DenseNet model incorporating hybrid attention mechanisms and clinical features for pancreatic cystic tumor classification.

PURPOSE: The aim of this study is to develop a deep learning model capable of discriminating between pancreatic plasma cystic neoplasms (SCN) and mucinous cystic neoplasms (MCN) by leveraging patient-specific clinical features and imaging outcomes. The intent is to offer valuable diagnostic support to clinicians in their clinical decision-making processes. METHODS: The construction of the deep learning model involved utilizing a dataset comprising abdominal magnetic resonance T2-weighted images obtained from patients diagnosed with pancreatic cystic tumors at Changhai Hospital. The dataset comprised 207 patients with SCN and 93 patients with MCN, encompassing a total of 1761 images. The foundational architecture employed was DenseNet-161, augmented with a hybrid attention mechanism module. This integration aimed to enhance the network's attentiveness toward channel and spatial features, thereby amplifying its performance. Additionally, clinical features were incorporated prior to the fully connected layer of the network to actively contribute to subsequent decision-making processes, thereby significantly augmenting the model's classification accuracy. The final patient classification outcomes were derived using a joint voting methodology, and the model underwent comprehensive evaluation. RESULTS: Using the five-fold cross validation, the accuracy of the classification model in this paper was 92.44%, with an AUC value of 0.971, a precision rate of 0.956, a recall rate of 0.919, a specificity of 0.933, and an F1-score of 0.936. CONCLUSION: This study demonstrates that the DenseNet model, which incorporates hybrid attention mechanisms and clinical features, is effective for distinguishing between SCN and MCN, and has potential application for the diagnosis of pancreatic cystic tumors in clinical practice.

Application of a pyramid pooling Unet model with integrated attention mechanism and Inception module in pancreatic tumor segmentation.

BACKGROUND: The segmentation and recognition of pancreatic tumors are crucial tasks in the diagnosis and treatment of pancreatic diseases. However, due to the relatively small proportion of the pancreas in the abdomen and significant shape and size variations, pancreatic tumor segmentation poses considerable challenges. PURPOSE: To construct a network model that combines a pyramid pooling module with Inception architecture and SE attention mechanism (PIS-Unet), and observe its effectiveness in pancreatic tumor images segmentation, thereby providing supportive recommendations for clinical practitioners. MATERIALS AND METHODS: A total of 303 patients with histologically confirmed pancreatic cystic neoplasm (PCN), including serous cystic neoplasm (SCN) and mucinous cystic neoplasm (MCN), from Shanghai Changhai Hospital between March 2011 and November 2021 were included. A total of 1792 T2-weighted imaging (T2WI) slices were used to build a CNN model. The model employed a pyramid pooling Inception module with a fused attention mechanism. The attention mechanism enhanced the network's focus on local features, while the Inception module and pyramid pooling allowed the network to extract features at different scales and improve the utilization efficiency of global information, thereby effectively enhancing network performance. RESULTS: Using three-fold cross-validation, the model constructed by us achieved a dice score of 85.49 ± 2.02 for SCN images segmentation, and a dice score of 87.90 ± 4.19 for MCN images segmentation. CONCLUSION: This study demonstrates that using deep learning networks for the segmentation of PCNs yields favorable results. Applying this network as an aid to physicians in PCN diagnosis shows potential for clinical applications.