Landmark identification on direct digital versus film-based cephalometric radiographs: a human skull study

Automatic cephalometric landmark identification with artificial intelligence: An umbrella review of systematic reviews

OBJECTIVES

The transition from manual to automatic cephalometric landmark identification has not yet reached a consensus for clinical application in orthodontic diagnosis. The present umbrella review aimed to assess artificial intelligence (AI) performance in automatic 2D and 3D cephalometric landmark identification.

DATA

A combination of free text words and MeSH keywords pooled by boolean operators: Automa* AND cephalo* AND ("artificial intelligence" OR "machine learning" OR "deep learning" OR "learning").

SOURCES

A search strategy without a timeframe setting was conducted on PubMed, Scopus, Web of Science, Cochrane Library and LILACS.

STUDY SELECTION

The study protocol followed the PRISMA guidelines and the PICO question was formulated according to the aim of the article. The database search led to the selection of 15 articles that were assessed for eligibility in full-text. Finally, 11 systematic reviews met the inclusion criteria and were analyzed according to the risk of bias in systematic reviews (ROBIS) tool.

CONCLUSIONS

AI was not able to identify the various cephalometric landmarks with the same accuracy. Since most of the included studies' conclusions were based on a wrong 2 mm cut-off difference between the AI automatic landmark location and that allocated by human operators, future research should focus on refining the most powerful architectures to improve the clinical relevance of AI-driven automatic cephalometric analysis.

CLINICAL SIGNIFICANCE

Despite a progressively improved performance, AI has exceeded the recommended magnitude of error for most cephalometric landmarks. Moreover, AI automatic landmarking on 3D CBCT appeared to be less accurate compared to that on 2D X-rays. To date, AI-driven cephalometric landmarking still requires the final supervision of an experienced orthodontist.

A novel approach to craniofacial analysis using automated 3D landmarking of the skull

Automatic dense 3D surface registration is a powerful technique for comprehensive 3D shape analysis that has found a successful application in human craniofacial morphology research, particularly within the mandibular and cranial vault regions. However, a notable gap exists when exploring the frontal aspect of the human skull, largely due to the intricate and unique nature of its cranial anatomy. To better examine this region, this study introduces a simplified single-surface craniofacial bone mask comprising of 6707 quasi-landmarks, which can aid in the classification and quantification of variation over human facial bone surfaces. Automatic craniofacial bone phenotyping was conducted on a dataset of 31 skull scans obtained through cone-beam computed tomography (CBCT) imaging. The MeshMonk framework facilitated the non-rigid alignment of the constructed craniofacial bone mask with each individual target mesh. To gauge the accuracy and reliability of this automated process, 20 anatomical facial landmarks were manually placed three times by three independent observers on the same set of images. Intra- and inter-observer error assessments were performed using root mean square (RMS) distances, revealing consistently low scores. Subsequently, the corresponding automatic landmarks were computed and juxtaposed with the manually placed landmarks. The average Euclidean distance between these two landmark sets was 1.5 mm, while centroid sizes exhibited noteworthy similarity. Intraclass coefficients (ICC) demonstrated a high level of concordance (> 0.988), with automatic landmarking showing significantly lower errors and variation. These results underscore the utility of this newly developed single-surface craniofacial bone mask, in conjunction with the MeshMonk framework, as a highly accurate and reliable method for automated phenotyping of the facial region of human skulls from CBCT and CT imagery. This craniofacial template bone mask expansion of the MeshMonk toolbox not only enhances our capacity to study craniofacial bone variation but also holds significant potential for shedding light on the genetic, developmental, and evolutionary underpinnings of the overall human craniofacial structure.

Accuracy and tracing time of cephalometric analyses on a tablet or desktop computer : A prospective study

BACKGROUND

This prospective study aimed to evaluate the influence of the computer type (tablet or desktop) on accuracy and tracing time of cephalometric analyses.

METHODS

Dental students used a web-based application specifically developed for this purpose to perform cephalometric analyses on tablet and desktop computers. Landmark locations and timestamps were exported to measure the accuracy, successful detection rate and tracing time. Reference landmarks were established by six experienced orthodontists. Statistical analysis included reliability assessment, descriptive statistics, and linear mixed effect models.

RESULTS

Over a period of 8 semesters a total of 277 cephalometric analyses by 161 students were included. The interrater reliability of the orthodontists establishing the reference coordinates was excellent (ICC > 0.9). For the students, the mean landmark deviation was 2.05 mm and the successful detection rate for the clinically acceptable threshold of 2 mm suggested in the literature was 68.6%, with large variations among landmarks. No effect of the computer type on accuracy and tracing time of the cephalometric analyses could be found.

CONCLUSION

The use of tablet computers for cephalometric analyses can be recommended.

Automated 3D Landmarking of the Skull: A Novel Approach for Craniofacial Analysis

Automatic dense 3D surface registration is a powerful technique for comprehensive 3D shape analysis that has found a successful application in human craniofacial morphology research, particularly within the mandibular and cranial vault regions. However, a notable gap exists when exploring the frontal aspect of the human skull, largely due to the intricate and unique nature of its cranial anatomy. To better examine this region, this study introduces a simplified single-surface craniofacial bone mask comprising 9,999 quasi-landmarks, which can aid in the classification and quantification of variation over human facial bone surfaces. Automatic craniofacial bone phenotyping was conducted on a dataset of 31 skull scans obtained through cone-beam computed tomography (CBCT) imaging. The MeshMonk framework facilitated the non-rigid alignment of the constructed craniofacial bone mask with each individual target mesh. To gauge the accuracy and reliability of this automated process, 20 anatomical facial landmarks were manually placed three times by three independent observers on the same set of images. Intra- and inter-observer error assessments were performed using root mean square (RMS) distances, revealing consistently low scores. Subsequently, the corresponding automatic landmarks were computed and juxtaposed with the manually placed landmarks. The average Euclidean distance between these two landmark sets was 1.5mm, while centroid sizes exhibited noteworthy similarity. Intraclass coefficients (ICC) demonstrated a high level of concordance (>0.988), and automatic landmarking showing significantly lower errors and variation. These results underscore the utility of this newly developed single-surface craniofacial bone mask, in conjunction with the MeshMonk framework, as a highly accurate and reliable method for automated phenotyping of the facial region of human skulls from CBCT and CT imagery. This craniofacial template bone mask expansion of the MeshMonk toolbox not only enhances our capacity to study craniofacial bone variation but also holds significant potential for shedding light on the genetic, developmental, and evolutionary underpinnings of the overall human craniofacial structure.

Web-based Fully Automated Cephalometric Analysis: Comparisons between App-aided, Computerized, and Manual Tracings

Objective

To compare the accuracy of cephalometric analyses made with fully automated tracings, computerized tracing, and app-aided tracings with equivalent hand-traced measurements, and to evaluate the tracing time for each cephalometric analysis method.

Methods

Pre-treatment lateral cephalometric radiographs of 40 patients were randomly selected. Eight angular and 4 linear parameters were measured by 1 operator using 3 methods: computerized tracing with software Dolphin Imaging 13.01(Dolphin Imaging and Management Solutions, Chatsworth, Calif, USA), app-aided tracing using the CephNinja 3.51 app (Cyncronus LLC, WA, USA), and web-based fully automated tracing with CephX (ORCA Dental AI, Las Vegas, NV). Correction of CephX landmarks was also made. Manual tracings were performed by 3 operators. Remeasurement of 15 radiographs was carried out to determine the intra-examiner and inter-examiner (manual tracings) correlation coefficient (ICC). Inter-group comparisons were made with one-way analysis of variance. The Tukey test was used for post hoc testing.

Results

Overall, greater variability was found with CephX compared with the other methods. Differences in GoGn-SN (°), I-NA (°), I-NB (°), I-NA (mm), and I-NB (mm) were statistically (p<0.05) and clinically significant using CephX, whereas CephNinja and Dolphin were comparable to manual tracings. Correction of CephX landmarks gave similar results to CephNinja and Dolphin. All the ICCs exceeded 0.85, except for I-NA (°), I-NB (°), and I-NB (mm), which were traced with CephX. The shortest analyzing time was obtained with CephX.

Conclusion

Fully automatic analysis with CephX needs to be more reliable. However, CephX analysis with manual correction is promising for use in clinical practice because it is comparable to CephNinja and Dolphin, and the analyzing time is significantly shorter.

Are orthodontic landmarks and variables in digital cephalometric radiography taken in fixed and natural head positions reliable?

OBJECTIVE

The purpose of the study was to evaluate the reliability of the most common landmarks and variables in digital lateral cephalometric radiographies in fixed and natural head positions.

MATERIALS AND METHODS

Twenty-one patients with anterior or distal displacement of the mandible treated in the Postgraduate Orthodontic Clinic of the Aristotle University of Thessaloniki had a digital lateral cephalometric radiography in fixed and in natural head position. The images where digitized. The main investigator and 6 examiners, digitized 61 landmarks and analysed 34 variables. We examined the intra-observer and inter-observer variability.

RESULTS

There was a significant difference in X-axis in the distribution (hence the mean results) according to the reliability of the landmarks and in only 2 variables in digital lateral cephalograms between fixed and natural head position.

CONCLUSIONS

Cephalometric landmarks and variables showed reliability in digital lateral cephalometric radiography in fixed and natural head position. In lateral cephalograms taken in fixed head position an anterior inclination of the head was noticed compared to those in natural head position.

Comparative evaluation of cephalometric measurements of monitor-displayed images by Nemoceph software and its hard copy by manual tracing

OBJECTIVE

The aim of this study was to evaluate and compare the cephalometric measurements obtained from computerized tracing of direct digital radiographs and hand tracing of their digital radiographic printouts.

MATERIAL AND METHODS

The soft- and hard-copies of pre-treatment lateral cephalograms of 40 subjects (both males and females) within the age group of 10-30 years, irrespective of the type of malocclusion were taken. Total 26 measurements (13 linear and 13 angular) were obtained using both the manual and the digital technique.

RESULTS

Amongst the linear measurements, Anterior facial height (AFH), Posterior facial height (PFH), Upper lip length (ULL), Lower lip length (LLL), Anterior cranial base length (ACBL), Posterior cranial base length (PCBL), Maxillary length (MxL), Mandibular length (MdL), Lower incisor to NB line (L1 to NB) and Lower lip protrusion (LLP) showed statistically significant difference between the two techniques but were clinically acceptable (difference between the digital and manual technique were less than 2 units (1 unit = 1 mm for linear measurements and 1° for angular measurements). While amongst the angular measurements, only occlusal plane angle showed statistically significant difference between the two techniques that was not clinically acceptable.

CONCLUSION

Digital measurements obtained from monitor-displayed images (soft copy) were found to be reproducible and comparable to the manual method done on its hard copy, for all the measurements except occlusal plane angle (SN-occlusal plane).

Influence of a programme of professional calibration in the variability of landmark identification using cone beam computed tomography-synthesized and conventional radiographic cephalograms

OBJECTIVES

The validity of any measurement obtained through a cephalogram largely depends on the reproducibility of the cephalometric landmarks. The purpose of this study is to evaluate the influence of a programme of professional calibration (PPC) on the variability of landmark identification comparing conventional radiographs and cone beam CT (CBCT)-synthesized cephalograms.

METHODS

5 graduate students in oral radiology identified 20 cephalometric landmarks from cephalograms generated from conventional radiographs (RADs), Ray-Sum CBCT-synthesized cephalograms (CBTs) and half-skull CBT (HSTs) from 10 patients. After a period of reinforcement on instruction and calibration with inter- and intraexaminer assessment of reproducibility (intraclass coefficient correlation scores > 0.75) for RADs, CBTs and HSTs obtained from 5 different patients, observers were asked to repeat the analysis of the first 10 patients under the same circumstances. Values in millimetres represented each landmark in a table of Cartesian co-ordinates (x- and y-axes).

RESULTS

ANOVA showed significant reduction in variability levels after the PPC, and there were no differences among the methods of image acquisition. Repeated measures ANOVA indicated that the PPC accounted for reduction in variability levels in 14 of 20 landmarks.

CONCLUSIONS

The results suggest that a PPC has more influence than the type of image acquisition on variability of landmark identification based on two-dimensional cephalometric analysis. Cephalograms obtained from RAD or CBCT can be considered equivalent for clinical and experimental applications.

Quantitative evaluation of patient movement during simulated acquisition of cephalometric radiographs

The objective of this study was to perform the quantitative three-dimensional analysis of the patients' movements at the different time points during the simulated acquisition of cephalometric radiographs. Fifty-three subjects (32 men, 21 women) were divided into four groups according to their age (Group 1: 9-12, Group 2: 13-19, Group 3: 20-25, and Group 4: 26-30 years old). The experiment (Exp) consisted in providing the subjects with three different kinds of verbal instructions as follows; Exp 1: they were simply instructed not to move, Exp 2: detailed instructions were provided, and Exp 3: they were specifically instructed to clench their molars. The amount of their movement during the 20 s of the cephalomatric X-ray exposure was measured using an optical marker and tracker. The maximum movement was analyzed three-dimensionally at 0.5, 2, 5, 10, 15, and 20 s. There was minimal vibrating movement every 0.3-0.5 s and relatively large movement every 3-5 s. The youngest group showed the largest amount of movement among the four age groups, and their movement was more significant in the up and down direction (p < 0.05). There was no significant difference in the amount of movement according to the instructions. The longer exposure time showed the larger amount of the movement. Children can show the significant movements during X-ray taking, and the longer exposure time can also result in the larger movement during acquisition of cephalometric radiographs. Therefore, the shorter exposure time is recommended in order to improve image quality.

Collimator with filtration compensator: clinical adaptation to meet European Union recommendation 4F on radiological protection for dental radiography

OBJECTIVE

Our aim was to develop a compensated filtration collimator for use in paediatric patients undergoing cephalometric radiography that reduces the radiation dose administered and fulfils recommendation 4F of the European guidelines on radiation protection in dental radiology.

METHODS

An easy to use filtration-compensated collimator was constructed of plastic, lead and aluminium and used randomly with a group of 32 children (mean age 11 years) undergoing cephalometric radiography before receiving orthodontic treatment. The radiation doses administered to patients (eye lens and thyroid, submandibular and parotid glands) and to the chassis of the radiographic equipment were determined.

RESULTS

The filtration-compensated collimator is easily fixed to the external surface of the radiographic equipment and results in (a) as collimator, a reduction of 40% in the surface irradiated in the children and of 61.4% in the dose administered to the thyroid glands (P<0.001); (b) as filtration compensator, a reduction of 32.8% administered to the eye lens (P<0.001), 31.45% to the submaxillary gland (P<0.01) and 11.4% to the parotid gland (P<0.05); there was no difference in the dose determined on the radiographic film.

CONCLUSIONS

A radiographic examination can be carried out with children using only a third of the dose normally used with no increase in the time or cost involved.

An evaluation of cellular neural networks for the automatic identification of cephalometric landmarks on digital images

Several efforts have been made to completely automate cephalometric analysis by automatic landmark search. However, accuracy obtained was worse than manual identification in every study. The analogue-to-digital conversion of X-ray has been claimed to be the main problem. Therefore the aim of this investigation was to evaluate the accuracy of the Cellular Neural Networks approach for automatic location of cephalometric landmarks on softcopy of direct digital cephalometric X-rays. Forty-one, direct-digital lateral cephalometric radiographs were obtained by a Siemens Orthophos DS Ceph and were used in this study and 10 landmarks (N, A Point, Ba, Po, Pt, B Point, Pg, PM, UIE, LIE) were the object of automatic landmark identification. The mean errors and standard deviations from the best estimate of cephalometric points were calculated for each landmark. Differences in the mean errors of automatic and manual landmarking were compared with a 1-way analysis of variance. The analyses indicated that the differences were very small, and they were found at most within 0.59 mm. Furthermore, only few of these differences were statistically significant, but differences were so small to be in most instances clinically meaningless. Therefore the use of X-ray files with respect to scanned X-ray improved landmark accuracy of automatic detection. Investigations on softcopy of digital cephalometric X-rays, to search more landmarks in order to enable a complete automatic cephalometric analysis, are strongly encouraged.

Image distortion and magnification of 3 digital CCD cephalometric systems

OBJECTIVE

The magnification and distortion of images made from scanning and nonscanning CCD cephalometric systems was evaluated.

STUDY DESIGN

Acrylic box and dry human skull phantoms were imaged using a conventional cephalostat, the nonscanning Kodak 8000C, the horizontally scanning GE/Instrumentarium OC100D, and the vertically scanning Sirona OrthophosDS. True linear and angular measurements of the phantoms were made using a coordinate measuring system and the Hitachi MercuRay cone beam system. The accuracy of linear and angular measurements was assessed, as were magnification and distortion, where appropriate.

RESULTS

Statistically significant differences between linear and angular measurements were found for almost all measurements for both phantoms. In general, the conventional cephalostat and Sirona OrthophosDS systems produced the greatest magnification and distortion while the GE/Instrumentarium OC 100D and Kodak 8000C systems, the least.

CONCLUSIONS

Measurement differences related to the beam geometries of these systems could not be predicted a priori. Unaccounted for, these differences could result in clinically significant consequences.

Comparison of hand-traced and computerized cephalograms: landmark identification, measurement, and superimposition accuracy

INTRODUCTION

The purposes of this study were (1) to investigate the variations of landmark identification between film and digital cephalometric tracings, (2) to compare the ability of Quick Ceph 2000 (Quick Ceph Systems, Inc, San Diego, Calif) to measure the linear and angular measurements with the hand-traced method, and (3) to compare Quick Ceph 2000 superimpositions to the hand-traced method of superimpositions that are currently accepted by the American Board of Orthodontics (ABO).

MATERIALS

We used 30 sets of serial cephalometric radiographs of growing patients from 1 orthodontic office. Fiduciary x- and y-axes were drawn in pencil on the T1 radiographs in the regions of the cranial base, the maxilla, and the mandible. The fiduciary lines were transferred to the digital and film serial cephalograms by regionally superimposing the tracings as described in the ABO Phase III examination handbook. A Mann-Whitney test was done to compare the median and Delta of the T1 and T2 values for each measurement acquired by hand and by Quick Ceph.

RESULTS AND CONCLUSIONS

There was no difference in the identification of cephalometric landmarks made manually vs digitally with Quick Ceph 2000. There was no difference in acquiring consistent cephalometric values for the measurements required by the ABO for the Phase III clinical examination manually vs digitally by using Quick Ceph 2000. There was no difference in the regional superimpositions of the mandible, the maxilla, and the cranial base, manually vs digitally with Quick Ceph 2000.

Accuracy of digital and analogue cephalometric measurements assessed with the sandwich technique

INTRODUCTION

The purpose of the study was to evaluate the accuracy of cephalometric measurements obtained with digital tracing software compared with equivalent hand-traced measurements. In the sandwich technique, a storage phosphor plate and a conventional radiographic film are placed in the same cassette and exposed simultaneously. The method eliminates positioning errors and potential differences associated with multiple radiographic exposures that affected previous studies. It was used to ensure the equivalence of the digital images to the hard copy radiographs. Cephalometric measurements instead of landmarks were the focus of this investigation in order to acquire data with direct clinical applications.

METHODS

The sample consisted of digital and analog radiographic images from 47 patients after orthodontic treatment. Nine cephalometric landmarks were identified and 13 measurements calculated by 1 operator, both manually and with digital tracing software. Measurement error was assessed for each method by duplicating measurements of 25 randomly selected radiographs and by using Pearson's correlation coefficient. A paired t test was used to detect differences between the manual and digital methods.

RESULTS

An overall greater variability in the digital cephalometric measurements was found. Differences between the 2 methods for SNA, ANB, S-Go:N-Me, U1/L1, L1-GoGn, and N-ANS:ANS-Me were statistically significant (P < .05). However, only the U1/L1 and S-Go:N-Me measurements showed differences greater than 2 SE (P < .0001).

CONCLUSIONS

The 2 tracing methods provide similar clinical results; therefore, efficient digital cephalometric software can be reliably chosen as a routine diagnostic tool. The user-friendly sandwich technique was effective as an option for interoffice communications.