Orthodontic treatment outcome predictive performance differences between artificial intelligence and conventional methods.

OBJECTIVES: To evaluate an artificial intelligence (AI) model in predicting soft tissue and alveolar bone changes following orthodontic treatment and compare the predictive performance of the AI model with conventional prediction models. MATERIALS AND METHODS: A total of 1774 lateral cephalograms of 887 adult patients who had undergone orthodontic treatment were collected. Patients who had orthognathic surgery were excluded. On each cephalogram, 78 landmarks were detected using PIPNet-based AI. Prediction models consisted of 132 predictor variables and 88 outcome variables. Predictor variables were demographics (age, sex), clinical (treatment time, premolar extraction), and Cartesian coordinates of the 64 anatomic landmarks. Outcome variables were Cartesian coordinates of the 22 soft tissue and 22 hard tissue landmarks after orthodontic treatment. The AI prediction model was based on the TabNet deep neural network. Two conventional statistical methods, multivariate multiple linear regression (MMLR) and partial least squares regression (PLSR), were each implemented for comparison. Prediction accuracy among the methods was compared. RESULTS: Overall, MMLR demonstrated the most accurate results, while AI was least accurate. AI showed superior predictions in only 5 of the 44 anatomic landmarks, all of which were soft tissue landmarks inferior to menton to the terminal point of the neck. CONCLUSIONS: When predicting changes following orthodontic treatment, AI was not as effective as conventional statistical methods. However, AI had an outstanding advantage in predicting soft tissue landmarks with substantial variability. Overall, results may indicate the need for a hybrid prediction model that combines conventional and AI methods.

Does artificial intelligence predict orthognathic surgical outcomes better than conventional linear regression methods?

OBJECTIVES: To evaluate the performance of an artificial intelligence (AI) model in predicting orthognathic surgical outcomes compared to conventional prediction methods. MATERIALS AND METHODS: Preoperative and posttreatment lateral cephalograms from 705 patients who underwent combined surgical-orthodontic treatment were collected. Predictors included 254 input variables, including preoperative skeletal and soft-tissue characteristics, as well as the extent of orthognathic surgical repositioning. Outcomes were 64 Cartesian coordinate variables of 32 soft-tissue landmarks after surgery. Conventional prediction models were built applying two linear regression methods: multivariate multiple linear regression (MLR) and multivariate partial least squares algorithm (PLS). The AI-based prediction model was based on the TabNet deep neural network. The prediction accuracy was compared, and the influencing factors were analyzed. RESULTS: In general, MLR demonstrated the poorest predictive performance. Among 32 soft-tissue landmarks, PLS showed more accurate prediction results in 16 soft-tissue landmarks above the upper lip, whereas AI outperformed in six landmarks located in the lower border of the mandible and neck area. The remaining 10 landmarks presented no significant difference between AI and PLS prediction models. CONCLUSIONS: AI predictions did not always outperform conventional methods. A combination of both methods may be more effective in predicting orthognathic surgical outcomes.

Evaluation of automated photograph-cephalogram image integration using artificial intelligence models.

OBJECTIVES: To develop and evaluate an automated method for combining a digital photograph with a lateral cephalogram. MATERIALS AND METHODS: A total of 985 digital photographs were collected and soft tissue landmarks were manually detected. Then 2500 lateral cephalograms were collected, and corresponding soft tissue landmarks were manually detected. Using the images and landmark identification information, two different artificial intelligence (AI) models-one for detecting soft tissue on photographs and the other for identifying soft tissue on cephalograms-were developed using different deep-learning algorithms. The digital photographs were rotated, scaled, and shifted to minimize the squared sum of distances between the soft tissue landmarks identified by the two different AI models. As a validation process, eight soft tissue landmarks were selected on digital photographs and lateral cephalometric radiographs from 100 additionally collected validation subjects. Paired t-tests were used to compare the accuracy of measures obtained between the automated and manual image integration methods. RESULTS: The validation results showed statistically significant differences between the automated and manual methods on the upper lip and soft tissue B point. Otherwise, no statistically significant difference was found. CONCLUSIONS: Automated photograph-cephalogram image integration using AI models seemed to be as reliable as manual superimposition procedures.

Factors influencing the development of artificial intelligence in orthodontics.

OBJECTIVES: Since developing AI procedures demands significant computing resources and time, the implementation of a careful experimental design is essential. The purpose of this study was to investigate factors influencing the development of AI in orthodontics. MATERIALS AND METHODS: A total of 162 AI models were developed, with various combinations of sample sizes (170, 340, 679), input variables (40, 80, 160), output variables (38, 76, 154), training sessions (100, 500, 1000), and computer specifications (new vs. old). The TabNet deep-learning algorithm was used to develop these AI models, and leave-one-out cross-validation was applied in training. The goodness-of-fit of the regression models was compared using the adjusted coefficient of determination values, and the best-fit model was selected accordingly. Multiple linear regression analyses were employed to investigate the relationship between the influencing factors. RESULTS: Increasing the number of training sessions enhanced the effectiveness of the AI models. The best-fit regression model for predicting the computational time of AI, which included logarithmic transformation of time, sample size, and training session variables, demonstrated an adjusted coefficient of determination of 0.99. CONCLUSION: The study results show that estimating the time required for AI development may be possible using logarithmic transformations of time, sample size, and training session variables, followed by applying coefficients estimated through several pilot studies with reduced sample sizes and reduced training sessions.

Comparison of individualized facial growth prediction models based on the partial least squares and artificial intelligence.

OBJECTIVES: To compare facial growth prediction models based on the partial least squares and artificial intelligence (AI). MATERIALS AND METHODS: Serial longitudinal lateral cephalograms from 410 patients who had not undergone orthodontic treatment but had taken serial cephalograms were collected from January 2002 to December 2022. On every image, 46 skeletal and 32 soft-tissue landmarks were identified manually. Growth prediction models were constructed using multivariate partial least squares regression (PLS) and a deep learning method based on the TabNet deep neural network incorporating 161 predictor, and 156 response, variables. The prediction accuracy between the two methods was compared. RESULTS: On average, AI showed less prediction error by 2.11 mm than PLS. Among the 78 landmarks, AI was more accurate in 63 landmarks, whereas PLS was more accurate in nine landmarks, including cranial base landmarks. The remaining six landmarks showed no statistical difference between the two methods. Overall, soft-tissue landmarks, landmarks in the mandible, and growth in the vertical direction showed greater prediction errors than hard-tissue landmarks, landmarks in the maxilla, and growth changes in the horizontal direction, respectively. CONCLUSIONS: PLS and AI methods seemed to be valuable tools for predicting growth. PLS accurately predicted landmarks with low variability in the cranial base. In general, however, AI outperformed, particularly for those landmarks in the maxilla and mandible. Applying AI for growth prediction might be more advantageous when uncertainty is considerable.

Evaluation of an individualized facial growth prediction model based on the multivariate partial least squares method.

OBJECTIVES: To develop a facial growth prediction model incorporating individual skeletal and soft tissue characteristics. MATERIALS AND METHODS: Serial longitudinal lateral cephalograms were collected from 303 children (166 girls and 137 boys), who had never undergone orthodontic treatment. A growth prediction model was devised by applying the multivariate partial least squares (PLS) algorithm, with 161 predictor variables. Response variables comprised 78 lateral cephalogram landmarks. Multiple linear regression analysis was performed to investigate factors influencing growth prediction errors. RESULTS: Using the leave-one-out cross-validation method, a PLS model with 30 components was developed. Younger age at prediction resulted in greater prediction error (0.03 mm/y). Further, prediction error increased in proportion to the growth prediction interval (0.24 mm/y). Girls, subjects with Class II malocclusion, growth in the vertical direction, skeletal landmarks, and landmarks on the maxilla were associated with more accurate prediction results than boys, subjects with Class I or III malocclusion, growth in the anteroposterior direction, soft tissue landmarks, and landmarks on the mandible, respectively. CONCLUSIONS: The prediction error of the prediction model was proportional to the remaining growth potential. PLS growth prediction seems to be a versatile approach that can incorporate large numbers of predictor variables to predict numerous landmarks for an individual subject.

Evaluation of an automated superimposition method based on multiple landmarks for growing patients.

OBJECTIVES: To determine if an automated superimposition method using six landmarks (Sella, Nasion, Porion, Orbitale, Basion, and Pterygoid) would be more suitable than the traditional Sella-Nasion (SN) method to evaluate growth changes. MATERIALS AND METHODS: Serial lateral cephalograms at an average interval of 2.7 years were taken on 268 growing children who had not undergone orthodontic treatment. The T1 and T2 lateral images were manually traced. Three different superimposition methods: Björk's structural method, conventional SN, and the multiple landmark (ML) superimposition methods were applied. Bjork's structural method was used as the gold standard. Comparisons among the superimposition methods were carried out by measuring the linear distances between Anterior Nasal Spine, point A, point B, and Pogonion using each superimposition method. Multiple linear regression analysis was performed to identify factors that could affect the accuracy of the superimpositions. RESULTS: The ML superimposition method demonstrated smaller differences from Björk's method than the conventional SN method did. Greater differences among the cephalometric landmarks tested resulted when: the designated point was farther from the cranial base, the T1 age was older, and the more time elapsed between T1 and T2. CONCLUSIONS: From the results of this study in growing patients, the ML superimposition method seems to be more similar to Björk's structural method than the SN superimposition method. A major advantage of the ML method is likely to be that it can be applied automatically and may be just as reliable as manual superimposition methods.

Virtually-Planned Orthognathic Surgery Achieves an Accurate Condylar Position.

PURPOSE: Accuracy in orthognathic surgery with virtual planning has been reported, but detailed analysis of accuracy according to anatomic location, including the mandibular condyle, is insufficient. The purpose of this study was to compare the virtual plan and surgical outcomes and analyze the degree and distribution of errors according to each anatomic location. PATIENTS AND METHODS: This retrospective cohort study evaluated skeletal class III patients, treated with bimaxillary surgery. The primary predictor was anatomic locations that consisted of right and left condyles, maxilla, and the distal segment of the mandible. Other variables were age and gender. The primary outcome was surgical accuracy, defined as mean 3-dimensional distance error, mean absolute error, and mean error along the horizontal, vertical, and anteroposterior axes between the virtual plan and surgical outcomes. Landmarks were compared using a computational method based on affine transformation with a 1-time landmark setting. The mean errors were visualized with multidimensional scattergrams. Bivariate and regression statistics were computed. RESULTS: This study included 52 patients, 26 men and 26 women, with a mean age of 21 years and 3 months. The mean 3D distance errors for condylar landmarks, maxillary landmarks, and landmarks on the distal segment of the mandible were 1.03, 1.25, and 2.24 mm, respectively. Condylar landmarks, maxillary landmarks, and the landmarks on the distal segment of the mandible were positioned at 0.49 mm inferior, 0.28 mm anterior, and 1.25 mm inferior, respectively. The landmark errors for the distal segment of the mandible exhibited a wider distribution than those for condylar and maxillary landmarks. CONCLUSIONS: Agreement between the planned and actual outcome aided by virtual surgical planning was highest for the condyles, followed by the maxilla, and the distal segment of the mandible. It is important to consider the tendency for surgical errors in each anatomic location during operations.

Evaluation of automated cephalometric analysis based on the latest deep learning method.

OBJECTIVES: To compare an automated cephalometric analysis based on the latest deep learning method of automatically identifying cephalometric landmarks (AI) with previously published AI according to the test style of the worldwide AI challenges at the International Symposium on Biomedical Imaging conferences held by the Institute of Electrical and Electronics Engineers (IEEE ISBI). MATERIALS AND METHODS: This latest AI was developed by using a total of 1983 cephalograms as training data. In the training procedures, a modification of a contemporary deep learning method, YOLO version 3 algorithm, was applied. Test data consisted of 200 cephalograms. To follow the same test style of the AI challenges at IEEE ISBI, a human examiner manually identified the IEEE ISBI-designated 19 cephalometric landmarks, both in training and test data sets, which were used as references for comparison. Then, the latest AI and another human examiner independently detected the same landmarks in the test data set. The test results were compared by the measures that appeared at IEEE ISBI: the success detection rate (SDR) and the success classification rates (SCR). RESULTS: SDR of the latest AI in the 2-mm range was 75.5% and SCR was 81.5%. These were greater than any other previous AIs. Compared to the human examiners, AI showed a superior success classification rate in some cephalometric analysis measures. CONCLUSIONS: This latest AI seems to have superior performance compared to previous AI methods. It also seems to demonstrate cephalometric analysis comparable to human examiners.

Evaluation of an automated superimposition method for computer-aided cephalometrics.

OBJECTIVES: To evaluate a new superimposition method compatible with computer-aided cephalometrics and to compare superimposition error to that of the conventional Sella-Nasion (SN) superimposition method. MATERIALS AND METHODS: A total of 283 lateral cephalometric radiographs were collected and cephalometric landmark identification was performed twice by the same examiner at a 3-month interval. The second tracing was superimposed on the first tracing by both the SN superimposition method and the new, proposed method. The proposed method not only relied on SN landmarks but also minimized the differences between four additional landmarks: Porion, Orbitale, Basion, and Pterygoid. The errors between the landmarks of the duplicate tracings oriented by the two superimposition methods were calculated at Anterior Nasal Spine, Point A, Point B, Pogonion, and Gonion. The paired t-test was used to find any statistical difference in the superimposition errors by the two superimposition methods and to investigate whether there existed clinically significant differences between the two methods. RESULTS: The proposed method demonstrated smaller superimposition errors than did the conventional SN superimposition method. When comparisons between the two superimposition methods were made with a 1-mm error range, there were clinically significant differences between them. CONCLUSIONS: The proposed method that was compatible with computer-aided cephalometrics might be a reliable superimposition method for superimposing serial cephalometric images.

How much deep learning is enough for automatic identification to be reliable?

OBJECTIVES: To determine the optimal quantity of learning data needed to develop artificial intelligence (AI) that can automatically identify cephalometric landmarks. MATERIALS AND METHODS: A total of 2400 cephalograms were collected, and 80 landmarks were manually identified by a human examiner. Of these, 2200 images were chosen as the learning data to train AI. The remaining 200 images were used as the test data. A total of 24 combinations of the quantity of learning data (50, 100, 200, 400, 800, 1600, and 2000) were selected by the random sampling method without replacement, and the number of detecting targets per image (19, 40, and 80) were used in the AI training procedures. The training procedures were repeated four times. A total of 96 different AIs were produced. The accuracy of each AI was evaluated in terms of radial error. RESULTS: The accuracy of AI increased linearly with the increasing number of learning data sets on a logarithmic scale. It decreased with increasing numbers of detection targets. To estimate the optimal quantity of learning data, a prediction model was built. At least 2300 sets of learning data appeared to be necessary to develop AI as accurate as human examiners. CONCLUSIONS: A considerably large quantity of learning data was necessary to develop accurate AI. The present study might provide a basis to determine how much learning data would be necessary in developing AI.

Redesign and reconstruction of a steviol-biosynthetic pathway for enhanced production of steviol in Escherichia coli.

BACKGROUND: Steviol glycosides such as stevioside have attracted the attention of the food and beverage industry. Recently, efforts were made to produce these natural sweeteners in microorganisms using metabolic engineering. Nonetheless, the steviol titer is relatively low in metabolically engineered microorganisms, and therefore a steviol-biosynthetic pathway in heterologous microorganisms needs to be metabolically optimized. The purpose of this study was to redesign and reconstruct a steviol-biosynthetic pathway via synthetic-biology approaches in order to overproduce steviol in Escherichia coli. RESULTS: A genome-engineered E. coli strain, which coexpressed 5' untranslated region (UTR)-engineered geranylgeranyl diphosphate synthase, copalyl diphosphate synthase, and kaurene synthase, produced 623.6 ± 3.0 mg/L ent-kaurene in batch fermentation. Overexpression of 5'-UTR-engineered, N-terminally modified kaurene oxidase of Arabidopsis thaliana yielded 41.4 ± 5 mg/L ent-kaurenoic acid. Enhanced ent-kaurenoic acid production (50.7 ± 9.8 mg/L) was achieved by increasing the cellular NADPH/NADP+ ratio. The expression of a fusion protein, UtrCYP714A2-AtCPR2 derived from A. thaliana, where trCYP714A2 was 5'-UTR-engineered and N-terminally modified, gave 38.4 ± 1.7 mg/L steviol in batch fermentation. CONCLUSIONS: 5'-UTR engineering, the fusion protein approach, and redox balancing improved the steviol titer in flask fermentation and bioreactor fermentation. The expression engineering of steviol-biosynthetic enzymes and the genome engineering described here can serve as the basis for producing terpenoids-including steviol glycosides and carotenoids-in microorganisms.

Automated identification of cephalometric landmarks: Part 2-Might it be better than human?

OBJECTIVES: To compare detection patterns of 80 cephalometric landmarks identified by an automated identification system (AI) based on a recently proposed deep-learning method, the You-Only-Look-Once version 3 (YOLOv3), with those identified by human examiners. MATERIALS AND METHODS: The YOLOv3 algorithm was implemented with custom modifications and trained on 1028 cephalograms. A total of 80 landmarks comprising two vertical reference points and 46 hard tissue and 32 soft tissue landmarks were identified. On the 283 test images, the same 80 landmarks were identified by AI and human examiners twice. Statistical analyses were conducted to detect whether any significant differences between AI and human examiners existed. Influence of image factors on those differences was also investigated. RESULTS: Upon repeated trials, AI always detected identical positions on each landmark, while the human intraexaminer variability of repeated manual detections demonstrated a detection error of 0.97 ± 1.03 mm. The mean detection error between AI and human was 1.46 ± 2.97 mm. The mean difference between human examiners was 1.50 ± 1.48 mm. In general, comparisons in the detection errors between AI and human examiners were less than 0.9 mm, which did not seem to be clinically significant. CONCLUSIONS: AI showed as accurate an identification of cephalometric landmarks as did human examiners. AI might be a viable option for repeatedly identifying multiple cephalometric landmarks.

Automated identification of cephalometric landmarks: Part 1-Comparisons between the latest deep-learning methods YOLOV3 and SSD.

OBJECTIVE: To compare the accuracy and computational efficiency of two of the latest deep-learning algorithms for automatic identification of cephalometric landmarks. MATERIALS AND METHODS: A total of 1028 cephalometric radiographic images were selected as learning data that trained You-Only-Look-Once version 3 (YOLOv3) and Single Shot Multibox Detector (SSD) methods. The number of target labeling was 80 landmarks. After the deep-learning process, the algorithms were tested using a new test data set composed of 283 images. Accuracy was determined by measuring the point-to-point error and success detection rate and was visualized by drawing scattergrams. The computational time of both algorithms was also recorded. RESULTS: The YOLOv3 algorithm outperformed SSD in accuracy for 38 of 80 landmarks. The other 42 of 80 landmarks did not show a statistically significant difference between YOLOv3 and SSD. Error plots of YOLOv3 showed not only a smaller error range but also a more isotropic tendency. The mean computational time spent per image was 0.05 seconds and 2.89 seconds for YOLOv3 and SSD, respectively. YOLOv3 showed approximately 5% higher accuracy compared with the top benchmarks in the literature. CONCLUSIONS: Between the two latest deep-learning methods applied, YOLOv3 seemed to be more promising as a fully automated cephalometric landmark identification system for use in clinical practice.