A5090557986- Radiomics and Machine Learning in Medical Imaging
- Single-cell and spatial transcriptomics
- Ferroptosis and cancer prognosis
- Cancer Immunotherapy and Biomarkers
- Cancer Genomics and Diagnostics
- Cancer Research and Treatments
- Cell Image Analysis Techniques
- Medical Imaging Techniques and Applications
- Nanoplatforms for cancer theranostics
- Gastric Cancer Management and Outcomes
- AI in cancer detection
- Colorectal Cancer Screening and Detection
- Cancer Cells and Metastasis
- Immune cells in cancer
- Monoclonal and Polyclonal Antibodies Research
- Gene expression and cancer classification
- Ubiquitin and proteasome pathways
- Bioinformatics and Genomic Networks
- Genetic factors in colorectal cancer
- Colorectal Cancer Surgical Treatments
- Cancer-related molecular mechanisms research
- Colorectal Cancer Treatments and Studies
The University of Queensland
2019-2025
Stanford University
2015-2023
Palo Alto University
2015
Abstract Purpose: The liver is the most frequent metastatic site for colorectal cancer. Its microenvironment modified to provide a niche that conducive cancer cell growth. This study focused on characterizing cellular changes in (mCRC) tumor (TME). Experimental Design: We analyzed series of microsatellite stable (MSS) mCRCs liver, paired normal tissue, and peripheral blood mononuclear cells using single-cell RNA sequencing (scRNA-seq). validated our findings multiplexed spatial imaging bulk...
Abstract Motivation Spatial transcriptomics (ST) technology is increasingly being applied because it enables the measurement of spatial gene expression in an intact tissue along with imaging morphology same tissue. However, current analysis methods for ST data do not use image pixel information, thus missing quantitative links between and morphology. Results We developed a user-friendly deep learning software, SpaCell, to integrate millions intensity values thousands measurements from...
Deep-learning classification systems have the potential to improve cancer diagnosis. However, development of these computational approaches so far depends on prior pathological annotations and large training datasets. The manual annotation is low-resolution, time-consuming, highly variable subject observer variance. To address this issue, we developed a method, H&E Molecular neural network (HEMnet). HEMnet utilizes immunohistochemistry as an initial molecular label for cells image trains...
Abstract Gastric cancer precursors demonstrate highly-variable rates of progression toward neoplasia. Certain high-risk precursors, such as gastric intestinal metaplasia with advanced histologic features, may be at up to 30-fold increased risk for compared lower-risk metaplasia. The biological differences between high- and low-risk lesions have been incompletely explored. In this study, we use several clinical cohorts characterize the microenvironment relative using bulk, spatial,...
Gastric intestinal metaplasia (
ABSTRACT Purpose The liver is the most frequent metastatic site for colorectal cancer ( CRC ). Its microenvironment modified to provide a niche that allows cell growth. This study focused on characterizing cellular changes in mCRC ) tumor TME Experimental Design We analyzed series of microsatellite stable (MSS) mCRCs liver, paired normal tissue and peripheral blood mononuclear cells using single RNA-seq scRNA-seq validated our findings multiplexed spatial imaging bulk gene expression with...
Abstract Gastric intestinal metaplasia (GIM) is a precursor lesion for the subtype of gastric cancer (GC). A risk stratification tool Operative Link on GIM (OLGIM), system that relies histopathologic annotation biopsies. Advanced OLGIM (stages III and IV) have highest progression to GC. Molecular biomarkers advanced are lacking. We explored transcriptomics aid assessment efforts among GC precursors.We used clinical genomic data from four cohorts: 1) GAPS, cohort OLGIM-staged patients...
We conducted a spatial analysis using imaging mass cytometry applied to stage III colorectal adenocarcinomas. This study used multiplexed markers distinguish individual cells and their organization from 52 cancers. determined the landscape features of cellular in CRC tumor microenvironment. single-cell identified 10 unique cell phenotypes microenvironment that included stromal immune with subset which had proliferative phenotype. These special neighborhood interactions between single as well...
Abstract Motivation Spatial transcriptomics technology is increasingly being applied because it enables the measurement of spatial gene expression in an intact tissue along with imaging morphology same tissue. However, current analysis methods for data do not use image pixel information, thus missing quantitative links between and morphology. Results We developed user-friendly deep learning software, SpaCell, to integrate millions intensity values thousands measurements from...
<h3>Background</h3> Cancer research experiments often require the dissociation of cells from their native tissue before molecular profiling, leading to loss spatial context. The cancer genomics has shifted mostly profiling tumour DNA mutations towards current frontier investigating individual genes and gene products in single immediate microenvironments. Information at this level with context enables us link cancer–causing environmental factors outcomes cell signalling, responses survival...
ABSTRACT Deep learning cancer classification systems have the potential to improve diagnosis. However, development of these computational approaches depends on prior annotation through a pathologist. This initial step relying manual, low-resolution, time-consuming process is highly variable and subject observer variance. To address this issue, we developed novel method, H&E Molecular neural network (HEMnet). two-step utilises immunohistochemistry as an molecular label for cells image...
Introduction: A substantial fraction of colonoscopies are performed for post-polypectomy surveillance. The 2012 MSTF Guidelines were the first to make explicit recommendations polyp surveillance after colonoscopy (i.e. when perform “repeat surveillance”) depending on history low-risk or high-risk adenomas. We hypothesized that documentation prior adenoma is frequently incomplete, and adherence with repeat poor in clinical practice. Our aims examine availability data patients' comprehensive...
<div>AbstractPurpose:<p> The liver is the most frequent metastatic site for colorectal cancer. Its microenvironment modified to provide a niche that conducive cancer cell growth. This study focused on characterizing cellular changes in (mCRC) tumor (TME).</p>Experimental Design:<p>We analyzed series of microsatellite stable (MSS) mCRCs liver, paired normal tissue, and peripheral blood mononuclear cells using single-cell RNA sequencing (scRNA-seq). We validated our...
<div>AbstractPurpose:<p> The liver is the most frequent metastatic site for colorectal cancer. Its microenvironment modified to provide a niche that conducive cancer cell growth. This study focused on characterizing cellular changes in (mCRC) tumor (TME).</p>Experimental Design:<p>We analyzed series of microsatellite stable (MSS) mCRCs liver, paired normal tissue, and peripheral blood mononuclear cells using single-cell RNA sequencing (scRNA-seq). We validated our...
<p>(A) Dot plot depicting expression levels of respective ligands and receptors together with the percentage cells expressing them. (B) Top 10 percent interactions inferred between cell types in TME. Edge weights are proportional to number interactions, circle sizes each group, edge color represents type as sender. For scale, autocrine CAFs 68. (A-B) Data from seven mCRCs five paired normal liver tissue.</p>
<p>(A) UMAP representation of dimensionally reduced data following batch-corrected graph-based clustering all tumor epithelial cells annotated by sample. (B) Heatmap depicting expression five highest significantly expressed genes (adjusted p-value <0.05) per patient. (C) inferred single-cell CNV profiles tumors and reference cells. (A-C) Data from seven mCRCs.</p>
<p>Heatmap depicting cell states from monocytes and macrophages identified using Ecotyper NMF analysis, annotated by condition genes with highest fold change in each state. Data seven mCRCs, five paired normal liver tissue two PBMCs.</p>
<p>(A) UMAP representation of dimensionally reduced data following batch-corrected graph-based clustering all tumor epithelial cells annotated by sample. (B) Heatmap depicting expression five highest significantly expressed genes (adjusted p-value <0.05) per patient. (C) inferred single-cell CNV profiles tumors and reference cells. (A-C) Data from seven mCRCs.</p>
<p>Supplemental Table 1: Antibodies, reporters, imaging order and exposure times used in CODEX imaging. EPCAM, SPP1, CD163 were excluded from downstream analysis due to non-specific staining; Supplemental 2: Sequencing metrics for scRNA-seq; 3: Absolute number of cells detected each sample per cell lineage; 4: Differentially expressed genes belonging respective macrophage clusters compared by site origin. pct.1, pct.2 indicate percentage expressing given gene the cluster interest all...
<p>Heatmap depicting cell states from monocytes and macrophages identified using Ecotyper NMF analysis, annotated by condition genes with highest fold change in each state. Data seven mCRCs, five paired normal liver tissue two PBMCs.</p>
<p>(A) Dot plot depicting expression levels of respective ligands and receptors together with the percentage cells expressing them. (B) Top 10 percent interactions inferred between cell types in TME. Edge weights are proportional to number interactions, circle sizes each group, edge color represents type as sender. For scale, autocrine CAFs 68. (A-B) Data from seven mCRCs five paired normal liver tissue.</p>
<p>(A-C) Comparisons with Pearson correlation between (A) proportions of cell lineages in five samples both scRNA-seq and CODEX data. (B) Average expression LGALS3 CD68 macrophages across all 15 patients. (C) COL4A1 ACTA2 CAFs patients.</p>
<p>Supplemental Table 1: Antibodies, reporters, imaging order and exposure times used in CODEX imaging. EPCAM, SPP1, CD163 were excluded from downstream analysis due to non-specific staining; Supplemental 2: Sequencing metrics for scRNA-seq; 3: Absolute number of cells detected each sample per cell lineage; 4: Differentially expressed genes belonging respective macrophage clusters compared by site origin. pct.1, pct.2 indicate percentage expressing given gene the cluster interest all...
<p>Graphical representation of cell types identified in CODEX analysis image data from respective patients.</p>