Patients and pretreatment evaluations
We performed a retrospective study of 331 patients that fulfilled the following criteria in our hospital between February 2012 and December 2016. Inclusion criteria were pathologically confirmed NPC, previously untreated, no evidence of distant metastases, receiving the whole course of radical IMRT, and full treatment plan data was available, including the isodose distribution and dose-volume histogram (DVH). Exclusion criteria included: prior or other current malignancy; prior RT, chemotherapy or surgery (except for diagnostic procedures) to the primary tumor or nodes.
Routine workup comprised complete medical history, physical and neurologic examinations, hematology and biochemistry profiles. MRI scans of the head and neck were performed to evaluate the extent of the locoregional disease. Chest and abdominal CT, as well as bone scintigraphy were performed to exclude distant metastasis. Medical records and imaging studies were analyzed retrospectively. All patients were restaged according to the 8th edition of the American Joint Committee on Cancer (AJCC) staging system for NPC.
MR scanning protocol
All MR images were acquired with the same 1.5 Tesla unit (Achieva, Philips, Best, Netherlands) using a head and neck coil. The plain scanning examination was performed with the following sequences：axial/coronal view: T1-weighted, short-term inversion recovery with T2-weighted fat suppression; sagittal view: T1- and T2-weighted imaging; slice thickness = 5mm; and spacing = 1mm. Enhanced scanning was performed as follows: axial, coronal, and sagittal fat-suppressed T1-weighted imaging after intravenous injection of 0.1mmol/kg Gd-DTPA (Bayer Pharma AG, Leverkusen, Germany).
All MR scans were evaluated by a multi-disciplinary treatment group of NPC, which included three radiation oncologists and two diagnostic radiologists; all disagreements were resolved by consensus. Radiologic criteria for the diagnosis of lymph node metastasis were based on the literature.
The diagnostic criteria for retropharyngeal lymph node (RPLN) and cervical lymph node (CLN) involvement included:[6-10] (1) any visible LN in the median RPLNs, a shortest axial dimension ≥ 5mm in the lateral RPLNs, ≥ 11mm for the jugulodigastric region and ≥ 10mm in other cervical regions, or a group of three LNs that were borderline in size; or (2) LNs of any size in the presence of necrosis or extracapsular spread (ES). The definition of central necrosis on MRI was a focal area of high signal intensity on T2-weighted images or a focal area of low signal intensity on T1-weighted images with or without a surrounding rim of enhancement. The criteria for ES were the presence of indistinct LN margins, irregular LN capsular enhancement, or infiltration into the adjacent fat or muscle. Lymph node locations were based on the International Consensus Guidelines for neck level delineation .
All patients received IMRT. Patients were immobilized in the supine position with a thermoplastic mask. Target volumes were defined by ICRU50 and 62 (International Commission on Radiation Units and Measurements) [11, 12]. The gross tumor volume (GTV) included the primary tumor (GTV-T) and metastatic lymph nodes (GTV-N). CTV-1 (defined as the high-risk clinical target volume) should include GTV plus 5- to 10-mm margin and also cover the entire nasopharynx, parapharyngeal space, and retropharyngeal nodal regions. CTV-2 (defined as the low-risk clinical target volume) included CTV-1 plus 5 mm margin and also encompassed the maxillary sinus (limited to 5 mm anterior to the posterior nasal aperture and maxillary mucosa), pterygopalatine fossa, posterior ethmoid sinus, parapharyngeal space, skull base, the anterior third of clivus and cervical vertebra, inferior sphenoid sinus, and cavernous sinus. CTV-N （the clinical target volume of the neck nodal regions）included bilateral coverage of levels Ⅱ, Ⅲ, Ⅳ, and Ⅴ, which were outlined according to the recommendation by the Radiation Therapy Oncology Group (RTOG)/European Organisation for Research and Treatment of Cancer (EORTC) delineation consensus for head and neck malignancies [3, 5]. The selection of level Ⅱb contouring methods was detailed below. Radiation was delivered using a simultaneous integrated boost-IMRT technique. The radiation dose prescribed evolved: a total dose of 66~70 Gy in fractions at 2.18 Gy/fraction to GTV-P and GTV-N, 60 Gy at 1.875 Gy/fraction to CTV-1, 50.4 Gy at 1.8 Gy/fraction to CTV-2, and 50.4~60 Gy to CTV-N in 28~32 fractions. The normal tissue constraints and plan evaluation were following the RTOG 0225 protocol . A total of 282 patients received 1~2cycles of cisplatin-based chemotherapy and 26 received 3~4 cycles. Whenever possible, salvage treatments (including boost irradiation, re-IMRT, surgery, and chemotherapy) were provided for patients who developed relapse or persistent disease.
Selection of level Ⅱb contouring methods
Two methods were used to contour the level Ⅱb. The first delineated the cranial border of level Ⅱb to the skull base, according to the guideline of RTOG 0615 (control group) . The other method contoured the cranial border of level Ⅱb to the lateral process of atlas for patients who meet the following criteria (modified group): ① the primary tumor has no expanded tendency to posterior and lateral direction on the ipsilateral side; ② no positive retropharyngeal LNs (LNRP) on the ipsilateral side; ③ on the ipsilateral side, the primary tumor do not invade the carotid sheath area, or invade the carotid sheath area but invasion <90 º(the degree of contact arch between the tumor and carotid artery is less than 90 º); ④ there is no positive lymph node in the level Ⅱ above the cranial edge of the second cervical vertebra (C2); ⑤there is no visible lymph node in level Ⅱfrom the skull base to the upper edge of C2.
Follow-up was measured from the first day of treatment to the day of last examination or death. Patients underwent weekly physical and hematology related examinations during the radiotherapy process. The follow-ups were conducted every 3 to 4 months during the first 2 years, then 6 to 12 months from year 3 to year 5 after radiotherapy. Follow-up examinations included a complete physical examination, blood tests, standard nasopharyngeal MRI scan, chest and upper abdominal enhanced CT scan (or chest radiography and abdominal ultrasound), bone scan, and fiber nasopharyngoscopy.
SPSS version 25.0 (SPSS Inc., Chicago, IL) and R version 3.0.2 (www.r-project.org) were used for data analysis. The Kaplan-Meier method was used for survival analysis, and the Log-rank test was applied to compare the difference. The 𝜒2 test was used for comparing categorical variables, and the independent t-test was used for comparing the means of continuous variables.
To balance the distribution of baseline characteristics, we used propensity score matching (PSM). PSM was performed by logistic regression analysis and included age, gender, AJCC staging, AJCC T classification, AJCC N classification and chemotherapy. Patients were matched 1:1 based on their propensity scores. All statistical tests were 2-tailed, with a level for significance set at <0.05.