Objective and Impact Statement:The multi-quantification of the distinct individualized maxillofacial traits,that is,quantifying multiple indices,is vital for diagnosis,decision-making,and prognosis of the maxillofacia...Objective and Impact Statement:The multi-quantification of the distinct individualized maxillofacial traits,that is,quantifying multiple indices,is vital for diagnosis,decision-making,and prognosis of the maxillofacial surgery.Introduction:While the discrete and demographically disproportionate distributions of the multiple indices restrict the generalization ability of artificial intelligence(AI)-based automatic analysis,this study presents a demographic-parity strategy for AI-based multi-quantification.Methods:In the aesthetic-concerning maxillary alveolar basal bone,which requires quantifying a total of 9 indices from length and width dimensional,this study collected a total of 4,000 cone-beam computed tomography(CBCT)sagittal images,and developed a deep learning model composed of a backbone and multiple regression heads with fully shared parameters to intelligently predict these quantitative metrics.Through auditing of the primary generalization result,the sensitive attribute was identified and the dataset was subdivided to train new submodels.Then,submodels trained from respective subsets were ensembled for final generalization.Results:The primary generalization result showed that the AI model underperformed in quantifying major basal bone indices.The sex factor was proved to be the sensitive attribute.The final model was ensembled by the male and female submodels,which yielded equal performance between genders,low error,high consistency,satisfying correlation coefficient,and highly focused attention.The ensemble model exhibited high similarity to clinicians with minor processing time.Conclusion:This work validates that the demographic parity strategy enables the AI algorithm with greater model generalization ability,even for the highly variable traits,which benefits for the appearance-concerning maxillofacial surgery.展开更多
Owing to the tooth-centered nature of most oral diseases,the tooth-centric radial plane of cone-beam computed tomography(CBCT)depicts the anatomical and pathological features along the long axis of the tooth,serving a...Owing to the tooth-centered nature of most oral diseases,the tooth-centric radial plane of cone-beam computed tomography(CBCT)depicts the anatomical and pathological features along the long axis of the tooth,serving as a crucial imaging modality in the diagnosis,treatment planning,and prognosis of multiple oral diseases.However,reconstructing these standard planes from CBCT is labor-intensive,time-consuming,and error-prone due to anatomical variances and multi-center discrepancies.This study proposes an expertise-inspired artificial intelligence(AI)pipeline for the reconstruction of the tooth-centric radial plane.By emulating expert's workflow,this AI pipeline acquires the optimized maxillary and mandibular cross sections,segments the teeth for dental arch curve depiction,and reconstructs dental arch-defined tooth-centric radial planes.A total of 420 CBCT scans from two independent centers,comprising both healthy and diseased subjects,were collected for model development and validation.Teeth on the optimized cross sections were explicitly segmented even in the presence of various complex diseases,resulting in precise dental arch curve depictions.The AI-reconstructed tooth-centric radial planes for all teeth exhibited low angular and distance errors compared with the ground truth planes.In terms of clinical utility,the AI-reconstructed planes demonstrated high image quality,accurately represented anatomical and pathological features,and facilitated precise dental biometrics measurement by both clinicians and downstream AI diagnostic tools.The expertise-inspired AI pipeline showcases outstanding performance in reconstructing tooth-centric radial planes and offers significant clinical utility for intelligent oral health management with high interpretability,robustness and generalization capabilities.展开更多
基金supported by the Guangzhou Science and Technology Project(no.2023B03J1232)National Natural Science Foundation of China(82301036)+1 种基金Special Funds for the Cultivation of Guangdong College Students’Scientific and Technological Innovation(no.pdjh2023b0013)Undergraduate Training Program for Innovation of Sun Yat-sen University(20240518).
文摘Objective and Impact Statement:The multi-quantification of the distinct individualized maxillofacial traits,that is,quantifying multiple indices,is vital for diagnosis,decision-making,and prognosis of the maxillofacial surgery.Introduction:While the discrete and demographically disproportionate distributions of the multiple indices restrict the generalization ability of artificial intelligence(AI)-based automatic analysis,this study presents a demographic-parity strategy for AI-based multi-quantification.Methods:In the aesthetic-concerning maxillary alveolar basal bone,which requires quantifying a total of 9 indices from length and width dimensional,this study collected a total of 4,000 cone-beam computed tomography(CBCT)sagittal images,and developed a deep learning model composed of a backbone and multiple regression heads with fully shared parameters to intelligently predict these quantitative metrics.Through auditing of the primary generalization result,the sensitive attribute was identified and the dataset was subdivided to train new submodels.Then,submodels trained from respective subsets were ensembled for final generalization.Results:The primary generalization result showed that the AI model underperformed in quantifying major basal bone indices.The sex factor was proved to be the sensitive attribute.The final model was ensembled by the male and female submodels,which yielded equal performance between genders,low error,high consistency,satisfying correlation coefficient,and highly focused attention.The ensemble model exhibited high similarity to clinicians with minor processing time.Conclusion:This work validates that the demographic parity strategy enables the AI algorithm with greater model generalization ability,even for the highly variable traits,which benefits for the appearance-concerning maxillofacial surgery.
基金National Natural Science Foundation of China,Grant/Award Number:82402380Undergraduate Training Program for Innovation of Sun Yat-sen University,Grant/Award Number:20240518+1 种基金Special Funds for the Cultivation of Guangdong College Students'Scientific and Technological Innovation,Grant/Award Number:pdjh2023b0013Guangzhou Science and Technology Project,Grant/Award Number:2023B03J1232。
文摘Owing to the tooth-centered nature of most oral diseases,the tooth-centric radial plane of cone-beam computed tomography(CBCT)depicts the anatomical and pathological features along the long axis of the tooth,serving as a crucial imaging modality in the diagnosis,treatment planning,and prognosis of multiple oral diseases.However,reconstructing these standard planes from CBCT is labor-intensive,time-consuming,and error-prone due to anatomical variances and multi-center discrepancies.This study proposes an expertise-inspired artificial intelligence(AI)pipeline for the reconstruction of the tooth-centric radial plane.By emulating expert's workflow,this AI pipeline acquires the optimized maxillary and mandibular cross sections,segments the teeth for dental arch curve depiction,and reconstructs dental arch-defined tooth-centric radial planes.A total of 420 CBCT scans from two independent centers,comprising both healthy and diseased subjects,were collected for model development and validation.Teeth on the optimized cross sections were explicitly segmented even in the presence of various complex diseases,resulting in precise dental arch curve depictions.The AI-reconstructed tooth-centric radial planes for all teeth exhibited low angular and distance errors compared with the ground truth planes.In terms of clinical utility,the AI-reconstructed planes demonstrated high image quality,accurately represented anatomical and pathological features,and facilitated precise dental biometrics measurement by both clinicians and downstream AI diagnostic tools.The expertise-inspired AI pipeline showcases outstanding performance in reconstructing tooth-centric radial planes and offers significant clinical utility for intelligent oral health management with high interpretability,robustness and generalization capabilities.
