Advanced search
Start date
Betweenand

Effect of the reduction of artefacts arising from high-density materials on cone-beam CT by artificial intelligence on the diagnosis of vertical root fracture and root perforation

Grant number: 21/01623-8
Support type:Scholarships in Brazil - Post-Doctorate
Effective date (Start): March 01, 2022
Status:Discontinued
Field of knowledge:Health Sciences - Dentistry - Endodontics
Principal researcher:Manoel Damiao de Sousa Neto
Grantee:Amanda Pelegrin Candemil
Home Institution: Faculdade de Odontologia de Ribeirão Preto (FORP). Universidade de São Paulo (USP). Ribeirão Preto , SP, Brazil
Associated scholarship(s):22/07081-5 - Effect of the reduction of artefacts arising from high-density materials on cone-beam ct by artificial intelligence on the diagnosis of vertical root fracture and root perforation, BE.EP.PD

Abstract

Cone-Beam Computed Tomography (CBCT) is an imaging exam used in endodontics to diagnose diseases and evaluate internal root anatomy, however, one of its limitations is the presence of artefacts that hamper the diagnostic efficacy. Although the literature presents different techniques to reduce artefacts, till this moment, there are no effective methods that contribute to its reduction. Deep convolutional neural network has shown promising approaches to reduce the presence of artefacts arising from high-density materials, however, besides the scientific literature only presents preliminary data, with neural networks with low training by few CBCT images, there is no information about its impact on diagnostic accuracy of clinical tasks such as root fracture and perforation. Therefore, this study aims to create a deep convolutional neural network to reduce artefacts arising from high-density materials and low X-ray dose, based in CBCT scans, evaluate the impact in the diagnostic accuracy of different clinical tasks and analyze and compare its efficacy with others artefact reduction methods. An imaging phantom will be custom made with a partially edentulous macerated human mandible. One hundred extracted human teeth will be endodontically instrumented and then divided in two steps: in the step 1, 40 instrumented and filled teeth will be used to simulate artefacts from high-density materials (scattering, beam hardening, starvation and blooming) to the training of the deep convolutional neural network. And, then, after the development of the deep convolutional neural network, in the step 2, the remain 60 teeth will be divided in three groups (n=20) in accordance with the proposed endodontic complication: vertical root fracture, root perforation and control. Each teeth group will be positioned individually in an empty socket of the human mandible and CBCT scans will be obtained with a field of view of 5 × 5 cm at two resolution protocols (high and low), with/without metal artefact reduction algorithm. Additional scans will be obtained from the teeth of interest in the empty sockets of the mandible to simulate different clinical conditions high-density materials (filling material, implant and intracanal post). Ad-hoc deep convolutional neural network will be developed based on the literature of the U-net network. It will be tailored to the properly prepared CBCT images of the imaging phantom that will be used for network validation and training. The CBCT images will be corrected by the trained deep convolutional neural network and will be analyzed subjectively by four observers to indicate the presence of vertical root fracture and root perforation with a 5-point scale. And, additionally, the images will be analyzed objectively by one observer, that will obtain measurements of the signal-to-noise ratio and gray value homogeneity of a homogeneous area. All the images will be evaluated in the CBCT software's OsiriX MD v.7.5.1 and e-Vol DX and will be posteriorly compared.

News published in Agência FAPESP Newsletter about the scholarship:
Articles published in other media outlets (0 total):
More itemsLess items
VEICULO: TITULO (DATA)
VEICULO: TITULO (DATA)

Please report errors in scientific publications list by writing to: cdi@fapesp.br.