1 E. RAPOSIO et al.: QUALI-QUANTITATIVE ANALYSIS I.J. of SIMULATION Vol. 9 No 1 ISSN 1473-804x online, 1473-8031 print QUALI-QUANTITATIVE ANALYSIS IN THE SPECTROPHOTOMETRIC DIAGNOSIS OF CUTANEOUS MELANOMA EDOARDO RAPOSIO, WALTER PATERNICH, GIACOMO ROBELLO, CLAUDIA PORZIO, ILARIA BALDELLI, MAURO TRUINI, PIERLUIGI SANTI. U.O. Chirurgia Plastica e Ricostruttiva, Università di Genova IST - Istituto Nazionale per la Ricerca sul Cancro Largo R. Benzi 10, 16132 Genova, Italy URL: http://www.istge.it/dip_toi/ch_pl/ch_pl.htm E-mail: patewa@virgilio.it CLAUDIO BOZZO, UGO TEDESCHI D'Appolonia S.p.A. Via S. Nazaro 19, 16145 Genova, Italy URL: http://www.dappolonia.it Abstract: This article presents a set of software tools which are used in conjunction with spectrophotometric technology to support the quali-quantitative analysis of pigmented lesions and their classification as either benign or malignant on the basis of their dominant spectral features. The software components interact with their hardware counterparts to acquire, reduce and analyse spectral images of the lesion directly from the patient’s skin. While the system is not intended to replace the specialist’s judgment in the diagnosis of melanoma, there is an indication that clearly negative cases can be discriminated against unclear cases requiring further investigation. Keywords: Spectrophotometry, Principal component analysis, Image processing, Melanoma, Diagnostic methods 1. INTRODUCTION Considering that early diagnosis represents the established ‘gold standard’ in the fight against melanoma, the Plastic Surgery Unit of the University of Genova has undertaken a research project named AISPEM, aimed at the spectrophotometric study of pigmented skin lesions. Specifically, this research project, started in December 2004, aims to develop, test, and validate a prototype system as an effective means to support doctors in the analysis of pigmented skin lesions. The system is intended to acquire and display images of such lesions within a spectral range of wave lengths defined between the visible (blu- violet) up to 1000 nm, and the near-infrared (NIR) starting at 400 nm, with a step of 5 nm. The aim of this article is to present the activities undertaken, the technology implemented, and the results obtained to date as part of the AISPEM project. During the course of this ongoing project a number of solutions have been considered and implemented to address research issues and opportunities as they arose, with the aim of providing a robust and comprehensive support system for the early diagnosis of melanoma, which can be effectively used in the daily medical practice. The activities concerning the hardware instrumentation have been devoted to the development and set-up of two prototypal devices capable of acquiring high spectral resolution images (Raposio et al., 2007). The first device, which could be considered as pre- prototypal, applies the spectral scanning technology to the acquisition of excised skin tissues. Although the ultimate goal of the project is to perform in vivo scans, this first device allowed for a preliminary collection of skin spectra, based on which suitable image processing algorithms could be developed. Meanwhile, the problem of in vivo scanning was addressed, resulting in a new prototype, capable of scanning and acquiring images directly from the patient’s body. Parallel work looked at the development of suitable software tools for data processing. In the development of such tools, the research team was faced with significant issues and difficulties linked to the complexity of the spectral acquisitions. However, referring to the study of acquisitions from excised tissues, taken as they became available in the early phases of the project, a joint development and investigative effort has led to the implementation of prototype software tools well- suited for the processing of such complex data. As a further objective, the research team worked on the expansion of the range of informatics tools currently available to clinical staff, with the aim of better