Nasopharyngeal carcinoma (NPC) is a malignant epithelial tumor originating from the nasopharynx, with a unique and unbalanced endemic distribution and high incidence in East and Southeast Asia, especially in South China (Chua et al., 2016; Chen et al., 2019). Owing to the peculiar anatomic location of the nasopharynx, 70% of NPC patients present with an advanced stage at initial diagnosis, and locally advanced nasopharyngeal carcinoma (LANPC) has a high metastasis rate and relatively poor prognosis (You et al., 2017). In 2016, 52,000 new cases and 26,700 deaths occurred from NPC in China, caused by interactions between Epstein-Barr virus infection and genetic and environmental factors, such as alcohol consumption and smoking (Tang et al., 2016; Zheng et al., 2022). NPC is a distinct type of head and neck cancer regarding treatment response because of its sensitivity to radiotherapy and chemotherapy (Hu et al., 2017). Induction chemotherapy based on cisplatin, 5-florouracil, and docetaxel can improve the survival outcome of patients with LANPC; however, a limited long-term treatment regimen necessitates new strategies for further improvement (Liu et al., 2018).

Intensity-modulated radiotherapy (IMRT) is a recognized and effective method for the radical treatment of NPC, and the 10-year survival rate of early-stage patients (stage I-II) after radical radiotherapy can reach 90% (Bossi et al., 2021). The Intergroup 0,099 trial found that the survival endpoint of concurrent chemoradiotherapy (CCRT) and adjuvant chemotherapy was superior to that of radiotherapy alone, thus establishing CCRT as the standard treatment for LANPC (stages III-IVa) (Al-Sarraf et al., 1998). A meta-analysis of individual data from 19 randomized controlled trials showed that CCRT (with or without adjuvant chemotherapy) significantly improved overall survival (OS) compared with adjuvant chemotherapy or induction chemotherapy plus radiotherapy (Blanchard et al., 2015). Therefore, CCRT is the core treatment for LANPC. The National Comprehensive Cancer Network Guidelines (Caudell et al., 2022) recommend using a cisplatin 100 mg/m² once every 3 weeks or 40 mg/m² once a week in conjunction with radiotherapy (Caudell et al., 2022). Although patients with LANPC have a clear survival benefit from cisplatin-based CCRT, it is often associated with poor treatment adherence and quality of life due to the known adverse effects of cisplatin, including hematologic toxicity, gastrointestinal reactions, nephrotoxicity, ototoxicity, and neurotoxicity (Ben Ayed et al., 2020; Wu et al., 2021). Cisplatin-based concurrent chemoradiotherapy requires pre- and post-treatment fluid replacement during cisplatin administration to protect renal function, prolonging the duration of hospital stay (Tang et al., 2018). A prospective study showed no difference in efficacy between carboplatin-based CCRT and cisplatin-based regimens in LANPC treatment, with better carboplatin tolerability (Chitapanarux et al., 2007). However, these results were unconfirmed in two subsequent trials, which found no further improvement in OS in patients receiving carboplatin compared to those receiving radiotherapy alone or cisplatin-based CCRT, thus concluding that the carboplatin regimen is unsuitable for use with concurrent radiotherapy (Yau et al., 2006; Huang et al., 2012). Nedaplatin, a reported alternative to cisplatin-based concurrent chemoradiotherapy, had unignorable toxicity (Tang et al., 2018). Therefore, new chemotherapeutic agents with similar efficacy but less toxicity than platinum are urgently warranted.

S-1 is composed of tegafur (a precursor of 5-fluorouracil), gimeracil, and potassium oxonate with less toxicity as anti-tumor drugs; its ability to effectively inhibit DNA synthesis has also been shown to improve the efficacy of radiotherapy (Takiuchi and Ajani, 1998). The most well-known indication is the combination of cisplatin in the treatment of unresectable locally advanced or metastatic gastric cancer. S-1 monotherapy is effective and safe for gastric and lung cancer treatment (Kuyama et al., 2017). Moreover, S-1 concurrent radiotherapy improved survival in head and neck cancer by preventing distant metastasis and was equally effective and well tolerated in older patients with gastric, pancreatic, and non-small cell lung cancer (Ikeda et al., 2013; Kim et al., 2018; Sasaki et al., 2018). Recently, a prospective phase II study of LANPC showed that the 3-year progression-free survival (PFS), OS, locoregional recurrence-free survival (LRRFS), and distant metastasis-free survival (DMFS) rates after S-1 concurrent radiotherapy were 87.4%, 95.7%, 94.7%, and 91.5%, respectively, with mild toxicity (Lv et al., 2019). These results suggest that S-1 may be a new drug alternative to platinum. However, studies on S-1 concurrent radiotherapy in patients with LANPC are still limited. The clinical benefits and toxicity tolerance of the S-1 regimen has clinical benefits in compared with platinum-based concurrent radiotherapy require further exploration. Therefore, we performed a retrospective propensity score-matched analysis to compare the efficacy and safety of concurrent radiotherapy with S-1 and platinum in patients with LANPC.

For patients with LANPC, CCRT is considered the standard treatment option, and its survival benefit with or without adjuvant chemotherapy is far greater than that of radiotherapy alone (Zhang et al., 2005; Lee et al., 2010). Several randomized trials have demonstrated the role of cisplatin-based CCRT in treating LANPC (Al-Sarraf et al., 2023). However, the comparable efficacy of S-1 as an oral chemotherapeutic agent remains unclear. Currently, only one prospective single-arm phase II study has evaluated the intervention effects of S-1 combined with IMRT for LANPC (Lv et al., 2019). Previous evidence suggests that randomized controlled trials primarily examine pharmaceutical intervention efficacy in controlled settings, whereas estimating true clinical effectiveness requires support from real-world evidence and network meta-analyses (Ford and Norrie, 2016). In the present study, we used propensity score matching analysis to eliminate the influence of confounding factors while assessing the curative effect and toxicity of CCRT with S-1 and platinum every 3 weeks for LANPC in the IMRT. To our knowledge, this is the first study to directly compare and evaluate the efficacy of S-1 and platinum in LANPC using real-world data, thus providing valuable insights into these treatment strategies in clinical practice.

Our retrospective study results showed that S-1 concurrent radiotherapy had comparable 5-year OS, PFS, LRRFS, and DMFS rates to platinum in whole and matched cohorts. These expected results have several plausible explanations. First, modern radiation techniques and equipment can improve treatment outcomes by delivering high radiation doses directly to target tissues (Beaton et al., 2019). A prospective study reported that IMRT has better local tumor control (91.7% vs 84.0%, P = 0.049) and survival (79.6% vs 67.1%, P = 0.001) rates than two-dimensional conventional radiation therapy (2D-CRT), especially in patients with LANPC (Peng et al., 2012). In addition, induction chemotherapy plus concurrent chemoradiotherapy has also been shown to be associated with favorable 5-year OS (85.6% vs 77.7%, P = 0.042), LRRFS (90.7% vs. 83.8%, P = 0.044), and DMFS (88.0% vs 79.8%, P = 0.030) rates (Li et al., 2019). Although more patients in the platinum-based group received induction chemotherapy (63.0% vs 34.9%, P = 0.001) than in the entire cohort, the survival benefit remained unchanged, possibly because of the greater proportion of stage IV patients (51.9% vs 34.0%, P = 0.072) in the platinum-based group. In our study, 5-year OS (HR = 0.91, 95% CI = 0.29–2.87; P = 0.876), PFS (HR = 0.93, 95% CI = 0.36–2.39; P = 0.883), LRRFS (HR = 0.85, 95% CI = 0.26–2.81; P = 0.789), and DMFS (HR = 0.91, 95% CI = 0.31–2.73; P = 0.871) were higher in the S-1 group than in the platinum, although this difference was not statistically significant. S-1 may improve short-term efficacy (ORR: 98.0% vs. 88.0%, P = 0.117), mainly owing to its radiosensitization or synergistic effect with radiotherapy, affecting long-term survival. This result contradicts common sense, which we considered possible because 4 weeks after treatment was not long enough for some patients to see significant change, and tumor shrinkage could continue for the first few weeks after completion of treatment.

Based on these findings, the therapeutic effects of the S-1 regimen should not be underestimated. A prospective study of 105 patients exploring the antitumor effect of 2D-CRT combined with S-1 in LANPC found that 2-year OS and PFS rates were 86.2% and 81.3%, respectively (Wen et al., 2015). Another study mentioned above showed that S-1 plus IMRT resulted in a 3-year OS of 95.7% and PFS of 87.4% (Lv et al., 2019). The marked difference in PFS and OS between these studies and the present study was chiefly attributed to the short follow-up period. Significant differences appear gradually; patients with NPC require more than 3 years of follow-up to determine the treatment effect, especially after 5 years (Liu et al., 2018). Another difference that must be addressed is drug toxicity. In the matched cohort of this study, 20% of the patients in the S-1 group developed grade 3 mucositis, which was higher than that reported in two previous studies. The main reason may be the serum 5-fluorouracil concentration increase due to a high dose of S-1. Therefore, further studies are needed to determine the optimal use of S-1 to balance toxicity and efficacy in the precision radiotherapy era.

Regarding adverse effects, the S-1-based CCRT regimen was well tolerated in this study compared to the platinum regimen. Many patients experience platinum-related toxicities and have relatively poor compliance. Tang et al. showed that grade 3 leukopenia, neutropenia, and nephrotoxicity were 22%, 12%, and 11% in the cisplatin group and 23%, 12%, and 4% in the nedaplatin group, respectively, similar to the results of our study (Tang et al., 2018). Leukopenia (6%), neutropenia (4%), and nephrotoxicity (0%) were uncommon in the S-1 group. Therefore, nearly all patients completed the planned treatment. Although mucositis was the most common restrictive factor for CCRT with S-1, this adverse effect was easily managed using growth factors and antibiotics (Elad et al., 2020). In recent years, docetaxel combined with radiotherapy is reportedly safe and effective for LANPC; however, the incidence of grade 3 mucositis is much higher than that with S-1 (74.5% vs 20.0%), limiting the widespread use of docetaxel-based CCRT in patients with LANPC (Liao et al., 2019). It has been reported that the anti-epidermal growth factor receptor (EGFR) monoclonal antibodies cetuximab and nimotuzumab combined with IMRT can maximize the survival rate of patients with LANPC, with a 3-year OS rate of 91.7% and a severe hematologic toxicity rate of only 3.4%–4.7% (You et al., 2017). Another large-data retrospective study showed that patients with stage III-IVB NPC who received IMRT plus nimotuzumab had a significantly higher 5-year OS rate than those who received IMRT alone (89.9% vs. 78.3%, P = 0.006) (Zhi-Qiang et al., 2019). However, cetuximab or nimotuzumab are too expensive to treat for most NPC patients in China, which limits the widespread use of concurrent chemoradiotherapy based on anti-EGFR antibodies in patients with LANPC. Additionally, oral administration is more convenient than intravenous infusion. Patients can take S-1 orally anytime and anywhere to complete the current treatment, saving the time for traveling to the hospital during the day to receive platinum infusions, which also substantially improves the compliance of outpatients and day ward patients. Therefore, S-1 combined with IMRT is safe and convenient for patients with platinum-intolerant LANPC, especially older patients and those who experience severe platinum-related adverse reactions during induction chemotherapy.

Since this retrospective study was based on real-world data, it had certain limitations. First, different clinicians use different treatment strategies for the same disease, and a certain degree of understandable deviation exists. Second, part of the collection of adverse events relies entirely on patient narratives, which may lead to the misgrading of toxic reactions. Although we eliminated selection bias using propensity score matching, such as age, sex, KPS, pre-treatment EBV DNA level, T stage, N stage, and clinical stage, it is unknown whether residual confounding factors remain. Further prospective studies with larger sample sizes are warranted.

CCRT based on S-1 regimen showed a non-inferior survival benefit compared with platinum in LANPC, and the toxicity was tolerable. The spectrum of adverse reactions was different between the two drugs, with platinum having a higher incidence of hematological toxicity, gastrointestinal reactions and nephrotoxicity, while S-1 showed significant mucositis.