This study aims to evaluate a personalized approach for treating patients with locoregionally advanced nasopharyngeal carcinoma (NPC). Currently, many high-risk patients receive additional treatment (adjuvant therapy) after standard chemoradiotherapy to prevent the cancer from returning. However, some patients may not actually need this extra treatment and could safely avoid its side effects.

This trial uses a novel risk prediction model called the Response-Adapted Individualized Risk Index (RAIRI). The RAIRI model evaluates how a patient's tumor and blood markers (such as Epstein-Barr Virus DNA) respond during and immediately after their initial chemoradiotherapy.

In this study, patients will be randomly assigned to one of two groups:

1. Standard Treatment Group: All patients will receive standard adjuvant therapy (either a PD-1 inhibitor or capecitabine) after completing their initial chemoradiotherapy.
2. RAIRI-Guided Group (Experimental): Patients will be evaluated using the RAIRI model after initial chemoradiotherapy. Only those identified as "high-risk" by the model will receive adjuvant therapy. Those identified as "low-risk" will be exempted from adjuvant therapy and will undergo regular observation.

The main goal of this study is to determine if using the RAIRI model to exempt low-risk patients from adjuvant therapy is as safe and effective as giving adjuvant therapy to everyone, measured by how long patients live without the disease returning or progressing.

Background and Rationale Locoregionally advanced nasopharyngeal carcinoma is commonly treated with definitive intensity-modulated radiotherapy combined with platinum-based chemotherapy, with or without induction therapy. Although modern radiotherapy and systemic treatment have substantially improved disease control, a considerable proportion of patients remain at risk of locoregional recurrence, distant metastasis, or death after standard treatment.

Adjuvant therapy, including metronomic capecitabine or PD-1 inhibitor monotherapy, has been shown to improve failure-free or event-free survival in selected high-risk patients. However, the absolute survival benefit is modest, and many patients may receive unnecessary adjuvant treatment, resulting in additional toxicity, prolonged treatment duration, increased financial burden, and impaired quality of life.

Current adjuvant treatment decisions are mainly based on pretreatment clinical risk factors, such as tumor stage, nodal stage, and baseline plasma Epstein-Barr virus DNA level. These static baseline factors do not fully capture interpatient heterogeneity in treatment response. Increasing evidence suggests that dynamic treatment-response markers, including longitudinal changes in plasma cell-free Epstein-Barr virus DNA and radiographic tumor regression during or after treatment, provide important prognostic information. A response-adapted strategy may therefore allow more precise identification of patients who can safely omit adjuvant treatment and those who may still require additional therapy.

RAIRI Model The response-adapted individualized risk index, or RAIRI, was developed to provide individualized dynamic risk prediction for patients with non-metastatic nasopharyngeal carcinoma. The model incorporates baseline clinical factors, longitudinal plasma cell-free Epstein-Barr virus DNA changes, and imaging-based tumor response assessments collected during treatment.

In prior validation cohorts, RAIRI showed favorable calibration and discrimination for predicting progression-free survival and overall survival. The model also demonstrated potential value for predicting the benefit of adjuvant chemotherapy. Patients categorized as RAIRI low-risk had favorable long-term outcomes and appeared to derive limited benefit from adjuvant capecitabine, whereas RAIRI high-risk patients were more likely to benefit from intensified adjuvant treatment.

Study Objective This trial is designed to prospectively evaluate whether RAIRI-guided adjuvant treatment decision-making can preserve survival outcomes while reducing unnecessary adjuvant therapy in patients with high-risk locoregionally advanced nasopharyngeal carcinoma.

The primary objective is to determine whether a RAIRI-guided adjuvant treatment strategy is non-inferior to standard adjuvant therapy with respect to 3-year failure-free survival. Additional objectives include evaluating the prognostic performance of RAIRI risk stratification and assessing whether observation in RAIRI low-risk patients can safely replace routine adjuvant therapy.

Study Design This is a prospective, multicenter, randomized, open-label, phase III non-inferiority trial. Approximately 651 eligible patients with high-risk locoregionally advanced, non-metastatic nasopharyngeal carcinoma will be enrolled.

Participants will be randomly assigned in a 1:1 ratio to either the experimental arm or the standard-treatment arm. Randomization will be stratified by study center, clinical stage, and use of neoadjuvant treatment.

Study Population Eligible patients will have histologically confirmed Epstein-Barr virus-encoded RNA-positive, non-metastatic, non-keratinizing nasopharyngeal carcinoma. Patients should have high-risk locoregionally advanced disease according to the protocol-defined staging and risk criteria, adequate organ function, Eastern Cooperative Oncology Group performance status of 0 or 1, available baseline plasma Epstein-Barr virus DNA testing, and measurable tumor lesions on baseline magnetic resonance imaging.

Patients with metastatic disease, insufficient baseline data, prior malignancies except selected low-risk cancers, severe uncontrolled comorbidities, active autoimmune disease requiring systemic immunosuppression, active infections requiring systemic treatment, pregnancy or lactation, or inability to comply with follow-up will be excluded.

Standard Treatment Before Adjuvant Therapy All participants will receive standard definitive chemoradiotherapy. Induction chemotherapy or chemoimmunotherapy may be administered before radiotherapy at the discretion of the treating physician based on disease characteristics and patient preference.

Radiotherapy will be delivered using intensity-modulated radiotherapy according to protocol-defined target-volume delineation, dose prescription, and organ-at-risk constraints. Concurrent chemotherapy will be mainly platinum-based. After completion of chemoradiotherapy, patients will undergo response assessment, including plasma cell-free Epstein-Barr virus DNA testing and nasopharyngeal/neck magnetic resonance imaging at protocol-specified time points.

Treatment Arms Experimental Arm: RAIRI-Guided Adjuvant Strategy In the experimental arm, adjuvant therapy will be individualized according to post-chemoradiotherapy RAIRI risk stratification.

Patients classified as RAIRI low-risk, defined as a predicted 5-year progression-free survival of 85% or higher, will not receive adjuvant therapy and will undergo observation and standard follow-up.

Patients classified as RAIRI high-risk, defined as a predicted 5-year progression-free survival of less than 85%, will receive adjuvant therapy with either a PD-1 inhibitor or metronomic capecitabine.

Standard-Treatment Arm In the standard-treatment arm, all patients will receive adjuvant therapy after chemoradiotherapy, regardless of RAIRI risk category. Adjuvant treatment will consist of either PD-1 inhibitor monotherapy or metronomic capecitabine, selected by the treating physician according to clinical circumstances and previous treatment exposure.

Patients receiving PD-1 inhibitor therapy will receive treatment every 3 weeks for up to 12 cycles unless disease progression, unacceptable toxicity, death, or patient refusal occurs. Patients receiving metronomic capecitabine will receive oral capecitabine twice daily for up to 1 year unless unacceptable toxicity, disease progression, withdrawal of consent, or other discontinuation criteria occur.

RAIRI Risk Assessment RAIRI risk stratification will be performed after completion of chemoradiotherapy using baseline clinical characteristics, longitudinal plasma cell-free Epstein-Barr virus DNA dynamics, and imaging-based tumor response assessments. The model will generate an individualized predicted 5-year progression-free survival probability.

A predicted 5-year progression-free survival of 85% or higher will define the RAIRI low-risk group. A predicted 5-year progression-free survival of less than 85% will define the RAIRI high-risk group.

To ensure standardized assessment, RAIRI scoring will be performed according to protocol-specified procedures. Clinical, imaging, and Epstein-Barr virus DNA data will be submitted for centralized review when required. Quality-control procedures will be implemented to ensure consistency across participating centers.

Primary Endpoint The primary endpoint is 3-year failure-free survival, defined as the time from randomization to locoregional recurrence, distant metastasis, or death from any cause, whichever occurs first.

Secondary Endpoints Secondary endpoints include 3-year overall survival, 3-year locoregional relapse-free survival, 3-year distant metastasis-free survival, complete response rate after chemoradiotherapy, the proportion of patients classified as RAIRI low-risk or high-risk after chemoradiotherapy, acute and late toxicities, adverse events, and changes in health-related quality of life.

Toxicities and adverse events will be graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0. Health-related quality of life will be evaluated using the EORTC QLQ-C30 and head-and-neck cancer-specific instruments at baseline and during follow-up.

Follow-up and Assessments Patients will be followed regularly after treatment. Follow-up assessments will include medical history, physical examination, laboratory tests including plasma Epstein-Barr virus DNA, nasopharyngoscopy, nasopharyngeal and neck imaging, chest imaging, abdominal imaging, and additional examinations when clinically indicated.

Suspected locoregional recurrence or distant metastasis will be confirmed by pathology whenever feasible. When biopsy is not possible, recurrence or metastasis may be diagnosed through multidisciplinary clinical and radiologic assessment.

Statistical Considerations The primary efficacy analysis will be performed in the intention-to-treat population. Failure-free survival and other time-to-event endpoints will be estimated using the Kaplan-Meier method. Treatment groups will be compared using stratified log-rank tests, and hazard ratios with corresponding confidence intervals will be estimated using Cox proportional hazards models.

A hierarchical testing strategy will be used. The first objective is to test whether the RAIRI-guided strategy is non-inferior to standard adjuvant treatment in the overall study population. If this objective is met, the study will further evaluate whether observation in RAIRI low-risk patients in the experimental arm is non-inferior to adjuvant therapy in RAIRI low-risk patients in the standard-treatment arm. The study will also independently assess whether RAIRI can reliably stratify patients into prognostically distinct low-risk and high-risk groups.

Safety analyses will include all patients who receive study treatment. Predefined subgroup analyses will evaluate the consistency of treatment effects across clinically relevant subgroups, including age, sex, Eastern Cooperative Oncology Group performance status, T category, N category, clinical stage, baseline Epstein-Barr virus DNA level, and use of neoadjuvant therapy.

Significance of the Study This study aims to establish a response-adapted precision treatment strategy for locoregionally advanced nasopharyngeal carcinoma. If successful, RAIRI-guided treatment may allow low-risk patients to avoid unnecessary adjuvant therapy while maintaining disease control, and may help identify high-risk patients who require additional treatment after chemoradiotherapy. This approach has the potential to improve individualized treatment decision-making, reduce treatment burden, and optimize the therapeutic value of adjuvant therapy in nasopharyngeal carcinoma.

Inclusion Criteria

1. Age 18-65 years, regardless of sex.
2. Histologically confirmed EBER-positive, non-metastatic, non-keratinizing nasopharyngeal carcinoma.
3. AJCC 9th edition stage II-III disease / AJCC 8th edition stage III-IVA disease, excluding T3-T4N0 and T3N1 disease; or baseline EBV DNA >4,000 copies/mL.
4. Eastern Cooperative Oncology Group performance status score of 0-1.
5. Availability of complete baseline pretreatment imaging data, including nasopharyngeal and neck MRI with functional MRI sequences, and at least one measurable tumor lesion.
6. Availability of pretreatment baseline plasma cfEBV-DNA measurement.
7. Patients must meet the following laboratory criteria: hemoglobin >120 g/L and white blood cell count ≥4 × 10⁹/L.
8. Platelet count ≥100 × 10⁹/L; liver and renal function parameters within 1.25 times the upper limit of normal; and no hearing impairment.
9. Ability to understand the study and provision of written informed consent.
10. Agreement to allow the use of personal data and biological samples, including blood and tissue samples, for research purposes.
11. Adequate function of major organs, except for abnormalities related to nasopharyngeal carcinoma.
12. Ability and willingness to comply with scheduled follow-up.

Exclusion Criteria

1. Absence of pretreatment cfEBV-DNA data or other essential baseline characteristic data.
2. AJCC 8th edition stage I-II or IVB disease / AJCC 9th edition stage I or IV disease, or T3-4N0 or T3N1 disease.
3. History of other malignancies, except stage I non-melanoma skin cancer or carcinoma in situ of the cervix.
4. Pregnant or lactating women, or women of childbearing potential who are not using contraception.
5. Current participation in another investigational drug trial.
6. Severe comorbidities, including myocardial infarction, severe arrhythmia, severe cerebrovascular disease, active ulcer disease, psychiatric illness, uncontrolled diabetes mellitus, active autoimmune disease, ongoing systemic immunosuppressive therapy, active infection requiring systemic treatment, history of human immunodeficiency virus infection, positive hepatitis B surface antigen, hepatitis B virus DNA >1 × 10³ copies/mL or >200 IU/mL, or positive hepatitis C virus antibody.
7. Inability to comply with regular follow-up.

De-identified individual participant data (IPD) that underlie the results reported in the published article will be shared.

Data will be available beginning 6 months following article publication.

Data will be shared with researchers whose proposed use of the data has been approved by an independent review committee identified for this purpose. Proposals should be directed to the corresponding author's email.