The Spine Journal 14 (2014) 2425–2433

Clinical Study

Axial plane analysis of Lenke 1A adolescent idiopathic scoliosis as an aid

to identify curve characteristics
Halil Atmaca, MDa, Mustafa Erkan Inanmaz, MDb, Emre Bal, MDb, Islam Caliskan, MDb,
Kamil Cagri Kose, MDb,*
a
Department of Orthopedics and Traumatology, Faculty of Medicine, Akdeniz University, Dumlupinar Avenue, 07058, Konyaalti, Antalya, Turkey
b
Department of Orthopedics and Traumatology, Faculty of Medicine, Sakarya University, Adnan Menderes street, 54100, Sakarya, Turkey
Received 6 June 2013; revised 27 January 2014; accepted 3 February 2014

Abstract BACKGROUND CONTEXT: Adolescent idiopathic scoliosis (AIS) is a complex three-

dimensional (3D) deformity of the spine involving deviations in the frontal plane, modifications
of the sagittal profile, and rotations in the transverse plane. Although Lenke classification system
is based on 2D radiographs and includes sagittal thoracic and coronal lumbar modifiers, Lenke
et al. suggested inclusion of axial thoracic and lumbar modifiers in the analysis.
PURPOSE: To analyze axial plane of Lenke 1A curves to identify curve characteristics.
STUDY DESIGN: Retrospective study.
PATIENT SAMPLE: Seventy patients (49 women, 21 men) with Lenke Type 1A idiopathic sco-
liosis were analyzed.
OUTCOME MEASURES: Coronal, sagittal, and axial parameters were measured from plain ra-
diographs that were obtained at initial medical examination of the patients.
METHODS: Coronal and sagittal plane and whole spine segmental vertebra rotations from thora-
cic 1 to lumbar 5 were evaluated in 70 AIS patients with Lenke 1A curves by using Drerup method.
Three different subgroups were identified according to magnitude and direction of lower end ver-
tebra (LEV) rotation.
RESULTS: In Group 1 (Lenke 1A1), the direction of LEV rotation was same with other vertebrae in the
main curve and the magnitude of the LEV rotation was less than
0.5 . In Group 2 (Lenke 1A2), the
rotation of LEV was between
0.5 and 0.5 and so was accepted as neutral. In Group 3 (Lenke
1A3), the rotation of LEV had opposite direction with vertebrae in the main curve and the magnitude
of LEV rotation was more than 0.5 . The mean thoracic Cobb angle of patients with Lenke 1A idiopathic
scoliosis was 51.1 (range 37 –80 ), whereas the mean lumbar Cobb angle was 16.4 (range 0 –32 ).
The mean angle of trunk rotation of the patients was 5.7 (range 1 –16 ). In terms of maximum thoracic
vertebra rotation, the mean rotation angle of Lenke 1A idiopathic curves was
18.9 (range
(9.8 –
44.7 )). The mean maximum lumbar vertebra rotation was 4.5 (range
7.2 to 15.1 ).
CONCLUSIONS: Addition of axial plane analysis to conventional coronal and sagittal evalua-
tions in patients with Lenke 1A curves may reveal inherent structural differences that are not ap-
parent in single planar radiographic assessments and may necessitate a different surgical
strategy. Ó 2014 Elsevier Inc. All rights reserved.
Keywords: Lenke 1A; Idiopathic scoliosis; Lower end vertebra; Axial plane analysis; Vertebra rotation; Trunk rotation

FDA device/drug status: Not applicable. Introduction

Author disclosures: HA: Nothing to disclose. MEI: Nothing to disclose.
EB: Nothing to disclose. IC: Nothing to disclose. KCK: Nothing to disclose. Adolescent idiopathic scoliosis (AIS) is a complex three-
Conflict of interest: No benefits in any form have been received or will dimensional (3D) deformity of the spine involving devia-
be received from a commercial party related directly or indirectly to the sub- tions in the frontal plane, modifications of the sagittal
ject of this article. profile, and rotations in the transverse plane [1–3]. Because
* Corresponding author. Department of Orthopedics and Traumatology,
Faculty of Medicine, Sakarya University, Adnan Menderes street, 54100,
of this altered morphology, axial plane analysis became
Sakarya, Turkey. Tel.: (90) 505-403-22-72; fax: (90) 264-275-91-92. popular [4–9]. Because Cobb [10] and Nash and Moe [11]
E-mail address: kacako@hotmail.com (K.C. Kose) presented documented methods for the measurement of
1529-9430/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.spinee.2014.02.015
Downloaded for Anonymous User (n/a) at Saint Joseph University from ClinicalKey.com by Elsevier on February 15,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
2426 H. Atmaca et al. / The Spine Journal 14 (2014) 2425–2433
axial vertebral rotation in coronal radiographic images,
studies aimed to evaluate vertebra rotation using plain radio-
graphs [4,5,7,9,12,13], 3D bone models [8,14], rasterster-
eography [15–17], and computed tomography (CT) [18–21].
Although Lenke classification system is based on 2D ra-
diographs and includes sagittal thoracic and coronal lumbar
modifiers, Lenke et al. suggested inclusion of axial thoracic
and lumbar modifiers in the analysis [22]. Notwithstanding
all previous studies had sufficient data regarding vertebra
rotation in patients with AIS, none of them mentioned ver-
tebral rotation with respect to the curve type in AIS.
The Lenke classification consists of six curve types, a
lumbar spine modifier, and a sagittal thoracic modifier. The
Lenke 1A curve was reported to be the most frequently eval-
uated curve type with a range of 25% of adolescent idiopathic
curves [22]. According to our hypotheses, Lenke 1A curves
differ from each other with respect to magnitude of segmen-
tal rotation of vertebrae although curves in this group have
similar coronal and sagittal curve characteristics. Also these
rotational differences may give ideas to surgeons whether
requirement of different surgical strategies to avoid early
and/or late postoperative complications. In this study, axial
plane of Lenke 1A curves were analyzed by using Drerup
method [12,13] to identify curve characteristics that would
help determine optimum surgical correction strategies.
Materials and methods
A retrospective review of radiographs of patients with
AIS at a single center was performed. Seventy patients
who had Lenke Type 1A curves with Cobb angles between
30 and 80 were included in the present study. Patients
with nonidiopathic or other types of idiopathic scoliosis,
patients older than 40 years, and patients who had a pre-
vious surgery were excluded.
X-ray and digital technique
The anteroposterior standing whole spine radiograph was
taken with the tube 1.5 m from the patient with the arms lying
comfortably at the sides. The hips were in slight external ro-
tation with the big toe pointing 15 outward. The heels were
20 cm apart. The lateral radiograph was taken with the tube
2 m from the patient whose arms were folded in front. The
radiograms were scanned to a computer workstation by
using a transparent media scanner (Mikrotek Scan Maker
9800XL Plus, MikrotekÒ, Hsinchu, Taiwan). Digital soft-
ware (Canvas 9.0, ACD Systems International Inc, Miami,
Florida, USA) was used for analyzing the measurements.
Curve classification
All curves were classified as Lenke Type 1A according
to the Lenke classification system; Type 1 is the main thora-
cic curve, and the proximal thoracic and thoracolumbar/
lumbar curves are minor nonstructural curves. The lumbar
Downloaded for Anonymous User (n/a) at Saint Joseph University from ClinicalKey.com by Elsevier on February 15,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
modifier A is used when the center sacral vertical line runs
between the lumbar pedicles and the level of the stable ver-
tebra (SV). The curve must have a thoracic apex at or ceph-
alad to the eleventh and twelfth thoracic disc levels.
Therefore, modifier A can be used only for a main thoracic
curve. It should also not be used when the center sacral ver-
tical line falls directly against the medial aspect of the lum-
bar apical pedicle [22].
Clinical measurement and radiological analysis of
rotation
The angle of trunk rotation (ATR) of all patients were
assessed at our department using the same scoliometer (Or-
thopedic Systems, Inc., Hayward, CA, USA). The patients
were instructed to bend forward, standing with their feet to-
gether, their knees straight, and their arms dependent with
hands together, palms and fingers in opposition. The exam-
iner obtained scoliometer measurements over the promi-
nence of the curve in the thoracic regions.
The radiographic measurements included coronal Cobb
angle of the structural curve (thoracic curve), the nonstruc-
tural curve (lumbar curve), and the T5–T12 thoracic sagit-
tal Cobb angle (kyphosis angle). Also the upper end
vertebrae (UEV) and lower end vertebra (LEV), number
of vertebrae in the main curve, and the apical region of
the thoracic curve were noted. The degree of vertebral rota-
tion was measured twice at each level from T1 to L5 inde-
pendently by two authors (HA, MEI) at the different time
periods using the technique described subsequently, and
the mean values were accepted.
Axial plane analysis was performed with the method de-
scribed by Drerup [12,13]. Drerup improved the Nash-Moe
method by modifying the measurement of the position of
the anatomic landmarks, that is, the projections of vertebral
pedicles. By using known predefined vertebral shape param-
eters, namely, the distance from the pedicle to the vertebral
body center, a trigonometrical model was used to measure
the axial vertebral rotation [23]. According to the Drerup
method, the width of the vertebral body (d) was measured
and the midpoint of vertebra was detected (d/2). Then the
medial borders of both pedicles were marked, and a line
was drawn from this point perpendicularly to line ‘‘d.’’
The distance from the medial border of the convex-sided
pedicle and the point of d/2 was measured as ‘‘S1,’’ and
the distance between the medial border of the concave-
sided pedicle and the point of d/2 was measured as ‘‘S2’’
(Fig. 1). All measures were given in the same unit, that is,
either millimeter or centimeter. In general, that is, if rotation
angles were not too small, S1 was positive and S2 was neg-
ative. Microsoft Office Excel 2007 was used for mathemat-
ical application of the Drerup formula. All measured data
were written to defined columns in Excel to find vertebra ro-
tations per level from T1 to L5, and also rotational curve of
the patients was obtained by using scatter chart tool of Mi-
crosoft Office Excel 2007. Rotation of the anterior part of
 H. Atmaca et al. / The Spine Journal 14 (2014) 2425–2433
Fig. 1. Measurement of vertebra rotation from standing radiographs. ‘‘d’’
(blue line) indicates the width of the vertebral body. Yellow line indicates
the midpoint of vertebra (d/2). The medial borders of both pedicles were
marked with red lines pendicularly to blue line. The distance from the me-
dial border of the convex-sided pedicle and the point of d/2 was measured
as ‘‘S1,’’ and the distance between the medial border of the concave-sided
pedicle and the point of d/2 was measured as ‘‘S2.’’
the vertebrae to the right side of the patient (anticlockwise
rotation) was defined as a negative angle, to the left clock-
wise rotation as a positive angle. The vertebrae with maxi-
mum rotation angles in thoracic and lumbar regions were
called maximum rotated vertebra—thoracic (MVR-T) and
maximum rotated vertebra—lumbar (MVR-L), respectively.
In critically analyzing Lenke 1A curves in the axial
plane, patients were divided into three groups depending
on LEV rotation. In Group 1 (Lenke 1A1), the direction
of LEV rotation was same with other vertebrae in the main
curve and the magnitude of the LEV rotation was less than

0.5 . In Group 2 (Lenke 1A2), the rotation of LEV was
between
0.5 and 0.5 and so was accepted as neutral.
In Group 3 (Lenke 1A3), the rotation of LEV had opposite
direction with vertebrae in the main curve and the magni-
tude of LEV rotation was more than 0.5 (Fig. 2).
Rotational curve reliability
All the rotational curve graphics were interpreted on 1
day and then reinterpreted 3 days later, in a different se-
quence. The two groups of reviewers were asked to choose
the appropriate curve type (1A1 through 1A3). The first
Downloaded for Anonymous User (n/a) at Saint Joseph University from ClinicalKey.com by Elsevier on February 15,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
2427
group consisted of three authors and one independent radiol-
ogist. The members of other group were randomly selected
from other orthopedic specialists in the same city. Inter-
and intraobserver reliabilities were estimated by calculating
the kappa coefficient values. The kappa value was the bal-
ance of the part of agreement that would occur by random
chance subtracted from the actual agreement. Thus, kappa
coefficients ranged from þ1 (perfect agreement), to 0
(chance agreement), to
1 (less agreement than expected
by chance). Svanholm et al. [24] suggested that kappa values
of more than 0.75 represent good or excellent reliability, 0.5
to 0.75 fair reliability, and less than 0.5 poor reliability.
Statistical analyses
Mean and standard deviation were calculated for de-
scriptive statistics of continuous variables and median val-
ues for discrete variables. Kolmogorov-Smirnov test was
used to analyze the normality of data. The means of groups
were analyzed using analysis of variance and then Bonfer-
roni post hoc test. Pearson correlation coefficient was used
to analyze whether significant correlation exists between
the parameters. Inter- and intraobserver rotational curve re-
liability was analyzed by cross-table test, and Kappa coef-
ficient was obtained. Two-tailed hypothesis was considered
in the analyses, and the significant differences were accep-
ted if p value was .05 or less. SPSS 15.0 software for Win-
dows (SPSS, Inc., Chicago, IL, USA) was used in the
evaluation of statistical analyses.
Results
A total of 70 patients (49 women, 21 men) with Lenke
1A idiopathic scoliosis were evaluated. The mean age of
the patients was 16.6 (range 11–31) years. Upper end ver-
tebra was T3 in 16, T4 in 32, and T5 in 22 patients. LEV
was T11 in 12, T12 in 20, L1 in 22, L2 in 9, and L3 in 7
patients. Thoracic apical vertebra was T7 in 6, T8 in 20,
T8–T9 disc in 9, T9 in 25, and T10 in 10 patients. With re-
spect to LEV rotation angle, patients were divided into
three groups. The demonstrative data of the patients in all
groups were given in Table 1.
Coronal plane analyses
The mean thoracic Cobb angle of patients with Lenke
1A idiopathic scoliosis was 51.1 (range 37 –80 ). The
mean Cobb angle was 52.1 (37.5 –68.7 ) in Group 1,
56.2 (44 –70.8 ) in Group 2, and 48.2 (37 –80 ) in Group
3. The differences were not significant (p5.054) (Table 1).
The mean lumbar Cobb angle of patients with Lenke 1A
idiopathic scoliosis was 16.4 (range 0 –32 ). The mean
Cobb angle was 13.5 (0 –27 ) in Group 1, 14.7 (7.5 –
24 ) in Group 2, and 22.8 (5 –32 ) in Group 3. There
was a statistically significant difference between the groups
(p5.008). Bonferroni test revealed that lumbar Cobb angle
2428 H. Atmaca et al. / The Spine Journal 14 (2014) 2425–2433

Fig. 2. Mean rotational curves of patients. (A) Patients with Lenke Type 1A1 curves, (B) patients with Lenke Type 1A2 curves, (C) patients with Lenke Type
1A3 curves, and (D) mean rotational curve of all patients with Lenke 1A adolescent idiopathic scoliosis.

was significantly higher in Group 3 compared with groups There was no statistically significant difference between
1 and 2 (p5.017 and .014, respectively) (Table 1). the groups (p5.137) (Table 1). In Lenke 1A curves, T5–
The mean number of vertebrae in the main curve was 7.1 T12 kyphosis angle was found to be significantly negative
(range 5–13), whereas it was 9.3 (7–13) in Group 1, 7.4 (6– correlated with thoracic Cobb angle (p5.026, r5
0.226).
9) in Group 2, and 5.9 (5–8) in Group 3. The number of ver-
tebra in the main curve was statistically significantly differ-
Axial plane analyses
ent in all groups (p5.000) (Table 1).
The mean ATR of the patients with Lenke 1A idiopathic
Sagittal plane analyses scoliosis was 5.7 (range 1 –16 ). The mean ATR was 3.7
(2 –6 ) in Group 1, 8 (4 –14 ) in Group 2, and 5.5 (1 –
The mean T5–T12 kyphosis angle of patients with 16 ) in Group 3. There was a statistically significant differ-
Lenke 1A idiopathic scoliosis was 24.2 (range 7 –60 ). ence between the groups with respect to ATR (p5.000)
The mean kyphosis angle was 19.6 (12 –25 ) in Group (Table 1). Angle of trunk rotation was significantly higher
1, 26.7 (8 –51 ) in Group 2, 25.2 (7 –60 ) in Group 3. in Group 2 compared with groups 1 and 3 (p5.000 and
Table 1 .02, respectively). The difference between Group 1 and
Demonstrative data of the patients Group 3 was not significant (p5.147). In Lenke 1A curves,
Group 1, Group 2, Group 3, Total, ATR was positively correlated with the thoracic Cobb angle
Mean N517 N517 N536 N570 p (p5.000, r50.720).
Age (y) 15.1 16.7 17.3 16.6 .149 A total of 1,161 vertebrae were measured from T1 to L5
Gender (F/M) 17/0 11/6 21/15 49/21 .006* on plain radiographs of 70 patients with Lenke 1A idio-
MVR-T
14.7
24.7
18.1
18.9 .002* pathic scoliosis. Mean values of axial plane rotation of ver-
MVR-L
1.1 2.8 8 4.5 .000* tebrae per level according to groups were given in Table 2.
UEV rotation (  ) 2.8 6.9
0.1 2.3 .003* In terms of MVR-T, the mean rotation angle of Lenke 1A
LEV rotation (  )
1.6
0.1 4.5 1.9 .000*
ATR (  ) 3.7 8 5.5 5.7 .000* idiopathic curves was
18.9 (range
9.8 to 44.7 ) in
Thoracic Cobb angle (  ) 52.1 56.2 48.2 51.1 .054 the whole study group. The mean MVR-T was
14.7
Lomber Cobb angle (  ) 13.5 22.8 14.7 16.4 .008* (10.1 –20.5 ) in Group 1,
24.7 (
16.2 to 37.2 ) in
T5–T12 Kyphosis 19.6 26.7 25.2 24.2 .137 Group 2, and
18.1 (
9.8 to 44.7 ) in Group 3. There
angle (  )
was a statistically significant difference between the groups
Number of vertebrae in 9.3 7.4 5.9 7.1 .000*
main curve with respect to MVR-T (p5.002) (Table 1). maximum ro-
tated vertebra—thoracic was significantly higher in Group
ATR, angle of trunk rotation; F, female; LEV, lower end vertebra; M,
male; MVR-L, maximum rotated vertebra-thoracic; MVR-T, maximum ro- 2 compared with groups 1 and 3 (p5.001 and .019, respec-
tated vertebra-lumbar; UEV, upper end vertebra. tively). There was no significant difference between Group
*
Significant differences were accepted if p#.05. 1 and Group 3 (p5.461). In Lenke 1A curves, MVR-T had

Downloaded for Anonymous User (n/a) at Saint Joseph University from ClinicalKey.com by Elsevier on February 15,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
 H. Atmaca et al. / The Spine Journal 14 (2014) 2425–2433 2429

Table 2 Table 4
Mean vertebra rotations of all patients per level according to grouped data Interobserver reliability of curve classification by five independent
Level Group 1 Group 2 Group 3 Total reviewers

Th1 2.2 2.2 2.4 2.3 Curve type Kappa coefficient

Th2 4.1 8.5 3.5 4.9 1A-1 0.72
Th3 6.3 8.6 5.7 6.6 1A-2 0.59
Th4 4.2 6.9 2.6 4.0 1A-3 0.51
Th5
2.2 3.2
1.5
0.5 Mean 0.61
Th6
4.8
10.9
5.8
6.8
Th7
10.0
21.3
12.1
13.8
Th8
12.5
23.7
14.9
16.5 of the Lenke 1A curves have proximal thoracic curves in
Th9
12.0
20.3
13.0
14.5 the opposite direction with the main thoracic curve. The

11.2
11.2
7.4
9.2
Th10
mean LEV rotation angle was 1.9 (
4 to 15.1 ) in the
Th11
6.7
8.1
1.6
4.4
Th12
5.4
2.4 3.7 0.0 study group. The mean LEV in was
1.6 (
4 to 0.7 )
L1
3.0 0.9 5.8 2.5 in Group 1,
0.1 (
0.2 to 0.1 ) in Group 2, and 4.5
L2 0.1 3.2 6.5 4.2 (0.6 –15.1 ) in Group 3. The differences were significant
L3 0.8 1.6 4.3 2.8 with respect to LEV (p5.000) (Table 1).
L4 0.1 0.3 0.6 0.1
L5 0.2 0.1 0.1 0.2
Note: (
) values indicate the direction of rotation with anticlockwise. Rotational curve classification
Among the three surgeons and one independent radiolog-
a significantly positive correlation with the thoracic Cobb ist who had observed the radiographs and rotational curve
angle (p5.000, r50.775). Maximum rotated vertebra— graphics, the mean interobserver reliability for determining
thoracic had a significantly positive correlation with ATR the curve type was 76.4% (range 65.7%–85.7%), with a
(p5.000, r50.831). mean kappa value of 0.64 (0.48–0.78), indicating fair reli-
The mean MVR-L of the patients with Lenke 1A idio- ability (Table 3). In the group of five reviewers who had
pathic scoliosis was 4.5 (range
7.2 to 15.1 ). The mean been randomly selected from other orthopedic specialists
MVR-L was
1.1 (
7.2 to 5.5 ) in Group 1, 2.8 (
2.4 in the same city; the mean kappa value for interobserver re-
to 7.6 ) in Group 2, and 8 (1.6 –15.1 ) in Group 3. The liability with the new classification was 0.61 (0.51–0.72)
differences between the groups with respect to MVR-L (Table 4). This value represents fair reliability except for
were significant (p5.000) (Table 1). Maximum rotated ver- the interobserver reliability for curve type. The intraob-
tebra—lumbar was significantly higher in Group 3 com- server reliability of the authors and independent reviewers
pared with groups 1 and 2 (p5.000 and .000, were good or excellent with a mean of 0.89 (0. 86–0.94)
respectively), and also, there was a statistically significant and 0.86 (0.83–0.90) kappa values, respectively (Table 5).
difference between Group 1 and Group 2 (p5.019). In
Lenke 1A curves, MVR-L had a significantly positive cor-
relation with the lumbar Cobb angle (p5.001, r50.373).
The mean UEV rotation angle was 2.3 (range
9.1 to Discussion
12.4 ) in the study group. The mean UEV was 2.8 (
5.9 Addition of axial plane analysis to conventional coronal
to 11.9 ) in Group 1, 6.9 (0 –12.4 ) in Group 2, and
0.1 and sagittal evaluations in patients with Lenke 1A curves
(
9.1 to 11.4 ) in Group 3. There was a statistically signif- may reveal inherent structural differences that are not appa-
icant difference between the groups with respect to UEV rent in single planar radiographic assessments and may ne-
(p5.003) (Table 1). Bonferroni test showed that this differ- cessitate different surgical strategy.
ence was dependent to a significant discrepancy between In this study, we have noted three different curve charac-
groups 2 and 3 (p5.002). These results indicated that most teristics with respect to LEV rotation. Lenke 1A1 curves
were including approximately 25% of all Lenke 1A curves
Table 3 with a relatively long segment main curve extending to L2
Interobserver reliability of curve classification by the four reviewers
Number Percentage of Table 5
of curves the curves Kappa Intraobserver reliability of curve classification by four reviewers and five
Reviewers classified same classified same (%) coefficient independent reviewers
1 and 2 60 85.7 0.78 Kappa coefficient
1 and 3 57 81.4 0.71
1 and 4 54 77.1 0.65 Curve type Authors Independent reviewers
2 and 3 54 77.1 0.65 1A-1 0.94 0.90
2 and 4 50 71.4 0.56 1A-2 0.88 0.85
3 and 4 46 65.7 0.48 1A-3 0.86 0.83
Mean 53.5 76.4 0.64 Mean 0.89 0.86

Downloaded for Anonymous User (n/a) at Saint Joseph University from ClinicalKey.com by Elsevier on February 15,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
2430 H. Atmaca et al. / The Spine Journal 14 (2014) 2425–2433
Fig. 3. Linear regression analysis. The correlation coefficient (r2) is 0.601
for MVR-T values versus thoracic Cobb angle values. (
) values for
MVR-T indicate the direction of rotation.
or L3. The magnitudes of ATR and MVR-T were signifi-
cantly lower in patients with Lenke 1A1 curves. The second
subgroup (Lenke 1A2) was as frequent as the first sub-
group. The patients with Lenke 1A2 curves had signifi-
cantly higher ATR and MVR-T magnitudes. The mean
Cobb angle was insignificantly higher in this group. LEV
was mostly T12–L1. The subgroup Lenke 1A3 was the
most common type (50%) that consisted of relatively small
number of vertebrae in the main curve (rotated to the same
direction). The LEV was usually T11–T12. Although the
mean Cobb angle of Lenke 1A3 curves was smaller than
Lenke 1A1 curves, patients in subgroup 1A3 had a rela-
tively higher ATR and MVR-T compared with subgroup
1A1.
Analysis of vertebral deformity in sagittal and coronal
planes from plain radiographs can be performed easily,
but the measurement of vertebral rotation from these radio-
graphs is more complex. Computed tomography has been
recognized as the most accurate method to determine verte-
bral rotation. But its high cost and inherently high radiation
dose preclude its use in sequential studies of adolescent pa-
tients [9]. An additional disadvantage of this method is that
it uses standing images to analyze sagittal or coronal de-
formities, whereas analysis of axial plane deformity by
CT is assessed in the supine position [25,26]. Yazici et al.
[9] aimed to assess the accuracy of vertebral rotation values
measured according to the Perdriolle method by comparing
them with those obtained on CT scans and to analyze the
influence of patient position on rotation measurements.
They reported that quantification of the transverse plane de-
formity is highly sensitive to the position of the patient. So
AIS should be evaluated on standing images for all three
planes.
Several methods for assessing vertebra rotation
using plain radiographs have been developed
[4,5,7,9,12,13,23,27–29]. Drerup improved the Nash-
Moe method that brought accuracy to approximately
65 in assessing vertebral rotation by modifying the
measurement of the position of anatomic landmarks, that
is, the projections of vertebral pedicles. By using known
Downloaded for Anonymous User (n/a) at Saint Joseph University from ClinicalKey.com by Elsevier on February 15,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
predefined vertebral shape parameters, namely, the angle
from the pedicle to the spinous process and the distance
from the pedicle to the vertebral body center, a trigono-
metrical model to measure axial vertebral rotation was
proposed [12,13,23]. The author stated that vertebral coro-
nal and sagittal tilt did not affect the rotation measure-
ments when they were referenced from the local
vertebral and not global radiographic coordinate system
[12,13].
Distal extent of fusion in Lenke 1A curves is still contro-
versial. Studies focusing on the adding-on phenomenon ob-
served with Lenke Type 1A curve have mostly investigated
the relationship between adding-on and the distal fusion
level [30–34]. Analyzing the coronal plane Lenke 1A
curves was further divided into groups. Depending on the
direction of L4 tilt, the Lenke Type 1A curve was subclas-
sified into 1AL and 1AR according to Miyanji et al. [35].
They suggested fusion to SV in L4 left tilt (1-AL) curves
similar to Lenke 1B curves, whereas generally a more distal
fusion was recommended in L4 right tilt (1-AR) curves
similar to King IV curves. In another study, authors classi-
fied Lenke Type 1A curve into four different types (A–D)
depending on the L3 and L4 vertebral tilts. They concluded
that the distal fusion level should be extended to at least
LEV-1 in Type 1A-A (neutral L3 and L4 vertebrae) and
Type 1A-D curves (the L4 vertebra is tilted to the left,
whereas the L3 vertebra is neutral) and that it should be ex-
tended to LEV in Type 1A-B (the L3 vertebra is tilted to the
right, whereas the L4 vertebra is neutral) and 1A-C (L3 and
L4 vertebrae are tilted to right) curves. They emphasized
the importance of the LEV as a reliable guide to determine
the distal fusion level in Lenke Type 1A curves [30].
Suk reported a high risk of postoperative adding on
when the distal fusion level was performed two or more
proximal to distal neutral vertebrae (NV) [31]. Determina-
tion of the optimal fusion level requires reliable apprecia-
tion of EV, NV, and SV. A previous study by Potter et al.
[36] has pointed out the low inter-intraobserver reliability
of NV selection. With contrast to the previous studies, the
present study clearly showed that detection of direction
and magnitude of LEV and/or NV was possible by using
the method of Drerup [12,13] on the available plain radio-
graphs without any necessity of additional scans such as CT
or magnetic resonance imaging that may cause extra costs,
radiation (for CT), and loss of time. Thus, it may help on
decision making the caudal extent of fusion to minimize
the risk of adding-on phenomenon.
Stokes reported that the magnitude of vertebral axial ro-
tation correlates with lateral deviation of vertebrae from
spinal axis, and the rotation is maximal near the curve apex
[37]. Xiong et al. [38] investigated the relation between
Cobb angle and vertebra rotation angle (VRA) in patients
with early scoliosis with Cobb angle between 0 and 30 .
They found a significant positive correlation between the
VRA and the Cobb angle. They claimed that a slight
VRA to the right is a common feature in the normal spine
 H. Atmaca et al. / The Spine Journal 14 (2014) 2425–2433 2431

Fig. 4. Mean intervertebral rotation of Lenke 1A curves per level. Intervertebral rotation was larger at T5–T6, T6–T7, and T9–T10 segments, whereas it was
smaller at T8–T9 and L2–L3 segments.

and that the VRA increases with progressive lateral devia-
tion of the spine. They concluded that the coronal plane de-
formity in early idiopathic scoliosis is accompanied and
probably coupled to vertebral rotation in the horizontal
plane. The present study showed statistically significant
positive correlation (p5.000, r250.601) between Cobb an-
gle and MVR-T in patients with Lenke 1A curves with
Cobb angle ranged 37 to 80 (Fig. 3). Maximum rotated
vertebra—thoracic was near the curve apex that was similar
to the findings of Stokes et al. [37].
Dubousset [39] also showed that the intervertebral axial
rotation reached its maximum at the ends of the curves and
its minimum at the apex. However, the simple finite ele-
ment models of only one case were used in the study, and
the method of measurement of the intervertebral deviation
was not described. Recently, Hattori et al. [8] reported that
the intervertebral rotation in the axial plane was smaller
near the apical region and larger near the junctional region.
The maximum intervertebral change at junctional region
was 12.6 (absolute value), and the minimum change at ap-
ical region was 0 in patients with AIS. Authors stated that
construction of bone models from CT images taken in su-
pine position was the main limitation of their study [8]. Ad-
ditionally, their study did not group the results according to
curve type because of small number of subjects. In the
present study, we investigated rotational discrepancies be-
tween two adjacent vertebrae in 70 patients with Lenke
1A AIS by using standing anteroposterior radiographs. As
a result of this evaluation, intervertebral rotation was larger
at T5–T6, T6–T7, and T9–T10 segments, whereas it was

Fig. 5. Linear regression analyses showing significant correlation with angle of trunk rotation (ATR) and either MVR-T or thoracic Cobb angle. (Left) The
correlation coefficient (r2) is 0.691 for ATR values versus MVR-T values. (
) values for MVR-T indicate the direction of rotation. (Right) The correlation
coefficient (r2) is 0.519 for ATR values versus thoracic Cobb angle values.

Downloaded for Anonymous User (n/a) at Saint Joseph University from ClinicalKey.com by Elsevier on February 15,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
2432 H. Atmaca et al. / The Spine Journal 14 (2014) 2425–2433
smaller at T8–T9 and L2–L3 segments (Fig. 4). Analysis of
the deformity in the standing position is a distinctive fea-
ture of our study.
With the increasing importance of cosmetic concerns,
rib hump might be an important element of body dissatis-
faction especially in adolescent patients with scoliotic de-
formity. Correction of the scoliotic deformity in coronal
and sagittal planes may not be enough for satisfactory re-
sults because of postoperative residual vertebral rotations
in the axial plane. The studies correlating vertebral rotation,
vertebral translation, and rib hump to Cobb angle reported
close correlation between these parameters [6,21,40]. In the
present study, ATR was found to be significantly correlated
with MVR-T and Cobb angle in Lenke 1A curves (Fig. 5).
Using direct vertebra rotation might be useful to restore ax-
ial plane to help decrease postoperative rib hump deformity
that may have a critical role on postoperative cosmetic
satisfaction.
The main limitation of our study is the evaluation of on-
ly the preoperative X-rays and lack of postoperative rota-
tional evaluation of Lenke 1A curves. In the postoperative
period, the anatomic landmarks cannot be localized be-
cause of the superposition of the implants, osteotomy lines,
and late fusion mass; thus, Drerup formula cannot be used
to analyze the postoperative X-rays. However, this techni-
que may use to evaluate whether adding-on phenomenon
or decompensation below the instrumented segments dur-
ing the follow-up period. In the future studies, the axial
plane analyses should be included in the final evaluation
of Lenke 1A curves. Our study is the first identifying Lenke
Type 1A curves in the axial plane using biplanar X-rays and
without exposing the patients to increased doses of
radiation.
Acknowledgment
The authors particularly emphasize their gratitude to
Prof. Dr. Burkhard Drerup (Univ. Klinikum M€unster Klinik
und Poliklinik f€
ur Technische Orthop€adie und Rehabilita-
tion, M€unster, Germany) for preparing the mathematical
formula and his ‘‘helpful comments and suggestions’’ dur-
ing the preparation of the article.
References
[1] Carpineta L, Labelle H. Evidence of three-dimensional variability in
scoliotic curves. Clin Orthop Relat Res 2003;412:139–48.
[2] Cruickshank JL, Koike M, Dickson RA. Curve patterns in idiopathic
scoliosis. A clinical and radiographic study. J Bone Joint Surg Br
1989;71:259–63.
[3] Perdriolle R, Vidal J. Morphology of scoliosis three-dimensional evo-
lution. Orthopedics 1987;10:909–15.
[4] Liu RW, Yaszay B, Glaser D, et al. A method for assessing axial ver-
tebral rotation based on differential rod curvature on the lateral radio-
graph. Spine 2012;37:E1120–5.
[5] Upasani VV, Chambers RC, Dalal AH, et al. Grading apical vertebral
rotation without a computed tomography scan: a clinically relevant
Downloaded for Anonymous User (n/a) at Saint Joseph University from ClinicalKey.com by Elsevier on February 15,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.
system based on the radiographic appearance of bilateral pedicle
screws. Spine 2009;34:1855–62.
[6] Kuklo TR, Potter BK, Lenke LG. Vertebral rotation and thoracic tor-
sion in adolescent idiopathic scoliosis: what is the best radiographic
correlate? J Spinal Disord Tech 2005;18:139–47.
[7] Chi WM, Cheng CW, Yeh WC, et al. Vertebral axial rotation meas-
urement method. Comput Methods Programs Biomed 2006;81:
8–17.
[8] Hattori T, Sakaura H, Iwasaki M, et al. In vivo three-dimensional seg-
mental analysis of adolescent idiopathic scoliosis. Eur Spine J
2011;20:1745–50.
[9] Yazici M, Acaroglu ER, Alanay A, et al. Measurement of vertebral
rotation in standing versus supine position in adolescent idiopathic
scoliosis. J Pediatr Orthop 2001;21:252–6.
[10] Cobb J. Outline for the study of scoliosis. Instr Course Lect 1948;5:
261–75.
[11] Nash C, Moe J. A study of vertebral rotation. J Bone Joint Surg Am
1969;51:223–9.
[12] Drerup B. Principles of measurement of vertebral rotation from fron-
tal projections of the pedicles. J Biomech 1984;17:923–35.
[13] Drerup B. Improvements in measuring vertebral rotation from the
projections of the pedicles. J Biomech 1984;18:369–78.
[14] Kadoury S, Labelle H. Classification of three-dimensional thoracic
deformities in adolescent idiopathic scoliosis from a multivariate
analysis. Eur Spine J 2012;21:40–9.
[15] Drerup B, Hierholzer E. Assessment of scoliotic deformity from back
shape asymmetry using an improved mathematical model. Clin Bio-
mech 1996;11:376–83.
[16] Hackenberg L, Hierholzer E, P€otzl W, et al. Rasterstereographic back
shape analysis in idiopathic scoliosis after anterior correction and fu-
sion. Clin Biomech 2003;18:1–8.
[17] Hackenberg L, Hierholzer E, P€otzl W, et al. Rasterstereographic back
shape analysis in idiopathicscoliosis after posterior correction and fu-
sion. Clin Biomech 2003;18:883–9.
[18] G€oçen S, Aksu MG, Baktiroglu L, et al. Evaluation of computed to-
mographic methods to measure vertebral rotation in adolescent idio-
pathic scoliosis: an intraobserver and interobserver analysis. J Spinal
Disord 1998;11:210–4.
[19] G€oçen S, Havitçioglu H, Alici E. A new method to measure vertebral
rotation from CT scans. Eur Spine J 1999;8:261–5.
[20] Krismer M, Sterzinger W, Haid C, et al. Axial rotation measurement
of scoliotic vertebrae by means of computed tomography scans.
Spine 1996;21:576–81.
[21] Aaro S, Dahlborn M. The longitudinal axis rotation of the apical ver-
tebra, the vertebral, spinal, and rib cage deformity in idiopathic sco-
liosis studied by computer tomography. Spine 1981;6:567–72.
[22] Lenke LG, Betz RR, Harms J, et al. Adolescent idiopathic scoliosis: a
new classification to determine extent of spinal arthrodesis. J Bone
Joint Surg Am 2001;83:1169–81.
[23] Vrtovec T, Pernus F, Likar B. A review of methods for quantitative
evaluation of axial vertebral rotation. Eur Spine J 2009;18:
1079–90.
[24] Svanholm H, Starklint H, Gundersen HJ, et al. Reproducibility of his-
tomorphologic diagnoses with special reference to the kappa statistic.
APMIS 1989;97:689–98.
[25] Russell GG, Raso VJ, Hill D, et al. Comparison of four computerized
methods for measuring vertebral rotation. Spine 1990;15:24–7.
[26] Skalli W, Lavaste F, Descrimes JL. Quantification of three-
dimensional vertebral rotations in scoliosis: what are the true values?
Spine 1995;20:546–53.
[27] Cimic M, Kovac V, Smiljanic I, et al. Relationship between clinical
contourometric measurements and vertebral rotation in adolescent
idiopathic scoliosis. Coll Antropol 2009;33:127–33.
[28] Kadoury S, Cheriet F, Beausejour M, et al. A three-dimensional ret-
rospective analysis of the evolution of spinal instrumentation for the
correction of adolescent idiopathic scoliosis. Eur Spine J 2009;18:
23–37.
 H. Atmaca et al. / The Spine Journal 14 (2014) 2425–2433 2433

[29] Pinheiro AP, Tanure MC, Oliveira AS. Validity and reliability of a
computer method to estimate vertebral axial rotation from digital ra-
diographs. Eur Spine J 2010;19:415–20.
[30] Sarlak AY, Atmaca H, Kim WJ, et al. Radiographic features of the
Lenke 1A curves to help to determine the optimum distal fusion level
selection. Spine 2011;36:1592–9.
[31] Suk SI, Lee SM, Chung ER, et al. Determination of distal fusion level
with segmental pedicle screw fixation in single thoracic idiopathic
scoliosis. Spine 2003;28:484–91.
[32] Tao F, Shi Z, Xie Y, et al. Determination of lowest instrumented ver-
tebra by the location of apical vertebra in Lenke type 1 adolescent
idiopathic scoliosis. Int Orthop 2011;35:561–7.
[33] Matsumoto M, Watanabe K, Hosogane N, et al. Postoperative distal
adding-on and related factors in Lenke type 1A curve. Spine
2013;38:737–44.
[34] Cho RH, Yaszay B, Bartley CE, et al. Which Lenke 1A curves are at
the greatest risk for adding-on . and why? Spine 2012;37:1384–90.
[35] Miyanji F, Pawelek JB, Van Valin SE, et al. Is the lumbar modifier
useful in surgical decision making?: defining two distinct Lenke
1A curve patterns. Spine 2008;33:2545–51.
[36] Potter BK, Rosner MK, Lehman RA Jr, et al. Reliability of end, neu-
tral, and stable vertebrae identification in adolescent idiopathic sco-
liosis. Spine 2005;30:1658–63.
[37] Stokes IA, Bigalow LC, Moreland MS. Three-dimensional spinal cur-
vature in idiopathic scoliosis. J Orthop Res 1987;5:102–13.
[38] Xiong B, Sevastik J, Hedlund R, et al. Segmental vertebral rotation in
early scoliosis. Eur Spine J 1993;2:37–41.
[39] Dubousset J. Three-dimensional analysis of the scoliotic deformity.
In: Weinstein SL, ed. The pediatric spine: principles and practices.
New York, NY: Raven Press, 1994:479–96.
[40] Easwar TR, Hong JY, Yang JH, et al. Does lateral vertebral transla-
tion correspond to Cobb angle and relate in the same way to axial ver-
tebral rotation and rib hump index? A radiographic analysis on
idiopathic scoliosis. Eur Spine J 2011;20:1095–105.

Downloaded for Anonymous User (n/a) at Saint Joseph University from ClinicalKey.com by Elsevier on February 15,
2022. For personal use only. No other uses without permission. Copyright ©2022. Elsevier Inc. All rights reserved.