Accurate prediction of recurrence risk is essential to devise effective and personalized treatment strategies for patients with soft tissue sarcoma (STS). This study aimed to develop and validate a multimodal deep learning framework that integrates clinical features, preoperative MR images, and hematoxylin and eosin-stained whole slide images (WSIs) to predict recurrence in patients with STS. A total of 323 patients with STS were retrospectively enrolled from two hospitals, serving as development and validation sets, respectively. The ShuffleNetV2 network was utilized to develop patch-level and WSI-level signatures. A convolutional neural network fusing the channel and spatial attention mechanisms was used to develop a radiology signature. The combined model was built by integrating clinical features, radiology signature score, and WSI-level signature score with Cox regression analysis. The combined model demonstrated superior performance in the validation set, achieving a C-index of 0.857 and a time-dependent area under the curve of 0.959. Class activation maps facilitated the monitoring of suspected regions to inform recurrence decisions. The recurrence-free survival times of the low- and high-risk cohorts were statistically different (p < 0.05). The proposed multimodal framework offers satisfactory accuracy for predicting recurrence risk in patients with STS and could guide the choice of treatment modality.

To develop and validate an MRI-derived radiomics model for the prediction of intratumoral tertiary lymphoid structures (TLSs) status of soft tissue sarcoma (STS) and explore its prognostic value. This study retrospectively included 302 patients of three cohorts who underwent surgical resection of STS from two medical centers. Radiomics features were derived for both intratumoral and peritumoral regions from preoperative axial fat-suppressed T2-weighted and T1-weighted imaging sequences. Intratumoral, peritumoral, and combined radiomics models were constructed using a logistic regression algorithm. The area under the receiver operator characteristic curve (AUC) and the DeLong test were utilized to assess and compare the performances of three radiomics models. By applying a linear combination of the chosen features, the Rad-score for the optimal radiomics model was computed. TLS positivity was identified in 114 (38

ABSTRACTBackgroundWith the elevated level of NAFLD prevalence, the incidence of diabetes, hypertension, metabolic syndrome and other diseases is also significantly elevated. GLP‐1RA can exert weight loss, glucose‐lowering effects and various nonglycaemic effects. However, the relationship between quantitative reduction in hepatic fat content and improvement of pancreatic islet function by GLP‐1RA is unclear.MethodsThis trial was a single‐arm open cohort study. A total of 38 patients with T2DM and NAFLD were enrolled in the GLP‐1RA treatment group. The included patients were tested for biochemical and blood glucose levels, adiponectin and FGF21 levels, and liver fat content was measured using MRI. Measure the above indicators again after at least 3 months of GLP‐1RA treatment. Divided into Q1 (average decrease of 0.37%) and Q2 (average decrease of 8.6%) groups based on the degree of reduction in liver fat content.ResultsQ2 group showed an average reduction in liver fat content of 8.6%, a decrease in glycated haemoglobin of 18.17%, a weight loss of 7.29% and an increase in fasting c‐peptide release by 1.03%, 1‐h and 2‐h postprandial c‐peptide release by 28.86% and 18.28% respectively. In contrast, Q1 group had an average reduction in liver fat content of 0.37%, a decrease in glycated haemoglobin of only 6.53%, a weight loss of 3.41%, a decrease in fasting c‐peptide release by 1.91% and an increase in 1‐h and 2‐h postprandial c‐peptide release by 19.18% and 11.66% respectively.ConclusionReduction in liver fat content effectively improves pancreatic islet function secretion, particularly postprandial c‐peptide secretion, especially in the first hour after a meal. This improvement leads to a decrease in glycated haemoglobin levels and promotes better compliance with blood glucose control.

Background: The use of histogram analysis of computed tomography (CT) values is a potential method for differentiating between benign osteoblastic lesions (BOLs) and malignant osteoblastic lesions (MOLs). Purpose: To explore the diagnostic efficacy of histogram analysis in accurately distinguishing between BOLs and MOLs based on CT values. Material and Methods: A total of 25 BOLs and 25 MOLs, which were confirmed through pathology or imaging followup, were included in this study. FireVoxel software was used to process the lesions and obtain various histogram parameters, including mean value, standard deviation, variance, coefficient of variation, skewness, kurtosis, entropy value, and percentiles ranging from 1st to 99th. Statistical tests, such as two independent-sample t-tests and the MannWhitney U test with Bonferroni correction, were employed to compare the differences in histogram parameters between BOLs and MOLs. A receiver operating characteristic (ROC) curve analysis was performed to evaluate the diagnostic efficacy of each parameter. Results: Significant differences were observed in several histogram parameters between BOLs and MOLs, including the mean value, coefficient of variation, skewness, and various percentiles. Notably, the 25th percentile demonstrated the highest diagnostic efficacy, as indicated by the largest area under the curve in the ROC curve analysis. Conclusion: Histogram analysis of CT values provides valuable diagnostic information for accurately differentiating between BOLs and MOLs. Among the different parameters, the 25th percentile parameter proves to be the most effective in this discrimination process.

Background To explore the potential of different quantitative dynamic contrast-enhanced (qDCE)-MRI tracer kinetic (TK) models and qDCE parameters in discriminating benign from malignant soft tissue tumors (STTs).Methods This research included 92 patients (41females, 51 males; age range 16-86 years, mean age 51.24 years) with STTs. The qDCE parameters (Ktrans, Kep, Ve, Vp, F, PS, MTT and E) for regions of interest of STTs were estimated by using the following TK models: Tofts (TOFTS), Extended Tofts (EXTOFTS), adiabatic tissue homogeneity (ATH), conventional compartmental (CC), and distributed parameter (DP). We established a comprehensive model combining the morphologic features, time-signal intensity curve shape, and optimal qDCE parameters. The capacities to identify benign and malignant STTs was evaluated using the area under the curve (AUC), degree of accuracy, and the analysis of the decision curve.Results TOFTS-Ktrans, EXTOFTS-Ktrans, EXTOFTS-Vp, CC-Vp and DP-Vp demonstrated good diagnostic performance among the qDCE parameters. Compared with the other TK models, the DP model has a higher AUC and a greater level of accuracy. The comprehensive model (AUC, 0.936, 0.884-0.988) demonstrated superiority in discriminating benign and malignant STTs, outperforming the qDCE models (AUC, 0.899-0.915) and the traditional imaging model (AUC, 0.802, 0.712-0.891) alone.Conclusions Various TK models successfully distinguish benign from malignant STTs. The comprehensive model is a noninvasive approach incorporating morphological imaging aspects and qDCE parameters, and shows significant potential for further development.

Objective: This study aimed to investigate the selected anatomical factors that can potentially influence temporomandibular joint (TMJ) clicking in young adults by assessing TMJ structures and lateral pterygoid muscle (LPM) function using magnetic resonance imaging (MRI).Methods: The patients were divided into four groups: the healthy control group; the clicking on mouth opening group; the clicking on mouth closing group; and the clicking on mouth opening and closing group. Additionally, we used clinical palpation to evaluate the masticatory muscles' functional state and employed MRI using the OCOR-T1WI-FSE-CLOSED, OSAG-PDW-FSE-CLOSED, and OSAG-PDW-FSE-OPEN sequences to analyze the texture of the lateral pterygoid muscle (LPM).Results: The proportion of any articular disc or condylar morphology class did not differ significantly between the TMJ clicking and HC groups. The articular disc position did not differ significantly between the TMJ clicking and HC groups. In the TMJ clicking group, the presence of masticatory muscle dysfunction differed significantly between the clicking and non-clicking sides. Moreover, the LPM accounted for the highest proportion among masticatory muscles with tenderness in all TMJ clicking subgroups (77.78%–100%). Therefore, in the TMJ clicking group, the LPM texture was less defined, more uniform in gray scale, and more similar to local texture (p < 0.0001).Conclusion: The occurrence of TMJ clicking in young adults is unrelated to the TMJ structure but related to the function of masticatory muscles, particularly the LPM.

[This corrects the article DOI: 10.3389/fpsyt.2023.1197987.].

Objectives: To image knee osteochondral specimens using magnetic resonance (MR) ultrashort echo time imaging with pointwise encoding time reduction with radial acquisition combined fat suppression (PETRA-FS) sequence to determine whether it can reveal non-calcified cartilage, including the deep radial layer, and to assess its effectiveness in cartilage damage diagnosis. Materials and methods: PETRA-FS imaging was performed on 58 osteochondral specimens of the lower femur and upper tibia to observe depth of cartilage damage, combined with histological results to observe signal intensity composition. Sensitivity, specificity, and reliability of PETRA-FS sequence for diagnosing cartilage damage were evaluated using histological results as the gold standard. Diagnostic efficacy was assessed using receiver operating characteristic (ROC) curve. Results: MR ultrashort echo time imaging PETRA-FS sequence showed non-calcified cartilage, including tangential, transitional, and radial layers, which showed a high signal. PETRA-FS sequence showed 37 cases of cartilage damage and 21 cases of no damage among 58 specimens, kappa value of 0.75. Histological analysis of the 58 osteochondral specimens revealed 38 cases of cartilage injury and 20 cases of undamaged cartilage. Using histological results as the gold standard, PETRA-FS sequence had a sensitivity of 87.00%, specificity of 80.00%, kappa value of 0.81, and an area under the ROC curve (AUC) of 0.83 for cartilage injury diagnosis. Conclusion: MR ultrashort echo time imaging PETRA-FS sequence can show non-calcified cartilage, including the deep radial layer (which cannot be shown by conventional MR), by exhibiting a high signal in knee osteo-chondral specimens. Thus, PETRA-FS sequences may have important diagnostic value for cartilage injury diagnosis.

Background Apolipoprotein E (APOE) ε2 and APOE ε4 are the most distinct alleles among the three APOE alleles, both structurally and functionally. However, differences in cognition, brain function, and brain structure between the two alleles have not been comprehensively reported in the literature, especially in non-demented elderly individuals. Methods A neuropsychological test battery was used to evaluate the differences in cognitive performance in five cognitive domains. Independent component analysis (ICA) and voxel-based morphometry (VBM) were used separately to analyze resting-state functional magnetic resonance imaging (rs-fMRI) data and the structure MRI data between the two groups. Finally, correlations between differential brain regions and neuropsychological tests were calculated. Results APOE ε2 carriers had better cognitive performance in general cognitive, memory, attention, and executive function than APOE ε4 carriers (all p < 0.05). In ICA analyses of rs-fMRI data, the difference in the resting-state functional connectivity (rsFC) between two groups is shown in 7 brain networks. In addition, VBM analyses of the T1-weighted image revealed that APOE ε2 carriers had a larger thalamus and right postcentral gyrus volume and a smaller bilateral putamen volume than APOE ε4 carriers. Finally, differences in brain function and structure may be the reason that APOE ε2 carriers are better than APOE ε4 carriers in cognitive performance. Conclusion These findings suggest that there are significant differences in brain function and structure between APOE ε2 carriers and APOE ε4 carriers, and these significant differences are closely related to their cognitive performance.

Postpartum osteoporosis (PO) is a rare condition characterized by low bone mineral density (BMD) and an increased risk of vertebral fragility fracture. We encountered a 34-year-old woman who developed back pain 1 week after delivery. Magnetic resonance imaging of the lumbar spine revealed three vertebral compression fractures. Pretreatment BMD evaluation by dual-energy X-ray absorptiometry revealed a low T-score and Z-score (−2.0 and −2.0, respectively; BMD, 0.876 g/cm 2 ) in the affected region of the spine. The patient was diagnosed with PO and treated with subcutaneous injection of denosumab 60 mg (Prolia; Amgen, Inc., Thousand Oaks, CA, USA) every 6 months. After two treatments, the BMD had significantly increased and the back pain was improved; the patient therefore decided to terminate the treatment. Two months later, her back pain worsened and BMD decreased as measured by dual-energy X-ray absorptiometry examination of the lumbar spine. Therefore, the patient resumed treatment with denosumab, and the BMD of the lumbar spine increased after another two treatments. Therefore, we consider denosumab to be promising in the management of PO with respect to increased BMD and decreased pain.

Background Conventional MR pulse sequences result in poor signal from low T2 cortical bone because the minimum achievable echo time is limited. A sequence resulting in improved bone contrast is desirable. Purpose To evaluate the image quality and diagnostic performance of grayscale inversion zero echo time imaging (GI‐ZTE) and grayscale inversion T1‐weighted imaging (GI‐T1WI) compared with computed tomography (CT). Study Type Prospective. Subjects A total of 50 patients with musculoskeletal tumors or tumor‐like diseases of the lower extremities having MRI and CT studies. Field Strength/Sequence GI‐T1WI and GI‐ZTE sequences at 1.5 T. Assessment Assessed cortical and medullary bone morphology abnormalities using CT as the reference standard. Three radiologists scored the images quality and recorded nine metrics to assess the diagnostic performance. Statistical Tests Differences in image quality were calculated using the Wilcoxon signed‐rank test. The intraclass correlation coefficient (ICC) was used to analyze the agreement of quantitative lesion parameters between CT and MR sequences, as well as the interobserver reliability. A P value <0.05 was considered statistically significant. Results Image quality score was significantly higher for CT images than GI‐TIWI images. Except for radiologist 3 [4(0) vs 4 (1)], there was no significant difference in scores between CT and GI‐ZTE [radiologist 1: 4 (0) vs 4 (0), P = 0.133; radiologist 2: 4 (0) vs 4 (0), P = 0.085]. There was good–excellent agreement between both MR sequences and CT for size, lesion number, location, sclerotic rim, expanded shell, destruction pattern, and matrix mineralization for all radiologists (ICC: 0.636–1.000). The consistency of periosteal reaction and penetration of the cortex was fair to good (0.481–0.729) between GI‐T1WI and CT and good to excellent between GI‐ZTE and CT (0.682–0.852). Data Conclusions GI‐ZTE images had superior intermodality agreement with CT images and allowed visualization of more cortical bone detail than GI‐T1WI images. Evidence Level 1. Technical Efficacy Stage 2.

RATIONALE AND OBJECTIVES:Accurate pretreatment assessment of histological differentiation grade of head and neck squamous cell carcinoma (HNSCC) is crucial for prognosis evaluation. This study aimed to construct and validate a contrast-enhanced computed tomography (CECT)-based deep learning radiomics nomogram (DLRN) to predict histological differentiation grades of HNSCC.MATERIALS AND METHODS:A total of 204 patients with HNSCC who underwent CECT scans were enrolled in this study. The participants recruited from two hospitals were split into a training set (n=124, 74 well/moderately differentiated and 50 poorly differentiated) of patients from one hospital and an external test set of patients from the other hospital (n=80, 49 well/moderately differentiated and 31 poorly differentiated). CECT-based manually-extracted radiomics (MER) features and deep learning (DL) features were extracted and selected. The selected MER features and DL features were then combined to construct a DLRN via multivariate logistic regression. The predictive performance of the DLRN was assessed using ROCs and decision curve analysis (DCA).RESULTS:Three MER features and seven DL features were finally selected. The DLRN incorporating the selected MER and DL features showed good predictive value for the histological differentiation grades of HNSCC (well/moderately differentiated vs. poorly differentiated) in both the training (AUC, 0.878) and test (AUC, 0.822) sets. DCA demonstrated that the DLRN was clinically useful for predicting histological differentiation grades of HNSCC.CONCLUSION:A CECT-based DLRN was constructed to predict histological differentiation grades of HNSCC. The DLRN showed good predictive efficacy and might be useful for prognostic evaluation of patients with HNSCC.

BACKGROUND AND PURPOSE:The Stage, Size, Grade and Necrosis (SSIGN) score is the most commonly used prognostic model in clear cell renal cell carcinoma (ccRCC) patients. It is a great challenge to preoperatively predict SSIGN score and outcome of ccRCC patients. The aim of this study was to develop and validate a CT-based deep learning radiomics model (DLRM) for predicting SSIGN score and outcome in localized ccRCC.METHODS:A multicenter 784 (training cohort/ test 1 cohort / test 2 cohort, 475/204/105) localized ccRCC patients were enrolled. Radiomics signature (RS), deep learning signature (DLS), and DLRM incorporating radiomics and deep learning features were developed for predicting SSIGN score. Model performance was evaluated with area under the receiver operating characteristic curve (AUC). Kaplan-Meier survival analysis was used to assess the association of the model-predicted SSIGN with cancer-specific survival (CSS). Harrell's concordance index (C-index) was calculated to assess the CSS predictive accuracy of these models.RESULTS:The DLRM achieved higher micro-average/macro-average AUCs (0.913/0.850, and 0.969/0.942, respectively in test 1 cohort and test 2 cohort) than the RS and DLS did for the prediction of SSIGN score. The CSS showed significant differences among the DLRM-predicted risk groups. The DLRM achieved higher C-indices (0.827 and 0.824, respectively in test 1 cohort and test 2 cohort) than the RS and DLS did in predicting CSS for localized ccRCC patients.CONCLUSION:The DLRM can accurately predict the SSIGN score and outcome in localized ccRCC.

The relationship between progesterone (P) and diabetic nephropathy (DKD) is unclear. Herein, we investigated the relationship between progesterone and DKD in men and postmenopausal women with type 2 diabetes mellitus.

目的 探讨鹅足滑囊影像学解剖及滑囊炎的MRI表现特点.方法 选取经影像及临床综合诊断为鹅足滑囊炎的83例患者,分析鹅足滑囊影像学解剖及总结鹅足滑囊炎MR特点.结果 鹅足滑囊位于内侧副韧带止点与鹅足腱之间.鹅足滑囊炎表现为滑囊内液体积聚,滑膜增厚,边界清(53/83)或欠清(30/83).其中1例伴有邻近骨质凹陷,边缘硬化;1例因外伤致滑囊破裂.鹅足滑囊积液时,形态呈条片状(43/83)、扁椭圆形(2/83)及分叶状(38/83).增厚的滑膜呈条状(67/83)、条状加结节状(5/83)及结节状(11/83).结论 鹅足滑囊位于内侧副韧带止点与鹅足腱之间.鹅足滑囊炎表现为滑囊内液体集聚,滑膜增厚;慢性鹅足滑囊炎可伴有邻近胫骨压迫性骨质吸收.

Objective. We aimed to evaluate the effectiveness of clinical examination combined with texture analysis of magnetic resonance imaging (MRI) and fasciculation patterns of the lateral pterygoid muscle (LPM) in distinguishing among the different anatomic causes of temporomandibular disorder.Methods. We divided the patients into four groups: healthy control (HC), disk without displacement (DWoD), disk displacement with reduction (DDWR), and disk displacement without reduction (DDWoR). Demographic information and clinical symptoms of patients in each group were recorded. LPM textures were compared among groups. LPM fasciculation was examined. P<0.05 indicated significant difference.Results. Several clinical symptoms and signs, but not age or sex, differed significantly among groups. Oblique sagittal planar MRI revealed significant differences in the parameters of Angular Second Moment, Contrast, Correlation, Inverse Difference Moment, and Entropy between the healthy controls and the 3 patient groups. MRI of the patients, both without and with disk displacement, demonstrated relative uniformity in gray distribution and correlation of gray values, with greater complexity but an unclear texture and no obvious regularity. The proportion of type B LPM fascicles was significantly higher in the DDWR and DDWoR groupsConclusion. Temporomandibular disorder, without and with disk displacement, is associated with clinical symptoms and texture analysis values that differ from healthy muscle. The types of LPM fascicles are related to the position of the articular disk. (Oral Surg Oral Med Oral Pathol Oral Radiol 2023;136:382-393)

Objectives To non-invasively predict the coexistence of isocitrate dehydrogenase (IDH) mutation and O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation in adult-type diffuse gliomas using apparent diffusion coefficient (ADC) histogram and direct ADC measurements and compare the diagnostic performances of the two methods. Materials and methods A total of 118 patients with adult-type diffuse glioma who underwent preoperative brain magnetic resonance imaging (MRI) and diffusion weighted imaging (DWI) were included in this retrospective study. The patient group included 40 patients with coexisting IDH mutation and MGMT promoter methylation (IDHmut/MGMTmet) and 78 patients with other molecular status, including 32 patients with IDH wildtype and MGMT promoter methylation (IDHwt/MGMTmet), one patient with IDH mutation and unmethylated MGMT promoter (IDHmut/MGMTunmet), and 45 patients with IDH wildtype and unmethylated MGMT promoter (IDHwt/MGMTunmet). ADC histogram parameters of gliomas were extracted by delineating the region of interest (ROI) in solid components of tumors. The minimum and mean ADC of direct ADC measurements were calculated by placing three rounded or elliptic ROIs in solid components of gliomas. Receiver operating characteristic (ROC) curve analysis and the area under the curve (AUC) were used to evaluate the diagnostic performances of the two methods. Results The 10th percentile, median, mean, root mean squared, 90th percentile, skewness, kurtosis, and minimum of ADC histogram analysis and minimum and mean ADC of direct measurements were significantly different between IDHmut/MGMTmet and the other glioma group (P < 0.001 to P = 0.003). In terms of single factors, 10th percentile of ADC histogram analysis had the best diagnostic efficiency (AUC = 0.860), followed by mean ADC obtained by direct measurements (AUC = 0.844). The logistic regression model combining ADC histogram parameters and direct measurements had the best diagnostic efficiency (AUC = 0.938), followed by the logistic regression model combining the ADC histogram parameters with statistically significant difference (AUC = 0.916) and the logistic regression model combining minimum ADC and mean ADC (AUC = 0.851). Conclusion Both ADC histogram analysis and direct measurements have potential value in predicting the coexistence of IDHmut and MGMTmet in adult-type diffuse glioma. The diagnostic performance of ADC histogram analysis was better than that of direct ADC measurements. The combination of the two methods showed the best diagnostic performance.

Background The purpose of this study was to investigate and validate multiparametric magnetic resonance imaging (MRI)-based machine learning classifiers for early identification of poor responders after neoadjuvant chemoradiotherapy (nCRT) in patients with locally advanced rectal cancer (LARC). Methods Patients with LARC who underwent nCRT were included in this retrospective study (207 patients). After preprocessing of multiparametric MRI, radiomics features were extracted and four feature selection methods were used to select robust features. The selected features were used to build five machine learning classifiers, and 20 (four feature selection methods × five machine learning classifiers) predictive models for the screening of poor responders were constructed. The predictive models were evaluated according to the area under the curve (AUC), F1 score, accuracy, sensitivity, and specificity. Results Eighty percent of all predictive models constructed achieved an AUC of more than 0.70. A predictive model using a support vector machine classifier with the minimum redundancy maximum relevance (mRMR) selection method followed by the least absolute shrinkage and selection operator (LASSO) selection method showed superior prediction performance, with an AUC of 0.923, an F1 score of 88.14%, and accuracy of 91.03%. The predictive performance of the constructed models was not improved by ComBat compensation. Conclusions In rectal cancer patients who underwent neoadjuvant chemoradiotherapy, machine learning classifiers with radiomics features extracted from multiparametric MRI were able to accurately discriminate poor responders from good responders. The techniques should provide additional information to guide patient-tailored treatment.

Purpose: To determine the accuracy, repeatability, and reproducibility of magnetic resonance imaging-based proton density fat fraction (MRI-PDFF) measurements of rotator cuff muscles between two readers and three different scanners.Methods: Twenty-seven volunteers underwent serial shoulder MRI examinations of both left and right sides on one 1.5-T MRI scanner and two 3.0-T MRI scanners. Two independent readers measured muscular PDFF of the supraspinatus, infraspinatus/teres minor muscle, and subscapularis. MR spectroscopy-based proton density fat fraction (MRS-PDFF) was regarded as the reference standard for assessing accuracy. A "coffee break" examination method was used to test the repeatability of each scanner. Bland-Altman plots, Pearson correlation, and linear regression analysis were used to assess bias and linearity. The Wilcoxon signed-rank test and Friedman test were applied to evaluate repeatability and reproducibility.Results: MRI-PDFF measurements indicated strong linearity (R2 = 0.749) and small bias (-0.18%) in comparison with the MRS-PDFF measurements. A very strong positive Pearson correlation (r = 0.955-0.986) between the PDFF estimates of the two repeat scans indicated excellent repeatability. The PDFF measurements showed high reproducibility, with a strong positive Pearson correlation (r = 0.668-0.698) and a small mean bias (-0.04 to -0.10%) across different scanners.Conclusion: MRI-PDFF measurements of rotator cuff muscles were highly accurate, repeatable, and reproducible across different readers and scanners, leading us to the conclusion that PDFF can be a reliable and robust quantitative imaging biomarker for longitudinal or multi-center studies.