Characterising Type I Collagen Architecture in Human Bone and Osteosarcoma with Second Harmonic Generation Microscopy
Creith, B. Y.; Harrison, J.; Johnson, P. B.; Oreffo, R. O. C.; Clarkin, C. E.; Mahajan, S.
Show abstract
Aberrant collagen matrix production is a hallmark of many cancers, including osteosarcoma (OS), the most common primary cancer of bone which remains poorly diagnosed. Here, the phenotyping potential of second harmonic generation (SHG) microscopy has been demonstrated through label-free imaging of pathologically altered collagen matrices within clinical biopsies. This work examined SHG imaging in quantifying OS-specific collagen (type I) signatures - focused on fibre length-based parameters - across clinical biopsies. We describe a novel SHG analysis workflow, following systematic optimisation of image resolution and field of view (FOV) size to enable robust quantification of collagen fibre-length metrics in human bone and OS. This included analyses of SHG images with a FOV spanning 350 {micro}m x 350 {micro}m to 1400 {micro}m x 1400 {micro}m, with the latter enabling whole-biopsy visualisation via stitching of up to 64 regions. In normal cortical (rib) bone biopsies, average fibre lengths did not significantly differ across increasing FOV size (26.44 {+/-} 1.75 {micro}m, 26.34 {+/-} 2.01 {micro}m, 27.25 {+/-} 2.47 {micro}m, 28.78 {+/-} 3.58 {micro}m). However, analysis of the top 50th percentile of collagen fibres revealed significant differences between the smallest FOVs; 350 {micro}m x 350 {micro}m and 466 {micro}m x 466 {micro}m, versus the largest 1400 {micro}m x 1400 {micro}m (p = 0.03 and p = 0.04 respectively), highlighting the requirement for larger FOVs for optimal characterisation of human bone collagen. Comparison of normal bone and OS (stage IIB) biopsies, the largest FOV revealed the largest differences, with OS exhibiting significantly reduced collagen fibre lengths (female OS: 23.18 {+/-} 0.56 {micro}m vs. bone: 29.39 {+/-} 2.17 {micro}m, p < 0.001; male OS: 24.39 {+/-} 1.32 {micro}m vs. bone: 27.87 {+/-} 1.05 {micro}m, p = 0.02) versus normal bone. Subsequent analyses across OS stages (IB - IVB) showed further evidence of pathological collagen signatures amplified with OS progression. These findings demonstrate that SHG microscopy enables robust, quantitative phenotyping of human bone and osteosarcoma collagen matrices, offering new potential for improved diagnosis and disease staging.
Matching journals
The top 9 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- The association between mineralised tissue formation and the mechanical local in vivo environment: Time-lapsed quantification of a mouse defect healing model 93%
- Unsupervised determination of lung tumor margin with widefield polarimetric second-harmonic generation microscopy 92%
- Method for in vivo assessment of cancer tissue inhomogeneity and accurate histology-like morphological segmentation based on Optical Coherence Elastography 92%
Similar papers in this journal
Similar papers in this journal
- The petrous bone contains high concentrations of osteocytes: one possible reason why ancient DNA is better preserved in this bone. 92%
- Dynamic imaging of myelin pathology in physiologically preserved human brain tissue using third harmonic generation microscopy 91%
- Effects of long-term in vivo micro-CT imaging on hallmarks of osteopenia and frailty in aging mice 91%
Similar papers in this journal
- Nanoscale prognosis of colorectal cancer metastasis from AFM image processing of histological sections 89%
- Prediction of radiation-induced hypothyroidism using radiomic data analysis does not show superiority over standard normal tissue complication models 89%
- Spatial omics imaging of fresh-frozen tissue and routine FFPE histopathology on a single cancer needle core biopsy: freezing device and multimodal workflow 88%
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.