Adaptive radiotherapy based on nodal response to induction chemotherapy in nasopharyngeal carcinoma: a retrospective dose de-escalation analysis

Introduction

Nasopharyngeal carcinoma (NPC) is endemic in southern China and Southeast Asia. Due to its anatomic location and intrinsic radiosensitivity, radiotherapy remains the primary treatment modality for this disease (1,2). For patients with locoregionally advanced nasopharyngeal carcinoma (LA-NPC), the standard treatment approach is induction chemotherapy (IC) followed by concurrent chemoradiotherapy. IC has been shown to reduce tumor burden and expand the spatial separation between target volumes and adjacent organs at risk (OARs) (3,4).

However, the growing adoption of IC raises questions regarding appropriate post-IC radiotherapy strategies, particularly target volume delineation and dose prescription. Despite significant tumor regression following IC, pre-IC imaging is still commonly used for radiotherapy planning, leading to potential overtreatment and unnecessary radiation-related toxicities, including mucositis, xerostomia, and ototoxicity (5). More than 70% of patients experience a >50% tumor volume reduction after IC (6), with some achieving radiological complete response (CR). Emerging data from clinical trials and real-world studies suggest that selected patients may benefit from de-escalated radiotherapy.

A phase III randomized trial showed that dose-reduced radiotherapy based on post-IC volume delineation achieved comparable 3-year local control rates (91.5% vs. 91.2%) with significantly fewer grade ≥3 toxicities (7). Similarly, in a phase II single-arm study which delivered 60 Gy to post-IC regressed areas, 68 Gy to residual primary lesions, and 62–66 Gy to lymph nodes, a 10-year local control rate of 89% was achieved with limited late toxicities (8). Specifically, post-IC tumor regression and Epstein-Barr virus (EBV) DNA clearance have been independently associated with favorable prognosis, reflecting tumor sensitivity and offering a basis for individualized radiotherapy (9,10). In a phase II trial involving stage III patients with baseline EBV DNA <4,000 copies/mL and favorable IC response (complete or partial), a de-escalated dose of 60 Gy resulted in a 2-year progression-free survival (PFS) of 94.8%, with no grade ≥3 late toxicity (11). Currently, an ongoing phase III trial (NCT04448522) is evaluating a regimen of 63.6 Gy in 30 fractions (2.12 Gy per fraction) following three cycles of IC and two cycles of concurrent chemoradiotherapy.

However, most de-escalation strategies have primarily focused on the primary nasopharyngeal tumor, whereas nodal disease has received comparatively little attention. Our study addressed this gap by evaluating the feasibility and outcomes of nodal dose de-escalation in LA-NPC patients who achieved complete nodal remission post-IC and exhibited no high-risk features in the involved cervical levels. We assessed long-term oncologic outcomes, EBV DNA dynamics, dosimetric profiles, and late toxicities to determine the safety and efficacy of this response-adapted strategy. We present this article in accordance with the STROBE reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0320/rc).

Methods

Patient selection

Between June 2018 and December 2020, the medical records of 316 patients diagnosed with NPC at the Department of Radiation Oncology were retrospectively reviewed. The inclusion criteria were as follows: (I) histologically confirmed nonkeratinizing undifferentiated carcinoma of the nasopharynx; (II) stage II–IVA disease according to the 2017 American Joint Committee on Cancer (AJCC) staging system; (III) treatment with IC followed by intensity-modulated radiotherapy (IMRT); (IV) no prior treatment; and (V) complete clinical records with regular follow-up. This retrospective study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments, and was approved by the Ethics Committee of Jiangsu Cancer Hospital (No. KB-2020-055). Written informed consent was provided by all participants.

Chemotherapy

All patients received 2–3 cycles of IC with either TP (80 mg/m² docetaxel and 80 mg/m² cisplatin on day 1) or TPF (60 mg/m² docetaxel and 60 mg/m² cisplatin on day 1, combined with fluorouracil at 600 mg/m²/day on days 1–5). Cycles were repeated every 3 weeks. Concurrent chemoradiotherapy was administered with weekly cisplatin at 40 mg/m². Some patients also received adjuvant chemotherapy using the same regimen as IC. Adjuvant chemotherapy was administered based on physician discretion and institutional guidelines, typically reserved for high-risk patients, such as those with bulky T4 primary tumors, advanced N3 disease, or persistently detectable EBV DNA post-radiotherapy.

Radiotherapy planning and delivery

Cervical lymph node metastases were identified according to established imaging criteria (12), including nodal size on contrast-enhanced magnetic resonance imaging/computed tomography (MRI/CT) or high fluorodeoxyglucose (FDG) uptake on positron emission tomography-computed tomography (PET-CT). The gross tumor volume (GTV1) included the primary nasopharyngeal tumor and retropharyngeal lymph nodes; residual (non-CR) nodes were delineated as GTV2. The clinical target volumes (CTVs) were delineated as follows: CTV1 comprised regions surrounding the primary tumor and high-risk nodal drainage areas, and CTV2 included low-risk nodal drainage areas. For nodes evaluated under the de-escalation protocol, the nodal gross tumor volume (GTVnd) was used for dose assignment (see Selection criteria for nodal dose de-escalation).

Radiotherapy was delivered using IMRT. Planning target volumes (PTVs) were generated by adding a 3 mm margin to both GTV and CTV. The prescribed doses were as follows: 66–75 Gy in 32–34 fractions to the PTV of GTV1; 66–70 Gy in 32–34 fractions to the PTV of GTV2; 60 or 64 Gy in 32 fractions to the PTV of GTVnd. The PTVs of CTV1 and CTV2 were prescribed 60 Gy in 32 fractions and 50.4 Gy in 28 fractions, respectively. GTV1 was not subjected to dose de-escalation and consistently received the standard prescribed dose of 66–75 Gy.

All patients received radiotherapy once daily, 5 days per week. Dose constraints for OARs adhered to Radiation Therapy Oncology Group (RTOG) guidelines. Replanning contrast-enhanced CT was performed before treatment and after the 5th, 15th, and 25th fractions to ensure target accuracy and the safety of dose de-escalation.

Selection criteria for nodal dose de-escalation

All pre- and post-IC MRI/CT scans were independently reviewed by two senior head and neck radiologists. Any discrepancies regarding the presence of high-risk features, such as extranodal extension (ENE) or necrosis, were resolved by consensus during our institutional multidisciplinary tumor board meetings to ensure standardized assessment.

Dose de-escalation was applied only to nodes achieving post-IC CR on MRI (short-axis <10 mm). Radiological CR was assessed on post-IC imaging, whereas the evaluation of high-risk exclusion features was based on pre-IC imaging. Nodes were excluded from de-escalation if any pre-IC high-risk features were present, including: (I) necrosis or cystic change; (II) ENE; (III) diameter ≥3 cm; and (IV) multiple nodes in the same drainage level. For eligible CR nodes, dose assignment was stratified by post-IC short-axis: nodes <5 mm were not delineated as GTVnd and were directly incorporated into CTV2; nodes measuring 5–8 mm and 8–10 mm were prescribed 60 and 64 Gy (32 fractions), respectively. All corresponding nodal drainage levels were delineated as CTV2 and uniformly prescribed 50.4 Gy.

For illustration, Figure 1 contrasts conventional IG-2018 (13) contouring (A-pre/B-pre) with the de-escalation contours used in this study (A-post/B-post, based on post-IC CT short-axis rules).

Figure 1 Comparison of conventional and de-escalated target delineation after induction chemotherapy. Conventional IG-2018 contours (A-pre/B-pre): nodes prescribed 66 Gy (blue), drainage levels prescribed 60 Gy (red). A-post/B-post show the study contours based on post-IC CT short-axis rules. (A) Within the left cervical level II region, a node with a post-IC short-axis diameter of 8–10 mm was delineated as GTVnd and prescribed 64 Gy (purple). The corresponding drainage level was included in CTV2 (orange, 50.4 Gy). (B) Within the left cervical level III region, a node with a post-IC CT short-axis <5 mm was incorporated directly into CTV2 (orange, 50.4 Gy) without GTVnd. CTV1 comprised regions surrounding the primary tumor and high-risk nodal drainage areas, and CTV2 included low-risk nodal drainage areas. CT, computed tomography; CTV, clinical target volume; GTVnd, nodal gross tumor volume; IC, induction chemotherapy.

Dosimetric evaluation

To evaluate the dosimetric impact of nodal dose de-escalation, 15 patients were selected from the 60-patient de-escalation cohort using a computer-generated random number sequence in SPSS. Baseline characteristics and target volumes of this subset were confirmed to have no significant differences compared to the rest of the cohort, ensuring high representativeness and minimizing selection bias. Statistical power analysis indicated that 15 paired plans were sufficient to achieve >90% power (α=0.05) in detecting significant differences for paired continuous variables. For each patient, two IMRT plans were generated using the Eclipse treatment planning system (v15.6, Varian Medical Systems, Palo Alto, CA, USA). All plans were created by a senior physicist with over 10 years of experience in head and neck radiotherapy.

In the conventional plan, 66 and 60 Gy were prescribed to GTVnd and CTV, respectively. In the de-escalated plan, the doses were reduced to 60 and 50.4 Gy. Both plans followed identical OAR optimization constraints and aimed for maximal conformity to eliminate planner-induced bias. Dose-volume histograms (DVHs) were generated for both plans. Dosimetric analysis of salivary glands focused on the mean dose (Dmean) and the volume percentage of an OAR receiving at least a specific dose (Vx).

Assessment and follow-up

Follow-up began on the first day of treatment and continued until the last visit or death. After 2–3 cycles of IC, nodal response was assessed according to the Response Evaluation Criteria in Solid Tumors (RECIST) (14) and categorized as CR, partial response (PR), stable disease (SD), or progressive disease (PD). CR was defined as the disappearance of nodal lesions with a short-axis diameter <10 mm on imaging.

All patients were evaluated weekly during treatment. Follow-up assessments were conducted every 3 months for the first 3 years, every 6 months for the next 3 years, and annually thereafter until death. Each visit included physical examination, nasopharyngoscopy, blood tests, contrast-enhanced MRI of the nasopharynx and neck, and contrast-enhanced CT of the chest and upper abdomen. Study endpoints included overall survival (OS), PFS, distant metastasis-free survival (DMFS), and locoregional relapse-free survival (LRRFS). Treatment-related toxicities were recorded throughout the treatment and follow-up period and graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0 (NCI CTCAE v5.0). Late radiation toxicities were assessed using RTOG criteria.

Statistical analysis

Continuous and categorical variables are presented as the median [interquartile range (IQR)] and count (percentage), respectively. Baseline characteristics were analyzed using descriptive statistics. Dosimetric differences between plans were compared by paired t-test. Propensity score matching (PSM) was performed between the optimized and matched control groups using a 1:1 nearest-neighbor matching algorithm. The matching variables included age, sex, T classification, N classification, clinical stage, IC cycles, IC regimen, and targeted therapy. Post-IC response variables were strictly excluded from the matching model to avoid violating PSM assumptions.

The median follow-up duration was estimated using the reverse Kaplan-Meier method. Survival outcomes (PFS, OS, DMFS, and LRRFS) were estimated using the Kaplan-Meier method, and the data are presented as the median and 95% confidence interval (CI). Survival differences between subgroups were assessed by log-rank test. The hazard ratio (HR) and 95% CI were estimated using Cox proportional hazards models. Logistic regression analyses were performed to evaluate associations between clinical variables and treatment response. For exploratory variable selection, least absolute shrinkage and selection operator (LASSO) Cox regression was performed utilizing a 10-fold cross-validation to select the optimal penalty parameter (λ). The λ value that minimized the partial likelihood deviance (λmin) was chosen to identify the most robust prognostic variables. All statistical analyses were conducted using SPSS software (version 26.0; IBM Corp., Armonk, NY, USA) and R software (version 4.1.0; R Foundation for Statistical Computing, Vienna, Austria). A P value <0.05 was considered statistically significant.

Results

Baseline characteristics

Of the 316 patients reviewed, 214 patients met all eligibility criteria and were included in the final analysis (Figure 2). The median age of the patients was 50 years (range, 38–56 years). The most commonly involved lymph node region was level II. The clinical stages at diagnosis were II, III, and IVA in 15.4%, 40.2%, and 44.4% of cases, respectively. Baseline clinical and treatment characteristics did not differ significantly between patients who received optimized nodal coverage and those who received standard treatment (Table 1).

Figure 2 Patient selection flowchart. AJCC, American Joint Committee on Cancer; CR, complete response; WHO, World Health Organization.

Response to IC and cohort optimization

Following IC, the median short-axis diameter of lymph nodes was decreased from 13.15 mm (IQR, 11.6–17 mm) to 8.95 mm (IQR, 7.28–12 mm), and the median long-axis diameter was reduced from 19 mm (IQR, 16–25 mm) to 13.85 mm (IQR, 10.3–17.1 mm). The median reduction in GTV was 72.7% (IQR, 56.6–83.0%). The objective and CR rates were 67.8% (145/214) and 33.6% (72/214), respectively. Changes in nodal dimensions are shown in Figure 3A.

Figure 3 Nodal response and survival outcomes. (A) Changes in short-axis diameter of cervical lymph nodes from baseline to post-IC in all patients (n=214). (B-E) Kaplan-Meier estimates of PFS (B), OS (C), LRRFS (D), and DMFS (E) in the de-escalation cohort (n=60). (F) Post-IC nodal CR was associated with better PFS compared to non-CR in all patients (n=214). Tumor response was assessed according to RECIST v1.1. CI, confidence interval; CR, complete response; DMFS, distant metastasis-free survival; HR, hazard ratio; IC, induction chemotherapy; LRRFS, locoregional relapse-free survival; OS, overall survival; PFS, progression-free survival; RECIST, Response Evaluation Criteria in Solid Tumors.

Based on post-IC images, patients with necrosis or ENE (31/214), bulky nodes ≥3 cm (9/214), multiple nodes within a single drainage region (40/214), or isolated non-CR (27/214) were excluded from nodal dose de-escalation. A total of 47 patients had multiple high-risk factors. Logistic regression analysis showed that, after adjustment for age and stage, multiple nodes within a single drainage region, nodal necrosis, and bulky nodal disease remained significantly associated with post-IC CR, whereas N stage and overall clinical stage were no longer independently associated (Table S1).

Ultimately, 60 patients met the criteria for de-escalation (Figure 2). A total of 65 nodal targets were optimized; 4 patients had bilateral involvement, and 1 patient required ipsilateral level II and III node optimization. Dose stratification for GTVnd was based on the residual nodal size after IC: 50.4 Gy for short-axis diameter <5 mm (n=3), 60 Gy for 5–8 mm (n=23), and 64 Gy for 8–10 mm (n=39). Elective CTV2 regions were prescribed 50.4 Gy in all cases.

Treatment outcomes

As of 31 May, 2025, the median follow-up duration was 45.0 months (IQR, 38.0–55.0 months). Overall, 50 patients (23.4%) experienced treatment failure, comprising 21 patients with locoregional recurrence only, 27 patients with distant metastasis only, and 2 patients with both. All locoregional failures occurred within regions with high-dose target volumes. The most common metastatic sites were bone and liver. A total of 10 patients (4.7%) died, with 8 and 2 deaths attributed to disease progression and unrelated causes, respectively. The rates of treatment failure, locoregional recurrence, distant metastasis, and death were 10.0%, 1.7%, 8.3%, and 3.3% in the optimized group, compared with 28.6%, 14.3%, 14.9%, and 5.2% in the conventional group, respectively.

In the optimized cohort (n=60), 6 patients experienced treatment failure (1 case of local recurrence and 5 cases of distant metastasis), with no nodal recurrences within de-escalated regions. Based on the exact binomial distribution, the 95% upper confidence limit for the in-field nodal failure rate in this optimized cohort was 4.9% (Figure 3B-3E). In comparison with LA-NPC patients who did not achieve CR, those who achieved CR after IC demonstrated better PFS (Figure 3F). In univariable analysis, nodal CR, T stage, and nodal necrosis were significantly associated with PFS. In multivariable analysis, only T stage and nodal CR remained independent predictors of PFS (Table S2).

Propensity score-matched survival analysis

To balance baseline characteristics between groups, PSM was performed using a 1:1 matching approach. After matching, 60 patients remained in each group, and baseline characteristics were well balanced (Table S3).

Kaplan-Meier analysis showed no significant differences between the optimized and control groups in 3-year OS, LRRFS, or DMFS (Figure 4). The 3-year PFS was 91.7% in the optimized group and 75.0% in the control group (P=0.02).

Figure 4 PSM-based survival analysis. (A-D) Kaplan-Meier estimates of PFS (A), OS (B), LRRFS (C), and DMFS (D) after 1:1 PSM in the optimized and matched control groups. CI, confidence interval; DMFS, distant metastasis-free survival; HR, hazard ratio; LRRFS, locoregional relapse-free survival; OS, overall survival; PFS, progression-free survival; PSM, propensity score matching.

EBV DNA kinetics in association with nodal response and survival

Dynamic monitoring of plasma EBV DNA was performed for all patients. At the end of IC, EBV DNA was undetectable in 64.5% of patients (138/214). The clearance rate of EBV DNA after IC was higher in patients achieving radiological CR than in those with non-CR (80.6% vs. 56.3%; P<0.001; Figure 5A). After the completion of radiotherapy, the EBV DNA clearance rate was further increased to 76.6%.

Figure 5 EBV DNA response patterns and association with survival outcomes. (A) Plasma EBV DNA clearance status distributed across nodal response groups (CR vs. non-CR), with a higher clearance rate observed in the CR group. (B-E) EBV DNA clearance after IC was associated with better PFS (B), OS (C), LRRFS (D), and DMFS (E) compared to non-clearance. CI, confidence interval; CR, complete response; DMFS, distant metastasis-free survival; EBV, Epstein-Barr virus; HR, hazard ratio; IC, induction chemotherapy; LRRFS, locoregional relapse-free survival; OS, overall survival; PFS, progression-free survival.

Failure to clear EBV DNA after IC was associated with significantly poorer PFS (HR 3.91; 95% CI: 2.19–6.97), OS (HR 4.10; 95% CI: 1.06–15.88), LRRFS (HR 4.44; 95% CI: 1.83–10.80), and DMFS (HR 3.02; 95% CI: 1.42–6.45) (Figure 5B-5E). Exploratory LASSO Cox regression analysis identified post-IC EBV DNA clearance as the most stable variable associated with PFS (Figure S1).

EBV DNA kinetics within 12 months after treatment were categorized as complete clearance or non-clearance, with the latter including both persistently positive and bounce patterns. In comparison with complete clearance, non-clearance was associated with significantly worse PFS, OS, DMFS, and LRRFS (Table 2, Figure S2).

Dosimetric impact of nodal de-escalation

To evaluate the dosimetric effect of the nodal de-escalation strategy, 15 patients from the de-escalation cohort were randomly selected for paired IMRT replanning. The conventional plan (CTV1 60 Gy; GTVnd 66 Gy) was compared with the optimized plan (CTV2 50.4 Gy; GTVnd 60 Gy). In comparison with the conventional plan, the optimized plan significantly reduced radiation doses prescribed to salivary glands. For the submandibular glands, the mean dose was reduced from 32.36 to 28.82 Gy, and V₃₉ was reduced from 28.14% to 16.29% (P<0.001). For the parotid glands, the mean dose was reduced from 29.07 to 23.85 Gy, and V₂₆ was reduced from 45.41% to 33.82% (P<0.001, Figure 6, Table 3).

Figure 6 Dosimetric advantages of the optimized plan compared with the conventional plan. (A,B) Dose distributions of the submandibular gland under the conventional (A) and optimized (B) plans. The submandibular gland is delineated by pink contours. (C,D) Dose distributions of the parotid gland under the conventional (C) and optimized (D) plans. The parotid gland is delineated by yellow contours. (E-H) Paired dosimetric comparisons of salivary glands. Lines connect paired plans from the same patient. ****, P<0.0001.

Toxicities in the de-escalation cohort

Treatment-related toxicities were assessed in all 60 patients who underwent nodal de-escalation, during both treatment and long-term follow-up. Acute toxicities associated with radiotherapy and/or chemotherapy were generally well tolerated. No grade 4 adverse events were observed, and no patients discontinued treatment due to toxicity. The most common acute toxicities were mucositis and xerostomia, followed by hematologic and gastrointestinal toxicities.

At the last follow-up, no persistent xerostomia or oral burning sensations were documented. According to RTOG/European Organization for Research and Treatment of Cancer (EORTC) criteria, late xerostomia was classified as grade 0, 1, and 2–3 in 13.3% (8/60), 83.3% (50/60), and 3.3% (2/60) of patients, respectively. The majority of cases were grade 1, indicating predominantly mild xerostomia.

Discussion

Despite the well-established benefits of IC for LA-NPC, the feasibility of reducing neck irradiation intensity without compromising disease control remains uncertain. We systematically evaluated an adaptive, response-guided nodal dose de-escalation strategy, optimizing both target volume and prescription dose for cervical lymph nodes achieving CR after IC and without high-risk features. Among 214 patients, 60 patients were selected for stratified dose reduction (50–64 Gy). No recurrences were observed within the de-escalated nodal regions in this subgroup, and the overall failure rate was 10.0%. Importantly, de-escalation led to significant salivary gland sparing and a reduction in late xerostomia, without adverse effects on regional control or OS outcomes. These findings support the use of an individualized, response-guided approach to selective nodal dose de-escalation.

Prospective studies have shown that post-IC contouring of the reduced GTVnd with lower prescribed doses mitigates high-grade toxicities without loss of local control, but nodal de-escalation has not been systematically evaluated.

In this study, GTVnd dose prescription was guided by post-IC nodal response. Current guidelines typically recommend 66–70 Gy for all metastatic nodes (15). However, delivering high doses to subclinical or regressed lesions after IC may be unnecessary. Accordingly, nodes showing CR and measuring <5 mm post-IC were incorporated into the low-risk CTV (50.4 Gy), whereas residual nodes ≥5 mm were prescribed 60–64 Gy, representing a 2–16 Gy reduction compared with IG-2018. This response-adapted approach minimized unnecessary high-dose exposure and expanded the safety margin between GTV and adjacent OARs. Although radiological evaluation cannot perfectly substitute for pathological diagnosis, pretreatment nodal biopsy is generally avoided in NPC because of concerns regarding tumor seeding and disruption of lymphatic drainage. Under this clinical context, comprehensive radiological assessment remains the most practical surrogate for response-adapted nodal selection.

For elective irradiation, current guidelines remain inconsistent regarding dose prescriptions for high-risk cervical CTV, recommending 59.4 Gy (RTOG) (16), 60 Gy (IG-2018) (13), or 59.4–63 Gy (National Comprehensive Cancer Network) (17). Whether high-dose irradiation is still necessary for regions showing radiologic CR after IC remains uncertain. Fowler’s radiobiological modeling (18) suggested that >90% tumor control could be achieved with 46–50 Gy, indicating that microscopic disease may not require doses ≥60 Gy. Consistently, a phase II trial that delivered 48 Gy to the low-risk prophylactic volume in stage III NPC patients with CR/PR and EBV DNA clearance achieved a 2-year PFS of 94.8% without grade ≥3 late toxicities (2). Moreover, approximately 60% of experts in the recent IG-2024 consensus supported reducing selected 60–70 Gy regions to 50 Gy (15), though clear criteria are still lacking. In this context, our study helps define a clinically applicable subgroup in whom 50 Gy was associated with reliable control of post-IC CR nodal regions and improved long-term safety. Our findings parallel the ongoing efforts of prospective trials like NCT04448522, which is exploring de-escalation across both gross and elective target volumes, with the high-risk elective volume reduced from 59.4 to 54 Gy and gross primary and nodal doses reduced from 69.96 to 63.6 Gy.

The concurrent reduction of both GTV and CTV optimized overall dose distribution, significantly lowering radiation exposure to normal tissues while maintaining tumor control and reducing treatment-related toxicity. The 3-year LRRFS was 94.6% for the entire cohort and 100% in the optimized group, comparable to historical data from standard full-dose nodal irradiation (19,20). After PSM, no significant differences were observed between the optimized and control groups in OS, LRRFS, or DMFS; the improvement in PFS was more likely attributable to treatment sensitivity and low-risk tumor phenotype, supporting the safety of the de-escalation strategy. Among the 60 patients receiving de-escalated radiotherapy, 5 (8.3%) developed distant metastases, and none experienced recurrence within the de-escalated nodal regions. As de-escalation was limited to the neck, these failures likely reflected systemic progression rather than regional insufficiency.

Dosimetric analysis further confirmed the benefits of this strategy. In 15 patients who underwent paired IMRT planning, synchronous GTV and CTV reduction significantly decreased mean doses to the parotid and submandibular glands. The resultant reduction in target volumes translated into a more favorable toxicity profile. Previous studies have established a clear dose–response relationship between parotid irradiation and long-term salivary recovery. Li et al. (21) reported that when the parotid mean dose is maintained below approximately 25–30 Gy, salivary function can substantially recover and approach pretreatment levels within 2 years. Eisbruch et al. (22) identified around 26 Gy as the threshold associated with preserved salivary flow, with little recovery observed above this dose. In our study, the optimized plans yielded a parotid mean dose of 23.85 Gy, lower than this favorable range. Based on follow-up records, most patients experienced only grade 1 late xerostomia, and grade ≥2 events were observed in 3.3% of cases, supporting the potential of nodal dose de-escalation to mitigate long-term salivary toxicity.

Notably, although approximately one-third of patients achieved radiological CR, only those meeting predefined clinical criteria—including CR and absence of necrosis, ENE, bulky nodes (≥3 cm), or multiple nodes in the same drainage region—were considered for dose de-escalation. These criteria aimed to minimize recurrence risk under reduced treatment intensity. Exclusion criteria for de-escalation were based on well-established prognostic indicators. Cervical nodal necrosis, an indicator of hypoxia-driven radioresistance, is independently associated with inferior OS, DMFS, and LRRFS, which can increase the risk of distant failure and mortality by 80–90% (23,24). ENE, defined as tumor spread beyond the nodal capsule, is an independent adverse prognostic factor for survival (25). Radiological ENE confers a 2.9-fold increased risk of distant metastasis, which increases to nearly 6-fold if soft tissue invasion is present (26). The 9th AJCC/Union for International Cancer Control (UICC) staging system (27) accordingly classifies advanced ENE as N3 disease. Bulky nodes are another marker of aggressive tumor biology. Although N3 disease is defined by a size ≥6 cm, recent studies have shown that a maximal axial diameter (MAD) ≥3 cm may independently predict inferior OS and recurrence-free survival and thus may warrant consideration in future staging revision (28). Beyond linear dimensions, cervical nodes with a volume ≥18 mL can substantially increase the risks of recurrence and death (29). Furthermore, the presence of multiple nodes within a single drainage region, especially when matted or confluent, has been independently linked to higher tumor burden and inferior survival outcomes (30). These findings provide strong rationale for excluding high-risk nodes from de-escalation protocols.

Our study further demonstrated that when de-escalation was confined to low-risk nodes meeting these clinical criteria, regional control remained excellent, with no in-field nodal failures observed. This favorable outcome may also reflect the early treatment response achieved by this subgroup. In the Cox analysis of the entire cohort, post-IC nodal CR remained an independent predictor of PFS after multivariable adjustment. Patients with post-IC nodal CR exhibited improved PFS and a higher EBV DNA clearance rate (80.6% vs. 56.3%). Although EBV DNA was neither a primary endpoint nor an eligibility criterion in this study, its clearance status showed a consistent favorable association with PFS, OS, DMFS, and LRRFS and closely paralleled radiologic CR. In exploratory LASSO regression analysis, post-IC EBV DNA clearance was stably retained as the variable most strongly associated with all survival outcomes. This concordance underscores the consistency between molecular and radiologic indicators of treatment response, suggesting that EBV DNA clearance may help identify the low-risk phenotype targeted by our response-adapted optimization and guide patient selection for de-escalation in future prospective trials.

This study has several limitations. First, it was a single-center retrospective analysis, and patients selected for de-escalation inherently had favorable tumor biology, including high chemosensitivity and the absence of high-risk features, introducing an intrinsic biological selection bias that likely contributed to their excellent outcomes. Second, all patients were treated in an endemic region, and the applicability of these findings to non-endemic, HPV-associated, or smoking-related NPC populations requires separate prospective validation. Third, although the median follow-up was 45.0 months and captured most recurrences, it may still be insufficient to detect very late relapses. Fourth, dosimetric comparison was limited to a randomly sampled subset of 15 patients due to the intensive computational and human resources required for paired IMRT replanning. Although the sample size was relatively small, the randomized selection process and subsequent power calculations supported the statistical robustness of the observed salivary gland sparing benefits. Finally, our toxicity analysis was limited to clinician-graded xerostomia; other relevant late toxicities, including dysphagia, hearing loss, and thyroid dysfunction, were not systematically evaluated due to the retrospective design. Consequently, we limited our conclusions to objective clinician-graded criteria and did not make subjective quality-of-life claims without validated patient-reported outcome (PRO) instruments.

Overall, the findings provide a strong rationale for prospective trials investigating response-adapted, risk-stratified radiotherapy in LA-NPC, with the potential to improve long-term quality of life without compromising tumor control.

Conclusions

In conclusion, adaptive radiotherapy based on nodal response to IC was feasible in patients with stage II–IVA NPC who achieved complete nodal response and had no high-risk features. This dose de-escalation strategy maintained locoregional control with no in-field nodal failures and reduced radiation exposure to the salivary glands, resulting in a low incidence of late xerostomia. These findings support selective nodal dose reduction in carefully selected patients without compromising disease control.

Acknowledgments

We thank the patients and investigators who participated in this study. During the preparation of this work, the authors used ChatGPT-4o to improve language and readability. No AI technologies were used for data analysis, statistical computation, or clinical interpretation. The authors reviewed and edited the content as needed and take full responsibility for the final published content.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0320/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0320/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0320/prf

Funding: This study was supported by a grant from National Natural Science Foundation of China (No. 82172804).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2026-1-0320/coif). All authors report that this study was supported by the grant from National Natural Science Foundation of China (No. 82172804). The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This retrospective study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments, and was approved by the Ethics Committee of Jiangsu Cancer Hospital (No. KB-2020-055). Written informed consent was provided by all participants.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Chen YP, Chan ATC, Le QT, et al. Nasopharyngeal carcinoma. Lancet 2019;394:64-80. [Crossref] [PubMed]
2. Guo SS, Yang JH, Sun XS, et al. Reduced-dose radiotherapy for Epstein-Barr virus DNA selected staged III nasopharyngeal carcinoma: A single-arm, phase 2 trial. Eur J Cancer 2023;194:113336. [Crossref] [PubMed]
3. Huang CL, Sun ZQ, Guo R, et al. Plasma Epstein-Barr Virus DNA Load After Induction Chemotherapy Predicts Outcome in Locoregionally Advanced Nasopharyngeal Carcinoma. Int J Radiat Oncol Biol Phys 2019;104:355-61. [Crossref] [PubMed]
4. Lee AW, Lau KY, Hung WM, et al. Potential improvement of tumor control probability by induction chemotherapy for advanced nasopharyngeal carcinoma. Radiother Oncol 2008;87:204-10. [Crossref] [PubMed]
5. Li Z, Chen Y, Ma J. Advances in basic and translational research into nasopharyngeal carcinoma. Oncol Transl Med 2025;11:10-6. [Crossref]
6. Petit C, Lee A, Ma J, et al. Role of chemotherapy in patients with nasopharynx carcinoma treated with radiotherapy (MAC-NPC): an updated individual patient data network meta-analysis. Lancet Oncol 2023;24:611-23. [Crossref] [PubMed]
7. Tang LL, Huang CL, Zhang N, et al. Elective upper-neck versus whole-neck irradiation of the uninvolved neck in patients with nasopharyngeal carcinoma: an open-label, non-inferiority, multicentre, randomised phase 3 trial. Lancet Oncol 2022;23:479-90. [Crossref] [PubMed]
8. Tang LL, Chen L, Xu GQ, et al. Reduced-volume radiotherapy versus conventional-volume radiotherapy after induction chemotherapy in nasopharyngeal carcinoma: An open-label, noninferiority, multicenter, randomized phase 3 trial. CA Cancer J Clin 2025;75:203-15. [Crossref] [PubMed]
9. Xiang ZZ, He T, Zeng YY, et al. Epstein-Barr virus DNA change level combined with tumor volume reduction ratio after inductive chemotherapy as a better prognostic predictor in locally advanced nasopharyngeal carcinoma. Cancer Med 2023;12:1102-13. [Crossref] [PubMed]
10. Yang Q, Cao SM, Guo L, et al. Induction chemotherapy followed by concurrent chemoradiotherapy versus concurrent chemoradiotherapy alone in locoregionally advanced nasopharyngeal carcinoma: long-term results of a phase III multicentre randomised controlled trial. Eur J Cancer 2019;119:87-96. [Crossref] [PubMed]
11. Zhao C, Miao JJ, Hua YJ, et al. Locoregional Control and Mild Late Toxicity After Reducing Target Volumes and Radiation Doses in Patients With Locoregionally Advanced Nasopharyngeal Carcinoma Treated With Induction Chemotherapy (IC) Followed by Concurrent Chemoradiotherapy: 10-Year Results of a Phase 2 Study. Int J Radiat Oncol Biol Phys 2019;104:836-44. [Crossref] [PubMed]
12. van den Brekel MW, Stel HV, Castelijns JA, et al. Cervical lymph node metastasis: assessment of radiologic criteria. Radiology 1990;177:379-84. [Crossref] [PubMed]
13. Lee AW, Ng WT, Pan JJ, et al. International guideline for the delineation of the clinical target volumes (CTV) for nasopharyngeal carcinoma. Radiother Oncol 2018;126:25-36. [Crossref] [PubMed]
14. Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 2009;45:228-47. [Crossref] [PubMed]
15. Lin SJ, Guo QJ, Liu Q, et al. International Consensus Guideline on Delineation of the Clinical Target Volumes at Different Dose Levels for Nasopharyngeal Carcinoma (2024 Version). Int J Radiat Oncol Biol Phys 2025;123:415-31. [Crossref] [PubMed]
16. Lee NY, Zhang Q, Pfister DG, et al. Addition of bevacizumab to standard chemoradiation for locoregionally advanced nasopharyngeal carcinoma (RTOG 0615): a phase 2 multi-institutional trial. Lancet Oncol 2012;13:172-80. [Crossref] [PubMed]
17. Colevas AD, Yom SS, Pfister DG, et al. NCCN Guidelines Insights: Head and Neck Cancers, Version 1.2018. J Natl Compr Canc Netw 2018;16:479-90. [Crossref] [PubMed]
18. Fowler JF. The linear-quadratic formula and progress in fractionated radiotherapy. Br J Radiol 1989;62:679-94. [Crossref] [PubMed]
19. Li XY, Luo DH, Guo L, et al. Deintensified Chemoradiotherapy for Pretreatment Epstein-Barr Virus DNA-Selected Low-Risk Locoregionally Advanced Nasopharyngeal Carcinoma: A Phase II Randomized Noninferiority Trial. J Clin Oncol 2022;40:1163-73. [Crossref] [PubMed]
20. Tang LQ, Chen DP, Guo L, et al. Concurrent chemoradiotherapy with nedaplatin versus cisplatin in stage II-IVB nasopharyngeal carcinoma: an open-label, non-inferiority, randomised phase 3 trial. Lancet Oncol 2018;19:461-73. [Crossref] [PubMed]
21. Li Y, Taylor JM, Ten Haken RK, et al. The impact of dose on parotid salivary recovery in head and neck cancer patients treated with radiation therapy. Int J Radiat Oncol Biol Phys 2007;67:660-9. [Crossref] [PubMed]
22. Eisbruch A, Kim HM, Terrell JE, et al. Xerostomia and its predictors following parotid-sparing irradiation of head-and-neck cancer. Int J Radiat Oncol Biol Phys 2001;50:695-704. [Crossref] [PubMed]
23. Fu YC, Liang SB, Huang WJ, et al. Prognostic Value of Lymph Node Necrosis at Different N Stages in Patients with Nasopharyngeal Carcinoma. J Cancer 2023;14:2085-92. [Crossref] [PubMed]
24. Ai QH, Hung KF, So TY, et al. Prognostic value of cervical nodal necrosis on staging imaging of nasopharyngeal carcinoma in era of intensity-modulated radiotherapy: a systematic review and meta-analysis. Cancer Imaging 2022;22:24. [Crossref] [PubMed]
25. Zhou P, Luo Y, Wang C, et al. The prognostic implications of radiologic extranodal extension in the lymph nodes of patients with nasopharyngeal cancer. Future Oncol 2024;20:2491-502. [Crossref] [PubMed]
26. Hu YJ, Lu TZ, Guo QJ, et al. The role of radiologic extranodal extension in predicting prognosis and chemotherapy benefit for T1-2 N1 nasopharyngeal carcinoma: A multicenter retrospective study. Radiother Oncol 2023;178:109436. [Crossref] [PubMed]
27. Pan JJ, Mai HQ, Ng WT, et al. Ninth Version of the AJCC and UICC Nasopharyngeal Cancer TNM Staging Classification. JAMA Oncol 2024;10:1627-35. [Crossref] [PubMed]
28. Peng WS, Xing X, Li YJ, et al. Prognostic nomograms for nasopharyngeal carcinoma with nodal features and potential indication for N staging system: Validation and comparison of seven N stage schemes. Oral Oncol 2023;144:106438. [Crossref] [PubMed]
29. Yuan H, Ai QY, Kwong DL, et al. Cervical nodal volume for prognostication and risk stratification of patients with nasopharyngeal carcinoma, and implications on the TNM-staging system. Sci Rep 2017;7:10387. [Crossref] [PubMed]
30. Dong A, Zhu S, Ma H, et al. Matted Lymph Nodes on MRI in Nasopharyngeal Carcinoma: Prognostic Factor and Potential Indication for Induction Chemotherapy Benefits. J Magn Reson Imaging 2024;59:1976-90. [Crossref] [PubMed]

(English Language Editor: J. Jones)