| Articular cartilage is hyaline cartilage, has the smooth surface, can reduce the friction between the adjacent two bones and buffers the vibration from the movement, It is pale blue and glossy and is a layer of special bright connective tissue collagen fiber structure overlaided on the epiphyseal end. When this structure subjected to pressure,it can also be a little deformation, alleviate the impact of the force transmission load and be the necessary to maintain joint activities of organizational structure. Therefore, before exercise it is a good idea to have the activities of the articulation, have the articulation been full of lubrication to prevent damaging the articular cartilage. But conversely, the movement of the articulation is necessary to play an important role for the maintenance of the normal articular cartilage structure. Because the cartilage cells scattered between the fibers maintains the normal metabolism of the articular cartilage, constituted by the elliptical or circular cells from the shallow to the deep gradually. Its nutrients and metabolic waste must be obtained or discharged from the synovial fluid and the nutrition metabolic pathway of the articular cartilage is the articular movement,to make the articular cartilage constantly afford pressure and stimulus. Osteoarthritis (Osteoarthritis, OA) is a very common age-related disease and second only to heart disease in the causing disability with the greatly impact of the socio-economic health care system. It will cause pain and show the rigid activities of loss elasticity on the articular movement, and if the pain and activity are reduced, the muscles and ligaments around the diseased articulation can be caused atrophy and lead to articulation deformity and dysfunction. The tear, rupture and thinning of the articular cartilage is the core area of causing the disease. Therefore, it is particularly important for early diagnosis and treatment of articular cartilage lesions.Currently, the OA treatment is mainly limited to the research of the control of symptoms and the treatment of the disease drug. Therefore, a lot of research on the analysis of articular cartilage and related disease progression has been invested. X-ray, CT and other methods can be used to evaluate the articular cartilage, but these methods can not be used in the early stages of cartilage lesions diagnosis, especially the change phase of biochemical composition before Morphological profile not been changed. With the rapid development of MR imaging technology, MR imaging can directly observe the changes of the structure of OA articular cartilage and have been increasingly used in the early stages of cartilage lesions diagnosis, especially the change phase of biochemical composition before Morphological profile not been changed for its multi-faceted, multi-sequence, multi-parameter and non-traumatic. T2(T2mapping) imaging as a new MR imaging technology has become a sensitive marker of articular cartilage lesions early diagnosis and condition monitoring.T2relaxation time (T2relaxation time) can quantify with T2mapping and is the time of decaying to37%of the maximum signal strength. The imaging principle: First, we use the multi-level and multi-echo spin-echo sequence and take the same TR time and different TE time for scanning. a range of T2-weighted images can be obtained; then calculate T2values of these images pixels, grayscale or color-order T2mapping quantitative analysis reconstructed.For the cartilage T2mapping segmentation, Rother provides a valid interactive segmentation algorithm based on imaging segmentation. The algorithm principle: Use the gray histogram to segment the region with covering a rectangular in the region to be segmented;2006, Tang proposed a semi-automatic gradient vector flow (Gradient the Vector Flow GVF) snake improved algorithm. The algorithm absorbed gradient direction information, modeled the cartilage boundary by B-spline model, achieved a more accurate cartilage boundary segmentation by the way to add control points; Folkesson and Dam proposed a voxel classification method by using kNN algorithm to separate the cartilage boundary and background in order to cartilage segmentation and later proposed an effective classification method for the computational complexity of the kNN algorithm. The principle is that a series of random sampled voxels were divided into the part of cartilage and other part of background pixels rather than each individual voxel was classified, It meant that if one voxel was classified as cartilage,continued to classify the adjacent pixels until no pixels were classified into cartilage.Although the above methods succeeded to extract the cartilage region, they did not achieve to divide automatically the cartilage region into different hierarchical or subregional areas. The structure of articular cartilage from the surface part to the deep part can be divided into four layers, respectively, including the superficial layer, transitional layer, radiation layer and calcified layer. Modl was the first to discover that the articular cartilage showed three layers on T2-weighted images and recognized the deep layer corresponding to the radiation layer and the calcified layer, the middle layer corresponding to the transitional layer and the superficial layer corresponding to the superficial layer. The structural characteristics of Layers:the collagen fibers in the superficial layer were thin and compact, parallelled to the articular surface with the highest water content; collagen fibers in the middle layer arranged irregularly, showed random oblique shape with water content in the second; the collagen fibers in the deep layer were thick, was perpendicular to the articular surface, the water content was at least, provided the maximum compressive force. A highly ordered arrangement of the collagen fibers of the articular cartilage load the structure anisotropy of the articular cartilage and the T2values of the articular cartilage were precisely depended on the structure anisotropy of cartilage tissue, collagen and its arrangement direction, the change of the water content. Thus, biochemical composition in one layer or one region of the articular cartilage had been changed, it meant that T2values of the layer or the area changed. When OA occured,The disappearance of the hierarchical structure of collagen is the morphological change of the earliest visible cartilage tissue. Therefore, the automatic achievement of articular cartilage layers was important significantly to analyze the signal strength of the different regions, investigate the function weight-bearing area of articular cartilage and early diagnosis of cartilage lesions. So we will achieve that the T2mapping of the knee articular cartilage can be automatically divided into three layers and nine regions.T2mapping of the article obtained by knee joint axial line scanned by Philip Intera1.5T superconducting MR SENSE-Flex-M coils with the scanning range from upper edge of the level on the patella to the tibial plateau, especially, put sandbags in the foot to fix the position of foot. T2mapping imaging scan parameters:8-echo SE sequence, TR1300ms, TE10ms-90ms, slice thickness3mm, layer spacing1.5mm, number of layers12, the field of view (FOV)180mm x140mm, pixel matrix512×512, acquisition time2, about7min scanning time. A total of15knees (left knee7, right knee8) were scanned, they were between23to42years old, the average age is28.6±5.3years, and all patients signed informed consent. A total of15sets of datas were obtained (T2mappings), then they were outlined by the clinical experience of the radiologist.The T2mapping of the knee cartilage outlined by a doctor would be become a binary image that the cartilage area was white and other regions were black. Then the edge of the cartilage region could be achieved by morphology function, which wasa closed area, that was to say, it was a contour map of the cartilage region. The central axis of the cartilage area was extracted by morphology function and burrs were removed.The horizontal and vertical coordinates of the central axis convoluted with the first-order differential Gaussian kernel, respectively. The slope of the tangent could be calculated, and then the slope of the normal line could be calculated by the slope of the tangent, the normal direction of the normal line were derived from the slope of the normal line, thereby each of the normal horizontal and vertical coordinates could be calculated.Each normal line and cartilage area edge line had two points of intersection, and then received two points which could averaged these two intersection’s distance, so that four points could be obtained, which were classified as A, B, C and D from top to bottom, then four dividing lines could be obtained by four points linear interpolation, the cartilage area could be divided into three layers including deep layer, middle layer and superficial layer. The minimum distance of the points from cartilage region to the four boundary line were calculated, and then determined two boundary lines in which this point was. and so forth, all the points of the cartilage region could be layered. Based on the cartilage area contour map divided into three layers two normals which could averaged the medial axis line could divided each layer into left, middle and right three regions, each region coordinate been put into the original knee cartilage T2mapping,then the T2mean value of region in the original knee cartilage T2mapping could be given. The trend of their spatial distribution was from deep layer of articular cartilage to superficial layer was first decreased and then increased, the closer to surface of the articular cartilage, the higher T2values were.Bland-Altman statistical analysis method was used to evaluate the consistency of measurement results between the two doctors and between manual and automatic method. The principle is that the mean and the standard deviation of difference of the two measurements were calculated, and then the95%limits of agreement could be calculated and the limits of agreement was reflected by graphical method, the limits of agreement was less than half past the minimum mean of two measurement results, it was determined that it had good consistency. Also we could evaluate the agreement of the two measurements by the coefficient of variation (Coefficient Of Variation, COV).Experiment measured results were that the mean of subregional measurement results of the doctor1and doctor2were39.06ms and39.71ms, respectly. The mean of difference was-0.66ms and the standard deviation of difference was2.6ms, the95%limits of agreement was-0.66±1.96×2.05ms, that was to say,-4.68ms to3.36ms and the coefficient of variation was COV=2.05ms/39.39ms=5.20%.Experiment measured results were that the mean of subregional measurement results of manually and automatic methods were39.06ms and39.71ms, respectly. The mean of difference was0.08ms and the standard deviation of difference was1.59ms, the95%limits of agreement was0.08±1.96×1.59ms, that was to say,-3.04ms to3.20ms. The limits of agreement was narrow (far less than half past the minimum mean of two sets of results mean,39.31ms) and the coefficient of variation was COV=1.59ms/39.35ms=4.04%.It showed that the consistency of the manual and automatical methods is good. within consistency limits, the maximum absolute value of the difference was3.20ms, the minimum average of doctors segmentation results is39.31ms, and the magnitude of difference was acceptable clinically,so the two methods had good consistency clinically.The automatic method could not only replace the manual method but also overcome the subjectivity of the manual method. In short, the proposed automatic method reduced the time of the knee articular cartilage subregion, improved efficiency, had better repeatability, would also contribute to analyze the signal strength of the different regions and cartilage tissue structure, investigate the function weight-bearing area of articular cartilage and early diagnosis of cartilage lesions. |
PDF Full Text Request