An Infrared Absorbance Sensor for the Detection of Melanoma in Skin Biopsies

Abstract

An infrared (IR) absorbance sensor has been designed, realized and tested with the aim of detecting malignant melanomas in human skin biopsies. The sensor has been designed to obtain fast measurements (80 s) of a biopsy using a small light spot (0.5 mm in diameter, typically five to 10 times smaller than the biopsy size) to investigate different biopsy areas. The sensor has been equipped with a monochromator to record the whole IR spectrum in the 3330-3570 nm wavelength range (where methylene and methyl stretching vibrations occur) for a qualitative spectral investigation. From the collected spectra, the CH₂ stretch ratio values (ratio of the absorption intensities of the symmetric to asymmetric CH₂ stretching peaks) are determined and studied as a cancer indicator. Melanoma areas exhibit different spectral shapes and significantly higher CH₂ stretch ratios when compared to healthy skin. The results of the infrared investigation are compared with standard histology. This study shows that the IR sensor is a promising supportive tool to improve the diagnosis of melanoma during histopathological analysis, decreasing the risk of misdiagnosis.

Keywords: absorbance spectroscopy; infrared sensor; melanoma; skin biopsy.

Figure A1

Flowchart of the MATLAB script structure. The flowchart is organized in two processes. The main process sets the user defined parameters, initializes the DAQ unit, performs the measurements, determines the IR absorbance spectrum and computes the CH₂ stretch ratio. The second process represents the measurement process driving the step motor of the monochromator, evaluating the pulse amplitude in voltage for each wavelength and for the user defined number of repetitions.

Figure 1

Typical infrared absorption spectrum of healthy skin tissue in the 3330–3570 nm wavelength range (2800–3003 cm⁻¹) acquired using a Bruker Hyperion 3000 IR microscope coupled to a Tensor 37 spectrometer. The spectrum was recorded at 4 cm⁻¹ resolution for a total of four scans. Methyl (CH₃) and methylene (CH₂) symmetric (ν_s) and asymmetric (ν_as) stretching modes are indicated. The CH₂ stretch ratio is computed from the ratio between the CH₂ symmetric peak at 3505 nm and the CH₂ asymmetric peak at 3420 nm after straight baseline subtraction.

Figure 2

(a) Schematic of the overall IR instrument setup. The system includes a thermal emitter as infrared light source, a monochromator for light filtering, an optical lens system to guide the light, and a photodiode for signal detection. The light beam is collimated and focused into the entrance of the monochromator by plano-convex CaF₂ lenses (f = 75 mm). After the monochromator, the light is collimated by a Si plano-convex lens (f = 25 mm) and vertically directed by an IR mirror. A second Si plano-convex lens focuses the light beam into the sample holder where the specimen is placed for IR analysis. The light transmitted by the sample is collected by a third Si plano-convex lens and focused onto the active area of the photodiode; (b) Photo of the IR sensor system.

Figure 3

H&E stained pictures of healthy skin tissue, nodular malignant melanoma, superficial spreading malignant melanoma and melanoma metastasis investigated by the IR sensor. Infrared spectral information and CH₂ stretch ratio values have been obtained from three different biopsy spots. The plotted infrared spectra originate from a sample area of ~0.5 mm in diameter. Spectral shapes of healthy and cancerous tissue exhibit significant differences. The CH₂ stretch ratio values of malignant melanomas and melanoma metastasis are increased compared to the healthy tissue of reference.