Application of deep learning and feature selection technique on external root resorption identification on CBCT images.

BACKGROUND: Artificial intelligence has been proven to improve the identification of various maxillofacial lesions. The aim of the current study is two-fold: to assess the performance of four deep learning models (DLM) in external root resorption (ERR) identification and to assess the effect of combining feature selection technique (FST) with DLM on their ability in ERR identification. METHODS: External root resorption was simulated on 88 extracted premolar teeth using tungsten bur in different depths (0.5 mm, 1 mm, and 2 mm). All teeth were scanned using a Cone beam CT (Carestream Dental, Atlanta, GA). Afterward, a training (70%), validation (10%), and test (20%) dataset were established. The performance of four DLMs including Random Forest (RF) + Visual Geometry Group 16 (VGG), RF + EfficienNetB4 (EFNET), Support Vector Machine (SVM) + VGG, and SVM + EFNET) and four hybrid models (DLM + FST: (i) FS + RF + VGG, (ii) FS + RF + EFNET, (iii) FS + SVM + VGG and (iv) FS + SVM + EFNET) was compared. Five performance parameters were assessed: classification accuracy, F1-score, precision, specificity, and error rate. FST algorithms (Boruta and Recursive Feature Selection) were combined with the DLMs to assess their performance. RESULTS: RF + VGG exhibited the highest performance in identifying ERR, followed by the other tested models. Similarly, FST combined with RF + VGG outperformed other models with classification accuracy, F1-score, precision, and specificity of 81.9%, weighted accuracy of 83%, and area under the curve (AUC) of 96%. Kruskal Wallis test revealed a significant difference (p = 0.008) in the prediction accuracy among the eight DLMs. CONCLUSION: In general, all DLMs have similar performance on ERR identification. However, the performance can be improved by combining FST with DLMs.

Three-dimensional surface area analyses of developing maxillary second premolar root apices for age estimation using CBCT images.

Three-dimensional surface area analyses of developing root apices for age estimation in children and young adults have shown promising results. The current study aimed to apply this three-dimensional method to develop a regression model for estimating age in Malaysian children aged 7 to 14 using developing maxillary second premolars. A training sample of 155 cone-beam computed tomography scans (83 Malays and 72 Chinese) was analysed, and the formula was subsequently validated on an independent sample of 92 cone-beam computed tomography scans (45 Malays and 47 Chinese). The results showed a strong correlation (r = 94 %) between the chronological age as a dependent variable and the predictor variables, including root surface area of the apex, sex, ethnicity, and root development status (open/closed apices). For this model, the predictor variables accounted for 88.4 % of the variation in age except sex and ethnicity. A mean absolute error value of 0.42 indicated that this model can be reliably used for Malaysian children. In conclusion, this study recognises the method of three-dimensional surface area analyses as a valuable tool for age estimation in forensic and clinical practice. Further studies are highly recommended to assess its effectiveness across different demographic groups.