A dual self-attentive transformer U-Net model for precise pancreatic segmentation and fat fraction estimation

• Published in: BMC medical imaging Vol. 25; no. 1; pp. 315 - 18
• Main Authors: Shanmugam, Ashok, Radhabai, Prianka Ramachandran, KVN, Kavitha, Imoize, Agbotiname Lucky
• Format: Journal Article
• Language: English
• Published: London BioMed Central 04.08.2025; BioMed Central Ltd; BMC
• Subjects: Accuracy; Adipose Tissue - diagnostic imaging; Artificial neural networks; Attention mechanism; Beta cells; Biomarkers; Computational linguistics; Computed tomography; Computed tomography (CT); CT imaging; Diabetes; Diabetes mellitus (non-insulin dependent); Electric transformers; Fat fraction estimation; Humans; Imaging; Insulin resistance; Language processing; Literature reviews; Medical imaging; Medicine; Medicine & Public Health; Metabolic disorders; Metabolic syndrome; Mortality; Natural language interfaces; Natural language processing; Neural networks; Neural Networks, Computer; Pancreas; Pancreas - diagnostic imaging; Pancreatic diseases; Pancreatic segmentation; Radiology; Segmentation; Semantics; Tomography; Tomography, X-Ray Computed - methods; Transformer Unet model; Type 2 diabetes; Pancreatic segmentation; Attention mechanism; Computed tomography (CT); Fat fraction estimation; Transformer Unet model
• ISSN: 1471-2342; 1471-2342
• DOI: 10.1186/s12880-025-01852-5

Accurately segmenting the pancreas from abdominal computed tomography (CT) images is crucial for detecting and managing pancreatic diseases, such as diabetes and tumors. Type 2 diabetes and metabolic syndrome are associated with pancreatic fat accumulation. Calculating the fat fraction aids in the investigation of β-cell malfunction and insulin resistance. The most widely used pancreas segmentation technique is a U-shaped network based on deep convolutional neural networks (DCNNs). They struggle to capture long-range biases in an image because they rely on local receptive fields. This research proposes a novel dual Self-attentive Transformer Unet (DSTUnet) model for accurate pancreatic segmentation, addressing this problem. This model incorporates dual self-attention Swin transformers on both the encoder and decoder sides to facilitate global context extraction and refine candidate regions. After segmenting the pancreas using a DSTUnet, a histogram analysis is used to estimate the fat fraction. The suggested method demonstrated excellent performance on the standard dataset, achieving a DSC of 93.7% and an HD of 2.7 mm. The average volume of the pancreas was 92.42, and its fat volume fraction (FVF) was 13.37%.