ABSTRACT
PURPOSE: To assess the value of orthogonal axial images (OAI) of MRI in gastric cancer T staging. METHODS: This retrospective study enrolled 133 patients (median age, 63 [range, 24-85] years) with gastric adenocarcinoma who underwent both CT and MRI followed by surgery. MRI lacking or incorporating OAI and CT images were evaluated, respectively. Diagnostic performance (accuracy, sensitivity, and specificity) for each T stage, overall diagnostic accuracy and rates of over- and understaging were quantified employing pathological T stage as a reference standard. The McNemar's test was performed to compare the overall accuracy. RESULTS: Among patients with pT1-pT4 disease, MRI with OAI (accuracy: 88.7-94.7%, sensitivity: 66.7-93.0%, specificity: 91.5-100.0%) exhibited superior diagnostic performance compared to MRI without OAI (accuracy: 81.2-88.7%, sensitivity: 46.2-83.1%, specificity: 85.5-99.1%) and CT (accuracy: 88.0-92.5%, sensitivity: 53.3-90.1%, specificity: 88.7-98.1%). The overall accuracy of MRI with OAI was significantly higher (83.5%) than that of MRI without OAI (67.7%) (p < .001). However, there was no significant difference in the overall accuracy of MRI with OAI and CT (78.9%) (p = .35). The over- and understaging rates of MRI with OAI (12.0, 4.5%) were lower than those of MRI without OAI (21.8, 10.5%) and CT (12.8, 8.3%). CONCLUSION: OAI play a pivotal role in the T staging of gastric cancer. MRI incorporating OAI demonstrated commendable performance for gastric cancer T-staging, with a slight tendency toward its superiority over CT.
ABSTRACT
Purpose: The detection of human epidermal growth factor receptor 2 (HER2) expression status is essential to determining the chemotherapy regimen for breast cancer patients and to improving their prognosis. We developed a deep learning radiomics (DLR) model combining time-frequency domain features of ultrasound (US) video of breast lesions with clinical parameters for predicting HER2 expression status. Patients and Methods: Data for this research was obtained from 807 breast cancer patients who visited from February 2019 to July 2020. Ultimately, 445 patients were included in the study. Pre-operative breast ultrasound examination videos were collected and split into a training set and a test set. Building a training set of DLR models combining time-frequency domain features and clinical features of ultrasound video of breast lesions based on the training set data to predict HER2 expression status. Test the performance of the model using test set data. The final models integrated with different classifiers are compared, and the best performing model is finally selected. Results: The best diagnostic performance in predicting HER2 expression status is provided by an Extreme Gradient Boosting (XGBoost)-based time-frequency domain feature classifier combined with a logistic regression (LR)-based clinical parameter classifier of clinical parameters combined DLR, particularly with a high specificity of 0.917. The area under the receiver operating characteristic curve (AUC) for the test cohort was 0.810. Conclusion: Our study provides a non-invasive imaging biomarker to predict HER2 expression status in breast cancer patients.
Subject(s)
Breast Neoplasms , Deep Learning , Humans , Female , Breast Neoplasms/diagnostic imaging , ROC CurveABSTRACT
ObjectiveTo investigate the capacity and diagnostic value of various MRI seguences in patients with trigeminal neuralgia.Methods MR images of 60 patients with trigeminal neuralgia were analyzed retrospectively.The sensitivity,specificity and diagnostic value of various MRI sequences were evaluated comparing with clinical data and operation results.All patients were scanned with conventional sequences including SE-T1WI,T2WI,FLAIR of head.Among them,9 cases were injected with contrast agent,49 cases were scanned with 3D-TOF and 3D-TSE sequences on cerebellopontine angle additinally.The sensitivity,specificity and accuracy of the 3D-TOF and 3D-TSE sequences were analyzed by using the x2 test.Results Six cases with tumor,3 cases with radiculitis and meningitis,1 case with multiple sclerosis and 1 case with pons infarction were diagnosed by conventional MR sequences.MRI of 49 cases with 3D-TOF and 3D-TSE showed neurovascular compress proved by operation.The sensitivity,specificity and accuracy of 3D-TOF and 3D-TSE were 95.3％ (41/43)vs.95.6％ (43/45),66.7％ (4/6)vs.50.0％(2/4) and 91.8％ (45/49)vs.91.8％ (45/49),and no significantly difference was found between the two sequences (x2 =0.13,0.19 and 0.17,P ＞0.5).ConclusionsMRI plays an important role in displaying the causes and diagnosis of trigeminal neuralgia.The conventional head MR sequences should be recommended for diagnosis of secondary TN caused by tumor and inflammation et al,and further 3D-TOF and 3D-TSE with high sensitivity and accuracy should be applied in patients with primary TN to display the neurovascular relationship.
