J Gandhara Med Dent Sci

ORIGINAL ARTICLE

COMPARISON OF ACCURACY AND RELIABILITY OF CEPHALOMETRIC LANDMARKS AMONG

DIGITAL SOFTWARE WEB CEPH AND CARE STREAM WITH TRADITIONAL MANUAL

TRACING

Ayesh a Iftikhar

Saman Zeb

Sidra Israr

ABSTRACT

OBJECTIVES

This study aimed to compare the accuracy and reliability of AI-assisted

software (WebCeph, Carestream) with manual tracing.

METHODOLOGY

This cross-sectional (analytical) study was conducted using the records of

patients presenting to the Orthodontic Department, Rehman College of

Dentistry (RCD), Peshawar, over a period of three months. Forty-two high-

quality lateral cephalograms were traced manually for four skeletal (SNA,

SNB, ANB, FMA), three dental (UI-SN, IMPA, Overjet), and three soft tissue

parameters (UL-Eline, LL-Eline, and NLA). The same cephalograms were

later digitally analyzed using AI-assisted software (Carestream and

WebCeph). One-way ANOVA test and post hoc Tukey test were applied to

assess any signicant dierences among the groups. Bonferroni correction

was employed to adjust the signicance level (alpha) for each test and

multiple comparisons. Pearson Correlation coecients were used

to determine the correlation among the various group values.

RESULTS

Web Ceph and manual tracing revealed a signicant dierence in UI -SN,

overjet, and UL-Eline (P<0.05). Comparison between manual tracing and

Carestream showed signicant dierences in FMA, IMPA, UL -E line, and

LL-Eline (P<0.05). While comparing Web Ceph and Carestream, signicant

dierences were observed in UL-Eline, IMPA, and NLA (P<0.05). It was

found that the dierences in the mean values between AI-driven software and

manual methods were clinically insignicant. However, dierences among

AI-driven methods were clinically signicant.

CONCLUSION

AI-driven cephalometric digital tracing software, Web Ceph, and Carestream

provide suciently reliable measurements for clinical application.

Web Ceph showed better results for specic values than Carestream. The

accuracy and consistency of automated software can be improved through

manual adjustments and verication. Despite the clinical applicability,

manual tracing cannot be replaced by computerized cephalometric.

KEYWORDS: Lateral Cephalometric Analysis, Articial Intelligence,

Carestream, Web Ceph

How to cite this article

Iftikhar A, Zeb S, Israr S. Comparison

of accuracy and Reliability of

Cephalometric Landmarks Among

Digital Software Web Ceph and Care

Stream with Traditional Manual

Tracing. J Gandhara Med Dent Sci.

2026;13(1):60-66.

Date of Submission: 20-10-2025

Date Revised: 19-12-2025

Date Acceptance: 20-12-2025

Year Resident, Department of

Orthodontics, Rehman College of

Correspondence

Ayesha Iftikhar, Associate Professor,

Department of Orthodontics, Rehman

College of Dentistry

+92-300-8554719

ayesha.iftikhar@rmi.edu.pk

Dentistry

INTRODUCTION

With the introduction of cephalometric analysis by

Broadbent in the 1930s, orthodontists have gained a

valuable diagnostic tool to quantitatively provide

information, evaluate skeletal and dental relationships,

determine growth and developmental abnormalities,

and assess orthodontic treatment and its outcomes.

1,2,3

For decades, traditional manual cephalometric analysis

has been considered the gold standard. However, this

method has been reported to be laborious, time-

consuming, and prone to measurement and calculation

errors.

Recently, Articial Intelligence (AI) has

revolutionized the dental eld and has been adopted

over the years for automated cephalometric

identification and analysis. With the advent of digital

technology, a variety of tools and software have

emerged, reported to provide greater accuracy and

efciency. The literature suggests that these digital

systems oer several advantages, including reduced

subjectivity-related errors. Time consumption,

facilitating faster data analysis and acquisition, image

storage, and sharing, as well as streamlining the

formulation of treatment plans.

5,6,7

AI-driven

cephalometric tracing system uses a deep convolutional

neural network (CNN) for automatic landmark

detection. For tracing soft tissue prole, the software

utilizes classical image processing, specically

mathematical morphology. This integration of AI-

https://doi.org/10.37762/jgmds.13-1.804

January - March 2026

J Gandhara Med Dent Sci

driven algorithms and traditional image processing

enhances the accuracy, speed, and consistency of

cephalometric analyses.

8,9,10

Today, a wide range of

software options is available, ranging from computer-

assisted applications like Dolphin imaging and

Viewbox to fully automated or semi-automated AI-

powered solutions such as CephX®, CEFBOT, Web

Ceph, and Carestream software, among others.11,12

These tools are transforming clinical workows by

reducing human error, saving time, and enhancing

diagnostic consistency. Web-Ceph is a web-based,

fully automated, and capable of performing nine

dierent cephalometric analyses and interpretations. It

can store, maintain, and archive patients’ cephalograms,

photographs, and orthopantomograms. Additionally, it

includes advanced features, such as visual treatment

simulation and superimposition, which are helpful in

day-to-day orthodontic practice.

13,14

Carestream, a

dental company, has also shown keen interest in

developing automatic cephalometric landmark

identification and tracking, which is widely used in

clinical practice. Nowadays, companies oer advanced

cephalometric imaging software that can fully trace any

cephalometric radiograph taken with a Carestream

cephalostat.

15,16,17

Orthodontists must locate

cephalometric landmarks for the eective diagnostic

outcome precisely.18 Numerous papers have been

published in recent years to evaluate the accuracy of

these AI- based software programs compared to the

traditional manual and digital methods. Niwat et al.

identified that 76% of cephalometric measurements

performed automatically by the dental imaging software

(Carestream Dental, Version 6.14) showed statistically

signicant dierences compared to the manual

method.19 Conversely, a study by Pamir et al

concluded that CephX® (a web-based cephalometric

software) was as accurate as the Ceph-Ninja 3.51 app

(app-aided and Dolphin Imaging 13.01 computerized

method after manual correction of the landmarks), and

it also provided shorter analysis time.

Tsolakis et al

also found fully automated cephalometric analysis by

CS imaging (Care Stream Version 8 ) to be as reliable

and accurate as the manually traced Dolphin 3D

Imaging program (version 11.0). While slight

dierences were observed in a few measurements

(FMA, L1-MP , ANS-PNS/Go-Gn, and U1-L1), they

were not considered clinically signicant.15 Similarly,

Ravi Kumar et al also suggested Web -Ceph (a fully

automated program) to be reasonably accurate, reliable,

and valid as compared to manual tracing of

cephalometric measurements.

To the best of our

knowledge, no published articles have been found

comparing the reliability and accuracy of articial

intelligence (AI)-based cephalometric analysis

programs with the manual tracing method within

Pakistan. Therefore, our study aimed to evaluate and

compare the accuracy and reliability of cephalometric

measurements obtained from two AI-based software

programs (Carestream and WebCeph) with traditional

manually traced cephalometric analysis. The null

hypothesis was that there are no statistical dierences

among the cephalometric analysis methods in terms of

accuracy and reliability.

METHODOLOGY

This cross-sectional (analytical) study was conducted

on the records of patients presenting at the Orthodontic

department, Rehman College of Dentistry (RCD),

Peshawar, over a period of three months following the

ethical approval of the research proposal from the

review committee of Rehman College of Dentistry, EC

Ref No: RCD-005-25-202. Informed consent was

obtained from each patient prior to the collection of

their records and data. The Sample size was calculated

using the G*Power sample size calculator, based on a

confidence interval of 95% and a statistical power of

80%.

A total of 42 samples were included in the

study. The sampling method employed was non-

probability convenience sampling. Lateral

cephalograms were obtained using a CS 9000 machine,

with the following acquisition and exposure parameters:

software version, CS Imaging v8.0.3, exposure time:

0.40seconds, tube voltage: 64kV. For an

optimal outcome, X-ray images were captured at the

maximum size (18 × 24 cm), in the lateral view. The

inclusion criteria for the study were high-quality

cephalograms with no artifacts, captured with the

patient in an upright standing position in centric

occlusion and with relaxed lips. Cephalograms of

patients with positional errors, unerupted teeth that

could interfere with the anatomical landmarks, and

craniofacial deformities (e.g., cleft lip and palate) were

excluded from the study. For manual tracing, the

cephalograms were printed at a 1:1 scale and traced

on an acetate paper using a 0.3 mm hard

lead black pencil. The same cephalogram images were

imported into CS 3600 software and saved in JPEG

format for subsequent automated analysis in

Web Ceph software. Figure 1 (A&B). Later, digital

analysis was performed using Carestream and

Web Ceph software. Figure 1. The study focused on

four skeletal, three dental, and three soft tissue

parameters. Table 1: All the landmarks

were identified by the trainee under the supervision of

the supervisor. The same operator performed both

manual and digital analyses. Manual adjustments of

landmarks were performed for a few landmarks where

required for both digital software. For the

manual method, no more than 5 tracings were

Comparison of Accuracy and Reliability of Cephalometric

January - March 2026

J Gandhara Med Dent Sci

performed on a single day to minimize human fatigue

and maintain consistency. The study was completed in

3 months. To assess the accuracy of the manual

measurements, 10 to 15 randomly selected lateral

cephalograms were reassessed by the same examiner

two weeks after the initial measurements. The

reliability of the measurements was evaluated using

intraclass correlation coecients. Data were recorded

and analyzed using SPSS version 25 (IBM Corp.,

Armonk, NY, USA. The signicance level for all

statistical tests was set at P=0.05.

Figure 1(a): Showing Landmarks and Web Cep Software

Analysis

Figure 1(b): Showing Landmarks and Care Stream Software

Analysis

RESULTS

Intra-examiner reliability for measurements was high

(ICC = 0.83), indicating a strong correlation between

the two scale items and eective measurement of the

underlying construct. The Shapiro-Wilk test (p > 0.05)

supported the normality assumption, justifying the use

of parametric tests. Analysis of 42 cephalograms

compared four skeletal, three dental, and three soft-

tissue parameters using manual, digital (CS3600 and

Web Ceph software) methods (Table 1). Descriptive

statistics are shown in Table 2. Repeated-measures

ANOV A revealed signicant dierences primarily in

dental and soft-tissue parameters across intra- and inter-

group comparisons (Table 3). Signicant dierences (p

< 0.05) were found between manual and Web Ceph

measurements for UI-SN, OVERJET, IMPA, and U L-

E line, but these dierences were clinically

insignificant (Tables 2 and 3). Manual and Care Stream

measurements showed signicant dierences (p < 0.05)

in FMA, IMPA, UL-E line, LL-E line, and Overjet.

Clinically signicant dierences were observed for

IMP A (101.7 ± 7.00 vs 94.69 ± 7.49), UL-E line (-2.80

±2.1 vs 3.81 ± 2.36), and LL-E line (-1.07 ± 2.4 vs 2.91

± 2.31) (Tables 2 and 3). Furthermore, clinically

significant dierences were found between Web-Ceph

and Care Stream measurements in UL-E line (-2.86

±2.43 Vs 3.81 ± 2.36), IMPA (98.93 ± 6.77 Vs 94.69 ±

7.49), and NLA (95.7 ± 13.25 Vs 100.9 ± 11.34) (p <

0.05). Correlation analysis showed strong, positive, and

significant correlations between Care Stream and Web

Ceph variables with manual measurements, except for

Overjet (r = 0.23, p = 0.13) and LL to E-line (r = -0.26,

p = 0.10) from Care Stream, which showed weak and

insignificant correlations (Table 4).

Table 1: Cephalometric Parameters Measured

Skeletal Dental Soft tissue

SNA U1 to SN UL to E line

SNB IMPA LL to E line

ANB OVERJET NLA

FMA

Table 2: Descriptive Data of All the Variables (Skeletal, Dental,

and Soft Tissue)

Variables N (42) Mean ± Std. Deviation

Skeletal

SNA

Manual 81.76 ± 3.35

Web Ceph 81.82 ± 3.65

CS 82.55± 4.65

SNB

Manual 76.30± 3.59

Web Ceph 76.32 ± 3.91

CS 76.91 ± 4.76

ANB

Manual 5.45± 2.51

Web Ceph 5.57 ± 2.69

CS 5.57 ± 2.75

FMA

Manual 25.50 ± 6.65

Web Ceph 24.05± 6.56

27.04 ± 5.01

Dental

OVER

JET

Manual 4.50 ± 3.77

Web Ceph 6.04 ± 3.15

CS 6.63 ± 3.39

UI to SN

Manual 107.47± 9.55

Web Ceph 105.21 ± 9.68

CS 105.45 ± 10.26

IMPA

Manual 101.7 ± 7.00

Web Ceph 98.93 ± 6.77

CS 94.69 ± 7.49

Soft

tissue

UL to

Eline

Manual -2.80 ±2.1

Web Ceph -2.86 ±2.43

CS 3.81 ± 2.36

LL to

Eline

Manual -1.07 ± 2.42

Web Ceph -1.56 ± 3.05

CS 2.91 ± 2.31

NLA

Manual 98.64 ± 13.20

Web Ceph 95.7 ± 13.25

CS 100.9 ± 11.34

Comparison of Accuracy and Reliability of Cephalometric

January - March 2026

J Gandhara Med Dent Sci

Table 3: ANOVA and Post Hoc-Tukey Test Showing Inter-Group

Comparisons

Variables Groups Groups P-value

SNA

Manual Web Ceph 1.00

CS 0.212

Web Ceph CS 0.161

SNB

Manual Web Ceph 1.00

CS 0.506

Web Ceph CS 0.192

ANB

Manual Web Ceph 1.000

CS 1.000

Web Ceph CS 1.000

FMA

Manual Web Ceph 0.068

CS 0.050

Web Ceph CS 0.000

OVERJET

Manual Web Ceph 0.000

CS 0.000

Web Ceph CS 0.217

Upper incisor

to SN plane

Manual Web Ceph 0.001

CS 0.145

Web Ceph CS 1.00

IMPA

Manual Web Ceph 0.000

CS 0.001

Webceph CS 0.002

Upper lip to E

line

Manual Web Ceph 0.00

CS 0.00

Web Ceph CS 0.02

Lower lip to

E line

Manual Web Ceph 0.62

CS 0.00

Web Ceph CS 0.08

NLA

Manual Web Ceph 0.10

CS 0.730

Web Ceph CS 0.044

Table 4: Showing Correlation Among Variables

Variables

Manual Web Ceph

P-Value

Care

Stream

P-Value

SNA 01

0.924

0.000

0.840

0.000

SNB 01

0.824

0.000

0.819

0.000

ANB 01

0.901

0.000

0.919

0.000

FMA 01

0.802

0.000

0.744

0.000

OVERJET 01

0.923

0.000

0.237

0.135

UI TO SN 01

0.907

0.000

0.824

0.000

IMPA 01

0.771

0.000

0.396

0.010

UL TO E LINE 01

0.827

0.000

-0.776

0.000

LL TO E LINE 01

0.619

0.000

-0.260

0.101

NLA 01

0.760

0.000

0.671

0.000

DISCUSSION

This study compared the accuracy and reliability of

manual cephalometric analysis with AI-driven software

(Web Ceph and Carestream). While statistically

signicant di erences were observed in some variable

measurements (FMA, UI-SN, Overjet, UL-E line, LL-E

line, NLA, IMPA) between the AI-based software

groups, though clinically insignicant, the dierences

were clinically signicant among the digital software

variables. However, strong positive correlations were

found among all variables from Carestream and Web

Ceph, as well as the manual cephalometric variables,

supporting the practical reliability and validity of these

automated methods in research and clinical settings.

Minor digital-to-manual dierences, often within

decimal points or a few degrees, further support the

clinical acceptability of AI-based programs. The

existing literature compares manual cephalometric

analysis with digital automated and semi-automated

software such as WebCeph, Carestream, AutoCAD,

OneCeph, and ViewBox. Software selection depends on

user familiarity, interface capabilities, and specic

features. The shift to digital software allows for

automated landmark identication, minimizing bias and

improving eciency and reproducibility, with the

potential for 3D analyses. Despite these benets, some

meta-analyses suggest that while digital tracing oers

reliable and ecient measurements, semi -automated

methods that allow landmark modications may be

preferable.

2,5,15

It can be concluded that automated

cephalometric measurements obtained from the

WebCeph program are reasonably accurate compared to

manual tracing. Likewise, a study by Ravi Kumar et al.

reported a strong correlation (r=0.75)

and good agreement between the two methods across

the parameters. Although it is to be noted that minor

dierences observed for these variables might be

attributed to an automated landmark identication

issue, despite manual adjustments.

According to the

current literature, a 2 mm variation is generally

clinically adequate, accurate, and, therefore, acceptable

and within the range of diagnostic accuracy.

Furthermore, each cephalometric parameter has a mean

norm value and a standard deviation value for every

measurement that acknowledges some degree of

variability acceptance. The literature reports that AI-

driven, fully automated landmark detection can be

misleading at times. Like, in our study, adjustments

were needed in the Web Ceph for specic landmarks,

i.e., the incisal edge and nasion point. Yassir A et al.

compared Web Ceph and computer-based AutoCAD

software; dierences were found among the soft-tissue

variables, especially the nasolabial angle (NLA).

The

dierence could be attributed to poor landmark

identification, mainly in soft tissue tracing that is

inherent to the automatic Web Ceph. This issue can be

mitigated by using semi-automatic software options that

utilize manual corrections. Similarly, a study by

Gokhan Serhat et al. evaluating Dolphin, orthodox, and

Web Ceph software found good, signicant consistency

in angular measurements, but linear and soft-tissue

Comparison of Accuracy and Reliability of Cephalometric

January - March 2026

J Gandhara Med Dent Sci

parameters showed low consistency. These ndings

align with the ndings of Khattri’s study,

comparing Web Ceph with V15.0 FACAD Manual

tracing, suggesting that fully automated versions are

not yet ready to replace semi-automated computer-

aided methods and reinforcing the importance of semi-

automated versions.

In our study, a comparison

between Manual tracing and Carestream (CS) software

measurements revealed that the mean values of only

IMP A, UL-E line, and LL-E line showed

a clinically signicant dierence despite manual

correction of the landmarks in the CS software. It can

be attributed to mathematical or directional errors, as

CS software does not indicate to the operator whether

the values are positive (+ive) or negative (-ive), which

can aect Interpretation. Some adjustments were

needed for Nasion, point A, point B, the Sella point,

and the incisal edge to improve accuracy. Ioannis A.

Tsolakis et al. compared a digital manual method

using (Dolphin 3D Imaging program version

11.0) with automated, CS imaging V8 software and

reported an excellent intra-class reliability (>0.97) and a

strong agreement between the two approaches, showing

statistically signicant dierences in the mean values,

FMA, L1-MP , ANS-PNS/GoGn, and U1-L1 (p <

0.0027) were observed,

that are consistent with the

findings of our study. Similar to the results of our study,

many other studies have reported signicant With the

introduction of cephalometric analysis by Broadbent in

the 1930s, orthodontists have gained a valuable

diagnostic tool to quantitatively provide information,

evaluate skeletal and dental relationships, determine

growth and developmental abnormalities, and assess

orthodontic treatment and its outcomes.

1,2,3

For decades,

traditional manual cephalometric analysis has been

considered the gold standard. However, this method has

been reported to be laborious, time-consuming, and

prone to measurement and calculation errors.

Recently,

Artificial Intelligence (AI) has revolutionized the dental

field and has been adopted over the years for automated

cephalometric identication and analysis. With the

advent of digital technology, a variety of tools and

software have emerged, reported to provide greater

accuracy and eciency. The literature suggests that

these digital systems oer several advantages,

including reduced subjectivity-related errors. Time

consumption, facilitating faster data analysis and

acquisition, image storage, and sharing, as well as

streamlining the formulation of treatment plans.

5,6,7

AI-

driven cephalometric tracing system uses a deep

convolutional neural network (CNN) for automatic

landmark detection. For tracing soft tissue prole, the

software utilizes classical image processing,

specifically mathematical morphology. This integration

of AI-driven algorithms and traditional image

processing enhances the accuracy, speed, and

consistency of cephalometric analyses.

8,9,10

Today, a

wide range of software options is available, ranging

from computer-assisted applications like Dolphin

imaging and Viewbox to fully automated or semi-

automated AI-powered solutions such as CephX®,

CEFBOT, Web Ceph, and Carestream software, among

others.

11,12

These tools are transforming clinical

workflows by reducing human error, saving time, and

enhancing diagnostic consistency. Web-Ceph is a web-

based, fully automated, and capable of performing nine

dierent cephalometric analyses and interpretations. It

can store, maintain, and archive patients' cephalograms,

photographs, and orthopantomograms. Additionally, it

includes advanced features, such as visual treatment

simulation and superimposition, which are helpful in

day-to-day orthodontic practice.

13,14

Carestream, a

dental company, has also shown keen interest in

developing automatic cephalometric landmark

identification and tracking, which is widely used in

clinical practice. Nowadays, companies oer advanced

cephalometric imaging software that can fully trace any

cephalometric radiograph taken with a Carestream

cephalostat.

15,16,17

Orthodontists must locate

cephalometric landmarks for the eective diagnostic

outcome precisely.

Numerous papers have been

published in recent years to evaluate the accuracy of

these AI- based software programs compared to the

traditional manual and digital methods. Niwat et al.

identified that 76% of cephalometric measurements

performed automatically by the dental imaging software

(Carestream Dental, Version 6.14) showed stat istically

significant dierences compared to the manual

method.19 Conversely, a study by Pamir et al

concluded that CephX® (a web-based cephalometric

software) was as accurate as the Ceph-Ninja 3.51 app

(app-aided and Dolphin Imaging 13.01 computerized

method after manual correction of the landmarks), and

it also provided shorter analysis time.20 Tsolakis et al

also found fully automated cephalometric analysis by

CS imaging (Care Stream Version 8 ) to be as reliable

and accurate as the manually traced Dolphin 3D

Imaging program (version 11.0). While slight

dierences were observed in a few measurements

(FMA, L1-MP , ANS-PNS/Go-Gn, and U1-L1), they

were not considered clinically signicant.15 Similarly,

Ravi Kumar et al also suggested Web -Ceph (a fully

automated program) to be reasonably accurate, reliable,

and valid as compared to manual tracing of

cephalometric measurements.

To the best of our

knowledge, no published articles have been found

comparing the reliability and accuracy of articial

intelligence (AI)-based cephalometric analysis

programs with the manual tracing method within

Pakistan. Therefore, our study aimed to evaluate and

Comparison of Accuracy and Reliability of Cephalometric

January - March 2026

J Gandhara Med Dent Sci

compare the accuracy and reliability of cephalometric

measurements obtained from two AI-based software

programs (Carestream and WebCeph) with traditional

manually traced cephalometric analysis. The null

hypothesis was that there are no statistical dierences

among the cephalometric analysis methods in terms of

accuracy and reliability. tracing. CephX analysis

combined with manual correction was found to be

promising for clinical use, as its accuracy is comparable

to CephNinja and Dolphin. In addition, the analysis

time was found to be signicantly shorter

for Ceph Ninja and Dolphin compared to manual

tracing.

In comparison with previous studies, both the

digital Web Ceph and Carestream (CS) software were

found to be accurate and reliable, although there were

significant dierences among a few variables, mainly

dental and soft tissue measurements. The literature

suggests that it is due to faulty landmark identication

by automated software programs that require manual

adjustments. Our study conrmed that

Web Ceph required less landmark adjustment than

the CS program. Besides quick cephalometric analyses

and Interpretation, features such as cloud-based storage

of patients' records, visual treatment simulation, and

superimposition can make "WebCeph"™ an ecient

and promising tool for routine clinical orthodontic

practice. AI-based methods show considerable

potential; they have been recommended for use as a

supportive rather than denitive diagnostic tool. Even

with automated methods and approaches, the operator

needs to review, verify, and adjust the software’s

landmark positions as required before nalizing the

cephalometric analysis to ensure accuracy.

LIMITATIONS

Future research can benet from a larger sample size

with additional parameters, involvement of multiple

investigators for the reduction of bias, and determining

inter-examiner reliability, and to further validate the

accuracy and reliability of AI-based software. Time

efciency was a key nding of our study: AI -based

digital software proved to be faster than manual tracing.

Future studies should compare time requirements for

dierent methods. Furthermore, the need for a stable

internet connection may limit the use of software in

specific clinical settings. Though it is pertinent to

identify that this software needs manual correction of

specific landmarks, and orthodontists should ideally

supervise that for accuracy.

CONCLUSIONS

Cephalometric measurements performed with Web

Ceph and Carestream demonstrated high reliability,

making them a viable option for integration into the

clinical routine; however, they should be used with

proper supervision and manual corrections as needed.

Web-Ceph was found to be better than Carestream

software, as Carestream software required more manual

corrections of the landmarks and did not indicate the

operator’s positive (+) or negative (–) values,

potentially aecting the results and Interpretation.

CONFLICT OF INTEREST: None

FUNDING SOURCES: None

REFERENCES

1. Nielsen IL. Cephalometric analysis: history and clinical

application. Taiwan J Orthod. 2022 Dec 27;34(4):1–10.

2. Schwendicke F, Chaurasia A, Arsiwala L, Lee JH, Elhennawy

K, Jost-Brinkmann PG, et al. Deep learning for cephalometric

landmark detection: systematic review and meta-analysis. Clin

Oral Investig.2021 Jul;25(7):4299-309.https://doi.org/10.1007/s

00784-021-03997-4.PMID:33886143

3. Sennimalai K, Selvaraj M, Kharbanda OP, Kandasamy D,

Mohaideen K. MRI-based cephalometrics: a scoping review of

current insights and future perspectives. Dentomaxillofac

Radiol.2023 Jul;52(5):20230106.https://doi.org/10.1259/dmfr.20

230106. PMID: 37365724

4. Alsubai S. A critical review on the 3D cephalometric analysis

using machine learning. Computers (Basel). 2022 Oct

28;11(11):154. https://doi.org/10.3390/computers11110154.

5. Narkhede S, Rao P, Sawant V, Sachdev SS, Arora S, Pawar

AM, et al. Digital versus manual tracing in cephalometric

analysis: a systematic review and meta-analysis. J Pers Med.

2024 May 25;14(6):566. https://doi.org/10.3390/jpm14060566.

PMID: 38892926

6. AlBarakati S, Kula K, Ghoneima A. The reliability and

reproducibility of cephalometric measurements: a comparison

of conventional and digital methods. Dentomaxillofac Radiol.

2012 Jan;41(1):11-7.https://doi.org/10.1259/dmfr/81804525.

PMID: 22116109

7. Subramanian AK, Chen Y, Almalki A, Sivamurthy G, Kafle D.

Cephalometric analysis in orthodontics using articial

intelligence-A comprehensive review. Biomed Res Int.

2022;2022:1-10.https://doi.org/10.1155/2022/9075785.PMID:35

223134

8. Gil SM, Kim I, Cho JH, Hong M, Kim M, Kim SJ, et al.

Accuracy of auto-identification of posteroanterior cephalometric

landmarks using a cascade convolutional neural network

algorithm. Am J Orthod Dentofacial Orthop.2022Apr;161(4):e3

61-71.https://doi.org/10.1016/j.ajodo.2021.08.026. PMID: 35364

377

9. Gurbade VV, Pacharaaney U, Puri C. Advancements in AI-

driven cephalometric analysis for orthodontic applications: a

review on automated landmark detection and angle assessment.

In: 2025 4th International Conference on Sentiment Analysis

and Deep Learning (ICSADL). IEEE;2025.p.1731-6.https://doi.

org/10.1109/ICSADL58257.2025.10496547.

10. Bulatova G, Kusnoto B, Grace V, Tsay TP, Avenetti DM,

Sanchez FJC. Assessment of automatic cephalometric landmark

identication using articial intelligence. Orthod Craniofac Res.

2021Dec;24(S2):37-42.https://doi.org/10.1111/ocr.12529.PMID:

34927440

Comparison of Accuracy and Reliability of Cephalometric

January - March 2026

J Gandhara Med Dent Sci

LICENSE: JGMDS publishes its articles under a Creative Commons Attribution Non-Commercial Share-Alike license (CC-BY-NC-SA 4.0).

COPYRIGHTS: Authors retain the rights without any restrictions to freely download, print, share and disseminate the article for any lawful purpose.

It includes scholarlynetworks such as Research Gate, Google Scholar, LinkedIn, Academia.edu, Twitter, and other academic or professional networking sites.

11. Foroozandeh M, Salemi F, Shokri A, Farhadian N, Aeini N,

Hassanzadeh R. Comparative accuracy of AI-based AudaxCeph

software, Dolphin software, and the manual technique for

orthodontic landmark identication. Imaging Sci Dent.

2024;55(1):11–8. https://doi.org/10.5624/isd.20240004.

12. Maniyar N, Prakash AT, Kumar HCK, Yavagal PC. Accuracy

of Dolphin Imaging software in predicting soft tissue changes of

Class II Division 1 cases treated by all rst premolar extraction.

J World Fed Orthod.2024 June;13(3):128-35.https://doi.org/10.1

016/j.ejwf.2024.02.003.

13. Raby I, Rojas V, Celis A, García-Duhalde C, Martinac M.

Accuracy and reliability of two articial intelligence platforms

for cephalometric analysis compared with a semi-automatic

computer program. Am J Orthod Dentofacial Orthop. 2025

Oct;168(4):505–14. https://doi.org/10.1016/j.ajodo.2025.03.019.

14. Kazimierczak W, Gawin G, Janiszewska-Olszowska J,

Dyszkiewicz-Konwińska M, Nowicki P, Kazimierczak N, et al.

Comparison of three commercially available AI-driven

cephalometric analysis tools in orthodontics. J Clin Med. 2024

Jun 26;13(13):3733. https://doi.org/10.3390/jcm13133733.

15. Tsolakis IA, Tsolakis AI, Elshebiny T, Matthaios S, Palomo

JM. Comparing a fully automated cephalometric tracing method

to a manual tracing method for orthodontic diagnosis. J Clin

Med.2022 Nov 20;11(22):6854.https://doi.org/10.3390/jcm1122

6854.PMID:36430506

16. Debelmas A, Keto S, Lanciaux S, Corre P, Friess M, Khonsari

RH. Reproducibility assessment of Delaire cephalometric

analysis using reconstructions from computed tomography. J

Stomatol Oral Maxillofac Surg. 2020 Feb;121(1):35–9.

https://doi.org/10.1016/j.jormas.2019.08.006. PMID: 31611054

17. Tanellari O, Toti C, Baruti Papa E, Ghanim S, Savin C,

Romanec C, et al. The link between obstructive sleep apnea

syndrome and cephalometric assessment of upper airways and

hyoid bone position. Medicina (Kaunas). 2022 Sep;58(9):1213.

https://doi.org/10.3390/medicina58091213. PMID: 36141968

18. Polizzi A, Leonardi R. Automatic cephalometric landmark

identication with articial intelligence: an umbrella review of

systematic reviews. J Dent 2024 Jul;146:105056.https://doi.org/1

0.1016/j.jdent.2024.105056.

19. Anuwongnukroh N, Dechkunakorn S, Damrongsri S, Nilwarat

C, Pudpong N, Radomsutthisarn W, et al. Assessment of the

reliability of automatic cephalometric analysis software. Int J

Mech Eng Robot Res. 2017;6(6):61–5.

20. Meric P, Naoumova J. Web-based fully automated cephalometric

analysis: comparisons between app-aided, computerized, and

manual tracings. Turk J Orthod.2020 Sep;33(3):142-9.https://doi

.org/10.5152/TurkJOrthod.2020.19072.

21. Mahto RK, Kae D, Giri A, Luintel S, Karki A. Evaluation of

fully automated cephalometric measurements obtained from

web-based articial intelligence–driven platform. BMC Oral

Health.2022 Dec;22(1):132.https://doi.org/10.1186/s12903-022-

02175-2.PMID:36531291

22. Yassir YA, Salman AR, Nabbat SA. The accuracy and

reliability of WebCeph for cephalometric analysis. J Taibah

Univ Med Sci.2022 Feb;17(1):57-66.https://doi.org/10.1016/j.jt

umed.2021.10.010.

23. Duran GS, Gökmen Ş, Topsakal KG, Görgülü S. Evaluation of

the accuracy of fully automatic cephalometric analysis software

with an articial intelligence algorithm. Orthod Craniofac Res.

2023 Aug;26(3):481-90.https://doi.org/10.1111/ocr.12644.PMID

:37409109

24. Khattri N, Nasim NN, S V, N T, J N. Comparison of reliability

and accuracy of AI-based fully automated cephalometric tracing

software with computer-aided semi-automatic cephalometric

software. J Dent Med Sci. 2023;22(1):26–33.

AUTHORS CONTRIBUTION

The authors accept responsibility for all aspects of the work

and will ensure that any concerns regarding the accuracy or

integrity of any part are properly investigated and resolved.

Ayesha Iftikhar - Concept & Design; Data Acquisition;

Data Analysis/Interpretation; Drafting Manuscript;

Critical Revision; Supervision; Final Approval

Saman Zeb - Concept & Design; Data Acquisition; Data

Analysis/Interpretation; Drafting Manuscript; Critical

Revision; Final Approval

Sidra Israr - Concept & Design; Data Acquisition; Data

Analysis/Interpretation; Drafting Manuscript; Final

Approval

Comparison of Accuracy and Reliability of Cephalometric

January - March 2026