Taken from the Angle Orthodontist on CD-ROM (Copyright © 1997 Angle

Orthodontist, Inc.), 1994 No. 4, 299 - 310: Commentary: Skeletal jaw

relationships Martin Fine.

--------------------------------

COMMENTARY

Commentary: Skeletal jaw relationships

Martin Fine, BDS, MSc

Cephalometrics in orthodontic practice is an established diagnostic tool

employed by clinicians worldwide. Conventional cephalometrics has served

orthodontic research and diagnosis since its standardization in 1931.1 It is

only in recent times that conventional cephalometric analysis has become

the subject of increased scientific scrutiny.

The orthodontic literature is replete with different analyses based upon

linear, angular and/or proportional measurement systems. When applied to

cephalometrics, these systems have little rigorous theoretical backing and

are based mainly upon convention.2 In fact, in six decades of cephalometric

usage, there has been relatively little scientific progress in the measurement

of cephalometric form or in the measurement of biological form in general.

The problem areas in cephalometrics can be divided into the following:

1.) Imaging difficulties: the reduction of a complex three-dimensional

craniofacial form into a two-dimensional projection is the first in a cascade

of steps which results in the indiscriminate loss of information in

cephalometry.

2.) Datum point selection: in conventional cephalometics irregular two-

dimensional form is reduced to a handful of datum points. Limited numbers

of datum points provide only a cursory description of craniofacial form,

yielding no data concerning the curvature of boundary outlines,3 resulting in

further indiscriminate data loss.

3) Measurement difficulties: the combination of the loss of the third

dimension and further reduction of data through the use of limited datum

point arrays is compounded by their summarization through inappropriate

measurement techniques.

Linear and angular techniques or their respective ratios are inadequate for

describing cephalometric form.4 Different combinations of datum points

may produce the same angle5 or linear distance.

Also, size and shape parameters cannot be discriminated from traditional

linear and/or angular cephalometric dimensions. Thus a change in the facial

angle or distance between gonion and condylion may reflect a size or shape

change, or more likely varying combinations of size and shape changes.

Conventional cephalometric analysis generally involves a univariate

approach of comparing individual measurements with corresponding

population means. This method is more appropriate for population studies

than for individuals.6 In addition, the variable correlation between different

conventional cephalometric measures renders them unsuitable for univariate

statistical analysis.7 Multivariate techniques are better suited to

cephalometric analysis and allow comparison of an array of measurements

as a whole as opposed to discrete parts.

In addition, the use of multiple discrete measurements in conventional

cephalometrics depends on their subjective analysis. It is difficult, if not

impossible, for a clinician to recount the logical steps made in arriving at a

cephalometric diagnosis from the array of measurements which make up a

conventional analysis.8

If traditional cephalometrics is fraught with so many problems, how has it

been possible for cephalometrics to produce any useful results?

Conventional cephalometric measurements are probably correlated with

more sophisticated forms of measurement to a greater or lesser degree. For

example, a patient with a large mandible (even if differently shaped than a

“normal” mandible) is likely to show increases in most linear measurements

of the mandible. Similarly, a “long face” is usually associated with an

increased vertical dimension.

Dr. Lowe and coworkers have addressed the concerns about conventional

cephalometrics by using a measurement technique (EFF) with a rigorous

scientific basis well-suited to the task of measuring irregular biological

forms. As opposed to the Finite Element Method (FEM, a different

rigorously-based method of measuring biological form) EFF facilitates the

measurement of outline form. They then analyzed the EFF data

appropriately using multivariate statistical techniques.

The difficulty with EFF (and FEM) is that its parameters are difficult to

understand (when compared with the relatively simple conventional

cephalometric measures). For example, we can all picture how the

mandibular plane angle will change as the mandible rotates open. What will

happen to EFF parameters in this scenario? At the present time we simply do

not have enough knowledge to elucidate how EFF parameters might vary to

reflect different skeletal morphological patterns.

One could argue that multivariate analysis will take care of this uncertainty.

However, it is important that the multivariate analysis be provided with

appropriate variables that reflect the important data. For example,

measurements of cranial base form are likely to be less important in

orthodontic A-P skeletal diagnosis than those of maxillo-mandibular form.

This factor can be taken into account by the differential variable weighting,

which can reduce misclassification in Cluster analysis.9–11 In fact, the

decision to include or exclude a variable is in itself a form of weighting.

This paper has taken steps to address fundamental problems in

cephalometrics. This could lead to further research which will provide for

more formal diagnostic techniques and therefore more logical objective

treatment planning.

Taken from the Angle Orthodontist on CD-ROM (Copyright © 1997 Angle

Orthodontist, Inc.), 1994 No. 6, 447 - 454: Landmark identification error in

posterior anterior cephalometrics Paul W. Major, Donald E. Johnson, Karen

L. Hesse,...

---------------------ORIGINAL ARTICLE

Landmark identification error in posterior anterior cephalometrics

Paul W. Major DDS,MSc.,MRCD(C);

Donald E. Johnson DDS,MSc;

Karen L. Hesse BSc.,DDS;

Kenneth E. Glover DDS,MSc.,MRCD(c)

Abstract :

This study was designed to quantify the intraexaminer and interexaminer

reliability of 52 commonly used posterior anterior cephalometric landmarks.

The horizontal and vertical identification errors were determined for a

sample of 33 skulls and 25 patients. The results show that there is a

considerable range in the magnitude of error with different horizontal and

vertical values. Interexaminer landmark identification error was significantly

larger than intraexaminer error for many landmarks. The identification error

was different for the skull sample compared to the patient sample for a

number of landmarks. The relevance of knowing the identification error for

each landmark being considered in a particular application was discussed.

Key Words :

Landmark identification error · Posterior anterior cephalometrics ·

Intraexaminer reliability · Interexaminer reliability

Since the introduction of a standardized method for obtaining skull

radiographs,1 cephalometrics has become one of the major diagnostic tools

in orthodontics.

The posterior anterior cephalogram contains diagnostic information not

readily available from other sources. This information allows the practitioner

to evaluate the width and angulation of the dental arches in relation to their

osseous bases in the transverse plane; evaluate the width and transverse

positions of the maxilla and mandible; evaluate the relative vertical

dimensions of bilateral osseous and dental structures; assess nasal cavity

width; and analyze vertical and/or transverse facial asymmetries.2–7

Regardless of the clinical or research application, it is critical to know the

reliability of the reference landmarks.

Baumrind and Frantz8 point out that there are two general classes of error

associated with cephalometric measurements. The first class of errors are

“projection” errors which arise from the geometry of the radiographic setup.

The fact that the x-ray beam originates from a source which has a finite size

leads to a penumbra effect or optical blurring.9,10 The x-ray beam diverges

as it moves away from the source, which results in an overall magnification

of the object being radiographed and a radial displacement of all points

which are not on the principal axis (central ray). The radiographic image is

distorted as points closer to the film are magnified less than points farther

from the film.

The second general class of landmark errors may be termed “errors of

identification,” and arise due to uncertainty involved in locating specific

anatomic landmarks on the radiograph. The precision with which any

landmark may be identified depends on a number of factors.8,11,12

Landmarks lying on a sharp curve or at the intersection of two curves are

generally easier to identify than points located on flat or broad curves. Points

located in areas of high contrast are easier to identify than points located in

areas of low contrast. Superimposition of other structures, including soft

tissue over the area of the landmark in question, reduces the ease of

identification. Precise written definitions describing the landmark reduces

the chance of interpretation error. Operator experience is an important factor

since increased knowledge of anatomy and familiarity with the radiographic

appearance of the subject reduces interpretive errors.

A literature review concerning the reliability of landmark identification in

posterior anterior cephalometrics revealed only one article, by El-Mangoury

et al.,12 which determined the horizontal, vertical and radial variability of 13

landmarks. They found that each landmark had its own characteristic

noncircular envelope of error, and that the variability is different in the

horizontal and vertical directions. Unfortunately, the majority of posterior

anterior cephalometric analyses use landmarks whose identification error has

not been independently reported.

The purpose of this study was to examine the reliability of posterior anterior

cephalometric landmarks. Skeletal and dental landmarks to be investigated

were chosen to include those most commonly used in published posterior

anterior cephalometric analyses,13,14–19 and those landmarks which can be

recognized on the posterior anterior cephalogram.20,21,22 Landmark

reliability for cephalograms taken both on dry skulls and living patients were

identified and compared.

Materials and methods

A sample of 33 dry adult skulls from the University of Alberta collection

with intact dentitions and no gross asymmetries were radiographed with a

standardized technique. The source-to-film distance was a constant 160 cm

and the distance from the middle of the earrods to the film was 17.5 cm. A

sample of 25 adult patient posterior anterior cephalograms based on the

absence of obvious skeletal or dental asymmetries, was chosen from

consecutive orthodontic records taken at a private radiology facility. All

patient cephalograms were taken using a Siemans OP10 x-ray machine with

standardized exposure and head positioning with Frankfort Horizontal

parallel to the floor. Source-to-earrod distance was 60 inches and earrod-to-

film distance was 5 inches.

Landmarks were digitized directly off the radiographs using a GP6 Sonic

Digitizer R in conjunction with an IBM-compatible computer and a custom

program developed using Basic TM. An individual coordinate system was

established for each radiograph by including two fiducial points which

consisted of a pinhole placed on each radiograph at the superior and medial

corner of both earrod markers. These two pinholes were digitized first which

enable the digitization program to calculate the slope of the line between the

two pinholes. This value was used as the X-axis of a cartesian coordinate

system. The Y-axis was calculated as the line perpendicular to the X-axis

originating at the midpoint of the line between the two pinholes. This

coordinate system eliminated the orientation of the radiograph on the

viewbox as a variable. Fifty-two commonly used landmarks were then

digitized including 36 bilateral skeletal landmarks.

The following landmarks (Figure 1) were identified on each radiograph:

A. Bilateral skeletal landmarks

1. Greater Wing Superior Orbit (GWSO) - the intersection of the

superior border of the greater wing of the sphenoid bone and lateral orbital

margin.

2. Greater Wing Inferior Orbit (GWI0) - the intersection of the inferior

border of the greater wing of the sphenoid bone and the lateral orbital

margin.

3. Lesser Wing Orbit (LWO) - the intersection of the superior border of

the lesser wing of the sphenoid bone and medial aspect of the orbital margin.

4. Orbitale (O) - the midpoint of the inferior orbital margin.

5. Lateral Orbit (LO) - the midpoint of the lateral orbital margin.

6. Medial Orbit (MO) - the midpoint of the medial orbital margin.

7. Superior Orbit (SO) - the midpoint of the superior orbital margin.

8. Zygomatic Frontal (ZF) - the intersection of the zygomaticofrontal

suture and the lateral orbital margin.

9. Zygomatic (Z) - the most lateral aspect of the zygomatic arch.

10. Foramen Rotundum (FR) - the center of foramen rotundum.

11. Condyle Superior (CS) - the most superior aspect of the condyle.

12. Center Condyle (CC) - the center of the condylar head of the condyle.

13. Mastoid Process (MP) - the most inferior point on the mastoid

process.

14. Malar (M) - the deepest point on the curvature of the malar process of

the maxilla.

15. Nasal Cavity (NC) - the most lateral point on the nasal cavity.

16. Mandible/Occiput (MBO) - the intersection of the mandibular ramus

and the base of the occiput.

17. Gonion (G) - the midpoint on the curvature at the angle of the

mandible (gonion).

18. Antegonial (AG) - the deepest point on the curvature of the antegonial

notch.

B. Midline skeletal landmarks

1. Crista Galli (CG) - the geometric center of the crista galli.

2. Sella Turcica (ST) - the most inferior point on the floor of sella

turcica.

3. Nasal Septum (NSM) - the approximated midpoint on the nasal

septum between crista galli and the anterior nasal spine.

4. Anterior Nasal Spine (ANS) - the center of the intersection of the

nasal septum and the palate.

5. Incisor Point (IPU) - the crest of the alveolus between the maxillary

central incisors.

6. Incisor Point (IPL) - the crest of the alveolus between the mandibular

central incisors.

7. Genial Tubercles (GT) - the center of the genial tubercles of the

mandible.

8. Menton (ME) - the midpoint on the inferior border of the mental

protuberance.

C. Bilateral dental landmarks

1. Maxillary Cuspid (MX3) - the incisal tip of the maxillary cuspid.

2. Maxillary Molar (MX6) - the midpoint on the buccal surface of the

maxillary first molar.

3. Mandibular Cuspid (MD3) - the incisal tip of the mandibular cuspid.

4. Mandibular Molar (MD6) - the midpoint on the buccal surface of the

mandibular first molar.

To determine intraexaminer landmark reliability, each radiograph was

digitized five times by the principle investigator. To avoid operator bias,

radiographs were digitized randomly and no individual radiograph was

digitized more than once in a day. The raw data was examined for any single

digitization which differed from the average of the other four by greater than

10 mm. Digitization of that particular radiograph was repeated, effectively

eliminating any instances where the wrong point was digitized by mistake.

Deviations from each landmark mean value were analyzed to give the

standard deviation of the mean, which was considered to be the landmark

identification error in millimeters.

To determine interexaminer landmark reliability each radiograph was

digitized one time by each of four operators with graduate level training in

cephalometrics. Each operator was provided with written descriptions and

diagrams of the landmark location for reference during digitization

procedures. Data analysis was completed using the procedure outlined for

intraexaminer landmark reliability.

The error of the method was established by repeated digitization of a

precisely defined point which consisted of a pinhole in the radiograph.

Results

A. Reliability of the Method

Reliability is a measure of the reproducibility or, in this case, the closeness

of the recorded coordinates for each particular landmark. In estimating the

reliability of the method, four contributing factors were identified.

1. Radiograph (R) - differences in landmark position between individual

skulls or patients.

2. Position (P) - differences between positions of different landmarks

within the same skull or patient.

3. Side (S) - differences in landmark position between the left and right

sides of the skull or patient.

4. Case (C) - differences between successive digitizations of the same

radiograph.

The reliability of the method was calculated using generalizability theory

which uses an analysis of variance to separate the total variance into its

component parts. The total variance is made up of the variation due to each

factor plus the variation due to all combinations of factors. Since reliability

is a measure of how reproducible the method is in repeated trials, any

variance between successive digitizations is considered undesirable. To

calculate the general reliability of the method, variance due to case and any

other variance in combination with case were subtracted from the total

variance, then this value was divided by the total variance.

where: R = reliability; VT = total variance; Vc = variance due to case; VcRP

= variance due to case in combination with radiograph position.

Because the sample was accepted on the criterion of good facial symmetry,

the relative contribution of side as a variable was not considered in the

estimation of reliability. The very high level of reliability [Rx(skull) = .9995,

Ry(skull) = .9992, Rx(patient) = .9910, Ry(patient) = .9985] indicates that

the relative contribution of multiple digitizations to the total variance is very

low.

B. Method error

The magnitude of error associated with the equipment (SDx = .13 mm, SDy

= .10 mm) was very close to the ± .1 mm accuracy of the digitizer claimed

by the equipment manufacturer.

C. Intraexaminer landmark error

The error associated with the identification of each landmark was calculated

for both the skull and patient samples (Tables 1 and 2). There was a wide

variation in the amount of identification error between landmarks, as well as

between the vertical and horizontal error for each particular landmark.

Visual inspection of the results indicates that the identification errors for the

skull and patient radiographs were similar, with the values generally larger

for the patient radiographs where soft tissue became a factor. Landmark

identification error for the skull sample and patient sample were compared

using a Student Newman Keuls comparison of means (P<.05). Horizontal

identification error was significantly greater in the patient sample for

Landmark Mandible/Occiput (MB0). Vertical identification error was

significantly greater in the patient sample for Landmark Maxillary Cuspid

(MX3) and Crista Galli (CG). Vertical identification error was significantly

greater in the skull sample for Landmark Zygomatic Frontal (ZF) and Nasal

Septum (NSM).

D. Interexaminer landmark error

The landmark identification errors for a single examiner and four examiners

were determined for a selected sample of 20 skull and patient radiographs

(Tables 3 to 6). The results indicate that landmark identification error was

generally larger when four examiners were used, with the error for the

patient sample larger than the skull sample.

A Student-Newman-Keuls comparison of means was used to compare the

identification errors of each sample.

The results listed in Tables 3 to 6 show the comparison between groups.

Horizontal interexaminer landmark identification error was significantly

larger than the intraexaminer error for four landmarks in the skull sample,

and 10 landmarks in the patient sample. Vertical interexaminer landmark

identification error was significantly larger than the intraexaminer error for

eight landmarks in the skull sample and 17 landmarks in the patient sample.

Horizontal interexaminer landmark identification error was larger in the

patient sample compared with the skull sample for landmarks Lateral Orbit

(LO), Foramen Rotundum (FR) and Malar (M). Vertical interexaminer

landmark identification error was larger in the patient sample compared to

the skull sample for Landmarks Orbital (O), Condyle Superior (CS),

Condyle Center (CC), Zygomatic Frontal (ZF), Foramen Rotundum (FR),

Maxillary Cuspid (MX3), Crista Galli (CG) and Genial Tubercles (GT).

Discussion

There was a great deal of variability in the magnitude of horizontal and

vertical landmark identification errors. This variability existed both within

each landmark and between different landmarks. This is in agreement with

the findings of other studies into landmark identification errors.8,11,12,23–

25 The range of values (in millimeters) for intraexaminer errors (0.28–2.23)

was of similar magnitude as that reported by Vincent and West11 (0.31–

2.09) who also used five digitizations. The El Mangoury et al.12 study into

Posterior Anterior Cephalometric landmark identification error reported a

range of error of 0.42 to 1.74. Her study used patient radiographs and when

the same landmarks were examined in this study, the range of error was of

similar magnitude (0.37–1.10).

The interexaminer identification errors showed a wide variation in

magnitude in both horizontal and vertical dimensions. The range of values

(0.31–4.79) was larger than in the intraexaminer portion of the study. This

difference can be attributed to interpretive differences between operators.

The study by El Mangoury et al.12 used only one operator and did not report

interexaminer error. The Baumrind and Frantz8 study on lateral

cephalograms used multiple operators and the range of error reported in their

study was 0.34 to 3.71, which is similar to the range found in this study.

The choice of landmarks used in any analysis will depend on the objective of

the analysis. Knowledge of the landmark identification error in both the

horizontal and vertical directions is essential in establishing a valid analysis.

Landmarks with a large horizontal identification error should be avoided in

transverse measurements. Similarly, landmarks with large vertical

identification error should be avoided in measuring vertical structural

relationships. Some landmarks will be useful for measurements in one

dimension but not in the other. For example, landmark Nasal Septum (NSM)

has a relatively small horizontal error (.49 in the skull sample) and large

vertical error (2.82 in the skull sample). Caution must be exercised when

comparing data taken from skull samples to patient samples. Most

landmarks had similar identification errors but there were exceptions.

Some landmarks may be quite useful in research trials where one examiner

takes repeated measurements, but less useful for clinical diagnosis where

differences in interpretation may be large. For example, landmark

Zygomatic (Z) had a relatively small intraexaminer error in both the

horizontal (0.29) and vertical (0.51) dimensions, but large interexaminer

errors in both the horizontal (2.42) and vertical (3.49) dimensions. This

particular landmark may be very useful in research but would have limited

value as part of a clinical diagnostic analysis.

The clinical significance of the magnitude of landmark identification error

will depend on the level of accuracy required. The landmark identification

errors reported in this study represent the standard deviation of error.

Landmarks with identification errors greater than 1.5 mm should probably

be avoided and landmarks with identification error greater than 2.5 mm are

inappropriate.

The reliability of landmarks for dried skulls was compared to live patients.

In general landmarks are less reliable on patient radiographs where soft

tissue reduces hard tissue image sharpness. These differences should be kept

in mind when applying data from dry skull studies to clinical settings.

The basis of cephalometrics in orthodontic diagnosis includes the use of

standardized and reproducible head position in relation to the x-ray source

and film. The cephalostat earrods minimize rotation about the vertical and

transverse axis. A third reference may be positioned against the nose to

prevent rotation about the anterior posterior axis.1 Rotations of the head can

potentially occur through soft tissue distortion or improper patient

positioning. This study did not investigate the effect of head rotation on

landmark identification.

Conclusion

The intraexaminer and interexaminer landmark identification errors

associated with 52 posterior anterior cephalometric landmarks were

presented. The magnitude of landmark identification error had a wide range

with the horizontal error often being different from the vertical error. Some

landmarks showed significantly different errors when taken from skull

radiographs versus patient radiographs. Interexaminer landmark

identification errors were generally larger and, in many cases, significantly

larger, than intraexaminer errors. Many of the proposed posterior anterior

cephalometric analyses use landmarks which have an unacceptable

magnitude of landmark identification error.

Taken from the Angle Orthodontist on CD-ROM (Copyright © 1997 Angle

Orthodontist, Inc.), 1987 No. 2, 168 - 175: Cephalometric Reliability A Full

ANOVA Model for the Estimation of True and Error Variance Peter H.

Buschang, Richar...

---------------------or variance has been method error.

Depending on the design of the analysis, method error alone could produce

inaccurate results ( BUSCHANG ET AL. 19844). Moreover, comparisons of

error variance are difficult to interpret due to the lack of standardization. In

contrast, the coefficient of reliability that is presented

Taken from the AJO-DO on CD-ROM (Copyright © 1997 AJO-DO),

Volume 1958 Dec (901 - 905): Résumé of the workshop and limitations of

the technique - Salzmann

--------------------------------

Ever since God created man in His image, man has been trying to change

man into his image. Attempts to change facial appearance are recounted

throughout recorded history. The question of what is a normal face, as that

of what constitutes beauty, will probably never be answered in a free

society. Orthodontists, in their attempts to change facio-oro-dental

deviations from accepted norms, have adopted cephalometric measurement,

a method long employed in physical anthropology. With the introduction of

roentgenography, it was inevitable that this procedure should be employed

as a medium for the purpose of roentgenographic cephalometrics.

Taken from the AJO-DO on CD-ROM (Copyright © 1997 AJO-DO),

Volume 1987 May (414 - 426): Normal radiographic anatomy and common

anomalies in cephalometrics - Kantor and Norton

--------------------------------

Normal radiographic anatomy and common anomalies seen in cephalometric

films

Mel L. Kantor, D.D.S., and Louis A. Norton, D.M.D.

Chapel Hill, N.C., and Farmington, Conn.

Lateral and posteroanterior cephalometric radiographs are used routinely in

the diagnosis and quantification of dentofacial anomalies that require

orthodontic treatment. The anatomic information that these films contain is

occasionally overlooked as the clinician prepares tracings and makes

measurements. With the increase of the average age of the orthodontic

patient population, there is greater likelihood of the presence of disease. This

article describes some important features of normal radiologic anatomy of

the head and neck so that a clinician can better recognize pathologic

changes. Common pathologic findings and anatomic anomalies are also

illustrated. (AM J ORTHOD DENTOFAC ORTHOP 1987;91:414-26.)

During the course of evaluation and treatment, the orthodontist often takes

cephalometric radiographs of the patient's skull. A mathematic analysis is

usually done to help diagnose and quantify skeletal and dental

malocclusions, make growth predictions, or monitor the patient's treatment

progress. However, fortuitous findings must not be overlooked or ignored.

The clinician should evaluate the skull radiographs for any abnormalities

that might be present. To assist the orthodontist with this responsibility, we

will review normal radiographic anatomy of the human skull emphasizing a

systematic approach to interpretation. Examples illustrating variations of

normal anatomy that may be mistaken for pathosis are provided as well as

examples of pathologic changes that are often overlooked. No attempt will

be made to illustrate the full range and distribution of normal anatomy in

this limited review. References dealing with this subject are cited.

FILM INTERPRETATION

The information content of a radiograph is a complex function of film/screen

selection, technique factors, processing, and patient anatomy. The first three

of these parameters can be controlled and should be optimized to ensure the

best radiographic image with the least patient exposure. However, once a

radiograph is processed the amount of information recorded in the image

does not change, but the amount of information that can be retrieved from

each image is greatly affected by the circumstances under which the film is

viewed.1,2 Reduced ambient lighting, quiet surroundings, and the

elimination of peripheral light improve visual acuity.3,4

Kundel and Nodine5 have described two modes of visual perception of

radiographs. First is "global perception" resulting from rapid parallel

processing of the entire retinal image by means of pattern recognition and

rapid association with previously acquired visual concepts. The second is

"analytic perception,'' which is based on the extraction of features from the

incoming visual data and the use of logical rules to combine them in a

meaningful way. This technique results in a gradual buildup of the

perception. They suggest that experienced radiologists perceive

abnormalities in a global manner and that specific features are perceived

secondarily. The experienced orthodontist can often rapidly scan a

cephalometric film and tell whether a patient has a dental or skeletal

problem or a combination of the two and what part of the anatomy is

contributing the most to the problem. The cephalometric analysis usually

corroborates this global impression and quantitates a qualitative judgment.

Christensen and associates6 evaluated the effect of search time on

perception and found that obvious abnormalities are detected almost

instantaneously but that the overall number of abnormalities identified

increased as the viewing time increased. The number of visual images that

are immediately recognizable is a function of experience and the analytic

approach is necessary to evaluate those images that represent uncommon

findings. Even the experienced radiologist can be seriously misled and draw

the wrong conclusion if pattern recognition is the primary mode of

radiographic interpretation.7

Bisk and Lee8 reviewed 513 lateral cephalometric head films, which

represented the total population of the orthodontic practice of the senior

author. Eighteen films (3.5%) were classified as having abnormalities or

pathosis present as follows: enlarged adenoids— 5, failure of segmentation

C4-CS— 1, impacted canine— 1, interstitial emphysema— 1, osteoma— 1,

sinus polyp— 1, and sinusitis— 8. Because abnormalities occur

infrequently, the orthodontist should carefully search the cephalometric

films for features that would suggest disease and warrent further

investigation. Nanda, Merow, and Martin9 reported four cases of significant

abnormalities that were incidental findings: (1) a foreign object in the right

nostril, (2) bilateral retention cyst in the maxillary sinuses, (3) unusual

intrasellar cyst with a tooth or dermoid and, (4) multiple cysts of the jaws as

part of the basal cell nervous syndrome. Although the first two observations

had little impact on the patients' health, the latter two findings could have

had a serious negative effect on the patients' well-being if they had been

overlooked.

CRANIUM

In evaluating the cranium, the method suggested by Meschan10 is

recommended.

1. Calvarium and base. Initially, the size and shape of the calvarium and

base should be evaluated. Gooding11 reviews some of the common

morphometric indices available and concludes that they are most valuable

for following changes once an abnormality has been identified and that

"with experience normal craniofacial proportions at different age levels are

appreciated, and deviation is recognized as an indication of intra-cranial

abnormality.''

The calvarium is divided into three layers; the inner and outer tables are

compact bone and the middle table is cancellous. Thickness varies widely in

individuals and this will be demonstrated as varying radiodensities on the

radiograph. The thickest part of normal vault should not exceed 1 cm, after

which some degree of cerebral underdevelopment or systemic disease should

be suspected.12

2. Lines, impressions, channels, and sutures. Examination of the inner

surface of the calvarium will show numerous lines, impressions, and

channels that reflect the structure of the brain and its meningeal covering

(Fig. 1, A).

a. Meningeal vessel grooves. The arteries and veins of the meninges are

closely adapted to the inner able of the calvarium resulting in lines readily

identifiable by their well-defined borders, smooth undulating course, and

characteristic location. The middle meningeal vessels are usually the most

prominent; they begin at foramen spinosum and branch out, tapering along

the way.

b. Diploic vein channels. The diploic veins are contained in channels within

the cancellous bone of the middle table or diplöe. They will appear as

radiolucent channels 2 to 3 mm wide, coursing in an irregular pattern over

the calvarium; they do not appear to taper as the meningeal vessels do.

When two or more of these veins anastomose, a diploic lake may be present.

The diploic venous lakes are irregular, usually less than 2 cm in size, and

have multiple diploic veins running into them. Awareness of the existence of

diploic venous lakes and the observation of diploic channels associated with

them will usually allow the clinician to recognize these for what they are and

not mistake them for osteolytic lesions, such as bone metastasis,

meningoceles, fibrous dysplasia or histiocytosis X.13

c. Sutures. The sutures form the articulation of the cranial bones. Many of

the sutures are closed by the second year of life. The spheno-occipital

synchondrosis begins to ossify at puberty; the coronal, lambdoidal, and

sagittal sutures persist through early adulthood.10,14 Premature closure of

the sutures may be a primary defect, a component of other known head and

neck syndromes, or associated with metabolic, osseous, or hematologic

disorders.15 Sutural widening is usually a result of increased intracranial

pressure or destruction of bone at the suture margins. Observation of any of

these findings warrants further studies and consultation with the patient's

physician is recommended. The coronal, lambdoidal, and squamosal sutures

can be seen on the lateral cephalograph; the sagittal and lambdoidal sutures

and their junction, lambda, are seen on the posteroanterior (PA)

cephalogram. The sutures appear as radiolucent serpentine lines in their

anatomically expected location. Occasionally, there are small independent

bones that persist within a suture; these are called wormian bones and the

lambda region is a common location for them (Fig. 1, B). Multiple wormian

bones may be associated with cleidocranial dysplasia, cretinism, or

osteogenesis imperfecta.13

It is important to recognize the radiolucent lines that represent

the meningeal vessel grooves, the diploic vein channels, and the sutures, and

to be able to distinguish them from fractures of the calvarium, especially

given a history of recent trauma.

d. Arachnoid (pacchionian) granulation impressions. The arachnoid

granulations are an out-pocketing of the arachnoid membrane and sub-

arachnoid space that may extend into the dural sinuses or the adjacent lacuna

laterales. When found in the latter region, they may present as irregularly

rounded, sharply radiolucent depressions of the inner table of the skull. They

are most commonly found just lateral to the superior sagittal sinus, although

they can be located in proximity to any of the dural sinuses.16 They may

also calcify and this presentation will be described in a later section.

e. Dural sinuses. The sinuses of the dura mater are the channels by which the

blood from the cerebral veins, and some of the meningeal and diploic veins

drain into the internal jugular veins. The superior sagittal, sphenoparietal,

transverse, and sigmoid sinuses groove the inner table of the calvarium

producing broad radiolucent channels.

f. Convolutional markings. Also called digital markings or brain markings,

the convolutional markings are impressions or thinning of the inner table of

the calvarium caused by pressure from the convolutions or gyri of the

growing brain. They are most prominent in the 3- to 12-year age group and

tend to regress with age.17,18 Absence of these markings in the young or

persistence into adulthood, especially when accompanied by neurologic

signs and symptoms or other cranial morphologic abnormalities, is a

significant pathologic finding.19,20

g. Artifacts. If the patient's hair is particularly thick, wet, or pulled taut, it

may cause linear streaks to appear over the calvarium (Fig. 1, C).

3. Calcification within the calvarium. There are a number of intracranial

structures that may calcify in the absence of any disease. Reiskin7 has

stressed the importance of multiple right-angle views for the localization and

evaluation of these structures as a necessary component to distinguish

between those structures that are normal or physiologic and those that are

pathologic. Meschan20 has described the normal structures within the

calvarium that may calcify. They can be summarized as follows:

a. Pineal gland. The incidence of pineal calcification varies from 33% to

76% in the North American white population; there is a considerably lower

incidence in Japanese (10%), Indians (8%), and Nigerians (5%). The size of

the calcification averages 5 mm in length and 3 mm in height and width.

When seen in the frontal projection, the pineal gland is a midline structure

and a shift of 3 mm or more from midline is considered significant (Fig. 2,

A). Numerous methods have been described to localize the pineal gland in

the lateral radiograph; in general, it will be found above and slightly behind

the petrous portion of the temporal bone (Fig. 2, B). Calcification of the

pineal in children is not as common as in adults, but it is not a rare

phenomenon. It may be observed in approximately 5% of white children

under 10 years of age.

b. The habenular commissure may calcify and it will appear as a C-shaped

radiodensity located a few millimeters anterior to the pineal gland in about

30% of the adult population (Fig. 2, C).

c. Meningeal calcifications. The falx cerebri is calcified in approximately

7% of adults and is usually shown to best advantage in the frontal projection

where it appears as a linear midline radiopacity (Fig. 2, D). Calcification of

the arachnoid granulation appears as uniform radiopacities near the

corresponding granulation impression in the calvarium.

d. Petroclinoid ligament and diaphragma sellae. Calcification of the

petroclinoid ligament occurs in approximately 12% of adults and appears as

a radiopaque line extending from the posterior clinoid process to the petrous

ridge. Calcification of the diaphragma sellae may give the appearance of a

separate enclosed pituitary fossa. However, it must be remembered that we

are only seeing a two-dimensional representation and, in fact, there is a

space between the interclinoid calcifications to accommodate the pituitary

stalk. Radiographically, this appearance is described as ''roofing" or

''bridging" of the sella (Fig. 2, E).

In the absence of any clinical neurologic signs or symptoms,

these calcifications may be considered normal; however, it is important to

remember that many pathologic processes can be associated with these

calcifications. A patient with a calcified pineal gland who is experiencing

headaches, nausea, and vomiting should not be ignored; appropriate referral

and follow-up are warranted.

Once again, the patient's hairstyle may create artifacts that

mimic real findings. For example, if the hair is gathered on the lateral

surface of the skull into pigtails, it may resemble intracranial calcification on

the lateral skull film (Fig. 2, F).

4. Size and shape of the sella turcica. The sella turcica is a saddle-shaped

formation of the sphenoid bone in the middle cranial fossa. When viewed in

the lateral radiograph, the anterior clinoid processes are usually

superimposed; the hypophyseal fossa appears as a single dense curved line

that merges posteriorly with the posterior clinoid processes of the dorsum

sellae. The clinoid process may range from short and rounded to long and

pointed. Normal variants include (1) a middle clinoid process, (2) extension

of the sphenoid sinus into the dorsum sellae, posterior clinoid process or

anterior process, and (3) bridging as previously described. Because the sella

turcica is a midline structure, the floor of the hypophyseal fossa usually

appears as a single line. A double-contoured appearance may represent a

variant of normal, an artifact of positioning, or a significant pathologic

change.21,22 When viewed in the sagittal plane, the normal range for the

greatest anteroposterior dimension is 5 to 16 mm (average 10.6 mm), and the

depth as measured from a line between the anterior and posterior clinoid

processes to the floor of the hypophyseal fossa ranges from 4 to 12 mm

(average 8.1 mm).23 Significant variation in the size, area, or volume of the

sella associated with a variation of two standard deviations in height and

weight as compared to age-matched cohorts suggests a pituitary abnormality

and the patient's physician should be alerted to this finding. Expansion or

erosion of the borders of the pituitary fossa, especially if accompanied by

neurologic findings such as headaches, blurred or double vision, or

dizziness, is a significant finding and the patient should be referred for a

thorough evaluation. The sella turcica is also seen in the PA view where it is

superimposed over the superior aspect of the nasal cavity. In this view the

floor of the sella is usually convex upward.

PARANASAL SINUSES

The paranasal sinuses develop as outpouchings of the mucous membrane of

the fetal nasal cavity that extend into the maxillary, sphenoid, frontal, and

ethmoid bones, and subsequently enlarge. In adulthood the sinuses

communicate with the nasal cavity through ostia, thus reflecting their

common embryologic origin. The maxillary, sphenoid, and ethmoid sinuses

begin to enlarge in utero and may occasionally be detected radiographically

at birth. The frontal sinuses do not begin to pneumatize until the second year

and are not usually visible on the radiograph until the sixth year.24 Hence,

all four sets of paranasal sinuses should be evident in the average

orthodontic patient. The variation in size of the normal sinus may be great.

1. Maxillary sinuses are seen in the PA, base, and lateral views. In the

standard PA view, the petrous portion of the temporal bone is superimposed

over the superior one third of the sinus. If disease is suspected, the best view

of the maxillary sinuses in the frontal plane is obtained with a Water's

projection. The lateral view will show the borders in the sagittal plane;

however, the right and left sinuses will be superimposed and often

indistinguishable. On films obtained in the erect position, soft-tissue

swelling can usually be differentiated from free fluid in the sinus by the

nature of the air-shadow interface. The air-fluid line will be straight and

paraliel to the floor (Fig. 3, A); a soft-tissue swelling will produce a shadow

that follows the bony contours or is convex (Fig. 3, B). Bone destruction is

an important radiographic sign that requires biopsy and/or culture.

2. Frontal sinuses are seen to best advantage in the PA and lateral views.

They vary greatly in size, are usually asymmetric, and may even be absent.

An osteoma of the frontal sinus is not a rare finding (Fig. 4); it may be an

isolated finding or part of a generalized process such as Gardner's

syndrome.25,26 If osteomas are identified in association with the sinuses or

anywhere else, inquiry into family history and examination of the skin for

sebaceous cysts are required. The patient's physician should be informed of

any positive findings.

3. Sphenoid sinuses appear as a single cavity in the sphenoid bone, inferior

to the sella turcica in the lateral film. Although identifiable in the frontal

projection, the superimposition of the nasal septum, lateral nasal wall, and

the medial wall of the orbits makes evaluation difficult. The lateral extension

of the sphenoid sinuses is easily seen on the base projection; it is known to

vary greatly and, in the absence of any other pathologic findings, should be

considered an insignificant incidental finding.27

4. The ethmoid sinuses, also known as the ethmoid air cells, form the medial

wall of the orbit and the lateral wall of the upper half of the nose. The

ethmoid sinuses are divided by numerous septa resulting in multiple

compartments. Of the radiographic projections typically obtained for

orthodontic treatment planning, the ethmoid sinuses are best seen on the

lateral and base views. In the frontal view, they are seen as a radiolucency

between the medial rim of the orbit and the nasal septum.

When evaluating the paranasal sinuses, the integrity of the bony borders and

adjacent structures and the degree of aeration must be established. In health,

the thin mucous membrane lining is not visible on the radiograph.

MASTOIDS

The mastoid air cells communicate indirectly with the nasal cavity via the

middle ear; however, embryologically they develop separately from the

paranasal sinuses. Nonetheless, the radiographic appearances of air-filled

cavities within the bone resemble the ethmoid air cells. The distribution and

pneumatization of the mastoid air cells are extremely variable; the cells are

located in the mastoid process and periauricular region and may extend as

far forward as the zygomatic process of the temporal bone.28

CERVICAL SPINE

The upper vertebrae are often visible on the lateral and PA cephalometric

radiographs. The atlas has no body or spinous process and has the form of a

ring. The axis has the fundamental structure of the cervical vertebra with the

addition of an upward projection called the dens or odontoid process. The

dens occupies the space where the body of the atlas would have developed;

it articulates with the posterior surface of the anterior arch of the atlas and

provides a pivot around which the atlas and skull rotate. The body of the axis

and the odontoid process have separate ossification centers23 and often do

not fuse until age 12.20 Therefore, a transverse radiolucency at the base of

the odontoid process in a young ambulatory patient with no history of

trauma should not be mistaken for a fracture.

The C-spine has a gentle curvature and is convex anteriorly when viewed

from the side. This normal lordotic curve is position-dependent and can be

altered as a result of failure to achieve natural head position when placing

the patient in the cephalometric head holder or as a result of muscle spasm

that causes the patient to posture the head in an effort to reduce pain and

discomfort.

Lines drawn along the anterior and posterior margins of the vertebral bodies

should be practically parallel. A straight line drawn along the front of the

odontoid process meets the anterior margin of the foramen magnum and lies

approximately 1 mm behind and away from the posterior border of the

anterior arch of the atlas. The normal dimension of the spinal canal ranges

from 18 to 27 mm at the first cervical vertebra to 15 to 20 mm at the seventh

cervical vertebra for children 15 years of age and less. For adults, the ranges

are 16 to 30 mm and 13 to 24 mm, respectively.20 In the PA view, the

lateral border of the vertebral body will be in alignment and the spinous

process will be visible. Frank displacement of a vertebra is a serious

abnormality that demands further investigation (Fig. 5).

The intervertebral disk is a fibrocartilaginous anulus with a gelatinous center

and is not visible on a conventional radiograph. However, we can make

inferential observations about the intervertebral disk by evaluating the

surrounding anatomy. The intervertebral disk space appears as a

radiolucency between the vertebral bodies defined by the relatively parallel

inferior and superior cortical margins. If the cortical margins appear

convergent or the disk space is narrowed, this may suggest a herniated disk.

UPPER AIRWAY AND NECK

The upper air passages— the nasal cavity, oral cavity, pharynx, and larynx

— appear radiolucent on the skull film. When sufficiently thick, the soft

tissues of the region will have an intermediate radiodensity between the

airway and skeleton.

The nasal air passages usually conform to the bony architecture as the

mucosal lining of the nasal cavity is usually less than 1 mm thick and does

not cast a radiographic shadow. Thickened membranes or linings can be

seen as an intermediate density between bone and air with proper exposure

factors. The cigar-shaped nasal conchae will be superimposed over the

airway; this will be discussed in greater detail in the next section.

The dimensions of the oral airway will vary depending on the position of the

tongue. If the tongue is elevated, it may contact the soft palate and their

radiographic shadows will merge. The palatine tonsils are situated between

the palatoglossal and palatopharyngeal folds in the lateral fauces. These can

sometimes be distinguished on the lateral film, especially if they are

inflamed and enlarged (Fig. 6).

On the superior aspect of the posterior wall of the nasopharynx, there is a

collection of lymphatic tissue (the nasopharyngeal tonsils or adenoids) that

may be quite large in children. This is usually easy to identify on the lateral

cephalometric film (Fig. 7). Changes in breathing patterns caused by

hypertrophied adenoids may affect facial growth patterns.29,30 The

lymphatic tissue tends to atrophy with age and will not be as prominent in

adult patients. The opening of the eustachian tubes on the lateral wall of the

nasopharynx just behind the inferior nasal conchae may be evident as a

round, relatively radiolucent area.20 These structures are difficult to see, but

may be discerned with certain anatomic and exposure factors. The soft

palate separates the nasopharynx from the oropharynx. At rest, it extends

from the posterior borders of the hard palate and arches inferiorly.

In the lateral projection, the hyoid bone is seen just below the angle of the

mandible. The thyroid, cricoid, and tracheal ring cartilage are usually not

visualized but may on occasion have areas of calcification that appear on the

radiographs. The epiglottis and the laryngeal folds are also seen.

The prevertebral soft tissue and muscles can be seen separating the airway

from the vertebral column. The retropharyngeal shadow at the line of C2

varies from 2 to 7 mm in children less than 15 years of age and from 1 to 7

mm in adults; the retrotracheal shadow at the level of the C6 varies from 5 to

14 and 9 to 22 mm, respectively.20 The soft-tissue shadow should have a

smooth anterior outline. In the PA view, the lateral wall of the

laryngopharynx and the larynx are seen; other parts of the airway are

obscured by superimposition of bony structures.

DENTOMAXILLOFACIAL COMPLEX

Orthodontists are most familiar with the facial portion of the skull as this is

the region they routinely treat. For our purposes we will consider the

dentomaxillofacial complex to include the orbits, nose, zygomatic arches,

and jaws. The paranasal sinuses have been dealt with separately in a

previous section.

1. Orbits. In the PA view, the rim of the orbit is seen as a smooth round

radiopaque line. There are a number of structures that appear within the orbit

and these should all be evaluated. The lesser wing of the sphenoid

contributes to the floor of the anterior cranial fossa and is seen as a

horizontal convex-down curvilinear radiodensity in the superior third of the

orbit. From the region where this line intersects the superolateral border of

the orbit, there is another linear radiopacity running downward and medially;

this is called the innominate line and represents a cuvature of the greater

wing of the sphenoid.

The optic foramen is a round radiolucency near the medial orbital wall. The

superior and inferior orbital fissures can be seen extending from this region

in lateral-upward and lateral-downward directions, respectively.

Occasionally, one can follow the path of the inferior orbital fissure as it

becomes the inferior orbital canal and emerges on the front of the face as the

infraorbital foramen. Just medial and slightly below the infraorbital foramen

is a somewhat larger well-defined circular radiolucency; this is foramen

rotundum through which the maxillary division of the trigeminal nerve

passes as it leaves the skull base. This may be a region deserving careful

scrutiny if the patient complains of pain over the area that this division

innervates. The vertical position of the foramen will vary depending upon

the tilt of the patient's head relative to the central ray of the beam. At the

junction of the middle and medial thirds of the superior rim of the orbit, the

supraorbital foramen may be seen as a small, round radiolucency (Fig. 8). In

the lateral view, the superior and inferior walls of the orbit are seen.

Likewise, the posterior and anterolateral margins of the orbit are visualized;

however, the superimposition of structures makes it difficult to distinguish

left from right. The zygomaticofrontal and maxillofrontal sutures may be

seen at the rim of the orbit and should not be mistaken for fractures.

2. The nose. In the PA view, the nasal septum, lateral walls, and conchae are

easily defined. The nasal septum should be positioned at the midline;

displacement from the midline may represent a congenitally deviated

septum, prior trauma, or the presence of a pathologic process causing the

displacement (Fig. 9). Extending medially from the lateral walls are the

nasal conchae or turbinates. The inferior and middle conchae are usually

seen, but the superior conchae may not be visualized. In the lateral views,

the inferior conchae appear as a cigar-shaped radiopacity. Often the posterior

extent of the conchae extends beyond the posterior border of the maxillary

sinus, which makes it radiographically difficult to distinguish from an

isolated radiopacity in the nasal cavity. If there is a question as to what this

radiographic shadow represents, establishing continuity of the outer

boundary of the radiopacity with the adjacent turbinate bone should confirm

its identity. Should a question persist, the posterior nasopharynx can be

visualized by indirect laryngoscopy using an angled mirror and proper

lighting.

3. Zygomatic arches. The zygomatic process arises from the maxillary bone

at the region of the first molar. The radiodensity, size, and shape of this

structure are variable and the structure often takes on a different form,

depending upon the angle of the directed x-ray beam. The zygomatic process

may appear quite radiolucent if the maxillary antrum extends into it. The

greater the extension of the maxillary sinus into the zygomatic process, the

greater the contrast of the dark radiolucent air spaces and the sharply defined

cortical walls of the process. Seen in the lateral cephalogram, the corticated

walls of the zygomatic process appear as a U-shaped radiopaque line known

as a key ridge. The definition of the molar apices superimposed on the

zygoma will vary with the amount of pneumatization that has occurred. If

aeration is minimal, molar apical and maxillary sinus anomalies may be

masked or ill-defined.

4. The jaws. Details of the teeth and their surrounding structures are difficult

to see on skull films because of superimposition of anatomic structures and

the inherent resolution limitation of screen film. Evaluation of the teeth and

periodontium is best accomplished by a periapical film. Most orthodontists

use these intraoral films in their diagnostic evaluations and treatment plans.

Misinterpretations can present problems here also. For example,

occasionally a double image of the lamina dura is seen that reflects the

normal concavities and fluting of the roots or the superimposition of

different roots of a multirooted tooth such as the maxillary first molar.

Superimposition of the lingual root surface and periodontal ligament space

of the first premolar onto the distal surface of the canine in the periapical

film should not be mistaken for a vertical root fracture of the canine. Care

should be taken to examine carefully for supernumerary teeth and evidence

of small developing bud follicles. They can be of great consequence if the

clinician is trying to move teeth into the space they occupy. If initially

overlooked and subsequently noted on follow-up radiographs, they are a

source of embarrassment at least, and iatrogenesis at worst (Fig. 10).

The trabecular pattern of the anterior maxilla is fine, granular, and dense.

The posterior maxilla shows a slightly less dense pattern with larger marrow

spaces. The trabeculae of the anterior mandible are thicker than the maxilla,

presenting a course pattern with large marrow spaces. The posterior

mandibular periapical trabeculae and marrow spaces are usually the largest

in the jaws. These can be variable in size and mimic pathologic lesions.

Changes in the density and pattern of the cancellous bone may result from

inflammation, systemic disease, or tumors (Fig. 11).

The mandibular symphysis frequently has a radiolucent line at the midline

suture that disappears at about 1 year postpartum. If this radiolucency is

found in older children or adults, it may suggest a fracture or cleft. The

genial tubercles are the bony projections of attachment of the genioglossus

and geniohyoid muscles. They often have a small radiolucent area in the

center (the lingual foramen) that is the point of exit of mandibular nerve.

Depending upon its size, this may be mistaken for incipient pathosis. The

mental fossa is a depression found in the labial aspect of the mandible. The

thinness of the hard tissue in this area may be mistaken for periapical disease

of the incisors. similarly, the- mental foramen, located between the first and

second premolars, can mimic periapical pathosis in this area. The

mandibular canal forrns a dark linear radiographic shadow with thin superior

and inferior opaque borders cast by its lamella boundaries. The molar teeth

apices are frequently projected over this canal, giving the illusion of a

discontinuous lamina dura surrounding these teeth. This is due to the

localized overexposure caused by this radiolucent linear structure. Finally,

the submandibular fossa is a depression on the lingual side of the mandible

below the mylohyoid ridge that accommodates the submandibular gland. It

will appear as a local radiolucency with scant or absent trabeculation. The

anterior and posterior aspects of this radiolucency will blend into the

surrounding bony pattern.

SUMMARY

We have presented a review of certain aspects of normal radiographic

anatomy, discussed range and distribution, and identified some common

errors in diagnosis. Nonetheless, this review has covered only a small

amount of the information available. It is up to clinicians through careful

study of the films, by use of available reference material, and by

consultation with colleagues in medical and dental radiology to constantly

expand and improve their knowledge of normal radiographic anatomy.

All radiographs of the head taken for orthodontic purposes should be

considered skull films before they are thought of as cephalograms. By

adopting this attitude, the orthodontist will be inclined to carefully review

these films for significant deviations from normal and evidence of pathosis.

Only after this responsibility has been met should cephalometric tracings or

other morphometric analysis be done.

The authors wish to thank Dr. Allan B. Reiskin for reviewing the

manuscript, and providing helpful comments and suggestions. We also

appreciate the expert assistance provided by the UNC School of Dentistry

Learning Resources Center, especially Mr. Warren McCollum for the

production of the illustrations and photography.

Taken from the Angle Orthodontist on CD-ROM (Copyright © 1997 Angle

Orthodontist, Inc.), 1997 No. 2, 83 - 85: Making sense of cephalometrics

Robert M. Rubin.

--------------------------------

EDITORIAL

Making sense of cephalometrics

Robert M. Rubin, DMD, MS

In the 60-year history since the development of cephalometric radiology,

literally undreds of methods of analysis have been proposed. Many of them

have contributed to a better understanding of the complexity of changes

associated with facial growth. Some analyses have been useful in identifying

how individual patients vary from norms that have been derived from large

numbers of cohorts. Some cephalometric analyses and methods of

superimposition are useful in monitoring the changes that are due to growth

or to a combination of growth and treatment.

Cephalometric measurements are also useful in descriptive communication.

Just as Angle’s classification describes a specific relationship between the

teeth in the maxilla and mandible, the Downs’ facial angle communicates a

picture of a relationship between the Frankfort horizontal and

nasion/menton.

Each method of analysis is based on certain assumptions, some expressed

and some implied. This essay examines several assumptions and evaluates

their strengths and weaknesses. In addition, there are two different uses for

assessment of the presenting patient. How does this patient vary from

recognized norms? This information allows the practitioner to focus on

where the patient’s anomalies exist, and allows him or her to plan for the

achievable ideal for the patient.

The second use of cephalometrics is to monitor changes due to growth or

treatment, or their combination. I propose that some measurements may be

well-suited for assessment but are poor choices for monitoring change.

Similarly, some measurements are poor choices for assessment but are

particularly well-suited for observing change. Failure to make this

distinction has led to confusion in treatment and absence of clarity in

communication in describing changes that occur with growth and/or

treatment.

Almost every article in the orthodontic literature begins with a section

describing the cephalometric system used for the evaluation that follows. It

would be more efficient if each writer did not have to define the method of

cephalometric assessment, as would be the case if there were agreement in

our profession on the measurements and their uses. Precedent for adopting

such agreements exists. In 1929, the world’s anthropologists met and agreed

on the definition of the Frankfort horizontal plane. Orthodontists were quick

to adopt that definition, and that agreement has contributed to better

communication in the anthropologic and orthodontic literature. Now, 60

years after its introduction, radiographic cephalometrics is overdue for an

agreement on how we assess craniofacial morphology and how we monitor

changes due to growth and/or treatment.

Consider the following analogy: The orthopedic surgeon, noting growth of

the femur, observes that the inferior epiphysial cartilage grows several

millimeters. The neurosurgeon may note oppositional growth between the

lumbar vertebrae. Neither of these physicians would suggest that the result

of these increments of growth would drive the feet into the ground. They are

in agreement that growth of the vertebrae or femur contributes to increased

height. This agreement may not seem remarkable because the obviousness of

it is so apparent. But, consider the possibilities if we lived in a weightless

environment. In that environment confusion about describing the results of

growth would be possible. Some would say that femoral growth carries the

ankle down; others would say it carries the pelvis up. Neurosurgeons might

describe vertebrae growth as moving the feet and head in opposite

directions. This is the sort of confusion we have in craniofacial growth.

The ease of identifying sella turcica led many cephalometric researchers to

choose the line from sella to nasion as a key line of registration. It turns out

to be a relatively poor choice because of the confusion it engenders. With

the head facing right, as is generally agreed in cephalo-metrics in this

hemisphere, we consider growth to move skeletal landmarks to the right and

down, away from sella. Confusion occurs when growth at the spheno-

occipital synchondrosis is considered. Its proliferation, which often

continues through puberty and can total more than 10 mm, carries the

glenoid fossa and the mandible to the left—the opposite direction that

condylar growth carries the mandible. This conflict is analogous to viewing

femur growth as carrying the feet into the ground. Agronin and Kokich

never mentioned the spheno-occipital synchondrosis in their report,

“Displacement of the glenoid fossa: a cephalometric evaluation of growth

during treatment.” (Am J Orthod Dentofac Orthop 1987;91:42-8.) They

stated that during craniofacial growth, articulare is displaced posteriorly and

inferiorly relative to the sphenoid bone. “The data support the premise that

changes in the spatial orientation of the glenoid fossa and temporal bone

may have an effect on mandibular position.” They were undermined by their

assumptions! A more accurate description is that growth of the spheno-

occipital synchondrosis carries the craniomaxillary complex superiorly and

to the right, making the case more Class II and increasing facial height.

There is a baseline available for viewing craniofacial growth that is

analogous to using the ground for a baseline for somatic growth. That

baseline is basion, the anterior edge of the foramen magnum. Using basion

as the base (aptly named), all craniofacial growth is seen as movement away

from the spinal cord, just as all skeletal growth is viewed as elevating the top

of the head.

A cephalometric analysis that uses this concept is the Coben basion

horizontal analysis, first presented in 1955. Orienting on basion and

maintaining the Frankfort plane parallel to the horizon, all growth of the

craniofacial skeletal is seen to carry structures to the right, away from the

vertebrae column.

Frankfort horizontal is a useful plane because it is believed to approximate

the optic axis, the plane that appears to be kept level throughout life. This is

important as it correlates the clinical appearance of the patient to his or her

cephalometric analysis. Analysis based on sella-nasion may not relate well

to the presenting patient if the anterior cranial base is steeply sloped. One

problem with the Frankfort plane is that it is not suitable for serial

evaluations. Coben handles this by using a constructed Frankfort on

subsequent tracings, drawn tangent to porion and at the same angle to sella-

nasion as the original tracing.

Some cephalometric measurements are excellent for assessment—that is, the

evaluation of the initial film to describe the problem. The same measurement

may, however, be a poor choice for monitoring change because an element

of it may be unstable. For example, upper incisor to occlusal plane is an

excellent assessment of the torque of the incisor. The norm is 65 degrees. It

is a poor choice to monitor torque achieved because its baseline, the occlusal

plane, can change during treatment. To monitor upper incisor change it

would be wiser to use upper incisor to sella-nasion. However, this is a poor

choice for assessment as it is remote from the occlusion and independently

related. A large upper incisor-sella-nasion angle can be due to a procumbent

incisor or a flat anterior cranial base.

It is not sufficient to rate a measurement as good or poor. It is important to

rate it as good or poor for assessment, and good or poor for monitoring

change. Table 1 shows some examples of commonly used cephalometric

measurements and an appraisal of their usefulness.

This essay proposes that superimposition on basion with the Frankfort plane

kept horizontal be adopted as the universal method of registration for

evaluating overall craniofacial changes due to growth and/or treatment. Area

superimpositions will, of course, still be necessary to determine the specific

sites of the changes. Such an agreement would eliminate the need to preface

every cephalometric study with an extensive section describing the method

of superimposition. The reduction in journal space and in reader’s time

would be an enormous savings for our specialty, and lead to a more efficient

comparison of studies. Frequently, it is impossible to compare similar

studies when different landmarks and methods of superimposition are used.

In addition, a glossary of measurements should be developed that not only

defines the measurement, but indicates if it is valid for assessment or for

documenting change. I believe these two measures would contribute to

increased clarity in our literature and enhanced coherence in the process of

planning treatment and evaluating progress and posttreatment records.

Orthodontics is marvelously complex. It is unnecessary to add to its

complexity by promulgating confusing and fuzzy assumptions that impair

accurate communication.