Design and Characterization of a Rman Based Online Sensor for Analysis of Tissue

This preliminary study demonstrated that Raman spectroscopy could be used effectively to differentiate between different types of tissue. We have accomplished work in the development of ex vivo Raman spectroscopic system, which includes a hardware instrumentation and software implementation.

The hardware instrumentations consisted of a tunable Lithrow-configuration diode laser emitting at 785 nm, high efficiency spectrometer (QE 65000), lenses and mirrors, and a combination of other optical components. Our Raman setup was successfully used to excite and collect signal from different types of pig tissues as fat, nerve, skin, and muscle. In contrast, the software contained a three main VIs for computerized motor control, acquired spectra from QE65000 spectrometer, and to do shift excitation difference spectroscopy. Our customized software was also successfully used for the control of the laser and the specotrimeter, and to collect Raman spectra.

It was found that the optimal set of parameters which can give the bet results for Raman signal detection were excitation wavelength 785 nm, for incident laser power of P = 200 mW at irradiation time t=0, and acquisition time = 2s.

Raman spectra were analyzed by shifted excitation Raman difference spectroscopy (SERDS) method and pllynomial fit method. SERDS method is based on the difference of the two spectra obtained with different excitation wavelenghts. Our SERDS spectra showed significant difference between different tissue types by simple visual inspection except for the fat/nerve tissue. A simple algorithm was used to reconstruct Raman spectra from its SERDS derivatives spectra.

Fluorescence rejection from Raman spectra by means of polynoial fit method is based on the principle that plynomial fit to the background tiessue fulorescence signal, and then a polynomial subtraction from the measured signal. A successful background removal from different types of tissue has been reported. These results showed significant diffierences between different tissue types by simple visual inspection.

The obtained results showed a good agreement between SERDS method and polynomial fit mehtod especially in fat and nerve tissues. However, it is clear that there are significant differences between the two methods, which are related to peak shapes and peak height. This confirmed what has been described previously, that it is not necessary to reconstruct the Raman spectra form its SERDS derivatives, and it is better to avoid further spectral treatment that may distort band intensity irregularly. SERDS method showed a better removal fulorescence background while the plynomial fit method showed a high noise level especially in skin and muscle tissues. Therefor, in order to compare which of these two methods can give better resutls, we must compare between SERDS derivative spectra and reconstructed Raman spectra from polynomial fit method. This consequently requires a more effective method than a simple visual inspection, which can be achieved by applying the statistical analysis (principal componentn analyis).

For future work, the effect of varying the laser line shift Δλ needs to be studied in more detail. Trying to reduce excitation laser power and signal collection time is an imoortant area for future development. Further development of the instrumentation and software needs to be carried out for real time clinical application. Instrumentation dvelopoment includes the use of optical fiber to excite the samples and collect the signal, while the software development includes signal collection and data analysis.