JOURNAL TRANSCRIPT
The Laryngoscope
C 2015 The American Laryngological, V
Rhinological and Otological Society, Inc.
Surgical Risk Factors for Recurrence of Inverted Papilloma David Y. Healy Jr., MD, CDR, MC, USN; Nipun Chhabra, MD; Ralph Metson, MD; Eric H. Holbrook, MD; Stacey T. Gray, MD Objectives/Hypothesis: To identify variations in surgical technique that impact the recurrence of inverted papilloma following endoscopic excision. Study Design: Retrospective cohort. Methods: Data from 127 consecutive patients who underwent endoscopic excision of inverted papilloma and oncocytic papilloma at a tertiary care medical center from 1998 to 2011 were reviewed. Patient demographics, comorbidities, tumor stage, and intraoperative details, including tumor location and management of the base, were evaluated to identify factors associated with tumor recurrence. Results: Recurrence of papilloma occurred in 16 patients (12.6%). Mean time to recurrence was 31.0 months (range, 5.2–110.0 months). Mucosal stripping alone was associated with a recurrence rate of 52.2% (12/23 patients), compared to 4.9% (3/61 patients) when the tumor base was drilled, 4.7% (1/21 patients) when it was cauterized, and 0.0% (0/22 patients) when it was completely excised (P 5 .001). Increased recurrence rate was associated with tumors located in the maxillary sinus (P 5 .03), as well as the performance of endoscopic medial maxillectomy (P 5 .001) and external frontal approaches (P 5 .02). Conclusions: Drilling, cauterizing, or completely excising the bone underlying the tumor base during endoscopic resection reduces the recurrence rate of inverted and oncocytic papilloma, when compared to mucosal stripping alone. Surgeons who perform endoscopic resection of these tumors should consider utilization of these techniques when possible. Key Words: Inverted papilloma, Schneiderian papilloma, paranasal sinus tumor, endoscopic sinus surgery, risk factors. Level of Evidence: 4 Laryngoscope, 126:796–801, 2016
INTRODUCTION Sinonasal inverted papilloma (IP) are benign neoplasms that demonstrate a slow but persistent growth pattern, can be locally destructive, and have potential for malignant transformation. Patients with IP typically present with unilateral nasal obstruction, but may have other symptoms such as rhinorrhea, hyposmia, epistaxis, or headache.1 IPs derive from the Schneiderian membrane, which is the mucosal lining of the nasal passage and paranasal sinuses; thus, the more descriptive nomenclature for this lesion is Schneiderian papilloma, inverted type. The tumor has a distinct histologic appearance of a thickened epithelium that lacks mucous-secreting cells From the Department of Otolaryngology (D.Y.H.), Naval Medical Center Portsmouth, Portsmouth, Virginia; Department of Surgery (N.C.), University of Illinois College of Medicine, Saint Anthony Medical Center, Rockford, Illinois; Department of Otolaryngology (R.M., E.H.H., S.T.G.), Massachusetts Eye and Ear Infirmary, Boston, Massachusetts; Department of Otology and Laryngology (R.M., E.H.H., S.T.G.), Harvard Medical School, Boston, Massachusetts, U.S.A. Editor’s Note: This Manuscript was accepted for publication August 19, 2015. The views expressed in this article are those of the author(s) and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, or the United States government. Research data were derived from a Massachusetts Eye & Ear Infirmary, Boston, Massachusetts, institutional review board protocol. The authors have no funding, financial relationships, or conflicts of interest to disclose. Send correspondence to David Y. Healy, Jr., MD, Department of Otolaryngology, Naval Medical Center Portsmouth, 620 John Paul Jones Circle, Portsmouth, VA 23708. E-mail:
[email protected] DOI: 10.1002/lary.25663
Laryngoscope 126: April 2016
796 28
and has “inverted” into the underlying stroma.2 A less common oncocytic type of Schneiderian papilloma has similar clinical presentation and behavior to IP, but has pseudostratified columnar epithelium containing eosinophilic cytoplasm on pathological examination. Optimal treatment of both inverted and oncocytic papilloma is complete surgical excision. The traditional external surgical approach of lateral rhinotomy with medial maxillectomy has given way to transnasal endoscopic resection. A systematic review of published series of IP excision demonstrated a recurrence rate of 12% for endoscopic approaches compared to 17% for external approaches.3 A meta-analysis comparing contemporary endoscopic (1992–2004) versus external approaches (1970–1995) demonstrated an improved recurrence rate in the endoscopic group (12% vs. 20%, respectively).4 Whether endoscopic or external approaches are used for access, the surgical principle is the same—complete excision of the tumor is necessary to avoid recurrence. Incomplete excision of these tumors invariably leads to recurrence, typically within 3 years.1 Modern surgical methods have managed to reduce, but not eradicate, recurrence of these tumors. Many studies have sought to identify risk factors for IP recurrence. One such risk factor appears to be tumor location. A systematic review of frontal sinus IPs demonstrated a recurrence rate of 22.4%.5 It was postulated by the authors that the technical challenges of accessing the frontal sinus caused the higher recurrence rate in this location. Other reported risk factors for recurrence Healy et al.: Surgical Risk Factors for Recurrence of IP
TABLE I. Patient Demographics. Patients
Total patients Gender Male Female Tobacco use History of smoking, quit
127
% of Total
100
Recurrence
P
16 (12.6%)
74
58.3
10 (13.5%)
NS
53
41.7
6 (11.3%)
NS
13
10.2
1 (7.7%)
NS
Current smoker
20
15.7
2 (10.0%)
NS
Tumor diagnosis Inverted papilloma
107
84.3
15 (14.0%)
NS
Oncocytic papilloma
15
11.8
0 (0.0%)
NS
IP with atypia/dysplasia Previous IP resection
5
3.9
1 (20.0%)
NS
0
94
74
1 2 or more
25 7
19.7 5.5
14 (14.9%) 1 (4.0%) 0 (0.0%)
0 1
45 49
35.4 38.6
6 (13.3%) 7 (14.2%)
2 or more
32
25.2
2 (6.3%)
Previous surgeries total* NS
*Includes all previous sinus surgeries. IP 5 inverted papilloma; NS 5 not significant.
include tobacco6; histologic parameters including hyperkeratosis, hyperplasia, and mitotic index7; and certain histochemical markers.1 Traditionally, tumor control was thought to be achieved by stripping all mucosa within the involved sinuses.8 An unproven but compelling surgical principle aimed at lowering recurrence rates is to achieve a deep oncologic margin by drilling or removing the underlying bone at the site of tumor attachment (i.e., the tumor base). Evidence that this surgical maneuver may be important is suggested from a pathologic analysis of the bone at the base of IPs. In one study, Chiu et al. found mucosal tissue to be embedded within a bony crevice in two of nine (22.2%) surgical IP specimens.9 Despite the theoretical advantage of using a drill or other ablative method along the IP base during surgery, these techniques have yet to be shown to reduce tumor recurrence rates. The purpose of this study was to determine whether or not the use of such methods intended to enhance surgical control of the deep tumor margins do, in fact, impact the rate of IP recurrence following endoscopic resection.
MATERIALS AND METHODS A retrospective analysis was conducted of all surgeries (n 5 135) performed to remove sinonasal inverted or oncocytic papillomas by three rhinologists (R.M., E.H.H., S.T.G..) at the Massachusetts Eye & Ear Infirmary (MEEI) from January 1998 through February 2012. Complete endoscopic excision or hybrid endoscopic/external approaches were utilized for all surgeries. Patients underwent at least 3 years of endoscopic tumor surveillance after surgery (mean length of follow-up 8.1 years). For the purpose of simplicity, inverted papilloma is used to
Laryngoscope 126: April 2016
mean both inverted and oncocytic papilloma unless otherwise specified. Exclusion criteria included: 1) surgeries that were known to be incomplete excisions at the conclusion of the procedure (n 5 3), 2) surgeries for malignant lesions or the presence of malignancy in pathologic specimen (n 5 1), 3) insufficient documentation (n 5 2), 4) no IP found in the pathologic specimen (n 5 1), 5) suspected but unproven recurrence due to loss of follow-up (n 5 1). In three cases, surgery was staged, because of the unexpected need for an external approach (n 5 1) or excessive blood loss (n 52). These staged procedures were treated as one surgery, with the higher estimated blood loss (EBL) used for statistical analysis. Surgical risk factors analyzed included method of tumor base deep margin control (drilling, removal, cauterization, or stripping of overlying mucosa alone), location of tumor, location of tumor base (as detailed in the operative report), previous surgeries, EBL, presence of bilateral polyps, and use of image guidance. If the operation report did not give the location of the tumor base, the location of the base was annotated as “not detailed,” and analyzed as a potential risk factor. Additional risk factors analyzed are given in Tables I and II. Four methods were utilized for surgical management of the tumor base: 1) mucosal stripping alone, 2) drilling the tumor base, 3) cauterization of the tumor base, or 4) completely resecting the tumor base (i.e., septectomy, middle turbinectomy, or removal of the lamina papyracea when the tumor is based on bony structures that can be completely excised).
Description of Statistical Methods Statistical analyses were performed using R software (R Foundation for Statistical Computing, Vienna, Austria). A stepwise multivariate logistic regression model was created with the binary outcome defined as no recurrence or recurrence of IP. Both the Fisher exact test and logistic regression analysis were used to analyze the effect of the various surgical risk factors. Institutional review board approval for this study was given by the Human Study Committee of the MEEI.
RESULTS The study population consisted of 127 patients with a mean age of 56.9 years (range, 22.0–82.3) as shown in Table I. Previous sinonasal surgery was performed prior to the diagnosis of IP in 65 (51.2%) patients. Previous surgery for a known diagnosis of IP was performed by outside surgeons in 32 (25.2%) patients. Forty-five patients (35.4%) had no prior history of endoscopic sinus surgery. The most common presenting symptom was nasal obstruction (63.3%), whereas 24.2% of patients were asymptomatic, with IP as an incidental finding on imaging or endoscopy. Other presenting symptoms included rhinorrhea/postnasal drip (8.7%), epistaxis (8.7%), pain or pressure (8.7%), sinusitis (7.9%), hyposmia (4.7%), visual disturbance (2.3%), and facial swelling (1.5%). Tumors were most commonly located (but not necessarily based) in the nasal cavity (58.3%), followed by the maxillary sinus (46.5%), the ethmoid sinuses (41.7%), the frontal sinus (17.3%), and the sphenoid sinus (11.0%). The tumor base (or the site of tumor attachment) was widely dispersed throughout the paranasal sinuses, with the classically reported location for IP based along the lateral Healy et al.: Surgical Risk Factors for Recurrence of IP
797 29
TABLE II. Location of Tumor and Base of Tumor. Patients
% of Total
Recurrence
P
Nasal cavity Ethmoid sinus
74 53
58.3 41.7
7 (9.5%) NS 9 (17.0%) .088
Maxillary sinus
59
46.5
10 (16.9%) .033
Sphenoid sinus Frontal sinus
14 26
11 20.5
1 (7.1%) 4 (15.3%)
NS NS
Nasal cavity Total
18
14.2
1 (5.6%)
NS
Septum
6
4.7
0 (0.0%)
NS
12
9.4
1 (8.3%)
NS
Total
53
41.7
4 (7.5%)
NS
Uncinate Lateral Wall
5 24
3.9 18.9
0 (0.0%) 3 (12.5%)
NS NS
Turbinate
15
11.8
1 (6.7%)
NS
Anterior sphenoid face Roof
5 14
3.9 11
0 (0.0%) 0 (0.0%)
NS NS
Total Anterior wall
56 15
44.1 11.8
6 (10.7%) 1 (6.7%)
NS NS
Posterior wall
11
8.7
1 (9.1%)
NS
Medial wall Lateral wall
16 8
12.6 6.3
4 (25.0%) 0 (0.0%)
NS NS
Superior wall
25
19.7
2 (8.0%)
NS
Floor Sphenoid sinus
4
3.1
1 (25.0%)
NS
Tumor location
Tumor base location
Lateral nasal wall Ethmoid cavity
frontal stenosis (5.5%). Uncommon complications included cerebrospinal fluid (CSF) leak, acute sinusitis, epistaxis, diplopia, facial cellulitis, sublabial hematoma, mucocele formation, and epiphora, which occurred in one or two patients each. Tumor recurrence occurred in 16 patients (12.6%) as shown in Table IV. Fifteen patients had recurrence of benign IP, and one patient had a malignant conversion to Schneiderian carcinoma. The average time from surgery to recurrence was 27 months (range, 6–110 months). Of the 15 patients with benign tumor recurrence, 12 underwent one revision surgery, and one underwent three revision surgeries. Both the patient with malignant transformation and the patient requiring multiple revision surgeries had extensive tumor involvement of the frontal sinus. Mucosal stripping alone without drilling, cauterizing, or removing the bone underlying the tumor base was associated with a recurrence rate of 52.2% (12/23 patients), compared to 4.9% (3/61) in patients where the
TABLE III. Surgical Details.
Maxillary sinus
Total
9
7.1
0 (0.0%)
NS
Anterior/inferior wall Posterior/superior wall
2 4
1.6 3.1
0 (0.0%) 0 (0.0%)
NS NS
Lateral wall
4
3.1
0 (0.0%)
NS
Medial wall Frontal sinus
1
0.8
0 (0.0%)
NS
26
20.5
4 (15.4%)
NS
Posterior wall Anterior wall
9 2
7.1 1.6
1 (11.1%) 0 (0.0%)
NS NS
Medial wall
1
0.8
0 (0.0%)
NS
1 11
0.8 8.7
0 (0.0%) NS 4 (36.7%) .032
Supraorbital cell
1
0.8
0 (0.0%)
Not detailed
8
6.3
5 (62.5%) .013
Total
Interfrontal cell Floor
NS
NS 5 not significant.
Total % of Total Recurrences
Surgery performed Frontal drillout (Draf III)
14
11
2 (14.3%)
Medial maxillectomy
30
23.6
8 (26.7%) .005
Denker’s procedure Caldwell-Luc
2 22
1.6 17.3
0 (0.0%) 3 (13.7%)
Trephine or osteoplastic flap
17
13.4
4 (23.5%) .024 12 (52.2%) .001*
Status of tumor base Mucosa stripped
Laryngoscope 126: April 2016
798 30
NS NS NS
23
18.1
Bone removed
22
17.3
0 (0.0%)
Bone drilled Bone cauterized
61 21
48 16.5
3 (4.9%) 1 (4.8%)
3 102
2.4 80.3
0 (0.0%) 15 (14.7%)
73 54
57.5 42.5
5 (6.8%) 11 (20.4%)
NS NS
108 17
85 13.4
12 (11.1%) 3 (17.7%)
NS
Bilateral
2
1.6
1 (50.0%)
Tumor stage‡ A
51
40.2
7 (13.7%)
74
58.3
8 (10.8%)
2
1.6
1 (50.0%)
Surgical details Staged surgery Image guidance used Mucosal margins† Not confirmed Confirmed negative Presence of polyps None Unilateral
B
nasal wall only occurring in 9.4% of patients in this series (Table II). All surgeries included a transnasal endoscopic approach (Table III). Thirty-nine patients (30.7%) underwent additional external procedures (Caldwell-Luc, external trephine, or osteoplastic flap without obliteration). Mean EBL was 279.1 mL (range, 10–1,350 mL, standard deviation 233.4 mL). Twenty-seven patients (21.2%) experienced complications, which included facial numbness (6.3%) and
P
C Lund-Mackay CT score
NS