Abstract
<i>In vivo</i> identification of early-stage cartilage degradation could positively impact disease progression in osteoarthritis, but to date remains a challenge. The primary goal of this study was to develop an infrared fiber-optic probe (IFOP) chemometric method using partial least squares (PLS1) to objectively determine the degree of cartilage degradation. Arthritic human tibial plateaus (<i>N</i> = 61) were obtained during knee replacement surgery and analyzed by IFOP. IFOP data were collected from multiple regions of each specimen and the cartilage graded according to the Collins Visual Grading Scale of 0, 1, 2, or 3. These grades correspond to cartilage morphology that displayed normal, swelling or softening, superficially slight fibrillation, and deeper fibrillation or serious fibrillation, respectively. The model focused on detecting early cartilage degradation and therefore utilized data from grades 0, 1, and 2. The best PLS1 calibration utilized the spectral range 1733–984 cm<sup>−1</sup>, and independent validation of the model utilizing 206 spectra to create a model and 105 independent test spectra resulted in a correlation between the predicted and actual Collins grade of <i>R</i><sup>2</sup> = 0.8228 with a standard error of prediction of 0.258 with a PLS1 rank of 15 PLS factors. A preliminary PLS1 calibration that utilized a cross-validation technique to investigate the possibility of correlation with histological tissue grade (33 spectra from 18 tissues) resulted in <i>R</i><sup>2</sup> = 0.8408 using only eight PLS factors, a very encouraging outcome. Thus, the groundwork for use of IFOP-based chemometric determination of early cartilage degradation has been established.
PDF Article
More Like This
Cited By
You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
Contact your librarian or system administrator
or
Login to access Optica Member Subscription