HA011 - Hatton
Área: 7 - Imaginologia

Objective assessment of the combined effect of exomass-related and motion artefacts in Cone beam CT
Candemil A P, Oliveira ML, Freitas DQ, Haiter Neto F, Wenzel A, Spin Neto R
Odontologia Restauradora - UNIVERSIDADE DE SÃO PAULO - RIBEIRÃO PRETO
Conflito de interesse: Não há conflito de interesse

The aim of this study was to assess the combined effect of cone-beam computed tomography (CBCT) exomass-related and patient motion artefacts. A cylindrical phantom containing 21 tubes filled with a radiopaque solution, allowing the inclusion of three titanium implants in the periphery to induce exomass-related artefacts, was mounted on a robot simulating 0.75-, 1.50-, and 3-mm movements (nodding/lateral rotation/tremor). CBCT images with/without exomass and with/without movements were acquired, in duplicate, in three units. Voxel value mean and standard deviation were assessed from each tube. For each CBCT volume, the 21 mean voxel values were averaged providing the overall mean voxel value (OMVV), and the standard deviation was calculated providing overall voxel value inhomogeneity (OVVI). The standard deviation from each of the 21 volumes-of-interest were averaged, providing overall image noise (ON). OMVV, OVVI, and ON were averaged for the duplicate acquisitions. The effect of the diverse tested conditions was inferred from a repeated-measures analysis of variance, followed by Sidak's test (α=0.05). Images acquired with exomass had lower OMVV, and higher OVVI and ON. Movement artefacts aggravated exomass-related alterations. OMVV and OVVI were mostly affected by 3-mm nodding movements. Motion-artefact correction was effective.
In conclusion, CBCT images are altered by exomass-related artefacts, and this finding is aggravated in the presence of motion artefacts. Motion-artefact correction eliminated the impact of movement.
(Apoio: CAPES  N° 001  |  Pró-Reitoria de Pós-Graduação at UNICAMP  N° 11/2019)

---

HA012 - Hatton
Área: 7 - Imaginologia

Automated Identification of Dental Implants by Using Artificial Intelligence
Santos RPM, Oliveira GAA, Aranha-Neto IS, Silva AIV, Alves TKC, Carmelo JC, Manzi FR
Odontologia - PONTIFÍCIA UNIVERSIDADE CATÓLICA DE MINAS GERAIS
Conflito de interesse: Não há conflito de interesse

The aim of this study was to develop and evaluate accuracy of a computer assisted system based on Artificial Intelligence (AI) for detecting and identifying automatically Dental Implants(DI) brands using digital periapical radiographs.A total amount of 1800 digital periapical radiographs with DI from three distinct manufacturers (f1,f2,f3) f1=600, f2=600 and f3=600 was split into training dataset (n = 1440 [80%]) and testing dataset (n = 360 [20%]) groups. The images were evaluated by a software developed through Convolutional Neural Networks (CNN) aimed to identify the manufacturer of IDs contained therein. Accuracy, Sensitivity, Specificity, positive and negative predictive values, and ROC curve were calculated for detection and diagnostic performance of CNN algorithm. At the final epoch (25), it was obtained 99.78% of system accuracy for training data, 99.36% for testing data and 85.29% for validation data. The latest one corresponds to the actual accuracy of dental implant manufacturer identification contained in digital periapical radiographs after the system learning process.
The results obtained in this study show that Deep CNN algorithm provides high accuracy for identifying dental implants by means of digital periapical radiographs, being a useful tool in odontological practice. With a more comprehensive data bank, this system may be widely used helping dentists to work with more predictability and to eliminate the challenge of discovering the implant model installed in patients when there is no previous treatment information.

---

HA013 - Hatton
Área: 7 - Patologia Oral

Phenotypic characterization of macrophages during induced apical periodontitis in mice
Pucinelli CM, Nelson-Filho P, León JE, Faccioli LH, Sorgi CA, Silva LAB, Segato RAB
Clínica Infantil - UNIVERSIDADE DE SÃO PAULO - RIBEIRÃO PRETO
Conflito de interesse: Não há conflito de interesse

The aim was to describe the M1 and M2 macrophages characterization by biomarker expression during the apical periodontitis (AP) formation. 130 wild-type mice were divided into control (sound teeth; n=50) and experimental (teeth with AP; n=80) groups. After 5 experimental periods, all animals were euthanized and the specimens submitted to histotechnical process in order to describe the apical and periapical tissue characteristics under conventional microscopy. Also, under fluorescence microscopy, morphometry was performed to measure the AP area. Were performed qRT-PCR for Cxcl10, CxCL9, iNos2, Arg1, Ym1, Fizz1 and MRC1 and Luminex® assay for GM-CSF, IFN-γ, IL-4, IL-13, IL-10, IL-6, IL-1β and TNF-α. ANOVA test and Tukey post-test were performed for morphometry. For qRT-PCR and Luminex®, Kruskal-Wallis test and Dunn post-test were performed. It was used GraphPad Prism 7.0a software (α=5%). The results showed a dynamically progression of AP, with a progressive destruction of the periodontal ligament, cement and alveolar bone. Markers for both, M1 and M2 macrophages, were present at different levels throughout the experimental periods, suggesting the occurrence of compensatory pathways.
In conclusion both macrophages were detected at different levels during all periods. It was observed a predominance of M1 phenotype in the early periods and M2 phenotype at 14 and 21 days, but returning to the M1 phenotype when the lesion was established.
(Apoio: Fapesp  N° 2016/24900-9  |  CAPES  N° 88881.190601/2018-01  |  CNPq  N° 140139/2019-2)

---

HA014 - Hatton
Área: 8 - Periodontia

Epigenetic regulation in human periodontal ligament cells with distinct osteogenic potential: Genome-Wide analysis
Assis RIF, Racca F, Ferreira RS, Nociti-Júnior FH, Ruiz KGS, Silva RA, Wiench M, Andia DC
Periodontia e Protese - FACULDADE DE ODONTOLOGIA DE PIRACICABA
Conflito de interesse: Não há conflito de interesse

Human periodontal ligament cells (PDLCs) can show distinct osteogenic potential, limiting their clinical applications. Through genomic approaches, our goal is to identify markers to assess osteogenic potential. PDLCs with high (h-) and low (l-) osteogenic potential were cultivated in control medium or osteogenic medium for 10 days, in vitro. DNA, chromatin and RNA were submitted to methylome, Assay for transposase accessible chromatin (ATAC) and RNA protocols, followed by next generation sequencing and bioinformatics analysis. After middle osteogenesis, results showed more differentially methylated probes (hyper/hypo) in l-PDLCs than in h- (74 x 41) with common methylated regions at gene body. There were more regions with accessible chromatin in the l-PDLCs, mainly in genes related to osteogenesis (RUNX2 l-43 x h-8; SP7 l-5 x h-1), transcription factors (Gli l-70 x h-17, Rab23) and pluripotency (OCT4 l-4 x h-0). The regulatory regions associated with cell identity (superenhancers) activated in l- were distinct from h-PDLCs. RNA-seq confirmed a different transcriptional pattern.
We unravel the epigenetic and transcriptional patterns of PDLCs with distinct osteogenic potential, showing distinct epigenetic regulation and transcriptional background, and this could affect gene markers and therefore, the clinical application outcome.
(Apoio: FAPs - FAPESP   N° 2019/01727-8  |  FAPs - FAPESP   N° 2017/12158-9  |  FAPs - FAPESP   N° 2017/07944-5)

---

HA015 - Hatton
Área: 8 - Periodontia

Interactions between hemodialysis vintage with xerostomia and tooth loss on oral health-related quality of life
Oliveira LM, Schöffer C, Santi SS, Antoniazzi RP, Zanatta FB
Estomatologia - UNIVERSIDADE FEDERAL DE SANTA MARIA
Conflito de interesse: Não há conflito de interesse

The aim of the present cross-sectional study was to evaluate whether tooth loss and xerostomia are independently associated with oral health-related quality of life (OHRQoL) in individuals with end-stage renal disease (ESRD) and whether hemodialysis vintage (HV) modifies the effect of these associations. One hundred eighty adults with ESRD were assessed by calibrated examiners. Oral examinations were performed for untreated dental caries, periodontitis, tooth loss (≤ 8 or > 8 teeth) and xerostomia (absent or present). HV was determined using hospital records and categorized as ≤ 12, 13-72 or ≥ 73 months. OHRQoL was assessed using the simplified version of the Oral Health Impact Profile (OHIP14) questionnaire. Adjusted multivariate negative binomial regression analysis was used to calculate the rate ratio (RR) of OHIP14 scores according to the exposures of interest and interactions with HV. In the adjusted model, xerostomia and tooth loss were significantly associated with poorer OHRQoL. Moreover, longer HV (≥ 73 months) negatively modified the effect of the associations between the exposures [xerostomia (RR = 2.96; 95% CI: 1.48 to 5.90); tooth loss (RR = 4.13; 95% CI: 2.00 to 8.51)] and outcome.
Xerostomia and tooth loss exert an influence on OHRQoL in individuals with end-stage renal disease, with a greater impact found among those on hemodialysis for a longer period of time. Policy makers should consider the greater impact of oral conditions on OHRQoL in extended HV by giving additional priority to patients on long-term pretransplant hemodialysis.
(Apoio: CAPES  N° 001)

---