1 February 2020 in Biomaterials
29 May 2024 in IEEE Journal of Biomedical and Health Informatics
The Segment Anything Model (SAM) is a foundational model that has demonstrated impressive results in the field of natural image segmentation. However, its performance remains suboptimal for medical image segmentation, particularly when delineating lesions with irregular shapes and low contrast. This can be attributed to the significant domain gap between medical images and natural images on which SAM was originally trained. In this paper, we propose an adaptation of SAM specifically tailored for lesion segmentation termed LeSAM. LeSAM first learns medical-specific domain knowledge through an efficient adaptation module and integrates it with the general knowledge obtained from the pre-trained SAM. Subsequently, we leverage this merged knowledge to generate lesion masks using a modified mask decoder implemented as a lightweight U-shaped network design. This modification enables better delineation of lesion boundaries while facilitating ease of training. We conduct comprehensive experiments on various lesion segmentation tasks involving different image modalities such as CT scans, MRI scans, ultrasound images, dermoscopic images, and endoscopic images. Our proposed method achieves superior performance compared to previous state-of-the-art methods in 8 out of 12 lesion segmentation tasks while achieving competitive performance in the remaining 4 datasets. Additionally, ablation studies are conducted to validate the effectiveness of our proposed adaptation modules and modified decoder.
6 November 2023 in IEEE Transactions on Instrumentation and Measurement
Due to the potential harmful risks caused by the excessive X-ray radiation dose, low-dose has emerged as one of the major rules of computed tomography (CT) for clinical applications. However, reducing radiation dose severely leads to degraded reconstructed CT image quality with noises and artifacts, and it may seriously hamper clinical diagnosis. Currently, model-driven deep learning (DL)-based dual-domain learning with the expert knowledge can reconstruct CT images from projections acquired in non-ideal conditions, and they have been demonstrated with inspired performance in Low-dose CT (LDCT) reconstruction. However, because of domain-progressive strategy, these dual-domain methods are prone to produce seriously secondary artifacts and suffer from subtle structural degeneration. In this paper, we propose a novel deep dual-domain united guiding learning framework, i.e., DUGL-Net. First, we put forward a domain-guiding strategy. The projection-domain network plays the guiding role on image-domain network to effectively enhance overall network stability and powerfully avoid secondary artifacts. Second, we elaborately design a Global-Local Transformer-Convolution U-Net, i.e., GL-TCUNet, as the base network in both projection-domain and image-domain for reinforcing dual-domain model learning. This efficiently leverage the advantages of Swin Transformer and CNN to suppress noises and artifacts. Finally, the proposed DUGL-Net is evaluated on the 2016 NIH- AAPM-Mayo clinic LDCT Grand Challenge dataset and real data, and our results validate that the proposed method outperforms the state-of-the-art competing methods. This efficient, accurate, and reliable LDCT denoising technique has a great potential for clinical applications.
9 June 2024 in Insights into Imaging
Objectives: To investigate the diagnostic performance of the apparent diffusion coefficient (ADC) for low to intermediate-risk prostate cancer (PCa), as well as its correlation with the prognostic Gleason score (GS). Materials and methods: Retrospective analysis of MRI images and relevant clinical data from patients with prostate disease. The differences in ADC between different GS groups were compared, and the efficacy of ADC in PCa diagnosis were analyzed. Furthermore, the diagnostic performance of the mean ADC (ADCmean) and minimum ADC (ADCmin) values was compared. Results: There were 1414 patients with 1631 lesions. In terms of GS, both ADCmin and ADCmean values of the GS 4 + 3 group were significantly lower than those of the GS 3 + 4 group, GS 3 + 3 group, and the benign group, with all differences being statistically significant (p < 0.01). The AUC values for diagnosing PCa based on ADCmin and ADCmean were 0.914 and 0.944, respectively. The corresponding diagnostic thresholds were 0.703 × 10−3 mm2/s for ADCmin and 0.927 × 10−3 mm2/s for ADCmean. The magnitudes of ADCmin and ADCmean values exhibited a negative correlation with GS values (ρ = −0.750, p < 0.001; ρ = −0.752, p < 0.001). Conclusions: ADC values demonstrate an inverse relationship with the invasiveness of PCa, indicating that higher invasiveness is associated with lower ADC values. Additionally, ADC values exhibit high diagnostic potential, sensitivity, and specificity for distinguishing between GS 3 + 4 and GS 4 + 3 lesions. Moreover, the diagnostic value of ADCmean is even more significant, highlighting its crucial role in the diagnosis of low to intermediate-risk PCa. Critical relevance statement: ADC values are a valuable tool for distinguishing different levels of aggressiveness in PCa. They help in the preoperative assessment of the biological characteristics of PCa, allowing clinicians to develop personalized treatment strategies, effectively mitigating the risk of unnecessary interventions. Key Points: The preoperative GS is crucial for planning the clinical treatment of PCa. The invasiveness of PCa is inversely correlated with ADC values. ADC values play a crucial role in the accurate preoperative evaluation of low to intermediate-risk PCa, thus aiding clinicians in developing tailored treatment plans. Graphical
27 May 2025 in ACS Applied Materials & Interfaces
18 April 2023 in Oncology Letters
Primary pulmonary intravascular large B‑cell lymphoma (IVLBCL) is a rare, malignant extranodal lymphoma. It is difficult to diagnose clinically as it requires a combination of clinical and computed tomography (CT) evaluations, as well as laboratory and pathological examinations. In the present study, 4 cases of primary pulmonary IVLBCL were reviewed. The patients' ages ranged from 60 to 69 years old. Of the 4 patients, 3 developed progressive dyspnea on exertion and intermittent fever. Other symptoms included coughing, chest tightness and weight loss. Laboratory data indicated that all patients had anemia, thrombocytopenia, hypoxemia, a markedly high serum lactate dehydrogenase level, elevated erythrocyte sedimentation rate and increased C‑reactive protein. CT demonstrated increased attenuation in bilateral lung parenchyma, especially in the upper lobes, with multiple ground‑glass opacities associated with small nodules in these patients. Initially, all 4 patients were misdiagnosed with pneumonia. However, none of them responded to anti‑inflammatory treatments. The pathologies of all patients were confirmed using lung biopsy. Only 1 patient received regular combination chemotherapy. Based on the observations of the present study, a standard regimen for lymphoma treatment may result in a notable clinical response.