Exploring deep learning radiomics for classifying osteoporotic vertebral fractures in X-ray images.

Purpose: To develop and validate a deep learning radiomics (DLR) model that uses X-ray images to predict the classification of osteoporotic vertebral fractures (OVFs). Material and methods: The study encompassed a cohort of 942 patients, involving examinations of 1076 vertebrae through X-ray, CT, and MRI across three distinct hospitals. The OVFs were categorized as class 0, 1, or 2 based on the Assessment System of Thoracolumbar Osteoporotic Fracture. The dataset was divided randomly into four distinct subsets: a training set comprising 712 samples, an internal validation set with 178 samples, an external validation set containing 111 samples, and a prospective validation set consisting of 75 samples. The ResNet-50 architectural model was used to implement deep transfer learning (DTL), undergoing -pre-training separately on the RadImageNet and ImageNet datasets. Features from DTL and radiomics were extracted and integrated using X-ray images. The optimal fusion feature model was identified through least absolute shrinkage and selection operator logistic regression. Evaluation of the predictive capabilities for OVFs classification involved eight machine learning models, assessed through receiver operating characteristic curves employing the "One-vs-Rest" strategy. The Delong test was applied to compare the predictive performance of the superior RadImageNet model against the ImageNet model. Results: Following pre-training separately on RadImageNet and ImageNet datasets, feature selection and fusion yielded 17 and 12 fusion features, respectively. Logistic regression emerged as the optimal machine learning algorithm for both DLR models. Across the training set, internal validation set, external validation set, and prospective validation set, the macro-average Area Under the Curve (AUC) based on the RadImageNet dataset surpassed those based on the ImageNet dataset, with statistically significant differences observed (P<0.05). Utilizing the binary "One-vs-Rest" strategy, the model based on the RadImageNet dataset demonstrated superior efficacy in predicting Class 0, achieving an AUC of 0.969 and accuracy of 0.863. Predicting Class 1 yielded an AUC of 0.945 and accuracy of 0.875, while for Class 2, the AUC and accuracy were 0.809 and 0.692, respectively. Conclusion: The DLR model, based on the RadImageNet dataset, outperformed the ImageNet model in predicting the classification of OVFs, with generalizability confirmed in the prospective validation set.

Early changes of bone metabolites and lymphocyte subsets may participate in osteoporosis onset: a preliminary study of a postmenopausal osteoporosis mouse model.

Purpose: Metabolic and immune changes in the early stages of osteoporosis are not well understood. This study aimed to explore the changes in bone metabolites and bone marrow lymphocyte subsets and their relationship during the osteoporosis onset. Methods: We established OVX and Sham mouse models. After 5, 15, and 40 days, five mice in each group were sacrificed. Humeri were analyzed by microCT. The bone marrow cells of the left femur and tibia were collected for flow cytometry analysis. The right femur and tibia were analyzed by LC-MS/MS for metabolomics analysis. Results: Bone microarchitecture was significantly deteriorated 15 days after OVX surgery. Analysis of bone metabolomics showed that obvious metabolite changes had happened since 5 days after surgery. Lipid metabolism was significant at the early stage of the osteoporosis. The proportion of immature B cells was increased, whereas the proportion of mature B cells was decreased in the OVX group. Metabolites were significantly correlated with the proportion of lymphocyte subsets at the early stage of the osteoporosis. Conclusion: Lipid metabolism was significant at the early stage of the osteoporosis. Bone metabolites may influence bone formation by interfering with bone marrow lymphocyte subsets.

Value of perfusion parameters from golden-angle radial sparse parallel dynamic contrast-enhanced magnetic resonance imaging in predicting pathological complete response after neoadjuvant chemoradiotherapy for locally advanced rectal cancer.

PURPOSE: Non-invasive methods for predicting pathological complete response (pCR) after neoadjuvant chemoradiotherapy (nCRT) can provide distinct leverage in the management of patients with locally advanced rectal cancer (LARC). This study aimed to investigate whether including the golden-angle radial sparse parallel (GRASP) dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) perfusion parameter (Ktrans), in addition to tumor regression grading (TRG) and apparent diffusion coefficient (ADC) values, can improve the predictive ability for pCR. METHODS: Patients with LARC who underwent nCRT and subsequent surgery were included. The imaging parameters were compared between patients with and without pCR. Receiver operating characteristic (ROC) curve analysis was used to evaluate the predictive ability of these parameters for pCR. RESULTS: A total of 111 patients were included in the study. A pCR was obtained in 32 patients (28.8%). MRI-based TRG (mrTRG) showed a negative correlation with pCR (r = -0.61, P < 0.001), and the average ADC value showed a positive correlation with pCR (r = 0.62, P < 0.001). Before nCRT, Ktrans in the pCR group was significantly higher than in the non-pCR group (1.30 ± 0.24 vs. 0.88 ± 0.34, P < 0.001), but no difference was identified after nCRT. Following ROC curve analysis, the area under the curve (AUC) of mrTRG (level 1-2), average ADC value, and Ktrans value for predicting pCR were 0.738 [95% confidence interval (CI): 0.65-0.82], 0.78 (95% CI: 0.69-0.86), and 0.84 (95% CI: 0.77-0.92), respectively. The model combining the three parameters had significantly higher predictive ability for pCR (AUC: 0.94, 95% CI: 0.88-0.98). CONCLUSION: The use of a combination of the GRASP DCE-MRI Ktrans with mrTRG and ADC can lead to a better pCR predictive performance.

A Contrast-Enhanced Tri-Modal MRI Technique for High-Performance Hypoxia Imaging of Breast Cancer.

Personalized radiotherapy strategies enabled by the construction of hypoxia-guided biological target volumes (BTVs) can overcome hypoxia-induced radioresistance by delivering high-dose radiotherapy to targeted hypoxic areas of the tumor. However, the construction of hypoxia-guided BTVs is difficult owing to lack of precise visualization of hypoxic areas. This study synthesizes a hypoxia-responsive T1 , T2 , T2 mapping tri-modal MRI molecular nanoprobe (SPION@ND) and provides precise imaging of hypoxic tumor areas by utilizing the advantageous features of tri-modal magnetic resonance imaging (MRI). SPION@ND exhibits hypoxia-triggered dispersion-aggregation structural transformation. Dispersed SPION@ND can be used for routine clinical BTV construction using T1 -contrast MRI. Conversely, aggregated SPION@ND can be used for tumor hypoxia imaging assessment using T2 -contrast MRI. Moreover, by introducing T2 mapping, this work designs a novel method (adjustable threshold-based hypoxia assessment) for the precise assessment of tumor hypoxia confidence area and hypoxia level. Eventually this work successfully obtains hypoxia tumor target and accurates hypoxia tumor target, and achieves a one-stop hypoxia-guided BTV construction. Compared to the positron emission tomography-based hypoxia assessment, SPION@ND provides a new method that allows safe and convenient imaging of hypoxic tumor areas in clinical settings.

Regional homogeneity alterations in patients with functional constipation and their associations with gene expression profiles.

Functional constipation, a highly prevalent functional gastrointestinal disorder, often accompanies by mental and psychological disorders. Previous neuroimaging studies have demonstrated brain functional and structural alterations in patients with functional constipation. However, little is known about whether and how regional homogeneity is altered in these patients. Moreover, the potential genetic mechanisms associated with these alterations remain largely unknown. The study included 73 patients with functional constipation and 68 healthy controls, and regional homogeneity comparison was conducted to identify the abnormal spontaneous brain activities in patients with functional constipation. Using Allen Human Brain Atlas, we further investigated gene expression profiles associated with regional homogeneity alterations in functional constipation patients with partial least squares regression analysis applied. Compared with healthy controls, functional constipation patients demonstrated significantly decreased regional homogeneity in both bilateral caudate nucleus, putamen, anterior insula, thalamus and right middle cingulate cortex, supplementary motor area, and increased regional homogeneity in the bilateral orbitofrontal cortex. Genes related to synaptic signaling, central nervous system development, fatty acid metabolism, and immunity were spatially correlated with abnormal regional homogeneity patterns. Our findings showed significant regional homogeneity alterations in functional constipation patients, and the changes may be caused by complex polygenetic and poly-pathway mechanisms, which provides a new perspective on functional constipation's pathophysiology.

The Clinical Value of Apparent Diffusion Coefficient of Readout Segmentation of Long Variable Echo Trains and Correlation With Ki-67 Expression in Distal Rectal Cancer.

OBJECTIVE: The aim of the study is to explore the clinical value of the apparent diffusion coefficient (ADC) derived from the readout segmentation of long variable echo trains (RESOLVE) technique for identifying clinicopathologic features of distal rectal cancer and correlations between ADC and Ki-67 expression. METHODS: The data of 112 patients with a proven pathology of distal rectal cancer who underwent preoperative magnetic resonance imaging were retrospectively analyzed. The mean ADC value was measured using the "full-layer and center" method. Differences in ADC values and Ki-67 expression in different clinical stages, pathological types, and tumor differentiation were compared using analysis of variance. Correlations between ADC value and clinicopathologic features were assessed using Spearman correlation analysis. RESULTS: Interobserver agreement of confidence levels from 2 radiologists was excellent for ADC measurement (k =  0.85). Patients with a lower clinical stage, well-differentiated adenocarcinomas, and a higher possibility of mucinous adenocarcinoma exhibited a positive correlation with higher ADC values, but these factors were negatively correlated with Ki-67 expression (all P < 0.05). We found that ADC value was negatively correlated with Ki-67 expression (r = -0.62, P < 0.001). CONCLUSIONS: The ADC value generated by RESOLVE sequences was significantly associated with clinicopathologic features and Ki-67 expression in patients with distal rectal cancer in this study. Thus, the ADC value could be considered a new noninvasive imaging biomarker that could be helpful in predicting the biological properties of distal rectal cancer.

Constructing a Deep Learning Radiomics Model Based on X-ray Images and Clinical Data for Predicting and Distinguishing Acute and Chronic Osteoporotic Vertebral Fractures: A Multicenter Study.

RATIONALE AND OBJECTIVES: To construct and validate a deep learning radiomics (DLR) model based on X-ray images for predicting and distinguishing acute and chronic osteoporotic vertebral fractures (OVFs). METHODS: A total of 942 cases (1076 vertebral bodies) with both vertebral X-ray examination and MRI scans were included in this study from three hospitals. They were divided into a training cohort (n = 712), an internal validation cohort (n = 178), an external validation cohort (n = 111), and a prospective validation cohort (n = 75). The ResNet-50 model architecture was used for deep transfer learning (DTL), with pre-training performed on RadImageNet and ImageNet datasets. DTL features and radiomics features were extracted from lateral X-ray images of OVFs patients and fused together. A logistic regression model with the least absolute shrinkage and selection operator was established, with MRI showing bone marrow edema as the gold standard for acute OVFs. The performance of the model was evaluated using receiver operating characteristic curves. Eight machine learning classification models were evaluated for their ability to distinguish between acute and chronic OVFs. The Nomogram was constructed by combining clinical baseline data to achieve visualized classification assessment. The predictive performance of the best RadImageNet model and ImageNet model was compared using the Delong test. The clinical value of the Nomogram was evaluated using decision curve analysis (DCA). RESULTS: Pre-training resulted in 34 and 39 fused features after feature selection and fusion. The most effective machine learning algorithm in both DLR models was Light Gradient Boosting Machine. Using the Delong test, the area under the curve (AUC) for distinguishing between acute and chronic OVFs in the training cohort was 0.979 and 0.972 for the RadImageNet and ImageNet models, respectively, with no statistically significant difference between them (P = 0.235). In the internal validation cohort, external validation cohort, and prospective validation cohort, the AUCs for the two models were 0.967 vs 0.629, 0.886 vs 0.817, and 0.933 vs 0.661, respectively, with statistically significant differences in all comparisons (P < 0.05). The deep learning radiomics nomogram (DLRN) was constructed by combining the predictive model of RadImageNet with clinical baseline features, resulting in AUCs of 0.981, 0.974, 0.895, and 0.902 in the training cohort, internal validation cohort, external validation cohort, and prospective validation cohort, respectively. Using the Delong test, the AUCs for the fused feature model and the DLRN in the training cohort were 0.979 and 0.981, respectively, with no statistically significant difference between them (P = 0.169). In the internal validation cohort, external validation cohort, and prospective validation cohort, the AUCs for the two models were 0.967 vs 0.974, 0.886 vs 0.895, and 0.933 vs 0.902, respectively, with statistically significant differences in all comparisons (P < 0.05). The Nomogram showed a slight improvement in predictive performance in the internal and external validation cohort, but a slight decrease in the prospective validation cohort (0.933 vs 0.902). DCA showed that the Nomogram provided more benefits to patients compared to the DLR models. CONCLUSION: Compared to the ImageNet model, the RadImageNet model has higher diagnostic value in distinguishing between acute and chronic OVFs. Furthermore, the diagnostic performance of the model is further improved when combined with clinical baseline features to construct the Nomogram.

The effect of cartilage dehydration and rehydration on quantitative ultrashort echo time biomarkers.

Background: The effect of dehydration of ex vivo cartilage samples and rehydration with native synovial fluid or normal saline on quantitative ultrashort echo time (UTE) biomarkers are unknown. We aimed to investigate the effect of cartilage dehydration-rehydration on UTE biomarkers and to compare the rehydration capabilities of native synovial fluid and normal saline. Methods: A total of 37 cartilage samples were harvested from patients (n=5) who underwent total knee replacement. Fresh cartilage samples were exposed to air to dehydrate for 2 hours after baseline magnetic resonance (MR) scanning, then randomly divided into two groups: one soaking in native synovial fluid (n=17) and the other in normal saline (n=20) to rehydrate for 4 hours. UTE-based biomarkers [T1, adiabatic T1r (AdiabT1r), macromolecular fraction (MMF), magnetization transfer ratio (MTR), and T2*] and sample weights were evaluated for fresh, dehydrated, and rehydrated cartilage samples. Differences and agreements between groups were assessed using the values of fresh cartilage samples as reference standard. Results: Dehydrating in air for 2 hours resulted in significant weight loss (P=0.000). T1, AdiabT1r, and T2* decreased significantly while MMF and MTR increased significantly (all P<0.02). Non-significant differences were observed in cartilage weights after rehydrating in both synovial fluid and normal saline, with P values being 0.204 and 0.769, respectively. There were no significant differences in T1, AdiabT1r, MMF, and MTR after rehydrating in synovial fluid (P>0.0167, with Bonferroni correction) while T2* (P=0.001) still had significant differences compared with fresh samples. However, no significant differences were detected for any of the evaluated UTE biomarkers after rehydrating in normal saline (all P>0.05). No differences were detected in the agreement of UTE biomarker measurements between fresh samples and samples rehydrated with synovial fluid and normal saline. Conclusions: Cartilage dehydration resulted in significant changes in UTE biomarkers. Rehydrating with synovial fluid or normal saline had non-significant effect on all the evaluated UTE biomarkers except T2* values, which still had significant differences compared with fresh samples after rehydrating with synovial fluid. No significant difference was observed in the rehydration capabilities of native synovial fluid and normal saline.

FFENet: frequency-spatial feature enhancement network for clothing classification.

Clothing analysis has garnered significant attention, and within this field, clothing classification plays a vital role as one of the fundamental technologies. Due to the inherent complexity of clothing scenes in real-world environments, the learning of clothing features in such complex scenes often encounters interference. Because clothing classification relies on the contour and texture information of clothing, clothing classification in real scenes may lead to poor classification results. Therefore, this paper proposes a clothing classification network based on frequency-spatial domain conversion. The proposed network combines frequency domain information with spatial information and does not compress channels. It aims to enhance the extraction of clothing features and improve the accuracy of clothing classification. In our work, (1) we combine the frequency domain information and spatial information to establish a clothing feature extraction clothing classification network without compressed feature map channels, (2) we use the frequency domain feature enhancement module to realize the preliminary extraction of clothing features, and (3) we introduce a clothing dataset in complex scenes (Clothing-8). Our network achieves a top-1 model accuracy of 93.4% on the Clothing-8 dataset and 94.62% on the Fashion-MNIST dataset. Additionally, it also achieves the best results in terms of top-3 and top-5 metrics on the DeepFashion dataset.

Bone marrow fatty acids affect osteoblastic differentiation through miR-92b-3p in the early stages of postmenopausal osteoporosis.

Osteoporosis is partially caused by dysfunctions in the commitment, differentiation or survival of osteoblasts. Bone marrow fatty acids affect bone resorption and formation. In this study, we aimed to explore the role of fatty acids in the early stages of postmenopausal osteoporosis and determine whether they influence osteogenic differentiation through microRNAs. A quantitative analysis of bone marrow fatty acids early after ovariectomy or sham surgery in a rat osteoporotic model was performed using gas chromatography/mass spectrometry. The results showed that palmitoleate was significantly decreased on postoperative day 3 while both pentadecanoate and palmitoleate were significantly decreased on postoperative day 5 in rats in the ovariectomized group compared with those in the sham group. Palmitoleate promotes osteogenic differentiation, whereas pentadecanoate inhibits this process. Palmitoleate levels were higher than those of pentadecanoate; therefore, the early overall effect of significant bone marrow fatty acid changes was a decrease in osteogenic differentiation. We also found that miR-92b-3p inhibited osteoblastogenesis via the miR-92b-3p/phosphatase and tensin homolog regulatory axis. Palmitoleate, pentadecanoate, and palmitate influenced the osteoblastogenesis of MC3T3-E1 cells through miR-92b-3p. Taken together, we propose that miR-92b-3p mediates the effect of bone marrow fatty acids on osteoblast differentiation in the early stages of osteoporosis. These findings may provide molecular insights for the treatment of osteoporosis.

Differentiation of acute and chronic vertebral compression fractures using conventional CT based on deep transfer learning features and hand-crafted radiomics features.

BACKGROUND: We evaluated the diagnostic efficacy of deep learning radiomics (DLR) and hand-crafted radiomics (HCR) features in differentiating acute and chronic vertebral compression fractures (VCFs). METHODS: A total of 365 patients with VCFs were retrospectively analysed based on their computed tomography (CT) scan data. All patients completed MRI examination within 2 weeks. There were 315 acute VCFs and 205 chronic VCFs. Deep transfer learning (DTL) features and HCR features were extracted from CT images of patients with VCFs using DLR and traditional radiomics, respectively, and feature fusion was performed to establish the least absolute shrinkage and selection operator. The MRI display of vertebral bone marrow oedema was used as the gold standard for acute VCF, and the model performance was evaluated using the receiver operating characteristic (ROC).To separately evaluate the effectiveness of DLR, traditional radiomics and feature fusion in the differential diagnosis of acute and chronic VCFs, we constructed a nomogram based on the clinical baseline data to visualize the classification evaluation. The predictive power of each model was compared using the Delong test, and the clinical value of the nomogram was evaluated using decision curve analysis (DCA). RESULTS: Fifty DTL features were obtained from DLR, 41 HCR features were obtained from traditional radiomics, and 77 features fusion were obtained after feature screening and fusion of the two. The area under the curve (AUC) of the DLR model in the training cohort and test cohort were 0.992 (95% confidence interval (CI), 0.983-0.999) and 0.871 (95% CI, 0.805-0.938), respectively. While the AUCs of the conventional radiomics model in the training cohort and test cohort were 0.973 (95% CI, 0.955-0.990) and 0.854 (95% CI, 0.773-0.934), respectively. The AUCs of the features fusion model in the training cohort and test cohort were 0.997 (95% CI, 0.994-0.999) and 0.915 (95% CI, 0.855-0.974), respectively. The AUCs of nomogram constructed by the features fusion in combination with clinical baseline data were 0.998 (95% CI, 0.996-0.999) and 0.946 (95% CI, 0.906-0.987) in the training cohort and test cohort, respectively. The Delong test showed that the differences between the features fusion model and the nomogram in the training cohort and the test cohort were not statistically significant (P values were 0.794 and 0.668, respectively), and the differences in the other prediction models in the training cohort and the test cohort were statistically significant (P < 0.05). DCA showed that the nomogram had high clinical value. CONCLUSION: The features fusion model can be used for the differential diagnosis of acute and chronic VCFs, and its differential diagnosis ability is improved when compared with that when either radiomics is used alone. At the same time, the nomogram has a high predictive value for acute and chronic VCFs and can be a potential decision-making tool to assist clinicians, especially when a patient is unable to undergo spinal MRI examination.

A preliminary study on degenerate characteristics of lumbar and abdominal muscles in middle-aged and elderly people with varying bone mass.

BACKGROUND: With the wide application of QCT in the clinical assessment of osteoporosis and sarcopenia, the characteristics of musculoskeletal degeneration in middle-aged and elderly people need to be further revealed. We aimed to investigate the degenerate characteristics of lumbar and abdominal muscles in middle-aged and elderly people with varying bone mass. METHODS: A total of 430 patients aged 40-88 years were divided into normal, osteopenia, and osteoporosis groups according to quantitative computed tomography (QCT) criteria. The skeletal muscular mass indexes (SMIs) of five muscles [abdominal wall muscles (AWM), rectus abdominis (RA), psoas major muscle (PMM), posterior vertebral muscles (PVM), and paravertebral muscles (PM)] included in lumbar and abdominal muscles were measured by QCT. Differences in SMIs among three groups, as well as the correlation between SMIs and volumetric bone mineral density (vBMD) were analyzed. The areas under the curves (AUCs) for SMIs for prediction of low bone mass and osteoporosis were calculated. RESULTS: In male group, SMIs of RA and PM in osteopenia group were significantly lower than those in the normal group (P = 0.001 and 0.023, respectively). In female group, only SMI of RA in osteopenia group was significantly lower than that in the normal group (P = 0.007). SMI of RA was positively correlated with vBMD with the highest coefficients in male and female groups (r = 0.309 and 0.444, respectively). SMIs of AWM and RA had higher AUCs varying from 0.613 to 0.737 for prediction of low bone mass and osteoporosis in both genders. CONCLUSIONS: The changes of SMIs of the lumbar and abdominal muscles in patients with varying bone mass are asynchronous. SMI of RA is expected to be a promising imaging marker for predicting abnormal bone mass. TRIAL REGISTRATION: ChiCTR1900024511 (Registered 13-07-2019).

Evaluation of left ventricular blood flow kinetic energy in patients with hypertension by four-dimensional flow cardiovascular magnetic resonance imaging: a preliminary study.

OBJECTIVES: To evaluate the intra-cavity left ventricular (LV) blood flow kinetic energy (KE) parameters using four-dimensional (4D) flow cardiovascular magnetic resonance (CMR) in patients with hypertension (HTN). METHODS: Forty-two HTN patients and twenty age-/gender-matched healthy controls who underwent CMR including cines, pre-/post-T1 mapping, and whole-heart 4D flow imaging were retrospectively evaluated. HTN patients were further divided into two subgroups: with preserved ejection fraction (HTN-pEF) and with reduced ejection fraction (HTN-rEF). KE parameters were indexed to LV end-diastolic volume (EDV) to obtain averaged LV, minimal, systolic, diastolic, peak E-wave, peak A-wave, E-wave, and A-wave KEiEDV, as well as the proportion of in-plane LV KE (%), the time difference (TD). These parameters were compared between the HTN group and healthy controls, also between two subgroups. The correlation of LV blood flow KE parameters with LV function and extracellular volume fraction (ECV) were analyzed in the HTN group using multivariate regression analysis. RESULTS: Peak E-wave KEiEDV in the HTN group was significantly lower (p = 0.01), while in-plane KE and TD were significantly higher (all p < 0.01) than those in healthy controls. Compared to the HTN-pEF subgroup, the proportion of in-plane KE and TD was significantly increased in the HTN-rEF subgroup (all p < 0.01). Only the proportion of in-plane KE demonstrated an independent correlation with ECV (ß* = 0.59, p < 0.01). CONCLUSIONS: The decreased peak E-wave KEiEDV and the increased proportion of in-plane KE, TD reflected the alterations of LV blood flow in HTN patients, and the proportion of in-plane KE was independently associated with ECV. KEY POINTS: • 4D flow CMR demonstrated that the peak E-wave KEiEDV was decreased, while the in-plane KE and time difference (TD) were increased in hypertensive (HTN) patients. • The proportion of in-plane KE and TD was further increased in HTN patients with reduced ejection fraction than in HTN patients with preserved ejection fraction, and the proportion of in-plane KE was independently associated with extracellular volume fraction in HTN patients. • 4D flow CMR intra-cavity blood flow KE parameters might reveal the LV hemodynamic status in preclinical HTN patients.

mDIXON-Quant technique diagnostic accuracy for assessing bone mineral density in male adult population.

BACKGROUND: To investigate the diagnostic efficacy of mDIXON-Quant technique for prediction of bone loss in male adults. METHODS: One hundred thirty-eight male adults were divided into normal, osteopenia, and osteoporosis groups based on DXA and QCT for the lumbar spine. Differences in mDIXON-Quant parameters [fat fraction (FF) and T2* value] among three groups, as well as the correlation of mDIXON-Quant parameters and bone mineral density (BMD) were analyzed. The areas under the curves (AUCs) for mDIXON-Quant parameters for prediction of low bone mass were calculated. RESULTS: According to DXA standard, FF and T2* value were significantly increased in osteoporosis group compared with normal group (P = 0.012 and P < 0.001). According to QCT standard, FF was significantly increased in osteopenia and osteoporosis groups compared with normal group (both P < 0.001). T2* values were significantly different among three groups (all P < 0.05). After correction for age and body mass index, FF was negatively correlated with areal BMD and volumetric BMD (r = -0.205 and -0.604, respectively; both P < 0.05), and so was T2* value (r = -0.324 and -0.444, respectively; both P < 0.05). The AUCs for predicting low bone mass according to DXA and QCT standards were 0.642 and 0.898 for FF, 0.648 and 0.740 for T2* value, and 0.677 and 0.920 for both combined, respectively. CONCLUSIONS: FF combined with T2* value has a better diagnostic efficacy than FF or T2* value alone in prediction of low bone mass in male adults, which is expected to be a promising MRI method for the screening of bone quality. TRIAL REGISTRATION: ChiCTR1900024511 (Registered 13-07-2019).

ICAM1-Targeting Theranostic Nanoparticles for Magnetic Resonance Imaging and Therapy of Triple-Negative Breast Cancer.

Purpose: Owing to the lack of effective biomarkers, triple-negative breast cancer (TNBC) has the worst prognosis among all subtypes of breast cancer. Meanwhile, tremendous progress has been made to identify biomarkers for TNBC. However, limited number of biomarkers still restrain the specifically targeting outcomes against TNBC. Here, to solve the obstacle, we designed and synthesized a new type of biocompatible nanoparticles to amplify the targeting effects for TNBC theranostics. Methods: To identify the biomarker of TNBC, the expression of intercellular adhesion molecule-1 (ICAM1) was assessed by real-time polymerase chain reaction and western blot among all subtypes of breast cancer and normal breast epithelium. Then, vesicular nanoparticles based on poly(ethylene glycol)-poly(ε-caprolactone) copolymers were prepared by the double emulsion method and modified with anti-ICAM1 antibodies through click chemistry to conjugate with related antigens on TNBC cell membranes and then loaded with magnetic resonance imaging (MRI) contrast agent gadolinium and chemotherapeutic drug doxorubicin. The targeting capability, diagnostic and therapeutic efficacy of this nanoparticle were validated through cell-based and tumor model-based experiments. Results: ICAM1 was expressed significantly higher on TNBC than on other subtypes of breast cancer and normal breast epithelium in both mRNA and protein level. Theranostic nanoparticle modified with anti-ICAM1 was proved to be able to specifically target to TNBC in vitro experiments. Such theranostic nanoparticle also displayed enhanced diagnostic and therapeutic efficacy by specifically targeting capability and extending circulation time in tumor models. The biocompatibility and biosafety of this nanoparticle was also confirmed in vitro and in vivo. Conclusion: Overall, this new nanoparticle has been demonstrated with effective therapeutic outcomes against TNBC, providing a promising theranostic approach for MRI-guided therapy of TNBC.

Value of texture analysis based on dynamic contrast-enhanced magnetic resonance imaging in preoperative assessment of extramural venous invasion in rectal cancer.

OBJECTIVE: Accurate preoperative assessment of extramural vascular invasion (EMVI) is critical for the treatment and prognosis of rectal cancer. The aim of our research was to develop an assessment model by texture analysis for preoperative prediction of EMVI. MATERIALS AND METHODS: This study enrolled 44 rectal patients as train cohort, 7 patients as validation cohort and 18 patients as test cohort. A total of 236 texture features from DCE MR imaging quantitative parameters were extracted for each patient (59 features of Ktrans, Kep, Ve and Vp), and key features were selected by least absolute shrinkage and selection operator regression (LASSO). Finally, clinical independent risk factors, conventional MRI assessment, and T-score were incorporated to construct an assessment model using multivariable logistic regression. RESULTS: The T-score calculated using the 4 selected key features were significantly correlated with EMVI (p < 0.010). The area under the receiver operating characteristic curve (AUC) was 0.797 for discriminating between EMVI-positive and EMVI-negative patients with a sensitivity of 88.2% and specificity of 70.4%. The conventional MRI assessment of EMVI had a sensitivity of 23.53% and a specificity of 96.30%. The assessment model showed a greatly improved performance with an AUC of 0.954 (sensitivity, 88.2%; specificity, 92.6%) in train cohort, 0.833 (sensitivity, 66.7%; specificity, 100%) in validation cohort and 0.877 in test cohort, respectively. CONCLUSIONS: The assessment model showed an excellent performance in preoperative assessment of EMVI. It demonstrates strong potential for improving the accuracy of EMVI assessment and provide a reliable basis for individualized treatment decisions.

Evaluation of bone mineral density and body compositions interrelation in young and middle-aged male patients with Crohn's disease by quantitative computed tomography.

Background: The aim of this study was to investigate the characteristics of bone mineral density (BMD) and body compositions, and the impact of body compositions on BMD in young and middle-aged male patients with Crohn's disease (CD). Methods: Patients with CD (n = 198) and normal controls (n = 123) underwent quantitative computed tomography (QCT) examination of lumbar vertebrae 1-3 (L1-3). The BMD and bone geometric parameters were measured and outputted by QCT post-process software. Meanwhile, body composition parameters, including subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT), lean mass (LM), and muscles mass around lumbar vertebrae were also acquired by QCT. Blood indicators [interleukin (IL)-6, IL-8, tumor necrosis factor alpha (TNF-α), C-reactive protein (CRP), Ca, and P] were collected from clinical medical records. Independent t-test was used to compare these variables between the CD group and the normal control group. Results: There was no significant difference in age, height, and weight between the CD group and the control group (p > 0.05), indicating that the sample size was relatively balanced. Mean BMD in the CD group were lower than those in the control group, but the difference was not statistically significant (p > 0.05). The bone geometric parameters of the CD group, including cortical area/density (Ct. Ar, Ct. BMD) and trabecular area/density (Tb. Ar and Tb. BMD), were significantly lower than those of the control group (p < 0.05), so were the body composition parameters including total adipose tissue (TAT), visceral adipose tissue (VAT), subcutaneous adipose tissue (SAT), lean mass (LM), and muscles mass (p < 0.05). In addition, the level of plasma IL-6, IL-8, CRP, and TNF-α of the CD group were higher than those of the control group (p < 0.05). On the contrary, the body mass index (BMI) and serum Ca and P levels of the CD group were lower than those of the control group (p < 0.05). Through multiple linear regression analysis, Tb. BMD, VAT, Ct. Ar, LM, Ca, and IL-8 entered the regression model and revealed a significant contribution to BMD. Conclusions: Patients with CD could suffer from reduction in BMD. However, the parameters of bone geometric parameters are more sensitive and accurate than BMD changes. Among them, Tb. BMD, VAT, Ct. Ar, and LM have significant effects on BMD reduction.

Preliminary study on the assessment of early cartilage degeneration by quantitative ultrashort echo time magnetic resonance imaging in vivo.

Background: To investigate the feasibility of quantitative ultrashort echo time magnetic resonance imaging (UTE-MRI) techniques for assessing early cartilage degeneration in vivo. Methods: A total of 46 patients with knee pain due to osteoarthritis (OA) as the main complaint were recruited into the study. We performed MRI examinations with different quantitative UTE-MRI techniques, including UTE-based magnetization transfer (MT), UTE-adiabaticT1ρ, and UTE-T2* mapping on a 3.0T clinical magnetic resonance (MR) scanner (MR750; GE Healthcare, Milwaukee, WI, USA). Three regions of interest (ROIs) were manually drawn on the medial and lateral femoral condyles and the corresponding medial and lateral tibial plateaus, respectively. A total of 561 ROIs (12 ROIs for each knee) were finally included and divided into 3 groups according to the MRI Osteoarthritis Knee Score (MOAKS): normal (MOAKS 0, n=175), mild degeneration (MOAKS 1, n=283), and moderate degeneration (MOAKS 2, n=103). One-way analysis of variance (ANOVA) and Tamhane's T2 test were used to compare the differences of quantitative UTE-biomarkers among different groups. The analysis of Spearman's correlation was used to assess the correlation between the UTE-biomarkers and MOAKS grading. The diagnostic efficacy of different quantitative UTE-MRI techniques for detecting mild cartilage degeneration was evaluated using the receiver operating characteristic (ROC) curve. Results: The UTE-MT ratio (UTE-MTR) and the UTE-adiabatic T1ρ values had a moderate correlation with the MOAKS grading (r=-0.523, P<0.001; r=0.531, P<0.001, respectively), while the UTE-T2* was weakly correlated with the MOAKS grading (r=-0.396, P<0.001). For the normal group (MOAKS 0) and the mild group (MOAKS 1), the UTE-MTR values were 21.09%±3.03% and 17.30%±3.22%, respectively. The UTE-adiabatic T1ρ values were 30.43±6.26 ms and 35.05±8.78 ms for the normal group (MOAKS 0) and the mild group (MOAKS 1), respectively. With respect to the UTE-T2* values, the normal group (MOAKS 0) values were 21.49±3.96 ms and the mild group (MOAKS 1) values were 19.86±3.08 ms. All the differences between the 2 groups of the 3 UTE-MRI values were significant. The AUCs of the UTE-MTR, UTE-adiabatic T1ρ, and UTE-T2* mapping were 0.794, 0.732, and 0.651, respectively. Conclusions: The quantitative UTE-MRI techniques (UTE-MT, UTE-adiabatic T1ρ, and UTE-T2* mapping) show great promise for assessing the early degeneration of articular cartilage in vivo, and the UTE-MT and UTE-adiabatic T1ρ values show better diagnostic efficacy than UTE-T2* mapping.

Correlation between musculoskeletal mass and perfusion in patients with gastrointestinal malignancy: a preliminary study based on quantitative CT and CT perfusion.

BACKGROUND: To investigate the correlation between musculoskeletal mass and perfusion using quantitative computer tomography (QCT) and CT perfusion (CTP) in patients with gastrointestinal malignancy. METHODS: In this prospective study, 96 patients (mean age 66 years, range 25-90; 63.5% male) with gastrointestinal malignancy underwent QCT and CTP between May 2019 and February 2021. Bone mineral density (BMD) and body composition [perivertebral muscular mass index (PMI), skeletal muscular mass index (SMI) and muscular fat fraction] were evaluated through QCT. Musculoskeletal perfusion parameters were measured by CTP. Differences in these parameters between (or among) two (or three) groups (grouped by BMD, SMI, and TNM staging) were analyzed. RESULTS: There were significant differences in PMI and muscular fat fraction among normal (n = 30), osteopenia (n = 43), and osteoporosis (n = 23) groups (both P < 0.001). Blood flow (r = 0.336, P = 0.001; adjusted for age and gender, r = 0.383, P < 0.001), blood volume (r = 0.238, P = 0.011; adjusted for age and gender, r = 0.329, P = 0.001), and flow extraction product (r = 0.217, P = 0.034; adjusted for age and gender, r = 0.320, P = 0.002) vaules of vertebral perfusion showed positive correlation with BMD. However, the relationships between PMI and perfusion parameters of perivertebral muscle were not significant. No significant differences were found in musculoskeletal mass and perfusion parameters between different TNM staging. CONCLUSIONS: The changes of bone mass and perivertebral muscular mass in patients with gastrointestinal malignancy are synchronous. Decreased vertebral bone mass is accompanied with reduced perivertebral muscular mass, increased muscular fat, and decreased bone perfusion. However, the changes of perfusion in vertebra and perivertebral muscles are asynchronous. Musculoskeletal mass and perfusion have no correlation with TNM staging of the patients with gastrointestinal malignancy. TRIAL REGISTRATION: SHSY-IEC-4.1/20-242/01 (Registered 09-12-2020, Retrospectively registered).