Mass Spectrometry Imaging for the Classification of Tumor Tissue

N.E. Mascini

Research output: ThesisDoctoral thesis 2 (Research NOT UU / Graduation UU)


Mass spectrometry imaging (MSI) can detect and identify many different molecules without the need for labeling. In addition, it can provide their spatial distributions as ‘molecular maps’. These features make MSI well suited for studying the molecular makeup of tumor tissue. Currently, there is an interest in using MSI to predict cancer progression and treatment response, which often remains a challenge in today’s clinical practice. Experimental evidence shows that tumor heterogeneity plays an important role in tumor biology and therefore in response to treatment. The molecular profiles generated by MSI reveal part of this heterogeneity. We describe the combination of matrix-assisted laser desorption/ionization (MALDI) MSI, histological tissue staining and multivariate data analysis to investigate inter- and intratumor heterogeneity.
The first part of this thesis focuses on the generation and use of MALDI-MSI data to predict treatment response and disease progression. Small pieces of formalin-fixed paraffin-embedded (FFPE) tissue were used, typically referred to as tissue microarrays (TMAs). We present an approach that combines MALDI-MSI on tissue microarrays with Principal Component Analysis and Linear Discriminant Analysis (PCA-LDA) to predict treatment response. The feasibility of this approach is evaluated on a set of 22 patient-derived xenograft models of triple-negative breast cancer. PCA-LDA is used to predict response to the chemotherapeutic drug cisplatin based on the proteomic information obtained with MALDI-MSI. A small predictive power is observed. In addition, the feasibility to predict lymph node metastasis and disease-specific survival from a sample set of 240 head and neck cancers is tested. We discuss how MALDI-MSI data from these samples is processed so that it could be used in five different classifiers: Linear and Quadratic Discriminant Analysis, a Naive Bayes classifier, Decision Tree classifier and Support Vector Machine. Only for some of the classifiers a small predictive power for disease-specific survival is shown.
The second part of this thesis focuses on intratumor heterogeneity. It presents a new method to detect and visualize hypoxic tumor regions and hypoxia-associated molecules in a breast tumor xenograft model. In hypoxic cells, the 2-nitroimidazole pimonidazole is reduced and forms reactive products that bind to cellular nucleophiles. We demonstrate that a reductively activated pimonidazole metabolite can be detected by MALDI-MSI. This pimonidazole metabolite is used to image hypoxic tissue regions and hypoxia-associated lipids and metabolites. Pimonidazole is a widely used hypoxia marker. We expect that the presented MALDI-MSI approach might be also applicable to other tissues from pimonidazole-injected animals or humans.
Perspectives for future research include further optimization of the sample preparation for FFPE tissue and the use of a modified chemical marker for the detection of hypoxia. As the focus shifts from MSI technology to application, research projects become increasingly multidisciplinary. This development calls for close collaboration of researchers with expertise in MSI, histo(patho)logy and data analysis.
Original languageEnglish
Awarding Institution
  • Utrecht University
  • Heeren, R.M.A., Primary supervisor
Award date20 Jan 2016
Print ISBNs978-90-77209-97-4
Publication statusPublished - 20 Jan 2016


  • Mass Spectrometry Imaging
  • Tumor
  • Hypoxia
  • Tissue microarray


Dive into the research topics of 'Mass Spectrometry Imaging for the Classification of Tumor Tissue'. Together they form a unique fingerprint.

Cite this