The aromatic profile of microbiologically contaminated canned tomatoes was analyzed by the dynamic headspace extraction technique coupled with gas chromatography-mass spectrometry. Canned tomatoes contaminated with Escherichia coli, Saccharomyces cerevisiae and Aspergillus carbonarius were analyzed after 2 and 7 days. About 100 volatiles were detected, among which alcohols, aldehydes and ketones were the most abundant compounds. Gas chromatographic peak areas were used for statistical purposes. First, principal component analysis was carried out in order to visualize data trends and clusters. Then, linear discriminant analysis was performed in order to detect the set of volatile compounds ables to differentiate groups of analyzed samples. Five volatile compounds, i.e. ethanol, beta-myrcene, o-methyl styrene, 6-methyl-5-hepten-2-ol and 1-octanol, were found to be able to better discriminate between uncontaminated and contaminated samples. Prediction ability of the calculated model was estimated to be 100% by the "leave-one-out" cross-validation. An electronic nose device was then used to analyze the same contaminated and not contaminated canned tomato samples. Preliminary results were compared with those obtained by dynamic headspace gas chromatography-mass spectrometry, showing a good agreement.