A MULTI-DISEASE STATISTICAL ANALYSIS OF FREE FATTY ACID SIGNATURES ASSESSED BY GAS CHROMATOGRAPHY-MASS SPECTROMETRY

Fatty acids are classified based on chain length into short-chain fatty acids (SCFAs), mediumchain fatty acids (MCFAs), and long-chain fatty acids (LCFAs), which are essential for various physiological processes, including gut health. The gut microbiota plays a significant role in SCFA production, and its imbalance (dysbiosis) is often linked to alterations in fatty acid profiles. On the other hand, both MCFAs and are generally encountered in the diet, especially from milk and dairy products. Overall, these metabolites exert various essential function for host health, such as the regulation of energy metabolism, gene expression, ion channels and pump activities, membrane trafficking, and the modulation of immune processes. In this study, through a dedicated gas chromatography-mass spectrometry protocol, we evaluated the serum of SCFAs, MCFAs a LCFAs abundances in 54 patients with adenomatous polyps, 55 colorectal cancer patients, 54 patients affected by celiac disease, 67 patients affected by Sars-Cov 2, 100 amiotrophic lateral sclerosis patietns and 100 healthy controls, in order to explore potential disease clusters based on fatty acid distribution. Multinomial logistic regression effectively distinguished healthy individuals, celiac disease patients, and those affected by COVID-19 and ALS, but faced challenges in differentiating adenoma from colorectal cancer patients. Fisher’s linear discriminant analysis (LDA) performed poorly, with an accuracy deficit in classifying ALS patients, who were misidentified as healthy. Additionally, LDA struggled with distinguishing between adenoma and colon cancer patients. Quadratic discriminant analysis (QDA) showed an improved ability to classify healthy controls and achieved better overall classification accuracy compared to LDA. However, QDA faced difficulties in differentiating adenoma, colon cancer, and cardiovascular disease patients, which were not confused in other models. Despite the varying performance of the models, one consistent finding was the difficulty in distinguishing adenoma from colorectal cancer. This similarity was attributed to the fact that adenoma is considered a precursor to colon cancer, and patients with adenomas exhibit fatty acid levels closely resembling those of colon cancer patients. Overall, the results suggest that while fatty acid profiling can differentiate several disease states, adenoma and colon cancer share substantial overlap in their fatty acid signatures, reflecting the pathological continuum between the two conditions. This study highlights the potential of fatty acid profiling as a diagnostic tool, while also illustrating the challenges in distinguishing conditions with overlapping metabolic features. Further research into the role of gut microbiota and fatty acid metabolism in these diseases may offer valuable insights for early diagnosis and targeted interventions.