Comparison of The 7th and 8th Editions of The Union for International Cancer Control /American Joint Committee on Cancer Staging System for Nonmetastatic Nasopharyngeal Carcinoma in The Intensity-Modulated Radiotherapy Era

Background: To compare the prognostic value of 7th and 8th editions of the Union for International Cancer Control /American Joint Committee on Cancer (AJCC) staging system for patients with nonmetastatic nasopharyngeal carcinoma (NPC) treated with intensity-modulated radiotherapy and simultaneous integrated boost– intensity-modulated radiation therapy (SIB-IMRT). Methods: Patients with NPC (n = 300) who received SIB-IMRT were included. Survival by T-classication, N-classication, and stage group of each staging system was assessed. Results: For T-classication, nonsignicant difference was observed between T1 and T2 and between T2 and T3 disease (P = 0.066 and 0.106, respectively) for overall survival (OS) in the 7th staging system, whereas all these differences were signicant in the 8th staging system (all P < 0.05). The survival curves for disease-free survival (DFS) and LRFS in both staging systems were similar, except for the comparison of T2 and T4 disease for LRFS (P = 0.07 for 7th edition; P = 0.011 for 8th edition). For N-classication, signicant differences were observed between N2 and N3 diseases after revision (P = 0.046 and P = 0.043 for OS and DFS, respectively). For staging system, no signicant difference was observed between IVA and IVB of 7th edition. Conclusion: The 8th AJCC staging system appeared to have superior prognosis value in the SIB-IMRT era compared with the 7th edition.


Introduction
Nasopharyngeal carcinoma (NPC) develops in the epithelial lining of the nasopharynx, the narrow tubular passage behind the nasal cavity, and radiation therapy (RT) is the primary treatment because of anatomical constraints and high radiosensitivity of this carcinoma.
Over the past 2 decades, NPC management has undergone substantial changes. Magnetic resonance imaging (MRI), which has been widely used in the clinical staging of NPC, has made it possible to de ne tumour volume precisely and allow the early detection of occult metastases [1][2][3]. In terms of disease modality, intensity-modulated radiation therapy (IMRT) could decrease the overall treatment time and increase the fractionation dose to planned target volume (PTV) with relatively less late toxicity compared with 2-dimensional conventional or 3-dimensional conformal radiotherapy [4][5][6][7][8]. Therefore, the new staging system should be based on up-to-date clinical data and maintain its relevance with current management approaches.
The classi cation of N3 in the 7th edition of the Union for International Cancer Control /American Joint Committee on Cancer (UICC/AJCC) NPC staging system is mainly based on anatomy, and its ability to predict prognosis is limited [9,10]. The UICC/AJCC released the 8th edition of this staging system in an attempt to further help clinicians opt for the best treatment for patients. The revised staging system was mainly derived from a study on 1609 patients with NPC on the basis of their MRI ndings and IMRT status [11]. A few key revisions in the 8 th edition are as follows: (1) for T-classi cation, patients with infratemporal fossa/masticator space involvement, which was considered in the T4 category in the 7th edition, has been replaced with a precise de nition; (2) medical pterygoid (MP) and lateral pterygoid (LP) muscles have been downgraded to the T2 category, whereas prevertebral muscles have been included in the T2 category; (3) For the N-classi cation, the supraclavicular fossa (SCF) has been replaced by the caudal border of the cricoid cartilage, and N3a and N3b in the 7th edition have been reclassi ed to N3; (4) T4 and N3 have been merged into IVA; (5) For the clinical stage, stages IVA and IVB in the 7th edition have been rede ned as IVA, and stage IVC has been reclassi ed as IVB in the 8th edition. Table 1 shows the classi cation criteria of the 7th and 8th editions of the UICC/AJCC nasopharyngeal carcinoma staging system.
Patients with NPC can receive IMRT via 2 push modes: the conventional sequential boost and simultaneous integrated boost. Compared with the conventional method, SIB-IMRT can be delivered in different target regions during the same treatment session and has a shorter waiting time. In addition, most clinical data revealed that SIB-IMRT had better sparing of the parotid glands and inner ear structures [12,13].
We performed this retrospective study to explore the clinical outcomes of SIB-IMRT, and compare the 7th and 8th editions of the UICC/AJCC staging system.

Patient characteristics
A total of 300 patients with newly diagnosed, pathologically proven, non-distant metastatic NPC who were treated with SIB-IMRT at West China Hospital between February 2009 and December 2013 were included in our study. Table 2 summarises the characteristics of all the patients. All the patients had no tumour history and did not receive any radiotherapy previously. The number of men was 215 (71.7%), whereas the number of women was 85 (28.3%). Median age was 47 years (range, 11-81 years). All patients were underwent the following pretreatment evaluations: recording of completed patient history, haematological and biochemical pro les, physical examination, exible beroptic endoscopic examination, MRI of the nasopharynx and neck, abdominal sonography, chest radiography or CT, and whole-body bone scan. All the patients were reclassi ed according to the 7th and 8th editions of the UICC/AJCC staging system by 2 clinicians. A third clinician was consulted to reach a consensus in case of disagreement.

RT
All patients completed radical SIB-IMRT at the Tumor Center of West China hospital according to the guidelines for RT based on reduced volume IMRT [14]. RT is carried out in accordance with the guidelines of NCCN radiotherapy for NPC. The primary nasopharynx gross tumour volume (GTVnx) and metastatic cervical lymph nodes (GTVnd) included all gross diseases observed in the MRI scan (GTVnx and GTVnd are usually referred to as 2 MRI scans of patients obtained before and after induction chemotherapy). CTV-1 was de ned as a high-risk region that included the primary nasopharynx tumour volume with a 5-10 mm margin and the entire nasopharynx. CTV-2 was de ned as potentially involved region that included the skull base, pterygopalatine fossa, pterygoid processes, anterior third of the clivus and cervical vertebra, inferior sphenoid sinus and cavernous sinus, nasopharyngeal cavity (including the posterior region of the nasal cavity), maxillary sinus (

Follow-up
The patients were followed up every 3 month during the rst 3 years and every 6 months thereafter or until death. Each follow-up consisted of physical examination, basic serum chemistry, exible beroptic endoscopy, MRI of the nasopharynx and neck, chest radiography or CT, abdominal sonography, and a whole-body bone scan. MRI was performed after SIB-IMRT was completed and to detect locoregional or distant relapse.

Statistical analysis
The endpoints of this study were overall survival (OS; time to death due to any cause), disease-free survival (DFS; time to treatment failure or death from any cause), distant failure-free survival (DMFS; time to distant metastasis), and locoregional recurrence-free survival (LRFS; time to locoregional persistence or recurrence). All the events were estimated from historical diagnosis. OS, DFS, DMFS, and LRRFS were calculated using the Kaplan-Meier method [15], and survival curves were estimated using log-rank tests [15]. Statistical Package for the Social Sciences, version 23.0, was used for statistical analysis.

T category classi cation
Of the 89 patients with stage T4 NPC according to the 7th edition, 61 were downgraded to T3 considering the 8th edition as the cancer had reached to the medial or LP muscles. A total of 82 patients with stage T1 NPC according to the 7th edition were reclassi ed as stage T2 considering the 8th edition on the basis of prevertebral muscle extension (Table 3). Table 4 presents the 5-year survival rates for different end points of T categories in the 7th and 8th editions. Figure 1A and 1B show the OS curves for the T categories in the 7th and 8th editions. There were signi cant differences between T4 and T1, T4 and T2, and T4 and T3 categories in the 7th and 8th editions (P < 0.001 for T4 and T1, T4 and T2; P = 0.007 for T4 and T3 in the 7th edition; P = 0.004 for T4 and T3 in the 8th edition). The OS rates between T3 and T2, and T3 and T1 categories in the 7th edition were not signi cantly different (P = 0.106 and P = 0.066, respectively), whereas these categories in the 8th edition were considerably different (P = 0.008 and P = 0.004, respectively). Figure 1C and 1D show the LRRFS rates of T categories in both the staging systems. In the 8th edition, the difference between T2 and T4 was statistically signi cant, while that in the 7th edition was not (P = 0.07 and P = 0.011, respectively). Therefore, the 8th edition had improved the prognosis value of NPC compared with the 7th edition. Figure 1E and 1F show the DFS curves; except for the difference between T1 and T2 categories, the difference between all other combinations of classi cations were statistically signi cant.

N category classi cation
In the 8th edition, the supraclavicular fossa (SCF) was replaced with the lower neck (below the caudal border of the cricoid cartilage), which led to the upstaging of 18 patients from N2 to N3 (Table 3). Table 4 presents the 5-year survival rates for different end points of N categories in the 7th and 8th editions. Figure 2 shows the OS, DFS, and DMFS survival curves for each staging system. The OS and DFS in the 7th edition system for N2 and N3a stages were not signi cantly different (P = 0.472 and P = 0.954, respectively, Figure 2A, 2C), whereas the OS and DFS for N2 and N3 using the 8th edition were statistically different (P = 0.046 and P = 0.043, respectively, Figure 2B, 2D). Thus, the 8th edition had a superior prognosis value compared with the 7th edition with respect to N category classi cation.
Additionally, there was no signi cant difference between classi cations N3a and N3b in the 7th edition system (P = 0.785 for OS, P = 0.241 for DFS, and P = 0.910 for DMFS; Figure 2A, 2C, 2E). The DFS curves for N3a and N3b even overlapped in the 7th edition ( Figure 2C). Therefore, merging N3a and N3b stages in the 8th edition was reasonable.

Stage group classi cation
Considering the 8th edition, 40 patients with stage IVA disease were downgraded to stage III (Table 3). No deaths were reported for stage I patients. Table 4a presents the 5-year survival rates for different end points of clinical stages in the 7th and 8th editions. Figure 3 shows the OS and DFS survival curves for each staging system. In both the staging systems, signi cant differences in OS and DFS were observed for clinical stages (P < 0.05) except for stages IVA and IVB, and I and II (P = 0.893 for OS and P = 0.711 for DFS; P = 0.549).

Discussion
Based on our study ndings, our data show that the 8th edition has a superior prognostic value for patients with NPC than the 7th edition.
In the treatment of NPC, IMRT has become the optimal radiation technique because of its clear advantage in target dose uniformity and better protection of adjacent organs at risk compared with 2dimensional radiotherapy (2D-RT) or 3-dimensional conformal radiotherapy (3D-CRT). It can be administered in 2 ways, sequential technology (SEQ-IMRT) [16,17] or SIB-IMRT [18]. Compared with SEQ-IMRT, SIB-IMRT simply uses a single radiation plan in the entire course of treatment, allowing the simultaneous delivery of different dose levels to different target volumes that reduces the treatment duration and enhances biologically equivalent dose (BED) [12,19]. The TNM staging system is crucial for predicting prognosis, guiding treatment decisions for different risk groups, assessing treatment e cacy, and evaluating clinical outcomes between different centres. Therefore, the TNM staging system should be updated based on the development of radiation technology. The 7th staging system was based on the information data form the 2D-RT era, and several trials have been conducted to determine its value considering the advent of IMRT [20][21][22]. Zong [20] et al analysed the data of 1241 NPC patients treated with IMRT and revealed that the differences in local relapse-free survival (LRRFS) between T1 and T2, and between T2 and T3 were not signi cantly different (P = 0.055 and 0.605, respectively). Additionally, they reported that the hazard ratios for OS and diseasespeci c survival between T1 and T2 were not statistically signi cant. The study considered that the TNM staging system should downgrade stage T2 patients to T1 patients. In a study performed by Chen et al [21] on 181 NPC patients with N0 stage, the authors reported that the difference in OS, LRFS, and PFS between T1 and T2, and between T3 and T4 was not statistically signi cant. In this study, we also con rmed that there were no differences in OS, DFS, and DMFS between T1 and T2 (P = 0.987, 0.984, and 0.191). Fortunately, the 8th staging system was revised after the introduction of IMRT as a treatment option and several previous studies [11,[23][24][25] have reported its superiority over the 7th edition staging system. Our data con rmed that the 8th edition had better prognostic performance than the 7 th edition.
For T categories, our data found that the T-classi cation in the 8th edition showed better separation between T3 and T2, and T3 and T1 compared to OS and LRRFS, while there were no signi cant differences in the T-classi cation in the 7 th edition. A retrospective study performed by Pan et al [11] on 1609 patients staged based on MRI ndings and treated with IMRT at 2 major centres in Hong Kong and Mainland China (median follow-up of 5 years) found that there were statistically signi cant differences among OS between T3 and T2 (P = 0.009). Additionally, OuYang et al [24] retrospectively studied 899 patients with NPC (from Hong Kong, Guangzhou, and Guangxi) who were staged based on MRI ndings and received IMRT; this study compared the 7th and 8th staging systems and reported that the 8th edition had better differentiation of OS between T3 and T2 (P = 0.003). All these data con rmed that it was reasonable to downstage MP and LP from T4 in the 7th edition to T2 in the 8th edition. This change has increased the survival difference values between T3 and T2, and also resulted in improved classi cation of patients with NPC.
In terms of N categories, replacing SCF with the lower neck region to differentiate N1-2 and N3 is the main revision in the 8th edition. Ng et al [10] rst explored the possibility of replacing the SCF by levels IV and Vb as a demarcating criterion for the N3 category, and found this method potentially useful. A few studies debated that the de nition of SCF involvement is primarily based on clinical examination and de ning SCF using clinical landmarks is di cult [6][7][8]. However, the lower neck, as an anatomical landmark, can be reliably de ned on the basis of both physical examination and cross-sectional images, thereby making it more convenient in clinical practice.
Several studies [23][24][25] have reported that the new staging system is useful in predicting outcomes with regard to N categories. In a study performed by Tang [23] that included 1790 NPC patients, the survival curves between different groups were accurately differentiated considering the 8th staging system. Another respective study also con rmed that the T-classi cation according to the 8th staging system showed better differentiation compared with that performed using the 7th edition [25]. Similarly, our results showed a clear difference between N2 and N3 among OS and DFS according to the new staging system. Moreover, we found no differences between N3a and N3b among OS, DFS, and DMFS considering the 7th staging system, indicating that this subgroup was unnecessary.
In terms of clinical stage, the 8th edition has upgraded IVC to IVB, and merged IVA and IVB from the 7th edition into IVA. Our data showed that the segregation of IVA and IVB in terms of survival was inaccurate in the 7th staging system as IVA and IVB share similar 5-year OS and DFS rates.
Our study included patients with NPC from a centre between year 2009 and 2013 with a relative long follow-up time. However, because of the radiation technique, only 300 patients in our study underwent SIB-IMRT. This small number of patients may result in low end-point events that may weaken the power to convince the differences between both the staging systems. Another limitation was the nature of the study (retrospective), and hence, prospective multicenter studies are required to be performed to con rm the results of our study.

Conclusion
The 8th edition of the UICC/AJCC staging system has a higher prognostic value and better classi cation compared with the 7th edition considering SIB-IMRT as the latest treatment option. The comparison results of both the staging systems should be further con rmed in multicenter, prospective trials with large sample sizes.

Declarations
Ethics approval and consent to participate The work described has been carried out in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) for experiments involving humans. The project was approved by the West China hospital ethics committee. All the partecipants signed an informed consent to partecipate in this study and a consent for the publication of the anonymized data has also been obtained.

Consent for publication
Not applicable.