Comparing Radiomics features of tumour and healthy liver tissue in a limited CT dataset: A machine learning study

IntroductionLiver cancer lesions on Computed Tomography (CT) withholds a great amount of data, which is not visible to the radiologists and radiographer. Radiomics features can be extracted from the lesions and used to train Machine Learning (ML) algorithms to predict between tumour and liver tissue. The purpose of this study was to investigate and classify Radiomics features extracted from liver tumours and normal liver tissue in a limited CT dataset.MethodsThe Liver Tumour Segmentation Benchmark (LiTS) dataset consisting of 131 CT scans of the liver with segmentations of tumour tissue and healthy liver was used to extract Radiomic features. Extracted Radiomic features included size, shape, and location extracted with morphological and statistical techniques according to the International Symposium on Biomedical Imaging manual. Relevant features was selected with chi2 correlation and principal component analysis (PCA) with tumour and healthy liver tissue as outcome according to a consensus between three experienced radiologists. Logistic regression, random forest and support vector machine was used to train and validate the dataset with a 10-fold cross-validation method and the Grid Search as hyper-parameter tuning. Performance was evaluated with sensitivity, specificity and accuracy.ResultsThe performance of the ML algorithms achieved sensitivities, specificities and accuracy ranging from 96.30% (95% CI: 81.03%–99.91%) to 100.00% (95% CI: 86.77%–100.00%), 91.30% (95% CI: 71.96%–98.93%) to 100.00% (95% CI: 83.89%–100.00%)and 94.00% (95% CI: 83.45%–98.75%) to 100.00% (95% CI: 92.45%–100.00%), respectively.ConclusionML algorithms classifies Radiomics features extracted from healthy liver and tumour tissue with perfect accuracy. The Radiomics signature allows for a prognostic biomarker for hepatic tumour screening on liver CT.Implications for practiceDifferentiation between tumour and liver tissue with Radiomics ML algorithms have the potential to increase the diagnostic accuracy, assist in the decision-making of supplementary multiphasic enhanced medical imaging, as well as for developing novel prognostic biomarkers for liver cancer patients.     

Do social support and community engagement act as mechanisms in the association between neighbourhood income inequality and the mental health of mothers in Calgary,Canada? A mediation analysis

PurposeAccording to the social determinants of health framework, income inequality is a potential risk factor for adverse mental health. However, few studies have explored the mechanisms suspected to mediate this relationship. The current study addresses this gap through a mediation analysis to determine if social support and community engagement act as mediators linking neighbourhood income inequality to maternal anxiety and depressive symptoms within a cohort of new mothers living in the City of Calgary, Canada.MethodsData collected at three years postpartum from mothers belonging to the All Our Families (AOF) cohort were used in the current study. Maternal data were collected between 2012 and 2015 and linked to neighbourhood socioeconomic data from the 2006 Canadian Census. Income inequality was measured using Gini coefficients derived from 2006 after-tax census data. Generalized structural equation models were used to quantify the associations between income inequality and mental health symptoms, and to assess the potential direct and indirect mediating effects of maternal social support and community engagement.ResultsIncome inequality was not significantly associated with higher depressive symptoms (β = 0.32, 95%CI = −0.067, 0.70), anxiety symptoms (β = 0.11, 95%CI = −0.39, 0.60), or lower social support. Income inequality was not associated with community engagement. For the depression models, higher social support was significantly associated with lower depressive symptoms (β = −0.13, 95%CI = −0.15, −0.097), while community engagement was not significantly associated with depressive symptoms (β = 0.059, 95%CI = −0.15, 0.27). Similarly, for the anxiety models, lower anxiety symptoms were significantly associated with higher levels of social support (β = −0.17, 95%CI = −0.20, −0.13) but not with higher levels of community engagement (β = 0.14, 95%CI = −0.14, 0.41).ConclusionThe current study did not find clear evidence for social support or community engagement mediating the relationship between neighbourhood income inequality and maternal mental health. Future investigations should employ a broader longitudinal approach to capture changes in income inequality, potential mediators, and mental health symptomatology over time.     

Pediatric button battery ingestion: A single center experience and risk score to predict severe outcomes

PurposeThe purpose of this study was to analyze the management and outcomes of primary button battery ingestions and their sequelae at a single high-volume center, and to propose a risk score to predict the likelihood of a severe outcome.MethodsThe medical record was queried for all patients under 21 years old evaluated at our institution for button battery ingestion from 2008 to 2021. A severe outcome was defined as having at least one of the following: deep/circumferential mucosal erosion, perforation, mediastinitis, vascular or airway injury/fistula, or development of esophageal stricture. From a selection of clinically relevant factors, logistic regression determined predictors of a severe outcome, which were incorporated into a risk model.Results143 patients evaluated for button battery ingestion were analyzed. 24 (17%) had a severe outcome. The independent predictors of a severe outcome in multivariate analysis were location of battery in the esophagus on imaging (96%), battery size >/ = 2 cm (95%), and presence of any symptoms on presentation (96%), with P < 0.001 in all cases. Predicted probability of a severe outcome ranged from 88% when all three risk factors were observed, to 0.3% when none were present.ConclusionWe report the presentation, management, and complication profiles of a large cohort of BB ingestions treated at a single institution. A risk score to predict severe outcomes may be used by providers initially evaluating patients with button battery ingestion in order to allocate resources and expedite transfer to a center with pediatric endoscopic and surgical capabilities.Level of evidenceLevel IV.Type of studyClinical Research Paper.     

特殊涉核环境人员黄斑区厚度研究

目的 探讨特殊涉核环境人员与普通环境非涉核人员黄斑区厚度的差异性。方法 在某基地年度体检时为特殊涉核环境人员（简称涉核人员）与普通环境非涉核人员（简称非涉核人员）增加眼科光学相干断层扫描（Optical coherence tomography，OCT）检查，测量黄斑区厚度值，每人测量3次，按照体检表末尾编号随机抽取204例编号为奇数的涉核人员和105例非涉核人员，取3次测量值的平均值进行比较。结果 涉核人员与非涉核人员两组年龄均值为（29.40 ±6.16）岁、（28.92 ±6.71）岁，P = 0.5325，无统计学差异。两组人员视力均值：涉核人员为（1.04 ±0.03）（右），（1.05 ±0.03）（左），非涉核人员为（1.00 ±0.05）（右），（1.02 ±0.05）（左），P = 0.5006（右），P = 0.5962（左），无统计学差异。涉核人员黄斑区厚度均值：（212.9 ±1.3） μm（右），（205.5 ±1.1） μm（左），P < 0.0001，两者比较差异有统计学意义；非涉核人员黄斑区厚度均值：（223.2 ±2.5） μm（右），（211.7 ±2.4） μm（左），P < 0.0001，两者比较有统计学差异。对涉核人员与非涉核人员右眼及左眼黄斑区厚度均值进行比较，P = 0.0003（右），P = 0.0217（左），均有统计学差异。结论 非涉核人员黄斑区厚度要厚于涉核人员。与涉核人员特殊工作环境中较多使用LED光源可能存在关联；其他因素包括涉核、气压、氧分压及人员的心理状态需要进一步的试验进行证明。     

Pharmacokinetics and pharmacodynamics of daptomycin in a clinical setting

We investigated achievement of a target 24-h area under the concentration-time curve to minimum inhibitory concentration ratio (AUC/MIC) ≥666 and the factors influencing this ratio in patients who received daptomycin (DAP) for infectious disease treatment in a clinical setting. The target AUC/MIC was obtained in 6 patients (35.3%) at a 4–6 mg/kg dose (Group_{_4–6 mg/kg}) and in 4 (18.2%) at a >6 mg/kg dose (Group_{_>6 mg/kg}). There was a significant difference in clearance of DAP (CL_{_DAP}) between these groups, but no other difference in characteristics. Multiple linear regression analysis was performed for prediction of AUC ≥666 based on patient factors and the presence or absence of sepsis. In a stepwise analysis, serum creatinine (SCr) was a significant predictor of AUC, but this parameter explained only 13% of the variance in achievement of the target AUC. These results show that the target AUC/MIC may or may not be achieved at the doses used in Group_{_4–6 mg/kg} and Group_{_>6 mg/kg}. Receiver operating characteristic analysis suggested that a CL_{_DAP} >0.450 L/hr may lead to failure to reach the target AUC/MIC. Therefore, regardless of dose, the efficacy of DAP should be monitored closely to prevent failure of infectious disease treatment, particularly because therapeutic drug monitoring of DAP is limited by difficulty measuring the DAP serum concentration at many medical facilities. Our findings are preliminary, and a further study is required to identify factors that increase CL_{_DAP} and to enable dose adjustment of DAP.     

Radiomics for classification of bone mineral loss: A machine learning study

PurposeThe purpose of this study was to develop predictive models to classify osteoporosis, osteopenia and normal patients using radiomics and machine learning approaches.Materials and methodsA total of 147 patients were included in this retrospective single-center study. There were 12 men and 135 women with a mean age of 56.88 ± 10.6 (SD) years (range: 28–87 years). For each patient, seven regions including four lumbar and three femoral including trochanteric, intertrochanteric and neck were segmented on bone mineral densitometry images and 54 texture features were extracted from the regions. The performance of four feature selection methods, including classifier attribute evaluation (CLAE), one rule attribute evaluation (ORAE), gain ratio attribute evaluation (GRAE) and principal components analysis (PRCA) along with four classification methods, including random forest (RF), random committee (RC), K-nearest neighbor (KN) and logit-boost (LB) were evaluated. Four classification categories, including osteopenia vs. normal, osteoporosis vs. normal, osteopenia vs. osteoporosis and osteoporosis + osteopenia vs. osteoporosis were examined for the defined seven regions. The classification model performances were evaluated using the area under the receiver operator characteristic curve (AUC).ResultsThe AUC values ranged from 0.50 to 0.78. The combination of methods RF + CLAE, RF + ORAE and RC + ORAE yielded highest performance (AUC = 0.78) in discriminating between osteoporosis and normal state in the trochanteric region. The combinations of RF + PRCA and LB + PRCA had the highest performance (AUC = 0.76) in discriminating between osteoporosis and normal state in the neck region.ConclusionThe machine learning radiomic approach can be considered as a new method for bone mineral deficiency disease classification using bone mineral densitometry image features.