Original Articles Validation of a Simple Method for Measuring Cranial

Original Articles Validation of a Simple Method for Measuring Cranial Deformities (Plagiocephalometry) Le´on N.A. van Adrichem, MD, PhD,* Leo A. van Vlimmeren, PhD,1 Dominika )adanova´, MD,* Paul J.M. Helders, PhD,2 Raoul H.H. Engelbert, PhD,2 Han (J) W. van Neck, PhD,* Anton H.J. Koning, PhD4 Rotterdam, The Netherlands

Craniofacial measuring is essential for diagnosis or evaluation of growth and therapies. Skull deformities in children are mainly caused by craniosynostosis or by external pressure in positional skull deformations. Traditional anthropometry does not sufficiently analyze craniofacial shape. In computed tomography (CT) scanning, radiation loads are considerable. Both CT and magnetic resonance imaging (MRI) scanning require anaesthesia in children for accurate imaging, due to their long acquisition time. This makes CT and MRI unsuitable for long term follow-up of pediatric patients unless there is a compelling reason. Other noninvasive three-dimensional (3D) surface scanners still have limited practical use. Van Vlimmeren et al6 presented plagiocephalometry (PCM) as a simple and versatile instrument to quantify skull deformities with high intrarater and interrater reliability, but no comparison was made with the actual skull shape. At the Erasmus University Medical Center Rotterdam, Sophia Children’s Hospital PCM was compared to 3D-CT scanning in 21 children with craniosynostosis early in life. The PCM ring proved to fit closely to the skin with mean differences less than 1 mm (P G 0.05). The shape of the PCM ring was not significantly changed when taken off the head (P 9 0.05). Finally, no significant differences

From the *Department of Plastic and Reconstructive Surgery; Dutch Craniofacial Center, Erasmus University Medical Center Rotterdam, Sophia Children’s Hospital, Rotterdam, the 1Department of Physical Therapy, Bernhoven Hospital, Veghel, the 2Department of Pediatric Physical Therapy and Exercise Physiology, University Medical Center, Wilhelmina Children’s Hospital, Utrecht; and the 4 Department of Bioinformatics, Erasmus University Medical Center Rotterdam, The Netherlands. Address correspondence and reprint requests to Dr. Le´on N.A. van Adrichem, Plastic Surgeon, Erasmus University Medical Center Rotterdam, Sophia Children’s Hospital Dr. Molewaterplein 60 3015 GJ Rotterdam, The Netherlands; E-mail: l.vanadrichem@ erasmusmc.nl

are shown between measurements on the skull (CT-scan) and PCM ring off the head (P 9 0.05). This study proves that PCM is a reliable method for analysis of skull deformities. The measurements are in agreement with 3D-CT scanning as golden standard. Although only 2-dimensional measurements are performed by PCM, the combination of simplicity, reliability, and validity make it a promising tool for daily practice. Key Words: Craniosynostosis, positional plagiocephaly, anthropometry, plagiocephalometry

M

easurement of craniofacial structures is essential for the diagnosis or evaluation of growth and subsequent intervention. Skull deformities in children are mainly caused by craniosynostosis or by external pressure in positional skull deformations. In traditional anthropometry, distances and angles are measured, but shape is not recorded. Furthermore, on the skull, especially in the deformed skull, no clear landmarks are present. Young children are lively, which complicates anthropometry. Plain X-rays of the skull do show cranial sutures, but shape is not sufficiently recorded.1 Ideally, sequential complete three-dimensional (3D) craniofacial pictures should be obtained. In computed tomography (CT) scanning, radiation loads are considerable and both CT and MRI scanning, due to their long acquisition time, require anaesthesia in children if an accurate picture is needed.2Y4 This makes CT and MRI unsuitable for long term follow-up of pediatric patients, unless there is a compelling reason to do so. Noninvasive 3D surface scanning might be a good solution, but although numerous articles using various forms of 3D surface data acquisition have been described since the late 1970s, these methods are of limited practical use. The acquisition time of these devices is in the range of several 15

Copyright @ 2008 Mutaz B. Habal, MD. Unauthorized reproduction of this article is prohibited.

THE JOURNAL OF CRANIOFACIAL SURGERY / VOLUME 19, NUMBER 1

seconds, not sufficiently short enough to capture a highly mobile child. Three-dimensional photogrammetry, however, is accurate, fast, noncontact, and noninvasive, and is a promising system to obtain sequential complete 3D surface data.5 Disadvantages are the price and the size of the system, so acquisition can only be done in specialized centers. Van Vlimmeren et al6 recently presented a simple and versatile instrument to quantify skull deformities. The plagiocephalometry (PCM) is performed with a strip of thermoplastic material, which is positioned around the infant’s head at the widest transverse circumference. Landmarks are the posterior edge of the tragus of the ear and the middle of the nose bridge. PCM turned out to be easy applicable, noninvasive at low costs, while intrarater and interrater reliability was high as illustrated by intraclass correlations and limits of agreement. Although the method is 2D, sufficient information was obtained to quantify the skull deformities. No validation was made, with 3D CT scanning being the gold standard of 3D monitoring. So, it is not shown that PCM corresponds with the actual shape of the head. The aim of this study is: 1. To investigate how closely the thermoplastic strip fits to the skin. 2. To investigate whether the thermoplastic strip retains its shape after taking it off the infants head.

January 2008

3. To determine to what extent the asymmetry, as acquired from the thermoplastic ring, corresponds with the actual asymmetry of the skull as acquired from CT scanning. MATERIALS

AND

METHODS

Patients

A

ll measurements were performed at the Erasmus University Medical Center Rotterdam, Sophia Children’s Hospital, from September 2004 till March 2005. All 21 included patients were scheduled for 3D-CT scanning of the head due to a serious suspicion of existing craniosynostosis (age in months: mean 12.6 +/j 12.7; median 10.5; minimum 3.8; maximum 62.7). The parents agreed that PCM with a thermoplastic strip was performed simultaneously. No children with positional skull deformations were included, because in these patients, CT scanning is redundant and not ethical due to anaesthesia. Methods

PCM is performed with a strip of thermoplastic material (3.2 mm thick) of 18 mm 50 cm (Thermo Extra-comfort, Non-Perfo\ by GeniMedical, Nieuwegein, The Netherlands), which is positioned around the infant’s head at the widest transverse circumference. In less than 2 minutes, the ring is fixed and the 3 landmarks (both ears and nose) are marked on the ring in a standardized manner. Both landmarks at the posterior edge of the tragus,

Fig 1 Illustration of plagiocephalometry: symmetric skull (3-month-old male): A) Photograph of the child with the thermoplastic ring fitted and landmarks. B) Paper copy of the same ring with drawn and measured lines.

16


VALIDATION OF PLAGIOCEPHALOMETRY / van Adrichem et al

correspond best with the meatus acousticus externus. The third landmark is traced off at the middle of the nose bridge. In this way it is possible to trace the position of the ears and the nose in relation to the transverse circumference and contours of the head. Afterwards, the ring is removed from the head and a fourth landmark is marked representing the middle of the posterior distance between the left and right ear, measured with a measuring tape. Using a standard copying machine, the upper side of the ring is copied on paper. Nine lines are drawn on the paper copy and measured to the nearest millimeter, by which the degree of asymmetry can simply be determined by calculating the differences between the lengths of the left and right lines. The clinically most important measures are arranged in three parts:

most outstanding.7 The ratio between the ODL and the ODR is calculated as the longest/shortest diameter 100% and is called oblique diameter difference index (ODDI). 3. Part 3: Transversal shape and proportion of the skull: the ratio between the sinistra-dextra (SD) and the anteroposterior (AP) is calculated as SD/ AP 100%, and is called the cranio proportional index (CPI) (Figs 1A, B and 2A, B). 4. A medical student (D.).) was trained to measure PCM using a standardized protocol. Prior to CT scanning, the ring was obtained. During CT scanning, the PCM ring was in place. After CT scanning, the PCM ring was copied on paper, and finally all lines were drawn and the distances were measured in millimeters.

1. Part 1: Position of the ears, nose and local flattening of the skull: ear deviation (ED), anteriorsinistra Y anterior-dextra (ASAD) and posteriordextra Y posterior-sinistra (PSPD). 2. Part 2: Diameter difference: oblique diameter difference (ODD). The oblique diameter left (ODL) and oblique diameter right (ODR) lines are drawn from points located 40- either side of the anteroposterior (AP) line. The ODD is calculated as ODL-ODR. The angle of 40- is chosen because this has been used by other authors formerly, probably because the differences between the diameters of the typical shape of the skull at these angles are the

Three-dimensional CT scanning was performed at the radiology department of the Erasmus University Medical Center Rotterdam, Sophia Children’s Hospital. All children were under general anaesthesia. The CT scanner equipment was Siemens Emotion 6 (Siemens Netherlands N.V., Den Haag, The Netherlands); a slice thickness of 2.5 mm with an interslice distance of 1.2 mm was used (Fig 3a and b). The 3D CT data sets were imported into BARCO I-Space (Barco, Kortrijk, Belgium) at the Department of Bioinformatics of the Erasmus University Medical Center Rotterdam. The I-Space is a CAVEi-like virtual reality system with 3D-processing tools.8

Fig 2 Illustration of plagiocephalometry: Asymmetric skull in positional plagiocephaly of right occiput (4-month-old male): A) Photograph of the child with the thermoplastic ring and landmarks. The digitally drawn lines are made to illustrate the agreement with the paper copy and to explain the names of the lines. B) Paper copy of the same ring with drawn and measured lines. AP indicates anterior-posterior; SD, sinistra-dextra; AS, anterior-sinistra; AD, anterior-dextra; PS, posterior-sinistra; PD, posterior-dextra; ED, ear deviation; ASAD, AS minus AD; PDPS, PD minus PS; ODL, oblique diameter left; ODR, oblique diameter right; ODD, ODL minus ODR.

17


THE JOURNAL OF CRANIOFACIAL SURGERY / VOLUME 19, NUMBER 1

January 2008

Fig 3 Example of a 3D-CT image with PCM ring on the head of an infant with trigonocephaly.

I-Space, equipped with 3D volume rendering software, uses eight projectors on 3 walls and the floor to create a true 3D image in a special viewing arena.9 Users can then interact with the image using a 6 degrees-of-freedom pointing device, and investigate the 3D data wearing a pair of glasses with polarizing lenses. The person interacting with the data is presented with the correct perspective by means of wireless head tracking. In I-Space, the plane of the PCM ring was determined visually by rotating and clipping the volume. In this plane, the same ear and nose markers of the PCM method were placed on the bone and visually checked for accuracy by looking at the

volume from different angles. After this, a screen dump was made and printed on A4 size paper, closely cropped on the skull. All the measurements described in the PCM part were performed on the bone. To investigate how close the PCM ring fits to the skin and the underlying bone, the distance between the PCM ring and the skin and between the PCM ring and the bone was measured at positions of 0-, 45-, 90-, 135-, 180-, 225-, 270-, and 315- (Fig 4A and B). A difference of 1 mm between the skin and the PCM ring was defined as an acceptable difference due to the interposing of hair, based on clinical findings in children of 1 year of age. At the

Fig 4 Example of the technique used for establishing the distance between: A) The strip and the center of the head at the various positions the strip and the skin at various positions. B) The strip and the skin at the various positions.

18



8 angles, the distance was measured with a ruler between the center (crossing of the anterior-posterior and sinistra-dextra lines) and the PCM ring on the head, the PCM ring off the head, the bone and skin. In a pilot study, the method appeared reproducible. Statistical Analysis All the obtained data were registered and statistically analyzed by making use of the Statistical Program for the Social Sciences (SPSS) version 12.0.1. The ‘‘t’’-test has been used for determining the significance of the obtained data.

Table 2. Distance Between PCM Ring and Skull. One-Sample Test Test Value = 4 95% Confidence Interval Angles

N

Mean

Std. Deviation

Sig. (2-tailed)

Lower

Upper

04590135180225270315-

21 21 21 21 21 21 21 21

4.51 3.21 3.69 3.84 3.85 3.81 3.86 3.61

1.37 0.92 0.93 0.75 0.97 0.70 0.63 0.87

0.11 0.00 0.14 0.34 0.49 0.23 0.32 0.05

3.88 2.79 3.26 3.50 3.41 3.49 3.57 3.22

5.13 3.63 4.11 4.18 4.29 4.13 4.14 4.01

RESULTS The Distance Between PCM Ring and Skin

T

he mean distance between PCM ring and skin is summarized in Table 1. The zero hypothesis is that the difference is equal to 1 mm. All mean values are less than 1 mm. The results of the One-Sample ‘‘t’’-Test show that in all but the 90- positions, the values are statistically significant (P G 0.05) different from 1 mm. Taking the mean values into account, this indicates that the distances are significantly smaller than 1 mm. At the 90- position, the value is not statistically significant (P 9 0.05), which means that the distance between the PCM ring and the skin matches our assumed value of 1 mm. In all positions, the distance between the PCM ring and skin fulfills the criterion of 1 mm or less.

the difference is equal to 4 mm, based on the clinical soft tissue-thickness of about 3 mm (measured during surgery in children of 1 year of age) combined with the expected distance of maximum 1 mm between the PCM ring and the skin. All mean values are between Table 3. Comparison of the PCM Ring On the Head and Off The Head

Pair 1

Pair 2

The Distance Between PCM Ring and Bone The mean distance between PCM ring and skull is summarized in Table 2. The zero-hypothesis was that

Pair 3

Pair 4

Table 1. Distance Between PCM Ring and Skin. One-Sample Test

Pair 5

Test Value = 1 95% Confidence Interval Angles

N

Mean

Std. Deviation

Sig. (2-tailed)

Lower

Upper

04590135180225270315-

21 21 21 21 21 21 21 21

0.48 0.58 0.82 0.43 0.53 0.42 0.73 0.64

0.60 0.48 0.61 0.38 0.44 0.43 0.49 0.49

0.00 0.00 0.20 0.00 0.00 0.00 0.02 0.00

0.21 0.36 0.54 0.26 0.34 0.22 0.51 0.42

0.76 0.80 1.10 0.60 0.73 0.61 0.96 0.87

Pair 6

Pair 7

Pair 8

Angles

Mean

N

Std. Deviation

Correlation

Sig.

Ring on the head at 0Ring off of the head at 0Ring on the head at 45Ring off of the head at 45Ring on the head at 90Ring off of the head at 90Ring on the head at 135Ring off of the head at 135Ring on the head at 180Ring off of the head at 180Ring on the head at 225Ring off of the head at 225Ring on the head at 270Ring off of the head at 270Ring on the head at 315Ring off of the head at 315-

76.72

21

8.92

1.00

0.00

76.70

21

8.80

67.67

21

7.71

0.98

0.00

67.29

21

8.08

62.92

21

5.01

0.95

0.00

62.95

21

5.05

70.17

21

6.81

0.98

0.00

70.40

21

6.78

76.72

21

8.92

1.00

0.00

76.68

21

8.81

69.60

21

5.47

0.96

0.00

69.76

21

5.70

63.02

21

4.48

0.94

0.00

63.50

21

4.28

67.66

21

5.83

0.97

0.00

67.45

21

6.63

19


THE JOURNAL OF CRANIOFACIAL SURGERY / VOLUME 19, NUMBER 1 Table 4. Difference Between PCM Ring On and Off the Head

January 2008

Table 6. Difference Between Measurements on the Skull and on the PCM Ring Off the Head

Paired Differences

Paired Differences

95% Confidence Interval of the Difference Std. Deviation

Angles

Mean

Pair 1

on head 0- - off head 0on head 45- - off head 45on head 90- - off head 90on head 135- - off head 135on head 180- - off head 180on head 225- - off head 225on head 270- - off head 270on head 315- - off head 315-

Pair 2 Pair 3 Pair 4 Pair 5 Pair 6 Pair 7 Pair 8

95% Confidence Interval of the Difference

Lower

Upper

Sig. (2-tailed)

0.02

0.51

j0.21

0.25

0.85

Pair 1

0.39

1.59

j0.33

1.11

0.28

Pair 2

j0.03

1.54

j0.73

0.67

0.93

Pair 3

j0.24

1.52

j0.93

0.46

0.48

Pair 4

0.05

0.47

j0.17

0.26

0.66

Pair 5

j0.16

1.53

j0.86

0.53

0.63

Pair 6

j0.48

1.56

j1.19

0.24

0.18

0.21

1.83

j0.63

1.04

0.61

3.2 to 4.5 mm. The results of the One-Sample ‘‘t’’-Test show that in all but the 45- positions the values are not statistically significant (P 9 0.05) different from 4 mm. No statistical significance (P 9 0.05) indicates that the distance between the PCM ring and the skull matches our assumed value of 4 mm. Taken into account the mean value at 45-, this indicates that the distance is significantly smaller than 4 mm (P G 0.05). In all positions, the distance between the PCM ring and skin fulfills the criterion of 4 mm or less.

Table 5. Comparison of Measurements on The Skull and on the PCM Ring Off the Head Measurements Pair 1 Pair 2 Pair 3 Pair 4 Pair 5 Pair 6

Mean

N

Std. Deviation

Correlation

Sig.

ASADstrip 0.33 ASADskull 0.46 PDPSstrip 0.02 PDPSskull 0.08 EDstrip 5.19 EDskull 5.06 ODDstrip 0.81 ODDskull 0.53 ODDIstrip (%) 103.03 ODDIskull (%) 103.38 CPIstrip (%) 83.93 CPIskull (%) 82.91

21 21 21 21 21 21 21 21 21 21 21 21

6.45 6.55 3.62 3.38 3.99 3.71 5.93 5.69 2.76 2.86 11.24 12.52

0.95

0.00

0.95

0.00

0.96

0.00

0.95

0.00

0.86

0.00

0.98

0.00

ASAD = anterior-sinistra Y anterior-dextra; PDPS = posterior-dextra Y posterior-sinistra; ED = ear deviation; ODD = oblique diameter difference (ODD); ODDI = oblique diameter difference index; CPI = cranio proportional index.

ASADstrip ASADskull PDPSstrip PDPSskull EDstrip EDskull ODDstrip ODDskull ODDIstrip Y ODDIskull (%) CPIstrip Y CPIskull (%)

Mean

Std. Deviation

Lower

Upper

Sig. (2-tailed)

j0.12

2.00

j1.03

0.79

0.78

j0.06

1.18

j0.59

0.48

0.83

0.13

1.06

j0.35

0.62

0.58

0.28

1.91

j0.59

1.15

0.51

j0.35

1.47

j1.02

0.32

0.29

1.02

2.85

j0.28

2.32

0.12


PCM Ring Taken Off the Head Compared to Ring On Head Comparison of the distance between the center (crossing of the anterior-posterior and sinistra-dextra lines) and the PCM ring at the 8 angles, when the PCM ring was on the head and off the head, showed a statistically significant correlation for every angle (Table 3). The mean differences between PCM ring on the head and off the head are summarized in Table 4. The results of the Paired-Samples ‘‘t’’-Test show that in all positions the values are not statistically significant (P 9 0.05). The 95% confidence intervals of difference are within 1.2 mm. So, only a negligible Table 7. Limits of Agreement in This Study and Intrarater PCM Study Skull Y PCM off the head (this study) ASAD PDPS ED ODD ODDI CPI

j4.12 till + 3.87 j2.42 till + 2.30 j1.99 till + 2.26 j3.55 till + 4.11 j3.30 till + 2.60 j4.69 till + 6.72

PCM Y PCM off the head (intrarater van Vlimmeren) j5.86 j2.88 j4.32 j3.37 j2.69 j5.32

till till till till till till

+ + + + + +

5.86 3.08 4.24 3.41 2.81 3.68


20



shape difference occurs when the PCM ring is taken off the head. Measurements on the Skull (CT Scan) Compared to the PCM Ring Off the Head Table 5 shows a statistically significant correlation between measurements on the skull (CT scan) and PCM ring off the head. The mean differences between measurements on the skull and on the PCM ring off the head are summarized in Table 6. The results of the PairedSamples ‘‘t’’-Test show that in all measurements the values are not statistically significant (P 9 0.05). The 95% confidence intervals of difference are within 2.3 mm. Table 7 shows the corresponding limits of agreement for this study and for the intrarater comparison of two PCM rings in the PCM validation study.6 The limits of agreement between skull and PCM off the head are comparable to the intrarater PCM study of van Vlimmeren et al.6 It can be concluded that the measurements on the PCM ring off the head correspond to the skull measurements on the CT scan. DISCUSSION

P

CM is an easy-to-apply, noninvasive measurement to assess skull shape with good clinical use and low application costs. Van Vlimmeren et al6 showed a statistically acceptable intrarater and interrater reliability for the application in clinical care and research. Our study proves that the measurements on the PCM ring yield a valid representation of the skull shape. The tape fits closely to the skin with mean differences less than 1 mm. Furthermore only a negligible shape difference occurs when the PCM ring is taken off the head. The distances between the PCM ring and the skull can be explained by the interposing soft tissues. Finally, no statistically significant difference is shown between measurements on the skull (CT scan) and the PCM ring off the head. Comparison of the measurements on the skull and on the PCM ring off the head show corresponding limits of agreement with the intrarater comparison of two PCM rings in the PCM

validation study.6 The fact that the CPI value shows slightly higher limits of agreement might be caused by the prone position of the infant during CT scanning with pressure at the back of the head. It can be concluded that PCM is a validated measuring method for skull deformities. We have to realize that the PCM is in fact a 2D measuring method. It only gives information on the transverse plane. No information is gathered on the height of the skull. Most skull deformities, however, can be recognized and analyzed in this transverse plane. True 3D measuring methods will provide more information, because more data are gathered. As explained earlier, no 3D system is available that is noninvasive, easy to apply and with low application costs. The available 3D surface monitoring systems are promising, but still reserved to specialized centers.5 This study proves that PCM actually reliably reproduces skull shape; the simplicity of the method makes it a promising tool for analysis of plagiocephaly in daily practice. REFERENCES 1. Pollack IF, Losken HW, Fasick P. Diagnosis and management of posterior plagiocephaly. Pediatrics 1997;99:180Y185 2. Binaghi S, Gudinchet F, Rilliet B. Three-dimensional spiral CT of craniofacial malformations in children. Pediatr Radiol 2000;30: 856Y860 3. Benson ML, Oliverio PJ, Yue NC, et al. Primary craniosynostosis: imaging features. AJR Am J Roentgenol 1996;166:697Y703 4. Darling CF, Byrd SE, Allen ED, et al. Three-dimensional computed tomography imaging in the evaluation of craniofacial abnormalities. J Natl Med Assoc 1994;86:676Y680 5. Riphagen JM, van Neck JW, van Adrichem LN. 30 years of 3D surface imaging in medicine: a review of working principles and implications for imaging the unsedated child. J Craniofacial Surg. In press 6. van Vlimmeren LA, Takken T, van Adrichem LN, et al. Plagiocephalometry: a non-invasive method to quantify asymmetry of the skull; a reliability study. Eur J Pediatr 2006; 165:149Y157. Epub 2005 Oct 7 7. Hutchison BL, Hutchison LA, Thompson JM, et al. Plagiocephaly and brachycephaly in the first two years of life: a prospective cohort study. Pediatrics 2004;114:970Y980 8. Cruz-Neira C, Sandin DJ, DeFanti T. Surround-screen projection-based virtual reality: the design and implementation of the CAVE. In SIGGRAPH ’93 Proceedings Association for Computing Machinery, August 1993 9. Koning AHJ. Applications of volume rendering in the CAVE. In: Enquist B., et al, ed. Simulation and visualization on the grid Paralleldatorcentrum, seventh annual conference. Stockholm, 2000:112Y121

21
