Spiral-CTs of bronchial tumors and obstructions and is well suited for the inspection of structures beyond the bronchialtree.

Three-dimensional analysis of the width of the subacromial space in healthy subjects and patients with impingement syndrome

OBJECTIVE

The aim of this study was to per form a three-dimensional analysis of the width of the subacromial space during passive and active arm abduction in healthy volunteers and patients with impingement syndrome.

SUBJECTS AND METHODS

The shoulders of 10 healthy subjects and 10 patients with impingement syndrome were imaged with an open MR system during abduction, with and without activation of the shoulder muscles. An apparatus was designed for applying an adduction force of 10 N to the distal humerus during image acquisition, and the minimal acromiohumeral distance was measured after three-dimensional reconstruction.

RESULTS

In the 10 healthy volunteers, muscle activity led to a significant decrease (-32%; p < .05) of the acromiohumeral distance at 60 degrees of abduction, whereas at 120 degrees of abduction the distance was significantly increased (+44%; p < .05). In these volunteers, muscle activation caused no significant effect at 90 degrees of abduction. However, in the 10 patients with impingement syndrome, muscle activity led to a significant decrease in the width of the subacromial space compared with that of the healthy contralateral side (-68%; p < .05).

CONCLUSION

Muscle activity and arm position were found to cause systematic changes in the width of the subacromial space. However, functional deficits of the supraspinous muscle in patients with early-stage impingement syndrome were not apparent during muscle relaxation.

A non-destructive technique for 3-D microstructural phenotypic characterisation of bones in genetically altered mice: preliminary data in growth hormone transgenic animals and normal controls

A non-destructive, three-dimensional technique for microstructural phenotypic characterisation of skeletal elements in genetically altered mice is presented. Preliminary data in bovine growth-hormone transgenic animals and control littermates are shown. The technique is based on microcomputed tomography (microCT) and digital postprocessing and allows for a differential quantitative analysis of the cortical and trabecular bone compartments in the axial and peripheral skeleton. The distal femora and the first lumbar vertebral bodies of six animals were CT scanned in the axial plane with an isotropic resolution of 20 microm. The periostal surface and the marrow spaces were segmented fully automatically, and the trabecular and cortical compartments were separated interactively. After 3-D reconstruction, various regions of interest (diaphyseal, metaphyseal and epiphyseal) were selected for the analysis. The femora and vertebrae of the transgenic animals showed obvious differences in size, shape, and trabecular arrangement compared with the control animals. The total bone mass was increased by a factor of two to three, but the trabecular bone was increased much more (up to 12 times) than the cortical bone. The transgenic animals showed an increased ratio of trabecular vs cortical bone (0.90 to 1.27 vs 0.14 to 0.36 in the femoral diaphysis) and an elevated trabecular bone volume fraction (49% to 73% vs 18% to 43% in the femoral metaphysis). The mean 3-D cortical thickness was similar in the normal and transgenic animals (values between 93 microm and 232 microm in the dia- and metaphyses), but the minimal cortical thickness was lower in the transgenic animals (22 to 31 microm vs 54 microm to 110 microm in the diaphysis). The technique presented is suitable for phenotypic characterisation of bone structure in genetically altered mice.

Hybrid rendering and virtual endoscopy of the auditory and vestibular system

A hybrid rendering method (color-coded 3D shaded-surface and volume display) with the possibility of virtual endoscopy using image data sets from HR-SCT was developed. To show the possible advantages and benefits of the improved rendering algorithm we have specifically highlighted the use in relation to the auditory and vestibular system. Postprocessing image visualization offers improved morphological analysis, and will benefit radiological diagnostics, medical education, surgical planning, surgical training and postoperative assessment.

[Virtual and three-dimensional bronchoscopy with spiral and electron beam computed tomography]

PURPOSE

To compare spiral computed tomography (CT) and electron-beam CT (EBT) for 3D and virtual CT-bronchoscopy.

MATERIALS AND METHODS

17 patients with various disorders of the tracheobronchial system were examined using fiberoptic bronchoscopy, spiral CT and EBT. 3D images were reconstructed from CT data sets using automated segmentation based on volume-growing methods. Surface-rendered, volume-rendered, and hybrid reconstructions were visualized in real time using a data helmet.

RESULTS

All data sets could be processed to high-quality three-dimensional (3D) and virtual reconstructions. The reduction of motion artifacts due to shorter scan times made EBT data sets better suited for automated segmentation and less susceptible to motion artifacts. 3D and virtual reconstructions did not increase the diagnostic sensitivity of CT compared to axial reconstructions alone.

CONCLUSIONS

Shorter scan times of CT imaging yield higher-quality 3D and virtual reconstructions. Modern reconstruction techniques are valuable visualization tools for select indications and are the prerequisite for future developments in computer-aided medicine.

[Computer-assisted diagnosis based on computer-based image interpretation and 3D-visualization]

PURPOSE

To survey methods for 3D data visualization and image analysis which can be used for computer based diagnostics.

MATERIAL AND METHODS

The methods available are explained in short terms and links to the literature are presented. Methods which allow basic manipulation of 3D data are windowing, rotation and clipping. More complex methods for visualization of 3D data are multiplanar reformation, volume projections (MIP, semi-transparent projections) and surface projections. Methods for image analysis comprise local data transformation (e.g. filtering) and definition and application of complex models (e.g. deformable models).

RESULTS

Volume projections produce an impression of the 3D data set without reducing the data amount. This supports the interpretation of the 3D data set and saves time in comparison to any investigation which requires examination of all slice images. More advanced techniques for visualization, e.g. surface projections and hybrid rendering visualize anatomical information to a very detailed extent, but both techniques require the segmentation of the structures of interest. Image analysis methods can be used to extract these structures (e.g. an organ) from the image data.

DISCUSSION

At the present time volume projections are robust and fast enough to be used routinely. Surface projections can be used to visualize complex and presegmented anatomical features.

Simultaneous occurrence of various mutations and polymorphisms in cis and in trans of the galactose-1-phosphate uridyltransferase gene in a Turkish family with classical galactosemia

Classical galactosemia, characterized clinically by acute hepatic dysfunction, sepsis, cataract, and failure to thrive, is caused by deficiency of galactose-1-phosphate uridyltransferase (GALT). Galactose restriction normalizes these acute symptoms; however, long-term complications such as intellectual deficits and ovarian failure are conspicuous in the majority of patients. Here we report two Turkish siblings with classical galactosemia. The clinical course of the two children differed markedly: only the older girl suffered from severe acute symptoms during the neonatal period, and she developed greater mental retardation than her younger affected brother. The functional activity of GALT was virtually absent in each affected children. The mother and two healthy siblings exhibited approximately 50% normal GALT activity and the father approximately 25%. Molecular analysis revealed that these two galactosemic siblings were homozygous for a stop codon mutation of E340X in GALT exon 10. Moreover, two additional mutations, a neutral polymorphism L218L and N314D, which are typical for the Duarte-I variant, were found in the same GALT allele. The two healthy siblings and the parents were heterozygous for these combinations of mutations. In addition, the father's second GALT allele revealed three intron mutations at nucleotide position 1105 (G-->C), 1323 (G-->A) and 1391 (G-->A) and the N314D mutation, which correspond to the mutations of Duarte-2 variant. Our findings indicate that in classical galactosemia several distinct mutations can be present in one allele (in cis) of the GALT gene. Therefore it seems to be necessary to examine all introns and exons of the GALT gene in galactosemic patients who do not carry the Q188R mutation or another frequent mutation in the GALT gene.

[Relevance of susceptibility-induced geometrical distortion for validity of MRI-based cartilage volume and density measurements of the knee joint]

The aim of the present study was to analyze the relevance of susceptibility-induced geometrical distortion to the accuracy of MR-based cartilage volume and thickness measurement in the human knee joint. Nine cadaveric knee joints were imaged in the sagittal plane with MRI at a resolution of 2 x 0.31 x 0.31 mm3, using a fat-suppressed gradient echo sequence, with a normal gradient orientation and also with the frequency- and phase-encoding directions changed. CT arthrographic data sets were then obtained. On the basis of 3-D constructions, we determined the cartilage volume and, with a 3-D minimal distance algorithm, the thickness distribution, of the patella, femur and tibia. Irrespective of the gradient orientation, good agreement was observed between MRI and CT arthrography in terms of cartilage volumes and maximum cartilage thickness. With a normal gradient orientation the volume was overestimated by 2.5% in MRI, and 2.3% when the gradients were changed. The maximum cartilage thickness was underestimated by 0.24 intervals (interval = 0.5 mm) with a normal gradient orientation, and by 0.22 intervals when the gradient orientation was changed. In none of the joint surfaces was a relevant difference between the two methods observed. It can be shown that, using high-resolution, fat-suppressed gradient-echo sequences--susceptibility-induced geometrical distortion has no significant effect on the accuracy of MR-based cartilage volume and thickness measurements. MRI would therefore appear suitable for the design of patient-specific finite element models with the aim of analysing load transmission in diarthrodial joints and planning surgical interventions.

Accuracy of cartilage volume and thickness measurements with magnetic resonance imaging

A noninvasive imaging technique for quantifying articular cartilage is needed for diagnosis, monitoring, and therapy control in osteoarthritis. In this study the accuracy of three-dimensional cartilage volume and thickness measurements in the knee with magnetic resonance imaging was analyzed. Eight cadaveric specimens had sagittal imaging with a fat suppressed gradient echo sequence. After a contrast agent was injected, two sagittal computed tomography data sets were obtained, with the knees being repositioned between the examinations. The cartilage thickness was determined, after three-dimensional reconstruction, using a minimal distance algorithm. The mean absolute volume deviation between magnetic resonance imaging and computed tomography arthrography was 3.3% and that between the two computed tomography data sets was 3.6%. The absolute error in determining the maximal cartilage thickness with magnetic resonance imaging was on average 0.6 intervals (of 0.5-mm thickness) and that between the computed tomography examinations was 0.5 intervals. In a patient with anterior knee pain, a focal cartilage defect was seen with magnetic resonance imaging, and this was verified by arthroscopic examination. Using three-dimensional image processing, magnetic resonance imaging can provide accurate data on cartilage volume and thickness in the human knee joint surfaces. This imaging technique potentially may be valuable in the treatment of patients with joint disease.

Quantitative relationships of normal cartilage volumes of the human knee joint--assessment by magnetic resonance imaging

The objective of this study was to assess the normal range of cartilage volumes in the knee joints of healthy adults, the ratio between the patellar, femoral, and tibial cartilages, and the correlation of the volumes with age, body weight, height, body mass index (obesity), patellar bone size, and the diameter of the tibial head. We examined the knee joints of nine healthy volunteers and eleven normal post-mortem specimens with an age range of 24 to 82 years. The cartilage volumes of the patella, femur, medial tibia and the lateral tibia were quantified, using a fat-suppressed FLASH-3D sequence (resolution 2x0.31x0.31 mm3) and digital postprocessing, involving three-dimensional reconstruction. The mean total volume of the knee joint cartilage was 23,245 mm3, the relative standard deviation (CV%) 19%, and the range 16,341 to 33,988 mm3. In the patella, femur and tibia, the CV% amounted to between 22 and 25%. These joint surfaces occupied a relatively variable proportion of the total knee joint volume, the percentage of the patella being 11 to 22%, that of the femur 54 to 69%, that of the medial tibia 7 to 12%, and that of lateral tibia 11 to 16%. The volumes of the lateral tibia were systematically higher than those of the medial tibia (P<0.001). There was no significant correlation of the knee joint cartilage volume with age (r=+0.05), body weight (r=+0.38), height (r=+0.39) or body mass index (r=+0.29), but a relatively high correlation with the diameter of the tibial head (r=+0.78, P<0.001). After normalising the volumes to this diameter, the CV% of the total knee joint cartilage volume was reduced to 13%, its variation being 12 to 21% in the patella, femur and tibia. MRI is available for measuring cartilage volume during growth, functional adaptation, and tissue loss in degenerative joint disease. The study shows that a wide variation of cartilage volumes exists in the knee joints of normal adults. To reduce the variability between individuals, the cartilage volumes may be normalised to the head of the tibial diameter.

Hybrid rendering of the cervico-cranial arteries using spiral computed tomography

PURPOSE

The purpose of this study was to develop a shaded-surface and volume display (hybrid rendering method) of the whole vascular system of the cervico-cranial arteries using spiral computed tomography (SCT).

MATERIALS AND METHODS

We examined 12 patients with anatomic abnormalities and pathological conditions of the arterial vascular system. The cervico-cranial arteries were segmented using an interactive threshold interval density volume-growing method and visualized with a color-coded shaded-surface display (SSD) rendering method. The adjacent bone structures were visualized using a transparent volume rendering method.

RESULTS

In all cases, the entire volume of the vascular system of the cervico-cranial arteries and the anatomic abnormalities and pathological conditions could be visualized.

CONCLUSION

Hybrid rendering of the circulation of the cervico-cranial arteries using image data sets from a subsecond spiral CT scanner is useful for the visualization of anatomical and pathological abnormalities and offer a promising minimally invasive alternative compared with other diagnostic procedures.

VR interaction techniques for medical imaging applications

Methods of virtual reality (VR) offer new ways of human-computer interaction. Medicine is predestined to benefit from this new technology in many ways. Virtual environments can support physicians in their work, alleviate communication between specialists from different fields or be established in educational and training applications. For the field of visualization and analysis of three-dimensional anatomical images (e.g. CT or MRI scans), an application is introduced which expedites recognition of spatial coherencies and the exploration and manipulation of the 3D data. To avoid long periods of learning and accustoming and to facilitate work in such an environment, a powerful human-oriented interface is required allowing interactions similar to the real world and utilization of our natural experiences. This paper shows the use of eye tracking parameters for a level-of-detail algorithm and the integration of a glove-based hand gesture recognition into the virtual environment as an essential component of the human-machine interface. Furthermore, virtual bronchoscopy and virtual angioscopy are presented as examples for the use of the virtual environment.

Effect of physical exercise on cartilage volume and thickness in vivo: MR imaging study

PURPOSE

To quantify, with magnetic resonance (MR) imaging, the in vivo changes in cartilage volume and thickness after physical exercise.

MATERIALS AND METHODS

The patellae of eight volunteers were imaged six times at physical test by using a spoiled fat-suppressed gradient-echo sequence with an acquisition time of 4.10 minutes. The volunteers then performed 50 knee bends, and two more data sets were acquired 3-7 minutes and 8-12 minutes after exercise. The patellar cartilage volume was determined after three-dimensional reconstruction, and the thickness was assessed with a three-dimensional minimal-distance algorithm.

RESULTS

Whereas repositioning had a small effect on the measurements (mean coefficient of variation, 1.4%), a statistically significant decrease in cartilage volume was observed 3-7 minutes (mean decrease, 6.0%; P < .05) and 8-12 minutes (mean decrease, 5.2%; P < .05) after exercise. The deformation was homogeneous throughout the joint surface. In one asymptomatic volunteer, a cartilage lesion became more pronounced after exercise.

CONCLUSIONS

MR imaging can be used to investigate the response of articular cartilage to physical exercise in vivo. Patients or volunteers should be allowed a sufficient period of physical rest if quantitative measurements of cartilage volume and thickness are to be undertaken in longitudinal studies.

In vivo reproducibility of three-dimensional cartilage volume and thickness measurements with MR imaging

OBJECTIVE

Previous studies suggest that MR imaging is capable of providing accurate data on knee joint cartilage volume and thickness in vitro, but the reproducibility of these data in living subjects has not been analyzed rigorously. Our aim was therefore to determine the in vivo reproducibility of volume and thickness measurements from replicated data sets, applying three-dimensional (3D) postprocessing methods.

SUBJECTS AND METHODS

Eight healthy volunteers were imaged six times at a resolution of 2 x 0.31 x 0.31 mm with a fat-suppressed fast low-angle shot 3D sequence, the knee being repositioned in between replicated examinations. Three-dimensional reconstructions of the articular cartilage surfaces were obtained from sagittal data sets, and the cartilage volumes were calculated. The thickness distribution was analyzed throughout the joint surfaces independent of the section orientation, using a previously validated 3D minimal-distance algorithm.

RESULTS

In the volunteers, the coefficient of variation for replicated volume measurements ranged from 1.3% (patella) to 3.4% (lateral tibia), and the standard deviation of the individual cartilage volumes ranged from +/- 16% (lateral tibia) to +/- 22% (femur). The intraclass correlation coefficient ranged from .959 (lateral tibia) to .995 (patella). The interobserver evaluation was similar to the interscan reproducibility. The mean interscan deviation of the maximal cartilage thickness interval ranged from 0.1 to 0.3 cartilage thickness intervals (of 0.5 mm); only in rare cases did we record deviations greater than one thickness interval.

CONCLUSION

MR imaging can be used to determine cartilage volume and thickness in the knee joints of living subjects with high precision, provided that a fat-suppressed gradient-echo sequence with adequate resolution and 3D digital image processing are used.

Spiral CT angiography and 3D reconstruction in patients with aortic coarctation

The objective of this study was to assess the reliability of spiral CT angiography (CTA) and 3D reconstruction in patients with aortic coarctation (CoA). Eighteen patients with suspected or surgically proven coarctation were examined by spiral CT. In addition to the axial slices, 3D reconstructions, such as shaded surface display (SSD) and maximum intensity projection (MIP), were used to determine the diameters of the CoA and the pre- and poststenotic aorta and to visualise the collateral vessels. Diameters derived from cardiac catheterization were compared with those from CTA in 8 patients. The degree of aortic stenosis was correlated with blood pressure gradients (BPG) in 12 patients. The difference between the diameters of the CoA and the pre- and poststenotic aorta derived from MIP and angiography was not statistically significant (p = 0.69). With SSD the internal thoracic artery was detected in 16 and the posterior intercostal artery in 13 cases. The degree of aortic stenosis correlated poorly with the BPG (r = 0.51, r2 = 0.26). CTA with 3D reconstruction represents a reliable noninvasive technique for the assessment of the degree of CoA and the visualisation of collateral vessels. It may serve as a follow-up investigation after intervention or surgical treatment.

[Principles and current possibilities of virtual scenarios for surgery planning]

This paper describes several new visualization and interaction techniques that enable the use of virtual environments for routine medical purposes. A new volume-rendering method supports shaded and transparent visualization of medical image sequences in real-time with an interactive threshold definition. Based on these rendering algorithms a segmentation approach offers intuitive assistance for a wide range of requirements in diagnosis and therapy planning. In addition, a hierarchical data representation for geometric surface descriptions guarantees optimal use of available hardware resources and prevents inaccurate visualization. Applications such as virtual endoscopy are described.

Repeatability of patellar cartilage thickness patterns in the living, using a fat-suppressed magnetic resonance imaging sequence with short acquisition time and three-dimensional data processing

A fast, reproducible, and noninvasive method is required for quantifying cartilage thickness clinically and for studying the deformation of articular cartilage during and after mechanical loading in vivo. The objective of the current investigation was to test the repeatability of regional distribution patterns of patellar cartilage thickness in the living on the basis of a fat-suppressed magnetic resonance imaging sequence with a short acquisition time and three-dimensional digital data processing. The knees of eight healthy volunteers were transversally imaged with a fat-suppressed FLASH-3D (fast low angle shot) sequence (acquisition time: 4 minutes and 10 seconds). In each case, the joint was newly positioned before each of the six replicate measurements was taken. The patellar cartilage was reconstructed three-dimensionally, and the distribution of cartilage thickness was determined with a three-dimensional minimal-distance algorithm. Whereas the cartilage volume ranged from 3,198 to 7,149 mm3, the mean coefficient of variation for the 6-fold volume measurement was 1.35%. On average, 75.1% (+/- 4.1%) of all test pixels could be attributed to the same cartilage thickness interval (0.5 mm) by image analysis; 14.8% (+/- 2.4%) deviated by one interval; 6.6% (+/- 1.5%), by two intervals; and 3.5% (+/- 1.8%), by more than two intervals. We conclude that, on the basis of a magnetic resonance imaging sequence with an acquisition time of less than 5 minutes, the quantitative distribution of cartilage thickness can be determined with high precision in vivo.

[Three-dimensional thickness and volume measurements of the knee joint cartilage using MRI: validation in an anatomical specimen by CT arthrography]

PURPOSE

In the present study we intended to validate knee joint cartilage volume and thickness measurements with MRI.

METHODS

Ten fresh cadaver knees (age 29 to 64 yrs.) were sagittally imaged, using a fat-suppressed FLASH-3D sequence with a resolution of 2 x 0.31 x 0.31 mm3. Then, a contrast agent was injected and the specimens submitted to CT arthrography. From both modalities the patellar, femoral, and tibial cartilages were segmented semiautomatically and reconstructed three-dimensionally. The cartilage thickness was determined independently of the sectional plane, based on a "minimal distance algorithm".

RESULTS

The volumes and the regional distribution patterns yielded a very high degree of similarity on direct comparison of both imaging modalities. The average volume error between MRI and CT was 3.8% (+/- 3.0%), the correlation 0.998, the slope of the regression line 1.04 and the gamma-intercept -80 mm3. The analysis yielded no significant differences between the two methods (Wilcoxon signed rank test, 5% level) in the patella, femur, medial, and lateral tibia.

CONCLUSION

The results suggest that, based on a fat-suppressed FLASH sequence with high resolution and three-dimensional concepts of digital image analysis, the cartilage volume and thickness can be analysed non-invasively and with high accuracy by MRI.

A non-invasive technique for 3-dimensional assessment of articular cartilage thickness based on MRI. Part 2: Validation using CT arthrography

Established methods for the measurement of articular cartilage thickness are invasive and cannot be sequentially applied in living subjects. In the present study, the distribution of cartilage thickness throughout entire joint surfaces was determined from MR images obtained with a fat-suppressed gradient-echo sequence at a resolution of 0.31 x 0.31 x 2.00 mm3, and compared to that derived from CT arthrography. A minimal distance algorithm was employed to produce 3D cartilage thickness maps of seven cadaveric human knee joints. The mean amount of deviation of the cartilage volumes was 5.6% (+/- 4.6), statistical analysis showing that there was high agreement between the two methods (r = 0.995, slope = 1.037, y-intercept = -90.5 mm3). The 3D thickness maps yielded a striking agreement between the two methods, the maximum values generally yielding a deviation of none or one thickness interval of 0.5 mm. This investigation shows that accurate 3D assessment of articular cartilage thickness can be performed with MRI, this technique having the advantage that it is suitable for investigating living subjects.

A non-invasive technique for 3-dimensional assessment of articular cartilage thickness based on MRI. Part 1: Development of a computational method

Articular cartilage thickness is of relevance in various fields in diagnostics and biomedical research. In view of recent improvements of MR cartilage imaging a computational method has been developed for three-dimensional determination of cartilage thickness from tomographic datasets. A correction algorithm that compensates for the error implied in the voxel based distance measurements is implemented. Four different thickness definitions have been applied to two numerical test structures in order to judge their usability in the medical realm. The results for each of the thickness measurement methods are shown as color-coded thickness maps wrapped round the test objects. An algorithm determining at each point the minimal distance from the articular surface to the bone-cartilage interface is suggested to give the most suitable demonstration of articular cartilage. This algorithm is successfully applied to a 3-dimensional data set of human knee joint cartilage obtained with a fat-suppressed gradient-echo sequence from a healthy volunteer. A non-invasive method for determining cartilage thickness could become a very valuable tool in diagnostic radiology, orthopaedic practice and biomechanics.

Virtual reality in medicine-computer graphics and interaction techniques

This paper describes several new visualization and interaction techniques that enable the use of virtual environments for routine medical purposes. A new volume-rendering method supports shaded and transparent visualization of medical image sequences in real-time with an interactive threshold definition. Based on these rendering algorithms two complementary segmentation approaches offer an intuitive assistance for a wide range of requirements in diagnosis and therapy planning. In addition, a hierarchical data representation for geometric surface descriptions guarantees an optimal use of available hardware resources and prevents inaccurate visualization. The combination of the presented techniques empowers the improved human-machine interface of virtual reality to support every interactive task in medical three-dimensional (3-D) image processing, from visualization of unsegmented data volumes up to the simulation of surgical procedures.

Hybrid rendering of multidimensional image data

The most important rendering methods applied in medical imaging are surface and volume rendering techniques. Each approach has its own advantages and limitations: Fast surface-oriented methods are able to support real-time interaction and manipulation. The underlying representation, however, is dependent on intensive image processing to extract the object surfaces. In contrast, volume visualization is not necessarily based on extensive image processing and interpretation. No data reduction to geometric primitives, such as polygons, is required. Therefore, the process of volume rendering is currently not operating in real time. In order to provide the radiological diagnosis with additional information as well as to enable simulation and preoperative treatment planning we developed a new hybrid rendering method which combines the advantages of surface and volume presentation, and minimizes the limitations of these approaches. We developed a common data representation method for both techniques. A preprocessing module enables the construction of a data volume by interpolation as well as the calculation of object surfaces by semiautomatic image interpretation and surface construction. The hybrid rendering system is based on transparency and texture mapping features. It is embedded in a user-friendly open system which enables the support of new application fields such as virtual reality and stereolithography. The efficiency of our new method is described for 3-D subtraction angiography and the visualization of morpho-functional relationships.

Determination of knee joint cartilage thickness using three-dimensional magnetic resonance chondro-crassometry (3D MR-CCM)

The objective of this article was to analyze the accuracy and precision with which the quantitative distribution of articular cartilage can be determined in the knee joint using MRI. A three-dimensional (3D) technique that accounts for the out-of-plane deviation of the interface normal in strongly curved joint surfaces (3D MR-CCM) has been developed for cartilage thickness measurements. Eight cadaveric knee-joint specimens and six volunteers were imaged using a fat-suppressed gradient-echo sequence at a resolution of 2 x 0.31 x 0.31 mm3. Cartilage volumes and topographical thickness maps were obtained and compared with those derived from anatomical sections by image analysis. The deviation of the MR volumes from those of the sections was 1-12%, the coefficient of variation after repositioning ranged from 2.9% (patella) to 8.2% (lateral tibial plateau). Between 60% and 80% of all image points could be attributed to identical thickness intervals, less than 20% deviating by more than 0.5 mm. The intraobserver and interobserver reproducibilities were very high in both the specimens and the volunteers. In the knee joint, 3D reconstructions of the cartilages, and measurements that take into account the out-of-plane deviation of the interface normals (3D MR-CCM), are required.

[A method of computer-assisted, 3-dimensional subtraction angiography using spiral roentgen computerized tomography]

The objective of this study was to develop a method for 3D subtraction CT angiography and to optimize the visualization after semi-automatic segmentation. Ten patients with aneurysms of the abdominal aorta were examined using spiral CT. To reconstruct the vessels, as well as adjacent organs such as the liver and kidneys, one image data volume was acquired before and after injection of the contrast agent. The CT scans were obtained with a Siemens Somatom Plus 4. To improve the results of automatic segmentation, as well as visualization by maximum intensity projection (i.e. removal of bony structures), subtraction of both image volumes is necessary. However, small translation shifts disturb the subtraction process and produce artificial contours. To calculate the disparities along the three coordinate axes of two corresponding image volumes, a cepstrum filter is applied to a pair of image volumes. After detection of the disparities, which manifest as bright spots, the real shift of the two subsignals can be calculated. Translation of the corresponding image volume pairs to their correct positions improves the subtraction process. In all cases the size of the aneurysm and the abdominal organs could be better segmented and visualized. Application of the cepstrum filter and subtraction of the image volumes before and after contrast medium injection completely removes the bony structures in the image data and results in superior visualization results.

Image analysis and synthesis of multimodal images in medicine

Radiologic and clinical practice can be enhanced by improved access to multimodal image information. Analysis, visualization, method characteristic image processing and image synthesis is needed not only for the interpretation of the images but also for performing effective consultations with clinical colleagues and computer supported therapy planning and control strategies. A method is presented which enables the fast display, three-dimensional visualization and the modality oriented analysis of multimodal image information. Based on a unique image format, modality specific procedures and two- or three-dimensional processing tools of image analysis produce the input data for therapy planning programs. The easy use of this multimedia visualization tool enables radiologists and clinicians to deal with their image data. The description of methods and procedures, as well as typical examples of radiologic practice will demonstrate the efficiency of the presented system.