The impact of the cumulative dose of cisplatin during concurrent chemoradiotherapy on the clinical outcomes of patients with advanced-stage nasopharyngeal carcinoma in an era of intensity-modulated radiotherapy

This retrospective study was approved by the Clinical Research Ethics Committee of the Sun Yat-sen University Cancer Center, and all the participants provided written informed consent before treatment. This study is in compliance with the Helsinki Declaration. Patient records were anonymized and de-identified prior to analysis.

This study retrospectively analysed data from 491 consecutive patients with histologically confirmed NPC who were treated with concurrent chemoradiotherapy between December 2006 and December 2010 at Sun Yat-sen University Cancer Center. Inclusion criteria for the patients consisted of (1) histologically confirmed NPC by biopsy of the nasopharynx, (2) no distant metastasis, (3) no treatment prior to admission, (4) no other tumour types or serious illnesses, (5) an Eastern Cooperative Oncology Group (ECOG) performance score ≤2, (6) received radical IMRT during the course of treatment, and (7) received concurrent chemotherapy with cisplatin. In all patients, the staging workup included an MRI of the head and neck, a chest radiograph, a bone scintigraphy, and an ultrasonography of the abdominal region. Patients who received neoadjuvant chemotherapy were ineligible. All participants were restaged according to the Seventh Edition of the American Joint Committee on Cancer (AJCC) staging system. There were 42, 328, and 121 patients with stage II, III, and IVa-b disease, respectively. Patients were divided into three groups, i.e., low-dose (cumulative dose ≤100 mg/m²), medium-dose (100 mg/m² < cumulative dose ≤ 200 mg/m²), and high-dose (cumulative dose >200 mg/m²), according to previous studies [12, 14]. Table 1 shows the clinicopathological features in the study population of 491 patients.

The target volumes were delineated using a previously described institutional treatment protocol [19], in accordance with the International Commission on Radiation Units and Measurements reports 50 and 62. All target volumes were delineated slice-by-slice on the treatment planning computed tomography scan. The primary nasopharyngeal gross tumour volume (GTVnx) and the involved cervical lymph nodes were determined based on the imaging, clinical, and endoscopic findings. The enlarged retropharyngeal nodes were outlined, together with primary gross tumour volume (GTV), as the GTVnx on the IMRT plans. The first clinical tumour volume (CTV1) was defined as the area from 0.5 to 1.0 cm outside the GTV, a site that involves potential sites of local infiltration. Clinical target volume 2 (CTV2) was defined as the margin from 0.5 to 1.0 cm around CTV1 and the lymph node draining area (Levels II, III, and IV). For stage N1–3 patients, the lower neck area received conventional anterior cervical field radiation with a midline shield to 50 Gy in daily fractions of 2 Gy. For patients with stage N0 disease, RT was not delivered to the lower neck area. The prescribed dose was 66–70 Gy to the planning target volume (PTV) of GTVnx (PTVnx), 60 Gy to PTV1, 54 Gy to PTV2, and 60–66 Gy to PTV of the involved cervical lymph nodes in 30 to 33 fractions. In total, 30–33 fractions were administered at 1 fraction per day, 5 days/week. The IMRT plan was designed in accordance with previous studies conducted at the Sun Yat-sen University Cancer Center [20, 21].

Concurrent cisplatin chemotherapy was delivered to all of the patients. Chemotherapy was initiated on the same day as IMRT, and the cisplatin regimen included intravenous infusion (IV) of 80–100 mg/m² cisplatin every 3 weeks or of 30–40 mg/m² IV cisplatin weekly. Among all of the 491 patients, 14 (2.9 %) had a cumulative dose of cisplatin less than or equal to 100 mg/m², 378 (77.0 %) had a cumulative dose of cisplatin >100 and ≤200 mg/m², and 99 (20.2 %) had a cumulative dose of cisplatin more than 200 mg/m² during treatment.

The follow-up duration was calculated from the first day of treatment to either the day of death or the day of the last examination. Patients were examined at least every 3 months during the first 2 years; thereafter, follow-up examinations were performed every 6 months for 3 years or until death. The median follow-up period for the entire patient cohort was 49 months (range 1–88 months).

The program Statistical Package for Social Sciences version 17 (SPSS Inc., Chicago, IL, USA) was used for analysis. The Kruskal–Wallis test and Fisher’s exact test were used to analyse the relationship among the low- (cumulative dose ≤100 mg/m²), medium- (100 mg/m² < cumulative dose ≤ 200 mg/m²), and high- (cumulative dose > 200 mg/m²) dose groups among all of the NPC patients. Survival curves were estimated using the product limit method of Kaplan–Meier with the log-rank test. Univariate analysis was conducted using the log-rank test, and multivariate analyses were calculated using the Cox proportional hazards regression model. The potentially important prognostic factors considered in the modelling process included the following: patient gender (1. female, 2. male), age (1. <45, 2. ≥45), T stage (1. T1, 2. T2, 3. T3, 4. T4), N stage (1. N0, 2. N1, 3. N2, 4. N3), Epstein–Barr virus deoxyribonucleic acid (EBV DNA) (1. <4000, 2. ≥4000), and cumulative dose of cisplatin (1. low-, 2. medium-, 3. high-dose group). The entire patient cohort was divided into high- and low-risk patients by pre-treatment with EBV DNA using a cut-off value of 4000 copies/ml, according to previous studies, which led to a distinct risk stratification [22, 23]. The following end-points (time to the first defining event) were assessed: overall survival (OS), disease-free survival (DFS, distant metastasis-free survival (DMFS), and locoregional relapse-free survival (LRFS). The OS was defined as the time from diagnosis of NPC to death from any cause or until the date of the last follow-up. DFS was defined as the time from the diagnosis of NPC to events that included death or disease progression at local, regional, or distant sites or until the date of the last follow-up. LRFS was defined as the time from the diagnosis of NPC to the absence of a primary site or neck lymph node relapse or until the date of the last follow-up. DMFS was defined as the time from the date of treatment to the date of the first observation of a distant metastases or until the date of the last follow-up. The primary endpoint was OS, and secondary endpoints were DFS, DMFS, and LRFS.

Multivariate analyses using the Cox proportional hazards regression model were utilised to test for independent significance. All P values were two-tailed; P ≤ 0.05 was considered statistically significant. Our report adheres to STROBE guidelines (http://​www.​strobe-statement.​org/​) for reporting observational research (Additional file 1).

NPC patients received low- (≤100 mg/m²), medium- (101–200 mg/m²), or high-doses (>200 mg/m²) of cumulative cisplatin. The clinical characteristics and treatment factors for the three groups (≤100 mg/m², 101–200 mg/m², >200 mg/m²) were well balanced. The 5-year OS rates of the low-, medium-, and high-dose groups were 64.1 %, 91.1 %, and 89.4 %, respectively (P = 0.002; Fig. 1). Multivariate analysis using the Cox proportional hazards regression model demonstrated that the cumulative dose of cisplatin was significantly associated with OS (Table 2), and the N stage was an independent prognostic factor for OS. Patients who were in the medium- and high-dose groups had lower odds of death than did the patients in the low-dose group, with odds ratios of 0.135 (95 % confidence intervals (CI) 0.045–0.405, P < 0.001) and 0.225 (95 % CI 0.069–0.734, P = 0.013), respectively. In addition, a significant difference in OS was observed on the N stage and EBV DNA. Patients with a N3 stage and EBV DNA ≥4000 copies/ml had an increased odd of death, with odds ratios of 7.404 (95 % CI 1.494–36.684, P = 0.014) and 4.953 (95 % CI 2.200–11.153, P < 0.001), respectively.

The 5-year DMFS rates of the low-, medium-, and high-dose groups were 69.2 %, 88.7 %, and 88.6 %, respectively; this difference was statistically significant (P = 0.027; Fig. 2). Multivariate analysis using the Cox proportional hazards model demonstrated that EBV DNA was the only independent prognostic factor associated with DMFS with an OR of 3.669 (95 % CI 2.058–6.540, P < 0.001; Table 2). The cumulative dose of cisplatin was not significantly associated with DFS or LRFS.

There were 300 (61.1 %) and 191 (38.9 %) patients with pre-treatment EBV DNA levels less than 4000 copies/ml or EBV DNA ≥4000 copies/ml, respectively. In the low-risk group, 8 (2.7 %) patients received less than 100 mg/m², 323 (77.3 %) patients received 101–200 mg/m², and 60 (20.0 %) patients received more than 200 mg/m². In the subgroup analysis for low-risk group patients (EBV DNA <4000 copies/ml), the cumulative dose of cisplatin was significantly associated with a lower OS based on univariate analysis (P < 0.001; Fig. 3). After multivariate analysis using the Cox proportional hazards regression model, the cumulative dose of cisplatin was significantly associated with OS (P = 0.009). The medium-dose group had reduced odds of death compared with the low-dose group, with an odds ratio of 0.062 (95 % CI 0.001–0.347, P = 0.002). The cumulative dose of cisplatin was significantly associated with DMFS (P = 0.034; Fig. 4). However, the cumulative dose of cisplatin was not significantly associated with DMFS by multivariate analysis. Moreover, the cumulative dose of cisplatin was not associated with OS or DMFS among the high-risk (EBV DNA ≥4000 copies/ml) patients by multivariate Cox regression analysis.