Comparison of transflection and transmission FTIR imaging measurements performed on differentially fixed tissue sections

Abstract

The widespread and cost-effective use of transflection substrates in Fourier transform infrared (FTIR) imaging of clinical samples is affected by the presence of artefacts including the electric field standing wave (EFSW) and contributions from light dispersion. For IR-based diagnostics, the manifestation of undesirable artifacts can distort the spectra and lead to erroneous diagnosis. Nevertheless, there is no clear consensus in the literature about the degree of influence of these effects. The aim of this work is to contribute to this discussion by comparing transflection and transmission images of the same tissue. For this purpose two adjacent sections of the same tissue (lymphoma sample) were fixed onto a CaF₂ window and a transflective slide for FTIR imaging. The samples in this case had a central area where based on morphology it was presumed the fixative did not penetrate to the same extent hence providing a comparable region for the two different substrates with a distinct physical/chemical difference. Transmission and transflection spectra from adjacent hyperspectral tissue images were combined in an extended dataset. Surprisingly, unsupervised hierarchical cluster analysis clustered together transflection and transmission spectra, being classified according to differences in tissue fixation instead of the geometry employed for the image acquisition. A more detailed examination of spectra from the peripheral zone of the tissue indicated that the main differences between the transflection and transmission spectra were: (1) a small shift of the amide I, (2) a larger “noise” component in the transflection spectra requiring more averaging to obtain representative spectra of tissue types, and (3) the phosphate bands were generally higher in absorbance in the transflection measurements compared to the transmission ones. The amide I shift and the larger spectral variance was consistent with results obtained in previous studies where the EFWS was present. The findings indicate that artifacts resulting from transflection measurements were small but consistent across the tissue, and therefore the use of transflection measurements could be employed for disease diagnosis. Accordingly, we recommend a straightforward multivariate comparison of images from transmission and transflection measurements in a combined data matrix obtained from adjacent sections of the tissue as a useful preliminary study for establishing the impact of the EFWS on the samples, before considering the routine use of transflection substrates for any new tissue studied.