Maxillary reconstruction using the scapular tip free flap: A radiologic comparison of 3D morphology

Size: px

Start display at page:

Download "Maxillary reconstruction using the scapular tip free flap: A radiologic comparison of 3D morphology"

Sybil Simpson
6 years ago
Views:

1 ORIGINAL ARTICLE Maxillary reconstruction using the scapular tip free flap: A radiologic comparison of 3D morphology Nitin A. Pagedar, MD, 1 Ralph W. Gilbert, MD, 2 * Harley Chan, PhD, 3 Michael J. Daly, 3 Jonathan C. Irish, MD, 2 Jeffrey H. Siewerdsen, PhD 3,4 1 Department of Otolaryngology Head and Neck Surgery, University of Iowa, Iowa City, Iowa, 2 Department of Otolaryngology Head and Neck Surgery, Wharton Head and Neck Centre, Princess Margaret Hospital, Toronto, Ontario, Canada, 3 Ontario Cancer Institute, Princess Margaret Hospital, Toronto, Ontario, Canada, 4 Department of Biomedical Engineering, Johns Hopkins University, Baltimore, Maryland. Accepted 3 August 2011 Published online 27 January 2012 in Wiley Online Library (wileyonlinelibrary.com). DOI /hed ABSTRACT: Background. Scapular tip osteomyogenous free flaps have been described for complex palate reconstruction. Minimal osteotomies are needed because of the similar shapes of the scapula and palate. We compared the bony morphology of the palate and scapular tip to determine the suitability of the scapular tip for palate reconstruction. Methods. We analyzed facial and chest CT images of 10 patients, comparing the morphology of 3 simulated palate resection specimens (total palate, subtotal palate, and hemipalate) with corresponding simulated scapular tip bone flaps from the same patient. Results. Conformance distances between palates and simulated flaps were small, indicating close shape similarity. Median conformance distances were 3.44 mm for hemipalatectomy, 3.56 mm for subtotal palatectomy, and 3.71 mm for total palatectomy. Six outlier observations accrued from 2 patients. Conclusions. Based on this analysis, there is close similarity between the shapes of the palate and the scapular tip. This similarity supports use of the scapular tip flap for selected palate defects. VC 2012 Wiley Periodicals, Inc. Head Neck 34: , 2012 KEY WORDS: maxillectomy, free flap, scapula, morphology, reconstruction Maxillary defects represent a unique challenge for reconstructive surgeons. Defects typically result in hypernasal speech, dysphagia, and malposition of the midfacial and orbital soft tissues. Historically, patients have been managed with maxillary prostheses, which separate the oral and nasal cavities. However, prostheses have several drawbacks. They require daily care and attention and, in extensive resections, are not stable enough to support mastication or provide optimal support of orbital and midfacial soft tissues. 1 For these reasons, options for surgical reconstruction have been increasingly explored. Multiple techniques have been described for bony reconstruction of maxillary defects. For defects involving the palate and maxillary infrastructure, the goals of reconstruction are to maintain separation of the oral and nasal cavities, allow fitting of a denture stable enough for mastication, and potentially provide for osseointegrated implant *Corresponding author: R. W. Gilbert, Department of Otolaryngology Head and Neck Surgery, Wharton Head and Neck Centre, Princess Margaret Hospital, Toronto, Ontario, Canada. ralph.gilbert@uhn.on.ca Contract grant sponsor: Princess Margaret Hospital Foundation; contract grant sponsor: National Institutes of Health; contract grant number: R Poster at the American Head and Neck Society Annual Meeting, May 27, 2009, Scottsdale, Arizona. Jeffrey H. Siewerdsen has a conflict of interest with Carestream Health Inc. in which he is on the Medical Advisory Board and had a research grant collaboration. He is also on the International Adivsory Board and had a research grant collaboration with Siemens Healthcare. placement. In addition, the reconstruction should be robust enough to withstand adjuvant radiotherapy or survive in a radiated wound, making vascularized bone a necessity. The radial forearm, 2 fibula, 3 and iliac crest 4 osteocutaneous free flaps have been described for this purpose. Complex osteotomies must be created in these bone flaps to re-create the 3-dimensional (3D) structure of the palate and maxilla. Recent work describes results using scapular tip myoosseous free flaps in selected patients. 5 The scapular tip, along with small portions of the infraspinatus and subscapularis muscles, is harvested based on the angular branch of the thoracodorsal artery and vena comitans (see Figure 1). The exposed muscle is left to mucosalize in the oral cavity. The bone of the lateral scapular border is in the position of the resected ipsilateral alveolar ridge, and might then accept osseointegrated implants. In our experience, an important advantage of the scapular tip flap is its morphologic similarity to the palate, which allows for effective bony reconstruction with a simple inset. We sought to demonstrate the 3D morphologic similarity of the palate and the scapula in individual patients by performing a quantitative conformance analysis based on CT representations of the bones. MATERIALS AND METHODS We performed a retrospective analysis of CT images obtained from 10 patients without known palatal or skeletal disease. We obtained approval of the Research Ethics HEAD & NECK DOI /HED OCTOBER

PAGEDAR ET AL. FIGURE 1. Sample location of scapular osteotomy used for hemipalate reconstruction. The right scapula is shown from a posterior view. The dashed line demonstrates the osteotomy.

Scans were obtained using multidetector CT scanners with standard protocols exhibiting nearly isotropic 3D spatial resolution for the facial bones and the chest, respectively.

2 PAGEDAR ET AL. FIGURE 1. Sample location of scapular osteotomy used for hemipalate reconstruction. The right scapula is shown from a posterior view. The dashed line demonstrates the osteotomy. Board of the University Health Network, Toronto, prior to performing this study. Each patient had a CT scan through the palate and of the chest on the same day. Scans were obtained using multidetector CT scanners with standard protocols exhibiting nearly isotropic 3D spatial resolution for the facial bones and the chest, respectively. The steps of the conformance analysis procedure are depicted for 1 patient in Figure 2. The images were imported to a workstation, and semiautomatic segmentation was performed using 3D visualization software (Mimics v12.0; Materialise, Ann Arbor, MI). First, bone and soft tissue were differentiated by intensity threshold. Regions of bone specific to the palate, left scapula, and right scapula were then defined automatically by regionbased segmentation and then smoothed and edited to remove obvious errors due to nonisotropic voxels, CT artifacts, and image noise. In dentate patients, the crowns of the teeth were manually excluded from the palate volume. The palate and scapula surfaces were then manually coregistered. The first criterion for registration was to obtain the best qualitative match of the anterior profile between the palate and scapular tip, matching the anterior maxillary spine to the tip of the scapular angle. Next, the registration was adjusted to obtain the best qualitative match to the location and curvature of the palate. After registration of the anterior and inferior surfaces, the posterior aspect of the scapula was digitally cut such that the scapular tip was of similar anterior posterior length as the palate (as viewed from above). We have previously described the procedure by which radiographic volumes are compared, by creation of conformance maps. 6 Briefly, the surfaces of the scapular and palate volumes are represented by triangular lattices. With the surfaces in their optimal coregistered positions, the distance between each vertex point on the palate surface and the nearest 3 vertex points on the scapular surface are computed, thus defining a conformance distance at each vertex point on the palate. The resulting distances are presented in a color-coded map referred to herein as a conformance map, depicting the morphologic discrepancy between the scapula and the palate. For each comparison, the root-mean-square (RMS) conformance was calculated as a summary metric for morphologic similarity. In the clinical setting of hemipalate and subtotal palate reconstruction, the orientation of the bone is determined by the vascular pedicle, which must emerge laterally. In these cases, the superficial surface of the scapula and infraspinatus muscle faces the oral cavity; the lateral border of the scapula is therefore placed in the position of the alveolar process. For the total palate case, the bone is placed with the opposite orientation, with the cut subscapularis muscle facing the oral cavity. Therefore, we FIGURE 2. Scheme of image registration procedure. In all illustrations, the palate is shown in magenta, the left scapula in yellow, and the right scapula in green HEAD & NECK DOI /HED OCTOBER 2012

3 MAXILLARY RECONSTRUCTION USING THE SCAPULAR TIP FREE FLAP FIGURE 3. Left: Coregistration of palate and both scapula volumes for patient 1. Right: Color-coded conformance map comparing palate and scapula volumes. analyzed 6 scenarios for each patient: right scapula to reconstruct right hemipalate, right subtotal palate (up to and including the left canine tooth), and total palate; and left scapula for reconstruction of the corresponding 3 leftsided defects. The other combinations were not considered clinically relevant. Mean conformances were compared with the Wilcoxon rank-sum test when appropriate. Conformance outliers were identified as those with conformance distance >1.5-fold the interquartile range for that comparison. 8 Outlier observations were included in all analyses. Descriptive statistics were computed using R version (R Foundation for Statistical Computing, Vienna, Austria, RESULTS Figure 3 demonstrates coregistration for a total palate defect using right and left scapulas in patient 1. The conformance map shows close conformance of the palate and both scapulas, typically approximately 1- to 4-mm distance between surfaces, with somewhat improved conformance suggested by the right scapula for this case. The areas of poorest conformance, shown in red, lie mainly in the inferior aspect of the alveolar ridge and the teeth, and are therefore not relevant to palate reconstruction. Similarly, Figures 4 and 5 demonstrate conformance maps for subtotal palate and hemipalate defects, respectively. In each case, most of the conformance discrepancy occurred at the alveolar ridges, and conformance was approximately 1 to 4 mm throughout most of the relevant volume. Qualitatively, the superimposed images demonstrate a reasonable morphologic similarity between the natural shapes of the palate and the scapular tip. Table 1 lists the RMS conformance distances for simulated reconstructions of total palate, subtotal palate, and hemipalate defects for each patient. Median RMS conformance over 10 patients was 3.71 mm for the left scapula and 3.85 mm for the right scapula. There was no statistically significant difference between mean conformances of right and left scapula for total palate reconstruction (p ¼.47). Patient 4 was an outlier for both left and right scapular reconstruction, as was the left scapula of patient 6. For subtotal defects, the left scapula exhibited median RMS conformance of 3.46 mm, whereas the right scapula showed median RMS conformance of 3.67 mm. For hemipalate defects, the median RMS conformance was 3.49 mm for the left side and 3.42 mm for the right side. Patient 6 was an outlier for left scapula subtotal palate. Both patients 4 and 6 were outliers for the right scapula hemipalate case. Examination of the coregistered shapes revealed that the scapula in patient P4 was smaller than the palate, whereas the opposite was the case for patient P6. HEAD & NECK DOI /HED OCTOBER

4 PAGEDAR ET AL. FIGURE 4. Coregistration and conformance maps for subtotal palate resection and both scapulas, patient 1. DISCUSSION There is no consensus on the optimal method for palate reconstruction. Free tissue transfer procedures can offer significant benefits to patients, including freedom from the need to manage a prosthesis, improved support of midfacial soft tissues, and potentially more effective mastication either by better denture stability or by osseointegrated implants. The scapular tip free flap provides a donor site with a favorable long-term morbidity profile. Flap anatomy is reliable, the vascular pedicle is of large caliber, and vessels do not usually suffer from atherosclerotic disease. Our group has previously reported on clinical outcomes with this flap, including excellent upper extremity related quality of life. 5 The purpose of the current study was to compare the morphology of the palate with that of the bony component of the scapular tip flap. Qualitatively, the coregistered images show a definite morphologic similarity, which translates clinically into a bone flap that effectively restores the contours of the palate and anteroinferior maxilla with little intraoperative or postoperative modification. Simple measurements of defect width and length are applied at the time of harvest. The myo-osseous nature of the flap means that the shape of the bone is the main determinant of the final reconstructed contour for infrastructure defects. Furthermore, the measured conformance distances of the simulated reconstructions likely overestimate the true morphologic difference between palate and scapular tip due to a conformance discrepancy in 2 areas: the alveolar ridges and the nasal surface of the maxilla. Both areas are functionally important, but the reconstruction need not replicate their morphology. The lateral border of the scapula can support standard dental prostheses and possibly osseointegrated implants, even though it is not shaped like an alveolar process. The nasal surface of the reconstruction does not have the contours of the maxilla, but these are not important for reconstruction. As described earlier, the anterior projection of the maxilla is reproduced by the reconstruction. If we were to deliberately exclude these areas from analysis, the conformance would be even greater. An important finding of this work was that 2 patients contributed outlier data points. The scapulas of patient 6 were wider than the palate, whereas those of patient 4 were small relative to the palate. Since this study used deidentified patient images, it was not possible to correlate outlier status with patient characteristics such as body mass index or dominant hand. However, our reported experience to date, with 39 patients, has not revealed size mismatch to be a significant problem. 8 Converting RMS conformance into a framework that can be appreciated clinically is difficult. Although these findings 1380 HEAD & NECK DOI /HED OCTOBER 2012

MAXILLARY RECONSTRUCTION USING THE SCAPULAR TIP FREE FLAP FIGURE 5. Coregistration and conformance maps for hemipalate resection and both scapulas, patient 1.

5 MAXILLARY RECONSTRUCTION USING THE SCAPULAR TIP FREE FLAP FIGURE 5. Coregistration and conformance maps for hemipalate resection and both scapulas, patient 1. support favorable morphology, placing a number on conformance may not help in the clinical setting. It is likely that a visual display, as shown in Figure 3, will help the operating surgeon more than a measurement, especially if TABLE 1. RMS conformance distances for left and right scapular reconstructions of 3 simulated palate resections in 10 patients. RMS conformance distances, mm Total palate Subtotal palate Hemipalate Left Right Left Right Left Right Patient scapula scapula scapula scapula scapula scapula P P P P P P P P P P Median Range SD Abbreviation: RMS, root-mean-square. intuitive manipulation of the contoured segments is available. The quantitative morphologic comparisons in this study are absolute in nature and do not support a relative comparison to alternative bone flap reconstructions. The complex osteotomies required for other bone flaps obscure the meaning of the comparison. The main value of this work is the quantitative confirmation of our clinical experience, that the scapular tip flap provides a close morphologic approximation of the palate. This study supports our clinical experience that the scapular tip provides a morphologically similar and effective reconstructive framework for palate reconstruction. The coregistration process described has potential to allow surgeons to refine surgical planning for complex bony reconstruction. Acknowledgements The authors thank the clinicians and scientists in the Guided Therapeutics Program at the University Health Network (Toronto, ON). REFERENCES 1. Cordeiro PG, Santamaria E, Kraus DH, Strong EW, Shah JP. Reconstruction of total maxillectomy defects with preservation of the orbital contents. Plast Reconstr Surg 1998;102: Cordeiro PG, Bacilious N, Schantz S, Spiro R. The radial forearm osteocutaneous sandwich free flap for reconstruction of the bilateral subtotal maxillectomy defect. Ann Plast Surg 1998;40: HEAD & NECK DOI /HED OCTOBER

6 PAGEDAR ET AL. 3. Chang DW, Langstein HN. Use of the free fibula flap for restoration of orbital support and midfacial projection following maxillectomy. J Reconstr Microsurg 2003;19: Brown JS, Jones DC, Summerwill A, et al. Vascularized iliac crest with internal oblique muscle for immediate reconstruction after maxillectomy. Br J Oral Maxillofac Surg 2002;40: Clark JR, Vesely M, Gilbert R. Scapular angle osteomyogenous flap in postmaxillectomy reconstruction: defect, reconstruction, shoulder function, and harvest technique. Head Neck 2008;30: Chan H, Gilbert RW, Pagedar NA, Daly MJ, Irish JC, Siewerdsen JH. A new method of morphological comparison for bony reconstructive surgery: maxillary reconstruction using scapular tip bone. SPIE Medical Imaging 2010: Visualization, Display, and Image-Guided Procedures 2010;7625: 76253B. 7. Tukey JW. Exploratory data analysis. Reading, MA: Addison Wesley; Miles B, Gilbert RW. Maxillary reconstruction with the scapular angle osteomyogenous free flap. Presented at: American Head and Neck Society Annual Meeting, Chicago, IL, April 27, HEAD & NECK DOI /HED OCTOBER 2012

Scapular & Parascapular flap FLAP TERRITORY ANATOMY. is normally accompanied by two venae comitantes.

Scapular & Parascapular flap FLAP TERRITORY This is a composite flap that is situated over the scapula with various incisional arrangements. It can be harvested as a skin and subcutaneous tissue flap,