Search

Abstract Background Evidence-based treatment decisions in medicine are made founded on population-level evidence obtained during randomized clinical trials. In an era of personalized medicine, these decisions should be based on the predicted benefit of a treatment on a patient-level. Survival prediction models play a central role as they incorporate the time-to-event and censoring. In medical applications uncertainty is critical especially when treatments differ in their side effect profiles or costs. Additionally, models must be adapted to local populations without diminishing performance and often without the original training data available due to privacy concern. Both points are supported by Bayesian models—yet they are rarely used. The aim of this work is to evaluate Bayesian parametric survival models on public datasets including cardiology, infectious diseases, and oncology. Materials and methods Bayesian parametric survival models based on the Exponential and Weibull distribution were implemented as a Python package. A linear combination and a neural network were used for predicting the parameters of the distributions. A superiority design was used to assess whether Bayesian models are better than commonly used models such as Cox Proportional Hazards, Random Survival Forest, and Neural Network-based Cox Proportional Hazards. In a secondary analysis, overfitting was compared between these models. An equivalence design was used to assess whether the prediction performance of Bayesian models after model updating using Bayes rule is equivalent to retraining on the full dataset. Results In this study, we found that Bayesian parametric survival models perform as good as state-of-the art models while requiring less hyperparameters to be tuned and providing a measure of the uncertainty of the predictions. In addition, these models were less prone to overfitting. Furthermore, we show that updating these models using Bayes rule yields equivalent performance compared to models trained on combined original and new datasets. Conclusions Bayesian parametric survival models are non-inferior to conventional survival models while requiring less hyperparameter tuning, being less prone to overfitting, and allowing model updating using Bayes rule. Further, the Bayesian models provide a measure of the uncertainty on the statistical inference, and, in particular, on the prediction.

Abstract Background To evaluate the feasibility of a novel approach for predicting hepatocellular carcinoma (HCC) response to drug-eluting beads transarterial chemoembolization (DEB-TACE) using computed tomography hepatic arteriography enhancement mapping (CTHA-EM) method. Methods This three-institution retrospective study included 29 patients with 46 HCCs treated with DEB-TACE between 2017 and 2020. Pre- and posttreatment CTHA-EM images were generated using a prototype deformable registration and subtraction software. Relative tumor enhancement (T Post/pre-RE) defined as the ratio of tumor enhancement to normal liver tissue was calculated to categorize tumor response as residual (T Post-RE > 1) versus non-residual (T Post-RE ≤ 1) enhancement, which was blinded compared to the response assessment on first follow-up imaging using modified RECIST criteria. Additionally, for tumors with residual enhancement, CTHA-EM was evaluated to identify its potential feeding arteries. Results CTHA-EM showed residual enhancement in 18/46 (39.1%) and non-residual enhancement in 28/46 (60.9%) HCCs, with significant differences on T Post-RE (3.05 ± 2.4 versus 0.48 ± 0.23, respectively; p < 0.001). The first follow-up imaging showed non-complete response (partial response or stable disease) in 19/46 (41.3%) and complete response in 27/46 (58.7%) HCCs. CTHA-EM had a response prediction sensitivity of 94.7% (95% CI, 74.0–99.9) and specificity of 100% (95% CI, 87.2–100). Feeding arteries to the residual enhancement areas were demonstrated in all 18 HCCs (20 arteries where DEB-TACE was delivered, 2 newly developed collaterals following DEB-TACE). Conclusion CTHA-EM method was highly accurate in predicting initial HCC response to DEB-TACE and identifying feeding arteries to the areas of residual arterial enhancement.

Image-guided percutaneous ablation techniques represent an attractive local therapy for the treatment of colorectal liver metastases (CLM) given its low risk of severe complications, which allows for early initiation of adjuvant therapies and spare functional liver parenchyma, allowing repeated treatments at the time of recurrence. However, ablation does not consistently achieve similar oncological outcomes to surgery, with the latter being currently considered the first-line local treatment modality in international guidelines. Recent application of computer-assisted ablation planning, guidance, and intra-procedural response assessment has improved percutaneous ablation outcomes. In addition, the evolving understanding of tumor molecular profiling has brought to light several biological factors associated with oncological outcomes following local therapies. The standardization of ablation procedures, the understanding of previously unknown biological factors affecting ablation outcomes, and the evidence by ongoing prospective clinical trials are poised to change the current perspective and indications on the use of ablation for CLM.

ObjectivesColorectal cancer (CRC), the third most common cancer in the USA, is a leading cause of cancer-related death worldwide. Up to 60% of patients develop liver metastasis (CRLM). Treatments like radiation and ablation therapies require disease segmentation for planning and therapy delivery. For ablation, ablation-zone segmentation is required to evaluate disease coverage. We hypothesize that fully convolutional (FC) neural networks, trained using novel methods, will provide rapid and accurate identification and segmentation of CRLM and ablation zones.MethodsFour FC model styles were investigated: Standard 3D-UNet, Residual 3D-UNet, Dense 3D-UNet, and Hybrid-WNet. Models were trained on 92 patients from the liver tumor segmentation (LiTS) challenge. For the evaluation, we acquired 15 patients from the 3D-IRCADb database, 18 patients from our institution (CRLM = 24, ablation-zone = 19), and those submitted to the LiTS challenge (n = 70). Qualitative evaluations of our institutional data were performed by two board-certified radiologists (interventional and diagnostic) and a radiology-trained physician fellow, using a Likert scale of 1–5.ResultsThe most accurate model was the Hybrid-WNet. On a patient-by-patient basis in the 3D-IRCADb dataset, the median (min–max) Dice similarity coefficient (DSC) was 0.73 (0.41–0.88), the median surface distance was 1.75 mm (0.57–7.63 mm), and the number of false positives was 1 (0–4). In the LiTS challenge (n = 70), the global DSC was 0.810. The model sensitivity was 98% (47/48) for sites ≥15 mm in diameter. Qualitatively, 100% (24/24; minority vote) of the CRLM and 84% (16/19; majority vote) of the ablation zones had Likert scores ≥4.ConclusionThe Hybrid-WNet model provided fast (

This retrospective study was aimed to evaluate the reduced iodine load on image quality and diagnostic performance in multiphasic hepatic CT using a novel monoenergetic reconstruction algorithm (nMERA) in assessment of local tumor progression after radiofrequency ablation (RFA) of hepatocellular carcinoma (HCC). Ninety patients who underwent CT 1 month after RFA of HCC. Forty-five patients had multiphasic hepatic dual-energy CT with a half-reduced contrast medium (HRCM) of 277.5 mg I/kg. The nMERA (40-70-keV) images were reconstructed in each phase. Another 45 patients received a standard contrast medium (SCM) of 555 mg I/kg, and the images were reconstructed as a simulated 120-kVp images. Primary outcome was accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) in assessment of local tumor progression. Additional advanced assessments included the image noise, attenuation value, contrast-to-noise ratio (CNR), and subjective image quality between the groups. The accuracy, sensitivity and specificity of nMERA HRCM images were 95.7%, 100% and 93.9% for 40 keV, 95.7%, 85.7% and 100% for 50 keV, 83.0%, 42.8% and 100% for 60 keV, and 83.0%, 42.9% and 100% for 70 keV. The AUROC was 0.99, 0.99, 0.94, and 0.93 for 40-70 keV nMERA HRCM images, respectively. Compared with simulated 120-kVp SCM images, nMERA HRCM images demonstrated comparable noise at 70-keV (P < 0.05), and comparable CNR at 40- and 50-keV (P < 0.05). nMERA DECT enables the contrast medium to be reduced to up to 50% in multiphasic hepatic CT while preserving diagnostic accuracy.

Objectives: Colorectal cancer (CRC), the third most common cancer in the USA, is a leading cause of cancer-related death worldwide. Up to 60% of patients develop liver metastasis (CRLM). Treatments like radiation and ablation therapies require disease segmentation for planning and therapy delivery. For ablation, ablation-zone segmentation is required to evaluate disease coverage. We hypothesize that fully convolutional (FC) neural networks, trained using novel methods, will provide rapid and accurate identification and segmentation of CRLM and ablation zones. Methods: Four FC model styles were investigated: Standard 3D-UNet, Residual 3D-UNet, Dense 3D-UNet, and Hybrid-WNet. Models were trained on 92 patients from the liver tumor segmentation (LiTS) challenge. For the evaluation, we acquired 15 patients from the 3D-IRCADb database, 18 patients from our institution (CRLM = 24, ablation-zone = 19), and those submitted to the LiTS challenge (n = 70). Qualitative evaluations of our institutional data were performed by two board-certified radiologists (interventional and diagnostic) and a radiology-trained physician fellow, using a Likert scale of 1-5. Results: The most accurate model was the Hybrid-WNet. On a patient-bypatient basis in the 3D-IRCADb dataset, the median (min-max) Dice similarity coefficient (DSC) was 0.73 (0.41-0.88), the median surface distance was 1.75 mm (0.57-7.63 mm), and the number of false positives was 1 (0-4). In the LiTS challenge (n = 70), the global DSC was 0.810. The model sensitivity was 98% (47/48) for sites =15 mm in diameter. Qualitatively, 100% (24/24; minority vote) of the CRLM and 84% (16/19; majority vote) of the ablation zones had Likert scores =4. Conclusion: The Hybrid-WNet model provided fast (<30 s) and accurate segmentations of CRLM and ablation zones on contrast-enhanced CT scans, with positive physician reviews. [ABSTRACT FROM AUTHOR]

Different kinds of mono- and bidentate phosphite ligands were used in Rh-catalyzed of styrene to illustrate the influence of steric and electronic properties of ligands on catalytic performance. High activity (99.9%) and good regioselectivity (85.4%) to the linear aldehyde were achieved under optimum conditions in the presence of Rh/bisphosphite complex (bisphosphite: 2,2′-bis(dipyrrolylphosphinooxy)-1,1′-(±)-binaphthyl). This system makes it possible to prepare functionalized terminal aldehydes from readily available styrene or its derivatives through with high linear selectivity. Copyright © 2013 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

A simple and highly efficient method for the cross-coupling of aromatics by C-H activation is described. Acetamide was found to be an effective cocatalyst for this generally applicable ruthenium(II)-catalyzed direct arylation reaction. A wide range of deactivated aryl chlorides could be efficiently coupled under the mild temperature of 101 °C. The new process is highly atom-economical and sustainable. [ABSTRACT FROM AUTHOR]

Rhodium-catalyzed hydroformylation of dicyclopentadiene using different kinds of mono- and bidentate phosphorus ligands was studied. The results demonstrated that the electronic and steric properties of ligand could influence the catalytic activity and the selectivity towards dialdehyde products. High conversion (99.9%) and good selectivity (95.0%) to dialdehydes were achieved under optimized reaction conditions in the presence of Rh/mixed mono- and bidentate phosphorus ligand complex (bidentate ligand: 2,2′-bis(dipyrrolylphosphinooxy)-1,1′-(±)-biphenyl (L1) and monodentate ligand: tri-(4-methoxylphenyl)phosphine (L8) (L1/L8 molar ratio 1/2)), which offers an easy approach to prepare dialdehydes from hydroformylation of cyclopentadiene under mild conditions (5.0MPa, 100°C). [ABSTRACT FROM AUTHOR]

Abstract: Six multidentate nitrogen ligands, 2,6-bis(1,5-diphenyl-1H-pyrazol-3-yl)pyridine (L1), 2,6-bis(1-methyl-5-phenyl-1H-pyrazol-3-yl)pyridine (L2), 2,6-bis(1,5-dimethyl-1H-pyrazol-3-yl)pyridine (L3), 2,6-bis(5-phenyl-1H-pyrazol-3-yl)pyridine (L4), 2,6-bis(1-p-triflurophenyl-5-phenyl-1H-pyrazol-3-yl)pyridine (L5), and 2,6-bis(1-m-methylphenyl-5-phenyl-1H-pyrazol-3-yl)pyridine (L6) were synthesized and characterized. Their palladium complexes catalyzed Suzuki–Miyaura cross-coupling revealed that the catalytic activities of these complexes were in the order of PdCl2-L5>PdCl2-L1>PdCl2-L6>PdCl2-L2>PdCl2-L3>PdCl2-L4. With a low Pd loading of 10−5 mol%, PdCl2-L1 showed a high activity. The turnover numbers (TONs) were up to 58,000,000 for 4-bromobenzotrifluoride. [Copyright &y& Elsevier]

Abstract: Ir/ZrO2·xH2O was prepared by co-precipitation and characterized by TEM, XRD, and XPS. It showed an excellent catalytic performance for the hydrogenation of haloaromatic nitro compounds to the corresponding amines in the mixture solvent of ethanol and water. The catalyst gave both conversion and selectivity to be over 99.9% for the hydrogenation of p-chloronitrobenzene to p-chloroaniline. Besides the intrinsic characterization of iridium in the hydrogenation of haloaromatic nitro compounds, the high activity and selectivity of Ir/ZrO2·xH2O is probably attributed to the formation of hydrogen bond between substrate and the hydroxyl groups on the surface of the hydrous catalyst or solvent water to activate the N nitro group. The activated N easy to be attacked by the activated hydrogen, so the hydrogenation is promoted. [Copyright &y& Elsevier]