14 August 2023 in BMC Cancer
Purpose: This study aims to investigate the predictive value of the combined index smni(skeletal muscle index (SMI)-prognostic nutrition index(PNI)) for the postoperative survival of patients with advanced gastric cancer(AGC). Methods: 650 patients with AGC from two centers (290 cases from the First Affiliated Hospital of Dalian University and 360 points from the Fujian Medical University Union Hospital) were selected as the study subjects based on unified screening criteria. Clinical data, preoperative abdominal CT images, results of hematology-related examinations, tumor-related characteristics, and surgical and follow-up data of the patients were collected and organized. The L3 vertebral level muscle area was measured using computer-assisted measurement techniques, and the skeletal muscle index(SMI) was calculated based on this measurement. The prognostic nutrition index (PNI) was calculated based on serum albumin and lymphocyte count indicators. The Kaplan-Meier survival analysis of data from the First Affiliated Hospital was used to determine that SMI and PNI are significantly correlated with the postoperative survival rate of patients with advanced gastric cancer. Based on this, a novel combined index smni was fitted and stratified for risk. Cox proportional hazards regression analysis was used to determine that the index smni is an independent prognostic risk factor for patients with AGC after surgery. The ROC curve was used to describe the predictive ability of the new combined index and its importance and predictive power in predicting postoperative survival of patients with AGC, which was verified in the data of Fujian Medical University Union Hospital. Result: The Kaplan-Meier curve analysis of the combined indicator smni Is clearly associated with long-term survival(3-year OS (P < 0.001) and DSS (P < 0.001)), univariate analysis and multivariate analysis showed that smni was an independent prognostic risk factor, The ROC curve for the first center 3-year OS(AUC = 0.678), DSS(AUC = 0.662) show good predictive ability and were validated in the second center. Conclusion: The combined index smni has a good predictive ability for the postoperative survival rate of patients with AGC and is expected to provide a new reference basis and more accurate and scientific guidance for the postoperative management and treatment of patients with AGC.
7 June 2024 in IEEE Transactions on Instrumentation and Measurement
Given the hurdles of limited data availability, annotation challenges, and constrained generalization capabilities in medical image segmentation, few-shot learning (FSL) has become a prominent approach due to its efficacy in learning new categories from minimal annotated samples. Existing methods predominantly adopt the support-guided paradigm, wherein support prototypes, derived from annotated support images, guide the segmentation of unlabeled query images. This approach, however, is prone to local information loss and classification bias. To counter these challenges, we introduce a novel Dual-Guided Prototype Alignment Network (DGPANet) that integrates a bidirectional segmentation architecture. The support-to-query branch incorporates query guidance information into the conventional support-guided approach for segmenting query images, whereas the complementary query-to-support branch segments support images in reverse with the dual guidance provided. This design not only optimizes our network but also alleviates segmentation biases pertaining to support category objects. Furthermore, we propose an adaptive prototype module to foster information amalgamation by adaptively clustering analogous elements. Moreover, we propose an adaptive prototype alignment loss to better improve the dual-guided optimization procedure. Extensive experiments show that our proposed DGPANet favors against state-of-the-art approaches on three public datasets, including Abdominal-CT, Abdominal-MRI, and Cardiac-MRI. Codes are at https://github.com/shiyi0306/DGPANet.
12 July 2024 in IEEE Journal of Biomedical and Health Informatics
The potential benefits of automatic radiology report generation, such as reducing misdiagnosis rates and enhancing clinical diagnosis efficiency, are significant. However, existing data-driven methods lack essential medical prior knowledge, which hampers their performance. Moreover, establishing global correspondences between radiology images and related reports, while achieving local alignments between images correlated with prior knowledge and text, remains a challenging task. To address these shortcomings, we introduce a novel Eye Gaze Guided Cross-modal Alignment Network (EGGCA-Net) for generating accurate medical reports. Our approach incorporates prior knowledge from radiologists' Eye Gaze Region (EGR) to refine the fidelity and comprehensibility of report generation. Specifically, we design a Dual Fine-Grained Branch (DFGB) and a Multi-Task Branch (MTB) to collaboratively ensure the alignment of visual and textual semantics across multiple levels. To establish fine-grained alignment between EGR-related images and sentences, we introduce the Sentence Fine-grained Prototype Module (SFPM) within DFGB to capture cross-modal information at different levels. Additionally, to learn the alignment of EGR-related image topics, we introduce the Multi-task Feature Fusion Module (MFFM) within MTB to refine the encoder output information. Finally, a specifically designed label matching mechanism is designed to generate reports that are consistent with the anticipated disease states. The experimental outcomes indicate that the introduced methodology surpasses previous advanced techniques, yielding enhanced performance on two extensively used benchmark datasets: Open-i and MIMIC-CXR.