Validation of a perturbative approach for the analysis of data obtained from diffuse optical monitoring of neoadjuvant chemotherapy in breast cancer patients: From simulations to tissue phantom tests.

A multi-wavelength optical mammograph that works under time domain diffuse optical spectroscopy, developed by our research group, is applied to monitor the breast cancer patients' response to neoadjuvant chemotherapy. So far, a simple homogeneous data model approach to estimate tissue composition averaged along compressed breast thickness proved sensitive to therapy induced changes. The application of a perturbative model, describing the lesion as a perturbation in healthy tissue, would allow a quantitative characterization during therapy. This study validates a perturbative fitting procedure through a systematic approach starting from simple to complex scenario: first on simulations, then phantoms ("stand-in" for human tissue). This kind of systematic assessment helps us to estimate the errors in the reconstructed parameters when there are uncertainties in the input, when applied to a real case scenario like that of $in vivo$ studies. Hence, by varying input parameters like perturbation volume, geometry, position and signal level, one at a time, their effect and the overall model efficiency are studied. The final aim would be to analyse the $in vivo$ measurements' data.