OGC O09 The clinical utilisation of tissue imaging Mass Spectrometry as a diagnostic model for Oesophageal Adenocarcinoma

Abstract Background The prevalence of Oesophageal adenocarcinoma (OA) in the United Kingdom has steadily risen since the 1970's. The majority of patients present with advanced disease. Early diagnosis utilising point of care testing may enable an increase in the number of patients managed on a potentially curative pathway. Tissue Imaging Mass Spectrometry by Desorption Electrospray (DESI-MS) is one such tool to differentiate the abundance of lipids in OA. This work presents the novel use of Mass Spectrometry (MS) imaging to identify variations in lipid abundance between OA and normal oesophagus (NO). Methods Tissue was sampled from specimens dissected immediately post-resection from the tumour and at 5-centimetre intervals in the proximal NO. Forty-one unique patient OA and NO tissue samples were analysed in triplicate by DESI-MS tissue imaging to differentiate Glycerophospholipids. A total of 246 DESI-MS images were correlated with paired formalin fixed Haematoxylin and Eosin stained tissue slides, the current gold standard cancer diagnostic test. Principle component analysis utilising a supervised model classified tissue types with a significance of p=0.01. Receiver operator curves (ROC) were used to determine the diagnostic sensitivity and specificity of DESI-MS for the identification of adenocarcinoma. Results Tissue analysis confirmed unsaturated Phosphatidylglycerols and Phosphatidic acids were significantly more abundant in OA compared to NO (p<0.001). Tissue analysis by a Partial Least Square (PLS) supervised linear regression model classified cancer and non-cancer tissue with a confusion matrix classification of 93.3% and 70.8% for OA and NO (Figure 1a-c). A diagnostic sensitivity of 85% and specificity of 88% was determined by ROC curve. The abundance of Phosphatidylglycerol and Phosphatidylserine decreased proportionally in analysed tissue sections at 5cm and 10cm from the primary tumour, as did the degree of lipid unsaturation. Conclusions This work presents a novel tissue imaging MS model to differentiate lipid abundance in OA. An abundance ratio of 1.17:1 (OA:NO) of unsaturated Phosphatidylglycerol was diagnostic of cancer by DESI-MS with a high sensitivity and specificity. This MS tissue analysis model has the potential to enable accurate point of care diagnostics for OA. In addition, we suggest oesophageal cancer induced perturbation of Glycerophospholipid metabolism may be a potential target for future chemotherapeutics.