BACKGROUND: Assessment of regional myocardial function at native pixel-level resolution can play a crucial role in recognizing the early signs of the decline in regional myocardial function. Extensive data processing in existing techniques limits the effective resolution and accuracy of the generated strain maps. The purpose of this study is to compute myocardial principal strain maps ε and ε from tagged MRI (tMRI) at the native image resolution using deep-learning local patch convolutional neural network (CNN) models (DeepStrain).

METHODS: For network training, validation, and testing, realistic tMRI datasets were generated and consisted of 53,606 cine images simulating the heart, the liver, blood pool, and backgrounds, including ranges of shapes, positions, motion patterns, noise, and strain. In addition, 102 in-vivo image datasets from three healthy subjects, and three Pulmonary Arterial Hypertension patients, were acquired and used to assess the network's in-vivo performance. Four convolutional neural networks were trained for mapping input tagging patterns to corresponding ground-truth principal strains using different cost functions. Strain maps using harmonic phase analysis (HARP) were obtained with various spectral filtering settings for comparison. CNN and HARP strain maps were compared at the pixel level versus the ground-truth and versus the least-loss in-vivo maps using Pearson correlation coefficients (R) and the median error and Inter-Quartile Range (IQR) histograms.

RESULTS: CNN-based local patch DeepStrain maps at a phantom resolution of 1.1mm × 1.1 mm and in-vivo resolution of 2.1mm × 1.6 mm were artifact-free with multiple fold improvement with ε ground-truth median error of 0.009(0.007) vs. 0.32(0.385) using HARP and ε ground-truth error of 0.016(0.021) vs. 0.181(0.08) using HARP. CNN-based strain maps showed substantially higher agreement with the ground-truth maps with correlation coefficients R > 0.91 for ε and ε compared to R < 0.21 and R < 0.82 for HARP-generated maps, respectively.

CONCLUSION: CNN-generated Eulerian strain mapping permits artifact-free visualization of myocardial function at the native image resolution.

Background and aim: Early detection of hepatocellular carcinoma (HCC) enhances eective and curative
management. New genetic markers with distinct diagnostic ability are required. Aim: determine the expression
of GPC3, PEG10, SERPINI1, MK and QP-C in the peripheral blood of HCC patients.
Methods: 74 HCV patients were recruited and divided into three groups; chronic hepatitis (I), liver cirrhosis
(II) and HCC (III). Demographics, laboratory and imaging data were collected. Child score and metastatic
work up were completed. e expression of the ve candidate genes in the peripheral blood was performed
by qRT-PCR assay.
Results: Groups were gender matched, age in group I was signicantly lower than in groups II and III (37.7
vs 50.4 and 55.6, p value <0.005). CHILD score; group II and III A/B/C = (7/5/6) and (20/6/3). AFP was
signicantly higher in group III than I and II (204 vs 3.9 and 6.9, p < 0.01). In HCC group 69% of the lesions
were < 5 cm, and had 1-2 nodules; 14% had metastases. GPC3, PEG10, SERPINI1 and MK mRNA were
signicantly higher in the HCC group compared to the other groups while QP-C mRNA was higher in chronic
hepatitis C group compared to other groups. e gene expression values in HCC patients were independent
of the tumor size, AFP levels or extrahepatic metastasis. Combined measurement of the ve gene markers
showed 100% sensitivity and 33% specicity, 48% PPV and 100% NPV.
Conclusion: GPC3, PEG10, SERPINI1 and MK are genetic markers that can represent a useful tool for
detection of HCC.

For the solar photovoltaic (PV) system to operate efficiently, it is necessary to effectively establish an equivalent model of PV cell and extract the relevant unknown model parameters accurately. This paper introduces a new metaheuristic algorithm, i.e., gaining-sharing knowledge based algorithm (GSK) to solve the solar PV model parameter extraction problem. This algorithm simulates the process of knowledge acquisition and sharing in the human life cycle and is with strong competitiveness in solving optimization problems. It includes two significant phases. The first phase is the beginner–intermediate or junior acquisition and sharing stage, and the second phase is the intermediate–expert or senior acquisition and sharing stage. In order to verify the effectiveness of GSK, it is applied to five PV models including the single diode model, double diode model, and three PV modules. The influence of population size on the algorithm performance is empirically investigated. Besides, it is further compared with some other excellent metaheuristic algorithms including basic algorithms and advanced algorithms. Among the five PV models, the root mean square error values between the measured data and the calculated data of GSK are 9.8602E−04 ± 2.18E−17, 9.8280E−04 ± 8.72E−07, 2.4251E−03 ± 1.04E−09, 1.7298E−03 ± 6.25E−18, and 1.6601E−02 ± 1.44E−16, respectively. The results show that GSK has overall better robustness, convergence, and accuracy.