Role of clinical and multidetector computed tomography (MDCT) features in the prediction of patients with intestinal lipoma developing intussusception.

Background: Intestinal lipoma is considered the most common benign tumor that causes intussusception. This retrospective case-control study aimed to present the clinical and multidetector computed tomography (MDCT) features between intestinal lipomas with and without intussusception and examine risk factors that predict intussusception caused by intestinal lipomas. Methods: We retrospectively analyzed 281 adult patients diagnosed with intestinal lipoma by radiologists using whole-abdominal MDCT between January 2015 and August 2022. Patients were divided into adult intussusception (AI) and non-AI groups based on MDCT images. Univariate logistic regression was performed to identify risk factors for intestinal lipoma-induced intussusception. Results: A total of 281 patients with intestinal lipomas were included in the study, with an average age of 68.0±11.3 years, and the male to female ratio was about 1:1.4. Among them, 24 patients developed lipoma-induced intussusception. Patients in the AI group presented with more abdominal pain (70.8% vs. 47.1%, P=0.03), nausea/vomiting (37.5% vs. 14.8%, P=0.009), hematochezia/melena (29.2% vs. 11.3%, P=0.02), and abdominal tenderness (66.7% vs. 24.9%, P<0.001). Lipomas were more common in the small bowel (224/281, 79.7%) than the large bowel (57/281, 20.3%). Lipomas in the AI group showed more heterogeneous hypodensity (41.7% vs. 15.6%, P=0.004), longer length (median, 2.2 vs. 1.2 cm, P<0.001), and larger volume (median, 4.1 vs. 0.6 cm3, P<0.001). In the univariate logistic regression, lipoma density [odds ratio (OR) =3.875, 95% confidence interval (CI): 1.609-9.331, P=0.003] and lipoma length (OR =3.216, 95% CI: 1.977-5.231, P<0.001) were risk factors for intestinal lipoma-induced intussusception. Conclusions: More patients in the AI group have digestive tract symptoms than those in the non-AI group. Lipoma density and length are risk factors for intussusception in patients with intestinal lipoma. In addition, the common site of intestinal lipoma may have changed from the colon to the small intestine.

Clinical Presentations, MDCT Features, and Treatment of Three Types of Adult Intussusceptions Based on the Location.

PURPOSE: This study aimed to explore the similarities and differences in clinical presentations, multidetector computed tomographic (MDCT) features, and treatment of three types of adult intussusceptions based on location. METHODS: We retrospectively reviewed 184 adult patients with 192 intussusceptions. Depending on the location, intussusceptions were classified as enteric, ileocolic, and colonic types. The similarities and differences of clinical presentations, MDCT features, and treatment of three types of adult intussusception were compared. Meanwhile, the three types of intussusceptions were further divided into surgical and conservative groups based on the treatment. Uni- and multivariate logistic analyses were used to identify risk factors for intussusception requiring surgery. RESULTS: Enteric and ileocolic intussusceptions were mainly presented with abdominal pain (78.46% and 85.71%). Hematochezia/melena (64.29%) was the main symptom of colonic intussusception. On MDCT, ileocolic intussusceptions were longer in length and had more signs of intestinal necrosis (hypodense layer, fluid collection and no/poor bowel wall enhancement) than enteric and colonic intussusceptions. Moreover, it was found that 93.88% (46/49) of ileocolic intussusception and 98.59% (70/71) of colonic intussusception belonged to the surgical group, whereas only 43.06% (31/72) of enteric intussusception belonged to the surgical group. Intussusception length (OR=1.171, P=0.028) and discernible lead point on MDCT (OR=21.003, P<0.001) were reliable indicators of enteric intussusception requiring surgery. CONCLUSION: Ileocolic intussusception may be more prone to intestinal necrosis than enteric and colonic intussusceptions, requiring more attention from clinicians. Surgery remains the treatment of choice for most ileocolic and colonic intussusceptions. Less than half of enteric intussusceptions require surgery, and MDCT features are effective in identifying them.

Neural-IMLS: Self-supervised Implicit Moving Least-Squares Network for Surface Reconstruction.

Surface reconstruction is a challenging task when input point clouds, especially real scans, are noisy and lack normals. Observing that the Multilayer Perceptron (MLP) and the implicit moving least-square function (IMLS) provide a dual representation of the underlying surface, we introduce Neural-IMLS, a novel approach that directly learns a noise-resistant signed distance function (SDF) from unoriented raw point clouds in a self-supervised manner. In particular, IMLS regularizes MLP by providing estimated SDFs near the surface and helps enhance its ability to represent geometric details and sharp features, while MLP regularizes IMLS by providing estimated normals. We prove that at convergence, our neural network produces a faithful SDF whose zero-level set approximates the underlying surface due to the mutual learning mechanism between the MLP and the IMLS. Extensive experiments on various benchmarks, including synthetic and real scans, show that Neural-IMLS can reconstruct faithful shapes even with noise and missing parts. The source code can be found at https://github.com/bearprin/Neural-IMLS.

Laplacian2Mesh: Laplacian-Based Mesh Understanding.

Geometric deep learning has sparked a rising interest in computer graphics to perform shape understanding tasks, such as shape classification and semantic segmentation. When the input is a polygonal surface, one has to suffer from the irregular mesh structure. Motivated by the geometric spectral theory, we introduce Laplacian2Mesh, a novel and flexible convolutional neural network (CNN) framework for coping with irregular triangle meshes (vertices may have any valence). By mapping the input mesh surface to the multi-dimensional Laplacian-Beltrami space, Laplacian2Mesh enables one to perform shape analysis tasks directly using the mature CNNs, without the need to deal with the irregular connectivity of the mesh structure. We further define a mesh pooling operation such that the receptive field of the network can be expanded while retaining the original vertex set as well as the connections between them. Besides, we introduce a channel-wise self-attention block to learn the individual importance of feature ingredients. Laplacian2Mesh not only decouples the geometry from the irregular connectivity of the mesh structure but also better captures the global features that are central to shape classification and segmentation. Extensive tests on various datasets demonstrate the effectiveness and efficiency of Laplacian2Mesh, particularly in terms of the capability of being vulnerable to noise to fulfill various learning tasks.

Predictors of malignant intussusception in adults using clinical manifestations and multidetector computed tomographic findings.

PURPOSE: To investigate the diagnostic performance of clinical manifestations and multidetector computed tomographic (MDCT) features in detecting predictors of malignant intussusception in adults. MATERIAL AND METHODS: We retrospectively reviewed 88 adults with 91 intussusceptions who were diagnosed by MDCT. Their clinical manifestations and MDCT features were reviewed and compared between the malignant and benign groups. Uni- and multivariate logistic regression analyses were used to identify independent predictors of malignant intussusception. RESULTS: There were 61 patients in the malignant group and 27 patients in the benign group. The malignant group had older age (mean, 62.61 vs 54.22 years, P = 0.014), more colon-related intussusception (89.06% vs 55.56%, P < 0.001), shorter intussusception length (median, 6.53 vs 9.73 cm, P = 0.009), higher maximum short axis diameter (mean, 4.85 vs 4.10 cm, P = 0.001), more enlarged lymph nodes (40.63% vs 11.11%, P = 0.006) than the benign group. Lead points were mainly presented as masses, which were irregular (44.74%) and lobular (28.95%) in the malignant group, and round or oval (92.00%) in the benign group. On the unenhanced MDCT, 90.62% of them in the malignant group showed non-hypodense. Multivariate analysis showed that intussusception length (P = 0.013), maximum short axis diameter (P = 0.007), non-round/oval lead point (P < 0.001) and non-hypodense lead point (P = 0.030) were independent factors of malignant intussusception. CONCLUSION: Malignant intussusception can be identified using independent predictors such as intussusception length, maximum short axis diameter, non-round/oval and non-hypodense lead point. When integrating these four factors, radiologists can make qualitative diagnoses withhigher sensitivity and specificity, allowing clinicians to develop more appropriate treatments.