We retrospectively analyzed the data on patients with pathological and newly diagnosed NPC in the First Affiliated Hospital of Nanchang University from January 2015 to March 2018. The inclusion criteria were as follows (1): the diagnoses of stage III-IVA according to the 8th edition of the Union for International Cancer Control/American Joint Committee on Cancer (UICC/AJCC) staging system; (2) scores ≥ 80 in the Karnofsky scale and were between ages 18 and 70; (3) received at least one cycle of IC prior to IMRT; (4) completed the treatment plan with complete clinical information and follow-up data; (5) exhibited normal heart, lung, liver, and renal function, without other cancers or pregnancy.

Based on these criteria, a total of 386 LA-NPC patients were selected for the study. The entire patient enrollment procedure is summarized in Figure 1 .

For all patients, pathological diagnoses were obtained by direct fiberoptic nasopharyngoscopy. According to the 2003 edition of the World Health Organization (WHO), pathological classifications were based on the following criteria: type I includes keratinizing squamous cell carcinoma; type IIA is non-keratinizing differentiated carcinoma and type IIB is non-keratinizing undifferentiated carcinoma; and type III is basal cell-like squamous cell carcinoma. The clinical stage was evaluated according to the eighth edition of AJCC/UICC, and the workup included the magnetic resonance imaging (MRI) of the head and neck, chest radiography (DR) or chest CT, abdominal ultrasound (B-US) or CT, whole-body bone scan (ECTs), and/or whole-body PET-CT examination.

All patients received one fraction of IMRT (6 MV photons) daily for 5 days each week. The primary nasopharyngeal gross tumor volume (GTVnx) and the involved cervical lymph nodes were determined based on MRI/CT and/or PET-CT imaging, clinical evaluation, and endoscopic findings. The target volumes were delineated following reports 50 and 62 by the International Commission on Radiation Units and Measurements. The prescribed radiation dose is defined as follows: 68–72 Gy to the planning target volume (PTV) of the GTVnx (including the primary tumor and enlarged lymph nodes), 64–70 Gy to the PTV of the nodal gross tumor volume (GTVnd; including enlarged lymph nodes in the neck), 60–63 Gy to the PTV of CTV1 (high-risk regions), and 54–60 Gy to PTV of CTV2 (low-risk regions) in 30–33 fractions.

GTVnx and the neck metastatic lymph node (GTVnd) were delineated based on the pre-treatment MRI or CT of the head and neck. CTV1 was delineated to cover sites with a high-risk of microscopic spread (including GTVnx with a margin of 5–10 mm). CTV2 was delineated to cover the low-risk sites of microscopic extension (including CTV1 with a margin of 5–10 mm) and potentially involved regions and lymphatic regions. After trimming, CTV1 or CTV2 can only be expanded by 2-3 mm adjacent to the brainstem and spinal cord.

In this study, all patients received at least 1 cycle of IC before radiotherapy. The regimens of AC and IC were same and included the following: TP (paclitaxel 135 mg/m² on the first day and cisplatin/nedaplatin 25 mg/m²/day on the first 3 days), DP (docetaxel 75 mg/m² on the first day and cisplatin/nedaplatin 25 mg/m²/day on the first 3 days), and TPF (cisplatin/docetaxel 60 mg/m² on the first day and fluorouracil 600 mg/m²/day as a continuous 120 h infusion on days 1–5), which were administered every 3 weeks. Concurrent chemotherapy (CCT) regimens consisted of a single-drug platinum or fluorouracil weekly for a maximum of seven cycles.

Follow-up was measured from the day of initial pathological diagnosis to last examination or the date of any endpoint event. Patients were assessed every 3 months during the first 2 years, then every 6 months thereafter, or until death. The follow-up endpoints were the OS rate, locoregional relapse-free survival (LRFS) rate, DMFS rate, and PFS rate.

Toxic reactions included blood toxicity (leukopenia, neutropenia, anemia, thrombocytopenia) and adverse events listed under the National Cancer Institute Common Terminology Criteria for Adverse Events NCI-CTCAE ver. 4.0.

All patients received IC for 1-6 cycles before radiotherapy. Since the median number of IC cycles was 2 cycles, patients given <3 cycles of IC were referred to as the IC<3 group, while patients administered with ≥3 cycles of IC were referred to as the IC≥3 group.

Among the total cohort, 224 (58%) received adjuvant therapy (AC) sequentially after CCRT/IMRT. Since patients in the IC<3 group received AC significantly more frequently than those in the IC≥3 group (68.4% vs. 39.6%, P<0.001), we therefore subdivided the patients into four treatment groups, that is, IC<3+AC, IC<3+non-AC, IC≥3+AC, and IC≥3+non-AC.

Continuous variables were analyzed by the T-test, and categorical variables were analyzed by the chi-square test. The Kaplan–Meier survival function was used to compare survival rates, draw survival curves, and perform log-rank tests of the survival rate for each group. Single and multivariate Cox regression analyses were applied to analyze the relationship between different variables and patient prognosis, and a multivariate Cox proportional hazard model was used to estimate correlations between the research variables and the prognosis and to evaluate the hazard ratio (HR) and 95% confidence interval (CI). All analyses were performed using SPSS25.0 software (SPSS Inc, Chicago, IL, USA). The differences of P<0.05 were considered statistically significant.

A total of 525 consecutive LA-NPC patients were treated in the First Affiliated Hospital of Nanchang University from January 2015 to March 2018. According to inclusion criteria, 386 patients were enrolled in this study and subsequently divided into IC<3 (247 patients) and IC≥3 (139 patients) treatment groups based on the numbers of IC cycles they received.

The baseline clinical characteristics of the different groups are shown in Table 1 . Notably, the IC<3 group had significantly more male patients (p=0.041), received more concurrent chemotherapy (CCT) (p<0.001), and included more AC patients (p<0.001) compared to the IC≥3 group. In addition, more patients who received the TP regimen of IC (P<0.001) were found in the IC≥3 group. There were no significant differences found between groups in other clinical characteristics, such as age, clinical stage, toxic reaction, T category, or N category. Furthermore, treatment failure was not significantly different between the IC<3 and IC≥3 groups.

The median follow-up time in this study was 53 months (range: 2–74 months). The 3-year OS, local recurrence-free survival (LRFS), DMFS, and PFS rates of the 386 patients were 92.2%, 88.9%, 87.2%, and 80.4%, respectively ( Figure 2 ).

We also assessed the relationship between the number of IC cycles and survival rates and found that the IC≥3 group had better survival outcomes. Specifically, the 3-year OS rate of the IC≥3 group was significantly higher than that of the IC<3 group (95.7% vs. 90.3%, p=0.020; Figure 3A ). However, no significant differences were observed in 3-year LRFS (p=0.212, Figure 3B ), DMFS (p= 0.112, Figure 3C ), or PFS (p=0.961, Figure 3D ) between the two groups.

To evaluate whether the IC cycle number was predictive of survival outcomes, we performed the univariate and multivariate analyses of all 386 patients. As shown in Table 2 , the univariate analysis indicated that the T stage and clinical stage were prognostic factors for OS (p=0.026 and p=0.017, respectively), LRFS (p=0.015 and p=0.015, respectively), DMFS (p=0.019 and p=0.028, respectively), and PFS (p=0.011 and p=0.001, respectively). However, the results of the multivariate analysis revealed that the clinical stage was only an independent factor for PFS (95% CI 0.343~0.844, HR=0.538, p=0.007; Table 3 ).

In addition, univariate analysis showed that the IC cycle number significantly affected OS (p=0.024) but not LRFS (p=0.215), DMFS (p=0.117), or PFS (p=0.961) ( Table 2 ). Similarly, multivariate analysis showed that the number of IC cycles was also an independent prognostic factor for OS (95% CI 0.145~0.733, HR= 0.326, p=0.007) but not for LRFS, DMFS, or PFS ( Table 3 ).

Both the univariate and multivariate analyses of the potential effects of gender, age, N category, CCT, AC, and IC regimen on survival outcomes indicated that none of these factors were associated with the prognosis ( Tables 2 , 3 ). These findings suggested that only the IC cycle was an independent factor for OS and that the clinical stage was an independent factor for PFS.

To evaluate the role of AC in the survival outcomes of patients with locally advanced nasopharyngeal carcinoma, we further subdivided the patient cohort into four groups: IC<3+non-AC (20.2%), IC<3+AC (43.8%), IC≥3+non-AC (21.8%), and IC≥3+AC (14.2%, Figure 4 ). We found that patients in the IC<3+non-AC group had significantly lower OS rates compared to the other three groups (all p<0.05; Figure 5A ). There were no significant differences among the four groups in LRFS, DMFS, and PFS (all p>0.05; Figures 5B–D , respectively). In particular, patients in the IC≥3+non-AC group had the highest OS rate (HR=0.208, 95% CI 0.069~0.631, p=0.006) ( Table 4 ) according to the OS curve and multivariate analysis. In addition, we performed an analysis to account for different potential contributing factors and found that gender, T and clinical stages, and CCT did not affect survival outcomes ( Table 4 ).

The comparison of treatment toxicities between each group showed no significant differences in the incidence of grade 3-4 hematological toxicity (i.e., leukopenia, neutropenia, anemia, and thrombocytopenia) among the four treatment modes. However, sequential AC after IC with CCRT/IMRT can increase the risk of 3-4 leukopenia and neutropenia, although no significant differences were found (p=0.761 and p=0.402, respectively; Table 5 ).