Computed tomography (CT)-guided pulmonary core biopsies of small pulmonary nodules less than 15 millimeters (mm) are challenging for radiologists, and their diagnostic accuracy has been shown to be variable in previous studies. Common complications after the procedure include pneumothorax and pulmonary hemorrhage. The present study compared the diagnostic accuracy of small and large lesions using CT-guided core biopsies and identified the risk factors associated with post-procedure complications. Between January 1, 2016, and December 31, 2017, 198 CT-guided core biopsies performed on 195 patients at our institution were retrospectively enrolled. The lesions were separated into group A (< or = 15 mm) and group B (> 15 mm) according to the longest diameter of the target lesions on CT. Seventeen-gauge introducer needles and 18-gauge automated biopsy instruments were coaxially used for the biopsy procedures. The accuracy and complications, including pneumothorax and pulmonary hemorrhage, of the procedures of each group were recorded. The risk factors for pneumothorax and pulmonary hemorrhage were determined using univariate analysis of variables. The diagnostic accuracies of group A (n = 43) and group B (n = 155) were 83.7 % and 96.8 %, respectively (p = 0.005). The risk factors associated with post-biopsy pneumothorax were longer needle path length from the pleura to the lesion (p = 0.020), lesion location in lower lobes (p = 0.002), and patients with obstructive lung function tests (p = 0.034). The risk factors associated with post-biopsy pulmonary hemorrhage were longer needle path length from the pleura to the lesion (p < 0.001), smaller lesions (p < 0.001), non-pleural contact lesions (p < 0.001), patients without restrictive lung function tests (p = 0.034), and patients in supine positions (p < 0.003). CT-guided biopsies of small nodules equal to or less than 15 mm using 17-gauge guiding needles and 18-gauge biopsy guns were accurate and safe. The biopsy results of small lesions were less accurate than those of large lesions, but the results were a reliable reference for clinical decision-making. Understanding the risk factors associated with the complications of CT-guided biopsies is necessary for pre-procedural planning and communication.

Background: A predictive model can provide physicians, relatives, and patients the accurate information regarding the severity of disease and its predicted outcome. In this study, we used an automated machine-learning-based approach to construct a prognostic model to predict the functional outcome in patients with primary intracerebral hemorrhage (ICH). Methods: We retrospectively collected data on demographic characteristics, laboratory studies and imaging findings of 333 patients with primary ICH. The functional outcomes at the 1st and 6th months after ICH were defined by the modified Rankin scale. All of the attributes were used for preprocessing and for automatic model selection with Automatic Waikato Environment for Knowledge Analysis. Confusion matrix and areas under the receiver operating characteristic curves (AUC) were used to test the predictive performance. Results: Among the models tested, the random forest provided the best predictive performance for functional outcome. The overall accuracy for predicting the 1st month outcome was 83.1%, with 77.4% sensitivity and 86.9% specificity, and the AUC was 0.899. The overall accuracy for predicting the 6th month outcome was 83.9%, with 72.5% sensitivity and 90.6% specificity, and the AUC was 0.917. Conclusions: Using an automatic machine learning technique to predict functional outcome after ICH is feasible, and the random forest model provides the best predictive performance across all tested models. This prediction model may provide information regarding functional outcome for clinicians that will help provide appropriate medical care for patients and information for their caregivers.

The aim of this study was to evaluate whether arterial spin labeling (ASL) perfusion magnetic resonance imaging (MRI) can reliably quantify perfusion deficit as compared to dynamic susceptibility contrast (DSC) perfusion MRI. Thirty-nine patients with acute ischemic stroke in the anterior circulation territory were recruited. All underwent ASL and DSC MRI perfusion scans within 30 hours after stroke onset and 31 patients underwent follow-up MRI scans. ASL cerebral blood flow (CBF) and DSC time to maximum (T_max) maps were used to calculate the perfusion defects. The ASL CBF lesion volume was compared to the DSC T_max lesion volume by Pearson's correlation coefficient and likewise the ASL CBF and DSC T_max lesion volumes were compared to the final infarct sizes respectively. A repeated measures analysis of variance and least significant difference post hoc test was used to compare the mean lesion volumes among ASL CBF, DSC T_max >4–6 s and final infarct. Mean patient age was 72.6 years. The average time from stroke onset to MRI was 13.9 hours. The ASL lesion volume showed significant correlation with the DSC lesion volume for T_max >4, 5 and 6 s (r = 0.81, 0.82 and 0.80; p5 s (29.2 ml, p6 s (21.8 ml, p5 or 6 s were close to mean final infarct size. Quantitative measurement of ASL perfusion is well correlated with DSC perfusion. However, ASL perfusion may overestimate the perfusion defects and therefore further refinement of the true penumbra threshold and improved ASL technique are necessary before applying ASL in therapeutic trials.

It is estimated that at least one million people die by suicide every year, showing the importance of suicide prevention and detection. In this study, an autoencoder and machine learning model was employed to predict people with suicidal ideation based on their structural brain imaging. The subjects in our generalized q-sampling imaging (GQI) dataset consisted of three groups: 41 depressive patients with suicidal ideation (SI), 54 depressive patients without suicidal thoughts (NS), and 58 healthy controls (HC). In the GQI dataset, indices of generalized fractional anisotropy (GFA), isotropic values of the orientation distribution function (ISO), and normalized quantitative anisotropy (NQA) were separately trained in different machine learning models. A convolutional neural network (CNN)-based autoencoder model, the supervised machine learning algorithm extreme gradient boosting (XGB), and logistic regression (LR) were used to discriminate SI subjects from NS and HC subjects. After five-fold cross validation, separate data were tested to obtain the accuracy, sensitivity, specificity, and area under the curve of each result. Our results showed that the best pattern of structure across multiple brain locations can classify suicidal ideates from NS and HC with a prediction accuracy of 85%, a specificity of 100% and a sensitivity of 75%. The algorithms developed here might provide an objective tool to help identify suicidal ideation risk among depressed patients alongside clinical assessment.

Our study aimed to clarify the neuroimaging correlates of suicide attempt by comparing differences in functional magnetic resonance imaging (fMRI) among depressed suicide attempters, depressed patients without suicide attempt history, and healthy controls through comprehensive and novel fMRI analyses and methods in the same study population. The association between depression severity and aspects of the brain imaging was also discussed. Our study recruited 109 participants who were assigned to three groups: 33 depressed patients with suicide attempt (SA), 32 depressed patients without suicide attempt (NS), and 44 healthy controls (HC). All participants were scanned using a 3 T MRI imaging system to obtain resting-state functional images. In seed-based correlation analysis, we found altered functional connectivity in some brain regions of the SA compared with the NS or HC, especially in the hippocampus and thalamus. In the voxel-based analysis, our results showed differential activation and regional homogeneity of the temporal lobe and several brain regions in the SA compared with the NS and HC. We also found that some brain areas correlated with the Hamilton Depression Rating Scale (HAM-D), anxiety, and depression scores, especially in the frontal and temporal lobes. Graph theoretical analysis (GTA) and network-based statistical (NBS) analyses revealed different topological organization as well as slightly better global integration and worse local segregation of the brain network (i.e., more like a random network) in depressed participants compared with healthy participants. We concluded that the brain function of major depressive disorders with and without suicide attempts changed compared with healthy participants.