Análise de Agenesia Dentária e Anomalias do Desenvolvimento em Pacientes Não Sindrômicos

Abstract

Dental agenesis is one of the most prevalent congenital anomalies of the oral cavity, characterized by the absence of one or more teeth due to alterations in the odontogenesis process. Its etiology is multifactorial, involving an interaction between genetic, epigenetic, and environmental factors. This study analyzed the genetic perspective, the association between dental agenesis and root morphology, its relationship with other developmental anomalies, and specific facial proportions. Chapter 1 explores, through a narrative review, the genetic basis of nonsyndromic dental agenesis, highlighting genes such as MSX1, PAX9, AXIN2, EDA, LRP6, WNT10A, and WNT10B, and the impact of genetic and epigenetic variations. Chapter 2 investigates the association between dental agenesis and root morphology through the analysis of 1,193 periapical radiographs. The prevalence of agenesis was 6.6%, with a higher occurrence in women. Patients with agenesis were more likely to present pipette-shaped roots (OR: 7.04; 95%CI: 4.23–11.72) and blunt roots (OR: 5.31; 95%CI: 2.62–10.75). Chapter 3 explores the relationship between dental agenesis and other developmental anomalies based on 4,990 panoramic radiographs, revealing significant associations with dental transposition (OR: 19.11), palatal displacement of canines (OR: 2.08), and molar infraocclusion (OR: 147.73). Machine Learning (ML) models such as Random Forest, XGBoost, SVM, and Deep Learning were applied and showed good predictive performance. Chapter 4 expands the diagnostic approach using facial measurements extracted from standardized photographs 2D. In a sample of 575 individuals (175 cases and 350 controls), measurements such as en-en, ex-ex, and zi-zi showed significant differences between groups (p < 0.001). Among the ML models tested, the Artificial Neural Network (ANN) performed best (AUC: 0.778; accuracy: 73.6%), followed by Random Forest (AUC: 0.760). The results suggest that specific facial proportions can be used as reliable predictors for the identification of dental agenesis through non-invasive approaches. Based on the results, it is concluded that nonsyndromic dental agenesis has a complex genetic basis, often associated with root morphological alterations, other developmental anomalies, and specific facial proportions. The integrated use of artificial intelligence, clinical, and biometric data significantly contributes to early diagnosis and multidisciplinary clinical planning.