1.1 Patient inclusion
From February 2017 to September 2019, 140 cases of stage IIIa-IVb locally advanced NPC treated in our department were collected. Inclusion criteria: (1) NPC with confirmed pathological diagnosis; (2) complete clinical data; (3) stage IIIa-IVb according to 8th edition of American Joint Committee on Cancer (AJCC) staging classifications; (4) The Karnofsky Performance Scale (KPS) ≥70 points; (5) no contraindication to radiotherapy and chemotherapy; (6) with written form of informed consent. The study was approved by the Ethical committee of the First Affiliated Hospital of Bengbu Medical College. Patients were divided into HT group and IMRT group according to radiotherapy methods, with 70 cases in each group. No significant difference was found on general information between the two groups (all p >0.05) (Table 1).
1.2 CT scan for simulation and immobilization
The patient was placed in supine position and was fixed with a head-neck-shoulder thermoplastic mask before subjection to a Spiral computed tomography (CT) simulator. The scanning range was 5 cm length from head to subclavian, and the layer thickness is 3 mm.
1.3 Target area and organs at risk (OAR) delineation
Target area was delineated according to the International Commission on Radiation Units & Measurements (ICRU) Reports 50 & 62. The primary nasopharyngeal tumor was defined as GTVnx and metastatic lymph nodes as GTVnd. CTV1 was defined as the high-risk areas of microscopic extension encompassing the entire nasopharynx, parapharyngeal space, retropharyngeal lymph node area, skull base, clivus, sphenoid sinus, pterygopalatine fossa, nasal cavity and 5 mm anterior to the posterior wall of the maxillary sinus, part of the oropharyngeal and superior neck lymph drainage area, including gross tumor (GTVnx) and regional metastatic nodes (GTVnd). CTV2 was defined as the low-risk region of inferior neck lymph drainage area below CTV1. PTV1 and PTV2 was respectively generated by 3 mm expansion outside of CTV1 and CTV2, and at least 3-mm distant away from the skin for protection purpose. OARs included brain stem, spinal cord, lens, optic nerves, optic chiasm, parotid glands, temporomandibular joints (TMJs), inner ear, parotid gland, oral cavity, spinal cord and other non-specified tissues. A re-delineation of the target area and OAR were processed when the larger volume of primary or metastatic lymph nodes were observed.
The Pinnacle 9.8 system was used for the target area and OAR delineation. IMRT plans and HT plans were respectively formulated on Pinnacle 9.8 system and TomoTherapy Hi-Art Planning Station by the same medical physicist.
1.4 Planning techniques
1.4.1 IMRT plan
IMRT plan were optimized by Dose Volume Optimize Version 11.0.31, and treatment dose was initially calculated using the Eclipse TPS algorithm. After interactively optimization by inverse planning software, the final dose was calculated using the Anisotropic Analytical Algorithm version 11.031 dose algorithm with a grid size of 2.5 mm.
1.4.2 HT plan
HT plan were optimized by TomoTherapy Hi-Art Software Version 2.07 (Accurray, Madison, WI, USA), and 3 major parameters were set as follows: 2.5 cm for field width, 0.287 for a pitch and 2.1-2.6 for modulation factor. Dose calculation was performed by a collapsed cone convolution model with a grid size of 1.95 mm.
1.5 Dose prescription
The prescribed dose was 69.96Gy/33F for PGTVnx, 66-70Gy/33F for PGTVnd, 60Gy/33F for PTV1, and 54Gy/33F for PTV2 respectively. Moreover, an at least 98% of the target area coverage, and >110% and > 93%of the prescribed dose was respectively cover <20% and <3% of planning target volume (PTV). OAR and target doses were restrained according to the protocol of Radiation Therapy Oncology Group (RTOG)-0615 as follows: parotid gland V30 <50%, temporal lobe <60Gy, brain stem <54Gy, spinal cord <45Gy, optic nerve and optic chiasm <54Gy, TMJ <60Gy, lens <5Gy, inner ear <50Gy.
1.6 Chemotherapy
In both HT and IMRT treatment groups, cisplatin-based concurrent chemotherapy was given during radiotherapy. The dose of cisplatin was 80 mg / m2 every 3 weeks for 1 cycle. During chemotherapy, symptomatic treatment was applied for anti-vomiting and the gastrointestinal mucosa protection.
1.7 Dosimetric comparisons
1.7.1 Conformity index (CI)
CI: measurement of how conformed the dose distribution is along target volume of PTV1 and PTV2. Formula: CI = (TVPV × TVPV)/(VPTV × VTV), whereas the definitions of TVPV , TVPV,VPTV and VTV were as follows: VPTV: volume of PTV, VTV: treatment volume of body receiving 95% of prescribed dose and TVPV: volume of VPTV within VTV. The closer of the CI value to 1, the better conformity of the tumor target area.
1.7.2 Homogeneity Index (HI)
Formula: HI = D5 / D95. In the formula, D5 is the lowest dose that covers the 5% PTV, and D95 is the lowest dose that covers the 95% PTV. The closer of the value to 1, the better conformity of the tumor target area.
1.8 Clinical monitoring
During the treatment, all patients were required to undergo blood routine tests every week, liver and kidney function tests every 2-3 weeks, head and neck enhanced magnetic resonance imaging and upper abdomen ultrasound examination 3 months after radiotherapy to evaluate the patients' short-term efficacy.
1.9 Evaluation of radiotherapy-related adverse reactions
The most commonly used American Radiation Therapy Oncology Group (RTOG)/European Organization for Research on Treatment of Cancer (EORTC) standards [9]were used here for evaluation.
1.10 Short-term efficacy evaluation
According to the solid tumor evaluation standard Response evaluation criteria in solid tumors (RECIST) 1.1 [10], the complete response (CR), partial remission (PR), stable disease (SD), progress disease (PD), and objective response (ORR), and the incidence of adverse reactions were used for between group comparison. ORR = (number of CR cases + number of PR cases) / total number of cases × 100%; observation period of the short-term efficacy was from the first day of the radiotherapy to 3 month after the end of the radiotherapy.
1.11 Statistical methods
Data were processed using SPSS version 19.0 software. The measurement data was expressed as mean ± standard deviation, and the t-test was used for between-group comparisons for the data with normal distribution, whereas the count data was expressed as the rate and compared using χ2 test. The category data was compared using Wilcoxon rank-sum test. p <0.05 was taken as statistically significant.