Representative Data Selection for Efficient Medical Incremental Learning.

To train a deep neural network relies on a large amount of annotated data. In special scenarios like industry defect detection and medical imaging, it is hard to collect sufficient labeled data all at once. Newly annotated data may arrive incrementally. In practice, we also prefer our target model to improve its capability gradually as new data comes in by quick re-training. This work tackles this problem from a data selection prospective by constraining ourselves to always retrain the target model with a fix amount of data after new data comes in. A variational autoencoder (VAE) and an adversarial network are combined for data selection, achieving fast model retraining. This enables the target model to continually learn from a small training set while not losing the information learned from previous iterations, thus incrementally adapting itself to new-coming data. We validate our framework on the LGG Segmentation dataset for the semantic segmentation task.Clinical relevance- The proposed VAE-based data selection model combined with adversarial training can choose a representative and reliable subset of data for time-efficient medical incremental learning. Users can immediately see the improvement of the medical segmentation model whenever new annotated images are contributed (after a few minutes of model retraining).

Automated Stanford classification of aortic dissection using a 2-step hierarchical neural network at computed tomography angiography.

OBJECTIVES: This study aimed to evaluate the feasibility of automatic Stanford classification of classic aortic dissection (AD) using a 2-step hierarchical neural network. METHODS: Between 2015 and 2019, 130 arterial phase series (57 type A, 43 type B, and 30 negative cases) in aortic CTA were collected for the training and validation. A 2-step hierarchical model was built including the first step detecting AD and the second step predicting the probability (0-1) of Stanford types. The model's performance was evaluated with an off-line prospective test in 2020. The sensitivity and specificity for Stanford type A, type B, and no AD (Sens A, B, N and Spec A, B, N, respectively) and Cohen's kappa were reported. RESULTS: Of 298 cases (22 with type A, 29 with type B, and 247 without AD) in the off-line prospective test, the Sens A, Sens B, and Sens N were 95.45% (95% confidence interval [CI], 77.16-99.88%), 79.31% (95% CI, 60.28-92.01%), and 93.52% (95% CI, 89.69-96.25%), respectively. The Spec A, Spec B, and Spec N were 98.55% (95% CI, 96.33-99.60%), 94.05% (95% CI, 90.52-96.56%), and 94.12% (95% CI, 83.76-98.77%), respectively. The classification rate achieved 92.28% (95% CI, 88.64-95.04%). The Cohen's kappa was 0.766 (95% CI, 0.68-0.85; p < 0.001). CONCLUSIONS: Stanford classification of classic AD can be determined by a 2-step hierarchical neural network with high sensitivity and specificity of type A and high specificity in type B and no AD. KEY POINTS: • The Stanford classification for aortic dissection is widely adopted and divides it into Stanford type A and type B based on the ascending thoracic aorta dissected or not. • The 2-step hierarchical neural network for Stanford classification of classic aortic dissection achieved high sensitivity (95.45%) and specificity (98.55%) of type A and high specificity in type B and no aortic dissection (94.05% and 94.12%, respectively) in 298 test cases. • The 2-step hierarchical neural network demonstrated moderate agreement (Cohen's kappa: 0.766, p < 0.001) with cardiovascular radiologists in detection and Stanford classification of classic aortic dissection in 298 test cases.

Sandwich Evaporation-Solvent Annealing Fabrication of Highly Crystalline MAPbIxCl3-x Perovskite Solar Cells.

Perovskites doped with chlorine (Cl-), which are usually fabricated using the solution process, can effectively improve the stability and carrier mobility. Compared with the low tolerance of the solution process that relies mostly on personal skill, thermal evaporation is an important method for large-scale production of perovskite solar cells but the production cost is high. In this study, the sandwich evaporation-solvent annealing (SE-SA) method is proposed. Using sandwich evaporation with a low-cost chamber of the sandwich evaporation technique (SET) made in the laboratory and with the help of DMSO steam-assisted crystallization, we have successfully produced chlorine-containing perovskite solar cells with a high crystallinity and a high efficiency of 15.1% with Voc = 0.98 V, Jsc = 21.94 mA/cm2, FF = 74.29%, and Rs = 3.66 Ω·cm2, which can greatly reduce the production cost. It is worth mentioning that all the processes are carried out outside a glove box, which makes it possible for large-scale production of chlorine-containing perovskite solar cells by evaporation.

Growth process control produces high-crystallinity and complete-reaction perovskite solar cells.

The growth process control (GPC) method, a new method which is better than thermal evaporation, for producing high-crystallinity perovskites by controlling the growth time in a low vacuum, is explored in this work. Inspired by evaporation technology, GPC is an effective method for modifying traditional thermal evaporation and for controlling the crystal growth of perovskite CH3NH3I3. Compared to fabrication with the process of co-evaporation, the MAPbI3 perovskite solar cell fabricated by GPC has high uniformity and film coverage. All of the manufacturing is carried out outside of the glove box. It provides an easy and effective way for perovskite fabrication for industrialization. Here, after using GPC to form perovskite solar cells, the residual methylammonium iodide (MAI) and PbI2 which is produced by the evaporation process can react completely, observed by time of flight secondary ion mass spectrometry (TOF-SIMS). Finally, formed by GPC, perovskite solar cells exhibit high performance and fewer crystal defects. The electron and hole recombination is greatly reduced. Through the GPC method, the J sc and the filling factor are improved with the increase of time after the fabrication. The power conversion efficiency was increased from 11.12% to 16.4%.

Penta-O-galloyl-ß-D-glucose suppresses EGF-induced eIF3i expression through inhibition of the PI3K/AKT/mTOR pathway in prostate cancer cells.

Approximately 70% of prostate cancer patients will develop bone metastasis in axial and other regions of the skeleton. Epidermal growth factor (EGF) generated from bone tissue contributes to prostate cancer metastasis. In a previous study, penta-O-galloyl-ß-D-glucose (PGG) suppressed androgen-independent prostate cancer bone metastasis by transcriptionally repressing EGF-induced MMP-9 expression. This study utilized proteomics to analyze the effects of PGG in EGF-induced prostate cancer bone metastasis. This study showed that PGG suppressed EGF-induced eIF3i expression in PC-3 cells. By transfection of eIF3i shRNA, it was observed that reduced eIF3i expression suppressed the invasion of PC-3 cells in vitro. PGG reduced EGF-induced eIF3i expression through inhibition of the PI3K/AKT/mTOR pathway. Therefore, PGG may be able to be used as a potential new therapeutic drug for prostate cancer bone metastasis.

Study of the Anti-Proliferative Activity of 5-Substituted 4,7-Dimethoxy-1,3-Benzodioxole Derivatives of SY-1 from Antrodia camphorata on Human COLO 205 Colon Cancer Cells.

A set of 10 4,7-dimethoxy-1,3-benzodioxole derivatives based on a lead compound previously discovered by our group, SY-1, which was isolated from Antrodia camphorata, were evaluated for their in vitro inhibitory activity on human colorectal carcinoma cells (COLO 205). Structure-activity relationship studies of the 10 compounds indicated the importance of the chain length of the alkyl group at the 5-position, and the 2-propenyl substituent named "apiole" exhibited the most potent inhibitory activity. In the present study, we demonstrate that the SY-1 analogue "apiole" decreased the proliferation of COLO 205 cells, but not that of normal human colonic epithelial cells (FHC). The G0/G1 cell cycle arrest induced by apiole (75-225 µM) was associated with significantly increased levels of p53, p21 and p27 and decreased levels of cyclin D1. Concerning COLO 205 cell apoptosis, apiole (>150 µM) treatment significantly increased the levels of cleaved caspases 3, 8, 9 and bax/bcl-2 ratio and induced ladder formation in DNA fragmentation assay and sub-G1 peak in flow cytometry analysis. These findings suggest that apiole can suppress COLO 205 cell growth; however, the detailed mechanisms of these processes require further investigation.

Inhibition of epidermal growth factor receptor signaling by Saussurea involucrata, a rare traditional Chinese medicinal herb, in human hormone-resistant prostate cancer PC-3 cells.

Prostate carcinoma is the most frequently diagnosed malignancy and the second leading cause of death of men in the United States. To date, no effective therapeutic treatment allows abrogation of the progression of prostate cancer to more invasive forms. In this study, we identified Saussurea involucrata Kar. et Kir., a rare traditional Chinese medicinal herb, as a potential agent for androgen-independent prostate cancer patients and investigated its biological mechanism as an antineoplastic agent. S. involucrata caused a concentration- and time-dependent inhibition of cell proliferation in human hormone-resistant prostate cancer PC-3 cells. Moreover, in vitro studies in a panel of several types of human cancer cell lines revealed that S. involucrata inhibited cell proliferation with high potency. To evaluate the bioactive compounds, we successively extracted the S. involucrata with fractions of methanol (SI-1), ethyl acetate (SI-2), n-butanol (SI-3), and water (SI-4). Among these extracts, SI-2 contains the most effective bioactivity. SI-2 treatment resulted in significant time-dependent growth inhibition together with G1 phase cell cycle arrest and apoptosis in PC3 cells. In addition, SI-2 treatment strongly induced p21WAF1/CIP and p27KIP1 expression, independent of the p53 pathway, and downregulated expression of cyclin D1 and cyclin-dependent kinase 4 (CDK4). SI-2 treatment increased levels of Bax, cytochrome c, activated caspase-3, and active caspase-9 and decreased Bcl-2 expression level. One of the major targets for the therapy in prostate cancer can be epidermal growth factor receptor (EGFR). SI-2 markedly reduced phosphorylation of EGFR and inhibited activation of AKT and STAT3. Moreover, p.o. administration of SI-2 induced a dose-dependent inhibition of PC-3 tumor growth in vivo. In summary, our study identifies S. involucrata as an effective inhibitor of EGFR signaling in human hormone-resistant prostate cancer PC-3 cells. We suggest that S. involucrata could be developed as an agent for the management of EGFR-positive human cancers.

Penta-O-galloyl-beta-D-glucose suppresses prostate cancer bone metastasis by transcriptionally repressing EGF-induced MMP-9 expression.

Prostate carcinoma is the most frequently diagnosed malignancy and the second leading cause of cancer-related death of men in the United States. Epidermal growth factor (EGF) generated from bone tissue contributes to prostate cancer metastasis through stimulating matrix metalloproteinase (MMP) secretions from prostate cancer cells. In this study, in vitro invasion assay was performed by incubating penta-O-galloyl-beta-D-glucose (5GG) at various concentrations with 2 x 10(4) PC-3 cells for 48 h. The anti-invasive and cytotoxic effects of 5GG were found and evaluated on the human androgen-independent prostate cancer PC-3 cell line by MTT assays and Western blot analyses. 5GG inhibited the EGF-induced cell invasiveness and MMP-9 expression in a dose- and time-dependent manner by reducing the MMP-9 transcriptional activity. To explore the mechanisms for the 5GG-mediated regulation of MMP-9, we further examined the effects of 5GG on transcription factors, including NF-kappaB, AP-1, and mitogen-activated protein kinase (MAPK) activities. The results showed that 5GG suppressed the EGF-induced NF-kappaB nuclear translocation and also abrogated the EGF-induced activation of c-jun N-terminal kinase (JNK), an upstream modulator of NF-kappaB. Moreover, we showed that 5GG reduced EGFR expression through the proteasome pathway. These results suggest that 5GG may exert at least part of its anti-invasive effect in androgen-independent prostate cancer by controlling MMP-9 expression through the suppression of the EGFR/JNK pathway. Finally, 5GG suppresses invasion and tumorigenesis in nude mice treatment with intratibia injection of PC-3 cells. These in vitro and in vivo results suggest that 5GG may be a therapeutic candidate for the treatment of advanced prostate cancer.