Background: In the past few years, U-Net based U-shaped architecture and skip-connections have made incredible progress in the field of medical image segmentation. U2-Net achieves good performance in computer vision. However, in the medical image segmentation task, U2-Net with over nesting is easy to overfit. Purpose: A 2D network structure TransU2-Net combining transformer and a lighter weight U2-Net is proposed for automatic segmentation of brain tumor magnetic resonance image (MRI). Methods: The light-weight U2-Net architecture not only obtains multi-scale information but also reduces redundant feature extraction. Meanwhile, the transformer block embedded in the stacked convolutional layer obtains more global information; the transformer with skip-connection enhances spatial domain information representation. A new multi-scale feature map fusion strategy as a postprocessing method was proposed for better fusing high and low-dimensional spatial information. Results: Our proposed model TransU2-Net achieves better segmentation results, on the BraTS2021 dataset, our method achieves an average dice coefficient of 88.17%; Evaluation on the publicly available MSD dataset, we perform tumor evaluation, we achieve a dice coefficient of 74.69%; in addition to comparing the TransU2-Net results are compared with previously proposed 2D segmentation methods. Conclusions: We propose an automatic medical image segmentation method combining transformers and U2-Net, which has good performance and is of clinical importance. The experimental results show that the proposed method outperforms other 2D medical image segmentation methods.

Background: Chest computed tomography (CT) is a critical tool in the diagnosis of pulmonary cryptococcosis as approximately 30% of normal immunity individuals may not exhibit any significant symptoms or laboratory findings. Pulmonary cryptococcosis granuloma and lung adenocarcinoma can appear similar on noncontrast chest CT. This study evaluates the use of an integrated model that was developed based on radiomic features combined with demographic and radiological features to differentiate pulmonary cryptococcosis nodules from lung adenocarcinomas. Methods: Preoperative chest CT images for 215 patients with solid pulmonary nodules with histopathologically confirmed lung adenocarcinoma and cryptococcosis infection were collected from two clinical centers (108 cases in the training set and 107 cases in the test set divided by the different hospitals). Radiomics models were constructed based on nodular lesion volume (LV), 5-mm extended lesion volume (ELV), and perilesion volume (PLV). A demoradiological model was constructed using logistic regression based on demographic information (age, sex) and 12 radiological features (location, number, shape and specific imaging signs). Both models were used to build an integrated model, the performance of which was assessed using the test set. A junior and a senior radiologist evaluated the nodules. Receiver operating characteristic (ROC) curve analysis was conducted, and areas under the curve (AUCs), sensitivity (SEN), and specificity (SPE) of the models were calculated and compared. Results: Among the radiomics models, AUCs of the LV, ELV, and PLV were 0.558, 0.757, and 0.470, respectively. Age, lesion number, and lobular sign were identified as independent discriminative features providing an AUC of 0.77 in the demoradiological model (SEN 0.815, SPE 0.642). The integrated model achieved the highest AUC of 0.801 (SEN 0.759, SPE 0.755), which was significantly higher than that obtained by a junior radiologist (AUC =0.689, P=0.024) but showed no significant difference from that of the senior radiologist (AUC =0.784, P=0.388). Conclusions: An integrated model with radiomics and demoradiological features improves discrimination of cryptococcosis granulomas from solid adenocarcinomas on noncontrast CT. This model may be an effective strategy for machine complementation to discrimination by radiologists, and whole-lung automated recognition methods might dominate in the future.

Background: It is difficult for radiologists, especially junior radiologists with limited experience to make differential diagnoses between mediastinal lymphomas and thymic epithelial tumors (TETs) due to the overlapping imaging features. The purpose of this study was to develop and validate a CT-based clinico-radiomics model for differentiating lymphomas from TETs and to investigate whether a human-machine hybrid system can assist junior radiologists in improving their diagnostic performance. Methods: The patients who underwent contrast-enhanced chest CT and pathologically confirmed with lymphoma or TET at two centers from January 2011 to December 2019 and from January 2017 to December 2021 were retrospectively included and split as training/validation set and external test set, respectively. Clinical and radiomic signatures were pre-selected by elastic-net, and the models were established with the selected signatures using ensemble learning. Three radiologists independently reviewed CT images and assessed each case of the external test set with knowledge of the relevant clinical information. The diagnoses of reader 1, reader 2, and reader 3 were compared with those of the models in the external test set and further separately input to the model’s ensemble process as a human-machine system to make final decisions in the external test set. The improvement of diagnostic performance of radiologists by human-machine system was evaluated by the area under the receiver operating characteristic curve and increase rate. Results: A total of 95 patients (51 with lymphomas and 44 with TETs) at Center 1 and 94 (52 with lymphomas and 42 with TETs) at Center 2 were enrolled and divided into training/validation sets and external test set, respectively. The diagnostic performance of the clinico-radiomics model has outperformed the junior radiologists and senior radiologist in AUC (clinico-radiomics model: 0.85 (0.76,0.92); reader 2: 0.70 (0.60,0.80); reader 3: 0.60 (0.49,0.71), reader 1: 0.76 (0.66,0.86), respectively) in the external test set. The human-machine hybrid system demonstrated significant increases in AUC (reader 1 + model: 0.87 (0.79,0.94), an increase of 14%; reader 2 + model: 0.86 (0.77,0.93), an increase of 23%; reader 3 + model: 0.84 (0.76,0.91), an increase of 40%), compared to the human performance alone. Conclusions: The clinico-radiomics model outperformed three radiologists in differentiating lymphomas from TETs on CT. The use of the human-machine hybrid system significantly improved the performance of radiologists, especially junior radiologists. It provides a real-time decision tool to reduce bias and mistakes in radiologist diagnosis and enhances the diagnostic confidence of junior radiologists. This attempt may lead to more human-machine hybrid systems being explored in the diagnosis of different diseases to drive future clinical applications.

Background: Lung adenocarcinoma associated with cystic airspace (LACA) was once considered an uncommon manifestation of lung adenocarcinoma (LUAD), and understandings of it are limited; however, it is being observed more frequently in clinical practice. This study sought to assess the prevalence of LACA, and compare the high-resolution computed tomography (HRCT) features of LACA in patients with varying degrees of invasiveness. Methods: This study retrospectively reviewed the HRCT scans of 1,525 patients with LUAD ≤3 cm in diameter at the Shanghai Chest Hospital between January 2016 and May 2016. Each nodule was examined to detect the presence of cystic airspace. Additionally, we analyzed the qualitative HRCT findings of the cystic airspaces, including the pattern, number, wall component density, distribution, inner surface, mural nodules, septa, and vessels passing through the cystic airspace using the Pearson χ² test or Fisher’s exact test as appropriate. We also analyzed the quantitative measurements, such as the cystic airspace diameter, wall thickness, and thin-wall proportion, using a one-way analysis of variance or the Kruskal-Wallis rank-sum test as appropriate. Results: LACAs were observed on HRCT in 11.5% (176/1,525) of the patients, of whom 7.1% (36/505) had pure ground-glass nodules, 13.5% (112/830) had mixed ground-glass nodules, and 14.7% (28/190) had solid nodules (P=0.001). The surgical procedures for LACAs varied (P=0.012). The incidence of LACAs increased as nodule diameter and invasiveness increased (both P<0.001). Statistically significant differences were observed in the wall component density, distribution, septa, vessels passing through the cystic airspace, cystic airspace diameter, wall thickness, and thin-wall proportion among the preinvasive lesion (PL), minimally invasive adenocarcinoma (MIA), and invasive adenocarcinoma (IAC) groups (P<0.001, P=0.024, P=0.001, P=0.025, P=0.001, P<0.001, and P<0.001, respectively). Wall component density increased as invasiveness increased (P<0.001). Unlike those in the MIAs and IACs, cystic airspaces in PLs typically lacked septa (P=0.001, and P<0.001, respectively). The IACs had larger cystic airspace diameters than the PLs (6.5 vs. 3.7 mm) (P<0.001). The IACs also had thicker wall thickness (11.8 vs. 6.8 mm, 11.8 vs. 8.3 mm) (P<0.001, and P<0.001, respectively) and smaller thin-wall proportions (181.5° vs. 264.8°, 181.5° vs. 223.8°) (P<0.001, and P=0.039, respectively) than the PLs and MIAs. Conclusions: The prevalence and characteristics of cystic airspaces on HRCT can be used to predict invasiveness in patients with LUADs ≤3 cm in diameter.