Using a semi-automatic layer detection algorithm to determine collagen fiber alignment in the human meniscus during loading

Abstract

Osteoarthritis (OA) is a chronic degenerative joint disease, linked to age, weight and joint overload. Typical symptoms include joint pain, loss of mobility and joint stiffness. The past two decades the amount of people suffering from OA has increased, and with the current trend of increasing age and weight it is expected to further increase. It is known that knee OA is correlated with the degeneration of the collagen structure in the meniscus, among other knee tissues. However, how the degeneration occurs inside the meniscus has not been extensively studied.


The meniscus is a crescent shaped disc of fibrocartilaginous tissue comprised mainly of water and collagen. The collagen is highly organized in layers, with fibers oriented circumferentially, radially or randomly depending on their location in the meniscus. This study aimed to investigate how the microstructural response of the meniscus changes during loading, by looking at the collagen fiber orientations during an axial stress-relaxation loading protocol. This was done by developing a layer detection algorithm, in combination with existing orientation quantification software. The orientations are determined in 3D as a combination of azimuth (longitudinal to the fiber planes) and elevation (in the fiber cross section direction) angles. The distribution of orientations in both azimuth and elevation directions were plotted as histograms. The peak of the histogram provide information about the main fiber orientation, and the full width half maximum was used to quantify the spread of fiber orientations (orientation heterogeneity).


The layer detection algorithm allowed to not only study the sample as a whole, but also quantify the behaviour of individual layers within the meniscus. The results showed that the stress is redistributed homogeneously throughout the meniscus, but the individual layers showed size changes up to 2.5\% of the sample height. Within the layers the results showed a few degrees shift in the peak angles for both azimuth and elevation angles during the two relaxation phases. The main differences were observed in the FWHM of the azimuth angles, which increased with 14-25\% during the second compression, indicating a decrease in orientation homogeneity. A general trend was that the major change occured after the second compression. The fiber density dropped during both compressions and gradually return to previous levels. The layer detection algorithm could correctly identify the boundaries between layers, but struggled to capture the fiber layers when they were tilted away from the horizontal plane.

In conclusion, the microstructural response to loading of the meniscus appears to be mainly related to azimuthal orientation changes, and fiber density. The drop in the fiber density could be interpreted as water being forced out of the extracellular matrix, and the gradual response be interpreted as the tissue reabsorbing the water, even under pressure. The layer detection algorithm was applied successfully on multiple samples of varying degree of tissue degeneration, and able to consistently capture the same layer between time points for all samples. The primary results from the individual layers suggest some differences in tissue response, depending on degree of degeneration.

Popular Abstract

Fiber alignment in the meniscus during loading.

The use of a custom written algorithm to analyze 3D x-ray images provides new insight into how the collagen fibers are arranged in human meniscus while loaded. The algorithm can identify layers of differently oriented fibers. It can be used for further research to study how damage to the meniscus, due to aging or osteoarthritis, affects the fiber orientations and its response to load.

The knee is the largest joint in the body, and it is the target of large forces daily. An important part of the knee is a pair of c-shaped discs made of collagen, known as the menisci. The c-shape allows the meniscus to alleviate stress on the underlying cartilage by spreading out the force placed on the meniscus along the c-shape. It is known that when the meniscus is damaged the collagen structure is affected, but it is not known exactly how.

Using 3D x-ray images, it could be seen that the orientations of the circumferential fibers are not homogeneous but structured in layers (See Figure 1). To properly investigate how the structure changes, these layers needed to be identified. For this purpose, a specific layer detection algorithm was written, and by analyzing the 3D x-ray images, we could gather a first insight into how the fiber orientations change during loading.

The algorithm consistently found the same collagen layers every time it was applied. The meniscus samples were subjected to vertical forces, and the fiber alignment was investigated within each collagen layer in both the horizontal and vertical plane. The major changes happened in the horizontal plane, as the previously aligned fibers became less aligned with increased pressure.
An interesting behaviour could be seen for the fiber density, as it decreased every time force was applied to the meniscus, to gradually increase back to its original value during relaxation.

The project serves two main purposes. It provides a first insight into dynamic fiber orientation change, a source of information regarding meniscus function and health. It also provides a stepping stone for further developments in developing a method that can be used to identify and extract information from collagen fiber layers.

• author: Abrahamsson, Martin
• supervisor: Maria Pierantoni; Hèctor Dejea Velardo; Hanna Isaksson
• organization: Department of Biomedical Engineering
• alternative title: Semi-automatisk algoritm för bildanalys av kollagenfiberstrukturen i human menisk under belastning
• course: BMEM01 20232
• year: 2023
• type: H2 - Master's Degree (Two Years)
• subject: Technology and Engineering
• language: English
• additional info: 2023-16
• id: 9141541
• date added to LUP: 2023-11-21 14:54:56
• date last changed: 2023-11-21 16:06:05