Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Jun 1.
Published in final edited form as: Ann Plast Surg. 2019 Jun;82(6):679–685. doi: 10.1097/SAP.0000000000001906

Long-term Characterization of Cranial Defects after Surgical Correction for Single-Suture Craniosynostosis

Gary B Skolnick 1, Murthy Sindhoora 2, Kamlesh B Patel 1, Huang Zhiyang 3, Sybill D Naidoo 1, Ju Tao 2, Matthew D Smyth 1, Albert S Woo 4
PMCID: PMC6519078  NIHMSID: NIHMS1519891  PMID: 31082848

Abstract

Introduction

Craniosynostosis is typically corrected surgically within the first year of life through cranial vault reconstruction. These procedures often leaves open calvarial defects at the time of surgery, which are anticipated to close over time in a large proportion of cases. However, residual calvarial defects may result as long-term sequelae from cranial vault remodeling. When larger defects are present, they may necessitate further reconstruction for closure.

Better understanding of the calvarial osseous healing process may help to identify which defects will resolve or shrink to acceptable size and which will require further surgery. Our study aims to assess the long-term changes in defect size after cranial vault reconstruction for craniosynostosis.

Methods

One year post-operative and long-term computed tomography scans were retrieved from the craniofacial anomalies archive. Analysis used custom software. All defects above the size of 1 cm2 were analyzed and tracked for calvarial location, surface area, and circularity. Monte Carlo simulation was performed to model the effect of initial defect size on the rate of defect closure.

Results

We analyzed a total of 74 defects. The average initial defect surface area was 3.27 ± 3.40 cm2. The average final defect surface area was 1.71 ± 2.54 cm2. The average percent decrease was 55.06 ± 28.99 %. There was a significant difference in the percentage decrease of defects in the parietal and fronto-parietal locations: 68.4% and 43.7%, respectively (p = 0.001). Monte Carlo simulation results suggest that less than 10% of defects above the size of 9 cm2 will close to the size of 2.5 cm2 or less.

Conclusions

We describe and make available a novel, validated method of measuring cranial defects. We find that the large majority of initial defects greater than 9cm2 remain at least one square inch in size (2.5cm2) one year post-operatively. Additionally, there appear to be regional differences in closure rates across the cranium, with front-parietal defects closing more slowly than those in the parietal region. This information will aid surgeons in the decision-making process regarding cranioplasty after craniosynostosis correction.

Introduction:

Calvarial vault remodeling remains the gold standard treatment for patients identified with craniosynostosis [1]. The goals of surgery are to normalize the appearance of the cranial vault and allow for normal brain development [2]. However, incomplete ossification is a common outcome following operative reconstruction. Indeed, incomplete healing following fronto-orbital advancement for unicoronal and metopic synostosis is well documented[3]. When such defects are considered to be clinically significant, additional surgical procedures may be necessary to correct for large persistent defects.

The presence of large calvarial defects following surgery can pose significant risks to those affected. Neurosurgical patients who undergo sizeable craniectomies require the use of protective helmets until reconstruction can be performed to minimize trauma to the brain. It is unclear, however, what defines a critical-sized defect of the skull requiring helmeting for protection or mandating reconstruction, especially in the craniosynostosis population. These decisions would be aided by more quantitative results elucidating the size of residual calvarial defects long-term in calvarial vault reconstruction patients.

While multiple reports explore the concept of cranioplasty for the correction of residual calvarial defects, clinical research into the natural healing of post-operative calvarial gaps and the effect of initial defect size and location in determining long-term closure remains incomplete. Instead, the literature generally focuses on either the materials and methods of cranioplasty[3–8] or the impact of age at initial repair on the need for cranioplasty[9]. Moreover, such studies rarely present quantitative analysis of bony regrowth. Rather, in part to minimize radiation exposure[7], many studies report post-surgical results obtained via X-ray imaging [10] or by clinical palpation [7, 11], precluding quantitative analysis of the defects.

Given the limitations in the medical literature, cranioplasty with bone grafts or implants may be recommended to patients with residual calvarial defects that might naturally heal to a safe size following surgery. Treating surgeons must weigh the risk of complications from cranioplasty, including postoperative epidural hematoma and infection[12], against the risks of maintaining an open calvarial defect in younger children. A better understanding of calvarial osseous healing would be beneficial to help predict outcomes and aid in decision-making. This study uses computer software to quantitatively assess the long-term changes in cranial defects following reconstruction of single-suture craniosynostosis.

Methods:

Patient Population:

After obtaining Institutional Review Board approval, patients for the study were identified from the craniofacial anomalies archive[13]. All subjects were required to meet the following criteria: 1) a date of birth on or after January 1st, 1999, 2) a correctional procedure for one of three single-suture craniosynostosis diagnoses: unicoronal, sagittal, or metopic, 3) an available 1 year post-operative cranial CT scan, 4) an available long-term post-operative cranial CT scan that was taken at least 15 months after the 1 year post-operational cranial CT scan, and 5) a non-syndromic diagnosis, to minimize concerns that calvarial defects may be the result of abnormal biology rather than the characteristics of the bony defect itself

Patients were excluded from the study if: 1) either of the two scans had a motion artifact that would interfere with measurement of cranial defects, 2) the subject had multiple-suture craniosynostosis, or 3) the patient had undergone an additional procedure on the skull, such as a second surgery or cranioplasty, between the initial surgery and the long-term post-operative scan. From these criteria, 37 children were included in the study. Out of the 37 patients, 14 had undergone endoscopic repair and 23 had an open reconstruction.

Surgical Technique:

Open Calvarial Vault Reconstruction:

Multiple variations exist among surgeons for surgical reconstruction of craniosynostosis. However, the procedures in this study could generally be classified as posterior calvarial vault remodeling or fronto-orbital advancement. When the sagittal or lambdoid sutures were affected, the posterior calvarial vault remodeling procedure was performed (Figure 1A). The techniques are well-described in the literature[14]. In brief, the parietal bones were typically remodeled with barrel stave osteotomies. An occipital craniectomy was sometimes performed with radial osteotomies to flatten and expand the posterior aspect of the skull. When the metopic or coronal sutures were affected, fronto-orbital advancement was undertaken (Figure 1B). With this procedure, a frontal craniotomy was performed and the supraorbital bandeau was then osteotomized, remodeled, and placed back into anatomic position. The frontal bone was similarly remodeled with radial osteotomies as necessary. Calvarial defects were common lateral to the frontal bone, especially when significant advancement was performed. While the authors currently employ particulate bone grafting techniques, none of the patients in this study had any significant bone grafting of the defects.

Figure 1.

Figure 1.

Open in a new tab

Perioperative CT images of A: open posterior calvarial vault remodeling for sagittal synostosis, B: open fronto-orbital advancement for metopic synostosis, and C: endoscopically assisted craniectomy for sagittal synostosis

Endoscopic Craniosynostosis Surgery:

The endoscopic procedures were largely performed in accordance with the procedure described by Jimenez and Barone[15, 16] (Figure 1C). Typically performed for sagittal synostosis, a 4–5 cm midline craniectomy was made extending from bregma to lambda. Lateral barrel stave osteotomies were then created to allow for additional lateral expansion of the parietal bones. When the metopic or coronal sutures were involved, the procedure consisted of a simple craniectomy of the affected suture, measuring no more than 1 cm in width.

Data Analysis:

Proprietary software was created by co-author Z.H. specifically to identify and measure cranial defects using three-dimensional CT data. The software used a triangulated surface to determine the defect borders through the method of morphological thinning [17]. Defect area was then calculated using minimum weight triangulation [18]. Details on the software reliability, validity and user interface are explained in the appendix. In order to limit analysis to clinically significant sizes, the minimum defect size for analysis was established to be 1 cm2. The software calculated defect area, perimeter, and circularity. Defects were characterized, tracked, and compared from the 1 year post-op imaging to the 5 year post-op scan (Figure 2).

Figure 2.

Figure 2.

Open in a new tab

The left pane shows a patient who underwent an endoscopic surgical procedure for sagittal synostosis correction with a 1 year post-op scan (top) and a long-term post-op scan (bottom); Defects A, B, and C were patent four years after the initial scan. The middle pane shows a patient who underwent an endoscopic strip craniectomy for metopic synostosis correction with a 1 year post-op scan (top) and a long-term year post-op scan (bottom); the smaller defects above defect D were not considered for analysis since they were below 1 cm2. Defect D was tracked from the 1 year post-op scan to the long-term post-op study. The right pane shows a patient who underwent a fronto-orbital advancement for right unicoronal synostosis correction with a 1 year post-op scan (top) and a long-term post-op result (bottom); Defect E closed in the middle to form two defects.

The same operator (S.M.) classified the defects into groups based on the location of the defects. In concordance with previously published methodology, calvarial defects were classified into 4 groups: frontal, fronto-parietal, parietal, and occipito-parietal [19].

Statistical Tests

Student’s two-tailed t-tests were used for comparisons of scalar variables between endoscopic and open repair. Analysis of variance and Fisher’s Exact testing were used in order to evaluate whether different procedures had defects in similar locations and proportions.

Multiple linear regression analysis was performed to determine factors influencing percent decrease in area with forced entry of covariates. Scan interval, circularity of defect, initial area of defect, and the age at surgery were included as factors in the model. To illustrate the results of the regression model, we performed a simulation to determine the percentages of defects with different initial sizes that would resolve to less than a standardized area (2.5cm2) [20] by age five. The Monte Carlo simulation was included 100,000 trials for each initial defect size. Age at surgery was fixed at six months and follow-up period was fixed at five years post-operative.

Fisher’s Exact testing was performed with an open-source script from MATLAB Central [21]. The Monte Carlo simulation and Fisher’s Exact testing were run using MATLAB R2011b (Mathworks, Natick, MA). All other statistical analyses were performed using SPSS version 22 (IBM SPSS, Chicago, IL). Significance was set a priori at p ≤ 0.05

Results:

Patients – Descriptive Statistics

Mean age at surgery was 9.6 ± 8.5 months (mean ± standard deviation). Age at the 1 year post-operative scan was 21.8 ± 7.9 months. Mean time between the 1 year post-op scan and the long-term post-op study was 47.4 ± 15.2 months. Twenty-three patients underwent an open repair: 12 sagittal, 7 metopic, and 4 unicoronal synostosis repairs. Fourteen patients underwent an endoscopic repair: 11 sagittal, 1 metopic, and 2 unicoronal synostosis repairs.

Defects – Descriptive Statistics

A total of 74 defects were identified and analyzed in 37 subjects on the 1 year post-operative scans. Mean initial defect surface area was 3.27 ± 3.40 cm2 (range of 1.02 to 22.59). Mean final defect surface area was 1.71 ± 2.54 cm2. The mean decrease in defect surface area was 1.56 ± 1.43 cm2, with an decrease in defect size of 55.1 ± 29.0%. Overall, 4 of the 37 patients (10.8%) had no clinically relevant defects (greater than 1 cm2) in their one-year post-operative scan. Five patients (13.5%) had defects in their long-term post-operative scan that healed completely (100% decrease in initial defect size). The number of defects per patient did not differ between the two repair types (1.9 defects/open repair, 2.2 defects/endoscopic repair, p = 0.419).

Defects -- Analytic Statistics

A multiple linear regression model with forced entry was used to analyze predictors of percent decrease of defect. We hypothesized that scan interval, circularity of defect, initial area of defect, and the age at surgery would affect the percent decrease and included these factors in the model. The coefficient of multiple determination for the model was 0.432 (F = 3.957, p = 0.006). The coefficients (β) for the co-variates were as follows: scan interval, 0.004/month (p = 0.044); circularity,- 0.29 (p = .342); initial area, −0.02cm2 (p = .050); and age at surgery, 0.007/month (p = 0.051).

To illustrate the results of the regression model, we performed a Monte Carlo simulation. Initial defect sizes from 3 to 13 cm2 were assessed. For each initial defect size, the simulation included 100,000 trials. Age at surgery was fixed at six months and follow-up period was fixed at five years post-operative. In the simulations, 59% of the defects 5 cm2 contracted to less than 2.5 cm2 but only 8% of defects initially 9 cm2 in area contracted to less than 2.5 cm2 (Figure 3).

Figure 3:

Figure 3:

Open in a new tab

Results of a Monte Carlo simulation illustrating results of the linear regression model. More than 50% of the defects 5 cm2 ended smaller than 2.5 cm2 (green bar). However, fewer than 10% of defects above 9 cm2 ended smaller than 2.5 cm2 (red bar).

Regional Differentiation

The cranial defects were stratified by defect location with respect to the cranial bones that encompassed the defects (Figure 4). This allowed consistent stratification of defects by location between different patients and also enabled the characterization of regional differentiation. There were 3 frontal, 4 occipito-parietal, 29 fronto-parietal, and 38 parietal defects. There was no difference in mean age at surgery between location groups (p = 0.982). A one-way ANOVA was performed to see if there were differences in measured parameters between different groups (Table 1).

Figure 4:

Figure 4:

Open in a new tab

The defects were stratified by which cranial bone or bones they spanned. This patient’s scan, for instance, shows a defect highlighted in green. Since this defect spanned the right parietal, left parietal, and occipital bones, we characterized it as an occipito-parietal defect.

Table 1:

Illustration of the difference in circularity, initial defect area, and percent decrease in area of cranial defects in different locations (e.g. frontal, fronto-parietal, parietal, occipito-parietal). The only variable that showed a significant difference between the different locations was the percent decrease in area.

Variable of Interest Location Mean Significant Difference Between Locations
Circularity frontal 0.263 ± 0.179 No, 0.199
fronto-parietal 0.244 ± 0.093
parietal 0.301 ± 0.117
occipito-parietal 0.249 ± 0.084
Total 0.274 ± 0.111
Initial Defect Area frontal 1.782 ± 0.779 No, 0.109
fronto-parietal 4.463 ± 4.807
parietal 2.571 ± 1.832
occipito-parietal 2.373 ± 1.172
Total 3.270 ± 3.405
Percent Decrease in Area frontal 56.74 ± 11.95 Yes, <0.001
fronto-parietal 43.66 ± 23.29
parietal 68.43 ± 26.23
occipito-parietal 9.52 ± 22.68
Total 55.06 ± 28.99
Open in a new tab

A post hoc test (Tukey HSD) was also performed to determine whether specific locations differed from one another with regard to percent decrease in area. There was a significant difference in the percent decrease between parietal and fronto-parietal locations of 0.684 and 0.437, respectively (p = 0.001). There was no statistically significant difference in the mean initial defect area between parietal and fronto-parietal locations.

A comparison of location versus synostosis type is shown in Figure 5. The location of cranial defects differed significantly based on the type of synostosis repaired (p < 0.001).

Figure 5.

Figure 5.

Open in a new tab

This bar graph visualizes the locations of the defects grouped by procedure. The location of cranial defects differed significantly based on the type of synostosis repaired (p = 0.0008)

Discussion:

Following surgical correction of craniosynostosis, cranial defects are frequently created with the expectation that such gaps naturally fill over time with osseous re-growth. In a subset of patients, however, larger clinically significant defects of the skull remain many years after surgery and require additional surgical intervention. Unfortunately, research on the closure of residual cranial defects in this population typically extends no more than one year after the operation. To remedy this issue, this study attempts to assess longer-term changes to cranial defects following surgery for reconstruction of single-suture craniosynostosis. One square centimeter was chosen arbitrarily as the initial minimum defect size for evaluation in this study. A literature search of the minimum size cranial defect indicated for cranioplasty revealed the smallest size noted in a repair measured 2.5 cm2 [20]. This value was subsequently utilized as a minimum size for clinical significance in a multiple linear regression model.

Presence of defects:

In our long-term study of cranial defects, 23 out of 37 patients (62.2%) had residual defects greater than 1 cm2 at a mean of 8.40 months postoperatively. Of these, 9 out of 37 patients (24.3%) had defects greater than 2.5 cm2 in size. In contrast, palpation in previous large-scale studies demonstrated non-closure rates one year following craniosynostosis surgery as high as 72.6% [10] versus numbers as low as 5.1%, in a separate study [11]. With so much variability in the available literature, more robust methods of characterizing and understanding cranial defect size would be valuable. Such data could provide surgeons critical information to help weigh the risks of injury in patients with large residual defects against the risks of surgical correction.

Notably, endoscopically-assisted strip craniectomy and open calvarial vault remodeling produced the same number of relevant defects. While the average size of the defects was smaller for patients undergoing endoscopic repair, these patients were also significantly younger at age of operation and age at post-operative scan making difficult direct comparisons in defect size between the two repair types.

Factors that affect defect closure

We examined multiple factors that might affect closure including defect shape and size, as well as patient age at repair. As there is a greater perimeter to area ratio in smaller defects, one would expect to see a greater percentage decrease in smaller defects. Our study bore out this hypothesis (p = 0.05). Still, only 5 out of 78 or (6.4%) of all defects closed completely, as determined by the long-term CT scan. Complete cranial defect closure after frontal orbital advancement has previously been documented to be between 8.0% and 53.8%, but was assessed through palpation and not confirmed with imaging studies [11]. In the same study, age at operation was found to have a statistically significant association with closure of bony defect [11]. In our sample, we found a strong trend (p = 0.051) toward greater percent closure in cranial defects with earlier repair, even after accounting for the variation in the duration between scans. This finding correlates well with results from other studies, which noted improved healing of calvarial defects in younger patients [22].

Regional Differentiation

Our data further show that there is a statistically significant difference in percentage decrease for different regions of the skull. In particular, we found that the parietal regions of the cranium had a statistically greater percent decrease in defect size than the fronto-parietal regions. In an earlier review of 592 patients who were operated on for craniosynostosis, it was found that forehead advancement was a parameter that decreased osseous wound healing [10]. The authors hypothesized that the absence of initial dural contact following surgery may play a role in this phenomenon. Our results are in line with these findings and may be related to the fact that the dura mater may be critical to osseous development in the skull [23] [24] [25] [26]. It is our supposition that with fronto-orbital advancement, the dura remains retracted particularly along the lateral fronto-parietal regions, leaving an empty space devoid of tissue until the dura slowly expands to meet the expanded bone segments (Figure 6). In contrast, parietal regions of the skull have little dead space following surgery, with good approximation of the dura to the superficial skin without a gap in between. Proximity to soft tissue (particularly the dura) and the absence of dead space may be critical elements to bony healing.

Figure 6.

Figure 6.

Open in a new tab

Axial slice of post-operative CT of a patient with corrected unicoronal synostosis demonstrating the space between the dura and the skull. The space has been highlighted in white, the skull is light grey, and the brain and other soft tissues are dark grey.

Limitations

A limitation to our study is the small number of patients for whom we had available 1 year post op and long-term CT scans. We also excluded patients who had secondary corrective surgery. Many of these patients likely had larger defects; excluding them may have skewed the distribution toward smaller defects. Due to the limitations of chart review, we were unable to evaluate possible causes of differences in bone growth due to potential confounding factors such as comorbidities or nutritional status.

Conclusion

Calvarial defects following surgery for the treatment of craniosynostosis demonstrate general patterns of healing. Overall, 13.5% of patients had complete closure roughly 5 years after surgery, and larger defects were less likely to close than smaller gaps. Posterior parietal defects were noted to close more readily than fronto-parietal defects. Age at surgery and initial area have negative effects on long-term defect closure. Monte Carlo methods were utilized to predict that less than 10% of defects above the size of 9 cm2 will close to the size of 2.5 cm2 or less.

Acknowledgements

This study was funded through the T35 NIH NHLBI Training Grant and the ICTS/CTSA, NIH Grant Number UL1 TR000448.

Appendix: Cranial Defect Identification and Measurement Software

User Interface

The interface included a view of a cross-section of an unaltered CT scan and a second view of the same cross section with a modifiable Hounsfield Unit (HU) threshold. Initially, the threshold was automatically set to a level that included the soft tissue of the cranium (Figure 1, Top Right). A single operator (S. M.) incrementally modified the threshold until the soft tissue of the eye and the pinna were not present, as demonstrated in Figure 7.

Figure 7.

Figure 7.

Open in a new tab

The software interface shows the 3D reconstruction on the left. The cross section of an unaltered CT scan (top right) and a second view of the same cross section with a modifiable Hounsfield Unit (HU) threshold (below right) is also shown.

Validation and Reliability

The software for measurement of calvarial defects was validated using a 3D printed model with defects of known size. The 3D model was made using OpenSCAD (version 2015.03–1). We calculated the theoretical surface areas using the spherical zone formula [27]. The model was printed three-dimensionally on an Objet Eden260VS printer (Stratasys Ltd., Eden Prairie, MN). We then scanned our model using a Somatom Definition Flash CT scanner (Siemens AG, Malvern, PA). The model featured 4 different sized defect sizes ranging from 1 cm2 to 28.3 cm2 with an average absolute error for these 4 defects of 3.003%.

Computed tomography scans of 10 patients were selected at random using a random number generator to test for intra-rater reliability. The statistical significance of intra-rater reliability was evaluated using the intraclass correlation coefficient. The single measure intraclass correlation coefficient for the area measurement was 0.977.

References

  • 1.Neligan P, Warren RJ, and Van Beek A, Plastic surgery 3rd ed. 2013, London; New York: Elsevier Saunders. [Google Scholar]
  • 2.Fearon JA, Evidence-based medicine: Craniosynostosis. Plast Reconstr Surg, 2014. 133(5): p. 1261–75. [DOI] [PubMed] [Google Scholar]
  • 3.Greene AK, et al. , Primary grafting with autologous cranial particulate bone prevents osseous defects following fronto-orbital advancement. Plast Reconstr Surg, 2007. 120(6): p. 1603–11. [DOI] [PubMed] [Google Scholar]
  • 4.Greene AK, et al. , Pediatric cranioplasty using particulate calvarial bone graft. Plast Reconstr Surg, 2008. 122(2): p. 563–71. [DOI] [PubMed] [Google Scholar]
  • 5.Chao MT, et al. , Demineralized bone matrix and resorbable mesh bilaminate cranioplasty: a novel method for reconstruction of large-scale defects in the pediatric calvaria. Plast Reconstr Surg, 2009. 123(3): p. 976–82. [DOI] [PubMed] [Google Scholar]
  • 6.Savolainen M, et al. , Promoting ossification of calvarial defects in craniosynostosis surgery by demineralized bone plate and bone dust in different age groups. J Plast Reconstr Aesthet Surg, 2017. 70(1): p. 110–119. [DOI] [PubMed] [Google Scholar]
  • 7.Biskup NI, et al. , Pediatric cranial vault defects: early experience with beta-tricalcium phosphate bone graft substitute. J Craniofac Surg, 2010. 21(2): p. 358–62. [DOI] [PubMed] [Google Scholar]
  • 8.Canzi G, et al. , Homologous Banked Bone Grafts for the Reconstruction of Large Cranial Defects in Pediatric Patients. J Craniofac Surg, 2018. 29(8): p. 2038–2042. [DOI] [PubMed] [Google Scholar]
  • 9.Utria AF, et al. , The importance of timing in optimizing cranial vault remodeling in syndromic craniosynostosis. Plast Reconstr Surg, 2015. 135(4): p. 1077–84. [DOI] [PubMed] [Google Scholar]
  • 10.Prevot M, Renier D, and Marchac D, Lack of ossification after cranioplasty for craniosynostosis: a review of relevant factors in 592 consecutive patients. J Craniofac Surg, 1993. 4(4): p. 247–54; discussion 255–6. [PubMed] [Google Scholar]
  • 11.Paige KT, et al. , Age-dependent closure of bony defects after frontal orbital advancement. Plast Reconstr Surg, 2006. 118(4): p. 977–84. [DOI] [PubMed] [Google Scholar]
  • 12.Klinger DR, et al. , Autologous and acrylic cranioplasty: a review of 10 years and 258 cases. World Neurosurg, 2014. 82(3–4): p. e525–30. [DOI] [PubMed] [Google Scholar]
  • 13.Perlyn CA, et al. , The craniofacial anomalies archive at St. Louis Children’s Hospital: 20 years of craniofacial imaging experience. Plast Reconstr Surg, 2001. 108(7): p. 1862–70. [DOI] [PubMed] [Google Scholar]
  • 14.van Veelen-Vincent MLC, et al. , Craniosynostosis, in Neurosurgery, Lumenta CB, et al. , Editors. 2010, Springer Berlin Heidelberg: Berlin, Heidelberg: p. 501–528. [Google Scholar]
  • 15.Ghenbot RG, et al. , Effects of open and endoscopic surgery on skull growth and calvarial vault volumes in sagittal synostosis. J Craniofac Surg, 2015. 26(1): p. 161–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Jimenez DF and Barone CM, Endoscopic technique for sagittal synostosis. Childs Nerv Syst, 2012. 28(9): p. 1333–9. [DOI] [PubMed] [Google Scholar]
  • 17.Liu L, Chambers EW, Letscher D, Ju T, A simple and robust thinning algorithm on cell complexes. Computer Graphics Forum, 2010. 29(7): p. 2253–60. [Google Scholar]
  • 18.Zou M, Ju T, Carr N, An algorithm for triangulating multiple 3D polygons. Computer Graphics Forum, 2013. 32(5): p. 157–66. [Google Scholar]
  • 19.Beidas OE, et al. , Safety and efficacy of recombinant human bone morphogenetic protein 2 on cranial defect closure in the pediatric population. J Craniofac Surg, 2013. 24(3): p. 917–22. [DOI] [PubMed] [Google Scholar]
  • 20.Posnick JC, et al. , Reconstruction of skull defects in children and adolescents by the use of fixed cranial bone grafts: long-term results. Neurosurgery, 1993. 32(5): p. 785–91; discussion 791. [DOI] [PubMed] [Google Scholar]
  • 21.Cardillo G MyFisher 2010; Available from: http://www.mathworks.com/matlabcentral/fileexchange/26883.
  • 22.Alleyne B, et al. , The Impact of Age Upon Healing: Absolute Quantification of Osteogenic Genes in Calvarial Critical-Sized Defects. J Craniofac Surg, 2016. 27(1): p. 258–63. [DOI] [PubMed] [Google Scholar]
  • 23.Levi B, et al. , Dura mater stimulates human adipose-derived stromal cells to undergo bone formation in mouse calvarial defects. Stem Cells, 2011. 29(8): p. 1241–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Slater BJ, et al. , The role of regional posterior frontal dura mater in the overlying suture morphology. Plast Reconstr Surg, 2009. 123(2): p. 463–9. [DOI] [PubMed] [Google Scholar]
  • 25.Raposo-Amaral CE and Raposo-Amaral CA, Aplasia cutis congenita: impact of early treatment on calvarial osteogenesis. J Plast Reconstr Aesthet Surg, 2011. 64(9): p. e237–40. [DOI] [PubMed] [Google Scholar]
  • 26.Ang BU, et al. , Dura in the pathogenesis of syndromic craniosynostosis: fibroblast growth factor receptor 2 mutations in dural cells promote osteogenic proliferation and differentiation of osteoblasts. J Craniofac Surg, 2010. 21(2): p. 462–7. [DOI] [PubMed] [Google Scholar]
  • 27.Weisstein EW. Spherical Cap. Available from: http://mathworld.wolfram.com/SphericalCap.html.

RESOURCES