Parapharyngeal Extension of Nasopharyngeal Carcinoma: Still a Significant Factor in Era of Modern Radiotherapy?

Int. J. Radiation Oncology Biol. Phys., Vol. 72, No. 4, pp. 1082–1089, 2008 Copyright Ó 2008 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/08/$–see front matter

doi:10.1016/j.ijrobp.2008.02.006

CLINICAL INVESTIGATION

Head and Neck

PARAPHARYNGEAL EXTENSION OF NASOPHARYNGEAL CARCINOMA: STILL A SIGNIFICANT FACTOR IN ERA OF MODERN RADIOTHERAPY? WAI T. NG, F.R.C.R.,* SIU H. CHAN, F.R.C.R.,* ANNE W. M. LEE, F.R.C.R.,* KAM Y. LAU, F.R.C.R.,y TZE K. YAU, F.R.C.R.,* WAI M. HUNG, C.M.D., M.SC.,* MICHAEL C. H. LEE, PH.D.,z AND CHEUK W. CHOI, M.SC.* Departments of *Clinical Oncology, y Diagnostic Radiology, and z Physics, Pamela Youde Nethersole Eastern Hospital, Hong Kong, People’s Republic of China Purpose: To retrospectively analyze the prognostic value of parapharyngeal space (PPS) extension after conformal radiotherapy for nasopharyngeal carcinoma. Patients and Methods: Between 1998 and 2005, 700 patients were treated with conformal radiotherapy at 2 Gy/ fraction daily to a total of 70 Gy. All patients underwent staging with magnetic resonance imaging. The incidence of PPS was determined, and the extent of involvement was further subclassified regarding the presence or absence of carotid space (CS) involvement. The prognostic parameters, including age, gender, stage, chemotherapy, additional boosting, and extent of PPS involvement, were analyzed by univariate and multivariate analyses. Results: The median duration of follow-up was 51 months, and the 5-year overall survival rate for the whole group was 73%. The overall incidence of PPS extension was high (74%), and 29% had additional extension to the CS. Multivariate analysis showed age, gender, chemotherapy, T stage, and N stage to be significant prognostic factors, but not PPS involvement with or without CS extension. In the subgroup of patients with Stage T2 disease (n = 242), the presence of PPS involvement alone or PPS plus CS extension had no statistically significant effect in terms of local control (p = 0.68), distant metastases (p = 0.34), or overall survival (p = 0.24) compared with those without PPS involvement (Stage T2a). Conclusions: With better tumor delineation by magnetic resonance imaging and improved coverage using modern radiotherapy techniques, PPS extension per se no longer predicts for disease outcome. Hence, subcategorizing Stage T2 disease is no longer important in future International Union Against Cancer/American Joint Committee on Cancer classifications. Ó 2008 Elsevier Inc. Nasopharyngeal carcinoma, Parapharyngeal space, Prognosis.

interruption provides a potential route for NPC to spread laterally (1, 2). The current American Joint Committee on Cancer (AJCC) staging system has subclassified nasopharyngeal carcinoma (NPC) Stage T2 into T2a (involvement of the oropharynx or nasal cavity without PPS extension) and T2b (with PPS involvement) (3). This subclassification was based on some early series demonstrating PPS involvement as a poor prognosticator for local control, distant metastases, and overall survival (OS) (4 6). In these studies, the patients underwent staging by computed tomography and were mainly treated with Ho’s method (7) or its modified version. This typical radiotherapy (RT) technique comprises two phases. In the first phase, the nasopharynx and upper neck are encompassed by two lateral opposing faciocervical fields and the lower neck is covered by a lower anterior cervical field. In the second phase, the nasopharynx is treated with two lateral opposing

INTRODUCTION The parapharyngeal space (PPS) is an inverted pyramidal fatfilled space in the lateral suprahyoid neck, with its base attaching to the skull base and the apex extending to the superior cornu of the hyoid bone. Anatomically, PPS is bordered anteriorly by the pterygomandibular raphe, anterolaterally by the medial pterygoid muscle, and posterolaterally by the deep lobe of the parotid gland. The definition of the posterior border is not unified: some regard the PPS as being completely separated from the carotid space (CS), and others consider the CS as part of the PPS. Medial to the PPS is the pharyngeal mucosa space, and they are separated by a strong aponeurosis known as the pharyngobasilar fascia, which connects the constrictor muscle to the base of the skull. This fascia is interrupted at the level of sinus Morgagni where the cartilaginous part of the eustachian tube and the levator veli palatine muscle enter the nasopharynx. This Reprint requests to: Wai T. Ng, F.R.C.R., Department of Clinical Oncology, Pamela Youde Nethersole Eastern Hospital, 3 Lok Man Rd., Chai Wan, Hong Kong, People’s Republic of China. Tel: (+852) 2595 4177; Fax: (+852) 2515 1266; E mail: [email protected]

Conflict of interest: none. Received Oct 30, 2007, and in revised form Jan 31, 2008. Accepted for publication Feb 13, 2008. 1082

NPC: prognostic factor on PPS extension d W. T. NG et al.

1083

facial fields and one anterior facial field, and the remaining neck is treated with a single anterior cervical field. However, the typical concave-shaped tumor volume often renders the tumor coverage with such a two-dimensional (2D) technique less than satisfactory. This is especially problematic when the tumor infiltrates posterolaterally to the PPS and wraps around the spinal cord, which could lead to significant underdosing of the posterolateral aspect of the gross tumor volume (GTV). With the advancement of three-dimensional (3D) conformal techniques and computer technologies, better dosimetric coverage of the PPS can now be achieved. We postulated that the poor clinical outcome of PPS extension was predominately related to the suboptimal dose and not other biologic factors. With adequate dose coverage using conformal techniques, the correlation with PPS involvement might no longer exist. Hence, we decided to analyze the outcomes of our patients who were treated with conformal RT and revisited the prognostic implication of PPS extension. PATIENTS AND METHODS Patient characteristics Between March 1998 and March 2005, 713 patients with NPC were treated with radical intent in our department. The median patient age was 49 years (range, 18 90), and 74% were male. Histologic examination showed that 99% had nonkeratinizing carcinoma, and only 1% had the keratinizing type. The median duration of follow up was 4.3 years (range, 0.2 9.2).

Clinical staging Magnetic resonance imaging (MRI) was performed to assess the locoregional tumor extent. A total of 700 patients underwent staging with MRI, and 13 patients did so with computed tomography (be cause of a contraindication to MRI) and were excluded from our analysis. The AJCC staging manual, 5th edition, was used (8). The 5th edition is basically the same as the 6th edition, except for addition of the masticator space as one of the staging criteria for Stage T4 in the latest edition. Only 111 patients (16%) had Stage I IIB disease, and the remainder had Stage III IVB disease. Using the AJCC definition, PPS involvement denoted posterolateral infil tration of the tumor beyond the pharyngobasilar fascia. In this anal ysis, we further subclassified the degree of extension according to the presence or absence of CS involvement.

Treatment and assessment All patients included in this study underwent RT with 6 MV pho tons using 3D conformal techniques throughout the whole course. All received a basic course of RT at 2 Gy/fraction daily to a total dose of 70 Gy. The GTV, including gross disease at the nasopharyn geal region and involved regional lymphatic nodes, was determined from the imaging and endoscopic findings at presentation. The clin ical target volume (CTV) for the 70 Gy dose included the GTV with a 5 10 mm margin (if possible) and the whole nasopharynx. The CTV for the 60 Gy dose covered high risk local structures (includ ing the PPS, posterior third of the nasal cavities and maxillary sinuses, pterygoid processes, base of the skull, lower one half of the sphenoid sinus, anterior one half of the clivus, and petrous tips), bilateral retropharyngeal nodes, and upper lymphatic regions (Fig. 1). The CTV for the 50 Gy dose covered the remaining

Fig. 1. Dose coverage using conformal radiotherapy for nasopha ryngeal carcinoma in presence of parapharyngeal space extension.

potential sites of local infiltration up to the roof of the sphenoid sinus and bilateral cervical lymphatics down to the supraclavicular fossae. The planning target volumes (PTVs) were determined from the CTVs, with a 2 mm margin for setup variations. All patients underwent treatment planning with the Helax TMS treatment planning system. The conformal treatment plan consisted of four phases. The first 40 Gy was delivered by three fields (one anterior and two lateral opposing fields) to encompass the primary tumor and upper neck, with two anteroposterior opposing fields used to cover the lower neck (abutting the upper fields at the level of hyoid bone). After the first 40 Gy, the lower cervical lymphatic was treated with an anterior field only (tilted about 30 inferiorly) to a total of 60 or 70 Gy, depending on whether any gross nodal involvement was present. The primary tumor and upper neck were usually treated with a four field arrangement consisting of one ante rior, two lateral opposing, and one vertex field for another 30 Gy. Progressive cone down and realignment of these radiation beams were arranged every 10 Gy to achieve the differential doses to the PTV (50 Gy, 60 Gy, and 70 Gy). In some cases with bulky postero lateral extension of the primary tumor, an arc field was used to give the required U shaped dose distribution. Our ideal goal was to de liver $95% of the intended dose to 100% of the respective target volume. Patients with Stage I IIB disease were treated with RT alone using conventional fractionation of five fractions weekly. An additional boost of 5 Gy (in two fractions) was given toward the end of the ba sic course; high dose rate brachytherapy using 192Ir was used to de liver this boost to Stage T1 T2a disease and stereotactic RT (SRT) for Stage T2b primary tumors. In our institute, Stage III IVB disease used to be treated with induction chemotherapy followed by concurrent chemotherapy and accelerated RT (2 Gy/fraction daily, six fractions weekly) according

1084

I. J. Radiation Oncology d Biology d Physics

to the departmental protocol. However, starting from 2000, suitable patients were recruited into two trials organized by the Hong Kong Nasopharyngeal Cancer Study Group. Those with Stage T1 T4N2 N3M0 disease were randomized to RT alone or chemoradiotherapy using the Intergroup 0099 regimen of cisplatin concurrent with RT followed by adjuvant cisplatin and 5 fluorouracil. Patients in both arms underwent RT with conventional fractionation (NPC 9901 trial) (9). Those with Stage T3 T4N0 N1M0 disease were random ized to RT alone at conventional fractionation, RT alone at acceler ated fractionation, chemoradiotherapy (using the Intergroup 0099 regimen) with RT at conventional fractionation or chemoradiother apy with RT at accelerated fractionation (NPC 9902 trial) (10). The first assessment of tumor response was performed 8 weeks after RT completion. All patients were assessed by complete physical examination and fiberoptic nasopharyngoscopy. The post treatment MRI scan has become part of the routine follow up investigation be ginning from 2004. Those with residual tumor were given additional RT by SRT at 2.5 Gy/fraction for another 15 20 Gy. Those with local persistence at 16 weeks after RT completion and those with recur rence after an initial remission were considered to have local failure. The patients were followed at least every 3 months during the first 3 years and every 6 months thereafter until death.

Statistical analysis All events were measured from the date of first treatment. The time to the first defining event was assessed for the following end points: local failure free survival (LFFS; persistence/recurrence at local site), distant failure free survival (DFFS; disease recurrence at distant sites), and OS (death from any cause). The actuarial rates were calculated with the Kaplan Meier method, the differences were compared with the log rank test, and Cox’s proportional hazards model was used to analyze the significance of the potential prognos tic factors. All statistical tests used two sided p values, with alpha level of 0.05 as significant. The Statistical Package for Social Sci ences computer program, version 12.0 (SPSS, Chicago, IL) was used for all statistical analyses.

RESULTS Table 1 lists the patient characteristics, tumor factors, and treatment parameters. A total of 700 patients received a total radiation dose of 70 Gy with the basic course of conformal RT. Of these patients, 347 were treated with RT alone, and 353 were given additional treatment with cisplatin-based chemotherapy, of whom 142 received chemotherapy as a concurrent with or without sequential sequence and 211 received induction chemotherapy with or without a concurrent sequence. A total of 320 patients were treated with accelerated RT with 6 fractions weekly. The median dose to the GTV without a boost was on average 72.4 Gy (range, 70.0 76.5), and the median dose to the PTV(70) was on average 72.3 Gy (range, 70.2 74.7). Because boost treatment was refused by, or not feasible for, some patients, only 91 patients received additional RT boost. Of these 91 patients, brachytherapy was used in 30, SRT in 63 (23 for local persistent disease after the basic course), and both methods were applied in 2 patients (with residual tumor despite boosting). At 8 weeks, 4.7% had persistent disease and 0.6% did so at 16 weeks. The 5-year LFFS, DFFS, and OS rate for the whole population was 88%, 82%, and 73%, respectively.

Volume 72, Number 4, 2008

Table 1. Patient characteristics and treatment factors (n = 700) Factor Patient Age (y) Median Range Male gender (%) Tumor (%) Differentiation (WHO type 1) Stage group I IIB III IVB T stage T1 T2a T2b T3 T4 Treatment Six fractions weekly (%) Median dose without boost (Gy) GTV PTV(70 Gy) Additional boost (%) HDR 5 Gy SRT 5 Gy SRT 15 20 Gy for residual tumor Chemotherapy (%) Concurrent + adjuvant Induction + concurrent

Value

49 18 90 74 1 16 84 4 35 35 26 46 72.4 (70.0 76.5) 72.3 (70.2 74.7) 4 6 3 20 30

Abbreviations: WHO = World Health Organization; GTV = gross tumor volume; PTV = planning target volume; HDR = high dose rate (brachytherapy); SRT = stereotactic radiotherapy. Data in parentheses are ranges.

Parapharyngeal involvement In the whole series, the overall incidence of PPS extension was high (74%). Of the patients, 200 patients (29%) had additional extension to the CS. On univariate analysis, the degree of PPS extension seemed to predict for OS, LFFS, and DFFS. However, after stratification according to T stage (T1 disease was excluded, because it did not include PPS extension by definition), the prognostic value of PPS extension was lost for LFFS and DFFS, and it only reached statistical significance in predicting OS for T3 patients (Figs. 2 4). However, this statistical significance was mainly attributed to the differences in the T3 subgroup; hence, additional adjustment of the potential confounding factors was done by multivariate analyses in the subsequent session. Treatment outcome of T2 disease Because the current AJCC staging system subdivided Stage T2 into T2a and T2b disease (with or without PPS extension), we specifically considered this subgroup of patients. A total of 242 patients (35%) had T2 disease, and PPS extension was noted in 156 of them. Figure 5 shows the treatment outcome according to the presence of PPS involvement. No statistical significance was noted for OS, LFFS, and DFFS. The 5-year LFFS, DFFS, and OS rate for Stage T2a was 94%, 87%, and 83%, respectively, and was 95%, 91%, and

NPC: prognostic factor on PPS extension d W. T. NG et al.

1085

Fig. 2. Local failure free survival (LFFS) (stratified by T stage) according to presence of parapharyngeal space (PPS) extension.

86%, respectively, for Stage T2b. Repeating the same analyses for CS involvement again showed no statistical significance for any of these clinical endpoints. The p value according to CS involvement alone for LFFS, DFFS, and OS was 0.7, 0.23, and 0.54, respectively. Furthermore, dosimetric analyses were performed in the subgroup of Stage T2 patients; the average minimal dose to the GTV was 70.5 Gy (range, 67 73). In fact, for Stage T2 patients, a minimum of 95% of the intended dose was delivered to 100% of the

respective target volume, showing that the results were not biased by inadequate coverage. Multivariate analysis To account for the difference in treatment strategies for the varying disease stages, Cox regression analysis was performed for factors, including gender, age, extent of PPS involvement, chemotherapy, brachytherapy boost, SRT boost, T stage (T2a and T2b were grouped together), and

Fig. 3. Distant failure free survival (DFFS) (stratified by T stage) according to presence of parapharyngeal space (PPS) extension.

1086

I. J. Radiation Oncology d Biology d Physics

Volume 72, Number 4, 2008

Fig. 4. Overall survival (stratified by T stage) according to presence of parapharyngeal space (PPS) extension.

N stage. A stepwise forward procedure was used, and the results are summarized in Table 2. The extent of PPS involvement was not an independent prognostic factor for OS, LFFS, or DFFS. DISCUSSION Incidence and definition The incidence and definition of PPS extension has varied in different series and with different imaging modalities.

Teo et al. (6) reported an incidence of 43.5% and defined PPS involvement as the presence of soft-tissue swelling deforming the parapharyngeal fibrofatty tissue plane between the pharyngeal constrictors and the pterygoids in two or more axial sections at the level of C1 vertebra. Sham and Choy (4) semiquantitatively classified PPS extension into three grades using the reference lines joining the pterygoid plates, carotid artery, styloid process, and the ascending ramus of the mandible; their incidence was 84%. Xiao et al. (11) subclassified PPS invasion into prestyloid and poststyloid compartments

Fig. 5. Treatment outcomes in subgroup with T2 disease. L FFS = local failure free survival; D FFS = distant failure free survival; OS = overall survival.

NPC: prognostic factor on PPS extension d W. T. NG et al.

Table 2. Significance of PPS extension on treatment outcome by multivariate analysis Endpoint Local failure Age, per year increase Gender (female vs. male) T stage T1 vs. T4 T2 vs. T4 T3 vs. T4 PPS extension PPS alone (yes vs. no) PPS + CS (yes vs. no) N stage N1 vs. N0 N2 vs. N0 N3 vs. N0 Brachytherapy (yes vs. no) SRT (yes vs. no) Chemotherapy (yes vs. no) Distant failure Age, per year increase Gender (female vs. male) T stage T1 vs. T4 T2 vs. T4 T3 vs. T4 PPS extension PPS alone (yes vs. no) PPS + CS (yes vs. no) N stage N1 vs. N0 N2 vs. N0 N3 vs. N0 Brachytherapy (yes vs. no) SRT (yes vs. no) Chemotherapy (yes vs. no) Overall survival Age, per year increase Gender (female vs. male) T stage T1 vs. T4 T2 vs. T4 T3 vs. T4 PPS extension PPS alone (yes vs. no) PPS + CS (yes vs. no) N stage N1 vs. N0 N2 vs. N0 N3 vs. N0 Brachytherapy (yes vs. no) SRT (yes vs. no) Chemotherapy (yes vs. no)

HR

95% CI

p

0.99 0.45

0.97 1.01 0.23 0.89

0.24 0.26 0.42

0.03 1.99 0.12 0.56 0.24 0.74

1.07 1.62

0.52 2.18 0.76 3.47

0.83 1.22 0.80 0.51 0.74 0.57

0.33 0.51 0.28 0.06 0.15 0.34

0.22 0.02