Dataset Information

Cancer tissue classification using supervised machine learning applied to MALDI mass spectrometry imaging

ABSTRACT: With the current clinical available technologies, not all cancers are staged accurately. Because of this a large percentage of patients are misclassified before treatment, leading to under or over treatment. Therefore, a classification system based on the molecular feature is required to deter-mine the tumour behavior and metastatic potential. Here, we have shown the diagnostic potential of MALDI MSI using supervised machine learning approach in distinguishing cancerous colorectal tissue from normal with an overall accuracy of 98%. Also, shown is the capability of the technique in predicting the presence of metastasis in endometrial cancer with an overall accuracy of 80%. The development of such a model can help in determining the optimum treatment for cancerous pa-tients, reduce morbidity and better patient outcome.

INSTRUMENT(S): ultraflex

ORGANISM(S): Homo Sapiens (human)

TISSUE(S): Uterus, Epithelial Cell Of Uterus

DISEASE(S): Uterine Cancer

SUBMITTER: Parul Mittal

LAB HEAD: Prof Peter Hoffmann

PROVIDER: PXD025594 | Pride | 2022-02-17

REPOSITORIES: Pride

ACCESS DATA

Publications

Cancer Tissue Classification Using Supervised Machine Learning Applied to MALDI Mass Spectrometry Imaging.

Mittal Paul P Condina Mark R MR Klingler-Hoffmann Manuela M Kaur Gurjeet G Oehler Martin K MK Sieber Oliver M OM Palmieri Michelle M Kommoss Stefan S Brucker Sara S McDonnell Mark D MD Hoffmann Peter P

Cancers 20211027 21

Matrix assisted laser desorption/ionization mass spectrometry imaging (MALDI MSI) can determine the spatial distribution of analytes such as protein distributions in a tissue section according to their mass-to-charge ratio. Here, we explored the clinical potential of machine learning (ML) applied to MALDI MSI data for cancer diagnostic classification using tissue microarrays (TMAs) on 302 colorectal (CRC) and 257 endometrial cancer (EC)) patients. ML based on deep neural networks discriminated c ...[more]

PMID: 34771551

Similar Datasets

Project description:It has been recognised for a long time that cell surface glycans plays a vital role in the biological process and their altered form can lead to cancer. However, the molecular details and regulatory mechanism between the glycans and cancer is yet not clear. Over the past decade mass spectrometry-based techniques has become a prominent method for analysing glycans especially N-glycan matrix assisted laser desorption/ionisation mass spectrometry imaging (MALDI MSI). It is a powerful technique that combines mass spectrometry with histology, enabling the visualisation and label free detection of N-linked glycans on a single tissue section. Here, we carried out N-glycan MALDI MSI on six endometrial cancer (EC) formalin fixed paraffin embedded (FFPE) tissue sections including the adjacent normal endometrium, and tissue microarrays (TMA) consisting of 8 EC patients with lymph node metastasis (LNM) and 20 without LNM. By doing that several m/z values were detected that can significantly distinguish normal endometrium from cancerous. Also, detected a m/z value that can discriminate the primary tumour with LNM from those without. Identification of those discriminative m/z values was performed using porous graphitized carbon liquid chromatography tandem mass spectrometry (PGC-LC-MS/MS). Overall, we observed higher abundance of oligomannose in tumour regions compared to normal with AUC ranges from 0.85-0.99. Whereas, complex N-glycans were detected in lower abundance with AUC ranges from 0.03-0.28. Comparison of N-glycans between the primary tumours with LNM and without LNM indicates reduced abundance of a complex core-fucosylated N-glycan in patients with metastasis relative to without. In summary, N-glycan MALDI MSI can be used to characterize the cancerous endometrium from the normal, also patients with LNM from those without. Identification of those discriminative m/z values was performed using porous graphitized carbon liquid chromatography tandem mass spectrometry (PGC-LC-MS/MS).

Project description:Expression profiling of breast cancer tumours, comparing 10 year survivors to deceased patients Background It is of great significance to find better markers to correctly distinguish between high-risk and low-risk breast cancer patients since the majority of breast cancer cases are at present being overtreated. Methods 46 tumours from node-negative breast cancer patients were studied with gene expression microarrays. A t-test was carried out in order to find a set of genes where the expression might predict clinical outcome. Two classifiers were used to evaluate the gene lists on the different data sets, a correlation-based classifier and a VFI (Voting Features Interval) classifier. We then evaluated the predictive accuracy of this expression signature on tumour sets from two similar studies on lymph-node negative patients which had developed gene expression signatures superior to current methods in classifying node-negative breast tumours. These two signatures were also tested on our material. Results A list of 51 genes whose expression profiles could predict clinical outcome with high accuracy in our material (96% or 89% accuracy in cross-validation, depending on type of classifier) was developed. When tested on two independent data sets, the expression signature based on the 51 identified genes had good predictive qualities in one of the data sets (74% accuracy), whereas their predictive value on the other data set were poor, presumably due to the fact that only 23 of the 51 genes were found in that material. We also found that previously developed expression signatures could predict clinical outcome well to moderately well in our material (72% and 61%, respectively). Conclusion The list of 51 genes derived in this study might have potential for clinical utility as a prognostic gene set, and may include candidate genes of potential relevance for clinical outcome in breast cancer. According to the predictions by this expression signature, 30 of the 46 patients should have had different adjuvant treatment than they did. Keywords: Expression Microarray, Lymph-node-negative Breast Cancer, Clinical Outcome, Classification 23 fresh frozen primary node negative breast cancer tumours from 10 year survivors were compared to 23 tumours from deceased patients in order to find genes where the expression differed between the two groups Microarrays were produced at the Swegene DNA Microarray Resource Center, Department of Oncology, Lund University, Sweden (http://swegene.onk.lu.se). Frozen tumours were homogenized together with TRIzol Reagent (Invitrogen, Carlsbad, CA, USA). From the cell-suspension total RNA was extracted using RNeasy mini kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s protocol. The quality of the RNA was evaluated with the Agilent 2100 BioAnalyser (Agilent Technologies, Palo Alto, CA, USA). Specimens where the 28S/16S ratio was lower than 1.0 or the RIN-value was lower than 6.7 were excluded from the study. For each sample, Cy3-dCTP (Amersham Biosciences, Buckinghamshire, UK) labelled probes were synthesized from 5 µg of the total tumour RNA and Cy5-dCTP (Amersham Biosciences) labelled reference were synthesized from 5 µg of commercial reference RNA (Universal Human Reference RNA, Stratagene, La Jolla, CA, USA) by reverse transcription. The probes were purified using ChipShot™ labelling cleanup system (Promega, Madison, WI, USA). The hybridizations were carried out using Pronto! Micro Array reagent systems (Corning Inc., Corning, NY, USA). For each sample, labelled tumour cDNA and reference cDNA were co-precipitated and hybridised to the microarray slide. The microarray slides were scanned with Agilent microarray scanner G2565AA (Agilent Technologies) and image analysis was performed using the Genepix 6.0.0.45 software (Axon Instruments, Union City, CA, USA).

Dataset Information

Cancer tissue classification using supervised machine learning applied to MALDI mass spectrometry imaging

Publications

Cancer Tissue Classification Using Supervised Machine Learning Applied to MALDI Mass Spectrometry Imaging.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure