A Fourier domain directional filtering method for analysis of collagen alignment in ligaments

Interstitial fluid flow induces myofibroblast differentiation and collagen alignment in vitro

The differentiation of fibroblasts to contractile myofibroblasts, which is characterized by de novo expression of alpha-smooth muscle actin (alpha-SMA), is crucial for wound healing and a hallmark of tissue scarring and fibrosis. These processes often follow inflammatory events, particularly in soft tissues such as skin, lung and liver. Although inflammatory cells and damaged epithelium can release transforming growth factor beta1 (TGF-beta1), which largely mediates myofibroblast differentiation, the biophysical environment of inflammation and tissue regeneration, namely increased interstitial flow owing to vessel hyperpermeability and/or angiogenesis, may also play a role. We demonstrate that low levels of interstitial (3D) flow induce fibroblast-to-myofibroblast differentiation as well as collagen alignment and fibroblast proliferation, all in the absence of exogenous mediators. These effects were associated with TGF-beta1 induction, and could be eliminated with TGF-beta1 blocking antibodies. Furthermore, alpha1beta1 integrin was seen to play an important role in the specific response to flow, as its inhibition prevented fibroblast differentiation and subsequent collagen alignment but did not block their ability to contract the gel in a separate floating gel assay. This study suggests that the biophysical environment that often precedes fibrosis, such as swelling, increased microvascular permeability and increased lymphatic drainage--all which involve interstitial fluid flow--may itself play an important role in fibrogenesis.

The framework of pathology: good laboratory practice by quantitative and molecular methods

Combined confocal laser scan microscopy (CLSM) and Fourier analysis (FA) by non-pathologists of dermal collagen bundle orientation recently gave results superior to subjective evaluation by experts. According to Good Laboratory Practice (GLP) criteria, combined CLSM/FA has not yet been adequately tested to replace current collagen evaluation, but this will not take long. Non-pathologists (clinicians) will then have taken over a laboratory test historically belonging to pathology. A general trend in this direction may develop, because pathologists seem not always to care enough about clinical significance, reproducibility and prognostic value, and new demands for innovative methods. Quantitative image analysis (QIA) and molecular methods are reproducible, inexpensive, and easy to perform; they often have greater value than classical evaluations and their cost-benefit ratio is good. However, their acceptance is not as widespread as one would expect and theoretical reasons which have been advanced do not provide a satisfactory explanation. A formal implementation study was therefore performed, in which an attempt was made to modernize a classical pathology laboratory. An external customer satisfaction investigation showed that 96% of the clinicians were 'very satisfied' (the highest rating possible) with the completed innovations, contrasting with low satisfaction at the beginning. Lack of primary innovative leadership among pathologists was judged to be the dominant cause preventing implementation. Pathologists should focus on carefully reacting to new clinical needs, using GLP criteria. Reproducibility and predictive accuracy should be major themes in any pathology practice.

Morphometry of dermal collagen orientation by Fourier analysis is superior to multi-observer assessment

In human dermis, collagen bundle architecture appears randomly organized, whereas in pathological conditions, such as scar tissue and connective tissue disorders, collagen bundle architecture is arranged in a more parallel fashion. Histological examination by one or two observers using polarized light is the most common method to determine collagen orientation. The hypothesis on which this study is based is that an objective image analysis technique, Fourier analysis, would improve the reliability (are the measurements reproducible?) and the accuracy (does the method measure what it is supposed to measure?) of collagen orientation assessment, compared with observer ratings. Fourier analysis was applied to 271 images of scar tissue and normal skin that were acquired by confocal laser-scanning microscopy. Observers rated the same areas using polarized light as well as the confocal microscopy images. Computer images consisting of different types of ellipses were generated with a fixed orientation. Observers and Fourier analysis evaluated the images to evaluate accuracy. The inter-observer reliability was acceptable when at least three observers rated polarized light images (r > 0.69), whereas two observers were sufficient for rating confocal microscopy images (r > 0.71). Fourier analysis correlated better with observer ratings of confocal microscopy images (r = 0.69) than with polarized light microscopy images (r = 0.42). Fourier analysis was more accurate than four observers for the evaluation of the 'true' orientation for almost all types of computer-generated images. For the first time it is shown that Fourier image analysis is suitable for the morphometry of dermal collagen orientation and leads to a superior measurement of collagen orientation compared with subjective histological evaluation by several experts. If an evaluation is performed by conventional light microscopy, at least three observers are required to attain an acceptable inter-observer reliability.

Microstructural morphology in the transition region between scar and intact residual segments of a healing rat medial collateral ligament

This study used a rat model to investigate the microstructural organization of collagen through the transition from scar to intact residual segments of a healing medial collateral ligament (MCL). Twenty-two male retired breeder Sprague-Dawley rats were randomly separated into two groups. Eleven underwent surgical transections of both MCLs and were allowed unrestricted cage activity until euthanized two weeks post surgery. The remaining eleven rats were used as normal controls. All 44 MCLs were harvested including intact femoral and tibial insertions and prepared for scanning electron microscopy (SEM) imaging. At harvest the scar region in the healing ligaments was more translucent than the normal tissue. Ligaments were viewed from femoral to tibial insertions at magnifications of 100X through 20,000X. Tissue away from the scar region in the transected MCLs was indistinguishable from normal tissue in uninjured ligaments. Collagen fibers and fibrils in these tissues were more aligned along the longitudinal axis of the ligament than in the scar tissue. Continuity of collagen fibers and fibrils were consistently observed from the residual portions of the transected ligament through the scar region. Bifurcations/fusions, but no anastomoses, in fibers and fibrils were observed in both normal and scar tissues of ligaments. Qualitatively, bifurcations were encountered more frequently in scar tissue. In the transition region, larger diameter fibers from the residual tissue bifurcated into smaller diameter fibrils in the scar. This connection between larger diameter and smaller diameter fibers and fibrils indicates that bifurcations/fusions are likely to be the dominant way in which force is transmitted from a region with larger fibrils (residual ligament) into and through a region with smaller fibrils (scar).

Analysis of asymmetry in mammograms via directional filtering with Gabor wavelets

This paper presents a procedure for the analysis of left-right (bilateral) asymmetry in mammograms. The procedure is based upon the detection of linear directional components by using a multiresolution representation based upon Gabor wavelets. A particular wavelet scheme with two-dimensional Gabor filters as elementary functions with varying tuning frequency and orientation, specifically designed in order to reduce the redundancy in the wavelet-based representation, is applied to the given image. The filter responses for different scales and orientation are analyzed by using the Karhunen-Loève (KL) transform and Otsu's method of thresholding. The KL transform is applied to select the principal components of the filter responses, preserving only the most relevant directional elements appearing at all scales. The selected principal components, thresholded by using Otsu's method, are used to obtain the magnitude and phase of the directional components of the image. Rose diagrams computed from the phase images and statistical measures computed thereof are used for quantitative and qualitative analysis of the oriented patterns. A total of 80 images from 20 normal cases, 14 asymmetric cases, and six architectural distortion cases from the Mini-MIAS (Mammographic Image Analysis Society, London, U.K.) database were used to evaluate the scheme using the leave-one-out methodology. Average classification accuracy rates of up to 74.4% were achieved.

Measuring collagen fiber orientation: a two-dimensional quantitative macroscopic technique

This paper describes the design, evaluation, and application of a new system for quantifying two-dimensional collagen fiber orientation in soft tissue. Series of transmitted polarized light images were collected using a custom-designed macroscope. Combined analysis of pixel brightness, and hue from images collected with a compensator plate, permitted the assignment of each pixel into the appropriate orientation band. Experiments were performed to quantify the linearity and noise of the system. Validation was performed on a specimen composed of strain-birefringent plastic strips at various orientations. Preliminary collagen fiber orientation data is presented from a tendon specimen. This study demonstrates the utility of this approach for studying collagen fiber orientation across large areas.

Assessment of stress fiber orientation during healing of radial keratotomy wounds using confocal microscopy

Radial keratotomy (RK) is a refractive surgical procedure in which partial thickness incisions are made in the cornea in order to alter its shape. Previous studies suggest that RK wounds undergo changes in wound gape in response to the ingrowth of myofibroblasts which mediate subsequent wound contraction and may modulate changes in corneal curvature seen after RK. A recent quantitative analysis of f-actin organization in full-thickness incisional wounds in the rabbit demonstrated that microfilament bundles (stress fibers) present in myofibroblasts align parallel to the long axis of the wound during wound contraction. To investigate whether the same pattern of alignment occurs after RK, a similar analysis of f-actin organization was undertaken using the cat RK model. Radial keratotomy wounds were studied from 10 to 28 days after surgery using en block staining with fluorescein isothiocyanate (FITC) phalloidin, and three-dimensional (3-D) datasets (z-series of en face optical sections) were collected using laser confocal microscopy at various regions within the wound. In addition, conventional en face sections were double-labeled using combinations of phalloidin and antibodies to fibronectin and alpha 5 beta 1 integrin. Myofibroblast ingrowth started in the bottom of the wound and progressed anteriorly. At 10 to 14 days, f-actin was predominantly distributed in long, thick bundles (stress fibers) within the wound. These fibers appeared to be randomly oriented anteriorly, but became progressively more aligned with the long axis of the wound posteriorly. At 21 days, the stress fibers were predominantly oriented parallel to the long axis of the wound at all levels. F-actin, fibronectin and integrin were coaligned at both the 14 and 21 day time points. Since the majority of wound closure occurs between 14 and 28 days after surgery, we conclude that parallel alignment of the actin filament-fibronectin-integrin assembly in the cat RK model is associated with wound contraction.

Evaluation and progression analysis of pulmonary tuberculosis from digital chest radiographs

Pulmonary tuberculosis is one of the most common infections in the third world countries. We propose a method for automated evaluation and progression analysis of this disease by processing a sequence of digital chest radiographs. The presence of radiological features characteristic to pulmonary tuberculosis is detected within the regions of interest using suitable image processing and pattern recognition techniques. The detected features are then quantified to evaluate the extent of the disease. The measures are evaluated for a sequence of radiographs of a patient and are normalized with respect to rib size measures. Temporal changes in these normalized measures indicate the progression of the disease during the period in which the sequence of radiographs is obtained. The details of the proposed technique and results of analysis of a number of tuberculous patients are given in the paper.

A small angle light scattering device for planar connective tissue microstructural analysis

The planar fibrous connective tissues of the body are composed of a dense extracellular network of collagen and elastin fibers embedded in a ground matrix, and thus can be thought of as biocomposites. Thus, the quantification of fiber architecture is an important step in developing an understanding of the mechanics of planar tissues in health and disease. We have used small angle light scattering (SALS) to map the gross fiber orientation of several soft membrane connective tissues. However, the device and analysis methods used in these studies required extensive manual intervention and were unsuitable for large-scale fiber architectural mapping studies. We have developed an improved SALS device that allows for rapid data acquisition, automated high spatial resolution specimen positioning, and new analysis methods suitable for large-scale mapping studies. Extensive validation experiments revealed that the SALS device can accurately measure fiber orientation for up to a tissue thickness of at least 500 microns to an angular resolution of approximately 1 degree and a spatial resolution of +/-254 microns. To demonstrate the new device's capabilities, structural measurements from porcine aortic valve leaflets are presented. Results indicate that the new SALS device provides an accurate method for rapid quantification of the gross fiber structure of planar connective tissues.

Collagen fiber architecture of a cultured dermal tissue

Advances in tissue engineering have led to the development of artificially grown dermal tissues for use in burn and ulcer treatments. An example of such an engineered tissue is Dermagraft, which is grown using human neonatal fibroblasts on rectangular sheets of biodegradable mesh. Using small angle light scattering (SALS), we quantified the collagen fiber architecture of Dermagraft with the mesh scaffold contributions removed through the use of a structurally based optical model. Dermagraft collagen fibers were found to have a preferred direction nearly parallel to the long dimension of the kite-shaped mesh opening with small spatial variations over the mesh. This study demonstrated the utility of SALS as a rapid and inexpensive technique for the evaluation of gross collagen fiber architecture in engineered tissues.

Soft-tissue "flaws" are associated with the material properties of the healing rabbit medial collateral ligament

This study evaluated microscopic flaws in the healing rabbit medial collateral ligament and their significance in terms of the material properties of this ligament during healing. A gap injury was created in the midsubstance of the medial collateral ligament in the right hindlimb of 15 skeletally mature (12 months old) New Zealand White rabbits. At postoperative intervals of 3, 6, or 14 weeks, histomorphometric analysis of the flaws was carried out in subgroups of animals. The medial collateral ligaments from four of the left hindlimbs (randomly selected) were used as uninjured contralateral controls. In one histologic section of each area of scar tissue and the analogous area in the controls, specified tissue flaws (blood vessels, fat cells, hypercellular areas, loose matrix, disorganized matrix, or a combination of these) were measured by four independent and blinded observers. The results showed that the mean total area of the flaws, as a percentage of the total section, and the mean area of the largest flaw decreased with healing time in each healing group but did not achieve control values by 14 weeks. Because it was not possible to test the healing medial collateral ligaments mechanically prior to measurement of the flaws (due to the destructive nature of failure testing), the data on the flaws were compared with the material strength and stiffness of a separate series of similarly injured and mechanically tested medial collateral ligaments (data published previously). A maximum likelihood statistical analysis showed a very strong functional association between the mean area of the largest flaw and the stress at failure (p < 0.004) and between the mean flaw area as a percentage of the total section area and the elastic modulus (p < 0.001). This study therefore demonstrates that it is possible to quantify material flaws in scar tissue in rabbit medial collateral ligaments, that these flaws become smaller with healing time as the scar remodels, and that flaws are functionally associated with the material properties of the ligament in this model (larger flaws with less tensile strength and more flaws with less stiffness).

Quantitative analysis of the fine vascular anatomy of articular ligaments

An image analysis technique has been developed to quantitatively describe the fine vascular patterns observed in ligament tissue. The longitudinal orientational distribution and total vessel volume of India-ink-perfused blood vessel segments in normal and healing ligaments were determined. The methods involved special vascular preparation of adult rabbit knee medial collateral ligaments (MCL) by India-ink perfusion. Black and white microscope images of ink-perfused tissue sections were subjected to a thresholding procedure to binarize digitized ligament images, which were then skeletonized and analyzed for directional distribution based on the least-squares technique. Analysis of medial collateral ligaments in New Zealand White rabbits using this method has shown that scarred tissue is more vascular and has a more chaotic angular distribution of blood-vessel segments than normal ligament tissue.

Statistical analysis of collagen alignment in ligaments by scale-space analysis

Injuries to ligaments in the knee are common in sports and other physical activities. Some clinical methods are available for qualitatively evaluating the degree of ligament injury and healing. It is, however, desirable to objectively assess the healing of ligaments and to predicate optimal treatment on quantitative measurements of their structure. Information such as areas of coverage and spatial orientations of collagen fibrils, for example, may provide important information about the internal structure of ligament tissues. Since normal ligament tissues are made up of collagen fibrils which are highly organized, they can be considered as oriented piecewise linear patterns. In this paper, we propose a computational technique for statistical analysis of collagen alignment in ligament images using the scale-space approach. In this method, a ligament image is preprocessed by a sequence of filters which are second derivatives of two-dimensional Gaussian functions with different scales. This gives a set of zero-crossing maps (the scale space) from which a stability map is generated. Significant linear patterns are captured by analyzing the stability map. The directional information in terms of orientation distributions of the collagen fibrils in the image and the area covered by the fibrils in specific directions are extracted for statistical analysis. Examples illustrating the performance of this method with scanning electron microscope images of the collagen fibrils in healing rabbit medial collateral ligaments are presented in this paper.

Scale space approach to directional analysis of images

In this paper a new technique for directional analysis of linear patterns in images is proposed based on the notion of scale space. A given image is preprocessed by a sequence of filters which are second derivatives of 2-D Gaussian functions with different scales. This gives a set of zero crossing maps (the scale space) from which a stability map is generated. Significant linear patterns are detected from measurements on the stability map. Information regarding orientation of the linear patterns in the image and the area covered by the patterns in specific directions is then computed. The performance of the method is illustrated through applications to synthetic patters and to scanning electron microscope images of collagen fibrils in rabbit ligaments.

A quantitative analysis of matrix alignment in ligament scars: a comparison of movement versus immobilization in an immature rabbit model

This investigation quantified the alignment of fibrillar matrix in normal rabbit medial collateral ligaments (MCLs) and in healing MCLs from animals treated with or without knee immobilization. Twenty-four immature female rabbits were given complete midsubstance injuries to their right MCLs. Fifteen of them had that knee pin immobilized in flexion, while the remaining nine were allowed unrestricted cage activity. Animals were sacrificed in groups of three at intervals of 3, 6, or 14 weeks after injury, and both healing MCLs and unoperated contralateral controls were fixed in situ for subsequent removal, freeze-fracture, and preparation for scanning electron microscopy (SEM). A random sampling of SEM photographs followed by automated, statistically validated image processing was used to quantify alignment of matrix in all samples. Results showed that nonimmobilized MCL scars in this model do remodel over 14 weeks of healing, returning to normal alignment values in that time. Surprisingly, MCL scars in immobilized knees were even better, with mean matrix alignments falling statistically within normal MCL limits at all healing intervals studied. If not due to an unknown sampling or fixation artifact, these results suggest that gross knee flexion and extension is not a prerequisite for scar matrix alignment in this immature model of ligament healing.

Detection of blood vessels in retinal images using two-dimensional matched filters

Blood vessels usually have poor local contrast, and the application of existing edge detection algorithms yield results which are not satisfactory. An operator for feature extraction based on the optical and spatial properties of objects to be recognized is introduced. The gray-level profile of the cross section of a blood vessel is approximated by a Gaussian-shaped curve. The concept of matched filter detection of signals is used to detect piecewise linear segments of blood vessels in these images. Twelve different templates that are used to search for vessel segments along all possible directions are constructed. Various issues related to the implementation of these matched filters are discussed. The results are compared to those obtained with other methods.