Related Experiment Video
Updated: Nov 26, 2025
High Resolution 3D Imaging of the Human Pancreas Neuro-insular Network
Published on: January 29, 2018
Image-Based Machine Learning Algorithms for Disease Characterization in the Human Type 1 Diabetes Pancreas
Xiaohan Tang1, Irina Kusmartseva2, Shweta Kulkarni2
1Department of Pathology, Immunology and Laboratory Medicine, College of Medicine, University of Florida Diabetes Institute, Gainesville, Florida; Department of Metabolism and Endocrinology, The Second Xiangya Hospital, Central South University, Changsha, China.
Type 1 diabetes impacts the pancreas beyond the islets of Langerhans, affecting exocrine tissue composition. These anatomical changes in the pancreas are linked to disease progression and are detectable using machine learning analysis.
Area of Science:
- Endocrinology
- Pathology
- Computational Biology
Background:
- Type 1 diabetes (T1D) is traditionally viewed as a disease primarily affecting pancreatic beta cells.
- Emerging evidence suggests T1D may also impact the exocrine pancreas, the tissue responsible for digestive enzyme production.
Purpose of the Study:
- To investigate anatomical differences in the exocrine pancreas between individuals with and without type 1 diabetes.
- To determine if machine learning can identify T1D-related changes in pancreatic tissue composition.
Main Methods:
- Digital pathology and machine learning algorithms were applied to whole-slide images of human pancreata from organ donors.
- Tissue composition, including acinar, endocrine, and ductal areas, cell size, and density, was analyzed in nondiabetic controls, individuals at risk for T1D, and those with T1D.
Main Results:
- Type 1 diabetes significantly alters exocrine pancreatic area, acinar cell density, and size.
- Observed pancreatic changes correlated with the presence or absence of remaining insulin-producing cells.
- Pre-disease onset changes in the exocrine pancreas were not detected in autoantibody-positive subjects.
Conclusions:
- Type 1 diabetes involves significant anatomical alterations in the exocrine pancreas.
- Machine learning is a viable tool for evaluating disease processes using cross-sectional pancreatic tissue data.
- These findings offer new insights into the comprehensive impact of T1D on pancreatic structure.
