- Email: [email protected]
Treatment of nasal cavity and paranasal sinus cancer with modern radiotherapy techniques in the postoperative setting—the MSKCC experience
Int. J. Radiation Oncology Biol. Phys., Vol. 67, No. 3, pp. 691–702, 2007 Copyright © 2007 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/07/$–see front matter
doi:10.1016/j.ijrobp.2006.09.023
CLINICAL INVESTIGATION
Head and Neck
TREATMENT OF NASAL CAVITY AND PARANASAL SINUS CANCER WITH MODERN RADIOTHERAPY TECHNIQUES IN THE POSTOPERATIVE SETTING—THE MSKCC EXPERIENCE BRADFORD S. HOPPE, M.D.,* LAUREN D. STEGMAN, M.D., PH.D.,* MICHAEL J. ZELEFSKY, M.D.,* KENNETH E. ROSENZWEIG, M.D.,* SUZANNE L. WOLDEN, M.D.,* SNEHAL G. PATEL, M.D.,† JATIN P. SHAH, M.D.,† DENNIS H. KRAUS, M.D.,† AND NANCY Y. LEE, M.D.* Departments of *Radiation Oncology and †Head and Neck Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY Purpose: To perform a retrospective analysis of patients with paranasal sinus (PNS) cancer treated with postoperative radiotherapy (RT) at Memorial Sloan-Kettering Cancer Center. Methods and Materials: Between January 1987 and July 2005, 85 patients with PNS and nasal cavity cancer underwent postoperative RT. Most patients had squamous cell carcinoma (49%; n ⴝ 42), T4 tumors (52%; n ⴝ 36), and the maxillary sinus (53%; n ⴝ 45) as the primary disease site. The median radiation dose was 63 Gy. Of the 85 patients, 76 underwent CT simulation and 53 were treated with either three-dimensional conformal RT (27%; n ⴝ 23) or intensity-modulated RT (35%; n ⴝ 30). Acute and late toxicities were scored according to the Radiation Therapy Oncology Group radiation morbidity scoring criteria. Results: With a median follow-up for surviving patients of 60 months, the 5-year estimates of local progressionfree, regional progression-free, distant metastasis-free, disease-free, and overall survival rates were 62%, 87%, 82%, 55%, and 67%, respectively. On multivariate analysis, squamous cell histology and cribriform plate involvement predicted for an increased likelihood of local recurrence, and squamous cell histologic features predicted for worse overall survival. None of the patients who underwent CT simulation and were treated with modern techniques developed a Grade 3-4 late complication of the eye. Conclusion: Complete surgical resection followed by adjuvant RT is an effective and safe approach in the treatment of PNS cancer. Emerging tools, such as three-dimensional conformal treatment and, in particular, intensity-modulated RT for PNS tumors, may minimize the occurrence of late complications associated with conventional RT techniques. Local recurrence remains a significant problem. © 2007 Elsevier Inc. Paranasal sinus, Surgery, Intensity-modulated radiotherapy, Radiotherapy, Cancer.
with chemotherapy. The outcomes of these unresectable patients treated with definitive RT with or without chemotherapy have been dismal, largely owing to the limitations in the radiation dose that can be safely delivered without causing injuries to the nearby critical normal tissues, such as the brainstem and optic structures (2–7). Radiotherapy techniques have evolved during the past two decades from conventional RT, to three-dimensional conformal RT (3D-CRT), to intensity-modulated RT (IMRT) with “dose-painting” and proton therapy. These newer techniques have allowed for improved dose distributions, with increased dose to the target volumes and reduced dose to the surrounding normal tissues compared with conventional treatment. One of the sites that can benefit from these advances in RT techniques is cancer involving the PNSs. Historical studies in the treatment of PNS cancers
INTRODUCTION Paranasal sinus (PNS) and nasal cavity cancers are rare, making up only 5% of head-and-neck cancers and ⬍1% of all malignancies (1). Because of the rarity of these tumors, clinical data regarding their management have been largely from small retrospective experiences from single institutions. These studies have shown surgical resection with clear margins offers an excellent chance of cure in earlystage tumors. However, in more advanced tumors, achieving clear margins is often difficult and, therefore, adjuvant radiotherapy (RT) is required. At the Memorial Sloan-Kettering Cancer Center (MSKCC), patients with PNS cancer who are surgical candidates are treated with surgery followed by RT. Tumors not amenable to complete surgical resection because of invasion of the dura, brain, orbit, or nasopharynx, are managed with RT, often in combination Reprint requests to: Nancy Y. Lee, M.D., Department of Radiation Oncology, Memorial Sloan-Kettering Cancer Center, 1275 York Ave., New York, NY 10021. Tel: (212) 639-3341; Fax: (212) 639-2417; E-mail: [email protected]
Conflict of interest: none. Received July 12, 2006, and in revised form Sept 11, 2006. Accepted for publication Sept 13, 2006. 691
692
I. J. Radiation Oncology
●
Biology
●
Physics
using conventional RT techniques have reported up to 35% of patients developing severe visual impairment as a consequence of RT (6). Recent studies of modern RT techniques have shown fewer patients developing visual complications; however, they lacked long-term follow-up (8 – 10). Visual impairment from RT and local failure often occur several years after treatment, and only with sufficient follow-up will these studies effectively assess the outcomes and severe toxicities. This study reports the MSKCC’s 18-year experience with PNS and nasal cavity tumors, treated with surgical resection and RT, predominantly using modern RT techniques, and compares the results with those of other PNS studies. Prognostic factors that may predict a worse outcome were evaluated to identify patients who may need more aggressive upfront therapy. Finally, we describe the current technique used at the MSKCC of IMRT dose painting for patients with PNS cancers. METHODS AND MATERIALS Between January 1987 and July 2005, 232 patients with a diagnosis of PNS cancer evaluated in the Department of Radiation Oncology were identified from the institution’s database. Patients were excluded from the analysis if they had received definitive RT or chemoradiotherapy (CTRT) (n ⫽ 51), had received their RT at another facility (n ⫽ 50), had the histologic features of melanoma, rhabdomyosarcoma, plasmacytoma, Merkel cell carcinoma, or noninvasive disease (in situ) (n ⫽ 40), had a history of retinoblastoma with previous RT (n ⫽ 4), or had not received RT to the primary disease site (n ⫽ 2). The present analysis included 85 PNS cancer patients, who had undergone surgical resection followed by postoperative RT.
Table 1. Patient characteristics Parameter Gender Male Female Ethnicity White Black Asian Other Site Maxillary sinus Nasal cavity Ethmoid sinus Frontal sinus Sphenoid sinus Histologic type Squamous cell carcinoma Adenoid cystic Sarcoma Esthesioneuroblastoma Adenocarcinoma Undifferentiated Neuroendocrine Other
n (%) 53 (62) 32 (38) 68 (80) 7 (8) 7 (8) 3 (4) 45 (53) 24 (28) 14 (16) 1 (1) 1 (1) 42 (49) 11 (13) 9 (11) 7 (8) 6 (7) 4 (5) 2 (3) 4 (5)
Volume 67, Number 3, 2007
Table 2. Staging (excluding sarcomas) Staging system AJCC (n ⫽ 69) Recurrent T stage T1 T2 T3 T4 N stage N0 N1 N2 Kadish (n ⫽ 7, esthesioneuroblastoma) A B C
n (%) 5 (7) 0 8 (12) 20 (29) 36 (52) 64 (93) 5 (7) 0 2 (29) 4 (57) 1 (14)
Abbreviation: AJCC ⫽ American Joint Committee on Cancer.
Patient characteristics The patient age-range was 10 – 82 years (median, 54), and the Karnofsky performance status was 70 –100 (median, 80). Other patient characteristics are listed in Table 1. Most patients were male (62%) and white (80%). The subsite of origin was determined as the region harboring the predominant bulk of disease and was most commonly in the maxillary sinus (53%). Squamous cell carcinoma (SCC) (49%) was the most frequent histologic type.
Evaluation and staging The pretreatment evaluation included a complete history and physical examination, direct flexible fiberoptic endoscopic examination, complete blood count, liver function tests, chest X-ray, CT, and dental evaluations. Of the 85 patients, 47 underwent additional MRI and 11 underwent 18F-fluorodeoxyglucose positron emission tomography (PET). For this study, patients were retrospectively restaged using the 2002 American Joint Committee on Cancer TNM staging system (11) (Table 2), with the exception of esthesioneuroblastomas, which underwent Kadish staging. Sarcomas were not staged. Patients predominantly presented with T4 disease (52%; n ⫽ 36). Five patients with local relapse after surgical resection alone were included in this analysis, including two adenoid cystic carcinomas, one SCC, one adenocarcinoma, and one myoepithelial carcinoma. At recurrence, these patients underwent radical surgical resection followed by postoperative RT. Clinical cervical lymph node metastasis was seen in 5 patients at presentation, including 4 patients with SCC and 1 with adenocarcinoma, and was located in the neck Level 1 region in 1 patient and the Level 2 region in 4.
Surgery All patients underwent gross surgical resection. The surgical techniques ranged from subtotal maxillectomy to cranial facial resection. Most patients had negative surgical margins (62%; n ⫽ 53). Surgical margins were recorded as negative, unless otherwise stated by the surgeon, pathologist, or treating radiation oncologist. Of the 85 patients, 21 had perineural invasion. Only the 5 patients who presented with clinical nodal involvement underwent neck dissection as part of their primary treatment. Thirteen patients underwent orbital exenteration as part of their primary surgery
Treatment of nasal cavity and PNS cancer with postoperative RT
Value
Surgery Craniofacial resection Orbital exenteration Yes No Neck dissection Yes No Chemotherapy Preoperative Concurrent RT technique Conventional Conventional ⫹ Chemotherapy 3D-CRT IMRT Radiation dose (Gy) Median Range Neck irradiated No Upper Whole
B. S. HOPPE et al.
693
patients had partial responses radiographically; for the third patient, no imaging was available before the start of chemotherapy. Disease was evident in the surgical specimen in all 3 cases. Two of these patients had microscopically positive margins. Two additional patients received concomitant CTRT after resection of their primary tumor (one because of rapid disease progression postoperatively). These patients received cisplatin alone or in combination with gemcitabine and had SCC histologic features.
Table 3. Treatment variables Variable
●
34 (40) 13 (15) 72 (85) 5 (6) 80 (94)
Radiotherapy
3 (4) 2 (2) 9 (11) 23 (27) 23 (27) 30 (35) 63 40–70 75 (88) 4 (5) 5 (6)
Abbreviations: RT ⫽ radiotherapy; 3D-CRT ⫽ three-dimensional conformal RT; IMRT ⫽ intensity-modulated RT. Data presented as number of patients, with percentages in parentheses, unless otherwise noted.
because of involvement of the orbital contents. The details of the surgical interventions are listed in Table 3.
Chemotherapy Systemic chemotherapy was used in the initial treatment of 5 patients (Table 3). In 3 pediatric patients, neoadjuvant chemotherapy was used, including combinations of cisplatin, gemcitabine, doxorubicin, methotrexate, cyclophosphamide, vincristine, and etoposide. These pediatric patients had osteosarcoma, sinonasal undifferentiated carcinoma, and esthesioneuroblastoma. Two
Radiotherapy was performed with conventional planning (11%, n ⫽ 9; median follow-up 174 months), CT simulation without dose–volume histograms (DVHs) (27%, n ⫽ 23; median follow-up 110 months), 3D-CRT (27%, n ⫽ 23; median follow-up 77 months), or IMRT (35%, n ⫽ 30; median follow-up 23 months), depending on the technology available at the time (Table 3). All treatment was administered with 6-MV photons from a linear accelerator. Patients were treated in the supine position using a variety of different immobilization devices, depending on the preferred method of treatment. Radiotherapy was delivered once daily, 5 d/wk, using 1.8 –2 Gy/fraction to an intended dose of 50 –70 Gy (median, 63 Gy). Only 2 patients did not receive the intended prescribed dose. One patient stopped treatment after 40 Gy because of psychosocial problems and another after 52 Gy because of severe mucositis. A twice-daily schedule was used in only 1 patient because of disease progression during RT. Four of the 5 patients with nodal involvement at presentation underwent RT to the neck after neck dissection; the fifth underwent only neck dissection. Five patients underwent elective neck irradiation (ENI) to the upper neck (n ⫽ 4) or the entire neck (n ⫽ 1), at the discretion of the treating physician. The median dose to the neck was 54 Gy (range, 42– 60 Gy). Conventional therapy was delivered using a wedge-pair or three-field technique to cover the PNS and the ipsilateral orbital floor. In 1993, CT simulation was introduced, and the design of the conventional portals were determined from the axial CT slices of the tumor bed and critical structures, with isodose curves drawn on two to three of these axial slices. Dose-volume histograms were not available with this type of treatment planning; however, isod-
Table 4. Doses to critical structures stratified by treatment planning technique Eye (Gy) RT technique Conventional Median Range CT based, no DVH Median Range 3D-CRT Median Range IMRT Median Range
Chiasm (Gy)
Optic nerve (Gy)
Dmax
Dm
D05
Dmax
Dm
D05
Dmax
Dm
D05
60 32–65
–– ––
–– ––
53 48–58
–– ––
–– ––
–– ––
–– ––
–– ––
56 18–63
–– ––
–– ––
52 27–61
–– ––
–– ––
56 36–63
–– ––
–– ––
48 8–74
8 3–43
29 5–68
51 34–59
52 19–49
46 32–55
52 36–79
32 21–70
51 34–78
47 14–72
16 3–41
3 7–67
40 5–51
48 6–53
37 6–81
51 11–100
53 4–105
53 4–105
Abbreviations: Dmax ⫽ maximal dose; Dm ⫽ median dose; D05 ⫽ maximal dose to 5% of structure; DVH ⫽ dose–volume histogram; other abbreviations as in Table 3.
694
I. J. Radiation Oncology
●
Biology
●
Physics
ose curves were generated, and prescription doses were generally prescribed to the 90% isodose line. Forward-planned 3D-CRT was introduced in 1995 and used multibeam techniques with the assistance of beam’s eye view reconstruction, 3D dose distributions, and DVH analysis. Three to nine different beams were used, and the dose was generally prescribed to the 90% isodose line. Our present approach uses IMRT, which was first introduced in 2000, with a simultaneous integrated boost termed “dose painting”. With this technique, different dose levels were prescribed to varying target volumes in a once-daily regimen. These target volumes and their respective dose specifications are described below.
Treatment setup for IMRT All patients treated with IMRT were immobilized in the supine position with custom Aquaplast masks (Aquaplast, Wycoff, NJ). Currently, we use thermoplastic masks that also immobilize the shoulders. Target localization was accomplished using CT simulation (AcQSim, Philips Medical Systems, Andover, MA). CT images indexed every 3 mm were obtained, extending from the vertex of the skull to 5 cm inferior to the clavicular heads. Treatment planning was performed using the inverse planning algorithm of Spirou and Chui, (12), using the proprietary MSKCC treatment planning system (13).
Volume 67, Number 3, 2007
Dosimetric analysis of treatment plans Dose-volume histograms of the treatment volumes and critical normal structures were analyzed. For PTVs, the volume, minimal dose, maximal point dose (Dmax), mean dose, volume covered by 95% of the prescription dose, and dose to 5% of the volume (D05) were evaluated. For the critical organs with functional subunits organized in series, the Dmax was examined, such as the brainstem (Dmax ⱕ50 Gy), spinal cord (Dmax ⱕ45 Gy), retina (Dmax ⱕ45 Gy), optic nerve (Dmax ⱕ54 Gy), and chiasm (Dmax ⱕ54 Gy). For critical organs with functional subunits organized in parallel, such as the parotid glands (i.e., entire gland, including deep and superficial lobes) and cochlea, the mean dose was examined. Table 4 lists the radiation dose to the critical optic structures in the patients in this series. The Dmax of the eye and optic chiasm were evaluated in all patients, and the D05 and mean dose could only be evaluated in patients treated with 3D-CRT or IMRT because they had calculated DVHs. Values for the dose to the lens were only available for 3D-CRT and IMRT. For 3D-CRT, these were the Dmax (median, 9 Gy, range, 4 –59), mean dose (median,
(a)
Local Relapse
Delineation of target volumes for IMRT After obtaining planning CT images in the treatment position (supine, neck hyperextended with a face mask), all target volumes were outlined slice by slice. The surgical bed defined the gross tumor volume. Whenever possible, a recent MRI and/or PET scan, in addition to the CT scan with contrast to the head and neck, was used to delineate the gross tumor volume with the assistance of a neuroradiologist and surgeon. The clinical target volume was defined as the gross tumor volume plus a margin for potential microscopic spread. A second clinical target volume was used in patients receiving nodal RT and encompassed the clinical lymph node regions. Expansion of the clinical target volumes, with a margin of 0.5–1 cm, created the planning target volumes (PTVs) for the dose-painting group. The surrounding organs at risk, including the brainstem, spinal cord, optic nerves, chiasm, parotid glands, mandible, and cochlea, were also outlined with 5-mm margins.
23 4
3
3
1
Distant Relapse
Lymph Node Relapse
(b)
Local Relapse
Dose specifications for IMRT
18
For dose painting, the prescription was specified to the normalized isodose line encompassing the PTV. The median prescribed dose was 60 Gy to PTV1 (range, 60 – 66) and 54 Gy to PTV2. The median dose per fraction was 2 Gy to PTV1 and 1.8 Gy to PTV2. Patients were treated once daily, 5 d/wk.
5
4 4
0
IMRT treatment delivery Intensity-modulated radiotherapy was delivered with dynamic multileaf collimation on a Varian accelerator (Varian Medical Systems, Palo Alto, CA). Dynamic leaf sequencing was accomplished using the algorithm of Spirou and Chui (14). Intensitymodulated radiotherapy beam arrangements consisted of equispaced coplanar beams.
1 0
Lymph Node Relapse
1
3 Distant Relapse
Fig. 1. (a) Site of first relapse and (b) all sites of relapse at last follow-up.
Treatment of nasal cavity and PNS cancer with postoperative RT
4.4 Gy, range, 2.5–34), and D05 (median, 5.7 Gy, range, 3.4 –50). For IMRT, they were the Dmax (median, 7.8 Gy, range, 2.4 – 44), mean dose (median, 5.4 Gy, 1.5–36), and D05 (median, 6.5 Gy, range, 2– 43).
●
B. S. HOPPE et al.
695
as described in the RT chart were graded according to the Radiation Therapy Oncology Group radiation morbidity scoring criteria (15).
Statistical analysis Follow-up The analysis of treatment outcomes was done using follow-up data available as of May 1, 2006. Patients were evaluated once a week during RT. Subsequently, they were seen 1–2 months after RT, every 6 months for the next 3 years, and then yearly. At each follow-up visit, a physical examination, including palpation of the neck, was performed. Routine imaging was done with MRI, CT, or PET every 3– 6 months, depending on physician discretion and the physical examination findings. Acute and late normal tissue effects
Descriptive statistics (mean, median, and proportions) were calculated to characterize the patient, disease, and treatment features, as well as the treatment toxicities. The 5-year local progression-free survival (LPFS), regional progression-free survival (RPFS), distant metastasis-free survival (DMFS), disease-free survival, and overall survival (OS) probabilities were estimated using the Kaplan-Meier product-limit method. Freedom from local progression was defined as the absence of primary tumor on physical examination and on any radiographic examination. Durations were calculated from the date of the last RT session. Univariate analysis
Fig. 2. Kaplan-Meier survival curves for (a) local-, regional-, and distant relapse-free survival and (b) disease-free and overall survival.
696
I. J. Radiation Oncology
●
Biology
●
Physics
was performed using WinSTAT Microsoft Excel software and the log–rank test. A Cox proportional hazard model was used for multivariate analysis (http://www.sph.emory.edu/⬃cdckms/CoxPH/ prophaz2.html).
RESULTS Treatment outcome The median follow-up for living patients was 60 months (range, 0 –225). The site of first failure and all sites of failure in patients are shown in Fig. 1. The LPFS, RPFS, DMFS, OS, and disease-free survival are shown in Fig. 2. Local failure Of the 85 patients, 31 had local failure, with a median time to local recurrence of 8 months (range, 0 –128). The
Volume 67, Number 3, 2007
5-year estimate of LPFS was 62%. In 28 cases, local recurrence was identified as a component of the failure, and in 3, it was a subsequent site of failure. Five local recurrences occurred outside the radiation field, with the others located within the previous radiation field. Local recurrences were found as a result of evaluation of a patient’s new symptom (n ⫽ 7), physical examination (n ⫽ 7), or routine CT (n ⫽ 10) or MRI (n ⫽ 7), in the absence of any physical findings. Regional failure Regional lymph node failure occurred in 10 patients, with a median time to relapse of 8 months (range, 0 –141). The 5-year RPFS rate was 87%. Lymph node failure developed in the Level 1 and/or 2 cervical nodes in 5, Level 1– 4 cervical nodes in 3, Level 1 and 5 cervical nodes in 1, and
Fig. 3. Kaplan-Meier survival curves for (a) local-, regional-, and distant relapse-free survival and (b) disease-free and overall survival in patients with squamous cell carcinoma histologic features.
Treatment of nasal cavity and PNS cancer with postoperative RT
●
B. S. HOPPE et al.
697
Table 5. Relapse and survival by histologic type and T stage Variable Histologic type All types SCC AC SA EN AD UN NE Other T stage T2 T3 T4 Recurrence
Alive
LR
RR
DR
Any
59 (69) 23 (55) 8 (73) 9 (100) 6 (86) 6 (100) 2 (50) 2 (100) 3 (75)
31 (36) 21 (50) 2 (18) 2 (22) 0 2 (33) 3 (75) 1 (50) 0
10 (12) 6 (14) 2 (18) 0 0 0 1 (25) 1 (50) 0
12 (14) 6 (19) 4 (36) 0 1 (14) 0 1 (25) 0 0
37 (44) 23 (55) 5 (45) 2 (22) 1 (14) 2 (33) 3 (75) 1 (50) 0
7 (88) 11 (55) 22 (61) 4 (80)
1 (12) 8 (40) 19 (53) 1 (20)
0 4 (20) 6 (17) 0
1 (12) 3 (15) 6 (17) 1 (20)
3 (38) 9 (45) 21 (58) 1 (20)
Abbreviations: LR ⫽ local relapse; RR ⫽ regional relapse; DR ⫽ distant relapse; Any ⫽ relapse anywhere; SCC ⫽ squamous cell carcinoma; AC ⫽ adenoid cystic; SA ⫽ sarcoma; EN ⫽ esthesioneuroblastoma; AD ⫽ adenocarcinoma; UN ⫽ sinonasal undifferentiated carcinoma; NE ⫽ neuroendocrine. Data presented as number of patients, with percentages in parentheses.
retropharyngeal lymph nodes in 1 patient. Two patients, who were diagnosed at presentation with clinically positive necks, developed regional recurrence despite receiving RT to the neck. None of the 5 patients who received ENI developed neck relapses. Regional relapse was the first site of recurrence in 8 patients and a subsequent site in 2 patients. All patients with neck relapse eventually developed local or distant relapse. Regional failure was found because of patients reporting a new neck mass (n ⫽ 4), routine physical examination (n ⫽ 3), or surveillance CT (n ⫽ 1), MRI (n ⫽ 1), or PET (n ⫽ 1). Distant metastasis Distant failure developed in 12 patients, with a median time to relapse of 21 months (range, 1–167). The 5-year DMFS rate was 82%. The location of the first metastasis was the lung in 7, liver in 2, brain in 1, bone in 1, and liver, lung, and spine in 1. Distant metastasis was the first site of failure in 5 patients and a subsequent site in 7 patients. Distant metastases were found on evaluation of a new symptom reported by the patient (n ⫽ 3) or routine chest
X-ray (n ⫽ 4), PET (n ⫽ 3), CT (n ⫽ 1), or MRI (n ⫽ 1), in the absence of other findings. Overall survival In this series, 26 patients died, and the 5-year estimate of OS was 67%. Twenty-four patients died of disease and two of neurologic deterioration without evidence of disease. Of the 5 patients who presented with lymph node involvement, 3 died of their disease, 1 was lost to follow-up after receiving emergent RT for local recurrence, and 1 was alive without evidence of disease 14 months after completing treatment. Histologic type, T stage, and lymph node involvement Because of the heterogeneity of the histologic types, a subgroup analysis was performed on the large subpopulation of patients with SCC and revealed a 5-year LPF, RPFS, DMF, disease-free survival, and OS rate of 49%, 86%, 72%, 47%, and 53%, respectively (Fig. 3). Table 5 lists the pattern of failure by histologic type and T stage. Patients with adenoid cystic carcinoma developed a relapse predom-
Table 6. Significant factors on univariate analysis All patients (n ⫽ 85)
SCC patients only (n ⫽ 42)
Predictor
LPFS
RPFS
DMFS
OS
LPFS
RPFS
DMFS
OS
Pathologic finding (SCC vs. other) Age (⬍60 vs. ⱖ60 y) Race (white vs. other) LNs (involved vs. not) Orbit (involved vs. not) Cribriform plate (involved vs. not)
0.01 0.1 0.37 0.09 0.04 0.01
0.48 0.23 0.55 0.01 0.02 0.86
0.86 0.22 0.59 ⬍0.005 0.01 0.13
0.01 0.59 0.02 0.04 0.01 0.05
NA 0.02 0.03 0.27 0.11 ⬍0.005
NA 0.03 0.13 0.03 0.02 0.25
NA 0.23 0.02 ⬍0.005 ⬍0.005 ⬍0.005
NA 0.06 ⬍0.005 0.25 0.27 0.01
Abbreviations: SCC ⫽ squamous cell carcinoma; LPFS ⫽ local progression-free survival; RPFS ⫽ regional progression-free survival; DMFS ⫽ distant metastasis-free survival; OS ⫽ overall survival; LNs ⫽ lymph nodes.
698
I. J. Radiation Oncology
●
Biology
●
Physics
Volume 67, Number 3, 2007
inantly in a distant site (n ⫽ 4; 36%). Patients with sarcomas did not develop relapse either regionally or distantly. Patients with sinonasal undifferentiated carcinoma (SNUC) predominantly developed local recurrence (n ⫽ 3; 75%). No patient with T2 disease developed a regional relapse. Patients with T3 and T4 disease developed local and regional relapses. Patients with lymph node involvement at presentation developed local (n ⫽ 3, 60%), regional (n ⫽ 2, 40%), and distant (n ⫽ 3, 60%) relapses. Prognostic factors The following factors were evaluated on univariate analysis for predictive value of LPFS, RPFS, DMFS, and OS in all patients and separately in patients with SCC histology: gender (male vs. female), age (⬍60 vs. ⱖ60 years), race (white vs. other), clinical T stage (T2-T3 vs. T4 and T2 vs. T3-T4), lymph node involvement, pathologic findings (SCC vs. other), PNS site (maxillary vs. other), margin status, presence of perineural invasion, orbital exenteration, RT technique (3D-CRT/IMRT vs. conventional), radiation dose (⬍63 vs. ⱖ63 Gy), orbital involvement, base of skull involvement, cribriform plate involvement, infratemporal fossa involvement, muscle involvement, nasopharynx involvement, skin involvement, and bone involvement. The significant factors and their p values are listed in Table 6. Significant variables on univariate analysis were used in a Cox proportional hazards regression multivariate analysis. In the evaluation of all patients, SCC (hazard ratio [HR] 2.66) and cribriform plate involvement (HR 2.99) were significant factors predicting for an increased likelihood of local recurrence, and SCC (HR 3.04) and nonwhite race (HR 3.13) were significant factors predicting for worse OS. In the evaluation of only patients with SCC histologic features, age ⬍60 years (HR 3.70), nonwhite race (HR 3.70), and cribriform plate involvement (HR 4.37) were significant factors predicting for local relapse, and nonwhite race (HR 3.03) and cribriform plate involvement (HR 3.89) were significant for worse OS. Acute and late toxicity The acute and late toxicity from RT by site and grade is detailed in Table 7. Eighty-three patients completed their Table 7. Radiation Therapy Oncology Group acute and late toxicity Grade Toxicity Acute (n ⫽ 85) Eye (n ⫽ 72) Mucous membrane Salivary gland Skin Late (n ⫽ 79) Eye (n ⫽ 66) Mucous membrane Salivary gland Skin
0
1
2
3
4
48 8 26 9
21 27 39 37
3 32 20 33
0 18 0 6
0 0 0 0
62 67 59 70
2 11 17 8
1 1 3 1
0 0 0 0
1 0 0 0
Fig. 4. Isodose curves for patient treated with conventionally planned radiotherapy who developed Grade 4 late visual complication to ipsilateral eye.
prescribed treatment course. Grade 1-2 mucositis was reported in 59 patients (69%) and Grade 3-4 in 18 (21%). Grade 1-2 xerostomia was seen in 59 patients (69%), but no patient developed Grade 3-4 xerostomia as an early complication of RT. Three patients were reported to have Grade 2 conjunctivitis during RT, requiring steroid eye drops, but these cases all resolved. Late toxicity could be evaluated in 79 patients with a minimum follow-up of 3 months. Twelve patients (15%) developed Grade 1-2 mucous membrane toxicity as a late effect of RT, but no patient developed Grade 3-4. Also, 20 patients (25%) developed Grade 1-2 salivary gland toxicity as a late effect, but no patient had Grade 3-4 toxicity. Three patients developed Grade 1-2 late ophthalmologic toxicity. An additional patient developed radiation-induced corneal ulceration, retinopathy, and cataract of his right eye 4 years after RT completion. This ultimately resulted in right unilateral blindness ⬎5 years after RT completion in that eye. This patient had been treated with conventional RT and received ⱕ63 Gy to the retina (Fig. 4). Salvage therapy Patients with relapse underwent a variety of different salvage therapies, depending on the relapse site and the patient’s performance status. Six patients with local relapse underwent surgical resection as a component of their salvage therapy. Only 2 of these patients died (8 and 69 months from the first failure); the other 4 patients were alive (range, 39 – 68 months from the first failure). Five patients underwent repeat RT alone or in combination with chemotherapy. Two of these patients were still alive (10 and 65 months from the first failure) and the other 3 patients died (range, 18 –35 months from the first failure). Chemotherapy alone was given to 7 patients, 6 of whom died (range, 8 –23
Treatment of nasal cavity and PNS cancer with postoperative RT
months from the first failure) and 1 of whom was still alive 17 months from the first failure. Patients receiving only supportive care died within 4 months of the first failure. Patients with only regional relapse as a site of first failure were treated with neck dissection followed by RT with or without chemotherapy. One patient survived, 32 months; the other 2 died 9 –12 months from the first failure. Four patients developed both local relapse and cervical adenopathy. One was treated with surgery and lived 42 months from the first failure, and other three received chemotherapy and eventually died 10 –20 months from the first failure. Patients who presented with metastases were generally treated with palliative RT or chemotherapy and eventually died (range, 4 –21 months from the first failure). One patient with adenoid cystic carcinoma developed a lung metastasis and underwent wedge resection and was still alive 147 months from the first failure. DISCUSSION No randomized controlled studies have evaluated treatment options for patients with PNS/nasal cavity cancer. Treatment recommendations have been based on small retrospective reviews by institutions with heterogeneous populations of patients treated with several different modalities and older RT techniques. In the present study, we reviewed the MSKCC experience with PNS cancer treated with surgery followed by RT. Compared with previous studies (Table 8) (2, 5, 7, 8, 16, 17), our study included the largest number of patients treated with postoperative RT using state-of-the-art RT techniques. The OS and local control were comparable with those of other studies, but the number of patients with Grade 3 or worse ophthalmic toxicity was lower than that seen in the older series using conventional RT techniques. Modern RT techniques using more precise imaging (MRI and PET) for target delineation, achieve a more rapid dose falloff around critical areas, including the retina, optic chiasm, and lens (18 –23) (Figs. 5 and 6) and, theoretically, should lead to fewer long-term risks and complications. Thus far, the outcomes from series using 3D-CRT and IMRT have had less ophthalmologic toxicity compared with
●
B. S. HOPPE et al.
699
older series. Roa et al. (10) studied 39 patients (40% unresectable) treated with 3D-CRT with 54 months of follow-up and found only 1 case of limited optic neuropathy and no case of blindness, with comparable local control and OS to historical studies. Pommier et al. (24) evaluated 40 patients (25% unresectable) treated with 3D treatment planning with a median follow-up of 19 months and reported only 1 patient developing ipsilateral blindness due to vascular glaucoma with a 2-year local control and OS rate of 66% and 73%, respectively. Padovani et al. (9) reported on 25 patients (12% unresectable) treated with 3D-CRT with 25 months of follow-up and found a local control and OS rate of 54% and 24%, respectively, with 2 patients developing uveitis and retinopathy. Most recently, Duthoy et al. (8) evaluated 39 patients with PNS cancer treated with postoperative IMRT with 31 months of follow-up and found only 2 patients with decreased vision after RT, no patient with blindness, and a 4-year local control and OS rate of 68% and 59%, respectively. In the present study, no severe early or late toxicities that affected vision were observed in any of the patients who were treated with modern RT techniques. Only 1 patient in our series developed blindness and that patient had been treated with conventionally planned RT without CT planning. The median Dmax to the eye in the patients treated in this series decreased with the introduction of each new RT technique, but remained constant for the optic chiasm. An example of an IMRT plan with dose painting is illustrated in Fig. 5, and the DVH from the plan is shown in Fig. 6. Local control and survival were not compromised by the newer RT techniques, with a 5-year rate of 62% and 67%, respectively. Further dose reduction to normal tissue may be possible with the use of proton beam therapy. Studies have been done comparing 3D-CRT, IMRT, and proton beam plans and have shown even better coverage of the PTV with lower doses to critical structures (21–23). The outcomes of clinical trials using proton therapy are eagerly awaited, and a headto-head clinical trial may be warranted in the future. Squamous cell carcinoma histologic type was an important prognostic factor predictive for local relapse and worse OS. These patients, as well as those with SNUC, develop recurrence more frequently locally, regionally, and distantly compared with patients with other histologic features (25).
Table 8. Outcomes in other paranasal sinus series using postoperative radiotherapy Investigator
n
Follow-up (mo)
Surgery ⫹ RT (%)
T3-T4 (%)
Maxillary (%)
⫹LNs (%)
SCC (%)
LC (%)
RC (%)
OS (%)
Grade 3–4 eye toxicity
Jiang et al. (16) Duthoy et al. (8) Myers et al. (17) Jansen et al. (5) Le et al. (7) Blanco et al. (2) Present study
73 39 141 73 97 106 85
83 31 69 66 78 60 60
100 100 52 68 63 65 100
74 59 88 86 92 86 88
100 15 70 NA 100 76 53
8 3 4 11 11 7 7
49 21 52 55 60 82 49
78 68 28 63 43 58 62
15 3 6 11 10 29 12
48 59 65 46 34 27 67
16 2 NA 24 4 3 1
Abbreviations: LC ⫽ local control; RC ⫽ regional control; other abbreviations as in Table 6.
700
I. J. Radiation Oncology
●
Biology
●
Physics
Volume 67, Number 3, 2007
neoadjuvant chemotherapy followed by surgery and postoperative RT. Our current approach is to treat these patients aggressively with surgery followed by CTRT. Patients with SCC also need to receive more aggressive therapy, especially when other poor prognostic factors are present. Neoadjuvant CTRT followed by surgery will be investigated in an upcoming North American Skull Base Society protocol. Another consideration could be to investigate targeted systemic therapies that have been shown to be beneficial in other head-and-neck epithelial cancers (29, 30). Lymph node involvement at presentation has been a poor prognostic factor in other studies (2, 5, 17, 25). In this study, lymph node involvement was a significant factor in predicting for regional relapse, distant metastasis, and OS on univariate analysis. These patients were treated with neck dissection followed by RT and had poor outcomes. Neoadjuvant chemotherapy or chemotherapy in combination with RT may improve the outcome for these patients. Patients with neck relapse also have a high risk of local and metastatic recurrences. Therefore, patients at greater risk of regional recurrence may benefit from ENI (16, 31, 32). In our series, none of the 5 patients receiving ENI developed regional relapse. Furthermore, patients who presented with lymph node involvement or neck relapse did so predominantly in the Level 1 or 2 regions. Other studies have found cervical neck relapses to occur in the same regions (31, 32). Currently, patients with T3 or T4 SCC and SNUC treated at MSKCC receive ENI. Cribriform plate involvement was also found on univariate and multivariate analyses to predict for local relapse. When PNS cancers involve the cribriform plate, adequate coverage is difficult without exceeding the RT tolerance for the nearby critical structures. Tsien et al. found a significant underdosage to the PTV in IMRT plans for PNS cancers when the left and right optic pathways were spared (33). However, PTV coverage was much better when only the contralateral optic pathway was spared. Therefore, when treating PNS cancer involving the cribriform plate, a critical evaluation of the treatment plan is necessary and one may need to consider compromising unilateral vision in favor of adequate PTV coverage. Salvage therapy was effective in selected patients. When surgery is a part of the salvage regimen, patients may survive for several years after their first relapse. In combination with RT or CTRT, salvage surgery may provide even longer survival. Although patients treated with chemotherapy alone for salvage have poor survival, they did better than patients receiving only supportive care. CONCLUSION Fig. 5. Isodose curves for intensity-modulated radiotherapy plan.
More aggressive approaches in the treatment of patients with SNUC, including induction chemotherapy followed by CTRT with or without full surgical resection have shown some benefit, with 2-year survival rates of 64 –75% (26 – 28). In this study, the only patient with SNUC who was alive at last follow-up without relapse was treated with
Surgical resection followed by postoperative RT is an effective treatment for cancer involving the PNSs, as well as the nasal cavity. Patients with cervical lymph node involvement, SCC and SNUC, and involvement of the cribriform plate have worse outcomes, thereby necessitating a more aggressive treatment approach. Modern RT techniques, such as IMRT, that can limit the radiation dose to nearby normal tissues do not appear to result in compromised tumor control.
Treatment of nasal cavity and PNS cancer with postoperative RT
●
B. S. HOPPE et al.
701
Fig. 6. Dose–volume histogram for intensity-modulated radiotherapy plan. PTV-DVH ⫽ PTV minus build-up region; OPNRV ⫽ optic nerve.
REFERENCES 1. Silverberg E, Grant RN. Cancer statistics, 1970. CA Cancer J Clin 1970;20:11–23. 2. Blanco AI, Chao KS, Ozyigit G, et al. Carcinoma of paranasal sinuses: Long-term outcomes with radiotherapy. Int J Radiat Oncol Biol Phys 2004;59:51–58. 3. Carrillo JF, Guemes A, Ramirez-Ortega MC, et al. Prognostic factors in maxillary sinus and nasal cavity carcinoma. Eur J Surg Oncol 2005;31:1206 –1212. 4. Harrison LB, Raben A, Pfister DG, et al. A prospective phase II trial of concomitant chemotherapy and radiotherapy with delayed accelerated fractionation in unresectable tumors of the head and neck. Head Neck 1998;20:497–503. 5. Jansen EP, Keus RB, Hilgers FJ, et al. Does the combination of radiotherapy and debulking surgery favor survival in paranasal sinus carcinoma? Int J Radiat Oncol Biol Phys 2000;48: 27–35. 6. Katz TS, Mendenhall WM, Morris CG, et al. Malignant tumors of the nasal cavity and paranasal sinuses. Head Neck 2002;24:821– 829. 7. Le QT, Fu KK, Kaplan M, et al. Treatment of maxillary sinus carcinoma: A comparison of the 1997 and 1977 American Joint Committee on cancer staging systems. Cancer 1999;86: 1700 –1711. 8. Duthoy W, Boterberg T, Claus F, et al. Postoperative intensity-modulated radiotherapy in sinonasal carcinoma: Clinical results in 39 patients. Cancer 2005;104:71– 82. 9. Padovani L, Pommier P, Clippe SS, et al. Three-dimensional conformal radiotherapy for paranasal sinus carcinoma: Clinical results for 25 patients. Int J Radiat Oncol Biol Phys 2003;56:169 –176. 10. Roa WH, Hazuka MB, Sandler HM, et al. Results of primary and adjuvant CT-based 3-dimensional radiotherapy for malignant tumors of the paranasal sinuses. Int J Radiat Oncol Biol Phys 1994;28:857– 865. 11. Greene FL, for the American Joint Committee on Cancer and American Cancer Society. AJCC cancer staging handbook: From the AJCC cancer staging manual. 6th ed. New York: Springer; 2002. 12. Spirou SV, Chui CS. A gradient inverse planning algo-
13. 14. 15.
16. 17. 18.
19.
20. 21.
22. 23.
rithm with dose-volume constraints. Med Phys 1998;25: 321–333. Mohan R, Barest G, Brewster LJ, et al. A comprehensive three-dimensional radiation treatment planning system. Int J Radiat Oncol Biol Phys 1988;15:481– 495. Spirou SV, Chui CS. Generation of arbitrary intensity profiles by combining the scanning beam with dynamic multileaf collimation. Med Phys 1996;23:1– 8. Cox JD, Stetz J, Pajak TF. Toxicity criteria of the Radiation Therapy Oncology Group (RTOG) and the European Organization for Research and Treatment of Cancer (EORTC). Int J Radiat Oncol Biol Phys 1995;31:1341–1346. Jiang GL, Ang KK, Peters LJ, et al. Maxillary sinus carcinomas: Natural history and results of postoperative radiotherapy. Radiother Oncol 1991;21:193–200. Myers LL, Nussenbaum B, Bradford CR, et al. Paranasal sinus malignancies: An 18-year single institution experience. Laryngoscope 2002;112:1964 –1969. Brizel DM, Light K, Zhou SM, et al. Conformal radiation therapy treatment planning reduces the dose to the optic structures for patients with tumors of the paranasal sinuses. Radiother Oncol 1999;51:215–218. Huang D, Xia P, Akazawa P, et al. Comparison of treatment plans using intensity-modulated radiotherapy and threedimensional conformal radiotherapy for paranasal sinus carcinoma. Int J Radiat Oncol Biol Phys 2003;56:158 – 168. Lohr F, Pirzkall A, Debus J, et al. Conformal three-dimensional photon radiotherapy for paranasal sinus tumors. Radiother Oncol 2000;56:227–231. Mock U, Georg D, Bogner J, et al. Treatment planning comparison of conventional, 3D conformal, and intensitymodulated photon (IMRT) and proton therapy for paranasal sinus carcinoma. Int J Radiat Oncol Biol Phys 2004;58: 147–154. Lomax AJ, Goitein M, Adams J. Intensity modulation in radiotherapy: Photons versus protons in the paranasal sinus. Radiother Oncol 2003;66:11–18. Miralbell R, Crowell C, Suit HD. Potential improvement of three dimension treatment planning and proton therapy in the
702
24.
25. 26. 27. 28.
I. J. Radiation Oncology
●
Biology
●
Physics
outcome of maxillary sinus cancer. Int J Radiat Oncol Biol Phys 1992;22:305–310. Pommier P, Ginestet C, Sunyach M, et al. Conformal radiotherapy for paranasal sinus and nasal cavity tumors: Three-dimensional treatment planning and preliminary results in 40 patients. Int J Radiat Oncol Biol Phys 2000;48: 485– 493. Bhattacharyya N. Cancer of the nasal cavity: Survival and factors influencing prognosis. Arch Otolaryngol Head Neck Surg 2002;128:1079 –1083. Rischin D, Porceddu S, Peters L, et al. Promising results with chemoradiation in patients with sinonasal undifferentiated carcinoma. Head Neck 2004;26:435– 441. Kim BS, Vongtama R, Juillard G. Sinonasal undifferentiated carcinoma: Case series and literature review. Am J Otolaryngol 2004;25:162–166. Rosenthal DI, Barker JL Jr, El-Naggar AK, et al. Sinonasal malignancies with neuroendocrine differentiation: Patterns of failure according to histologic phenotype. Cancer 2004;101: 2567–2573.
Volume 67, Number 3, 2007
29. Bonner JA, Harari PM, Giralt J, et al. Radiotherapy plus cetuximab for squamous-cell carcinoma of the head and neck. N Engl J Med 2006;354:567–578. 30. Pfister DG, Su YB, Kraus DH, et al. Concurrent cetuximab, cisplatin, and concomitant boost radiotherapy for locoregionally advanced, squamous cell head and neck cancer: A pilot phase II study of a new combined-modality paradigm. J Clin Oncol 2006;24:1072–1078. 31. Le QT, Fu KK, Kaplan MJ, et al. Lymph node metastasis in maxillary sinus carcinoma. Int J Radiat Oncol Biol Phys 2000;46:541–549. 32. Paulino AC, Fisher SG, Marks JE. Is prophylactic neck irradiation indicated in patients with squamous cell carcinoma of the maxillary sinus? Int J Radiat Oncol Biol Phys 1997;39: 283–289. 33. Tsien C, Eisbruch A, McShan D, et al. Intensity-modulated radiation therapy (IMRT) for locally advanced paranasal sinus tumors: Incorporating clinical decisions in the optimization process. Int J Radiat Oncol Biol Phys 2003;55:776 –784.
doi:10.1016/j.ijrobp.2006.09.023
CLINICAL INVESTIGATION
Head and Neck
TREATMENT OF NASAL CAVITY AND PARANASAL SINUS CANCER WITH MODERN RADIOTHERAPY TECHNIQUES IN THE POSTOPERATIVE SETTING—THE MSKCC EXPERIENCE BRADFORD S. HOPPE, M.D.,* LAUREN D. STEGMAN, M.D., PH.D.,* MICHAEL J. ZELEFSKY, M.D.,* KENNETH E. ROSENZWEIG, M.D.,* SUZANNE L. WOLDEN, M.D.,* SNEHAL G. PATEL, M.D.,† JATIN P. SHAH, M.D.,† DENNIS H. KRAUS, M.D.,† AND NANCY Y. LEE, M.D.* Departments of *Radiation Oncology and †Head and Neck Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY Purpose: To perform a retrospective analysis of patients with paranasal sinus (PNS) cancer treated with postoperative radiotherapy (RT) at Memorial Sloan-Kettering Cancer Center. Methods and Materials: Between January 1987 and July 2005, 85 patients with PNS and nasal cavity cancer underwent postoperative RT. Most patients had squamous cell carcinoma (49%; n ⴝ 42), T4 tumors (52%; n ⴝ 36), and the maxillary sinus (53%; n ⴝ 45) as the primary disease site. The median radiation dose was 63 Gy. Of the 85 patients, 76 underwent CT simulation and 53 were treated with either three-dimensional conformal RT (27%; n ⴝ 23) or intensity-modulated RT (35%; n ⴝ 30). Acute and late toxicities were scored according to the Radiation Therapy Oncology Group radiation morbidity scoring criteria. Results: With a median follow-up for surviving patients of 60 months, the 5-year estimates of local progressionfree, regional progression-free, distant metastasis-free, disease-free, and overall survival rates were 62%, 87%, 82%, 55%, and 67%, respectively. On multivariate analysis, squamous cell histology and cribriform plate involvement predicted for an increased likelihood of local recurrence, and squamous cell histologic features predicted for worse overall survival. None of the patients who underwent CT simulation and were treated with modern techniques developed a Grade 3-4 late complication of the eye. Conclusion: Complete surgical resection followed by adjuvant RT is an effective and safe approach in the treatment of PNS cancer. Emerging tools, such as three-dimensional conformal treatment and, in particular, intensity-modulated RT for PNS tumors, may minimize the occurrence of late complications associated with conventional RT techniques. Local recurrence remains a significant problem. © 2007 Elsevier Inc. Paranasal sinus, Surgery, Intensity-modulated radiotherapy, Radiotherapy, Cancer.
with chemotherapy. The outcomes of these unresectable patients treated with definitive RT with or without chemotherapy have been dismal, largely owing to the limitations in the radiation dose that can be safely delivered without causing injuries to the nearby critical normal tissues, such as the brainstem and optic structures (2–7). Radiotherapy techniques have evolved during the past two decades from conventional RT, to three-dimensional conformal RT (3D-CRT), to intensity-modulated RT (IMRT) with “dose-painting” and proton therapy. These newer techniques have allowed for improved dose distributions, with increased dose to the target volumes and reduced dose to the surrounding normal tissues compared with conventional treatment. One of the sites that can benefit from these advances in RT techniques is cancer involving the PNSs. Historical studies in the treatment of PNS cancers
INTRODUCTION Paranasal sinus (PNS) and nasal cavity cancers are rare, making up only 5% of head-and-neck cancers and ⬍1% of all malignancies (1). Because of the rarity of these tumors, clinical data regarding their management have been largely from small retrospective experiences from single institutions. These studies have shown surgical resection with clear margins offers an excellent chance of cure in earlystage tumors. However, in more advanced tumors, achieving clear margins is often difficult and, therefore, adjuvant radiotherapy (RT) is required. At the Memorial Sloan-Kettering Cancer Center (MSKCC), patients with PNS cancer who are surgical candidates are treated with surgery followed by RT. Tumors not amenable to complete surgical resection because of invasion of the dura, brain, orbit, or nasopharynx, are managed with RT, often in combination Reprint requests to: Nancy Y. Lee, M.D., Department of Radiation Oncology, Memorial Sloan-Kettering Cancer Center, 1275 York Ave., New York, NY 10021. Tel: (212) 639-3341; Fax: (212) 639-2417; E-mail: [email protected]
Conflict of interest: none. Received July 12, 2006, and in revised form Sept 11, 2006. Accepted for publication Sept 13, 2006. 691
692
I. J. Radiation Oncology
●
Biology
●
Physics
using conventional RT techniques have reported up to 35% of patients developing severe visual impairment as a consequence of RT (6). Recent studies of modern RT techniques have shown fewer patients developing visual complications; however, they lacked long-term follow-up (8 – 10). Visual impairment from RT and local failure often occur several years after treatment, and only with sufficient follow-up will these studies effectively assess the outcomes and severe toxicities. This study reports the MSKCC’s 18-year experience with PNS and nasal cavity tumors, treated with surgical resection and RT, predominantly using modern RT techniques, and compares the results with those of other PNS studies. Prognostic factors that may predict a worse outcome were evaluated to identify patients who may need more aggressive upfront therapy. Finally, we describe the current technique used at the MSKCC of IMRT dose painting for patients with PNS cancers. METHODS AND MATERIALS Between January 1987 and July 2005, 232 patients with a diagnosis of PNS cancer evaluated in the Department of Radiation Oncology were identified from the institution’s database. Patients were excluded from the analysis if they had received definitive RT or chemoradiotherapy (CTRT) (n ⫽ 51), had received their RT at another facility (n ⫽ 50), had the histologic features of melanoma, rhabdomyosarcoma, plasmacytoma, Merkel cell carcinoma, or noninvasive disease (in situ) (n ⫽ 40), had a history of retinoblastoma with previous RT (n ⫽ 4), or had not received RT to the primary disease site (n ⫽ 2). The present analysis included 85 PNS cancer patients, who had undergone surgical resection followed by postoperative RT.
Table 1. Patient characteristics Parameter Gender Male Female Ethnicity White Black Asian Other Site Maxillary sinus Nasal cavity Ethmoid sinus Frontal sinus Sphenoid sinus Histologic type Squamous cell carcinoma Adenoid cystic Sarcoma Esthesioneuroblastoma Adenocarcinoma Undifferentiated Neuroendocrine Other
n (%) 53 (62) 32 (38) 68 (80) 7 (8) 7 (8) 3 (4) 45 (53) 24 (28) 14 (16) 1 (1) 1 (1) 42 (49) 11 (13) 9 (11) 7 (8) 6 (7) 4 (5) 2 (3) 4 (5)
Volume 67, Number 3, 2007
Table 2. Staging (excluding sarcomas) Staging system AJCC (n ⫽ 69) Recurrent T stage T1 T2 T3 T4 N stage N0 N1 N2 Kadish (n ⫽ 7, esthesioneuroblastoma) A B C
n (%) 5 (7) 0 8 (12) 20 (29) 36 (52) 64 (93) 5 (7) 0 2 (29) 4 (57) 1 (14)
Abbreviation: AJCC ⫽ American Joint Committee on Cancer.
Patient characteristics The patient age-range was 10 – 82 years (median, 54), and the Karnofsky performance status was 70 –100 (median, 80). Other patient characteristics are listed in Table 1. Most patients were male (62%) and white (80%). The subsite of origin was determined as the region harboring the predominant bulk of disease and was most commonly in the maxillary sinus (53%). Squamous cell carcinoma (SCC) (49%) was the most frequent histologic type.
Evaluation and staging The pretreatment evaluation included a complete history and physical examination, direct flexible fiberoptic endoscopic examination, complete blood count, liver function tests, chest X-ray, CT, and dental evaluations. Of the 85 patients, 47 underwent additional MRI and 11 underwent 18F-fluorodeoxyglucose positron emission tomography (PET). For this study, patients were retrospectively restaged using the 2002 American Joint Committee on Cancer TNM staging system (11) (Table 2), with the exception of esthesioneuroblastomas, which underwent Kadish staging. Sarcomas were not staged. Patients predominantly presented with T4 disease (52%; n ⫽ 36). Five patients with local relapse after surgical resection alone were included in this analysis, including two adenoid cystic carcinomas, one SCC, one adenocarcinoma, and one myoepithelial carcinoma. At recurrence, these patients underwent radical surgical resection followed by postoperative RT. Clinical cervical lymph node metastasis was seen in 5 patients at presentation, including 4 patients with SCC and 1 with adenocarcinoma, and was located in the neck Level 1 region in 1 patient and the Level 2 region in 4.
Surgery All patients underwent gross surgical resection. The surgical techniques ranged from subtotal maxillectomy to cranial facial resection. Most patients had negative surgical margins (62%; n ⫽ 53). Surgical margins were recorded as negative, unless otherwise stated by the surgeon, pathologist, or treating radiation oncologist. Of the 85 patients, 21 had perineural invasion. Only the 5 patients who presented with clinical nodal involvement underwent neck dissection as part of their primary treatment. Thirteen patients underwent orbital exenteration as part of their primary surgery
Treatment of nasal cavity and PNS cancer with postoperative RT
Value
Surgery Craniofacial resection Orbital exenteration Yes No Neck dissection Yes No Chemotherapy Preoperative Concurrent RT technique Conventional Conventional ⫹ Chemotherapy 3D-CRT IMRT Radiation dose (Gy) Median Range Neck irradiated No Upper Whole
B. S. HOPPE et al.
693
patients had partial responses radiographically; for the third patient, no imaging was available before the start of chemotherapy. Disease was evident in the surgical specimen in all 3 cases. Two of these patients had microscopically positive margins. Two additional patients received concomitant CTRT after resection of their primary tumor (one because of rapid disease progression postoperatively). These patients received cisplatin alone or in combination with gemcitabine and had SCC histologic features.
Table 3. Treatment variables Variable
●
34 (40) 13 (15) 72 (85) 5 (6) 80 (94)
Radiotherapy
3 (4) 2 (2) 9 (11) 23 (27) 23 (27) 30 (35) 63 40–70 75 (88) 4 (5) 5 (6)
Abbreviations: RT ⫽ radiotherapy; 3D-CRT ⫽ three-dimensional conformal RT; IMRT ⫽ intensity-modulated RT. Data presented as number of patients, with percentages in parentheses, unless otherwise noted.
because of involvement of the orbital contents. The details of the surgical interventions are listed in Table 3.
Chemotherapy Systemic chemotherapy was used in the initial treatment of 5 patients (Table 3). In 3 pediatric patients, neoadjuvant chemotherapy was used, including combinations of cisplatin, gemcitabine, doxorubicin, methotrexate, cyclophosphamide, vincristine, and etoposide. These pediatric patients had osteosarcoma, sinonasal undifferentiated carcinoma, and esthesioneuroblastoma. Two
Radiotherapy was performed with conventional planning (11%, n ⫽ 9; median follow-up 174 months), CT simulation without dose–volume histograms (DVHs) (27%, n ⫽ 23; median follow-up 110 months), 3D-CRT (27%, n ⫽ 23; median follow-up 77 months), or IMRT (35%, n ⫽ 30; median follow-up 23 months), depending on the technology available at the time (Table 3). All treatment was administered with 6-MV photons from a linear accelerator. Patients were treated in the supine position using a variety of different immobilization devices, depending on the preferred method of treatment. Radiotherapy was delivered once daily, 5 d/wk, using 1.8 –2 Gy/fraction to an intended dose of 50 –70 Gy (median, 63 Gy). Only 2 patients did not receive the intended prescribed dose. One patient stopped treatment after 40 Gy because of psychosocial problems and another after 52 Gy because of severe mucositis. A twice-daily schedule was used in only 1 patient because of disease progression during RT. Four of the 5 patients with nodal involvement at presentation underwent RT to the neck after neck dissection; the fifth underwent only neck dissection. Five patients underwent elective neck irradiation (ENI) to the upper neck (n ⫽ 4) or the entire neck (n ⫽ 1), at the discretion of the treating physician. The median dose to the neck was 54 Gy (range, 42– 60 Gy). Conventional therapy was delivered using a wedge-pair or three-field technique to cover the PNS and the ipsilateral orbital floor. In 1993, CT simulation was introduced, and the design of the conventional portals were determined from the axial CT slices of the tumor bed and critical structures, with isodose curves drawn on two to three of these axial slices. Dose-volume histograms were not available with this type of treatment planning; however, isod-
Table 4. Doses to critical structures stratified by treatment planning technique Eye (Gy) RT technique Conventional Median Range CT based, no DVH Median Range 3D-CRT Median Range IMRT Median Range
Chiasm (Gy)
Optic nerve (Gy)
Dmax
Dm
D05
Dmax
Dm
D05
Dmax
Dm
D05
60 32–65
–– ––
–– ––
53 48–58
–– ––
–– ––
–– ––
–– ––
–– ––
56 18–63
–– ––
–– ––
52 27–61
–– ––
–– ––
56 36–63
–– ––
–– ––
48 8–74
8 3–43
29 5–68
51 34–59
52 19–49
46 32–55
52 36–79
32 21–70
51 34–78
47 14–72
16 3–41
3 7–67
40 5–51
48 6–53
37 6–81
51 11–100
53 4–105
53 4–105
Abbreviations: Dmax ⫽ maximal dose; Dm ⫽ median dose; D05 ⫽ maximal dose to 5% of structure; DVH ⫽ dose–volume histogram; other abbreviations as in Table 3.
694
I. J. Radiation Oncology
●
Biology
●
Physics
ose curves were generated, and prescription doses were generally prescribed to the 90% isodose line. Forward-planned 3D-CRT was introduced in 1995 and used multibeam techniques with the assistance of beam’s eye view reconstruction, 3D dose distributions, and DVH analysis. Three to nine different beams were used, and the dose was generally prescribed to the 90% isodose line. Our present approach uses IMRT, which was first introduced in 2000, with a simultaneous integrated boost termed “dose painting”. With this technique, different dose levels were prescribed to varying target volumes in a once-daily regimen. These target volumes and their respective dose specifications are described below.
Treatment setup for IMRT All patients treated with IMRT were immobilized in the supine position with custom Aquaplast masks (Aquaplast, Wycoff, NJ). Currently, we use thermoplastic masks that also immobilize the shoulders. Target localization was accomplished using CT simulation (AcQSim, Philips Medical Systems, Andover, MA). CT images indexed every 3 mm were obtained, extending from the vertex of the skull to 5 cm inferior to the clavicular heads. Treatment planning was performed using the inverse planning algorithm of Spirou and Chui, (12), using the proprietary MSKCC treatment planning system (13).
Volume 67, Number 3, 2007
Dosimetric analysis of treatment plans Dose-volume histograms of the treatment volumes and critical normal structures were analyzed. For PTVs, the volume, minimal dose, maximal point dose (Dmax), mean dose, volume covered by 95% of the prescription dose, and dose to 5% of the volume (D05) were evaluated. For the critical organs with functional subunits organized in series, the Dmax was examined, such as the brainstem (Dmax ⱕ50 Gy), spinal cord (Dmax ⱕ45 Gy), retina (Dmax ⱕ45 Gy), optic nerve (Dmax ⱕ54 Gy), and chiasm (Dmax ⱕ54 Gy). For critical organs with functional subunits organized in parallel, such as the parotid glands (i.e., entire gland, including deep and superficial lobes) and cochlea, the mean dose was examined. Table 4 lists the radiation dose to the critical optic structures in the patients in this series. The Dmax of the eye and optic chiasm were evaluated in all patients, and the D05 and mean dose could only be evaluated in patients treated with 3D-CRT or IMRT because they had calculated DVHs. Values for the dose to the lens were only available for 3D-CRT and IMRT. For 3D-CRT, these were the Dmax (median, 9 Gy, range, 4 –59), mean dose (median,
(a)
Local Relapse
Delineation of target volumes for IMRT After obtaining planning CT images in the treatment position (supine, neck hyperextended with a face mask), all target volumes were outlined slice by slice. The surgical bed defined the gross tumor volume. Whenever possible, a recent MRI and/or PET scan, in addition to the CT scan with contrast to the head and neck, was used to delineate the gross tumor volume with the assistance of a neuroradiologist and surgeon. The clinical target volume was defined as the gross tumor volume plus a margin for potential microscopic spread. A second clinical target volume was used in patients receiving nodal RT and encompassed the clinical lymph node regions. Expansion of the clinical target volumes, with a margin of 0.5–1 cm, created the planning target volumes (PTVs) for the dose-painting group. The surrounding organs at risk, including the brainstem, spinal cord, optic nerves, chiasm, parotid glands, mandible, and cochlea, were also outlined with 5-mm margins.
23 4
3
3
1
Distant Relapse
Lymph Node Relapse
(b)
Local Relapse
Dose specifications for IMRT
18
For dose painting, the prescription was specified to the normalized isodose line encompassing the PTV. The median prescribed dose was 60 Gy to PTV1 (range, 60 – 66) and 54 Gy to PTV2. The median dose per fraction was 2 Gy to PTV1 and 1.8 Gy to PTV2. Patients were treated once daily, 5 d/wk.
5
4 4
0
IMRT treatment delivery Intensity-modulated radiotherapy was delivered with dynamic multileaf collimation on a Varian accelerator (Varian Medical Systems, Palo Alto, CA). Dynamic leaf sequencing was accomplished using the algorithm of Spirou and Chui (14). Intensitymodulated radiotherapy beam arrangements consisted of equispaced coplanar beams.
1 0
Lymph Node Relapse
1
3 Distant Relapse
Fig. 1. (a) Site of first relapse and (b) all sites of relapse at last follow-up.
Treatment of nasal cavity and PNS cancer with postoperative RT
4.4 Gy, range, 2.5–34), and D05 (median, 5.7 Gy, range, 3.4 –50). For IMRT, they were the Dmax (median, 7.8 Gy, range, 2.4 – 44), mean dose (median, 5.4 Gy, 1.5–36), and D05 (median, 6.5 Gy, range, 2– 43).
●
B. S. HOPPE et al.
695
as described in the RT chart were graded according to the Radiation Therapy Oncology Group radiation morbidity scoring criteria (15).
Statistical analysis Follow-up The analysis of treatment outcomes was done using follow-up data available as of May 1, 2006. Patients were evaluated once a week during RT. Subsequently, they were seen 1–2 months after RT, every 6 months for the next 3 years, and then yearly. At each follow-up visit, a physical examination, including palpation of the neck, was performed. Routine imaging was done with MRI, CT, or PET every 3– 6 months, depending on physician discretion and the physical examination findings. Acute and late normal tissue effects
Descriptive statistics (mean, median, and proportions) were calculated to characterize the patient, disease, and treatment features, as well as the treatment toxicities. The 5-year local progression-free survival (LPFS), regional progression-free survival (RPFS), distant metastasis-free survival (DMFS), disease-free survival, and overall survival (OS) probabilities were estimated using the Kaplan-Meier product-limit method. Freedom from local progression was defined as the absence of primary tumor on physical examination and on any radiographic examination. Durations were calculated from the date of the last RT session. Univariate analysis
Fig. 2. Kaplan-Meier survival curves for (a) local-, regional-, and distant relapse-free survival and (b) disease-free and overall survival.
696
I. J. Radiation Oncology
●
Biology
●
Physics
was performed using WinSTAT Microsoft Excel software and the log–rank test. A Cox proportional hazard model was used for multivariate analysis (http://www.sph.emory.edu/⬃cdckms/CoxPH/ prophaz2.html).
RESULTS Treatment outcome The median follow-up for living patients was 60 months (range, 0 –225). The site of first failure and all sites of failure in patients are shown in Fig. 1. The LPFS, RPFS, DMFS, OS, and disease-free survival are shown in Fig. 2. Local failure Of the 85 patients, 31 had local failure, with a median time to local recurrence of 8 months (range, 0 –128). The
Volume 67, Number 3, 2007
5-year estimate of LPFS was 62%. In 28 cases, local recurrence was identified as a component of the failure, and in 3, it was a subsequent site of failure. Five local recurrences occurred outside the radiation field, with the others located within the previous radiation field. Local recurrences were found as a result of evaluation of a patient’s new symptom (n ⫽ 7), physical examination (n ⫽ 7), or routine CT (n ⫽ 10) or MRI (n ⫽ 7), in the absence of any physical findings. Regional failure Regional lymph node failure occurred in 10 patients, with a median time to relapse of 8 months (range, 0 –141). The 5-year RPFS rate was 87%. Lymph node failure developed in the Level 1 and/or 2 cervical nodes in 5, Level 1– 4 cervical nodes in 3, Level 1 and 5 cervical nodes in 1, and
Fig. 3. Kaplan-Meier survival curves for (a) local-, regional-, and distant relapse-free survival and (b) disease-free and overall survival in patients with squamous cell carcinoma histologic features.
Treatment of nasal cavity and PNS cancer with postoperative RT
●
B. S. HOPPE et al.
697
Table 5. Relapse and survival by histologic type and T stage Variable Histologic type All types SCC AC SA EN AD UN NE Other T stage T2 T3 T4 Recurrence
Alive
LR
RR
DR
Any
59 (69) 23 (55) 8 (73) 9 (100) 6 (86) 6 (100) 2 (50) 2 (100) 3 (75)
31 (36) 21 (50) 2 (18) 2 (22) 0 2 (33) 3 (75) 1 (50) 0
10 (12) 6 (14) 2 (18) 0 0 0 1 (25) 1 (50) 0
12 (14) 6 (19) 4 (36) 0 1 (14) 0 1 (25) 0 0
37 (44) 23 (55) 5 (45) 2 (22) 1 (14) 2 (33) 3 (75) 1 (50) 0
7 (88) 11 (55) 22 (61) 4 (80)
1 (12) 8 (40) 19 (53) 1 (20)
0 4 (20) 6 (17) 0
1 (12) 3 (15) 6 (17) 1 (20)
3 (38) 9 (45) 21 (58) 1 (20)
Abbreviations: LR ⫽ local relapse; RR ⫽ regional relapse; DR ⫽ distant relapse; Any ⫽ relapse anywhere; SCC ⫽ squamous cell carcinoma; AC ⫽ adenoid cystic; SA ⫽ sarcoma; EN ⫽ esthesioneuroblastoma; AD ⫽ adenocarcinoma; UN ⫽ sinonasal undifferentiated carcinoma; NE ⫽ neuroendocrine. Data presented as number of patients, with percentages in parentheses.
retropharyngeal lymph nodes in 1 patient. Two patients, who were diagnosed at presentation with clinically positive necks, developed regional recurrence despite receiving RT to the neck. None of the 5 patients who received ENI developed neck relapses. Regional relapse was the first site of recurrence in 8 patients and a subsequent site in 2 patients. All patients with neck relapse eventually developed local or distant relapse. Regional failure was found because of patients reporting a new neck mass (n ⫽ 4), routine physical examination (n ⫽ 3), or surveillance CT (n ⫽ 1), MRI (n ⫽ 1), or PET (n ⫽ 1). Distant metastasis Distant failure developed in 12 patients, with a median time to relapse of 21 months (range, 1–167). The 5-year DMFS rate was 82%. The location of the first metastasis was the lung in 7, liver in 2, brain in 1, bone in 1, and liver, lung, and spine in 1. Distant metastasis was the first site of failure in 5 patients and a subsequent site in 7 patients. Distant metastases were found on evaluation of a new symptom reported by the patient (n ⫽ 3) or routine chest
X-ray (n ⫽ 4), PET (n ⫽ 3), CT (n ⫽ 1), or MRI (n ⫽ 1), in the absence of other findings. Overall survival In this series, 26 patients died, and the 5-year estimate of OS was 67%. Twenty-four patients died of disease and two of neurologic deterioration without evidence of disease. Of the 5 patients who presented with lymph node involvement, 3 died of their disease, 1 was lost to follow-up after receiving emergent RT for local recurrence, and 1 was alive without evidence of disease 14 months after completing treatment. Histologic type, T stage, and lymph node involvement Because of the heterogeneity of the histologic types, a subgroup analysis was performed on the large subpopulation of patients with SCC and revealed a 5-year LPF, RPFS, DMF, disease-free survival, and OS rate of 49%, 86%, 72%, 47%, and 53%, respectively (Fig. 3). Table 5 lists the pattern of failure by histologic type and T stage. Patients with adenoid cystic carcinoma developed a relapse predom-
Table 6. Significant factors on univariate analysis All patients (n ⫽ 85)
SCC patients only (n ⫽ 42)
Predictor
LPFS
RPFS
DMFS
OS
LPFS
RPFS
DMFS
OS
Pathologic finding (SCC vs. other) Age (⬍60 vs. ⱖ60 y) Race (white vs. other) LNs (involved vs. not) Orbit (involved vs. not) Cribriform plate (involved vs. not)
0.01 0.1 0.37 0.09 0.04 0.01
0.48 0.23 0.55 0.01 0.02 0.86
0.86 0.22 0.59 ⬍0.005 0.01 0.13
0.01 0.59 0.02 0.04 0.01 0.05
NA 0.02 0.03 0.27 0.11 ⬍0.005
NA 0.03 0.13 0.03 0.02 0.25
NA 0.23 0.02 ⬍0.005 ⬍0.005 ⬍0.005
NA 0.06 ⬍0.005 0.25 0.27 0.01
Abbreviations: SCC ⫽ squamous cell carcinoma; LPFS ⫽ local progression-free survival; RPFS ⫽ regional progression-free survival; DMFS ⫽ distant metastasis-free survival; OS ⫽ overall survival; LNs ⫽ lymph nodes.
698
I. J. Radiation Oncology
●
Biology
●
Physics
Volume 67, Number 3, 2007
inantly in a distant site (n ⫽ 4; 36%). Patients with sarcomas did not develop relapse either regionally or distantly. Patients with sinonasal undifferentiated carcinoma (SNUC) predominantly developed local recurrence (n ⫽ 3; 75%). No patient with T2 disease developed a regional relapse. Patients with T3 and T4 disease developed local and regional relapses. Patients with lymph node involvement at presentation developed local (n ⫽ 3, 60%), regional (n ⫽ 2, 40%), and distant (n ⫽ 3, 60%) relapses. Prognostic factors The following factors were evaluated on univariate analysis for predictive value of LPFS, RPFS, DMFS, and OS in all patients and separately in patients with SCC histology: gender (male vs. female), age (⬍60 vs. ⱖ60 years), race (white vs. other), clinical T stage (T2-T3 vs. T4 and T2 vs. T3-T4), lymph node involvement, pathologic findings (SCC vs. other), PNS site (maxillary vs. other), margin status, presence of perineural invasion, orbital exenteration, RT technique (3D-CRT/IMRT vs. conventional), radiation dose (⬍63 vs. ⱖ63 Gy), orbital involvement, base of skull involvement, cribriform plate involvement, infratemporal fossa involvement, muscle involvement, nasopharynx involvement, skin involvement, and bone involvement. The significant factors and their p values are listed in Table 6. Significant variables on univariate analysis were used in a Cox proportional hazards regression multivariate analysis. In the evaluation of all patients, SCC (hazard ratio [HR] 2.66) and cribriform plate involvement (HR 2.99) were significant factors predicting for an increased likelihood of local recurrence, and SCC (HR 3.04) and nonwhite race (HR 3.13) were significant factors predicting for worse OS. In the evaluation of only patients with SCC histologic features, age ⬍60 years (HR 3.70), nonwhite race (HR 3.70), and cribriform plate involvement (HR 4.37) were significant factors predicting for local relapse, and nonwhite race (HR 3.03) and cribriform plate involvement (HR 3.89) were significant for worse OS. Acute and late toxicity The acute and late toxicity from RT by site and grade is detailed in Table 7. Eighty-three patients completed their Table 7. Radiation Therapy Oncology Group acute and late toxicity Grade Toxicity Acute (n ⫽ 85) Eye (n ⫽ 72) Mucous membrane Salivary gland Skin Late (n ⫽ 79) Eye (n ⫽ 66) Mucous membrane Salivary gland Skin
0
1
2
3
4
48 8 26 9
21 27 39 37
3 32 20 33
0 18 0 6
0 0 0 0
62 67 59 70
2 11 17 8
1 1 3 1
0 0 0 0
1 0 0 0
Fig. 4. Isodose curves for patient treated with conventionally planned radiotherapy who developed Grade 4 late visual complication to ipsilateral eye.
prescribed treatment course. Grade 1-2 mucositis was reported in 59 patients (69%) and Grade 3-4 in 18 (21%). Grade 1-2 xerostomia was seen in 59 patients (69%), but no patient developed Grade 3-4 xerostomia as an early complication of RT. Three patients were reported to have Grade 2 conjunctivitis during RT, requiring steroid eye drops, but these cases all resolved. Late toxicity could be evaluated in 79 patients with a minimum follow-up of 3 months. Twelve patients (15%) developed Grade 1-2 mucous membrane toxicity as a late effect of RT, but no patient developed Grade 3-4. Also, 20 patients (25%) developed Grade 1-2 salivary gland toxicity as a late effect, but no patient had Grade 3-4 toxicity. Three patients developed Grade 1-2 late ophthalmologic toxicity. An additional patient developed radiation-induced corneal ulceration, retinopathy, and cataract of his right eye 4 years after RT completion. This ultimately resulted in right unilateral blindness ⬎5 years after RT completion in that eye. This patient had been treated with conventional RT and received ⱕ63 Gy to the retina (Fig. 4). Salvage therapy Patients with relapse underwent a variety of different salvage therapies, depending on the relapse site and the patient’s performance status. Six patients with local relapse underwent surgical resection as a component of their salvage therapy. Only 2 of these patients died (8 and 69 months from the first failure); the other 4 patients were alive (range, 39 – 68 months from the first failure). Five patients underwent repeat RT alone or in combination with chemotherapy. Two of these patients were still alive (10 and 65 months from the first failure) and the other 3 patients died (range, 18 –35 months from the first failure). Chemotherapy alone was given to 7 patients, 6 of whom died (range, 8 –23
Treatment of nasal cavity and PNS cancer with postoperative RT
months from the first failure) and 1 of whom was still alive 17 months from the first failure. Patients receiving only supportive care died within 4 months of the first failure. Patients with only regional relapse as a site of first failure were treated with neck dissection followed by RT with or without chemotherapy. One patient survived, 32 months; the other 2 died 9 –12 months from the first failure. Four patients developed both local relapse and cervical adenopathy. One was treated with surgery and lived 42 months from the first failure, and other three received chemotherapy and eventually died 10 –20 months from the first failure. Patients who presented with metastases were generally treated with palliative RT or chemotherapy and eventually died (range, 4 –21 months from the first failure). One patient with adenoid cystic carcinoma developed a lung metastasis and underwent wedge resection and was still alive 147 months from the first failure. DISCUSSION No randomized controlled studies have evaluated treatment options for patients with PNS/nasal cavity cancer. Treatment recommendations have been based on small retrospective reviews by institutions with heterogeneous populations of patients treated with several different modalities and older RT techniques. In the present study, we reviewed the MSKCC experience with PNS cancer treated with surgery followed by RT. Compared with previous studies (Table 8) (2, 5, 7, 8, 16, 17), our study included the largest number of patients treated with postoperative RT using state-of-the-art RT techniques. The OS and local control were comparable with those of other studies, but the number of patients with Grade 3 or worse ophthalmic toxicity was lower than that seen in the older series using conventional RT techniques. Modern RT techniques using more precise imaging (MRI and PET) for target delineation, achieve a more rapid dose falloff around critical areas, including the retina, optic chiasm, and lens (18 –23) (Figs. 5 and 6) and, theoretically, should lead to fewer long-term risks and complications. Thus far, the outcomes from series using 3D-CRT and IMRT have had less ophthalmologic toxicity compared with
●
B. S. HOPPE et al.
699
older series. Roa et al. (10) studied 39 patients (40% unresectable) treated with 3D-CRT with 54 months of follow-up and found only 1 case of limited optic neuropathy and no case of blindness, with comparable local control and OS to historical studies. Pommier et al. (24) evaluated 40 patients (25% unresectable) treated with 3D treatment planning with a median follow-up of 19 months and reported only 1 patient developing ipsilateral blindness due to vascular glaucoma with a 2-year local control and OS rate of 66% and 73%, respectively. Padovani et al. (9) reported on 25 patients (12% unresectable) treated with 3D-CRT with 25 months of follow-up and found a local control and OS rate of 54% and 24%, respectively, with 2 patients developing uveitis and retinopathy. Most recently, Duthoy et al. (8) evaluated 39 patients with PNS cancer treated with postoperative IMRT with 31 months of follow-up and found only 2 patients with decreased vision after RT, no patient with blindness, and a 4-year local control and OS rate of 68% and 59%, respectively. In the present study, no severe early or late toxicities that affected vision were observed in any of the patients who were treated with modern RT techniques. Only 1 patient in our series developed blindness and that patient had been treated with conventionally planned RT without CT planning. The median Dmax to the eye in the patients treated in this series decreased with the introduction of each new RT technique, but remained constant for the optic chiasm. An example of an IMRT plan with dose painting is illustrated in Fig. 5, and the DVH from the plan is shown in Fig. 6. Local control and survival were not compromised by the newer RT techniques, with a 5-year rate of 62% and 67%, respectively. Further dose reduction to normal tissue may be possible with the use of proton beam therapy. Studies have been done comparing 3D-CRT, IMRT, and proton beam plans and have shown even better coverage of the PTV with lower doses to critical structures (21–23). The outcomes of clinical trials using proton therapy are eagerly awaited, and a headto-head clinical trial may be warranted in the future. Squamous cell carcinoma histologic type was an important prognostic factor predictive for local relapse and worse OS. These patients, as well as those with SNUC, develop recurrence more frequently locally, regionally, and distantly compared with patients with other histologic features (25).
Table 8. Outcomes in other paranasal sinus series using postoperative radiotherapy Investigator
n
Follow-up (mo)
Surgery ⫹ RT (%)
T3-T4 (%)
Maxillary (%)
⫹LNs (%)
SCC (%)
LC (%)
RC (%)
OS (%)
Grade 3–4 eye toxicity
Jiang et al. (16) Duthoy et al. (8) Myers et al. (17) Jansen et al. (5) Le et al. (7) Blanco et al. (2) Present study
73 39 141 73 97 106 85
83 31 69 66 78 60 60
100 100 52 68 63 65 100
74 59 88 86 92 86 88
100 15 70 NA 100 76 53
8 3 4 11 11 7 7
49 21 52 55 60 82 49
78 68 28 63 43 58 62
15 3 6 11 10 29 12
48 59 65 46 34 27 67
16 2 NA 24 4 3 1
Abbreviations: LC ⫽ local control; RC ⫽ regional control; other abbreviations as in Table 6.
700
I. J. Radiation Oncology
●
Biology
●
Physics
Volume 67, Number 3, 2007
neoadjuvant chemotherapy followed by surgery and postoperative RT. Our current approach is to treat these patients aggressively with surgery followed by CTRT. Patients with SCC also need to receive more aggressive therapy, especially when other poor prognostic factors are present. Neoadjuvant CTRT followed by surgery will be investigated in an upcoming North American Skull Base Society protocol. Another consideration could be to investigate targeted systemic therapies that have been shown to be beneficial in other head-and-neck epithelial cancers (29, 30). Lymph node involvement at presentation has been a poor prognostic factor in other studies (2, 5, 17, 25). In this study, lymph node involvement was a significant factor in predicting for regional relapse, distant metastasis, and OS on univariate analysis. These patients were treated with neck dissection followed by RT and had poor outcomes. Neoadjuvant chemotherapy or chemotherapy in combination with RT may improve the outcome for these patients. Patients with neck relapse also have a high risk of local and metastatic recurrences. Therefore, patients at greater risk of regional recurrence may benefit from ENI (16, 31, 32). In our series, none of the 5 patients receiving ENI developed regional relapse. Furthermore, patients who presented with lymph node involvement or neck relapse did so predominantly in the Level 1 or 2 regions. Other studies have found cervical neck relapses to occur in the same regions (31, 32). Currently, patients with T3 or T4 SCC and SNUC treated at MSKCC receive ENI. Cribriform plate involvement was also found on univariate and multivariate analyses to predict for local relapse. When PNS cancers involve the cribriform plate, adequate coverage is difficult without exceeding the RT tolerance for the nearby critical structures. Tsien et al. found a significant underdosage to the PTV in IMRT plans for PNS cancers when the left and right optic pathways were spared (33). However, PTV coverage was much better when only the contralateral optic pathway was spared. Therefore, when treating PNS cancer involving the cribriform plate, a critical evaluation of the treatment plan is necessary and one may need to consider compromising unilateral vision in favor of adequate PTV coverage. Salvage therapy was effective in selected patients. When surgery is a part of the salvage regimen, patients may survive for several years after their first relapse. In combination with RT or CTRT, salvage surgery may provide even longer survival. Although patients treated with chemotherapy alone for salvage have poor survival, they did better than patients receiving only supportive care. CONCLUSION Fig. 5. Isodose curves for intensity-modulated radiotherapy plan.
More aggressive approaches in the treatment of patients with SNUC, including induction chemotherapy followed by CTRT with or without full surgical resection have shown some benefit, with 2-year survival rates of 64 –75% (26 – 28). In this study, the only patient with SNUC who was alive at last follow-up without relapse was treated with
Surgical resection followed by postoperative RT is an effective treatment for cancer involving the PNSs, as well as the nasal cavity. Patients with cervical lymph node involvement, SCC and SNUC, and involvement of the cribriform plate have worse outcomes, thereby necessitating a more aggressive treatment approach. Modern RT techniques, such as IMRT, that can limit the radiation dose to nearby normal tissues do not appear to result in compromised tumor control.
Treatment of nasal cavity and PNS cancer with postoperative RT
●
B. S. HOPPE et al.
701
Fig. 6. Dose–volume histogram for intensity-modulated radiotherapy plan. PTV-DVH ⫽ PTV minus build-up region; OPNRV ⫽ optic nerve.
REFERENCES 1. Silverberg E, Grant RN. Cancer statistics, 1970. CA Cancer J Clin 1970;20:11–23. 2. Blanco AI, Chao KS, Ozyigit G, et al. Carcinoma of paranasal sinuses: Long-term outcomes with radiotherapy. Int J Radiat Oncol Biol Phys 2004;59:51–58. 3. Carrillo JF, Guemes A, Ramirez-Ortega MC, et al. Prognostic factors in maxillary sinus and nasal cavity carcinoma. Eur J Surg Oncol 2005;31:1206 –1212. 4. Harrison LB, Raben A, Pfister DG, et al. A prospective phase II trial of concomitant chemotherapy and radiotherapy with delayed accelerated fractionation in unresectable tumors of the head and neck. Head Neck 1998;20:497–503. 5. Jansen EP, Keus RB, Hilgers FJ, et al. Does the combination of radiotherapy and debulking surgery favor survival in paranasal sinus carcinoma? Int J Radiat Oncol Biol Phys 2000;48: 27–35. 6. Katz TS, Mendenhall WM, Morris CG, et al. Malignant tumors of the nasal cavity and paranasal sinuses. Head Neck 2002;24:821– 829. 7. Le QT, Fu KK, Kaplan M, et al. Treatment of maxillary sinus carcinoma: A comparison of the 1997 and 1977 American Joint Committee on cancer staging systems. Cancer 1999;86: 1700 –1711. 8. Duthoy W, Boterberg T, Claus F, et al. Postoperative intensity-modulated radiotherapy in sinonasal carcinoma: Clinical results in 39 patients. Cancer 2005;104:71– 82. 9. Padovani L, Pommier P, Clippe SS, et al. Three-dimensional conformal radiotherapy for paranasal sinus carcinoma: Clinical results for 25 patients. Int J Radiat Oncol Biol Phys 2003;56:169 –176. 10. Roa WH, Hazuka MB, Sandler HM, et al. Results of primary and adjuvant CT-based 3-dimensional radiotherapy for malignant tumors of the paranasal sinuses. Int J Radiat Oncol Biol Phys 1994;28:857– 865. 11. Greene FL, for the American Joint Committee on Cancer and American Cancer Society. AJCC cancer staging handbook: From the AJCC cancer staging manual. 6th ed. New York: Springer; 2002. 12. Spirou SV, Chui CS. A gradient inverse planning algo-
13. 14. 15.
16. 17. 18.
19.
20. 21.
22. 23.
rithm with dose-volume constraints. Med Phys 1998;25: 321–333. Mohan R, Barest G, Brewster LJ, et al. A comprehensive three-dimensional radiation treatment planning system. Int J Radiat Oncol Biol Phys 1988;15:481– 495. Spirou SV, Chui CS. Generation of arbitrary intensity profiles by combining the scanning beam with dynamic multileaf collimation. Med Phys 1996;23:1– 8. Cox JD, Stetz J, Pajak TF. Toxicity criteria of the Radiation Therapy Oncology Group (RTOG) and the European Organization for Research and Treatment of Cancer (EORTC). Int J Radiat Oncol Biol Phys 1995;31:1341–1346. Jiang GL, Ang KK, Peters LJ, et al. Maxillary sinus carcinomas: Natural history and results of postoperative radiotherapy. Radiother Oncol 1991;21:193–200. Myers LL, Nussenbaum B, Bradford CR, et al. Paranasal sinus malignancies: An 18-year single institution experience. Laryngoscope 2002;112:1964 –1969. Brizel DM, Light K, Zhou SM, et al. Conformal radiation therapy treatment planning reduces the dose to the optic structures for patients with tumors of the paranasal sinuses. Radiother Oncol 1999;51:215–218. Huang D, Xia P, Akazawa P, et al. Comparison of treatment plans using intensity-modulated radiotherapy and threedimensional conformal radiotherapy for paranasal sinus carcinoma. Int J Radiat Oncol Biol Phys 2003;56:158 – 168. Lohr F, Pirzkall A, Debus J, et al. Conformal three-dimensional photon radiotherapy for paranasal sinus tumors. Radiother Oncol 2000;56:227–231. Mock U, Georg D, Bogner J, et al. Treatment planning comparison of conventional, 3D conformal, and intensitymodulated photon (IMRT) and proton therapy for paranasal sinus carcinoma. Int J Radiat Oncol Biol Phys 2004;58: 147–154. Lomax AJ, Goitein M, Adams J. Intensity modulation in radiotherapy: Photons versus protons in the paranasal sinus. Radiother Oncol 2003;66:11–18. Miralbell R, Crowell C, Suit HD. Potential improvement of three dimension treatment planning and proton therapy in the
702
24.
25. 26. 27. 28.
I. J. Radiation Oncology
●
Biology
●
Physics
outcome of maxillary sinus cancer. Int J Radiat Oncol Biol Phys 1992;22:305–310. Pommier P, Ginestet C, Sunyach M, et al. Conformal radiotherapy for paranasal sinus and nasal cavity tumors: Three-dimensional treatment planning and preliminary results in 40 patients. Int J Radiat Oncol Biol Phys 2000;48: 485– 493. Bhattacharyya N. Cancer of the nasal cavity: Survival and factors influencing prognosis. Arch Otolaryngol Head Neck Surg 2002;128:1079 –1083. Rischin D, Porceddu S, Peters L, et al. Promising results with chemoradiation in patients with sinonasal undifferentiated carcinoma. Head Neck 2004;26:435– 441. Kim BS, Vongtama R, Juillard G. Sinonasal undifferentiated carcinoma: Case series and literature review. Am J Otolaryngol 2004;25:162–166. Rosenthal DI, Barker JL Jr, El-Naggar AK, et al. Sinonasal malignancies with neuroendocrine differentiation: Patterns of failure according to histologic phenotype. Cancer 2004;101: 2567–2573.
Volume 67, Number 3, 2007
29. Bonner JA, Harari PM, Giralt J, et al. Radiotherapy plus cetuximab for squamous-cell carcinoma of the head and neck. N Engl J Med 2006;354:567–578. 30. Pfister DG, Su YB, Kraus DH, et al. Concurrent cetuximab, cisplatin, and concomitant boost radiotherapy for locoregionally advanced, squamous cell head and neck cancer: A pilot phase II study of a new combined-modality paradigm. J Clin Oncol 2006;24:1072–1078. 31. Le QT, Fu KK, Kaplan MJ, et al. Lymph node metastasis in maxillary sinus carcinoma. Int J Radiat Oncol Biol Phys 2000;46:541–549. 32. Paulino AC, Fisher SG, Marks JE. Is prophylactic neck irradiation indicated in patients with squamous cell carcinoma of the maxillary sinus? Int J Radiat Oncol Biol Phys 1997;39: 283–289. 33. Tsien C, Eisbruch A, McShan D, et al. Intensity-modulated radiation therapy (IMRT) for locally advanced paranasal sinus tumors: Incorporating clinical decisions in the optimization process. Int J Radiat Oncol Biol Phys 2003;55:776 –784.