Cephalometric Superimposition on the Occipital Condyles as a Longitudinal Growth Assessment Reference: I-Point and I-Curve



This retrospective study tests the hypothesis that superimposition referenced at the occipital condyles (defined as I-point, I-curve) and oriented to the anterior cranial base (ACB) will display a growth pattern that is more consistent with independent evaluations, such as the Melsen necropsy specimens and the Bjork implant studies, when compared with traditional superimpositions referenced at sella turcica. Twenty-eight sets of serial lateral cephalometric radiographs were selected from an archived growth study. The apparent facial growth was compared using polar coordinate analysis from superimposition tracings of the serial films for each subject. The two superimposition methods were compared. The traditional method, ACB registered on the anterior curvature of sella turcica, versus registration on I-point while maintaining ACB parallel. I-point registered superimpositions consistently displayed a facial growth pattern that was more consistent with the classic necropsy specimens of children and the cephalometric studies superimposing on implant markers. Traditional ACB superimposition suggests that airway is restricted by normal growth. This apparent physiologic artifact does not occur when superimpositions are registered on I-point. Sella turcica displays vertical movement that is consistent with brain growth. These data indicate that registration on I-point is a more accurate physiologic representation of facial growth than the traditional ACB superimpositions. When compared with the traditional registration at sella turcica, I-point superimposition better elucidates physiologic growth patterns. As cephalometrics evolve from a two to a three dimensional science, it is important to use a more biologically valid registration for evaluating therapeutics and facial growth patterns Anat Rec, 2008. © 2008 Wiley-Liss, Inc.

The occipital condyles (OC) are the anatomically stable structures at the base of the skull that articulate with the spinal column (atlanto-occipital joint). They are formed early in the development of the head and are relatively stable because anabolic modeling (bone formation) progresses away from articular surfaces of endochondral bones (Roberts et al., 2004). Despite the biological appeal of the OC as a cephalometric landmark, it is difficult to use for two dimensional (2D) longitudinal growth studies because the OC, in norma lateralis, are obscured by dense cranial and vertebral structures in many routine cephalometric radiographs. Anterior cranial base (ACB) structures evolved as convenience landmarks for evaluating facial growth because they are visible on almost all routine 2D films and are internally stable after fusion of the sphenoethmoid, frontoethmoid, and intersphenoid synchondroses by about 8 years of age. (Melsen, 1974; Hoyte, 1991). This ACB superimposition of longitudinal films results in a biologically distorted view of the actual growth process. Furthermore, cephalometric studies are inconsistent with anatomically independent methods such as bone surface activity in necropsy specimens (Melsen, 1974) and jaw rotation demonstrated during growth by endosseous metal implants used as references for superimposition of longitudinal cephalograms (Bjork and Skieller, 1977, 1983). ACB superimposition ignores the contribution of the sphenooccipital synchondrosis (SOS) relative to increases in distance between anterior and posterior cranial bases.

Landmarks for the use of the OC as a posterior cranial base reference are I-point, U-point, and I-curve (Ic). The latter is introduced to supplement the traditional anterior curvature of sella turcica-anterior cranial base (SACB) reference. I-point (Fig. 1) is the most antero-inferior point on the OC in norma lateralis. U-point is the midpoint between I-point and O prime (O′) as defined by Frankel (Frankel, 1980) (Fig. 1). Ic is the external contour from U-point to the midpoint of the inferior surface of the OC in norma lateralis.

Figure 1.

(a) Dry skull in norma lateralis radiograph; arrows pointing to the superimposition of the occipital condyles (bottom arrow) and remnant of the sphenooccipital synchondrosis (top arrow). (b) Tracing of the superimposed occipital condyles as seen on a lateral cephalometric radiograph: I point is the most anterior-inferior point on the condyle; U point is the midpoint between O' and I-point; I Curve from U point to the midpoint of the inferior contour (SIA or remnant).

It is hypothesized that the OC referenced at I-point/Ic and oriented to the ACB superimposition (OC-ACB) will more accurately display longitudinal growth patterns and differences between orthodontic treatment effects consistent with previous growth studies using implants as radiographic markers. It is suggested that the OC are directly under the craniofacial center of gravity, and that the growth pattern observed with superimposition referenced at I-point/Ic is a continuation of the fetal unfolding growth pattern.


  • 1The center of gravity is centered over the ventral portion (basioccipital) of the OC on lateral cephalometric films; that reference serves as the base for centripetal growth of the crania.
  • 2Growth of the craniofacial complex is an extension of the pattern established in early embryologic development.
  • 3The human brain, continuing its morphological development through the age of about 20 years, is instrumental in craniofacial growth via tension developed on the galea aponeurotica.
  • 4The normal forward rotation of the face is related to the coordination of neural growth, tension on the epicranial aponeurosis, connective tissue and muscle maturation, hinge movement of cranial sutures and joints, somatotype expression, muscle tonicity, and the relative effects of gravity on the hydrated components of the tissues and the craniofacial musculoaponeurotic system (CFMAS). (Standerwick and Roberts, (Submitted for publication)).
  • 5Forward rotation of the craniofacial complex provides the increased airway capacity required during growth to accommodate increases in stature, muscle mass, and metabolic demand.


Twenty-eight patients with serial lateral cephalometric radiographs of adequate quality were selected, from The Iowa Facial Growth Study (University of Iowa College of Dentistry; Higley, 1954). The sample consisted of 19 (8 to 14 years old) females and 9 (8 to 16 years old) males and utilized the film at 8 years of age as time zero (T0). Twenty-three landmarks were identified on all cephalometric radiographs (Table 1).

Table 1. Twenty three landmarks identified on all cephalometric radiographs
Inferior internal symphysisAnterior cranial base/ethmoidale 
Anterior nasal spine (ANS)Posterior nasal spine (PNS) 
Anterior curvature of sella turcica (Sella)
Inferior contour of the inferior alveolar canal

The radiographs were traced three times (Pilot ENO 0.7 mm pencils, Pilot HCR-197-l, Pilot HCR-197-R, Tokyo, Japan; 3M Unitek tracing acetate, Monrovia, California) to compare the accuracy of the landmarks. The linear distances were compared and analyzed by superimposing transparencies of the cephalometric tracings (Table 2). Reproducibility of SNA, SNB, ANB, palatal plane, and SN-GoGn angles was determined by tracing sample radiographs three times. The sample was neutral divergent (SN-GoGn 27° < χ < 37°) with a normal ANB relationship (Table 3).

Table 2. Repeatability: Average distance of landmark placement between repeated tracings (millimeters)
LocationAvg. distance betweenSD of distance between points
A pt220.810.370.191.50220.420.260.050.95
B pt220.680.440.121.69220.430.280.051.17
I point271.010.730.133.50270.640.390.101.47
PT point270.430.350.021.30270.360.370.031.84
Table 3. Summary statistics for the sample (degrees)

To compare the superimposition methods, the serial cephalometric films were first referenced at the SACB. The serial radiographs were then compared with the measurements from the same radiographs oriented to the OC-ACB for which the ACB was kept parallel to itself in the superimposition series. The relative movement of the cephalometric points of condylion (Cond), manibular internal symphysis (IS), and the inferior alveolar canal (IAC) were evaluated using an OC and ACB derived polar coordinate system; described below (Fig. 2). The same T0 coordinate system was used for both superimpositions, registered at the anterior curvature of sella turcica (S) and I-point, respectively (see below), so that the T1 overlay position during superimposition was the only variable. To construct the coordinate system, S and the ACB were traced from the T0 radiographs, copied and transferred for the T1 radiograph tracing. The polar coordinate was constructed using the T0 SACB line as the horizontal axis, and a perpendicular line through the SACB line which bisected I-point (Fig. 2) as the vertical axis. Lines tangent to the SACB line and I point perpendicular were made to bisect the T0 landmarks: IS, IAC, and Cond (Fig. 3). T1 landmarks were located with a line originating from the T0 origin at the SACB line and then I-point perpendicular which were made to bisect the T1 landmark (Fig. 3). The angle between the T0 and T1 line was measured to determine horizontal (ventral/dorsal) and perpendicular (cephalad, caudad) movements. Cephalad and dorsal movement were recorded as negative, whereas caudad and ventral movement were recorded as positive.

Figure 2.

Landmark location (Cond, IAC, and IS are denoted by arrows). Tangental and bisecting lines used to locate IAC are shown. Displayed are the anterior curvature of sella turcica—anterior cranial base line with associated I point perpendicular used to construct the arbitrary grid.

Figure 3.

(a) Growth changes shown by superimposition referenced on sella and anterior cranial base in a representative subject, as an example of the triangulation method. Cephalad and dorsal movements are negative, whereas caudad and ventral movements are positive. (b) Growth changes shown by reference at the occipital condyles and anterior cranial base in the same subject as (a); the triangulation method is the same as for Figure 3a. Represented are the SACB line and I-point perpendicular line (black), Cond (red), IAC (blue), IS (green) with representative ages. Greater ventral movement of IAC is observed in (b) which is more consistent with airway demands during growth. Sella turcica is observed to move cephalad and ventral when referenced at I-point, which is consistent with previous microscopic observations of bone modeling. Greater dorsal movement of Cond, IAC, and IS is observed with SACB. Less caudad movement of IAC with I-point is important for airway development.

Point location for the inferior contour of IS was determined by the deepest point in the concavity on the inferior aspect of the internal mandibular symphysis.

Cond was defined as the most posterior superior aspect of the mandibular condyle head. IAC was determined by locating gonion as the point bisecting the angle created by lines tangent to the descending posterior ramus and the inferior mandibular border posterior to the antigonial notch. The midpoint angle created by these tangents was bisected through the inferior cortical contour of the inferior alveolar canal; this point bisecting the inferior border was selected for IAC (Fig. 2).

Statistical Analysis

To assess the repeatability of points in the three tracings, the distance between the points on each pair of tracings was calculated as well as the averages and standard deviations of the distances between the points (Table 2). Means, standard deviations, ranges, and 95% confidence intervals for the mean were calculated to summarize the measurements made using I-point and SACB as references. Agreement between measurements calculated using I-point and SACB as references were assessed using intraclass correlation coefficients and paired t-tests to determine if the measurements were significantly different using the two methods.


Summary data for the sample are in Table 3.

Measurements using I-point and SACB as references are summarized in Table 4. Confidence intervals that are not zero indicate significant movements in the specified direction. For example, the Cond horizontal measurement using I-point as the reference had a confidence interval that included zero, so the movement could not be considered to be significant in one direction or the other. However, the Cond horizontal measurement using SACB as the reference had a confidence interval that was entirely below zero, so the movement could be considered to be significant in the dorsal direction.

Table 4. Measurements using I and S as references (degrees)
MeasurementReferenceNMeanSDMin.Max.95% CI for mean
Cond horizSACB28−9.886.22−212−12.29−7.47
Cond perpSACB287.2321.91−3776−1.2615.73
IAC horizSACB28−1.952.19−7.53−2.80−1.10
IAC perpSACB2811.918.91−9308.4615.37
IS horizSACB281.52.49−4.550.532.47
IS perpSACB288.2914.57181.52.6313.94
S horizI-point2835.8640.83−829020.0251.69
S perpI-point28−4.544.13−151−6.14−2.94

Agreement between measurements using I and S as references: There were significant differences in the measurements made using the two reference points and the ICCs were low except for IS horizontal (Table 5), so the measurements made using the two reference points did not have good agreement. The conclusions made using the confidence intervals for the mean regarding significance of movement would be similar using either reference point for IAC perpendicular, IS horizontal, and IS perpendicular, whereas the conclusions regarding significance of movement would be different for the two reference points for Cond horizontal, Cond perpendicular, and IAC horizontal.

Table 5. Agreement between measurements using I and S as references
Cond horizS − I28−7.889.63−30120.00020.34
Cond perpS − I2822.2733.53−71220.00160.00
IAC horizS − I28−1.792.06−63<0.00010.51
IAC perpS − I285.965.02−2.515.5<0.00010.64
IS horizS − I28−1.051.24−3.520.00010.80
IS perpS − I284.3813.26−0.571.50.09210.19

I-point was shown to be accurately located when compared with other cephalometric points (Table 2). I-point was within the range of other commonly used landmarks such as: A-point, B-point, Gonion, Menton, Nasion, and Pterygomaxillare (Table 2). The facial divergency and sagittal relationship of the sample was calculated and was shown to be representative of neutral divergency and slightly increased ANB angle (Table 3). Therefore, an average internal forward rotation of 15°, external/matrix backward rotation of 11 to 12°, and resulting average 3–4° of forward rotation is representative of this sample (Bjork and Skieller, 1983; Proffit and Fields, 2000).

Triangulation of points at IS, IAC, and Cond resulted in the data in Tables 4 and 5. Sella remained static with SACB superimposition, but displayed ventral and cephalad movement with OC superimposition. Cond remained static horizontally and moved cephalad with respect to I-point, whereas dorsal and caudad movement was observed relative to SACB. IAC moved caudad for both I-point and SACB (almost twice as much for SACB), and dorsally for SACB. IS displayed similar ventral movement, whereas greater caudad movement was observed with SACB. The representative patterns can be seen in Fig. 4; rotational pattern associated with I-point/curve reference can be seen in Fig. 4a, when compared with the traditional SACB, Fig. 4b.

Figure 4.

(a) Growth changes are shown by superimpositions referenced on the occipital condyles while maintaining a parallel relationship for the anterior cranial base in a representative subject. (b) Tracings of the same series of films are superimposed on the anterior cranial base; black is age 4y9m, blue is 6y6m, red is 10y6m, and green is 14y0m. Notice that physiologic movement of sella turcica, development of the airway and proportional craniofacial development is better displayed with I-point superimposition.

In Fig. 4a, sella turcica is not static; therefore, the condyle moves vertically in conjunction with sella turcica. The orientation of the grid measurements did not allow for linear comparison of movement; the closer the position of a landmark to the horizontal or vertical reference line, the greater is the change observed. The angular measurements were chosen to decrease the effect of any magnification error in the sample.


Craniofacial growth investigators have long grappled with the difficulty of finding a stable radiographic reference for documenting longitudinal growth patterns. Isaacson (Isaacson, 1996) described the frame of reference as the fundamental limitation of cephalometric radiograph superimposition. The sella-nasion line has received much justified criticism as a longitudinal reference (Ricketts, 1975, 1976; Ghafari et al., 1987). Nasion and sella turcica are not suitable as stable longitudinal references because they are not fixed in the sagittal dimension (Baume, 1961; Latham, 1972; Melsen, 1974; Bjork and Skieller, 1983; Ghafari et al., 1987; Ranly, 2000).

The SOS, brain and orbit are horizontal and vertical growth centers (Ricketts, 1976; Enlow and Hans, 1996; Dixon et al., 1997; Roberts et al., 2004). Although growth within the SOS is bi-directional, it is not necessarily equal and opposite, nor symmetric (Baume, 1961; Melsen, 1971, 1974; Vilmann et al., 1980; Coben, 1998). This poses a problem for any growth pattern study based on sella turcica because the SOS is typically located between the mandibular condyle and the sella turcica in the sagittal plane.

A new orientation plane is proposed: ACB oriented parallel and referenced at I-point/I curve on the OC, from which other landmarks are measured. This “point set” of structures was chosen as orientation error is more likely to be amplified if the reference structures are close together. To accommodate the growth between the two structures, it is necessary to superimpose on the OC and maintain the ACB parallel.

Developmentally, the OC begin to ossify at 13–14 weeks of gestation, (Mauser et al., 1975; Nemzek et al., 2000) and achieve adult form (complete ossification) by 4–5 years of age (Frankel, 1980). Early development of the OC facilitate growth and support the cranium (Frankel, 1980). The synchondrosis intraoccipitalis anterior (SIA) has fused by about age 4–7 (Tilmann and Lorenz, 1978; Hoyte, 1991).

The OC are proximate to the sagittal midline and are located near the coronal plane that is through the estimated center of gravity of the head (Vital and Senegas, 1986; Preston, 2005). The zone between the OC is the midline of the cranium from which growth radiates (Frankel, 1980). Therefore, the OC are an appropriate reference for longitudinal growth from the inception of cranial development. Similar to the articular surfaces of all bones formed from cartilage analage, the inferior margin of the OC cannot grow by surface apposition because of the high pressure gradient associated with the weight of the head. Growth against a pressure gradient requires an epiphyseal growth plate composed of hyaline cartilage (Baume, 1968; Roberts et al., 2004). Therefore, I-point and Ic are recommended as relatively stable landmarks for longitudinal reference as the head utilizes the OC as the foundation supporting growth of the craniofacial complex.

The results of this study are similar to previous authors Bergersen (Bergersen, 1966), Broadbent (Broadbent, 1937), Coben (Coben, 1961, 1998), Frankel (Frankel, 1980), Kanomi (K point), and Ricketts (Ricketts, 1975, 1976). Cephalad growth of sella turcica is observed, because they all incorporate the posterior cranial base in their analyses. Coben described the growth at the SOS as being ventral and cephalad due to the orientation of the SOS (Coben, 1998). The present study corroborates evidence that caudad and ventral facial growth is compounded by cephalad and ventral displacement of the ACB. As sella turcica moves cephalad, there is a pattern more consistent with the rotational facial growth observed by Bjork and Skeiller (Bjork, 1955b, 1969; Bjork and Skieller, 1983; Skieller et al., 1984). Rotational pattern associated with I-point/Ic reference can be seen in Fig. 4a, when compared with the traditional SACB, Fig. 4b.

If the OC are an axis of growth, a center of gravity located near the coronal plane through the OC (Vital and Senegas, 1986; Lieberman et al., 2000; Preston, 2005) is important if the head displays centripetal growth. The center of mass for the head has been estimated slightly forward of the OC, however, center of mass may be more difficult to ascertain than originally thought, because of postmortem swelling and/or shrinkage artifacts introduced during brain fixation (Quester and Schroder, 1997; Courchesne, et al., 2000).

Courchesne (Courchesne et al., 2000) utilized MRI technology to calculate brain volume and weight and found that their estimates were 4% greater than postmortem values, and postulated the difference may be that, in the living brain, 2–3% of its volume is blood, most of which is in the gray matter capillaries. Therefore, the gray matter trend observed by Giedd (Giedd et al., 1999; Gogtay et al., 2004; Lenroot and Giedd, 2006) gives credence to a rotational pattern of the brain and crania (Bjork, 1955a; Giedd et al., 1999). The centripetal growth of the crania (Ranly, 1980) around the OC reflects the pattern of brain growth relative to the brainstem, (Ranly, 1980; Lieberman et al., 2000) which in combination with airorhynchy (Lieberman et al., 2000) is a continuation of the embryologic pattern of head “unfolding” associated with facial development. The increase in facial bulk with growth is offset by progressive pneumatization of the facial bones allowing balance to be maintained.

Counter-clockwise rotation of the brain and crania is masked by the selection of reference structures in this study but alluded to when considering functional anisotropy of brain growth, airorhynchy, and the centripetal growth at the SOS (Baume, 1961; Melsen, 1971, 1974; Vilmann et al., 1980; Giedd et al., 1999; Lieberman et al., 2000; McCarthy and Lieberman, 2001; Gogtay et al., 2004; Evans, 2006). Giedd found that, changes in cortical gray matter were regionally specific, with developmental curves for the frontal and parietal lobe peaking at about age 12, the temporal lobe at about age 16, and through 20 years of age for the occipital lobe (Giedd et al., 1999). Interestingly, the frontal and parietal gray matter peak approximately 1 year earlier in females, corresponding with the earlier age of onset of puberty which suggests a possible influence of gonadal hormones (Giedd et al., 1999). This later brain growth may have particular relevance to the work of Behrents (Behrents, 1985). The temporal lobe increase is consistent with observation that temporal lobe and middle fossa continue to enlarge for several more years than the frontal lobe and anterior cranial fossa (Enlow and Hans, 1996). Sella turcica, including the anterior curvature, is not a stable landmark to be used for superimposition due to observed circumpubertal apposition (Melsen, 1974; Ghafari et al., 1987).

Sella turcica is expected to move rather than remain static as with the SACB analysis. Nasion grows more consistently with the ACB and face, and the sella-nasion line to cribriform plate is a relatively stable relationship (Bjork, 1955a). The variability of nasion with growth, while maintaining a stable relationship to ethmoidale, is due to cranial rotation. As well, a spurt in brain growth may uncouple the paced rotation of sella turcica with cranial rotation due to the rapid change at the hypophysis (pituitary gland). This separation causes the deposition seen at sella turcica (Melsen, 1974; Enlow and Hans, 1996).

Airway is disregarded with traditional superimposition on the SACB, (Buschang and Santos-Pinto, 1998) creates a relative dorsal direction of mandibular growth. This growth direction carries the posterior of the tongue closer to the posterior pharyngeal wall which tends to restrict the airway, theoretically leading to airway obstruction, without a compensatory neuromuscular response (Buschang and Santos-Pinto, 1998). The lingual tonsil (part of Waldeyer's Ring located in the base of the tongue posterior to foramen cecum) enlargement associated other lymphoid tissues (nasopharyngeal adenoids, palatine tonsil) (Tourne, 1991) must be considered when observing the tongue position relative to mandibular growth rotation. Superimposition at I-point elucidates growth rotation and the vertical ascent of sella turcica (Figs. 3 and 4). IS displays less caudad movement for OC-ACB or thought of another way, displays greater cephalad movement through forward mandibular rotation. The vertical ascent is important to show that any impingement on airway by posterior condylar and posterior ramal displacement is compensated by the movement of sella. Some posterior condylar movement would be expected especially with forward condylar growth and a counter-clockwise rotation of the mandible (Standerwick and Roberts, (Submitted for publication)).

This study lends support to an aponeurotic tension model of craniofacial growth (Standerwick and Roberts, (Submitted for publication)), which describes vertical development of the crania as a result of brain growth. This vertical development places pressure on the galea aponeurotica or epicranial aponeurosis. Tension on the aponeurosis is transmitted through the CFMAS and creates the rotational patterns observed in the maxilla and mandible. The effect of brain growth is superimposed with cranial rotation (Bjork, 1955a), airorhynchy (Lieberman et al., 2000; McCarthy and Lieberman, 2001) and is a continuation of the embryonic growth pattern as a result of brain flexure.

Registration at SACB is simple, convenient, and suitable for diagnosis; however, superimposition for treatment and growth evaluation skews the proportional changes. Critical evaluation of the traditional view of relative growth proportion (2 year old child cranial to facial proportion of 6:1 versus adult 2 or 3:1) displays that proportional growth is better maintained than previously thought (3.7:1 versus 2.2:1) (Trenouth and Joshi, 2006). Registration at OC is simple, convenient, and better suited for evaluation of treatment and growth change because the physiologic pattern of growth is better displayed. With advancements in radiographic enhancement and cone beam computed tomography (CBCT), the condyle is easier to locate and is therefore recommended as a new gold standard for longitudinal superimposition. With CBCT, a reference based solely on the OC and Ic can be applied. This would allow the full effect of cranial rotation to be observed, which was obscured by the current use of the ACB in parallel. Proximity of landmarks points is an issue; the closer a reference is to its orientation point, the greater will be the effect of small errors in landmark placement and superimposition. To create a distant orientation point, still referenced solely on the OC, it is suggested that CBCT be used. A transverse line can be drawn through I point or SIA on each occipital condyle and then ventrally to the superior mid-sagittal surface of the SOS. The expected accuracy of point location should reduce the error caused by close proximity of landmarks.

It is recommended that I-point be combined with structural (best fit) superimposition of Ic, which is expected to decrease the variability in landmark location (Baumrind et al., 1976).


The OC, registered at I point and I curve, are simple and convenient to locate accurately. They are better suited for evaluation of treatment and growth change than registration at sella turcica, as growth patterns more similar to in the Bjork implant studies are evident. Advancement of this superimposition method is recommended with the emergence of cone beam CT technology. This method warrants further investigation to improve the technique.

The findings of this project lend support to an alternate model of craniofacial growth: the aponeurotic tension model of craniofacial growth.


We thank Dr. J. Baldwin and Dr. J. Shanks for their contribution to the committee in preparation of this manuscript; Dr. Tom Southard and the Iowa University Orthodontic Department for the use of the Iowa Growth Study material, George Eckert for statistical analysis, and Dolphin Imaging for generating the I point landmark for use with the software.