Measurement Errors

1
Measurement Errors in Acetabular Index and
Center-Edge Angle According to Pelvis Position -
3D Image Analysis -
Tae-Joon Cho, Hwan Seong Cho, In Ho Choi,
ChinYoub Chung, Won Joon Yoo Seoul National
University Childrens Hospital, Seoul,
Korea National Cancer Center, Kyeonggi, Korea
2
INTRODUCTION

• Acetabular index(AI) and center-edge (CE) angle
  are the most important parameters in evaluating
  pediatric hip dysplasia.

Measurement errors from - intra or inter
observer variability - irregularity in
ossification - non-standardized pelvic
positioning

• Effect of hip disease e.g.)
  flexion contracture
• Asymmetrical growth of the pelvis
• Hip joints not in focus e.g.) simple
  abdomen, IVP
• Technical error in shooting

• We investigated the change in measurements of of
  acetabular index and CE angle according to pelvic
  rotation and tilting using 3-D CT images, and
  delineated the acceptable range of rotation for
  acetabular index measurement.

3
MATERIALS and METHODS

• 21 LCP patients
• M / F 18/ 3
• Age avg. 7.5 years (range, 4 11 years)
• 3D-CT software V-worksTM 5.0 (Cybermed Inc.,
  Seoul, Korea)
• 3-D maximum intensity projection (MIP)
• set at the anatomical position
• rotated around the superoinferior axis in 3
  increments to 15
• tilted around the mediolateral axis in 3
  increment to 9
• AI, CE angle, foramen obturator index (FOI) and
  pelvic tilt index (PTI) were measured in each
  positions, and compared with those measured in
  the anatomical position.

4
MATERIALS and METHODS
3D MIP image
9 Rt. rotation
9 Lt. rotation
pelvic tilt index (PTI) r / t
AI
t
CE angle
r
a
b
foramen obturator index (FOI) a / b
9 Post. tilt
9 Ant. tilt
5
RESULTS
FOI (foreamen obturator index) was correlated
well with the amount of pelvic rotation. But PTI
(pelvic tilt index) was inappropriate to estimate
the amount of pelvic tilt due to wide range of
individual variation.
Rt. rotation 9
Lt. rotation 9
FOI 0.5
FOI 2.0
Rt
Lt
Linear relationship between logFOI and rotation
6
RESULTS

• Measurement error in acetabular index

AI was over-estimated when pelvis was internally
rotated and tilted backward. Change in AI
measurements according to pelvic rotation was
less when pelvis was posteriorly tilted.
overestimated
external rotation
anterior tilting
internal rotation
underestimated
posterior tilting
?AIError in acetabular index measurement
7
RESULTS
Measurement error in acetabular index
Pelvic rotation by 6 or less (FOI 0.661.43)
produced measurement error of AI less than 2
(95 confidence interval).
overestimated
external rotation
internal rotation
underestimated
cor0AI when pelvis is not tilted lowerlower
limit of 95 confidence interval upperupper
limit of 95 confidence interval.
8
RESULTS
Measurement error in acetabular index
Change of AI measurement according to the
rotation remained relatively consistent when
posteriorly tilted. When pelvis was posteriorly
tilted by 9, AI was slightly overestimated and
varied within only 2 by pelvic rotation of 30
range.
overestimated
external rotation
internal rotation
cor9 when pelvis is posteriorly tilted by
9 lower lower limit of 95 confidence
interval upper upper limit of 95 confidence
interval
underestimated
9
RESULTS
Measurement error in C-E angle
CE angle was over-estimated, when pelvis is
internally rotated and posteriorly tilted. CE
angle change according to pelvic rotation and
tilting was exaggerated on the affected side.
overestimated
overestimated
underestimated
underestimated
?CEerror in CE angle measurement
affected side (LCP) contralateral normal side
10
SUMMARY

• Measurement errors in AI and CE angle according
  to the pelvis position were investigated using
  3D-CT of 21 LCP patients at average age of 7.5
  years.
• FOI is a good estimator of pelvic rotation,
  but PTI is not of pelvic tilt.
• Both AI and CE angle are over-estimated when
  internally rotated and vice versa.
• Large obturator foramen, large AI large CE
  angle!
• Pelvis rotation within 6 (FOI 0.66 1.43)
  renders measurement error of the acetabular index
  less than 2.
• Posteriorly tilted pelvis (as in Laugenstein
  view) renders slightly over-estimated but
  consistent measurement of AI regardless of pelvic
  rotation.
• Measurement errors of these parameters should
  be taken into consideration when evaluating hip
  dysplasia.

Correspondence to Tae-Joon Cho, MD
(tjcho_at_snu.ac.kr) Seoul National University
Childrens Hospital Seoul, Korea