CBCT Ceph Atlas

Rule (Andrews, Six Elements, Element II; sourced via Wu et al. 2023 and the Andrews 2008/2015 papers). The Goal Anterior Limit Line is a true vertical in natural head position. It passes through the forehead's FFA point (the midpoint of the clinical forehead) when the forehead inclination is ≤ 7°; for every degree beyond 7° it moves 0.6 mm forward, but never beyond glabella. The FA point of the maxillary central incisor should sit on it.

Applied here. Forehead inclination ° (scanner vertical) / ° (Frankfort frame). The adjustment ( mm) exceeds the FFA-to-glabella gap ( mm), so the cap applies: the GALL is the glabellar vertical, mm anterior of nasion. The cap is reached at any inclination above 18°, so the GALL does not depend on the forehead being clipped by the field of view above mm (a trichion-based inclination of 25–29° would move it by at most ~1 mm).

Considered and rejected

• Mandibular setback. To close the bite alone it takes B 5.7 mm and pogonion 5.1 mm back — over the airway limit — and leaves the midface flat and the face no longer.
• Chin advancement. The chin is already the most forward point of the face.
• Le Fort II / III. Over-projects the upper midface (orbital rim to +1.6–2.7 SD) with a 10 mm nasofrontal step; the rims are not truly deficient — a 3 mm onlay covers them.
• High (quadrangular) Le Fort I. Subpupil area overshoots; reported vertical relapse 62–158 %.
• Rotation-only plans (downgraft without advance). Chin 4–7 mm back, bite still Class III, airway worse.
• Inverted-L ramus lengthening. Flattens an already flat jaw plane on an already long ramus.
• Malar implants. Cheekbone projection is already at the norm.

Before anything is fixed

• Natural head position — one profile photograph with a plumb line. The advance is right for ~4° chin-down; at a level Frankfort posture it should drop to 4.5–5 mm.
• A scan after expansion — whatever the expander actually drops and drifts the maxilla changes the Le Fort I by several millimetres; totals stay the same.
• A sleep study — moderate or severe sleep apnoea would favour a larger bimaxillary advance at the cost of the chin and incisor targets.

Sequence: expansion (≈1 month active, 6 months retention) → orthodontics 9–14 months → one operation → orthodontics 6–9 months → stage-2 implants and nose 6–12 months later. About 2–2.5 years.

Why it stops there

• The missing height is above the jaws. Midface (soft nasion to subnasale) and upper-lip length are the short segments; no jaw operation lengthens either, and the lip gains about 1 mm.
• Lip seal is the binding limit. Each extra millimetre of downgraft parts the lips about 0.6 mm; the stack sits at a 3.3 mm gap, borderline at the worst soft-tissue bound, settling to 2.4.
• Tooth show is capped at about 4 mm; more downgraft needs more intrusion than orthodontics delivers.
• The chin is at its norm at 7 mm; more makes it long.
• The ramus cannot follow a larger parallel downgraft without a 4 mm step or a stretched sling.
• About a quarter of the gain is the lips parting rather than tissue, and every millimetre lands below the nose, so the upper-to-lower third ratio moves from 0.89 to 0.80.

The index against several male norms

Part of the gap is definition: measured from bony-nasion level the same face reads 88.7 at surgery. The rest is soft cheekbone width (144.8 mm), not height. Leanness (≈3 mm less soft width) would add ≈1.7 index points; zygomatic reduction was rejected because it narrows the cheekbones the plan keeps.

Screw opening is not bone. Meta-analyses of bone-borne expanders in late adolescents and adults: the suture splits in ≈93 % of cases; the gain is ≈49 % basal bone, ≈8 % alveolar and ≈43 % tooth tipping, with a typical basal gain of 2.3 mm (1.7–3.0). Surgically assisted expansion in adults gives ≈7.5 mm at the molars with ≈3.5 mm skeletal and ≈30 % dental relapse. Ten millimetres of activation therefore puts roughly five into the basal bone; closing the full transverse deficit may need over-expansion and a second step.

What happens to the rest of the face depends on how far the maxilla drops and drifts while it widens. With the mandible left to find its own bite, every millimetre of drop props the jaw open on the back teeth, swinging the chin down and back. That looks like progress — a longer face, a smaller Class III — but the front teeth stop meeting and the tongue base moves back. The airway limit here is 2 mm.

Pipeline

• Volume: 668×668×445 voxels, 0.45 mm isotropic, HU-calibrated. Bone threshold 400 HU, skin −450 HU.
• Landmarking: for each of four regions, two agents located every landmark independently — one by reading thin slices in three planes, one algorithmically from the volume — without access to any earlier result.
• Verification: one adversarial agent per landmark re-rendered both candidates in three planes, re-derived the point from its definition, and chose or corrected it; each verifier's note records the disagreement between the two candidates and what structure the final point sits on.
• Frame: Frankfort plane by least squares through both porions and orbitale (residuals ≤ 0.2 mm); midsagittal plane perpendicular to FH through nasion and sella.
• Norms: adult-male means and SDs with sources (Steiner, McNamara, Ricketts, Holdaway, Jarabak/Björk, Tweed, Legan–Burstone, Arnett STCA, Farkas, Penn, Andrews). Where only pooled or range values exist, the table says so and gives no z.
• Checks: two independent agents re-derived the frame, every coordinate and every headline number from the landmark voxels. Round 1 found eleven issues (occlusal-plane contact point, an alar landmark definition mismatch, unverifiable SDs for U1-NA/L1-NB and the Jarabak ratio, a missing Ricketts row, the Holdaway chin-thickness definition, the U1-to-A-vertical measuring point, the provisional facial axis, the facial-index classification, a yaw sign and an overbite sign in notes); all were corrected in round 2. A third pass checked this page against the analysis file.

Corrections versus the first version of this study

• Arnett midface norms: the first version compared the orbital rim, cheekbone and subpupil points with the female norms (−18.7, −20.6, −14.8). The male norms are −22.4, −25.2, −18.4. The "5–9 mm infraorbital deficit" shrinks to 2–3 mm at the rim (about 1 SD) and ~4 mm under the pupil (4.6 R / 3.9 L); the cheekbone point reads normal but sits 5 mm from the subpupil point on the same slope, so the honest summary is one infraorbital reading 1–2 SD behind.
• Pogonion was 2.8 mm too posterior; nasion is now on the frontonasal suture, 4.5 mm above the concavity point used before; zygion was on temporal bone (bony Zy–Zy 132.1, not 141); overjet/overbite are 0.6 / −0.6 (edge-to-edge), not 1.5 / 3.5; chin pad is 14–15 mm (normal), not 17; nasolabial angle ≈ 65° (range 59–73 by construction), not 97°.
• Ramus: Ar–Go 58 mm is +2.8 SD on Björk's norm — long, not "normal". Gonial angle Co-Go-Me 121°, not 117°; Ar-Go-Me 128° (norm 130 ± 6, normal).
• Diagnosis moves from "mandible normal, maxilla neutral" to an 8–12 mm Class III jaw discrepancy: mildly retrusive maxilla (SNA 79.4, A −1.6 to N-perp, −10.9 to the GALL where a balanced male sits at −7 to −10), a mandible of normal length (Co–Gn 130 mm ×1.08 vs 134 ± 7) rotated forward by a horizontal growth pattern (FMA 14.5°, Björk sum 383°, S-Go:N-Me 73 %) so that the chin reads +11.8 mm to N-perp in the Frankfort frame, ≈ +4.8 mm at the scanned pitch, and +2.5 mm to the GALL where a balanced male sits at ≈ −9; brachyfacial; maxillary transverse deficiency; narrow airway. Whether the chin is called "prognathic" depends on the reference; on every reference with a norm it is 4–12 mm ahead, and the jaw discrepancy does not change.
• Remaining low-confidence items: columella point (conf 0.5, ±2 mm along the columella, so the nasolabial angle is approximate), Pt point (facial axis provisional), antegonial notches (absent), left alar base (alare landmarked, Arnett norm not applied), thin bone (ANS, sella floor) not fully rendered in the 3D model.