Modeling transcutaneous Vis-NIR spectroscopy of metacarpal bone

Abstract

The ability to perform routine monitoring of bone quality is crucial for patients with bone diseases such as osteoporosis. Current assessments of bone quality are expensive and cannot be used regularly without exposing patients to ionizing radiation. Alternatively, visible-near infrared (Vis-NIR) spectroscopy is a non-invasive, non-ionizing technique that can be used to assess the compositional properties of bone. Recently, studies have reported agreement between transcutaneous Vis-NIR spectroscopic measures of bone quality and conventional radiographic measures collected from the second metacarpal bone of the hand. Computational simulations using Monte-Carlo (MC) modeling offer a valuable tool to better understand the relative contributions from the underlying bone in comparison with the superficial skin, as well as to investigate the relative benefits of specific fiberoptic illumination/collection geometries for transcutaneous measurement of metacarpal bone. To inform the model, skin from above the 2 nd metacarpal bone and the bone itself were dissected from human cadaver hands. Reflectance and transmittance measurements of the skin and bone tissues were taken using an integrating sphere setup in the range of 400 nm-1800 nm. Optical properties were estimated using the Inverse Adding Doubling (IAD) technique. MC models of skin-bone tissues were created using these estimated optical properties as well as physical measurements of tissue thickness, and simulations of fiber-optic Vis-NIR measurements were performed. Results indicate up to 30% of the absorbance signal arises from contributions from the bone in specific spectral ranges.