Lymph node metastasis prediction and biological pathway associations underlying DCE-MRI deep learning radiomics in invasive breast cancer.

BACKGROUND: The relationship between the biological pathways related to deep learning radiomics (DLR) and lymph node metastasis (LNM) of breast cancer is still poorly understood. This study explored the value of DLR based on dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in LNM of invasive breast cancer. It also analyzed the biological significance of DLR phenotype based on genomics. METHODS: Two cohorts from the Cancer Imaging Archive project were used, one as the training cohort (TCGA-Breast, n = 88) and one as the validation cohort (Breast-MRI-NACT Pilot, n = 57). Radiomics and deep learning features were extracted from preoperative DCE-MRI. After dual selection by principal components analysis (PCA) and relief methods, radiomics and deep learning models for predicting LNM were constructed by the random forest (RF) method. A post-fusion strategy was used to construct the DLR nomograms (DLRNs) for predicting LNM. The performance of the models was evaluated using the receiver operating characteristic (ROC) curve and Delong test. In the training cohort, transcriptome data were downloaded from the UCSC Xena online database, and biological pathways related to the DLR phenotypes were identified. Finally, hub genes were identified to obtain DLR gene expression (RadDeepGene) scores. RESULTS: DLRNs were based on area under curve (AUC) evaluation (training cohort, AUC = 0.98; validation cohort, AUC = 0.87), which were higher than single radiomics models or GoogLeNet models. The Delong test (radiomics model, P = 0.04; GoogLeNet model, P = 0.01) also validated the above results in the training cohorts, but they were not statistically significant in the validation cohort. The GoogLeNet phenotypes were related to multiple classical tumor signaling pathways, characterizing the biological significance of immune response, signal transduction, and cell death. In all, 20 genes related to GoogLeNet phenotypes were identified, and the RadDeepGene score represented a high risk of LNM (odd ratio = 164.00, P < 0.001). CONCLUSIONS: DLRNs combining radiomics and deep learning features of DCE-MRI images improved the preoperative prediction of LNM in breast cancer, and the potential biological characteristics of DLRN were identified through genomics.

Prediction of Lymphovascular invasion status in breast cancer based on magnetic resonance imaging radiomics features.

OBJECTIVE: This study intended to investigate the feasibility and effectiveness of using clinical magnetic resonance imaging (MRI) radiomics features to predict lymphovascular invasion (LVI) status in breast cancer (BC) patients. METHODS: A total of 182 BC patients were retrospectively collected and randomly divided into a training set (n = 127) and a validation set (n = 55) in a 7:3 ratio. Based on pathological examination results, the training set was further divided into LVI group (n = 60) and non-LVI group (n = 67), and the validation set was divided into LVI group (n = 24) and non-LVI group (n = 31). General data and MRI examination indicators were compared. Multivariate logistic regression was utilized to analyze MRI radiomics features and clinically relevant indicators that were significant in the baseline information of patients in training set, independent risk factors were identified, and a logistic regression model was built. The accuracy of logistic model was validated using ROC curves in training and validation sets. RESULTS: Age, pathohistological classification, tumor length, tumor width, presence or absence of Magnetic Resonance Spectroscopy (MRS) cho peak, presence or absence of spicule sign, peritumoral enhancement, and peritumoral edema were statistically significant (P < 0.05) between the two groups. Multivariate logistic regression analysis presented that spicule and peritumoral edema were independent risk factors for LVI in BC patients (P < 0.05). The ROC curve illustrated that AUC of the logistic regression model in the training set was 0.807 (95%CI: 0.730-0.885) and that in the validation set was 0.837 (95%CI: 0.731-0.944). CONCLUSION: Radiomics features of spicule sign and peritumoral edema were independent risk factors for LVI in BC patients. A logistic regression model based on these factors, along with age, could accurately predict LVI occurrence in BC patients, providing data support for diagnosis and modeling of LVI in BC patients.

Automated ASPECTS calculation may equal the performance of experienced clinicians: a machine learning study based on a large cohort.

OBJECTIVES: The Alberta Stroke Program Early CT Score (ASPECTS) is a semi-quantitative method to evaluate the severity of early ischemic change on non-contrast computed tomography (NCCT) in patients with acute ischemic stroke (AIS). In this work, we propose an automated ASPECTS method based on large cohort of data and machine learning. METHODS: For this study, we collected 3626 NCCT cases from multiple centers and annotated directly on this dataset by neurologists. Based on image analysis and machine learning methods, we constructed a two-stage machine learning model. The validity and reliability of this automated ASPECTS method were tested on an independent external validation set of 300 cases. Statistical analyses on the total ASPECTS, dichotomized ASPECTS, and region-level ASPECTS were presented. RESULTS: On an independent external validation set of 300 cases, for the total ASPECTS results, the intraclass correlation coefficient between automated ASPECTS and expert-rated was 0.842. The agreement between ASPECTS threshold of ≥ 6 versus < 6 using a dichotomized method was moderate (κ = 0.438, 0.391-0.477), and the detection rate (sensitivity) was 86.5% for patients with ASPECTS threshold of ≥ 6. Compared with the results of previous studies, our method achieved a slight lead in sensitivity (67.8%) and AUC (0.845), with comparable accuracy (78.9%) and specificity (81.2%). CONCLUSION: The proposed automated ASPECTS method driven by a large cohort of NCCT images performed equally well compared with expert-rated ASPECTS. This work further demonstrates the validity and reliability of automated ASPECTS evaluation method. CLINICAL RELEVANCE STATEMENT: The automated ASPECTS method proposed by this study may help AIS patients to receive rapid intervention, but should not be used as a stand-alone diagnostic basis. KEY POINTS: NCCT-based manual ASPECTS scores were poorly consistent. Machine learning can automate the ASPECTS scoring process. Machine learning model design based on large cohort data can effectively improve the consistency of ASPECTS scores.

Genetic contribution of synapse-associated protein 97 to cerebellar functional connectivity changes in first-episode schizophrenia.

Our previous study data suggested that the synapse-associated protein 97 (SAP97) rs3915512 polymorphism is significantly related to clinical performance in schizophrenia. The cerebellum exhibits abundant expression of SAP97, which is involved with negative symptoms, cognition and emotion in schizophrenia. As functional dysconnectivity with the cortical-subcortical-cerebellar circuitry has been widely shown in patients with schizophrenia, cortical-subcortical-cerebellar dysconnectivity can therefore be considered a possible intermediate phenotype that connects risk genes with schizophrenia. In this study, resting-state functional magnetic resonance imaging (fMRI) was applied to evaluate whether the SAP97 rs3915512 polymorphism changes cortical/subcortical-cerebellar resting-state functional connectivity (RSFC) in 104 Han Chinese subjects (52 first-episode schizophrenia (FES) patients and 52 matched healthy controls (HCs)). To examine RSFC between cortical/subcortical regions and the cerebellum, a ROI (region of interest)-wise functional connectivity analysis was conducted. The association between abnormal cortical/subcortical-cerebellar connectivity and clinical manifestation was further assessed in FES patients with different genotypes. The interactive effect of disease and genotype on RSFC was found between the frontal gyrus (rectus) and cerebellum. A positive correlation was suggested between RSFC in the cerebellum and the hostility scores in FES patients with the A allele, and no correlation was found in FES patients with the TT genotype. The current findings identified that SAP97 may be involved in the process of mental symptoms in FES patients via cerebellar connectivity depending on the rs3915512 polymorphism genotype.

Feasibility evaluation of intravoxel incoherent motion diffusion-weighted imaging in the diagnosis of skull-base invasion in nasopharyngeal carcinoma.

Objective: This study aimed to evaluate the feasibility of intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) in the diagnosis of skull-base invasion (SBI) in nasopharyngeal carcinoma (NPC). Materials and methods: A total of 50 patients pathologically diagnosed with NPC and a group of 40 controls comprised of those with either normal nasopharynx or patients with nasopharyngitis underwent conventional MRI and IVIM-DWI scans with 3 different groups of b values. Among the 50 patients, 36 patients diagnosed with SBI in NPC were included in the case group according to SBI criteria. All subjects (including those in the control group and case group) were divided into the b1, b2, and b3 groups based on their b values. The pure diffusion coefficient (D), perfusion-related incoherent microcirculation (D*), and microvascular volume fraction (f) values obtained in each measurement area of each group were tested for variance. Next,2 groups of b-value parameters with statistically significant data in the 3 groups were randomly selected for use in both the control group and the case group. A t-test was performed on the D, D*, and f values obtained by measuring each area of the skull base, and the area under the curve (AUC) of the receiver operating characteristic (ROC) was used to evaluate the diagnostic efficacy of the D, D*, and f values. Results: There was no statistical significance among the D, D*, and f values of the b1 and b3 groups (P>0.05), and the differences in parameters between the b1 and b2 groups were statistically significant(P < 0.05),and the differences in parameters between the b3 and b2 groups were also statistically significant(P < 0.05).The f value of the case group, which was obtained using the b1 and b2 parameters in each area of the skull base, was lower than that of the control group (P <0.05).The D, D*, and f values of the case group obtained by the b1 and b2 parameters in the pars petrosa of the temporal bone (including the foramen lacerum) were lower than those of the control group (P<0.05).When the parameters of the b1 group were used in the corpus of sphenoid bone (including the foramen ovale), the D, D*, and f values of the control group and the case group were compared, yielding a statistically significant difference (P<0.05).When the parameters of the b1 group were used, the diagnostic efficacy of the f value in each area of the skull base was the highest (AUC=0.908-0.991), followed by the D* value (AUC=0.624-0.692). Conclusion: When the number of b values <200 s/mm2 in IVIM-DWI accounts for more than half of the selected b values, IVIM-DWI is highly stable for the diagnosis of SBI in NPC. The D, D*, and f values of the bone and muscle areas of the skull base in patients with SBI of NPC showed a downward trend, and the f value had the best diagnostic performance, followed by the D* value, while the D value had the worst. Thus, IVIM-DWI can be used as a noninvasive method in the diagnosis of SBI in NPC.

Genetic variations in DOCK4 contribute to schizophrenia susceptibility in a Chinese cohort: A genetic neuroimaging study.

BACKGROUND: Emerging evidence suggests that the DOCK4 gene increases susceptibility to schizophrenia. However, no study has hitherto repeated this association in Chinese, and further investigated the relationship between DOCK4 and clinical symptoms in schizophrenic patients using clinical scales and functional magnetic resonance imaging (fMRI). METHODS: In this study, we genotyped three single nucleotide polymorphisms (SNPs) (rs2074127, rs2217262, and rs2074130) within the DOCK4 gene using a case-control design (including 1289 healthy controls and 1351 patients with schizophrenia). 55 first-episode schizophrenia (FES) patients and 59 healthy participants were divided by the genotypes of rs2074130 into CC and CT+TT groups. We further investigated the association with clinical symptoms and neural characteristics (brain activation/connectivity and nodal network metrics). RESULTS: Our results showed significant associations between all selected SNPs and schizophrenia (all P < 0.05). In patients, letter fluency and motor speed scores of T allele carriers were significantly higher than the CC group (all P < 0.05). Interestingly, greater brain activity, functional connectivity, and betweenness centrality (BC) in language processing and motor coordination were also observed in the corresponding brain zones in patients with the T allele based on a two-way ANCOVA model. Moreover, a potential positive correlation was found between brain activity/connectivity of these brain regions and verbal fluency and motor speed. CONCLUSION: Our findings suggest that the DOCK4 gene may contribute to the onset of schizophrenia and lead to language processing and motor coordination dysfunction in this patient population from China.

Glyoxalase 1 Confers Susceptibility to Schizophrenia: From Genetic Variants to Phenotypes of Neural Function.

This research aimed to investigate the role of glyoxalase 1 (Glo-1) polymorphisms in the susceptibility of schizophrenia. Using the real-time polymerase chain reaction (PCR) and spectrophotometric assays technology, significant differences in Glo-1 messenger ribonucleic acid (mRNA) expression (P = 3.98 × 10-5) and enzymatic activity (P = 1.40 × 10-6) were found in peripheral blood of first-onset antipsychotic-naïve patients with schizophrenia and controls. The following receiver operating characteristic (ROC) curves analysis showed that Glo-1 could predict the schizophrenia risk (P = 4.75 × 10-6 in mRNA, P = 1.43 × 10-7 in enzymatic activity, respectively). To identify the genetic source of Glo-1 risk in schizophrenia, Glo-1 polymorphisms (rs1781735, rs1130534, rs4746, and rs9470916) were genotyped with SNaPshot technology in 1,069 patients with schizophrenia and 1,023 healthy individuals. Then, the impact of risk polymorphism on the promoter activity, mRNA expression, and enzymatic activity was analyzed. The results revealed significant differences in the distributions of genotype (P = 0.020, false discovery rate (FDR) correction) and allele (P = 0.020, FDR correction) in rs1781735, in which G > T mutation significantly showed reduction in the promoter activity (P = 0.016), mRNA expression, and enzymatic activity (P = 0.001 and P = 0.015, respectively, GG vs. TT, in peripheral blood of patients with schizophrenia) of Glo-1. The expression quantitative trait locus (eQTL) findings were followed up with the resting-state functional magnetic resonance imaging (fMRI) analysis. The TT genotype of rs1781735, associated with lower RNA expression in the brain (P < 0.05), showed decreased neuronal activation in the left middle frontal gyrus in schizophrenia (P < 0.001). In aggregate, this study for the first time demonstrates how the genetic and biochemical basis of Glo-1 polymorphism culminates in the brain function changes associated with increased schizophrenia risk. Thus, establishing a combination of multiple levels of changes ranging from genetic variants, transcription, protein function, and brain function changes is a better predictor of schizophrenia risk.

Genetic Contribution of Synapse-Associated Protein 97 to Orbitofrontal-Striatal-Thalamic Circuitry Connectivity Changes in First-Episode Schizophrenia.

Functional and structural disturbances in the orbitofrontal-striatal-thalamic circuitry are thought to be associated with mental symptoms and neurocognitive impairments in schizophrenia. This study tested whether synapse-associated protein 97 (SAP97), a reasonable candidate gene for schizophrenia, is related to orbitofrontal-striatal-thalamic connection changes in first-episode schizophrenia (FES) patients and the clinical performance of schizophrenic patients by affecting this integrity. Fifty-two FES patients and 52 matched healthy controls were recruited. All subjects underwent genotyping via the improved multiplex ligation detection reaction technique and scanning with magnetic resonance imaging (MRI) to provide orbitofrontal-striatal-thalamic functional and structural imaging data. A two-way analysis of covariance model was employed to examine abnormal brain connectivities, and Spearman correlations were applied to estimate the relationships between brain connectivity and clinical manifestations. In the FES group, those with the SAP97 rs3915512 TT genotype showed lower structural and functional connectivity than A allele carriers between the orbitofrontal gyrus and striatum/thalamus. In the FES group, negative correlations were found between resting-state functional connectivity (RSFC) in the orbitofrontal gyrus and thalamus, and positive symptoms between structural connections in the orbitofrontal gyrus and striatum and cognitive functions, and positive correlations were suggested between RSFC in the orbitofrontal gyrus and thalamus and negative symptoms. Our findings suggested that the SAP97 rs3915512 polymorphism may be involved in mental symptoms and cognitive dysfunction in FES patients by influencing structural and functional connectivity of the orbitofrontal-striatal and orbitofrontal-thalamic regions.

SAP97 rs3915512 Polymorphism Affects the Neurocognition of Schizophrenic Patients: A Genetic Neuroimaging Study.

Our previous study suggested that the synapse-associated protein 97 (SAP97) gene rs3915512 polymorphism may influence neurocognition in schizophrenia patients. Neuroimaging studies have shown a possible association between cognitive function and brain activity/connectivity. Considering the poor understanding of whether the disease state and SAP97 rs3915512 polymorphism have interactive effects on brain activity/connectivity, 52 first-episode schizophrenia (FES) patients and 52 healthy controls were genotyped using blood DNA samples and underwent magnetic resonance imaging scanning. A two-way ANCOVA model was performed with rs3915512 genotypes and disease state as the between-subject factors. A significant disease × SAP97 interactive effect was found for the amplitude of low-frequency fluctuation (ALFF) in the right supplementary motor area, left rolandic opercularis area (ROC-L), and bilateral middle occipital gyrus (MOG). In addition, among auditory/visual-related brain areas, a significant interactive effect was found for resting-state functional connectivity (RSFC) between the MOG-L and bilateral superior temporal gyrus (STG) in the STG-L with ROC-R, right cuneus (Cu-R), left fusiform (Fu-L), and left lingual gyrus (LG-L). Positive correlations were found between ALFF in the ROC-L and motor speed scores, between RSFC in the STG-L and LG-L and between Brief Assessment of Cognition in Schizophrenia verbal memory scores in FES. The SAP97 rs3915512 polymorphism may affect neurocognitive function in patients with schizophrenia by changing the brain activity and connectivity of auditory/visual-related brain areas.

Clinical presentation and imaging characteristics of occult lung cancer associated ischemic stroke.

We investigated the clinical and imaging characteristics of initial and recurrent strokes in patients with occult lung cancer associated ischemic stroke (OLCA-stroke). A retrospective review of all ischemic stroke patients with occult lung cancer in the absence of conventional stroke etiologies between 2005 and 2013 was conducted. We compared the initial and recurrent lesion patterns on diffusion-weighted MRI in patients with OLCA-stroke, with respect to vascular territory involved, number and size of lesions, clinical presentation, cancer subtypes, recurrences and fatalities, and outcome of survivors. Thirteen patients with confirmed OLCA-stroke were identified. All had elevated D-dimer levels, six had central lung cancer and seven had peripheral lung cancer. Eight (62%) had adenocarcinoma, and nine (69%) had metastasis. Ten (77%) patients had multiple lesions in multiple vascular territories. Twelve (92%) patients suffered recurrent strokes. Multiple small and large disseminated lesions in multiple vascular territories were more frequent in recurrent strokes in comparison with initial strokes. The middle cerebral artery was most frequently involved in recurrent strokes, followed by the posterior circulation territory and anterior cerebral artery, which were of similar frequency as initial strokes. Overall, 58% of patients had their first recurrent stroke within the first month, and 69% had a poor outcome, especially for those with multiple recurrent strokes and metastases. Occult cancer should be considered in the setting of multiple and recurrent embolic strokes within the short term in the absence of conventional stroke etiologies. The severity of malignancy and cancer treatments and stroke influenced the recurrences and outcome.