Design and Realization of Fluorescent Images Analysis Algorithms for Noninvasive Tumor Diagnosis

Abstract

The diagnosis of malignant tissues with fluorescent probes excited by visible or even infrared radiation has been fostered by the concurrent use of fluorescent substances and laser sources in photodynamic therapy of tumors. Various techniques have so far been proposed to implement diagnostic systems for malignant tissues, especially for endoscopy.^1,2 One of the most promising among them in terms of selectivity is the so-called Time Gated fluorescence Imaging (TGI) technique which has been recently proposed by the authors.^3,4 The approach followed relies on the detection of an exogenous fluorophore that localizes in malignant tissues selectively with respect to surrounding healthy tissues. The fluorophore of election for the experiments is Hematoporphyrin Derivative (HpD), a photosensitizer extensively used in the clinical practice. For diagnostic purposes, HpD can be profitably excited at a wavelength around 400 nm. It emits a red fluorescence (615–630 nm), having a lifetime of about 15 ns. The tissue auto fluorescence, which is also excited by the blue light, gives a background fluorescence that mainly extinguishes within 3–5 ns. The difference in the decay time of the two components can be used to suppress the auto fluorescence background. To this purpose the fluorescence of the suspected portion of tissue is acquired after a delay greater than 15 ns with respect to the excitation. Assuming a selective localization of HpD in the tumor, the image of the long living fluorescence maps the extension of the neoplasia.