A comparison of machine learning methods for radiomics modeling in prediction of occult lymph node metastasis in clinical stage IA lung adenocarcinoma patients.

Background: Accurate prediction of occult lymph node metastasis (ONM) is an important basis for determining whether lymph node (LN) dissection is necessary in clinical stage IA lung adenocarcinoma patients. The aim of this study is to determine the best machine learning algorithm for radiomics modeling and to compare the performances of the radiomics model, the clinical-radilogical model and the combined model incorporate both radiomics features and clinical-radilogical features in preoperatively predicting ONM in clinical stage IA lung adenocarcinoma patients. Methods: Patients with clinical stage IA lung adenocarcinoma undergoing curative surgery from one institution were retrospectively recruited and assigned to training and test cohorts. Radiomics features were extracted from the preoperative computed tomography (CT) images of the primary tumor. Seven machine learning algorithms were used to construct radiomics models, and the model with the best performance, evaluated using the area under the curve (AUC), was selected. Univariate and multivariate logistic regression analyses were performed on the clinical-radiological features to identify statistically significant features and to develop a clinical model. The optimal radiomics and clinical models were integrated to build a combined model, and its predictive performance was assessed using receiver operating characteristic curves, Brier score, and decision curve analysis (DCA). Results: This study included 258 patients who underwent resection (training cohort, n=182; test cohort, n=76). Six radiomics features were identified. Among the seven machine learning algorithms, extreme gradient boosting (XGB) demonstrated the highest performance for radiomics modeling, with an AUC of 0.917. The combined model improved the AUC to 0.933 and achieved a Brier score of 0.092. DCA revealed that the combined model had optimal clinical efficacy. Conclusions: The superior performance of the combined model, based on XGB algorithm in predicting ONM in patients with clinical stage IA lung adenocarcinoma, might aid surgeons in deciding whether to conduct mediastinal LN dissection and contribute to improve patients' prognosis.

Dual-energy CT-based radiomics in predicting EGFR mutation status non-invasively in lung adenocarcinoma.

Objectives: Patients with epidermal growth factor receptor (EGFR) mutations in lung adenocarcinoma (LUAD) can benefit from individualized targeted therapy. This study aims to develop, compare, analyse prediction models based on dual-energy spectral computed tomography (DESCT) and CT-based radiomic features to non-invasively predict EGFR mutation status in LUAD. Materials and methods: Patients with LUAD (n = 175), including 111 patients with and 64 patients without EGFR mutations, were enrolled in the current study. All patients were randomly divided into a training dataset (122 cases) and validation dataset (53 cases) at a ratio of 7:3. After extracting CT-based radiomic, DESCT and clinical features, we built seven prediction models and a nomogram of the best prediction. Receiver operating characteristic (ROC) curves and the mean area under the curve (AUC) values were used for comparisons among the models to obtain the best prediction model for predicting EGFR mutations. Results: The best distinguishing ability is the combined model incorporating radiomic, DESCT and clinical features for predicting the EGFR mutation status with an AUC of 0.86 (95 % CI: 0.79-0.92) in the training group and an AUC value of 0.83 (95 % CI: 0.73, 0.96) in the validation group. Conclusions: Our study provides a predictive nomogram non-invasively with a combination of CT-based radiomic, DESCT and clinical features, which can provide image-based biological information for targeted therapy of LUAD with EGFR mutations.

The Value of Deep Learning Image Reconstruction in Improving the Quality of Low-Dose Chest CT Images.

This study aimed to evaluate the value of the deep learning image reconstruction (DLIR) algorithm (GE Healthcare's TrueFidelity™) in improving the image quality of low-dose computed tomography (LDCT) of the chest. First, we retrospectively extracted raw data of chest LDCT from 50 patients and reconstructed them by using model-based adaptive statistical iterative reconstruction-Veo at 50% (ASIR-V 50%) and DLIR at medium and high strengths (DLIR-M and DLIR-H). Three sets of images were obtained. Next, two radiographers measured the mean CT value/image signal and standard deviation (SD) in Hounsfield units at the region of interest (ROI) and calculated the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR). Two radiologists subjectively evaluated the image quality using a 5-point Likert scale. The differences between the groups of data were analyzed through a repeated measures ANOVA or the Friedman test. Last, our result show that the three reconstructions did not differ significantly in signal (p > 0.05) but had significant differences in noise, SNR, and CNR (p < 0.001). The subjective scores significantly differed among the three reconstruction modalities in soft tissue (p < 0.001) but not in lung tissue (p > 0.05). DLIR-H had the best noise reduction ability and improved SNR and CNR without distorting the image texture, followed by DLIR-M and ASIR-V 50%. In summary, DLIR can provide a higher image quality at the same dose, enhancing the physicians' diagnostic confidence and improving the diagnostic efficacy of LDCT for lung cancer screening.

Clinical and imaging features of desmoid tumors of the extremities.

BACKGROUND: Desmoid fibroma is a rare soft tissue tumor originating from the aponeurosis, fascia, and muscle, and it is also known as aponeurotic fibroma, invasive fibroma, or ligamentous fibroma. AIM: To investigate the clinical and imaging features of desmoid tumors of the extremities. METHODS: Thirteen patients with desmoid fibroma of the extremities admitted to our hospital from October 2016 to March 2021 were included. All patients underwent computed tomography (CT), magnetic resonance imaging (MRI), and pathological examination of the lesion. Data on the diameter and distribution of the lesion, the relationship between the lesion morphology and surrounding structures, MRI and CT findings, and pathological features were statistically analyzed. RESULTS: The lesion diameter ranged from 1.7 to 8.9 cm, with an average of 5.35 ± 2.39 cm. All lesions were located in the deep muscular space, with the left and right forearm each accounting for 23.08% of cases. Among the 13 patients with desmoid fibroma of the extremities, the lesions were "patchy" in 1 case, irregular in 10, and quasi-round in 2. The boundary between the lesion and surrounding soft tissue was blurred in 10 cases, and the focus infiltrated along the tissue space and invaded the adjacent structures. Furthermore, the edge of the lesion showed "beard-like" infiltration in 2 cases; bone resorption and damage were found in 8, and bending of the bone was present in 2; the boundary of the focus was clear in 1. According to the MRI examination, the lesions were larger than 5 cm (61.54%), round or fusiform in shape (84.62%), had an unclear boundary (76.92%), showed uniform signal (69.23%), inhomogeneous enhancement (84.62%), and "root" or "claw" infiltration (69.23%). Neurovascular tract invasion was present in 30.77% of cases. CT examination showed that the desmoid tumors had slightly a lower density (69.23%), higher enhancement (61.54%), and unclear boundary (84.62%); a CT value < 50 Hu was present in 53.85% of lesions, and the enhancement was uneven in 53.85% of cases. Microscopically, fibroblasts and myofibroblasts were arranged in strands and bundles, without obvious atypia but with occasional karyotyping; cells were surrounded by collagen tissue. There were disparities in the proportion of collagen tissue in different regions, with abundant collagen tissue and few tumor cells in some areas, similar to the structure of aponeuroses or ligaments, and tumor cells invading the surrounding tissues. CONCLUSION: Desmoid tumors of the extremities have certain imaging features on CT and MRI. The two imaging techniques can be combined to improve the diagnostic accuracy, achieve a comprehensive diagnosis of the disease in the clinical practice, and reduce the risk of missed diagnosis or misdiagnosis. In addition, their use can ensure timely diagnosis and treatment.