Evaluation Of Surgical Safety Margin Of Oral Squamous Cell Carcinoma Based On Metabolomics

Surgery is the main method for the treatment of oral squamous cell carcinoma(OSCC),and obtaining the surgical safety margin by removing the tumor completely at one time is the key factor for controlling the recurrence of OSCC.But it is still controversy about the guidline of evaluation standard for surgical safety margin of OSCC in the world.According to the NCCN guidelines,the safety distance of OSCC is more than 5mm between the tumor infiltration front and the surgical boudary.At present,the "gold standard" of surgical safety margin evaluation of OSCC is still based on frozen pathological diagnosis during the operation,and the diagnosis is made according to the presence of tumor cells or dysplastic cells under the microscope.But more and more research indicated that patients with pathological "negative margin" still had recurrence rate as high as 10-30%.These results indicated that the diagnosis of surgical margin based on histopathological changes has obvious defects,so,it is necessary to find a new method to evaluate the surgical margin status of OSCC.With the development of cell and molecular biology,it was found that the changes of genes,proteins and metabolites were earlier than cell histomorphology.These results suggest that we can use new techniques to detect the levels of cancer-related genes or metabolic molecules that have been altered in the surgical margin tissue,which combined with the traditional histopathological methods to diagnose margin status.So,the purpose of the study is to use the traditional gas chromatography-mass spectrometer(GC-MS)/ ultrahigh performance liquid chromatography-tandem mass spectrometry(UHPLC-MS/MS)and DESI-MSI ambient mass spectrometry to study the mucosal margin molecular markers of OSCC on amino acid metabolomics and lipid metabolomics,respectively.Part One Untargeted screening of amino acid metabolic markers in OSCC tumor and different surgical-margin status by GC-MS[Objective]In order to study the differential metabolites of different mucosal margin status of OSCC.[Methods]We analyze the metabolic characteristics of 32 matched tumor tissue and marginal tissue samples by GC-MS(including 8 OSCC cases).Each case contained tumor and at least 15 mm of continuous mucosal margin,and divided into 4 sections: tumor,margin-1(0-5mm,dysplastic marginal tissue),margin-2(5-10 mm,negative marginal tissue),and margin-3(10-15 mm,normal tissue).All 32 samples were divided into 4 groups: Tumor(T),margin-1(M1),margin-2(M2),and margin-3(M3).One-way anova,Tukey's test and student's t test were used for statistical analysis.[Results]A total of 208 metabolites with known names were obtained,which were mainly amino acids,nucleotides and carbohydrate.The metabolites between the four groups were analyzed by one-way anova and Tukey's test,16 differential metabolites were screened out,8 of which were amino acids.The use of 8 differential amino acids for heat map analysis could clearly distinguish T,M1,M2 and M3 groups.In order to further determine the amino acid molecular markers for the negative margin,16 amino acids were preliminarily screened by student's t test of T vs M2(p<0.05),and the ability of these amino acids to identify T vs M2 was detected by ROC curve.Finally,10 amino acids were determined as candidate markers of the negative surgical margin(p<0.05,AUC > 0.90).By the same method,9 amino acids were finally determined as candidate markers for the dysplastic margin(p<0.05,AUC > 0.90)by comparing M1 vs M2.[Conclusions]Using GC-MS non targeted high-throughput method,preliminary determine the 10 amino acids(lysine,proline and alanine,serine,threonine,leucine,histidine,valine,glutamic acid,aspartic acid)and 9 amino acids(alanine,proline,serine,aspartic acid,glutamic acid and glutamine,histidine,asparagine,ornithine)as negative surgical margin and dysplastic margin candidate metabolic markers,respectively.Part Two UHPLC-MS/MS targeted metabolomics quantitatively analysis of the amino acid metabolic markers of OSCC tumor and different surgical-margin status[Objective]UHPLC-MS/MS was used to conduct targeted quantitative analysis of the candidate amino acid markers determined by GC-MS untargeted analysis.[Methods]Another 80 matched tumor and mucosa margin tissue samples(validation group)were recruited in the study(including 20 OSCC cases).Each case also included 4 sections: tumor,margin-1,margin-2,and margin-3.Student's t test was used to test the differences of the two groups(T vs M2,M1 vs M2),and p<0.05 was considered statistically significant.Meanwhile,the significance of specific amino acid metabolic markers for the prognosis of OSCC patients was analyzed.[Results]There are nine of ten candidate amino acid markers were detected again and has statistical significance(p < 0.05)in validation group,four of these amino acids(aspartic acid,glutamic acid,proline,valine)can better distinguish T vs M2(AUC > 0.90).Moreover,using the above four amino acids for logistic regression analysis,obtain the best specificity and sensitivity for T vs M2(AUC=0.98).Similarly,There are 6 of 9 candidate amino acid markers were detected again and has statistical significance(p < 0.05)in validation group(AUC: 0.65-0.75),the above six amino acids has different discriminating abilities for M1 vs M2.However,the six aminoacids combinations for logistic regression analysis,obtain the best specificity and sensitivity for M1 vs M2(AUC=0.81).In order to verify the clinical significance of amino acid markers of surgical margin,aspartic acid and asparagine were selected,which were markers of dysplastic margin and related to carcinogenesis.We found that the levels of aspartic acid and asparagine gradually increased from normal tissue to tumor tissue.KEGG metabolic pathway analysis also revealed that there were active aspartic acid metabolic pathways in OSCC.Asparagine synthetase(ASNS),a key metabolic enzyme of aspartic acid metabolism which can catalyse aspartic acid to asparagine.We further examined the expression level of ASNS protein in dysplastic margin of OSCC by immunohistochemistry(IHC),the results showed that the expression of ASNS protein in the dysplastic margin increased with the worsening of the degree of dysplasia and was correlated with the poor prognosis of OSCC.[Conclusions]Four amino acids(aspartic acid,glutamic acid,proline and valine)were determined as markers of negative margin,and six amino acids(aspartic acid,asparagine,proline,glutamic acid,histidine and serine)were selected as markers of dysplastic margin.Asparagine synthase(ASNS)may be related to the development of oral squamous cell carcinoma,the specific mechanism of promoting carcinogenesis is worthy of further study.Part Three DESI-MSI visualized in situ determined lipid metabolism markers of surgic al-margin in oral squamous cell carcinoma[Objective]DESI-MSI technique was used to accurately determine the mucosal safety margin distance and margin status of OSCC at the level of lipid omics,and further determine lipid-markers of positive margin and negative margin on fresh frozen section.[Methods]We first used DESI-MSI to analyze frozen sections of 22 cases of OSCC which containing tumor and continuous mucosal margin.Mass spectrometry data wereextracted and characteristic molecules that could distinguish OSCC tumor from normal tissues were selected from numerous metabolite ions by lasso machine learning method.Furthermore,these characteristic molecules are further optimized to construct a panel of visual molecular diagnostic models that can be used for the diagnosis of margin status and the assessment of mucosal marign distance.Student's t test were used to determine lipid-markers of positive margin and negative margin.[Results]Using lasso machine learning method,179 characteristic metabolites that could distinguish oral squamous cell carcinoma from normal tissues were selected from 22,955 metabolites.These characteristic molecules were further optimized,ultimately,14 ions were determined as a combination of molecular diagnostic models for surgical margin status diagnosis and mucosal margin distance assessment.It was found that the accuracy of the DESI-MSI molecular diagnostic model in predicting different surgical margin status such as tumor region,positive margin and negative margin reached 100%,88.89% and 88.89% respectively,while the accuracy of close margin was only 33.33%,reflecting the complexity and controversy of the definition of close margin.The DESI-MSI molecular diagnostic model was used to measure the mucosal safety distance of OSCC in frozen sections.The results showed that the mucosa safety distance of most cases were located in the close margin area(2-10mm),and only 2 cases were located in the negative margin area(> 10mm).A total of 647 m/z average mass spectrogram were extracted from five regions(T,M1,M2,M3,N).Nine lipids(365.33 m/z,836.53 m/z,524.29 m/z,331.26 m/z,359.29 m/z,337.31 m/z,393.37 m/z,573.48 m/z,834.51m/z)and one lipid(331.26 m/z)were determined as markers of negative margin and positive margin respectively.The above results provide a theoretical basis for DESI-MSI in situ imaging technique to evaluate the surgical safety margin of OSCC,which has great potential for clinical application.[Conclusions]We successfully constructed a molecular diagnostic model,which can accurately distinguish tumor from normal tissues.The ion mass spectrometry imaging of the diagnostic model can visually distinguish tumor tissues from the mucosal margin tissues,and can directly delineate the safety margin distance on frozen sections and predict different surgical margin status.Meanwhile,9 lipid molecules were selected as negative margin markers and 1 lipid molecule as positive margin markers of OSCC.[Overall Conclusions]In this paper,molecular markers of surgical-margin of OSCC were studied by mass spectrometry based on amino acid metabolism and lipid metabolism.Firstly,four amino acids(aspartic acid,glutamic acid,proline and valine)and six amino acids(aspartic acid,asparagine,proline,glutamic acid,histidine and serine)were determined as markers of negative margin and dysplastic margin respectively,by non-in-situ mass spectrometry technique.Further validated aspartic acid metabolism may be involved in the development of OSCC.The expression of asparagine synthase(key enzyme of aspartic acid metabolism)in dysplastic margin was related to the poor prognosis of OSCC.Next,DESI-MSI technique was used to visualize in situ lipid markers of surgical-margin in oral squamous cell carcinoma.14 lipid-ions were determined by meachine learning as a combination of molecular diagnostic models for surgical margin status diagnosis and mucosal margin distance assessment.Nine lipids(365.33 m/z,836.53 m/z,524.29 m/z,331.26 m/z,359.29 m/z,337.31 m/z,393.37 m/z,573.48 m/z,834.51m/z)and one lipid(331.26 m/z)were determined as markers of negative margin and positive margin respectively.The results are helpful to improve the diagnosis of surgical safety margin and prognosis of oral squamous cell carcinoma.