Int. J. Radiation Oncology Biol. Phys., Vol. 46, No. 3, pp. 541–549, 2000 Copyright © 2000 Elsevier Science Inc. Printed in the USA. All rights reserved 0360-3016/00/$–see front matter
PII S0360-3016(99)00453-8
CLINICAL INVESTIGATION
Head and Neck
LYMPH NODE METASTASIS IN MAXILLARY SINUS CARCINOMA QUYNH-THU LE, M.D.,* KAREN K. FU, M.D.‡, MICHAEL J. KAPLAN, M.D.,§ DAVID J. TERRIS, M.D.,† WILLARD E. FEE, M.D.,† AND DON R. GOFFINET, M.D.* Departments of *Radiation Oncology and †Otolaryngology, Stanford University, Stanford, CA; Departments of ‡Radiation Oncology and §Otolaryngology, University of California, San Francisco, San Francisco, CA Purpose: To evaluate the incidence and prognostic significance of lymph node metastasis in maxillary sinus carcinoma. Methods and Materials: We reviewed the records of 97 patients treated for maxillary sinus carcinoma with radiotherapy at Stanford University and at the University of California, San Francisco between 1959 and 1996. Fifty-eight patients had squamous cell carcinoma (SCC), 4 had adenocarcinoma (ADE), 16 had undifferentiated carcinoma (UC), and 19 had adenoid cystic carcinoma (AC). Eight patients had T2, 36 had T3, and 53 had T4 tumors according to the 1997 AJCC staging system. Eleven patients had nodal involvement at diagnosis: 9 with SCC, 1 with UC, and 1 with AC. The most common sites of nodal involvement were ipsilateral level 1 and 2 lymph nodes. Thirty-six patients were treated with definitive radiotherapy alone, and 61 received a combination of surgical and radiation treatment. Thirty-six patients had neck irradiation, 25 of whom received elective neck irradiation (ENI) for N0 necks. The median follow-up for alive patients was 78 months. Results: The median survival for all patients was 22 months (range: 2.4 –356 months). The 5- and 10-year actuarial survivals were 34% and 31%, respectively. Ten patients relapsed in the neck, with a 5-year actuarial risk of nodal relapse of 12%. The 5-year risk of neck relapse was 14% for SCC, 25% for ADE, and 7% for both UC and ACC. The overall risk of nodal involvement at either diagnosis or on follow-up was 28% for SCC, 25% for ADE, 12% for UC, and 10% for AC. All patients with nodal involvement had T3– 4, and none had T2 tumors. ENI effectively prevented nodal relapse in patients with SCC and N0 neck; the 5-year actuarial risk of nodal relapse was 20% for patients without ENI and 0% for those with elective neck therapy. There was no correlation between neck relapse and primary tumor control or tumor extension into areas containing a rich lymphatic network. The most common sites of nodal relapse were in the ipsilateral level 1–2 nodal regions (11/13). Patients with nodal relapse had a significantly higher risk of distant metastasis on both univariate (p ⴝ 0.02) and multivariate analysis (hazard ratio ⴝ 4.5, p ⴝ 0.006). The 5-year actuarial risk of distant relapse was 29% for patients with neck control versus 81% for patients with neck failure. There was also a trend for decreased survival with nodal relapse. The 5-year actuarial survival was 37% for patients with neck control and 0% for patients with neck relapse. Conclusion: The overall incidence of lymph node involvement at diagnosis in patients with maxillary sinus carcinoma was 9%. Following treatment, the 5-year risk of nodal relapse was 12%. SCC histology was associated with a high incidence of initial nodal involvement and nodal relapse. None of the patients presenting with SCC histology and N0 necks had nodal relapse after elective neck irradiation. Patients who had nodal relapse had a higher risk of distant metastasis and poorer survival. Therefore, our present policy is to consider elective neck irradiation in patients with T3– 4 SCC of the maxillary sinus. © 2000 Elsevier Science Inc. Maxillary sinus carcinoma, Histology, Nodal metastasis, Survival, Radiotherapy, Surgery.
INTRODUCTION
belief that nodal metastasis is uncommon in tumors without extensive lymphatic involvement, most authors have advocated no elective neck irradiation (ENI) for patients with maxillary sinus carcinoma (7–9). This long-held conviction has been recently challenged by Jiang et al. (6) and Paulino et al. (4). Both authors advocated elective ipsilateral neck irradiation in patients with squamous cell carcinoma (SCC) due to the high incidence of neck relapse associated with this histology (28.9 –33%) and the inferior survival of those who relapsed in the neck. At Stanford University (SUH) and
Carcinomas of the maxillary sinus are uncommon, with an annual incidence of ⬍ 1:100,000 in the United States. They comprise 0.2– 0.5% of all cancers, 3% of all head and neck cancers, and 80% of all paranasal sinus cancers (1). In general, they represent a challenging therapeutic dilemma because of their rarity and proximity to critical structures. The incidence of primary nodal metastasis at presentation ranges from 3.3% to 26% (2– 6). Because of the general Reprint requests to: Quynh-Thu Le, M.D., Department of Radiation Oncology, Stanford University, 300 Pasteur Dr., MC 5302, Stanford, CA 94305-5302. E-mail:
[email protected]
Accepted for publication 8 October 1999.
541
542
I. J. Radiation Oncology
●
Biology
●
Physics
at the University of California, San Francisco (UCSF), there has not been a uniform policy on the use of ENI for maxillary sinus carcinoma. The technique and the extent of neck irradiation has also varied depending on the treating physician and the period of treatment. To determine the incidence and the pattern of nodal relapse, its influence on the overall tumor control and survival, and the effectiveness of ENI in preventing nodal metastasis, we undertook a retrospective review of the patients treated definitively for maxillary sinus carcinoma at both institutions. METHODS AND MATERIALS Patient characteristics Between January 1, 1959 and December 31, 1996, 97 patients were treated definitely for carcinoma of the maxillary sinus at SUH and UCSF. There were 67 male and 30 female patients. The median age at diagnosis was 58 (range: 20 – 85). The histologic distribution is as follows: SCC, 58; adenocarcinoma (ADE), 4; undifferentiated carcinoma (UC), 16; and adenoid cystic carcinoma (AC), 19. The median follow-up for all patients was 28 months (range: 2.4 months–30.4 years). The median follow-up for alive patients was 78 months (range: 18 months–23 years). Ninetythree percent and 60% of alive patients had follow-up ⬎ 2 and ⬎ 5 years, respectively. The patient evaluation consisted of a complete history and physical examination with emphasis on the head and neck evaluation, a complete blood count, liver function tests, renal function tests, and chest radiograph. Between 1959 and 1977, plain films and tomograms were used to image the primary sites (47 patients). After 1977, computed tomography (CT) and more recently a combination of CT scans and magnetic resonance imaging (MRI) were added as part of the routine assessment (50 patients). In all patients, the disease was restaged clinically using all the information obtained prior to surgery or radiotherapy according to the 1997 American Joint Committee on Cancer (AJCC) TNM staging system (10). Overall, there were 7 patients with Stage II, 33 with Stage III, and 59 with Stage IV neoplasms. Eight patients had T2, 36 T3, and 53 T4 tumors. Eleven patients had documented nodal involvement at diagnosis. Table 1 shows the N-stage distribution, the sites of nodal involvement, and the nodal distribution according to tumor histology. Treatment of primary tumors Fifty-six patients were treated with surgical resection followed by postoperative radiotherapy; five had preoperative radiation and surgery; and 36 had radiation treatment alone. A total or radical maxillectomy was performed in 40 patients, radical maxillectomy and orbital exenteration in 16 patients, and partial maxillectomy in 5 patients. Of the 61 patients treated with surgery, 19 had negative surgical margins, 29 had microscopically positive margins, and 13 had grossly positive margins. Radiotherapy was delivered with megavoltage energy by either telecobalt machines (12 pa-
Volume 46, Number 3, 2000
Table 1. N-stage distribution, the sites of nodal involvement, and the nodal distribution according to tumor histology
Parameter N-stage N0 N1 N2b N2c Nodal region Ipsilateral level I Ipsilateral level II Contralateral level II Histology Squamous cell carcinoma Adenocarcinoma Undifferentiated carcinoma Adenoid cystic carcinoma
No. of pts
No. of pts N⫹ at diagnosis (%)
86 6 3 2 3 8 2 58 4 16 19
9 (16) 0 1 (6) 1 (5)
Pts ⫽ patients; N ⫽ nodal.
tients) or linear accelerators (85 patients). Three patients also received brachytherapy boosts (2 with I125 and 1 with Ir192), and 2 had particle therapy boosts (1 with a neon beam, and 1 with a helium beam). In the majority of the patients, external beam therapy was delivered via a 3-field technique to cover the paranasal sinuses and the ipsilateral orbital floor. In a small number of patients, a wedge-pair technique was employed. The field size ranged from 4.0 ⫻ 4.6 cm to 14 ⫻ 14 cm (median 9 ⫻ 8 cm). A wedge-pair or a 3-field technique was employed in all cases. All but 2 patients were treated with conventional once-daily fraction. The fraction size ranged from 1.5–3.0 Gy/day (median 2 Gy); 5 patients received fraction size ⬍ 1.8 Gy and 12 had fraction size ⬎ 2 Gy. One patients was treated with 1.2 Gy twice daily to a total dose of 64.8 Gy, and one was treated on a hypofractionated protocol to a total dose of 48 Gy in 12 fractions over 5 weeks (4 Gy/fraction, 2–3 fractions/week) with concurrent misonidazole. The median tumor dose for patients treated with radiotherapy alone was 64.8 Gy (range: 46.4 –72.1 Gy); the median dose for patients receiving postoperative and preoperative radiotherapy was 64.8 Gy (range: 46 –76 Gy) and 60.4 Gy (range 50 – 65 Gy), respectively. Treatment of neck nodes Of the 11 patients with involved nodes at presentation, one had combined radical neck dissection and postoperative neck irradiation to a total dose of 65 Gy. The remainder received neck irradiation alone with a median dose of 63 Gy (range: 48 –73.4 Gy). Twenty-six patients had ENI for N0 necks. The upper neck was treated with opposed lateral fields matching on to the sinus fields. A small block was placed over the spinal cord at the field junction to protect the spinal cord from overdosage. An anteroposterior field was used to treat the supraclavicular area with the dose prescribed to a depth of 3 cm. A midline larynx block was
Lymph node metastasis in maxillary sinus carcinoma
●
Q.-T. LE et al.
543
Fig. 1. Overall survival for all patients.
placed a the junction of the upper and lower neck fields. The majority of these patients (20/26) received 50 Gy to the neck (range: 34 – 65 Gy). ENI was given to 13 patients with SCC, 5 with UC, 6 with AC, and 2 with ADE. Of the patients who received ENI, 2 had T2, 9 had T3, and 15 had T4 tumors. Systemic treatment Fourteen patients received systemic therapy before, during, or after radiotherapy. Eleven were treated with chemotherapy, and 3 received misonidazole for radiosensitization. The most common chemotherapeutic agents used were cisplatin and 5-fluorouracil (5-FU) (6 patients). Statistical analysis Estimates of freedom from local relapse, freedom from regional relapse, freedom from distant metastases (FFDM), and overall survival (OS) were computed using the KaplanMeier product limit method (11). For nodal control analysis, the first nodal relapse was scored. Patients who died of local recurrence or distant metastasis or intercurrent illnesses without nodal relapse were censored at the last follow-up. Log-rank statistics were employed to identify important prognostic factors for regional control, and FFDM, and survival. Analyzed variables included age, gender, tumor histology, T-stage, N-stage, local therapy, radiation dose, radiotherapy overall time, chemotherapy, local control, and nodal control. Cox proportional hazard models using the forward stepwise approach were applied to all potentially significant prognostic variables for multivariate analysis (12). Since there was a strong interaction between local therapy and local control, only local therapy was used in the
multivariate analysis for survival and distant metastasis. Stata Statistical Software was used for statistical evaluations (13).
RESULTS Survival The estimated 5- and 10-year survival for all patients was 31% and 34%, respectively. The median survival was 22 months. Figure 1 shows the Kaplan-Meier survival curve for all patients. Pattern of relapse At the time of this analysis, 65 patients have relapsed. Eighty-six percent (56/65) recurred within 2 years of diagnosis. Six of 9 patients who failed more than 2 years from diagnosis had AC histology. The sites of initial relapse after treatment and subsequent relapse after salvage therapy are shown in Table 2. Overall, local relapse was the most common pattern of recurrence, accounting for 61% of all sites of relapse.
Table 2. Pattern of relapse Number of patients Site
Initial relapse
Subsequent relapse
Total
Primary only Primary ⫹ Lymph node Lymph node only Distant metastasis
45 1 7 12
6 0 2 11
51 1 9 23
544
I. J. Radiation Oncology
●
Biology
●
Physics
Volume 46, Number 3, 2000
Fig. 2. (a) Pattern of nodal relapse. (b) Pattern of nodal involvement either at diagnosis or on follow-up.
Neck failure At the latest follow-up, 10 patients have relapsed in the neck: 9 with initial N0 necks, and 1 with initial N1 neck. The estimated risk of developing a neck recurrence at 5 years was 12%. The median time to neck relapse was 7 months (range: 2–16 months). One patient relapsed in the neck alone; one had simultaneous primary and regional failures; 2 failed locally followed by regional relapse; 1 failed regionally followed by a local failure; 2 failed region-
ally followed by local and distant relapses; and 3 had regional failures followed by distant metastases. Figures 2a and 2b show the sites of neck recurrence and overall nodal involvement (at diagnosis and recurrence), respectively. Eight patients suffered a nodal relapse at a single site, 1 relapsed in both ipsilateral submandibular (level I) and superior posterior cervical lymph nodes (level V), and 1 patient had a bilateral nodal recurrence, both in the jugulo-digastric areas (level II). As shown in Figs. 2a
Lymph node metastasis in maxillary sinus carcinoma
●
Q.-T. LE et al.
545
Table 3. 5-year estimates of neck relapse according to potential prognostic factors Parameter T-stage 2 3 4 N-stage 0 1–2 Histology SCC ADE UC AC NC involved Yes No NP involved Yes No OC involved Yes No SC involved Yes No Surgical resection Yes No Neck RT Yes No Systemic therapy Yes No Primary control Yes No
No. of pts
% 5-yr neck relapse
8 36 53
0 16 12
86 11
13 10
0.98
58 4 16 19
14 25 6 5
0.62
29 68
4 14
0.14
23 74
19 12
0.32
23 74
12 13
0.75
7 90
14 12
0.77
61 36
15 6
0.39
37 60
10 14
0.53
83 14
13 10
0.81
46 51
10 14
0.44
p value
0.51
NC ⫽ nasal cavity; NP ⫽ nasopharynx; OC ⫽ oral cavity; SC ⫽ subcutaneous; RT ⫽ radiotherapy.
and 2b, the most common sites of initial nodal involvement and nodal recurrence were ipsilateral level I and II nodes. Table 3 shows the 5-year risk of neck relapse in relation to
Fig. 3. Freedom from regional relapse in patients with N0 necks and SCC histology by elective nodal irradiation.
a number of tumor- and treatment-related parameters. None of the factors evaluated reached statistical significance, most likely due to the small number of neck relapses in this series. There was a suggestion of increased neck relapse with increasing tumor stage. All of the neck relapses were observed in patients with T3– 4 neoplasms, and none was seen for the 8 patients with T2 tumors. Similarly, there was an association between tumor histology and neck relapse. SCC and ADE had the highest risk of nodal involvement and relapse in comparison to other histologies. Patients treated with combined surgery and radiotherapy relapsed in the neck more often than those treated with radiotherapy alone, and tumor invasion into the nasopharynx was associated with a higher risk of neck failure. Tumor extension to other areas containing rich lymphatic network, N-stage at presentation, and primary tumor control did not appear to predict for neck failure. The use of systemic therapy also did not affect the risk of nodal relapse. Table 4 shows the pattern of nodal relapse by T-stage, ENI, primary tumor control, and the overall incidence of nodal involvement for each tumor histology. Because there
Table 4. Estimate of nodal relapse by T-stage, elective neck irradiation, and local control according to individual histology in patients with N0 neck at presentation % 5-year risk of nodal relapse (number failed/total)
No. pts
T2
T3
T4
Yes
No
Yes
No
% Overall risk of nodal involvement
49 4 15 18
0 (0/5) — 0 (0/1) 0 (0/2)
15 (2/14) 33 (1/3) 17 (1/6) 0 (0/5)
17 (4/30) 0 (0/1) 0 (0/8) 13 (1/11)
0 (0/13) 50 (1/2) 20 (1/5) 0 (0/5)
20 (6/26) 0 (0/2) 0 (0/10) 12 (1/13)
13 (4/26) 50 (1/2) 0 (0/5) 0 (0/9)
15 (2/23) 0 (0/2) 12 (1/10) 17 (1/9)
28 25 12 10
T-stage Histology SCC ADE UC AC
ENI
LC
ENI ⫽ elective nodal irradiation; LC ⫽ local control; SCC ⫽ squamous cell carcinoma; ADE ⫽ adenocarcinoma; UC ⫽ undifferentiated carcinoma; AC ⫽ adenoid cystic carcinoma.
546
I. J. Radiation Oncology
●
Biology
●
Physics
Volume 46, Number 3, 2000
Fig. 4. FFDM by regional control.
were only 4 patients with ADE histology, further analysis of this subgroup is not feasible. Of the patients with SCC histology, all neck relapses occurred in those with T3– 4 tumors (Table 4), and ENI effectively prevented nodal relapse (Fig. 3). Salvage therapy for regional relapse Salvage therapy for patients with nodal relapse consisted of neck dissection alone in 4 patients, neck dissection and postoperative neck irradiation in 4 patients, and neck radiotherapy plus hyperthermia plus chemotherapy in 1 patient. One patient received no treatment. Only 1 patient with regional recurrence alone was effectively salvaged and remained disease-free. One patient had persistent neck disease at death; the remainder progressed with either local or distant relapses and died of their neoplasms shortly after regional recurrences. Relationship between regional relapse and distant metastasis and survival There was a significant association between neck control and the risk of distant relapse. The 5-year estimates of FFDM were 71% for patients with regional control and 18% for those with regional failure (p ⫽ 0.02, Fig. 4). The only other factor that was important for FFDM on univariate analysis was N-stage (favoring N0 neck, p ⫽ 0.0005). T-stage was of borderline significance (p ⫽ 0.06). Age, gender, tumor histology, local therapy, and local control did not achieve statistical significance (p value ⬎ 0.10 for all variables). On multivariate analysis, T-stage (hazard ratio
[HR] ⫽ 1.3, p ⫽ 0.007), N-stage (HR ⫽ 7.9, p ⬍ 0.001), and regional nodal control (HR ⫽ 4.5, p ⫽ 0.006) were significant prognostic factors for FFDM. Similar prognostic factors were analyzed for overall survival. Age (p ⫽ 0.009), T-stage (p ⫽ 0.02), N-stage (p ⫽ 0.007), local therapy (p ⫽ 0.006), radiotherapy overall time (p ⫽ 0.002), and local tumor control (p ⬍ 0.001) achieved statistical significance for survival on univariate analysis. There was also a trend toward decreased survival in male patients (p ⫽ 0.06). Patients with regional failure appeared to have lower survival than those without. The 5-year survival estimates were 37% for those with neck control and 0% for those with nodal relapse; however, the difference did not reach statistical significance on univariate analysis (p ⫽ 0.3). Tumor histology, the use of systemic therapy, and radiation dose did not achieve statistical significance for survival. Because there was a strong interaction between local control and local therapy, local control was not used for multivariate analysis. Age (favoring younger patients, HR ⫽ 1.03/year, p ⬍ 0.001), T-stage (favoring T2–3, HR ⫽ 0.37, p ⫽ 0.001), N-stage (favoring N0 necks, HR ⫽ 0.31, p ⫽ 0.003), local therapy (favoring patients treated with combined surgery and radiotherapy, HR ⫽ 0.45, p ⫽ 0.002), gender (favoring female patients, HR ⫽ 0.55, p ⫽ 0.04), and neck control (favoring patients with neck control, HR ⫽ 0.43, p ⫽ 0.04) were important prognostic factors for survival on multivariate analysis. In addition, when the overall survival for patients with nodal involvement either at presentation or at relapse was compared to that for patients with neck control, the difference was statistically
Lymph node metastasis in maxillary sinus carcinoma
●
Q.-T. LE et al.
547
Fig. 5. Overall survival by nodal involvement either at diagnosis or on follow-up.
significant (p ⫽ 0.01, Fig. 5). The 5-year survival estimate was 40% for patient with regional control and 8% for those with neck disease either on presentation or at relapse. Complications Sixteen patients developed severe, late treatment-related complications, 15 of which were secondary to the treatment of the primary site. Only 1 patient developed late complications related to the neck irradiation. This patient was treated with definitive radiotherapy in 1966 for a T4N0 UC of the maxillary sinus. He received 70 Gy to the primary site via a 3-field technique, and 60 Gy to the left neck with an AP:PA technique using 4 MV X-rays. He developed ipsilateral blindness, severe trismus, left brachial plexopathy, and severe neck fibrosis 18 months after the completion of treatment. He died a year later with local recurrence.
DISCUSSION Carcinoma of the maxillary sinus is a rare condition in the United States. Because of the tumor location and the lack of early symptoms, patients usually present with advanced disease, resulting in poor local control and survival. In recent years, with the use of combined modality therapy employing modern surgical and radiotherapeutic approaches and chemotherapy, investigators have noted an improvement in treatment results (14). With better local control, more attention is being paid to regional and distant metastases. Recently, some studies have demonstrated an association between tumor histology and the incidence of nodal relapse (6, 15). Many have also shown poorer survival or disease-specific survival with regional failure (4, 8). The issue of ENI for these tumors has been controversial. Fletcher et al. (16), in their textbook, recommended ENI for
Table 5. Incidence of nodal relapse by tumor histology without elective nodal irradiation reported in the literature % Neck relapse Author
SCC
ADE
UC
AC
ME
Korzeniowski (15) Jiang (6) Paulino (20) Shidnia (5) Le et al.
21 33 29 16 20
12 20 N/A N/A 0
20 50 N/A N/A 0
N/A 5 N/A N/A 14
0 0 N/A N/A N/A
N/A ⫽ not available; UC ⫽ undifferentiated carcinoma; ACC ⫽ adenoid cystic carcinoma; ME ⫽ mucoepidermoid carcinoma.
548
I. J. Radiation Oncology
●
Biology
●
Physics
patients with T3– 4 tumors, a view supported by Bataini and Ennuyer (2). In contrast, others do not advocate ENI due to the low incidence of nodal failure, poor local control, and low survival rates in these patients (7, 17). Recently, Jiang et al. (6) revisited the issue, and recommended ENI for patients with SCC and UC histologies based on a ⬎ 30% incidence of neck relapse in N0 patients when the neck is not treated. Table 5 is a summary of the incidence of nodal relapse by tumor histology reported in the literature for maxillary sinus carcinoma when no ENI was given. In patients treated without ENI, Jiang et al. (6) reported a 33% risk of neck relapse for SCC, 50% for UC, and 5% for both AC and mucoepidermoid histologies. Patients with regional failure had an inferior 10-year disease-specific survival rate compared to those without (34% vs. 58%). Based on these findings, the authors advocated elective nodal treatment for patients with T2– 4 squamous cell or UC. Korzeniowski et al. (15) noted a similar influence of histology on regional control. In their report, the frequency of nodal metastasis at presentation and during follow-up was 60% in patients with UC, 30% in patients with nonkeratinizing SCC, and 13% for patients with keratinizing SCC. They recommended ENI for the UC and nonkeratinizing SCC histology. More recently, Paulino et al. (4) reviewed 42 patients with SCC of the maxillary sinus treated at Loyola University with either radiotherapy alone or with combined surgery and radiotherapy. Of 38 patients with initial node-negative necks, 11 (28.9%) developed nodal relapse. Their analysis showed that only tumor stage was predictive of neck relapse with T1–2 cancers doing worse than T3– 4. Tumor location (supra vs. infrastructure), extension to areas containing rich lymphatic networks, and local control were not significant prognostic factors for nodal relapse. They also noted an inferior median survival in patients with neck node metastases either on initial presentation or on follow-up when compared to those without nodal disease (25 months vs. 80 months). The authors, therefore, recommended elective ipsilateral neck irradiation for all patients with SCC of the maxillary sinus. In our study, the overall risk of neck relapse calculated by the Kaplan-Meier curve is 12% at 5 years, which is lower than observed in previous studies. However, 31% of the patients had ENI for N0 necks. We observed a higher incidence of neck involvement and neck relapse in patients with SCC and ADE histologies. The number of patients with ADE histology in the series was too small to draw definitive conclusions. As shown in Tables 3 and 4, all of the neck relapses were associated with T3– 4 cancers, and ENI effectively prevented nodal relapse in patients with SCC histology. The most common sites of nodal relapse and initial nodal spread were ipsilateral level I–II neck nodes. Similar to previous reports (4, 8), nodal metastasis at diagnosis and nodal relapse were associated with poor survival in this study. While 37% of patients with neck control were alive at 5 years, none of those with neck relapse were alive.
Volume 46, Number 3, 2000
We also observed a strong association between nodal relapse and the development of distant metastasis (Fig. 4). In fact, nodal relapse was an independent prognostic predictor for distant tumor spread on multivariate analysis. Previous studies have suggested an increased risk of neck relapse with tumor extension into areas with rich lymphatic networks (18, 19). More recently, this notion has been questioned by Paulino et al. (4), who found no association between nodal relapse and tumor involvement of the nasal cavity, nasopharynx, oral cavity, and oropharynx. The results of the present series confirmed this lack of association, with the exception of a nonsignificant trend toward increased neck recurrence with nasopharyngeal involvement. Aside from the suggestion of increased nodal relapse with T3– 4 tumors and SCC/ADE histologies, we also found no correlation between nodal relapse and either N-stage or local tumor control. In the present study, late complications related to neck therapy were rare. Only 1 of 36 patients receiving neck irradiation developed treatment-related toxicity (ipsilateral brachial plexopathy and severe neck fibrosis). This same patient also developed ipsilateral blindness and severe trismus from treatment of the primary site. Review of his treatment data showed appropriate tumor doses and dosimetry. It is possible that this patient, who was diagnosed with an UC at age 20, had a genetic abnormality in his DNA repair processes that increased his susceptibility to radiation damage. Unfortunately, fresh tissues from the patient are not available to analyze for such abnormality. A similar low rate of complications related to neck treatment was noted in the M. D. Anderson series (6). Nearly all of their described complications were secondary to treatment of the primary site. They noted one case of mandibular osteoradionecrosis secondary to ENI. Because most of the neck failures occurred in the ipsilateral level I and II lymph nodes both in the present study and in the study reported by Paulino et al. (4), treatment toxicity may be minimized by considering elective radiation only to the ipsilateral upper neck. We recognize that this is a retrospective study with all of the potential inherent biases associated with this type of review. However, this is one of the larger series in this country with significant long-term follow-up for alive patients. Over half of the alive patients were followed for more than 5 years. Our data suggest that nodal relapse was strongly associated with a higher risk of distant metastasis and poorer overall survival. In addition, it can be effectively prevented with minimal morbidity by elective nodal irradiation in patients at high risk for regional failure. A multicenter prospective randomized study would be the best approach to evaluate the effectiveness of elective nodal irradiation in this setting; unfortunately, the rarity of these tumors would make it difficult for such a study to be completed in a timely fashion. Meanwhile, our treatment policy is to consider elective ipsilateral neck irradiation for patients with T3– 4 SCC of the maxillary sinus.
Lymph node metastasis in maxillary sinus carcinoma
CONCLUSION The 5-year risk of nodal relapse following therapy was 12% in this study. T3– 4 cancers with SCC histology were associated with a high incidence of initial nodal involvement and
●
Q.-T. LE et al.
549
nodal relapse. None of the patients presenting with SCC histology and N0 necks had nodal recurrence after ENI. Patients who had nodal relapse had a higher risk of distant metastasis and poorer survival. Therefore, our policy is to consider ENI in patients with T3– 4 SCC of the maxillary sinus.
REFERENCES 1. Grant RN. Cancer statistics 1970. Ca Cancer J Clin 1970;8: 6 –20. 2. Bataini JP, Ennuyer A. Advanced carcinoma of the maxillary antrum treated by cobalt teletherapy and electron beam irradiation. Br J Radiol 1971;44:590 –598. 3. Cheng VS, Wang CC. Carcinomas of the paranasal sinuses: A study of sixty-six cases. Cancer 1977;40:3038 –3041. 4. 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. 5. Shidnia H, Hornback NB, Saghafi N, et al. The role of radiation therapy in treatment of malignant tumors of the paranasal sinuses. Laryngoscope 1984;94:102–106. 6. Jiang GL, Ang KK, Peters LJ, et al. Maxillary sinus carcinomas: Natural history and results of postoperative radiotherapy. Radiother Oncol 1991;21:193–200. 7. Pezner RD, Moss WT, Tong D, et al. Cervical lymph node metastases in patients with squamous cell carcinoma of the maxillary antrum: The role of elective irradiation of the clinically negative neck. Int J Radiat Oncol Biol Phys 1979;5: 1977–1980. 8. Giri SP, Reddy EK, Gemer LS, et al. Management of advanced squamous cell carcinomas of the maxillary sinus. Cancer 1992;69:657– 661. 9. Lavertu P, Roberts JK, Kraus DH, et al. Squamous cell carcinoma of the paranasal sinuses: The Cleveland Clinic experience 1977–1986. Laryngoscope 1989;99:1130 –1136. 10. Paranasal sinuses. In: Fleming RD, Cooper JS, Henson DE,
11. 12. 13. 14. 15.
16.
17. 18. 19. 20.
Hutter RVP, Kennedy BJ, Murphy GP, O’Sullivan B, Sobin LH, editors. American Joint Committee on Cancer staging manual. Philadelphia: Lippincott-Raven; 1992. p. 47–52. Kaplan ES, Meier P. Non-parametric estimations from incomplete observations. Am Stat Assoc J 1958;53:457– 482. Cox DR. Regression models and life tables. J R Stat Soc 1972;34:187–220. Statistics with Stata, Version 3. Computing Resource Center. Santa Monica, 1992. Shibuya H, Horiuchi JI, Suzuki S, et al. Maxillary sinus carcinoma: Results of radiation therapy. Int J Radiat Oncol Biol Phys 1984;10:1021–1026. Korzeniowski S, Reinfuss M, Skolyszewski J. The evaluation of radiotherapy after incomplete surgery in patients with carcinoma of the maxillary sinus. Int J Radiat Oncol Biol Phys 1985;11:505–509. Fletcher GH. Nasal and paranasal sinus carcinoma. In: Fletcher GH, Goepfert H, Jesse RHJ, editors. Textbook of radiotherapy. Philadelphia: Lea and Febiger; 1980. p. 408 – 425. Bush SE, Bagshaw MA. Carcinoma of the paranasal sinuses. Cancer 1982;50:154 –158. Dalley VM. Malignant disease of the antrum. Br J Radiol 1959;32:378 –385. Del-Regato JA. Roentgentherapy in epitheliomas of the maxillary sinus. Surg Gynecol Obstet 1937;65:657– 665. Paulino AC, Marks JE, Bricker P, et al. Results of treatment of patients with maxillary sinus carcinoma. Cancer 1998;83: 457– 465.