Multicenter Evaluation of a Weakly Supervised Deep Learning Model for Lymph Node Diagnosis in Rectal Cancer at MRI.

Purpose To develop a Weakly supervISed model DevelOpment fraMework (WISDOM) model to construct a lymph node (LN) diagnosis model for patients with rectal cancer (RC) that uses preoperative MRI data coupled with postoperative patient-level pathologic information. Materials and Methods In this retrospective study, the WISDOM model was built using MRI (T2-weighted and diffusion-weighted imaging) and patient-level pathologic information (the number of postoperatively confirmed metastatic LNs and resected LNs) based on the data of patients with RC between January 2016 and November 2017. The incremental value of the model in assisting radiologists was investigated. The performances in binary and ternary N staging were evaluated using area under the receiver operating characteristic curve (AUC) and the concordance index (C index), respectively. Results A total of 1014 patients (median age, 62 years; IQR, 54-68 years; 590 male) were analyzed, including the training cohort (n = 589) and internal test cohort (n = 146) from center 1 and two external test cohorts (cohort 1: 117; cohort 2: 162) from centers 2 and 3. The WISDOM model yielded an overall AUC of 0.81 and C index of 0.765, significantly outperforming junior radiologists (AUC = 0.69, P < .001; C index = 0.689, P < .001) and performing comparably with senior radiologists (AUC = 0.79, P = .21; C index = 0.788, P = .22). Moreover, the model significantly improved the performance of junior radiologists (AUC = 0.80, P < .001; C index = 0.798, P < .001) and senior radiologists (AUC = 0.88, P < .001; C index = 0.869, P < .001). Conclusion This study demonstrates the potential of WISDOM as a useful LN diagnosis method using routine rectal MRI data. The improved radiologist performance observed with model assistance highlights the potential clinical utility of WISDOM in practice. Keywords: MR Imaging, Abdomen/GI, Rectum, Computer Applications-Detection/Diagnosis Supplemental material is available for this article. Published under a CC BY 4.0 license.

Ultrafine nanoporous intermetallic catalysts by high-temperature liquid metal dealloying for electrochemical hydrogen production.

Intermetallic compounds formed from non-precious transition metals are promising cost-effective and robust catalysts for electrochemical hydrogen production. However, the development of monolithic nanoporous intermetallics, with ample active sites and sufficient electrocatalytic activity, remains a challenge. Here we report the fabrication of nanoporous Co7Mo6 and Fe7Mo6 intermetallic compounds via liquid metal dealloying. Along with the development of three-dimensional bicontinuous open porosity, high-temperature dealloying overcomes the kinetic energy barrier, enabling the direct formation of chemically ordered intermetallic phases. Unprecedented small characteristic lengths are observed for the nanoporous intermetallic compounds, resulting from an intermetallic effect whereby the chemical ordering during nanopore formation lowers surface diffusivity and significantly suppresses the thermal coarsening of dealloyed nanostructure. The resulting ultrafine nanoporous Co7Mo6 exhibits high catalytic activity and durability in electrochemical hydrogen evolution reactions. This study sheds light on the previously unexplored intermetallic effect in dealloying and facilitates the development of advanced intermetallic catalysts for energy applications.

Deep learning predicts resistance to neoadjuvant chemotherapy for locally advanced gastric cancer: a multicenter study.

BACKGROUND: Accurate pre-treatment prediction of neoadjuvant chemotherapy (NACT) resistance in patients with locally advanced gastric cancer (LAGC) is essential for timely surgeries and optimized treatments. We aim to evaluate the effectiveness of deep learning (DL) on computed tomography (CT) images in predicting NACT resistance in LAGC patients. METHODS: A total of 633 LAGC patients receiving NACT from three hospitals were included in this retrospective study. The training and internal validation cohorts were randomly selected from center 1, comprising 242 and 104 patients, respectively. The external validation cohort 1 comprised 128 patients from center 2, and the external validation cohort 2 comprised 159 patients from center 3. First, a DL model was developed using ResNet-50 to predict NACT resistance in LAGC patients, and the gradient-weighted class activation mapping (Grad-CAM) was assessed for visualization. Then, an integrated model was constructed by combing the DL signature and clinical characteristics. Finally, the performance was tested in internal and external validation cohorts using area under the receiver operating characteristic (ROC) curves (AUC). RESULTS: The DL model achieved AUCs of 0.808 (95% CI 0.724-0.893), 0.755 (95% CI 0.660-0.850), and 0.752 (95% CI 0.678-0.825) in validation cohorts, respectively, which were higher than those of the clinical model. Furthermore, the integrated model performed significantly better than the clinical model (P < 0.05). CONCLUSIONS: A CT-based model using DL showed promising performance for predicting NACT resistance in LAGC patients, which could provide valuable information in terms of individualized treatment.

CT-based radiomics analysis in the prediction of response to neoadjuvant chemotherapy in locally advanced gastric cancer: A dual-center study.

BACKGROUND: To investigate the ability of the CT-based radiomics models for pretreatment prediction of the response to neoadjuvant chemotherapy (NAC) in patients with locally advanced gastric cancer (LAGC). METHODS: This retrospective analysis included 279 consecutive LAGC patients from center I (training cohort, n = 196; internal validation cohort, n = 83) who were examined by contrast-enhanced CT before treatment and 211 consecutive patients from center II who were recruited as an external validation cohort. A total of 102 features were extracted from the portal venous phase CT images, and feature selection was further subjected to three-step procedures. Next, five classifications, including Logistic Regression (LR), Naive Bayes, Random forest (RF), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGB) algorithms, were applied to construct radiomics models for predicting the good-responder (GR) to NAC in the training cohort. The prediction performances were evaluated using ROC and decision curve analysis (DCA). RESULTS: No statistically significant difference was detected for all clinicopathological characteristics. Additionally, allsix key features were significantly different between GR and poor-responder (PR). Compared to models from other classifiers, the model obtained with XGB showed promising prediction performance with the highest AUC of 0.790(95%CI: 0.700-0.880) in the training cohort. The corresponding AUCs were 0.784(95%CI, 0.659-0.908) and 0.803(95%CI, 0.717-0.888) in the internal and external validation cohorts, respectively. DCA confirmed the clinical utility. CONCLUSIONS: The proposed pretreatment CT-based radiomics models revealed good performances in predicting response to NAC and thus may be used to improve clinical treatment in LAGC patients.

Catenary-based phase change metasurfaces for mid-infrared switchable wavefront control.

Active wave manipulation by ultracompact meta-devices is highly embraced in recent years, but a major concern still exists due to the lack of functional reconfigurability. Moreover, the phase or amplitude discontinuities introduced by collective response of discrete meta-atoms make current meta-devices far from practical applications. Here, we demonstrate actively tunable wavefront control with high-efficiency by combining catenary-based meta-atoms for intrinsic continuous phase regulation with the chalcogenide phase change material (PCM) of Ge2Sb2Te5. First, switchable beam deflection is demonstrated in a wide mid-IR range between 8 µm and 9.5 µm with 'on' and 'off' states for beam steering between anomalous and normal specular reflections. Second, a switchable meta-axicon for zero order Bessel beam generation is demonstrated with full width at half maximum (FWHM) as small as ∼0.41 λ (λ = 12 µm). As a result, our scheme for active and continuous phase control potentially paves an avenue to construct active photonic devices especially for applications where large contrast ratio is highly desirable, such as optoelectronic integration, wavefront engineering and so on.

Silica dust exposure induces autophagy in alveolar macrophages through switching Beclin1 affinity from Bcl-2 to PIK3C3.

Increased deposition of silica dust in pulmonary interstitial tissues leads to silicosis, in which autophagy plays a defensive role in silica dust-associated stress response and cell death. Our previous studies revealed that silica dust exposure contributed to autophagy in pulmonary macrophages in vivo, while the specific regulatory mechanism is still unclear. This study aimed to figure out the regulatory mechanism as well as the role of autophagy in the pathogenesis of experimental silicosis. We used 3-methyladenine (3-MA) and ABT-737 to suppress the expression of phosphatidylinositol 3-kinase catalytic subunit type 3 (PIK3C3) and B cell leukemia/lymphoma 2 (Bcl-2), two critical initiators of autophagy, and detected and evaluated the autophagy in NR8383 cells with or without silica dust exposure. We found that exposure of silica dust increased autophagy in NR8383 cells and elevated the expression of Beclin1 and PIK3C3, but it reduced the expression of Bcl-2. The relationship among Beclin1, PIK3C3, and Bcl-2 were then investigated using immunoprecipitation analysis, and we found that suppression of PIK3C3 and/or Bcl-2 using 3-MA and/or ABT-737 could alter the autophagy induced by silica dust in NR8383 cells, and the complexes of Beclin1/PIK3C3 and Beclin1/Bcl-2 were both downregulated, which may be that inhibition of PIK3C3 and Bcl-2 altered the affinity of Beclin1 with PIK3C3 and Bcl-2 and lead to the silence of PIK3C3 signaling. These findings indicate that silica dust exposure induces autophagy via changing the connectivity of Beclin1 from Bcl-2 to PIK3C3.

Optical Sensor Assistant with Voltage Enrichment for Ultrasensitive Detection of Mercury Ions.

Voltage enrichment-assisted surface-enhanced resonance Raman scattering (SERRS) was employed for monitoring mercury(II) ions based on a nanoporous gold (NPG)/aptamer hybrid sensor. The hybrid sensor based on the coordination chemistry of thymine-Hg2+-thymine interaction and an applied voltage that pushed mercury ions toward the NPG film improved both the speed and sensitivity of Hg2+ ion detection. The detection limit can reach 0.1 pM, even in the aqueous solution containing 13 metal ions, and the sensor also possesses a fine reproducibility and stability, promising great potential in real-time sensing applications.

Genetic manipulation of Escherichia coli central carbon metabolism for efficient production of fumaric acid.

Fumaric acid is one of the top 12-biomass building-block chemicals. In this study, we reported manipulation of E. coli central carbon metabolism with the aim to decrease the by-products and improve fumaric acid production. PEP-dependent glucose phosphotransferase system was replaced with a galactose translocation system to minimize the consumption of phosphoenolpyruvate. Engineering anaplerotic pathway (phosphoenolpyruvate carboxylase) was employed to redistribute carbon flux from glycolysis to Krebs cycle. Deletion of malate dehydrogenase and overexpression of acetyl-CoA synthase could decrease the byproducts malic acid and acetic acid. The combined strategies led to fumaric acid yield up to 1.53 g/g dry cell weight, a 50% increase compared with the parental strain. The result demonstrated that these genetic modifications were effective strategies for improving the production of fumaric acid and the engineered strain may serve a platform microbial cell factory for efficient production of fumaric acid or other dicarboxylic acids.

A Brain Region-Based Deep Medullary Veins Visual Score on Susceptibility Weighted Imaging.

Cerebral venous collagenosis played a role in the pathogenesis of white matter hyperintensities (WMHs) through venous ischemia. Since pathological changes of veins from intramural stenosis to luminal occlusion is a dynamic process, we aimed to create a deep medullary veins (DMVs) visual grade on susceptibility-weighted images (SWI) and explore the relationship of DMVs and WMHs based on venous drainage regions. We reviewed clinical, laboratory and imaging data from 268 consecutive WMHs patients and 20 controls. SWI images were used to observe characteristics of DMVs and a brain region-based DMVs visual score was given by two experienced neuroradiologists. Fluid attenuated inversion recovery (FLAIR) images were used to calculate WMHs volume. Logistic-regression analysis and partial Pearson's correlation analysis were used to examine the association between the DMVs score and WMHs volume. We found that the DMVs score was significantly higher in WMHs patients than in controls (p < 0.001). Increased DMVs score was independently associated with higher WMHs volume after adjusting for total cholesterol level and number of lacunes (p < 0.001). Particularly, DMVs scores were correlated with regional PVHs volumes in the same brain region most. The newly proposed DMVs grading method allows the clinician to monitor the course of DMVs disruption. Our findings of cerebral venous insufficiency in WMHs patients may help to elucidate the pathogenic mechanisms and progression of WMHs.

Changes in structure and perfusion of grey matter tissues during recovery from Ischaemic subcortical stroke: a longitudinal MRI study.

Stroke recovery with changes in volume and perfusion of grey matter (GM) tissues remains largely unknown. We hypothesized that GM atrophy co-existed with GM plasticity presenting with increased volume and perfusion in specific regions in the period of post-stroke recovery. Twelve well-recovered stroke patients with pure subcortical lesions in the middle cerebral artery-perfused zone were included. All of them underwent structural and perfusion magnetic resonance imaging (MRI) examinations at admission and a mean of 6 months after stroke onset. Differences in GM volume (GMV) on structural images and cerebral blood flow (CBF) derived from perfusion images between two examinations were compared using voxel-based morphometry. The associations between changes in GMV and CBF with clinical scores were analysed. Decreased GMV was found in post-central gyrus, pre-central gyrus, precuneus, angular gyrus, insula, thalamus and cerebellum, and increased GMV was found in hippocampus, orbital gyrus and lingual gyrus (all corrected P < 0.05) at the follow-up examination. Increased CBF was found in subcallosal cingulate gyrus, hippocampus and lingual gyrus (all corrected P < 0.05) at the follow-up examination. Only decreased GMV in the anterior lobe of cerebellum was negatively associated with improvement of Barthel index (ß = -0.683, P = 0.014). Our study provides the imaging evidence of GM atrophy co-existing with GM plasticity involving in increased volume and perfusion in specific regions (including cognition, vision and emotion) in well-recovered stroke patients, which advances our understanding of neurobiology of stroke recovery.

Different iron deposition patterns in early- and middle-late-onset Parkinson's disease.

BACKGROUND: Iron deposition may contribute to the clinical symptoms in Parkinson's disease (PD). With partial different clinical manifestations, the iron deposition patterns between patients with early-onset Parkinson's disease (EOPD) and middle-late-onset Parkinson's disease (M-LOPD) are still unclear. This study was designed to investigate the patterns of iron deposition and their clinical relevance in EOPD and M-LOPD patients, using quantitative susceptibility mapping technique. MATERIALS AND METHODS: Thirty-five EOPD patients and 24 matched young controls, 33 M-LOPD patients and 22 matched older controls were recruited in the study. The iron content in the deep grey matter nuclei in the basal ganglia and midbrain were measured, and compared between patients and their corresponding controls. The correlations of regional iron content and clinical features were explored in patient groups. RESULTS: Both M-LOPD and EOPD patients showed increased iron content in the substantia nigra (SN) pars compacta and SN pars reticulata. Increased iron content in the putamen was only observed in M-LOPD patients. The relationship between the increased iron content and disease severity (H&Y stages, UPDRS II scores and UPDRS III scores) was observed in M-LOPD patients, but not in EOPD patients. CONCLUSION: Our study suggested that the iron deposition pattern was greatly influenced by the age of PD onset, which increases our understanding of the different pathological underpinnings of EOPD and M-LOPD patients.

Giant cell tumor of the tendon sheath: Magnetic resonance imaging findings in 38 patients.

The present study aimed to investigate the value of magnetic resonance imaging (MRI) in the diagnosis of giant cell tumor of the tendon sheath (GCTTS), including localized (L-) and diffuse (D-) types. A total of 38 patients with GCTTS, including 31 with L-GCTTS and 7 with D-GCTTS, diagnosed by surgery and pathology, were retrospectively analyzed. All patients underwent MRI examination. Of the 31 patients with L-GCTTS, the tumors were located in the hand and wrist (18 patients), the ankle and foot (10 cases), the knee joint (2 cases) and the temporomandibular joint (1 case). All 31 lesions were either located in relation to a tendon or were partially/completely enveloping it and all were well marginated. With respect to the 7 D-GCTTS patients, the tumors were located in the ankle and foot (6 cases) or the hand and wrist (1 cases). All 7 lesions presented as an aggressive soft tissue mass infiltrating the tendon sheath and adipose tissue around the joint. The characteristic internal signal of GCTTS, including L-GCTTS and D-GCTTS, was demonstrated by MRI examination. MRI is currently the optimal modality for preoperative assessment of tumor size, extent and invasion of adjacent joint and tenosynovial space.

The relationship between microvasculature in white matter hyperintensities and cognitive function.

White matter hyperintensities (WMHs) are associated with cognitive decline, but less is known about pathophysiology of cognitive decline in patients with WMHs. We investigated microvasculature and microstructure in WMHs using intravoxel incoherent motion (IVIM) and their associations with cognitive function. Thirty-two subjects with WMHs were enrolled in our study. Fast diffusion coefficient (D*), perfusion fraction (f) and slow diffusion coefficient (D) from IVIM model were compared between regions of WMHs (periventricular WMHs, PWMHs and deep WMHs, DWMHs) and surrounding normal white matter. Multivariate linear model was used to determine the independent factors associated with cognitive function assessed by the Mini Mental State Examination (MMSE) and the standardized coefficient (ß) of factors was estimated. D* was significantly lower (4.95 × 10-3 mm2/s versus 8.36 × 10-3 mm2/s in PWMHs and 5.04 × 10-3 mm2/s versus 8.67 × 10-3 mm2/s in DWMHs, both P < 0.001), and f (14.64 % versus 12.01 % in PWMHs and 14.26 % versus 11.31 % in DWMHs, both P < 0.001) and D (1.02 × 10-3 mm2/s versus 0.73 × 10-3 mm2/s in PWMHs and 0.86 × 10-3 mm2/s versus 0.70 × 10-3 mm2/s in DWMHs, both P < 0.001) were significantly higher in WMHs. Only f in PWMHs was independently associated with MMSE (ß = 0.443, P = 0.016). The decreased D* and increased D in WMHs were similar to previous findings. The increased f in PWMHs relating with better cognition provides the pathophysiological basis in understanding cognitive decline in patients with WMHs.

White matter injury induced by diabetes in acute stroke is clinically relevant: A preliminary study.

The importance of white matter injury induced by diabetes in stroke severity and prognosis is largely unknown. We aimed to investigate the relationship between diabetes-related white matter injury beyond stroke lesions with acute neurological deficits and clinical outcome after stroke. In total, 36 stroke patients within 3-7 days after onset were enrolled. Neurological deficits on admission were assessed by National Institute of Health Stroke Score, and poor outcome at 3 months was defined as modified Rankin score >2. White matter tracts were compared between patients with diabetic and non-diabetic stroke using fractional anisotropy from diffusion tensor imaging. Regional white matter abnormality with decreased fractional anisotropy was observed in diabetic patients (n = 18) when compared to non-diabetic patients (n = 18). Decreased fractional anisotropy in ipsilesional distal corticospinal tract was independently associated with higher National Institute of Health Stroke Score motor component score (ß = -0.444, p = 0.005), and decreased fractional anisotropy in contralesional superior longitudinal fasciculus I was independently related to poor outcome (odds ratio, 0.900; p = 0.033). Our findings suggested that only white matter injury induced by diabetes in specific tracts like corticospinal tract and superior longitudinal fasciculus beyond stroke lesions has clinically relevant, providing insight into the mechanism of stroke recovery under the diabetic condition.

A Comparison Study of Single-Echo Susceptibility Weighted Imaging and Combined Multi-Echo Susceptibility Weighted Imaging in Visualizing Asymmetric Medullary Veins in Stroke Patients.

BACKGROUND: Asymmetric medullary veins (AMV) are frequently observed in stroke patients and single-echo susceptibility weighted imaging (SWIs) is the main technique in detecting AMV. Our study aimed to investigate which echo time (TE) on single-echo susceptibility is the optimal echo for visualizing AMV and to compare the ability in detecting AMV in stroke patients between SWIs and multi-echo susceptibility weighted imaging (SWIc). MATERIALS AND METHODS: Twenty patients with middle cerebral artery stroke were included. SWI was acquired by using a multi-echo gradient-echo sequence with six echoes ranging from 5 ms to 35.240 ms. Three different echoes of SWIs including SWIs1 (TE = 23.144 ms), SWIs2 (TE = 29.192 ms) and SWIs3 (TE = 35.240 ms) were reconstructed. SWIc was averaged using the three echoes of SWIs. Image quality and venous contrast of medullary veins were compared between SWIs and SWIc using peak signal-to-noise ratio (PSNR), mean opinion score (MOS), contrast-to-noise ratio (CNR) and signal-to-noise ratio (SNR). The presence of AMV was evaluated in each SWIs (1-3) and SWIc. RESULTS: SWIs2 had the highest PSNR, MOS and CNR and SWIs1 had the highest SNR among three different echoes of SWIs. No significant difference was found in SNR between SWIs1 and SWIs2. PSNR, MOS and CNR in SWIc were significantly increased by 27.9%, 28.2% and 17.2% compared with SWIs2 and SNR in SWIc was significantly increased by 32.4% compared with SWIs1. 55% of patients with AMV were detected in SWIs2, SWIs3 and SWIc, while 50% AMV were found in SWIs1. CONCLUSIONS: SWIs using TE around 29ms was optimal in visualizing AMV. SWIc could improve image quality and venous contrast, but was equal to SWIs using a relative long TE in evaluating AMV. These results provide the technique basis for further research of AMV in stroke.

Decreased Inter-Hemispheric Functional Connectivity in Cognitively Intact Elderly APOE ɛ4 Carriers: A Preliminary Study.

The apolipoprotein E (APOE) ɛ4 allele is the best-known genetic risk factor for developing sporadic Alzheimer's disease (AD). According to neuroimaging studies, the APOE ɛ4 allele is associated with localized altered brain function. However, in long-range circuitry, APOE ɛ4 allele-related alterations in functional communication between hemispheres have rarely been directly investigated. We examined the alteration of resting-state functional connectivity (RSFC) between inter-hemispheric homotopic regions in cognitively intact, elderly APOE ɛ4 carriers. The voxel-mirrored homotopic connectivity method was used to assess the inter-hemispheric RSFC. The current study included 13 cognitively intact, elderly APOE ɛ4 carriers (with at least one copy of APOE ɛ4 allele) and 22 well-matched ɛ3 homozygotes. Comparisons between the two groups were conducted, and subsequently, the correlation between the differential inter-hemispheric RSFC and cognitive ability was analyzed. Compared with ɛ3 homozygotes, APOE ɛ4 carriers showed decreased inter-hemispheric RSFC in the bilateral medial temporal lobe (MTL) and orbital frontal cortex (OFC). Moreover, in APOE ɛ4 carriers, the inter-hemispheric RSFC of the MTL correlated with the Wechsler Memory Scale-Logical Memory (WMS-LM) (immediate and delayed performance, r = 0.64, p <  0.05; r = 0.65, p <  0.05, respectively), and the inter-hemispheric RSFC of the OFC correlated with the WMS-LM delayed performance (r = 0.71, p <  0.05). In our study, the presence of the APOE ɛ4 allele was linked with decreased inter-hemispheric RSFC, which was attributed to memory performance in carriers.

[Cerebral lesions of DWI hyperintensity in patients with subacute stroke assessed by intravoxel incoherent motion technique].

OBJECTIVE: To investigate the cerebral lesions of diffusion weighted imaging (DWI) hyperintensity in patients with subacute stroke with intravoxel incoherent motion (IVIM) technique. METHODS: The clinical data of 20 patients with ischemic stroke (3 to 7 d after onset) who underwent DWI and IVIM scanning between June 2014 and July 2015, were retrospectively analyzed. The parameters from IVIM including slow diffusion coefficient (D), fast diffusion coefficient (D(*)) and perfusion fraction (f) were processed. DWI hyperintensity was segmented by its signal intensity greater than the mean+2 standard deviations of the value in the homologous contralateral region. Then, DWI hyperintensity was classified into two regions of interest (ROIs): infarction core and peri-core with the ADC threshold of 0.55 × 10⁻³ mm²/s. The mirrored ROIs of infarction core and peri-core were also obtained. Then, we measured the values of ADC and D, D(*) and f in these ROIs. The ratios of ADC (rADC), D (rD), D(*) (rD(*)) and f (rf) were also calculated (e.g., rADC=ADCinfarction core/ADCmirrored region). RESULTS: Compared with mirrored region, ADC, D and f in the infarction core region decreased by 45% (P<0.001), 42% (P<0.001) and 32% (P<0.001), respectively; while ADC, D and f in the peri-core region decreased by 22% (P<0.001), 32% (P<0.001) and 8% (P=0.009), respectively. The values of rADC, rD, rD(*) and rf in the infarction core region were significantly lower than those in the peri-core region (all P<0.001). Pearson analysis showed that rADC was positively correlated with rf in the peri-core region (r=0.467, P=0.038). CONCLUSION: During subacute stage of stroke, compared to the infarction core region within DWI hyperintensity, D and f increase in the peri-core region of DWI hyperintensity, reflecting the increased water diffusion in microstructure and perfusion volume in microvasculature. This result shows that the potential reason for the heterogeneous ADC signal is associated with the disappearance of cellular edema and microvascular compensatory with increased blood volume.