The 5-year relative survival rate for all NPC patients has increased from approximately 50% to 75% over the past ten years [15, 18]. Advances in technology and the increasing availability of both diagnostic and therapeutic facilities have made significant contributions to this achievement. However, the introduction of new therapeutic interventions or technologies requires a reevaluation of the treatment outcomes and the patterns of failure in NPC patients with N0 disease after IMRT, in order to effectively guide treatment in the future.
RLN metastasis classification
According to the sixth edition of the AJCC staging system, Tang et al.  and Tham et al.  proved that N0 NPC patients with RLN alone have a similar risk of distant metastasis to patients with N1 disease. The proposal that patients with RLN should be classified with N1 disease formed the basis of the revisions to the N0/N1 classifications in the seventh edition of the AJCC staging system. In this study, all patients underwent magnetic resonance imaging (MRI) examinations and received IMRT as their primary treatment. In patients with negative cervical lymph node involvement, the presence of RLN metastasis was found to negatively affect both DMFS, indicating it is still appropriate to classify RLN metastasis as N1 disease, even in the era of improved NPC treatment and diagnosis.
IMRT is considered to be an advantageous radiation treatment technique compared to conventional two-dimensional radiation. IMRT can deliver high-dose irradiation to defined tumor targets, while minimizing the dose delivered to the surrounding normal organs and tissues, thereby improving the therapeutic ratio of radiation therapy . IMRT has improved the treatment outcome with respect to locoregional control, and in this study 5-year LRFS for all patients was 93.7%, similar to the local control rates of 90%-97% achieved in other studies using IMRT [8–11, 19]. In N0 disease, we achieved an excellent local control rate of 96.3%.
In NPC patients treated with two-dimensional conventional radiotherapy, prognostic factors such as parapharyngeal space extension, cranial nerve palsy and intracranial extension are validated prognostic factors for distant metastasis . However, in our cohort of N0 patients treated with IMRT, no prognostic factor could significantly determine DMFS, which may be due to the excellent local control offered by IMRT, which reduces the rate of metastasis and may weaken the significance of potential prognostic factors.
Radiation oncologists have previously treated all neck node levels in NPC comprehensively with definitive intent radiotherapy or prophylactic irradiation . Several investigators confirmed that elective level II, III, and VA irradiation is suitable for NPC without lymph node metastasis [4, 21]. Therefore, the volume of N0 disease that requires irradiation can vary among different doctors. As a result, only 43.6% N0 disease patients, who had no cervical lymph nodes or RLN metastasis based on MRI, received prophylactic irradiation to the upper neck lymph drainage region in this study. Only one failure was observed in a patient who developed Level II lymph node recurrence after receiving prophylactic irradiation of the retropharyngeal area and Level II, III, IV, and V lymph node regions. It should be emphasized that this patient developed a local recurrence, and so it is unclear whether the regions of lymph node recurrence received efferent lymphatics from the local recurrence
Using MRI, false negative diagnoses of metastases in the neck regions are relatively rare, with an occurrence rate of 0.5% . Furthermore, high-dose irradiation of the neck area is associated with side effects, such as soft tissue fibrosis, which may adversely affect the patients' quality of life . Our data confirmed that prophylactic irradiation which excludes the Level IV and the supraclavicular region does not increase the risk of regional recurrence in patients classified with N0 disease according to the seventh edition of the AJCC staging system using MRI again.
Survival of N0 disease according to AJCC T-classification
T1N0, T2N0 and T3N0 showed excellent survival rates in this study. Tham et al.  reported excellent locoregional control in NPC, with no significant difference in the LRFS rates of T1, T2 and T3 patients. We focused on patients with N0 disease, in which the prognostic factor of lymph nodes metastasis was excluded, and found that the current TNM staging T-classification is becoming less powerful for segregation of patients into risk groups for local control and overall survival. In practice, locoregional control in T1-T3 patients should no longer be a major problem due to the improved outcomes after IMRT treatment, accurate geographic coverage of tumors assisted by CT-guided radiation treatment planning, increased diagnostic accuracy provided by MRI and PET and the intensive use of chemotherapy. Therefore, due to improved survival, it will become increasingly important to pay attention to the long-term complications of NPC treatment.
In this study, T4 patients had the poorest prognosis, and T4 disease was the most challenging to treat. Ng et al. also reported that advanced T4 disease remains difficult to treat . Perhaps there are two main reasons. Firstly, intracranial extension is a significant independent predictor of LRFS in this study, as the adjacent brainstem and spinal cord cannot tolerate radiation doses at levels of 64 Gy in small volumes (1-10 mL) . When IMRT is administered proximal to critical neurological structures, such as the brain stem, the very steep dose gradient contributes to an inadequate tumor dose and increases the risk of marginal failures. The risk of marginal failures may be reduced by increasing the precision of treatment using an adequate and very tight dose coverage of target volumes. Image-guided radiotherapy (IGRT) can potentially improve the accuracy of radiotherapy treatment delivery, particularly within high-dose gradients . IMRT with IGRT could theoretically improve local control in T4N0 disease by reducing the incidence of marginal failures, and this hypothesis should be tested in the clinic. Secondly, another possible explanation is that larger tumors of T4 disease need higher radiation doses for tumor control due to the log-cell-kill principle of radiation treatment. This need could be solved by increasing the physical dose of radiation to an optimal level and/or by administering accelerated fractionation to the tumor. Whole-field simultaneous integrated-boost intensity-modulated radiotherapy with a dose > 70 Gy achieved excellent locoregional control, without an excess incidence of severe . However, it must be emphasized that this was a retrospective study, and our conclusions need to be confirmed by future prospective studies.