Hyperspectral imaging and optical profilometry of murine tumor models

Cancer is one of the leading causes of death worldwide. There is a pressing need to diagnose cancer early and monitor it as the disease progresses or recedes. Tissue physiology and morphology are significantly different in tumors compared to healthy tissues, and these changes could be detected noninvasively using optical imaging techniques.

In this study, we combined hyperspectral imaging (HSI) with 3D optical profilometry (OP) to image multiple subcutaneously grown murine tumor models. Tissue properties such as melanin volume fraction and blood oxygenation were extracted from HSI data using the inverse adding-doubling (IAD) algorithm. Blood vessels were segmented from tissue oxygenation maps using the B-COSFIRE algorithm, and tumor volumes were calculated from 3D OP data and compared to manual measurements by a caliper.

We confirm that tumors can be discriminated from healthy tissue based on the differences in the measured reflectance spectra and extracted tissue parameters (p<0.05). The results also indicate that blood oxygenation is 50% higher in the blood vessels than in the surrounding tissue. Finally, we show that the tumor volumes estimated using OP deviate up to 26% from the manual measurements. Our study demonstrates that the combined HSI and OP imaging can successfully monitor the evolution of tumors in murine models.