At Pittcon 2008 (Mar. 2–7; New Orleans, LA), Rohit Bhargava from the University of Illinois (Urbana-Champaign) discussed his group’s work with Fourier-transform infrared (FTIR) spectroscopic imaging to automate the analysis and recognition of tissue structure and disease without the need for dyes, molecular probes, or human intervention. According to Bhargava, FTIR imaging is emerging as a preferred technique for high-throughput biomedical sampling applications, combining the molecular selectivity of spectroscopy with the spatial specificity of optical microscopy.
Working closely with researchers from the National Institutes of Health (Bethesda, MD), Bhargava and colleagues are developing a high-fidelity spectrometer that combines a focal-plane-array detector with a conventional interferometer to potentially increase the speed of data acquisition several-fold. He says they have achieved mathematical noise reduction that enables an approximately 50- to 100-fold higher rate of data collection and have also used optical modeling to elucidate and resolve various confounding optical effects. The team’s technique has been used to automate human cancer diagnoses and grading for different tissue types by applying it to routine archival tissue samples. According to Bhargava, well-defined tests of statistical significance can provide insight into the conditions required to enable the prediction of patient and disease outcomes.