Combining multimodal magnetic resonance brain imaging and machine learning to unravel neurocognitive function in non-neuropsychiatric systemic lupus erythematosus.

OBJECTIVE: To study whether multimodal brain MRI comprising permeability and perfusion measures coupled with machine learning can predict neurocognitive function in young patients with SLE without neuropsychiatric manifestations. METHODS: SLE patients and healthy controls (HCs) (≤40 years of age) underwent multimodal structural brain MRI that comprised voxel-based morphometry (VBM), magnetization transfer ratio (MTR) and dynamic contrast-enhanced (DCE) MRI in this cross-sectional study. Neurocognitive function assessed by Automated Neuropsychological Assessment Metrics was reported as the total throughput score (TTS). Olfactory function was assessed. A machine learning-based model (i.e. glmnet) was constructed to predict TTS. RESULTS: Thirty SLE patients and 10 HCs were studied. Both groups had comparable VBM, MTR, olfactory bulb volume (OBV), olfactory function and TTS. While after correction for multiple comparisons the uncorrected increase in the blood-brain barrier (BBB) permeability parameters compared with HCs did not remain evident in SLE patients, DCE-MRI perfusion parameters, notably an increase in right amygdala perfusion, was positively correlated with TTS in SLE patients (r = 0.636, false discovery rate P < 0.05). A machine learning-trained multimodal MRI model comprising alterations of VBM, MTR, OBV and DCE-MRI parameters mainly in the limbic system regions predicted TTS in SLE patients (r = 0.644, P < 0.0005). CONCLUSION: Multimodal brain MRI demonstrated increased right amygdala perfusion that was associated with better neurocognitive performance in young SLE patients without statistically significant BBB leakage and microstructural abnormalities. A machine learning-constructed multimodal model comprising microstructural, perfusion and permeability parameters accurately predicted neurocognitive performance in SLE patients.

Comparison of quantitative parameters and radiomic features as inputs into machine learning models to predict the Gleason score of prostate cancer lesions.

INTRODUCTION: The classification of prostate cancer (PCa) lesions using Prostate Imaging Reporting and Data System (PI-RADS) suffers from poor inter-reader agreement. This study compared quantitative parameters or radiomic features from multiparametric magnetic resonance imaging (mpMRI) or positron emission tomography (PET), as inputs into machine learning (ML) to predict the Gleason scores (GS) of detected lesions for improved PCa lesion classification. METHODS: 20 biopsy-confirmed PCa subjects underwent imaging before radical prostatectomy. A pathologist assigned GS from tumour tissue. Two radiologists and one nuclear medicine physician delineated the lesions on the mpMR and PET images, yielding 45 lesion inputs. Seven quantitative parameters were extracted from the lesions, namely T2-weighted (T2w) image intensity, apparent diffusion coefficient (ADC), transfer constant (KTRANS), efflux rate constant (Kep), and extracellular volume ratio (Ve) from mpMR images, and SUVmean and SUVmax from PET images. Eight radiomic features were selected out of 109 radiomic features from T2w, ADC and PET images. Quantitative parameters or radiomic features, with risk factors of age, prostate-specific antigen (PSA), PSA density and volume, of 45 different lesion inputs were input in different combinations into four ML models - Decision Tree (DT), Support Vector Machine (SVM), k-Nearest-Neighbour (kNN), Ensembles model (EM). RESULTS: SUVmax yielded the highest accuracy in discriminating detected lesions. Among the 4 ML models, kNN yielded the highest accuracies of 0.929 using either quantitative parameters or radiomic features with risk factors as input. CONCLUSIONS: ML models' performance is dependent on the input combinations and risk factors further improve ML classification accuracy.

Quantifying the changes in the tumour vascular micro-environment in spinal metastases treated with stereotactic body radiotherapy - a single arm prospective study.

BACKGROUND: The primary objective was to quantify changes in vascular micro-environment in spinal metastases (SM) patients treated with stereotactic body radiotherapy (SBRT) with multi-parametric dynamic contrast enhanced (DCE) magnetic resonance imaging (MRI). The secondary objective was to study plasma biomarkers related to endothelial apoptosis. PATIENTS AND METHODS: Patients were imaged with DCE-MRI at baseline/1-week/12-weeks post-SBRT. Metrics including normalised time-dependent leakage (Ktrans), permeability surface product (PS), fractional plasma volume (Vp), extracellular volume (Ve) and perfusion (F) were estimated using distributed parameter model. Serum acid sphingomyelinase (ASM) and sphingosine-1-phosphate (S1P) were quantified using ELISA. Clinical outcomes including physician-scored and patient-reported toxicity were collected. RESULTS: Twelve patients (with varying primary histology) were recruited, of whom 10 underwent SBRT. Nine patients (with 10 lesions) completed all 3 imaging assessment timepoints. One patient died due to pneumonia (unrelated) before follow-up scans were performed. Median SBRT dose was 27 Gy (range: 24-27) over 3 fractions (range: 2-3). Median follow-up for alive patients was 42-months (range: 22.3-54.3), with local control rate of 90% and one grade 2 or higher toxicity (vertebral compression fracture). In general, we found an overall trend of reduction at 12-weeks in all parameters (Ktrans/PS/Vp/Ve/F). Ktrans and PS showed a reduction as early as 1-week. Ve/Vp/F exhibited a slight rise 1-week post-SBRT before reducing below the baseline value. There were no significant changes, post-SBRT, in plasma biomarkers (ASM/S1P). CONCLUSIONS: Tumour vascular micro-environment (measured by various metrics) showed a general trend towards downregulation post-SBRT. It is likely that vascular-mediated cell killing contributes to excellent local control rates seen with SBRT. Future studies should evaluate the effect of SBRT on primary-specific spinal metastases (e.g., renal cell carcinoma).

Synchronous airway lesions in children: an analysis of characteristics and comorbidities.

OBJECTIVES: To analyze the characteristics and the associated medical co-morbidities in children with synchronous airway lesions (SALs) found during rigid bronchoscopy. METHODS: Retrospective case series and chart review of patients who were found to have more than one airway lesion after undergoing airway evaluation via rigid endoscopy at a tertiary care pediatric hospital between 2001 and 2011. Patient demographics, presence of associated non-airway pathologies, and the number and types of airway lesions were collected. For analysis, airway lesions were classified based on the anatomical subsites involved (supraglottic, glottic, subglottic, tracheal and bronchial). RESULTS: Out of 592 rigid bronchoscopies performed, there were 73 cases with SALs (12.3%). Of these, only 20% of patients were term infants without associated congenital anomalies. Over 70% of patients with SALs have combinations of lesions involving the trachea, subglottis and supraglottis. Neurological anomalies and GERD were both independently associated with a three-time increase in the odds of having synchronous involvement of these three anatomical subsites (OR 3.15, 95% CI 1.06-9.41; OR 3.0, 95% CI 1.05-8.50, respectively). Glottic lesions were present in 28.7% of patients. Prematurity and cardiac anomalies were both associated with tendency of doubling the odds of glottic lesions (OR 2.34, 95% CI 0.84-6.52; OR 2.0, 95% CI 0.76-5.60, respectively). Overall, almost 10% of newly diagnosed lesions in context of SALs required an additional intervention. CONCLUSIONS: The majority of patients with SALs are either born prematurely or have associated congenital anomalies. In SAL patients with associated neurological anomalies or GERD, the lesions are more likely to be localized to the supraglottis, subglottis and trachea whereas prematurity and cardiac anomalies could both be increasing the odds of a glottic lesion. High suspicious index should be kept in mind when rigid bronchoscopy is performed to not miss an associated lesion.

Pharmacological AMP-kinase activators have compartment-specific effects on cell physiology.

5'-AMP-activated kinase (AMPK) regulates numerous biological events and is an essential target for the treatment of type 2 diabetes. The objectives of the present study were first to determine the compartment-specific effects of three established AMPK activators on Thr172 phosphorylation of the α-subunit, an indicator of AMPK activation. Second, we examined how cytoplasmic and nuclear processes are modulated by pharmacological AMPK activators. Specifically, the impact of phenformin, resveratrol, and 5-aminoimidazole-4-carboxamide riboside (AICAR) on Thr172 phosphorylation in the cytoplasm and nucleus was quantified by different methods. To analyze how these activators change cell physiology, we measured the inactivation of acetyl-CoA-carboxylase 1, a predominantly cytoplasmic enzyme that is crucial for lipid metabolism. As a criterion for activities associated with the nucleus, de novo RNA synthesis in nucleoli was quantified. Our studies demonstrate that pharmacological activators of AMPK can alter the balance between nuclear and cytoplasmic AMPK pools. Thus, phenformin and resveratrol caused a strong activation of AMPK in the cytoplasm, whereas the effect was less pronounced in nuclei. By contrast, AICAR elicited a comparable rise in Thr172 phosphorylation in both compartments. Notably, these activators differed drastically in their effects on physiological processes that are located in distinct subcellular compartments. All compounds led to a substantial inactivation of acetyl-CoA-carboxylase 1 in the cytoplasm, with only minor changes to the nuclear enzyme. In the nucleolus, transcription was strongly inhibited by resveratrol, while a moderate inhibition was observed with phenformin and AICAR. Taken together, the compartment-specific phosphorylation of AMPK and downstream events are determined by the activator.

Diffusion tensor tractography: corticospinal tract fiber reduction is associated with temporary hemiparesis in benign extracerebral lesions.

OBJECTIVE: Benign extracerebral lesions such as meningiomas may cause hemiparesis by compression and deviation without infiltrating the white matter. We used magnetic resonance diffusion tensor imaging and diffusion tensor tractography to investigate the effects of benign extracerebral lesions on the corticospinal tract (CST). METHODS: Thirteen patients with extracerebral lesions (11 benign meningiomas and 2 benign cysts) underwent magnetic resonance diffusion tensor imaging and diffusion tensor tractography of the CST using fiber assignment by continuous tractography. The CST was reconstructed and assessed by comparing the ipsilateral and unaffected contralateral fibers. The tumor volume, relative fractional anisotropy, fiber deviation, relative fiber number, and relative fiber per voxel were compared between patients without and with temporary presurgical hemiparesis. RESULTS: Seven patients without hemiparesis and five patients with temporary hemiparesis were analyzed; one patient had permanent weakness and was excluded from analysis. There was no significant difference in the tumor volume, relative fractional anisotropy, presence of cerebral edema, or CST deviation between groups. In patients with temporary hemiparesis, the median relative fiber number (mean, 0.35 +/- 0.32) and relative fiber per voxel (mean, 0.49 +/- 0.14) were significantly reduced compared with patients without hemiparesis (0.92 +/- 0.55, P = 0.04; and 0.96 +/- 0.28, P < 0.01, respectively). CONCLUSION: In patients with benign extracerebral lesions, reduction in fiber number and fiber per voxel, but not fiber deviation, correlated with temporary hemiparesis. Clinical recovery was possible even if the CST fibers detected by diffusion tensor tractography were reduced by benign extracerebral lesions.