DeepKEGG: a multi-omics data integration framework with biological insights for cancer recurrence prediction and biomarker discovery.

Deep learning-based multi-omics data integration methods have the capability to reveal the mechanisms of cancer development, discover cancer biomarkers and identify pathogenic targets. However, current methods ignore the potential correlations between samples in integrating multi-omics data. In addition, providing accurate biological explanations still poses significant challenges due to the complexity of deep learning models. Therefore, there is an urgent need for a deep learning-based multi-omics integration method to explore the potential correlations between samples and provide model interpretability. Herein, we propose a novel interpretable multi-omics data integration method (DeepKEGG) for cancer recurrence prediction and biomarker discovery. In DeepKEGG, a biological hierarchical module is designed for local connections of neuron nodes and model interpretability based on the biological relationship between genes/miRNAs and pathways. In addition, a pathway self-attention module is constructed to explore the correlation between different samples and generate the potential pathway feature representation for enhancing the prediction performance of the model. Lastly, an attribution-based feature importance calculation method is utilized to discover biomarkers related to cancer recurrence and provide a biological interpretation of the model. Experimental results demonstrate that DeepKEGG outperforms other state-of-the-art methods in 5-fold cross validation. Furthermore, case studies also indicate that DeepKEGG serves as an effective tool for biomarker discovery. The code is available at https://github.com/lanbiolab/DeepKEGG.

IBPGNET: lung adenocarcinoma recurrence prediction based on neural network interpretability.

Lung adenocarcinoma (LUAD) is the most common histologic subtype of lung cancer. Early-stage patients have a 30-50% probability of metastatic recurrence after surgical treatment. Here, we propose a new computational framework, Interpretable Biological Pathway Graph Neural Networks (IBPGNET), based on pathway hierarchy relationships to predict LUAD recurrence and explore the internal regulatory mechanisms of LUAD. IBPGNET can integrate different omics data efficiently and provide global interpretability. In addition, our experimental results show that IBPGNET outperforms other classification methods in 5-fold cross-validation. IBPGNET identified PSMC1 and PSMD11 as genes associated with LUAD recurrence, and their expression levels were significantly higher in LUAD cells than in normal cells. The knockdown of PSMC1 and PSMD11 in LUAD cells increased their sensitivity to afatinib and decreased cell migration, invasion and proliferation. In addition, the cells showed significantly lower EGFR expression, indicating that PSMC1 and PSMD11 may mediate therapeutic sensitivity through EGFR expression.

[The value of diffusion tensor imaging and fiber tractography in cervical spondylotic myelopathy].

OBJECTIVE: To study the diagnostic value of diffusion tensor imaging (DTI) in cervical spondylotic myelopathy. METHODS: Twenty healthy volunteers and fifty patients with cervical spondylotic myelopathy underwent DTI in the Affiliated Hospital of Medical College of Ningbo University from January 2014 to April 2015. Healthy volunteers served as controls. Fifty patients were divided into three groups (group A , B, C) according to cervical MRI scan standard. Group A (17 cases) had only the dura mater spinalis compressed; Group B (23 cases) showed the cervical spinal cord compressed, but no high signal in it; Group C (10 cases) had the cervical spinal cord compressed with high signal in the same level. The average apparent diffusion coefficients(ADC) and fractional anisotropy (FA)values in these examinee were analyzed and all subjects were performed fiber tracking. RESULTS: There was no statistically significant differences in ADC and FA values in C2/C3, C3/C4, C4/C5, C5/C6, C6/C7 of control group (P>0.05). The average ADC and FA values in control group were (0.875 +/- 0.096) x10(3) mm2/s and 0.720 +/- 0.051, respectively; compared with group A,there was no statistically significant difference; compared with group B and C, there was significant difference; comparison among group A, B, C, there was significant differences. CONCLUSION: DTI can early and accurately quantify the changes of microstructure in cervical spondylotic myelopathy. Fiber tracking can show the damage range of spinal cord lesions.

Using R2* values to evaluate brain tumours on magnetic resonance imaging: preliminary results.

OBJECTIVE: To determine the usefulness of the R2* value in assessing the histopathological grade of glioma at magnetic resonance imaging and differentiating various brain tumours. METHODS: Sixty-four patients with brain tumours underwent R2* mapping and diffusion-weighted imaging examinations. ANOVA was performed to analyse R2* values among four groups of glioma and among high-grade gliomas (grades III and IV), low-grade gliomas (grades I and II), meningiomas, and brain metastasis. Spearman's correlation coefficients were used to determine the relationships between the R2* values or apparent diffusion coefficient (ADC) and the histopathological grade of gliomas. R2* values of low- and high-grade gliomas were analysed with the receiver-operator characteristic curve. RESULTS: R2* values were significantly different among high-grade gliomas, low-grade gliomas, meningiomas, and brain metastasis, but not between grade I and grade II or between grade III and grade IV. The R2* value (18.73) of high-grade gliomas provided a very high sensitivity and specificity for differentiating low-grade gliomas. A strong correlation existed between the R2* value and the pathological grade of gliomas. CONCLUSIONS: R2* mapping is a useful sequence for determining grade of gliomas and in distinguishing benign from malignant tumours. R2* values are better than ADC for characterising gliomas. KEY POINTS: • Magnetic resonance imaging parameters are increasingly used to assess cerebral lesions. • R2* values are better than diffusion weighting for characterising gliomas. • R2* values can help distinguish among different grades of glioma. • Significant difference existed in R2* values between high- and low-grade gliomas.

A case-control study on risk factors of breast cancer in China.

INTRODUCTION: To screen the risk factors associated with breast cancer among Chinese women in order to evaluate the individual risk of developing breast cancer among women in China. MATERIAL AND METHODS: A case-control study on 416 breast cancer patients and 1156 matched controls was conducted in 14 hospitals in 8 provinces of China in 2008. Controls were age- and region-matched to the cases. Clinicians conducted in-person interviews with the subjects to collect information on demographics and suspected risk factors for breast cancer that are known worldwide. Conditional logistic regression was used to derive odds ratios (OR) and 95% confidence intervals (CI) for the associations between risk factors and breast cancer. RESULTS: Compared with matched controls, women with breast cancer were significantly more likely to have higher body mass index (BMI, OR = 4.07, 95% CI: 2.98-5.55), history of benign breast disease (BBD) biopsy (OR = 1.68, 95% CI: 1.19-2.38), older age of menarche (AOM) (OR = 1.41, 95% CI: 1.07-1.87), stress anticipation (SA), for grade 1-4, OR = 2.15, 95% CI: 1.26-3.66; for grade 5-9, OR = 3.48, 95% CI: 2.03-5.95) and menopause (OR = 2.22, 95% CI: 1.50-3.282) at the level of p < 0.05. Family history of breast cancer (FHBC) in first-degree relatives (OR = 1.66, 95% CI: 0.77-3.59) and use of oral contraceptives (OC) (OR = 1.59, 95% CI: 0.83-3.05) were associated with an increased risk of breast cancer at the level of p < 0.20. CONCLUSIONS: Our results showed that BMI, history of BBD biopsy, older AOM, SA and menopause were associated with increased risk of breast cancer among Chinese women. The findings derived from the study provided some suggestions for population-based prevention and control of breast cancer in China.