Correspondence-Aware Manifold Learning for Microscopic and Spatial Omics Imaging: A Novel Data Fusion Method Bringing Mass Spectrometry Imaging to a Cellular Resolution

Abstract

High-dimensional molecular measurements are transforming the field of pathology into a data-driven discipline. While hematoxylin and eosin (H&E) stainings are still the gold standard to diagnose diseases, the integration of microscopic and molecular information is becoming crucial to advance our understanding of tissue heterogeneity. To this end, we propose a data fusion method that integrates spatial omics and microscopic data obtained from the same tissue slide. Through correspondence-aware manifold learning, we can visualize the biological trends observed in the high-dimensional omics data at microscopic resolution. While data fusion enables the detection of elements that would not be detected taking into account the separate data modalities individually, out-of-sample prediction makes it possible to predict molecular trends outside of the measured tissue area. The proposed dimensionality reduction-based data fusion paradigm will therefore be helpful in deciphering molecular heterogeneity by bringing molecular measurements such as mass spectrometry imaging (MSI) to the cellular resolution.

Tracking Control of the Dynamic Input-Output Economic System Based on Data Fusion

“…e so-called data fusion is the association, correlation, and integration of data and data obtained from single and multiple data sources to obtain accurate location and identity estimates, as well as the situation, characteristics, attributes, trends, and threats and their importance, with a data processing process for comprehensive and timely evaluation; this process is a continuous refinement process for its estimation, evaluation, and evaluation of additional data source needs [16,17]. It is also a process of continuous self-correction in the data processing process to obtain improved results.…”

Emerging Computational Methods in Mass Spectrometry Imaging

“…Other computational methods that have been developed to sharpen MSI images based on complementary information from other modalities include pan‐sharpening algorithm, [ 196 ] Laplacian pyramid fusion, [ 197 ] patch‐based super resolution, [ 180 ] and correspondence‐aware manifold learning. [ 184 ] Other microsampling‐based multimodal approaches that do not require special computations are not highlighted here. For additional information on both method development and application of multimodal experiments, the readers are referred to detailed reviews.…”

“…Data from other modalities provide complementary molecular or spatial information, which may be used to enhance the quality of MSI data or provide additional insights using more targeted analytical tools. MSI data have been coupled with histology, [ 91 , 133 , 178 , 179 , 180 , 181 , 182 , 183 , 184 , 185 , 186 , 187 ] fluorescence microscopy, [ 188 , 189 , 190 , 191 ] Allen brain atlas, [ 107 , 192 ] topology, [ 193 , 194 , 195 ] electron microscopy, [ 196 , 197 ] Raman spectroscopy, [ 198 , 199 , 200 , 201 ] infrared spectroscopy, [ 202 ] magnetic resonance imaging, [ 93 , 203 ] and microsampling LC‐MS/MS analysis of proteins, transcripts, and genes. [ 164 , 204 , 205 , 206 , 207 , 208 , 209 , 210 ] In these studies, computational approaches have been used for registration of the multimodal data.…”

Construction of a Diagnostic Model for Lymph Node Metastasis of the Papillary Thyroid Carcinoma Using Preoperative Ultrasound Features and Imaging Omics

“…Imaging omics was proved to be objective in image extraction of lymph node features in PTC and had important implications for prediction of clinical outcome [11][12][13][14][15]. Since imaging omics has been successfully applied to the diagnosis of thyroid cancer, lung cancer, liver cancer, breast cancer, and other diseases [16][17][18][19][20][21][22], it will also be employed in the present study.…”