Automated Diagnosis of Bone Metastasis by Classifying Bone Scintigrams Using a Self-defined Deep Learning Model.

BACKGROUND: Patients with cancer can develop bone metastasis when a solid tumor invades the bone, which is the third most commonly affected site by metastatic cancer, after the lung and liver. The early detection of bone metastases is crucial for making appropriate treatment decisions and increasing survival rates. Deep learning, a mainstream branch of machine learning, has rapidly become an effective approach to analyzing medical images. OBJECTIVE: To automatically diagnose bone metastasis with bone scintigraphy, in this work, we proposed to cast the bone metastasis diagnosis problem into automated image classification by developing a deep learning-based automated classification model. METHODS: A self-defined convolutional neural network consisting of a feature extraction sub-network and feature classification sub-network was proposed to automatically detect lung cancer bone metastasis, with a feature extraction sub-network extracting hierarchal features from SPECT bone scintigrams and feature classification sub-network classifying high-level features into two categories (i.e., images with metastasis and without metastasis). RESULTS: Using clinical data of SPECT bone scintigrams, the proposed model was evaluated to examine its detection accuracy. The best performance was achieved if the two images (i.e., anterior and posterior scans) acquired from each patient were fused using pixel-wise addition operation on the bladder-excluded images, obtaining the best scores of 0.8038, 0.8051, 0.8039, 0.8039, 0.8036, and 0.8489 for accuracy, precision, recall, specificity, F-1 score, and AUC value, respectively. CONCLUSION: The proposed two-class classification network can predict whether an image contains lung cancer bone metastasis with the best performance as compared to existing classical deep learning models. The high accumulation of 99mTc MDP in the urinary bladder has a negative impact on automated diagnosis of bone metastasis. It is recommended to remove the urinary bladder before automated analysis.

Detecting multiple lesions of lung cancer-caused metastasis with bone scans using a self-defined object detection model based on SSD framework.

Objective.To facilitate manual diagnosis of lung cancer-caused metastasis, in this work, we propose a deep learning-based method to automatically identify and locate the hotspots in a bone scan image which denote the lesions metastasized from lung cancer.Approach.An end-to-end metastasis lesion detection model is proposed by following the classical object detection framework single shot multibox object detector (SSD). The proposed model casts lesion detection problem into automatically learning the hierarchal representations of lesion features, locating the spatial position of lesion areas, and boxing the detected lesions.Main results.Experimental evaluation conducted on clinical data of retrospective bone scans shows the comparable performance with a mean score of 0.7911 for average precision. A comparative analysis between our network and others including SSD shows the feasibility of the proposed detection network on automatically detecting multiple lesions of metastasis lesions caused by lung cancer.Significance.The proposed method has the potential to be used as an auxiliary tool for improving the accuracy and efficiency of metastasis diagnosis routinely conducted by nuclear medicine physicians.

Automated detection of lung cancer-caused metastasis by classifying scintigraphic images using convolutional neural network with residual connection and hybrid attention mechanism.

BACKGROUND: Whole-body bone scan is the widely used tool for surveying bone metastases caused by various primary solid tumors including lung cancer. Scintigraphic images are characterized by low specificity, bringing a significant challenge to manual analysis of images by nuclear medicine physicians. Convolutional neural network can be used to develop automated classification of images by automatically extracting hierarchal features and classifying high-level features into classes. RESULTS: Using convolutional neural network, a multi-class classification model has been developed to detect skeletal metastasis caused by lung cancer using clinical whole-body scintigraphic images. The proposed method consisted of image aggregation, hierarchal feature extraction, and high-level feature classification. Experimental evaluations on a set of clinical scintigraphic images have shown that the proposed multi-class classification network is workable for automated detection of lung cancer-caused metastasis, with achieving average scores of 0.7782, 0.7799, 0.7823, 0.7764, and 0.8364 for accuracy, precision, recall, F-1 score, and AUC value, respectively. CONCLUSIONS: The proposed multi-class classification model can not only predict whether an image contains lung cancer-caused metastasis, but also differentiate between subclasses of lung cancer (i.e., adenocarcinoma and non-adenocarcinoma). On the context of two-class (i.e., the metastatic and non-metastatic) classification, the proposed model obtained a higher score of 0.8310 for accuracy metric.

Automated detection of skeletal metastasis of lung cancer with bone scans using convolutional nuclear network.

A bone scan is widely used for surveying bone metastases caused by various solid tumors. Scintigraphic images are characterized by inferior spatial resolution, bringing a significant challenge to manual analysis of images by nuclear medicine physicians. We present in this work a new framework for automatically classifying scintigraphic images collected from patients clinically diagnosed with lung cancer. The framework consists of data preparation and image classification. In the data preparation stage, data augmentation is used to enlarge the dataset, followed by image fusion and thoracic region extraction. In the image classification stage, we use a self-defined convolutional neural network consisting of feature extraction, feature aggregation, and feature classification sub-networks. The developed multi-class classification network can not only predict whether a bone scan image contains bone metastasis but also tell which subcategory of lung cancer that a bone metastasis metastasized from is present in the image. Experimental evaluations on a set of clinical bone scan images have shown that the proposed multi-class classification network is workable for automated classification of metastatic images, with achieving average scores of 0.7392, 0.7592, 0.7242, and 0.7292 for accuracy, precision, recall, and F-1 score, respectively.

Tumor-targeting and imaging micelles for pH-triggered anticancer drug release and combined photodynamic therapy.

Herein, we construct a charge - switchable polymer nano micelles poly (2-(hexamethyl eneimino) ethyl methacrylate) - b - poly (ethylene glycol) monomethyl ether methacrylate) - b - poly (diethyl enetriaminepentaacetic acid methacrylate) - b - poly (1-vinyl imidazole) - b - poly (4-vinyl phenylboronic acid) (PC7A-PEG-DTPA-VI-PBA) in different pH solutions. DOX released faster from micelles in a weakly acidic environment (pH 5.0) than at pH 7.4. In order to enhance the anti-tumor effect, the imidazole functional groups in the polymer were used to coordinate CdSeTe quantum dots (QDs) for photodynamic treatment (PDT). In addition, the surfaces of the micelles were further decorated with phenylboronic acidas a targeting group, using DTPA chelating 99mTc for SPECT imaging.It has been successfully demonstrated that the nanoparticles have a good cumulative effect on the tumor site.The structure of the polymer was characterized by 1HNMR. The morphology and particle size of the micelles were characterized by transmission electron microscopy (TEM) and dynamic light scattering (DLS). The drug loading capacity (DLC) and drug loading efficiency (DLE) of the micelles were analyzed by ultraviolet visible spectroscopy. And the pH-sensitive drug release and cytotoxicity of the micelles were verified in vitro. In vitro experiments showed that the nano micelles were noncytotoxic to different cell lines, while DOX@CdSeTe@PC7A-PEG-DTPA-VI-PBA inhibited the proliferation and promoted the apoptosis of B16F10 cells. An in vivo study with C57BL tumor-bearing mice indicated that DOX@CdSeTe@PC7A-PEG-DTPA-VI-PBA nano micelles efficiently inhibited tumor growth. Results showed that the nano micelles had good pH responsibility and biocompatibility, and the loaded DOX could be released in the weak acidic environment of tumor cells, and it was expected to be a good drug delivery system.

A novel amino acid site closely associated with the neurovirulence of live, attenuated Japanese encephalitis vaccine (SA14-14-2 strain).

Japanese encephalitis (JE) poses a serious threat to the world's public health yet without a cure, the only way to prevent Japanese encephalitis virus (JEV) infection is vaccination. Live attenuated vaccine (SA14-14-2 strain) is the most widely used JE vaccine, and clinical data have confirmed its safety and effectiveness. Eight sitesassociated with virulence in the Envelope (E) protein are often the focus of quality control of JE vaccine. However, sequences retrieved from NCBI, as well as our previous results showed that the wild strain SA14 may harbor two different amino acids at amino acid residue 244 of the E glycoprotein (E244), and it may be related to virulence. In this study, we introduced a single mutation at nt1708 (G â†’ A) in the full-length cDNA clone of SA14-14-2, replacing a Gly with Glu at amino acid residue 244 of the E glycoprotein, and successfully constructed the mutant virus (JEV E244). JEV E244 exhibited a similar plaque morphology and growth characteristics to JEV SA14-14-2 in cell culture. However, it had lethal neurovirulence in mice and could enter the brain following intraperitoneal inoculation. Moreover, the virulence of JEV E244 in the context of vaccine in mice is significantly different from that of the JEV E244 alone. These results suggested that E244 site should be included in the assessment of the genetic stability of the attenuated JE vaccine. The detection of minor mutations in vaccine population and influence on the safety of vaccine is discussed.

Japanese encephalitis virus/yellow fever virus chimera is safe and confers full protection against yellow fever virus in intracerebrally challenged mice.

Yellow fever (YF) is an acute viral haemorrhagic disease caused by the yellow fever virus (YFV), which remains a potential threat to public health. The live-attenuated YF vaccine (17D strain) is a safe and highly effective measure against YF. However, increasing adverse events have been associated with YF vaccinations in recent years; thus, safer, alternative vaccines are needed. In this study, using the Japanese encephalitis live vaccine strain SA14-14-2 as a backbone, a novel chimeric virus was constructed by replacing the pre-membrane (prM) and envelope (E) genes with their YFV 17D counterparts.The chimeric virus exhibited a reduced growth rate and a much smaller plaque morphology than did either parental virus. Furthermore, the chimera was much less neurovirulent than was YF17D and protected mice that were challenged with a lethal dose of the YF virus. These results suggest that this chimera has potential as a novel attenuated YF vaccine.

Envelope Protein Mutations L107F and E138K Are Important for Neurovirulence Attenuation for Japanese Encephalitis Virus SA14-14-2 Strain.

The attenuated Japanese encephalitis virus (JEV) strain SA14-14-2 has been successfully utilized to prevent JEV infection; however, the attenuation determinants have not been fully elucidated. The envelope (E) protein of the attenuated JEV SA14-14-2 strain differs from that of the virulent parental SA14 strain at eight amino acid positions (E107, E138, E176, E177, E264, E279, E315, and E439). Here, we investigated the SA14-14-2-attenuation determinants by mutating E107, E138, E176, E177, and E279 in SA14-14-2 to their status in the parental virulent strain and tested the replication capacity, neurovirulence, neuroinvasiveness, and mortality associated with the mutated viruses in mice, as compared with those of JEV SA14-14-2 and SA14. Our findings indicated that revertant mutations at the E138 or E107 position significantly increased SA14-14-2 virulence, whereas other revertant mutations exhibited significant increases in neurovirulence only when combined with E138, E107, and other mutations. Revertant mutations at all eight positions in the E protein resulted in the highest degree of SA14-14-2 virulence, although this was still lower than that observed in SA14. These results demonstrated the critical role of the viral E protein in controlling JEV virulence and identified the amino acids at the E107 and E138 positions as the key determinants of SA14-14-2 neurovirulence.

A novel dengue virus serotype 1 vaccine candidate based on Japanese encephalitis virus vaccine strain SA14-14-2 as the backbone.

To develop a potential dengue vaccine candidate, a full-length cDNA clone of a novel chimeric virus was constructed using recombinant DNA technology, with Japanese encephalitis virus (JEV) vaccine strain SA14-14-2 as the backbone, with its premembrane (prM) and envelope (E) genes substituted by their counterparts from dengue virus type 1 (DENV1). The chimeric virus (JEV/DENV1) was successfully recovered from primary hamster kidney (PHK) cells by transfection with the in vitro transcription products of JEV/DENV1 cDNA and was identified by complete genome sequencing and immunofluorescent staining. No neuroinvasiveness of this chimeric virus was observed in mice inoculated by the subcutaneous route (s.c.) or by the intraperitoneal route (i.p.), while some neurovirulence was displayed in mice that were inoculated directly by the intracerebral route (i.c.). The chimeric virus was able to stimulate high-titer production of antibodies against DENV1 and provided protection against lethal challenge with neuroadapted dengue virus in mice. These results suggest that the chimeric virus is a promising dengue vaccine candidate.

Construction and preliminary investigation of a novel dengue serotype 4 chimeric virus using Japanese encephalitis vaccine strain SA14-14-2 as the backbone.

For the purpose of developing a novel dengue vaccine candidate, recombinant plasmids were constructed which contained the full length cDNA clone of Japanese encephalitis (JE) vaccine strain SA14-14-2 with its premembrane (PreM) and envelope (E) genes replaced by the counterparts of dengue virus type 4 (DENV4). By transfecting the in vitro transcription products of the recombinant plasmids into BHK-21 cells, a chimeric virus JEV/DENV4 was successfully recovered. The chimeric virus was identified by complete genome sequencing, Western blot and immunofluorescent staining. Growth characteristics revealed it was well adapted to primary hamster kidney (PHK) cells. Its genetic stability was investigated and only one unintentional mutation in 5'-untranslated region (5'-UTR) was found after 20 passages in PHK cells. Neurotropism, neurovirulence and immunogenicity of the chimeric virus were tested in mice. Besides, the influence of JE vaccine pre-immunization on the neutralizing antibody level induced by the chimeric virus was illuminated. To our knowledge, this is the first chimeric virus incorporating the JE vaccine stain SA14-14-2 and DENV4. It is probably a potential candidate to compose a tetravalent dengue chimeric vaccine.

Passive and active hepatoma tumor targeting of new N-(2-hydroxypropyl)methacrylamide copolymer conjugates: synthesis, characterization, and evaluation in vitro and in vivo.

Human hepatocellular carcinoma (HCC) is one of the major causes of death worldwide. To investigate the relative importance of active and passive targeting strategies, the synthesis, characterization, in vitro uptake, and in vivo biodistribution of specific sulfapyridine HPMA (HPMA: N-(2-hydroxypropyl methacrylamide)) copolymer (sulfapyridine: SPD) conjugates, nonspecific HPMA copolymer conjugates, and DTPA are described in this study. The poly(HPMA)-SPD-DTPA (DTPA: diethylenetriaminepentaacetic acid), poly(HPMA)-DTPA, and DTPA conjugates were radiolabeled with the radionuclide (99m)Tc and tested for uptake by cultured H22 cells. The cellular accumulation of poly(HPMA)-SPD-DTPA-(99m)Tc complex was found to be time-dependent. The poly(HPMA)-SPD-DTPA-(99m)Tc tracer exhibited rapid uptake kinetics in cell culture with a t(1/2) of ~5 min. The uptake of poly(HPMA)-SPD-DTPA-(99m)Tc was significantly higher than that of poly(HPMA)-DTPA-(99m)Tc, indicating that the uptake of the poly(HPMA)-SPD-DTPA-(99m)T was active binding. The uptake of poly(HPMA)-DTPA-(99m)Tc was significantly higher than that of DTPA-(99m)Tc, suggesting that the uptake of the poly(HPMA)-DTPA-(99m)T was passive binding. Twenty-four hour necropsy data in the hepatocellular carcinoma tumor model showed significantly higher (p < 0.001) tumor localization for poly(HPMA)-SPD-DTPA-(99m)Tc (4.98 ± 0.48%ID/g [percentage injected dose per gram tissue]) compared with poly(HPMA)-DTPA-(99m)Tc (2.69 ± 0.15% ID/g) and DTPA-(99m)Tc (0.83 ± 0.03%ID/g). Moreover, higher T/B for poly(HPMA)-SPD-DTPA-(99m)Tc indicated reduced extravazation of the targeted polymeric conjugates in normal tissues. Specific molecular targeting and nonspecific vascular permeability are both significant in the relative tumor localization of poly(HPMA)-SPD-DTPA-(99m)Tc. Extravascular leak in nonspecific organs appears to be a major factor in reducing the T/B for the sulfapyridine molecules. Thus, the poly(HPMA)-SPD-DTPA is expected to be used as the potential macromolecular targeting carrier for hepatoma carcinoma in mice.

Live attenuated measles virus vaccine induces apoptosis and promotes tumor regression in lung cancer.

Although the treatment of lung carcinoma has improved, at least 65% of patients with this tumor succumb to progressive disease. Measles virus oncolytic therapy has been reported to be effective in reducing tumor burden in immunocompetent or nude mice; however, its potential to reduce tumor burden in lung carcinoma remains to be determined. Herein, we report the potent antitumor effects of a live attenuated measles vaccine virus Hu-191 strain (MV) against lung carcinoma. Immunocompetent C57BL/6 mice bearing Lewis lung carcinoma (LLC) cells were treated with MV (1x104 to 1x106 CCID50/ml) once every other day for 10 days. Our results showed that treatment with MV effectively suppressed tumor growth and significantly prolonged the survival time of tumor-bearing animals. Histological examination revealed that the antitumor effects of MV therapy may result from increased induction of apoptosis, tumor necrosis and elevated lymphocyte infiltration. Our data suggest that MV, one of the widely used vaccines in China, has the ability to inhibit the growth of mouse lung carcinoma and may prove useful in the further exploration of the application of this approach in the treatment of human advanced lung cancer.

A cis element within flowering locus T mRNA determines its mobility and facilitates trafficking of heterologous viral RNA.

The Arabidopsis flowering locus T (FT) gene encodes the mobile florigen essential for floral induction. While movement of the FT protein has been shown to occur within plants, systemic spread of FT mRNA remains to be unequivocally demonstrated. Utilizing novel RNA mobility assay vectors based on two distinct movement-defective viruses, Potato virus X and Turnip crinkle virus, and an agroinfiltration assay, we demonstrate that nontranslatable FT mRNA, independent of the FT protein, moves throughout Nicotiana benthamiana and mutant Arabidopsis plants and promotes systemic trafficking of viral and green fluorescence protein RNAs. Viral ectopic expression of FT induced flowering in the short-day N. tabacum Maryland Mammoth tobacco under long-day conditions. Recombinant Potato virus X bearing FT RNA spread and established systemic infection more quickly than the parental virus. The cis-acting element essential for RNA movement was mapped to the nucleotides 1 to 102 of the FT mRNA coding sequence. These data demonstrate that a plant self-mobile RNA molecule can mediate long-distance trafficking of heterologous RNAs and raise the possibility that FT RNA, along with the FT protein, may be involved in the spread of the floral stimulus throughout the plant.

[Study on the gene mutation of quinolone-resistant Mycobacterium tuberculosis isolated from Sichuan Province].

OBJECTIVE: To investigate the molecular mechanism of quinolone-resistance of M. tuberculosis and characterize the gene mutation in Sichuan Province. METHODS: Susceptibility of the clinical isolates to quinolones (ofloxacin, ciprofloxacin and sparfloxacin) was tested by the absolute concentration method. GyrA gene quinolone reasistance-determining region (QRDR) mutations M. tuberculosis were detected with PCR-SSCP and DNA sequencing. RESULTS: Of 68 clinical isolates of M. tuberculosis, 25 high-resistant, 11 low-resistant and 10 sensitive isolates were noted to have abnormal gyrA SSCP profile and different gyrA sequences from the standard strain H37Rv, and 14 sensitive and 8 low-resistant isolates were found with no mutation of gyrA gene. DNA sequencing unveiled Ser-->Thr mutation at codon 95, Asp-->Gly at codon 94, Ala-->Val at codon 90, and Ala-->Val at codon 83. CONCLUSION: This study confirmed the strong correlation between the quinolone-resistance and the mutation of gyrA gene, which might be a major molecular mechanism of quinolone-resistance in M. tuberculosis. The types of mutations exhibit no difference between Sichuan Province and other areas in China.