Your browser doesn't support javascript.
loading
Show: 20 | 50 | 100
Results 1 - 4 de 4
Filter
Add more filters










Database
Language
Publication year range
1.
J Digit Imaging ; 35(5): 1101-1110, 2022 Oct.
Article in English | MEDLINE | ID: mdl-35478060

ABSTRACT

To visualise the tumours inside the body on a screen, a long and thin tube is inserted with a light source and a camera at the tip to obtain video frames inside organs in endoscopy. However, multiple artefacts exist in these video frames that cause difficulty during the diagnosis of cancers. In this research, deep learning was applied to detect eight kinds of artefacts: specularity, bubbles, saturation, contrast, blood, instrument, blur, and imaging artefacts. Based on transfer learning with pre-trained parameters and fine-tuning, two state-of-the-art methods were applied for detection: faster region-based convolutional neural networks (Faster R-CNN) and EfficientDet. Experiments were implemented on the grand challenge dataset, Endoscopy Artefact Detection and Segmentation (EAD2020). To validate our approach in this study, we used phase I of 2,200 frames and phase II of 331 frames in the original training dataset with ground-truth annotations as training and testing dataset, respectively. Among the tested methods, EfficientDet-D2 achieves a score of 0.2008 (mAPd[Formula: see text]0.6+mIoUd[Formula: see text]0.4) on the dataset that is better than three other baselines: Faster-RCNN, YOLOv3, and RetinaNet, and competitive to the best non-baseline result scored 0.25123 on the leaderboard although our testing was on phase II of 331 frames instead of the original 200 testing frames. Without extra improvement techniques beyond basic neural networks such as test-time augmentation, we showed that a simple baseline could achieve state-of-the-art performance in detecting artefacts in endoscopy. In conclusion, we proposed the combination of EfficientDet-D2 with suitable data augmentation and pre-trained parameters during fine-tuning training to detect the artefacts in endoscopy.


Subject(s)
Artifacts , Neural Networks, Computer , Humans , Endoscopy , Machine Learning
2.
Biomaterials ; 252: 120094, 2020 09.
Article in English | MEDLINE | ID: mdl-32422495

ABSTRACT

Healing of large calvarial bone defects remains a challenging task in the clinical setting. Although BMP2 (bone morphogenetic protein 2) is a potent growth factor that can induce bone repair, BMP2 provokes the expression of antagonist Noggin that self-restricts its bioactivity. CRISPR interference (CRISPRi) is a technology for programmable gene suppression but its application in regenerative medicine is still in its infancy. We reasoned that Nog inhibition, concurrent with BMP2 overexpression, can promote the osteogenesis of adipose-derived stem cells (ASC) and improve calvarial bone healing. We designed and exploited a hybrid baculovirus (BV) system for the delivery of BMP2 gene and CRISPRi system targeting Nog. After BV-mediated co-delivery into ASC, the system conferred prolonged BMP2 expression and stimulated Nog expression while the CRISPRi system effectively repressed Nog upregulation for at least 14 days. The CRISPRi-mediated Nog knockdown, along with BMP2 overexpression, additively stimulated the osteogenic differentiation of ASC. Implantation of the CRISPRi-engineered ASC into the critical size defects at the calvaria significantly enhanced the calvarial bone healing and matrix mineralization. These data altogether implicate the potentials of CRISPRi-mediated gene knockdown for cell fate regulation and tissue regeneration.


Subject(s)
Bone Morphogenetic Protein 2 , Osteogenesis , Bone Morphogenetic Protein 2/genetics , Bone Regeneration , Cell Differentiation , Clustered Regularly Interspaced Short Palindromic Repeats , Skull , Stem Cells
3.
Mol Ther ; 28(2): 441-451, 2020 02 05.
Article in English | MEDLINE | ID: mdl-31882321

ABSTRACT

CRISPR activation (CRISPRa) is a burgeoning technology for programmable gene activation, but its potential for tissue regeneration has yet to be fully explored. Bone marrow-derived mesenchymal stem cells (BMSCs) can differentiate into osteogenic or adipogenic pathways, which are governed by the Wnt (Wingless-related integration site) signaling cascade. To promote BMSC differentiation toward osteogenesis and improve calvarial bone healing by BMSCs, we harnessed a highly efficient hybrid baculovirus vector for gene delivery and exploited a synergistic activation mediator (SAM)-based CRISPRa system to activate Wnt10b (that triggers the canonical Wnt pathway) and forkhead c2 (Foxc2) (that elicits the noncanonical Wnt pathway) in BMSCs. We constructed a Bac-CRISPRa vector to deliver the SAM-based CRISPRa system into rat BMSCs. We showed that Bac-CRISPRa enabled CRISPRa delivery and potently activated endogenous Wnt10b and Foxc2 expression in BMSCs for >14 days. Activation of Wnt10b or Foxc2 alone was sufficient to promote osteogenesis and repress adipogenesis in vitro. Furthermore, the robust and prolonged coactivation of both Wnt10b and Foxc2 additively enhanced osteogenic differentiation while inhibiting adipogenic differentiation of BMSCs. The CRISPRa-engineered BMSCs with activated Wnt10b and Foxc2 remarkably improved the calvarial bone healing after implantation into the critical-sized calvarial defects in rats. These data implicate the potentials of CRISPRa technology for bone tissue regeneration.


Subject(s)
Bone Regeneration/genetics , Forkhead Transcription Factors/genetics , Mesenchymal Stem Cells/metabolism , Osteogenesis/genetics , Transcriptional Activation , Wnt Proteins/genetics , Adipogenesis , Animals , Calcification, Physiologic , Calcium/metabolism , Cell Differentiation/genetics , Cells, Cultured , Rats , Skull/diagnostic imaging , Skull/metabolism , Wnt Signaling Pathway , X-Ray Microtomography
4.
Theranostics ; 9(21): 6099-6111, 2019.
Article in English | MEDLINE | ID: mdl-31534539

ABSTRACT

Background: Peripheral nerve regeneration requires coordinated functions of neurotrophic factors and neuronal cells. CRISPR activation (CRISPRa) is a powerful tool that exploits inactive Cas9 (dCas9), single guide RNA (sgRNA) and transcription activator for gene activation, but has yet to be harnessed for tissue regeneration. Methods: We developed a hybrid baculovirus (BV) vector to harbor and deliver the CRISPRa system for multiplexed activation of 3 neurotrophic factor genes (BDNF, GDNF and NGF). The hybrid BV was used to transduce rat adipose-derived stem cells (ASC) and functionalize the ASC sheets. We further implanted the ASC sheets into sciatic nerve injury sites in rats. Results: Transduction of rat ASC with the hybrid BV vector enabled robust, simultaneous and prolonged activation of the 3 neurotrophic factors for at least 21 days. The CRISPRa-engineered ASC sheets were able to promote Schwann cell (SC) migration, neuron proliferation and neurite outgrowth in vitro. The CRISPRa-engineered ASC sheets further enhanced in vivo functional recovery, nerve reinnervation, axon regeneration and remyelination. Conclusion: These data collectively implicated the potentials of the hybrid BV-delivered CRISPRa system for multiplexed activation of endogenous neurotrophic factor genes in ASC sheets to promote peripheral nerve regeneration.


Subject(s)
Clustered Regularly Interspaced Short Palindromic Repeats , Nerve Growth Factors/metabolism , Nerve Regeneration , Adipocytes/physiology , Adipose Tissue , Animals , Axons/physiology , Baculoviridae/genetics , Cell Movement , Cell Proliferation , Female , Mesenchymal Stem Cells , Nerve Growth Factors/genetics , Neurons/physiology , Peripheral Nerves/physiology , RNA, Guide, Kinetoplastida/genetics , Rats , Rats, Sprague-Dawley , Recovery of Function , Schwann Cells/physiology , Streptococcus pyogenes/enzymology , Streptococcus pyogenes/genetics
SELECTION OF CITATIONS
SEARCH DETAIL
...