ABSTRACT
Photoacoustic imaging (PAI) is a relatively new imaging modality that blends the strong optical absorption contrast of biological tissue with the high spatial resolution capabilities of ultrasound. e high contrast provided by endogenous chromophores alone has allowed PAI to image anatomy (Maslov et al. 2008), brain structure (Wang et al. 2003), functional organization of the cerebral cortex (Wang et al. 2003; Zhang et al. 2006; Liao, Li et al. 2010), and to detect the breast cancer in humans (Oraevsky et al. 2001; Manohar et al. 2007) and melanoma cells in rats (Zhang et al. 2006). With the advent of bioconjugated, tunable optical contrast agents (e.g., gold nanoparticles or carbon nanotubes), molecular PAI is also possible (Copland et al. 2004; Portnov et al. 2005; De La Zerda et al. 2008; Li et al. 2008; Kim et al. 2009; Pan et al. 2009; Xiang et al. 2009; Kim et al. 2010; Ntziachristos and Razansky 2010). However, the ability to interpret the molecular or functional contrast (the photoacoustic images) depends on the reliability of the PAI absorption spectroscopy (Laufer et al. 2005). If the absorption spectrum of a chromophore is known, then it is possible to calculate the concentration. For example, if the optical absorption is dominated by hemoglobin, then the absorption coecient can be written as
µ λ ε λ ε λα( , ) ( )[ ] ( )[ ].x = +HbO HbRHbO HbR2 2 (16.1)
Traditional spectroscopy requires the measurements at multiple wavelengths to invert the above equation to arrive at the concentrations of oxy-and deoxyhemoglobin. Current
photoacoustic spectroscopic methods assume that the recovered initial pressure distributions (the PA data) reect accurately (at least to within a multiplicative constant) the absorption properties of the medium under study, but his may not always be a safe assumption. Without a light transport model, the uence is unknown, and the uence depends on the chromophore concentrations that one is attempting to recover in addition to depending on the absorption and scattering spectra of everything that the light has passed through to get to the point of interest (Cox et al. 2008, 2009). Because the relative photoacoustic (PA) signal is not an intrinsic property of a medium, but is instead a product of the optical absorption coecient (the quantity of interest) and the local light uence, spatial and spectral inhomogeneities in the uence may undermine the spectral interpretation of PA images.
