Abstract
Background
Orbital decompression is frequently performed in the management of patients with sight-threatening and disfiguring Graves’ ophthalmopathy. The quantitative measurements of the change in orbital volume after orbital decompression procedures are not definitively known. Furthermore, the quantitative effect of septal deviation on volume change has not been previously analyzed.
Objectives
To provide quantitative measurement of orbital volume change after medial and inferior endoscopic decompression and describe a straightforward method of measuring this change using open-source technologies. A secondary objective was to assess the effect of septal deviation on orbital volume change.
Methods
A retrospective review was performed on all patients undergoing medial and inferior endoscopic orbital decompression for Graves’ ophthalmopathy at a tertiary care academic medical center. Pre-operative and post-operative orbital volumes were calculated from computed tomography (CT) data using a semi-automated segmenting technique and Osirix™, an open-source DICOM reader. Data were collected for pre-operative and post-operative orbital volumes, degree of septal deviation, time to follow-up scan, and individual patient Hertel scores.
Results
Nine patients (12 orbits) were imaged before and after decompression. Mean pre-operative orbital volume was 26.99 cm 3 (SD = 2.86 cm 3 ). Mean post-operative volume was 33.07 cm 3 (SD = 3.96 cm 3 ). The mean change in volume was 6.08 cm 3 (SD = 2.31 cm 3 ). The mean change in Hertel score was 4.83 (SD = 0.75). Regression analysis of change in volume versus follow-up time to imaging indicates that follow-up time to imaging has little effect on change in volume ( R = − 0.2), and overall mean maximal septal deviation toward the operative side was − 0.5 mm. Negative values were attributed to deviation away form the operative site. A significant correlation was demonstrated between change in orbital volume and septal deviation distance site ( R = 0.66), as well as between change in orbital volume and septal deviation angle ( R = 0.67). Greater volume changes were associated with greater degree of septal deviation away from the surgical site, whereas smaller volume changes were associated with greater degree of septal deviation toward the surgical site.
Conclusion
A straightforward, semi-automated segmenting technique for measuring change in volume following endoscopic orbital decompression is described. This method proved useful in determining that a mean increase of approximately 6 cm in volume was achieved in this group of patients undergoing medial and inferior orbital decompression. Septal deviation appears to have an effect on the surgical outcome and should be considered during operative planning.
1
Introduction
Surgical decompression is a common treatment for many patients with cosmetic or functional complications of Graves’ ophthalmopathy. In 1956 Drs. Walsh and Ogura first performed the transantral Caldwell–Luc inferior and medial wall decompression that remained the mainstay of decompression surgery for many years . However, the high incidences of postoperative diplopia and infraorbital paresthesia associated with the Walsh–Ogura technique prompted the search for more effective and less morbid treatments . Numerous advances in technology and surgical techniques have subsequently increased the options available to surgeons and patients with Graves’ disease .
Surgical decompression for Graves’ ophthalmopathy involves the removal or remodeling of any combination of one or more of the medial, lateral and inferior orbital walls, and/or the removal of intraorbital fat . A number of different surgical techniques are utilized for orbital decompression. The surgical approaches can be open, through a variety of different incisions, or endoscopic transnasally. Historically, the type of decompression offered to patients is often based upon surgeon or institutional preference rather than on individual patient characteristics. As a result, decompression surgery is rarely tailored to an individual patient’s anatomy and pathology, which can be quite variable. The degree of exophthalmos can vary greatly from patient to patient, and each patient’s osseous anatomy is unique. Therefore, applying the identical surgical technique to all patients may yield vastly different results. The body of literature surrounding customized orbital decompression, tailored to particular patient anatomical and clinical characteristics, is growing . However, quantitative volumetric analysis of the change in orbital volume achieved with these tailored techniques is lacking.
Orbital volumes calculated using CT have been reported on patients with known Graves’ ophthalmopathy, and quantitative measurements correlate with clinical findings . It has been demonstrated that the increase in orbital volume correlates with the reduction in proptosis, and subsequently the patient’s cosmetic deformity, after orbital decompression . Despite the usefulness of these data, definitive measurement of the change in orbital volume following orbital decompression is lacking in the literature.
Quantitative data regarding the change in orbital volume as a result of a particular surgical technique may be useful in informing the surgical decision-making process for each patient. There exists a need for a no-cost technique that is accessible to physicians who may not necessarily have training in the manipulation and analysis of radiographic films. Osirix™ is available freely for download and provides a straightforward method to analyze orbital volumes. Its accuracy in measuring orbital volume has been demonstrated in the orbital trauma literature . Information regarding the change in orbital volume might also be useful to document the definitive outcomes achieved by the surgical procedure. Knowing the change in orbital volume after orbital decompression informs the surgeon’s decision-making process for the individual patient’s care.
While the mean change in Hertel measurement has been well document for medial and inferior orbital decompression , and while CT has been used in patients with zygomatic fractures to measure proptosis , to the best of our knowledge the mean change in orbital volume following medial and inferior endoscopic orbital decompression has never been reported. Additionally, significant septal deviation is a known challenge for any form of endoscopic nasal surgery, but the effect of septal deviation on orbital decompression outcomes has not been studied.
The purpose of this study is to describe an accessible technique for the measurement of change in orbital volumes, document the approximate change in orbital volume and analyze the effects of septal deviation on outcomes following medial and inferior endoscopic orbital decompression.
2
Methods
2.1
Patient selection
All patients who underwent medial and inferior orbital decompression for Graves’ disease by a single team of surgeons during the study period from the Department of Otorhinolaryngology–Head and Neck Surgery and Ophthalmology at Albert Einstein College were eligible. Decrease in vision and desire for improved cosmesis were the surgical indications for all patients. Approval for the study was obtained from the Montefiore Medical Center Institutional Review Board. A total of 39 orbits in 28 patients met these criteria. All patients received pre-operative CT scans of the orbit. Post-operative scans were obtained for a subset of patients with the following indications: pre-operative scanning for surgery on the contralateral orbit, evaluation of rhinosinusitis, or incidentally for other medical conditions. The remainder of the patients was excluded because post-operative scans were not available. Complete data were available for nine patients (12 orbits), who underwent both pre and post-operative CT imaging of the orbits.
2.2
Surgical technique
In each of these patients, orbital decompression was performed in the same manner utilizing image guidance (Medtronic™ Fusion™ Navigation System) with the patient under general anesthesia. After topical decompression of the nasal cavity with Oxymetazoline soaked pledgets, injection of lidocaine with epinephrine (1:100,000) was performed under direct visualization to the middle turbinate and its attachment to the lateral nasal wall. A complete uncinectomy was then performed, and the maxillary sinus entered. A wide maxillary antrostomy was performed, extending posteriorly to the posterior wall of the maxillary sinus, superiorly to the floor of the orbit, inferiorly to the attachment of the inferior turbinate and anteriorly to the lacrimal duct. The bulla ethmoidalis was removed and the posterior ethmoids entered through the basal lamella. All anterior and posterior ethmoid cells were removed and the face of the sphenoid sinus and skull base clearly identified. The agger nasi air cell, and any associated frontal sinus cells extending into the natural drainage pathway of the frontal sinus were then removed, and the frontal sinus ostium was entered to insure adequate drainage postoperatively. The lamina papyracea was removed from the face of the sphenoid posteriorly toward the lacrimal duct anteriorly and from the skull base to the strut between the lamina and the floor of the orbit. The strut between the lamina and the orbital floor was then removed and the floor of the orbit removed to the infraorbital nerve. The periorbita was incised superiorly, posteriorly and inferiorly and removed en-bloc from posterior to anterior. After removal of the periorbita, the globe was massaged and the periorbital fat gently teased with a ball probe to maximize intraorbital fat prolapse into the nasal cavity.
2.3
Imaging techniques
Helical CT imaging of the orbits was performed pre-operatively and post-operatively in the following manner. Images were obtained in the axial plane and acquired on a General Electric (Fairfield, CT, USA) LightSpeedTM multidetector CT scanner (rotation time = 0.5 seconds, 120 kV, 150 mA, 0.625 slice collimation, pitch = 0.562, zero gantry angulation with either 0.625 mm, 1 mm, or 1.250 mm cuts). CT digital imaging and communications in medicine (DICOM) images were then transferred for post-processing to a Macintosh computer and post-processing was performed using Osirix™ (a division of Pixmeo, Bernex, Switzerland), open-source DICOM reader. Regions of interests were manually selected on each axial slice through the orbit.
2.4
Calculation of orbital volume
Using a semi-automated segmenting technique, cross-sectional areas and volume of the bony orbit were calculated using the following method. Closed polygon regions of interest (ROI A) are placed by a blinded observer grossly overestimating the bony orbit on each axial section through the orbit of interest. Care is done to draw ROI A across the orbital rim (specifically excluding globe, orbital fat, and other soft tissue protruding anteriorly out of the bony confines of the orbit), across the superior and inferior orbital fissures, and across the anterior-most aperture of the optic canal. The excluded areas are then subtracted from the entire series of images leaving only the drawn regions of interest (ROI A) by setting the pixel values outside the regions of interest to − 1000 HU. Next, a second set of regions of interest (ROI B) are generated using the grow region tool from the ROI menu by first placing the cursor in the middle of the orbit and growing a region of interest using a lower threshold of − 200 HU and an upper threshold of 150 HU. This effectively selects all soft tissue and fat and excludes bone. The excluded areas are again subtracted from the series leaving only the segmented regions of interest (ROI B) by again setting the pixel values outside the regions of interest to − 1000 HU. Volumes are computed automatically from the remaining ROIs ( Fig. 1 ).
2.5
Septal deviation analysis
Nasal septal deviation angle and distance were measured on preoperative coronal images between the midline demarcated by the crista galli and the most prominent point of deviation. Septal deviation away from operative side was assigned a negative number, whereas deviation toward the operative side was assigned a positive number. Pearson coefficients were calculated to determine a correlation between nasal septal deviation measurements and change in orbital volumes. An example of angle and distance measurements is demonstrated in Fig. 2 .
2.6
Statistical analysis
Paired two-tailed Student’s t-test was used to compare preoperative and postoperative orbital volumes with a significance level of p < 0.05. The interval from pre-operative scanning to surgery, as well as from surgery to post-operative scanning, was recorded. Calculation of correlation between time from surgical intervention to follow-up imaging and change in orbital volume was done to determine if differences in follow-up time affect measurements in post-operative orbital volume.
2
Methods
2.1
Patient selection
All patients who underwent medial and inferior orbital decompression for Graves’ disease by a single team of surgeons during the study period from the Department of Otorhinolaryngology–Head and Neck Surgery and Ophthalmology at Albert Einstein College were eligible. Decrease in vision and desire for improved cosmesis were the surgical indications for all patients. Approval for the study was obtained from the Montefiore Medical Center Institutional Review Board. A total of 39 orbits in 28 patients met these criteria. All patients received pre-operative CT scans of the orbit. Post-operative scans were obtained for a subset of patients with the following indications: pre-operative scanning for surgery on the contralateral orbit, evaluation of rhinosinusitis, or incidentally for other medical conditions. The remainder of the patients was excluded because post-operative scans were not available. Complete data were available for nine patients (12 orbits), who underwent both pre and post-operative CT imaging of the orbits.
2.2
Surgical technique
In each of these patients, orbital decompression was performed in the same manner utilizing image guidance (Medtronic™ Fusion™ Navigation System) with the patient under general anesthesia. After topical decompression of the nasal cavity with Oxymetazoline soaked pledgets, injection of lidocaine with epinephrine (1:100,000) was performed under direct visualization to the middle turbinate and its attachment to the lateral nasal wall. A complete uncinectomy was then performed, and the maxillary sinus entered. A wide maxillary antrostomy was performed, extending posteriorly to the posterior wall of the maxillary sinus, superiorly to the floor of the orbit, inferiorly to the attachment of the inferior turbinate and anteriorly to the lacrimal duct. The bulla ethmoidalis was removed and the posterior ethmoids entered through the basal lamella. All anterior and posterior ethmoid cells were removed and the face of the sphenoid sinus and skull base clearly identified. The agger nasi air cell, and any associated frontal sinus cells extending into the natural drainage pathway of the frontal sinus were then removed, and the frontal sinus ostium was entered to insure adequate drainage postoperatively. The lamina papyracea was removed from the face of the sphenoid posteriorly toward the lacrimal duct anteriorly and from the skull base to the strut between the lamina and the floor of the orbit. The strut between the lamina and the orbital floor was then removed and the floor of the orbit removed to the infraorbital nerve. The periorbita was incised superiorly, posteriorly and inferiorly and removed en-bloc from posterior to anterior. After removal of the periorbita, the globe was massaged and the periorbital fat gently teased with a ball probe to maximize intraorbital fat prolapse into the nasal cavity.
2.3
Imaging techniques
Helical CT imaging of the orbits was performed pre-operatively and post-operatively in the following manner. Images were obtained in the axial plane and acquired on a General Electric (Fairfield, CT, USA) LightSpeedTM multidetector CT scanner (rotation time = 0.5 seconds, 120 kV, 150 mA, 0.625 slice collimation, pitch = 0.562, zero gantry angulation with either 0.625 mm, 1 mm, or 1.250 mm cuts). CT digital imaging and communications in medicine (DICOM) images were then transferred for post-processing to a Macintosh computer and post-processing was performed using Osirix™ (a division of Pixmeo, Bernex, Switzerland), open-source DICOM reader. Regions of interests were manually selected on each axial slice through the orbit.
2.4
Calculation of orbital volume
Using a semi-automated segmenting technique, cross-sectional areas and volume of the bony orbit were calculated using the following method. Closed polygon regions of interest (ROI A) are placed by a blinded observer grossly overestimating the bony orbit on each axial section through the orbit of interest. Care is done to draw ROI A across the orbital rim (specifically excluding globe, orbital fat, and other soft tissue protruding anteriorly out of the bony confines of the orbit), across the superior and inferior orbital fissures, and across the anterior-most aperture of the optic canal. The excluded areas are then subtracted from the entire series of images leaving only the drawn regions of interest (ROI A) by setting the pixel values outside the regions of interest to − 1000 HU. Next, a second set of regions of interest (ROI B) are generated using the grow region tool from the ROI menu by first placing the cursor in the middle of the orbit and growing a region of interest using a lower threshold of − 200 HU and an upper threshold of 150 HU. This effectively selects all soft tissue and fat and excludes bone. The excluded areas are again subtracted from the series leaving only the segmented regions of interest (ROI B) by again setting the pixel values outside the regions of interest to − 1000 HU. Volumes are computed automatically from the remaining ROIs ( Fig. 1 ).
