BPG is committed to discovery and dissemination of knowledge
Retrospective Cohort Study Open Access
Copyright: ©Author(s) 2026. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution-NonCommercial (CC BY-NC 4.0) license. No commercial re-use. See permissions. Published by Baishideng Publishing Group Inc.
World J Clin Oncol. Aug 24, 2026; 17(8): 124227
Published online Aug 24, 2026. doi: 10.5306/wjco.124227
Timing matters: Early vs late recurrence in nasopharyngeal carcinoma and its impact on survival
Jing-Bo Wang, Yang Meng, Mei-Lin He, Xiao-Dong Huang, Ye Zhang, Run-Ye Wu, Kai Wang, Xue-Song Chen, Yuan Qu, Jiang-Hu Zhang, Qing-Feng Liu, Jing-Wei Luo, Li Gao, Jian-Ping Xiao, Ye-Xiong Li, Jun-Lin Yi, Department of Radiation Oncology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100021, China
Jing-Bo Wang, Department of Oncology, The First Hospital of Qiqihar, Qiqihar 161005, Heilongjiang Province, China
Si-Jia Zhang, Cancer Center, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430022, Hubei Province, China
Li-Hua Luo, Department of Oncology, Central Hospital of Enshi Autonomous Prefecture, Enshi Clinical College of Wuhan University, Enshi 445000, Hubei Province, China
ORCID number: Jing-Bo Wang (0000-0002-5233-5211); Yang Meng (0009-0007-9197-0465); Ye-Xiong Li (0000-0003-0985-235X); Jun-Lin Yi (0000-0002-7083-0216).
Co-first authors: Jing-Bo Wang and Yang Meng.
Co-corresponding authors: Li-Hua Luo and Jun-Lin Yi.
Author contributions: Wang JB, Meng Y, Luo LH and Yi JL designed the research study; He ML, Zhang SJ, Huang XD, Zhang Y, Wu RY, Wang K, Chen XS, Qu Y, Zhang JH, Liu QF, Luo JW and Xiao JP performed the research and collected the clinical data; Wang JB, Meng Y and He ML analyzed the data and performed statistical analysis; Wang JB and Meng Y drafted the manuscript; Li YX, Luo LH and Yi JL critically revised the manuscript and supervised the study. All authors have read and agreed to the published version of the manuscript. Wang JB and Meng Y contributed equally to this work. We designated Yi JL and Luo LH as co-corresponding authors to accurately reflect their joint leadership in executing this study. Although the analyzed patient cohort originated exclusively from the National Cancer Center, Luo LH participated in an extended period of intensive on-site collaboration within our primary research group. During this time, both senior authors jointly managed the research workflow, maintaining strict quality control over clinical data curation, methodological implementation, and manuscript development. Their combined oversight and shared administrative efforts were vital for ensuring the scientific rigor of this investigation. Therefore, shared corresponding authorship correctly acknowledges their equal accountability for overseeing the research process, addressing potential inquiries, and guaranteeing the overall integrity of the published manuscript.
AI contribution statement: Portions of this manuscript were edited using AI tools solely for language refinement. The authors carefully reviewed and verified all AI-assisted outputs and take full responsibility for the scientific content of the manuscript.
Supported by Chinese Academy of Medical Sciences (CAMS) Innovation Fund for Medical Sciences, No. 2024-I2M-C&T-B-063 and No. 2024-I2M-C&T-A-006; Beijing Hope Run Special Fund of Cancer Foundation of China, No. LC2021 L06; National Natural Science Foundation of China, No. 81172125; and National Key Research and Development Program of China, No. 2023YFC2411602.
Institutional review board statement: This research was conducted in accordance with the principles outlined in the Declaration of Helsinki and obtained ethical approval from the Ethics Committee of our institution (NCC: 23/353-4095).
Informed consent statement: Given its observational nature using retrospective clinical data, informed consent was waived.
Conflict-of-interest statement: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
STROBE statement: The authors have read the STROBE Statement-checklist of items, and the manuscript was prepared and revised according to the STROBE Statement-checklist of items.
Data sharing statement: The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Corresponding author: Jun-Lin Yi, Department of Radiation Oncology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, No. 17 Panjiayuan Nanli, Chaoyang District, Beijing 100021, China. yijunlin1969@163.com
Received: June 15, 2026
Revised: July 27, 2026
Accepted: August 20, 2026
Published online: August 24, 2026
Processing time: 72 Days and 1.5 Hours

Abstract
BACKGROUND

Despite the wide utilization of intensity-modulated radiation therapy and other advanced treatment modalities, approximately 10% of patients with nasopharyngeal carcinoma (NPC) still suffer from local or regional disease recurrence, leading to deteriorated survival outcome. Particularly, early recurrence is often associated with more aggressive biological behavior and greater resistance to initial treatments, subsequently resulting in poorer survival outcomes.

AIM

To investigate the temporal distribution of site-specific recurrence and evaluate the prognostic impact of early vs late recurrence in NPC.

METHODS

This study retrospectively included 330 recurrent NPC patients who received initial treatment of radical radiotherapy from January 1990 to December 2020 in our center. Patients were categorized into early and late recurrence groups using the maximally selected rank method. Post-recurrence overall survival (OS) was evaluated utilizing Kaplan-Meier estimates and multivariable Cox regression analysis.

RESULTS

Among the 330 eligible patients, 167 developed pure local recurrence, 86 had pure regional recurrence, and 77 had locoregional recurrence. With a median follow-up duration of 97.8 months, there was no significant difference in timing distribution across various patterns of failure. The peak incidence of progression was observed in the 2nd year, contributing 29.4% of all progression events, with a cumulative 5-year proportion of 82.8%. The optimal cutoff for differentiating early and late recurrence was identified as 24 months. Advanced T (P = 0.008) and N (P = 0.028) stages were predisposing factors for early recurrence. Survival analysis manifested that late locoregional recurrence was associated with markedly better post-recurrence OS [hazard ratio (HR) = 0.37, P = 0.006], whereas such scenario was not observed in entire cohort or sub-cohorts with pure local or regional failure. The complementary 24-month landmark analysis further unveiled that late recurrence was associated with superior post-landmark OS across the entire cohort (HR = 0.38, P < 0.001) and all failure subgroups.

CONCLUSION

Recurrence timing is a pivotal prognostic factor in recurrent NPC. Early locoregional recurrence within 24 months is indicative of significantly inferior post-recurrence OS, warranting more intensive surveillance and aggressive salvage strategies for this subset.

Key Words: Nasopharyngeal carcinoma; Recurrence; Prognosis; Post-recurrence survival; Radiotherapy

Core Tip: This is the core tip of our study focusing on recurrent nasopharyngeal carcinoma. In 330 recurrent patients, recurrence events peaked in the second year. Recurrence timing is a pivotal prognostic factor. Early locoregional recurrence within 24 months is indicative of significantly inferior post-recurrence overall survival, warranting more intensive surveillance and aggressive salvage strategies for this subset.



INTRODUCTION

In recent decades, the outcome of nasopharyngeal carcinoma (NPC) has improved remarkably due to the advancements in treatment modalities, especially the wide utilization of intensity-modulated radiation therapy as well as the integration of chemotherapy and immunotherapy[1]. However, still around 10% patients suffer from local or regional disease recurrence, leading to the deterioration in survival outcomes[2]. This highlights the necessity for a comprehensive understanding of the factors influencing survival after recurrence, which could ultimately optimize the management of recurrence. The therapeutic strategies for recurrent NPC partially depend on the recurrence site. For local recurrence, re-irradiation or surgical resection is typically recommended, whereas surgery is generally preferred for regional recurrence[3]. In cases where patients are unsuitable for radiotherapy (RT) or surgery, chemotherapy or best supportive treatment may be considered[4].

Beyond the recurrence site, the timing of recurrence also plays a critical role in shaping the overall prognosis and determining the treatment strategies. Generally, early recurrence is often associated with more aggressive biological behavior and greater resistance to the initial treatment, subsequently leading to poorer survival outcomes[5,6]. On the other hand, therapeutic approaches for recurrent NPC also vary between early and late recurrences[3]. For example, re-irradiation is generally not advised for individuals with a recurrence latency less than one year after completion of primary RT[4].

However, the temporal and locational features of recurrence and the following prognostic consequence remain poorly understood. This study sought to: (1) Depict the distribution pattern of recurrence timing and site in a large recurrent NPC cohort; (2) Further identify the predisposing factors for various recurrence timings and sites; and (3) Determine the prognostic factors for recurrent NPC.

MATERIALS AND METHODS
Patient selection

Between January 1990 and December 2020, 330 NPC patients with complete medical records who received radical RT initially at our center and experienced local or regional recurrence were retrospectively reviewed and analyzed. All patients were histologically confirmed to have NPC with no evidence of distant metastases. All patients underwent comprehensive pre-treatment evaluation and were restaged using the 8th AJCC staging system[7]. This research was conducted in accordance with the principles outlined in the Declaration of Helsinki and obtained ethical approval from the Ethics Committee of our institution (NCC: 23/353-4095). Given its observational nature using retrospective clinical data, informed consent was waived.

Follow-up schedule

Patients were advised to attend follow-up appointments at least every 3 months in the first 2 years, every 6 months during the third to fifth years, and subsequently annually until death. Assessment procedures included a thorough physical examination, fiberoptic nasopharyngoscopy, magnetic resonance for nasopharynx and neck, computed tomography (CT) of the chest, abdominal ultrasonography, as well as bone scans and PET/CT when indicated. Recurrence was confirmed by at least two sets of above-mentioned examinations, with histological or cytological confirmation strongly preferred if feasible.

Definitions of recurrence

The optimum cutoff timing for the definition of early and late recurrence was determined by adopting the maximally selected rank method, facilitated by the survminer package in R[8]. To evaluate the stability of this cutoff, internal validation using 1000 bootstrap resamples was performed. After defining the “early” and “late” recurrence, patients were categorized into six groups: Early local recurrence (ETR), late local recurrence (LTR), early regional recurrence (ENR), late regional recurrence (LNR), early locoregional recurrence (ETNR), and late locoregional recurrence (LTNR).

Statistical analysis

Post-recurrence overall survival (OS) was evaluated as the duration between confirmation of specific recurrence and any cause of death or last follow up. Furthermore, 24-month landmark OS was also explored as a complementary assessment. χ2 test was employed to analyze clinical characteristics and treatment modalities among subgroups. A Kernel density function was generated to compare the timing distribution across various patterns of failure (POF). Kaplan-Meier approach was utilized to estimate survival, with log-rank test applied to assess differences between survival curves. Multivariable Cox regression analysis was conducted to evaluate prognostic factors affecting survival. The variables of interest included age, gender, initial TNM, recurrent TNM and recurrence timing. A significance threshold of P < 0.05 was applied to determine statistical significance. Data analysis was performed with R 3.6.2.

RESULTS
General characteristics

A total of 330 recurrent patients were eligible for final analysis, including 167 patients (50.6%) with pure local recurrence, 86 patients (26.1%) with pure regional recurrence, and 77 patients (23.3%) with locoregional recurrence. In the entire group, the median follow-up duration was 97.8 months [95% confidence interval (CI): 83.9-111.7], while the median time to recurrence was 24.7 months (95%CI: 22.7-27.4). Detailed characteristics and demographics are presented in Table 1. The median OS of the entire cohort was 88.4 months (95%CI: 76.7-100.1), while the median post-recurrence OS was 35.4 months (95%CI: 27.2-43.6).

Table 1 General characteristics of the entire, local, regional and locoregional recurrence group, n (%).
Variables
Total (n = 330)
Local (n = 167)
Regional (n = 86)
Locoregional (n = 77)
P value
Gender0.615
    Male257 (77.88)131 (78.44)69 (80.23)57 (74.03)
    Female73 (22.12)36 (21.56)17 (19.77)20 (25.97)
Age0.374
    < 60271 (82.12)133 (79.64)71 (82.56)67 (87.01)
    ≥ 6059 (17.88)34 (20.36)15 (17.44)10 (12.99)
Smoking status0.112
    No202 (61.21)94 (56.29)54 (62.79)54 (70.13)
    Yes128 (38.79)73 (43.71)32 (37.21)23 (29.87)
Drinking status0.663
    No223 (67.58)109 (65.27)60 (69.77)54 (70.13)
    Yes107 (32.42)58 (34.73)26 (30.23)23 (29.87)
KPS status0.783
    60-7011 (3.33)7 (4.19)2 (2.33)2 (2.60)
    80-90319 (96.67)160 (95.81)84 (97.67)75 (97.40)
Weight loss status0.044
    < 10%308 (93.33)154 (92.22)85 (98.84)69 (89.61)
    ≥ 10%22 (6.67)13 (7.78)1 (1.16)8 (10.39)
T< 0.001
    1 or 2150 (45.45)57 (34.13)58 (67.44)35 (45.45)
    3 or 4180 (54.55)110 (65.87)28 (32.56)42 (54.55)
N< 0.001
    0 or 1150 (45.45)98 (58.68)24 (27.91)28 (36.36)
    2 or 3180 (54.55)69 (41.32)62 (72.09)49 (63.64)
TNM0.853
    I-II63 (19.09)33 (19.76)17 (19.77)13 (16.88)
    III-IV267 (80.91)134 (80.24)69 (80.23)64 (83.12)
Radiotherapy technique0.863
    Non-IMRT196 (59.39)101 (60.48)49 (56.98)46 (59.74)
    IMRT134 (40.61)66 (39.52)37 (43.02)31 (40.26)
Time of treatment0.632
    1990-1999113 (34.24)57 (34.13)25 (29.07)31 (40.26)
    2000-2009126 (38.18)65 (38.92)36 (41.86)25 (32.47)
    2010-202091 (27.58)45 (26.95)25 (29.07)21 (27.27)
Chemotherapy0.149
    No206 (62.42)112 (67.07)47 (54.65)47 (61.04)
    Yes124 (37.58)55 (32.93)39 (45.35)30 (38.96)
Early or late recurrence0.109
    Early156 (47.27)74 (44.31)49 (56.98)33 (42.86)
    Late174 (52.73)93 (55.69)37 (43.02)44 (57.14)
rT< 0.001
    086 (26.06)0 (0.00)86 (100.00)0 (0.00)
    1 or 2119 (36.06)74 (44.31)0 (0.00)45 (58.44)
    3 or 4125 (37.88)93 (55.69)0 (0.00)32 (41.56)
rN< 0.001
    0167 (50.61)167 (100.00)0 (0.00)0 (0.00)
    1 or 2150 (45.45)0 (0.00)77 (89.53)73 (94.81)
    313 (3.94)0 (0.00)9 (10.47)4 (5.19)
rTNM< 0.001
    I-II158 (47.88)74 (44.31)57 (66.28)27 (35.06)
    III-IV172 (52.12)93 (55.69)29 (33.72)50 (64.94)
Post recurrence treatment< 0.001
    Chemotherapy36 (10.91)19 (11.38)7 (8.14)10 (12.99)
    RT150 (45.45)90 (53.89)19 (22.09)41 (53.25)
    Surgery107 (32.42)33 (19.76)56 (65.12)18 (23.38)
    BST37 (11.21)25 (14.97)4 (4.65)8 (10.39)

Table 1 compares the baseline characteristics of patients among the local, regional, and locoregional recurrence groups. Not surprisingly, the local recurrence group harbored more aggressive T stage at the initial diagnosis, and the regional recurrence group exhibited a more advanced initial N stage. There were also significant distinctions with regard to the post-recurrence management. More patients in the local and locoregional recurrence groups underwent re-irradiation, whereas the regional recurrence group received surgical resections more frequently. No differences between groups were found regarding gender, age, smoking habits, drinking habits, RT techniques or treatment era.

Chronological distribution of recurrences

The detailed distribution of time to recurrence is illustrated in Figure 1A-C. In the overall cohort, the peak incidence of recurrence was observed in the 2nd year, contributing 29.4% of all recurrence events. Afterwards, a gradual decline in the annual number of recurrence was observed, resulting in the scenario that recurrences occurring within 5 years accounted for 82.8% (Figure 1A). Specifically for local recurrence, the highest recurrence (27.9%) was also noted in the 2nd year, with cumulative rates reaching 83.7% at the fifth year (Figure 1B). In cases of regional recurrence, the peak period was almost evenly distributed within the 1st and 2nd year, achieving a cumulative percentage of 81.0% by the 5th year (Figure 1C). There was no significant difference in timing distribution across various POF (Supplementary Figure 1).

Figure 1
Figure 1 Chronological distribution of recurrence types. A: Local or regional recurrence; B: Local recurrence; C: Regional recurrence; D: Proportion of recurrence types among patients. Recurrence incidence peaked during the 2nd year for local or regional, local, and regional recurrence. ENR: Early regional recurrence; ETNR: Early locoregional recurrence; ETR: Early local recurrence; LNR: Late regional recurrence; LTNR: Late locoregional recurrence; LTR: Late local recurrence.
Identification of early and late recurrence

The 24-month interval of time to recurrence was identified as the optimal timing cutoff to define early and late recurrence with most significantly distinct post-recurrence OS. Internal bootstrap validation (1000 resamples) confirmed the robustness of this 24-month cutoff. Accordingly, 156 patients (47.3%) were categorized into the early recurrence group, with a median time to recurrence of 14.3 months (95%CI: 13.4-15.6), while 174 patients (52.7%) were categorized into the late recurrence group, exhibiting a median time to recurrence of 47.7 months (95%CI: 41.4-53.1). Within the local recurrence subgroup, 74 patients (44.3%) exhibited ETR, and 93 patients (55.7%) developed LTR. Among those with regional recurrence, 49 patients (57.0%) experienced ENR, and 37 patients (43.0%) had LNR. In the locoregional recurrence group, 33 patients (42.9%) developed ETNR, and 44 patients (57.1%) developed LTNR. The composition of patients is illustrated in Figure 1D.

Factors associated with early vs late recurrence

Table 2 compares the characteristics of patients among 4 pairs of groups. More advanced T stage was significantly linked with overall early recurrence and ETR, but not with ENR or ETNR. Accordingly, more advanced N stage was significantly related with overall early recurrence, but it did not correlate with ETR, ENR or ETNR. More aggressive TNM stage was significantly associated with overall early recurrence and ETR, yet not correlated with ENR or ETNR. The distribution of treatment era differed significantly between early and late recurrence in the overall cohort (P = 0.008) and the regional recurrence subgroup (P = 0.032), where earlier treatment eras (1990-1999) accounted for a higher proportion of late recurrences. Dichotomous factors including gender, age, Karnofsky Performance status, weight loss status, RT techniques and recurrence stage did not demonstrate any difference among subgroups with early and late recurrence.

Table 2 Comparison of general characteristics between early vs late recurrence in terms of the entire, local, regional and locoregional recurrence group, n (%).
Variables
Total (n = 330)
Local (n = 167)
Regional (n = 86)
Locoregional (n = 77)
Early (n = 156)
Late (n = 174)
P value
ETR (n = 74)
LTR (n = 93)
P value
ENR (n = 49)
LNR (n = 37)
P value
ETNR (n = 33)
LTNR (n = 44)
P value
Gender0.6890.7180.7080.822
    Male123 (78.85)134 (77.01)59 (79.73)72 (77.42)40 (81.63)29 (78.38)24 (72.73)33 (75.00)
    Female33 (21.15)40 (22.99)15 (20.27)21 (22.58)9 (18.37)8 (21.62)9 (27.27)11 (25.00)
Age0.9750.4240.7540.129
    < 60128 (82.05)143 (82.18)61 (82.43)72 (77.42)41 (83.67)30 (81.08)26 (78.79)41 (93.18)
    ≥ 6028 (17.95)31 (17.82)13 (17.57)21 (22.58)8 (16.33)7 (18.92)7 (21.21)3 (6.82)
Smoking status0.0900.1440.7290.281
    No88 (56.41)114 (65.52)37 (50.00)57 (61.29)30 (61.22)24 (64.86)21 (63.64)33 (75.00)
    Yes68 (43.59)60 (34.48)37 (50.00)36 (38.71)19 (38.78)13 (35.14)12 (36.36)11 (25.00)
Drinking status0.9220.9220.3000.350
    No105 (67.31)118 (67.82)48 (64.86)61 (65.59)32 (65.31)28 (75.68)25 (75.76)29 (65.91)
    Yes51 (32.69)56 (32.18)26 (35.14)32 (34.41)17 (34.69)9 (24.32)8 (24.24)15 (34.09)
KPS status0.1770.2130.5040.504
    60-703 (1.92)8 (4.60)1 (1.35)6 (6.45)2 (4.08)0 (0.00)0 (0.00)2 (4.55)
    80-90153 (98.08)166 (95.40)73 (98.65)87 (93.55)47 (95.92)37 (100.00)33 (100.00)42 (95.45)
Weight loss status0.2500.4711.0000.419
    < 10%143 (91.67)165 (94.83)67 (90.54)87 (93.55)48 (97.96)37 (100.00)28 (84.85)41 (93.18)
    ≥ 10%13 (8.33)9 (5.17)7 (9.46)6 (6.45)1 (2.04)0 (0.00)5 (15.15)3 (6.82)
T0.0080.0020.6270.064
    1 or 259 (37.82)91 (52.30)16 (21.62)41 (44.09)32 (65.31)26 (70.27)11 (33.33)24 (54.55)
    3 or 497 (62.18)83 (47.70)58 (78.38)52 (55.91)17 (34.69)11 (29.73)22 (66.67)20 (45.45)
N0.0280.1620.4160.338
    0 or 161 (39.10)89 (51.15)39 (52.70)59 (63.44)12 (24.49)12 (32.43)10 (30.30)18 (40.91)
    2 or 395 (60.90)85 (48.85)35 (47.30)34 (36.56)37 (75.51)25 (67.57)23 (69.70)26 (59.09)
TNM0.0020.0030.7080.114
    I-II19 (12.18)44 (25.29)7 (9.46)26 (27.96)9 (18.37)8 (21.62)3 (9.09)10 (22.73)
    III-IV137 (87.82)130 (74.71)67 (90.54)67 (72.04)40 (81.63)29 (78.38)30 (90.91)34 (77.27)
Radiotherapy technique0.3510.4731.0000.569
    Non-IMRT88 (56.41)108 (62.07)42 (56.76)59 (63.44)28 (57.14)21 (56.76)18 (54.55)28 (63.64)
    IMRT68 (43.59)66 (37.93)32 (43.24)34 (36.56)21 (42.86)16 (43.24)15 (45.45)16 (36.36)
Time of treatment0.0080.2050.0320.531
    1990-199941 (26.28)72 (41.38)20 (27.03)37 (39.78)9 (18.37)16 (43.24)12 (36.36)19 (43.18)
    2000-200971 (45.51)55 (31.61)33 (44.59)32 (34.41)25 (51.02)11 (29.73)13 (39.39)12 (27.27)
    2010-202044 (28.21)47 (27.01)21 (28.38)24 (25.81)15 (30.61)10 (27.03)8 (24.24)13 (29.55)
Chemotherapy0.7530.9020.7330.946
    No96 (61.54)110 (63.22)50 (67.57)62 (66.67)26 (53.06)21 (56.76)20 (60.61)27 (61.36)
    Yes60 (38.46)64 (36.78)24 (32.43)31 (33.33)23 (46.94)16 (43.24)13 (39.39)17 (38.64)
rT0.0690.488-0.423
    049 (31.41)37 (21.26)0 (0.00)0 (0.00)49 (100.00)37 (100.00)0 (0.00)0 (0.00)
    1 or 256 (35.90)63 (36.21)35 (47.30)39 (41.94)0 (0.00)0 (0.00)21 (63.64)24 (54.55)
    3 or 451 (32.69)74 (42.53)39 (52.70)54 (58.06)0 (0.00)0 (0.00)12 (36.36)20 (45.45)
rN0.532-0.7910.824
    074 (47.44)93 (53.45)74 (100.00)93 (100.00)0 (0.00)0 (0.00)0 (0.00)0 (0.00)
    1 or 275 (48.08)75 (43.10)0 (0.00)0 (0.00)43 (87.76)34 (91.89)32 (96.97)41 (93.18)
    37 (4.49)6 (3.45)0 (0.00)0 (0.00)6 (12.24)3 (8.11)1 (3.03)3 (6.82)
rTNM0.3420.4880.8260.836
    I-II79 (50.64)79 (45.40)35 (47.30)39 (41.94)32 (65.31)25 (67.57)12 (36.36)15 (34.09)
    III-IV77 (49.36)95 (54.60)39 (52.70)54 (58.06)17 (34.69)12 (32.43)21 (63.64)29 (65.91)
Post recurrence treatment0.0540.2830.5160.573
    Chemotherapy18 (11.54)18 (10.34)11 (14.86)8 (8.60)3 (6.12)4 (10.81)4 (12.12)6 (13.64)
    RT64 (41.03)86 (49.43)38 (51.35)52 (55.91)11 (22.45)8 (21.62)15 (45.45)26 (59.09)
    Surgery61 (39.10)46 (26.44)17 (22.97)16 (17.20)34 (69.39)22 (59.46)10 (30.30)8 (18.18)
    BST13 (8.33)24 (13.79)8 (10.81)17 (18.28)1 (2.04)3 (8.11)4 (12.12)4 (9.09)
Prognostic factors associated with post-recurrence OS

Multivariable analysis of post-recurrence OS is exhibited in Table 3. Age over 60 was associated with an unfavorable prognosis in the entire group [hazard ratio (HR) = 1.73, 95%CI: 1.21-2.48, P = 0.003] and regional recurrence group (HR = 2.53, 95%CI: 1.20-5.31, P = 0.014). In the regional failure group, female gender was an independent factor associated with more favorable post-recurrence OS (HR = 0.29, 95%CI: 0.11-0.76, P = 0.012). More advanced rTNM stage was related to worse post-recurrence OS in the entire group (HR = 1.88, 95%CI: 1.38-2.56, P < 0.001) and the regional recurrence group (HR = 3.26, 95%CI: 1.68-6.34, P < 0.001). In terms of recurrence timing, late recurrence was linked with improved post-recurrence OS in the locoregional recurrence subgroup (HR = 0.37, 95%CI: 0.19-0.76, P = 0.006), whereas no similar trend was observed in other subgroups. Accordingly, patients with late recurrence exhibited notably better post-recurrence OS in the locoregional recurrence cohort (median: 45.4 months vs 17.2 months, P = 0.004) (Figure 2). No significant differences were found between early and late recurrence subgroups among the entire cohort (median: 45.4 months vs 29.7 months, P = 0.401), LTR and ETR (median: 42.1 months vs 29.7 months, P = 0.783), LNR and ENR (median: 60.0 months vs 59.2 months, P = 0.817), respectively.

Figure 2
Figure 2 Post-recurrence overall survival between early and late recurrence. A: Overall cohort; B: Local recurrence cohort; C: Regional recurrence cohort; D: Locoregional recurrence cohort. No significant difference in post-recurrence overall survival (OS) was observed between early and late recurrence in the overall, local, and regional cohorts. However, late locoregional recurrence significantly improved post-recurrence OS. ENR: Early regional recurrence; ETNR: Early locoregional recurrence; ETR: Early local recurrence; LNR: Late regional recurrence; LTNR: Late locoregional recurrence; LTR: Late local recurrence.
Table 3 Multivariable analysis of early vs late recurrence as a post-recurrence prognostic factor in the entire, local, regional and locoregional recurrence group.
Variables
Total (n = 330)
Local (n = 167)
Regional (n = 86)
Locoregional (n = 77)
P value
HR (95%CI)
P value
HR (95%CI)
P value
HR (95%CI)
P value
HR (95%CI)
Gender
    Male1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
    Female0.1350.76 (0.53-1.09)0.3580.79 (0.47-1.31)0.0120.29 (0.11-0.76)0.3941.35 (0.68-2.66)
Age
    < 601.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
    ≥ 600.0031.73 (1.21-2.48)0.0561.61 (0.99-2.61)0.0142.53 (1.20-5.31)0.3851.50 (0.60-3.78)
TNM
    I-II1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
    III-IV0.2691.26 (0.84-1.88)0.0531.77 (0.99-3.17)0.5530.79 (0.36-1.74)0.4220.65 (0.23-1.85)
Early or late recurrence
    Early1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
    Late0.4610.89 (0.66-1.21)0.8721.04 (0.67-1.59)0.5611.21 (0.64-2.28)0.0060.37 (0.19-0.76)
rTNM
    I-II1.00 (reference)1.00 (reference)1.00 (reference)1.00 (reference)
    III-IV< 0.0011.88 (1.38-2.56)0.1321.39 (0.91-2.13)< 0.0013.26 (1.68-6.34)0.4511.32 (0.64-2.74)
Landmark OS analysis

After excluding 26 patients who died or were censored prior to the 24-month timepoint, 304 patients remained eligible for the landmark OS analysis. Multivariable Cox regression suggested that late recurrence independently predicted superior post-landmark survival in the entire cohort (HR = 0.38, 95%CI: 0.27-0.53, P < 0.001) and across all failure subgroups: Local (HR = 0.45, 95%CI: 0.29-0.72, P < 0.001), regional (HR = 0.44, 95%CI: 0.22-0.88, P = 0.021), and locoregional recurrence (HR = 0.11, 95%CI: 0.04-0.26, P < 0.001) (Supplementary Table 1). Relevant survival comparisons are exhibited in Supplementary Figure 2.

DISCUSSION

Based on 330 recurrent NPC patients, we observed that the peak recurrence period was in the second year after the completion of primary treatment. The time to recurrence of 24 months was identified as the optimal cutoff for differentiating early from late recurrence. Patients experiencing early recurrence were prone to harbor more advanced T, N and TNM stage of disease. Early recurrence demonstrated markedly worse post-recurrence OS in the cohort with locoregional recurrence whereas no significantly inferior survival was found in patients with early recurrence in exclusive local or regional site.

In the present study, we first illustrated the chronological distribution of various recurrence patterns in NPC. The first 2 year was revealed to be the peak period of local and regional recurrence, which was consistent with corresponding reports from endemic centers[9-11] and our previous report regarding the dynamic hazard of progression[12]. In addition, we further unveiled a basically consistent distribution of the failure time-frame among various recurrence pattern. This result validates the paradigm that primary tumor and cervical regions could share a uniform post-treatment monitoring schedule. These findings also highlight the importance of dynamic risk-adapted surveillance strategies. For instance, intensified follow-up during the first 2-3 years is highly recommended to timely detect recurrence and provide salvage therapies to mitigate the devastating effect of tumor relapse on OS.

To understand the predisposing factors associated with the site-specific recurrence, we compared the clinical characteristics between different recurrence sites. Patients suffering from local recurrence tended to have more advanced initial T stage, while those experiencing regional recurrence had more unfavorable initial N stage. Additionally, we also analyzed the predictors that distinguish early and late recurrence. Patients with more advanced T stage or overall TN stage were more likely to suffer from early recurrence and ETR. More advanced N stage was more frequently found in entire early recurrence cohort whereas no such trend was observed in other three sub-cohorts. This indicates that the ETR may result from the initial primary tumor burden. On the contrary, LNR may be more related to historical treatment modalities or the host factors, such as immune status or genetic polymorphisms[13], while less related to the gross nodal burden.

Although studies variably defined “early” recurrence in NPC, the cutoff of 24 months was most frequently rendered[9,14,15]. However, most studies identified the early recurrence as ≤ 24 months mainly on the basis of arbitrary or empirical definition. In our study, we attempted to identify the optimal cutoff timing by using a more rational approach, the maximally selected rank method, to maximize the post-recurrence OS difference. Coincidentally, it was still the 24-month that was considered the most suitable cutoff to distinguish early and late recurrence. The underlying mechanisms of early or late recurrences remain inconclusive. Biologically, the early progression indicated resistance to RT and chemotherapy, inadequate intensity of therapy, or accelerated re-population after initial treatment[16]. Instead, decreased anti-tumor immune response or enhanced tumor stem cell properties may, in part, explain the occurrence of late relapse[17-20].

By rendering 24-month as the cutoff for defining early and late recurrence, several studies consistently reported significantly inferior OS in patients who had early recurrence within 2 years after treatment[9,14,15]. In general, the reported OS for early recurrence was around 30%-40% and that for late recurrence was about 40%-70%. However, when survival is calculated from initial diagnosis, this observation is mathematically confounded by time bias, as late-relapsing patients are guaranteed to survive through their longer recurrence-free latency period. To diminish the guarantee-time bias, we conducted a complementary 24-month landmark OS analysis and revealed that late recurrence was independently related with superior post-landmark survival across all failure patterns. These results supported the hypothesis that late recurrence may be biologically less aggressive and could achieve more favorable survival outcome compared with those with early relapse.

Nevertheless, post-recurrence OS represents a clinically more meaningful assessment since it could better guide salvage decision-making. From the perspective of biological behavior, early recurrent tumors’ resistance to initial RT or chemotherapy indicated reduced sensitivity to subsequent treatments, further resulting in worse prognosis. Moreover, the early recurrence may also lead to more challenges in salvage management, which may consequently translate to unfavorable post-recurrence survival. Therefore, it is plausible to expect a more adverse post-recurrence OS for patients experiencing early recurrence. However, few studies compared the post-recurrence survival between the patients with early and late recurrence, presenting conflicting results in a limited number of studies. Based on cohorts from endemic regions, Li et al[9] demonstrated almost equivalent post-recurrence OS between patients who experienced recurrence within and after 24 months, regardless of local recurrence, regional recurrence or locoregional recurrence. Nevertheless, another study, also from an endemic region, reported significantly reduced post-recurrence OS in patients who had recurrence within 24 months[10]. In the current study, we didn’t observe post-recurrence OS difference between early and late recurrence in terms of certain POF, including overall local or regional recurrence, pure local recurrence as well as pure regional recurrence. However, for patients with both local and regional relapse, early recurrence was significantly associated with worse post-recurrence OS, which has also been ascertained as an independent adverse factor by the multivariable analysis.

This intriguing disparity may be explained by underlying tumor biology, particularly intratumor heterogeneity and clonal evolution[21,22]. Simultaneous locoregional recurrence represents a highly aggressive phenotype characterized by widespread spatial dissemination and multiclonal seeding. Under the intense selection pressure of definitive chemoradiotherapy, early relapsing tumors likely emerge through the rapid clonal selection of intrinsically radioresistant and chemoresistant subclones, resulting in cross-resistance to subsequent salvage therapies[22]. Conversely, late relapses occurring after 24 months may arise from dormant, less aggressive subclones with slower evolutionary kinetics, thereby retaining greater therapeutic vulnerability to salvage surgery or re-irradiation[21]. The integration of immunotherapy may provide insights into the salvage treatment for this tough subset of patients.

Current study has several limitations. Firstly, its retrospective design introduces unavoidable inherent biases. Secondly, some prognostic factors, such as EBV-DNA and LDH, were missing since the involved patients were treated across a three-decade time span. Thirdly, all patients in our study were diagnosed with recurrent disease before 2020, when immunotherapy was not yet widely available for recurrent or metastatic NPC. Despite these limitations, the findings of the current study may still offer insights into the investigation on recurrent NPC treated with immunotherapy.

CONCLUSION

In conclusion, the peak recurrence period of NPC was within the first two years after the completion of primary treatment. Our study further identified 24-month as a landmark to distinguish early and late recurrence. Early locoregional recurrence within 24 months was indicative of worse post-recurrence OS, warranting more intensive and efficacious management for this subset. This study enhances the understanding of the significance of recurrence timing and highlights its prognostic impact, shedding light on the improvement of individualized surveillance and salvage therapy for recurrent NPC.

ACKNOWLEDGEMENTS

We thank all the patients and their families for their contribution to this study.

References
1.  Lai SZ, Li WF, Chen L, Luo W, Chen YY, Liu LZ, Sun Y, Lin AH, Liu MZ, Ma J. How does intensity-modulated radiotherapy versus conventional two-dimensional radiotherapy influence the treatment results in nasopharyngeal carcinoma patients? Int J Radiat Oncol Biol Phys. 2011;80:661-668.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Cited by in Crossref: 436]  [Cited by in RCA: 538]  [Article Influence: 35.9]  [Reference Citation Analysis (1)]
2.  Wu LR, Liu YT, Jiang N, Fan YX, Wen J, Huang SF, Guo WJ, Bian XH, Wang FJ, Li F, Song D, Wu JF, Jiang XS, Liu JY, He X. Ten-year survival outcomes for patients with nasopharyngeal carcinoma receiving intensity-modulated radiotherapy: An analysis of 614 patients from a single center. Oral Oncol. 2017;69:26-32.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Cited by in Crossref: 80]  [Cited by in RCA: 88]  [Article Influence: 9.8]  [Reference Citation Analysis (0)]
3.  Lee AWM, Ng WT, Chan JYW, Corry J, Mäkitie A, Mendenhall WM, Rinaldo A, Rodrigo JP, Saba NF, Strojan P, Suárez C, Vermorken JB, Yom SS, Ferlito A. Management of locally recurrent nasopharyngeal carcinoma. Cancer Treat Rev. 2019;79:101890.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Cited by in Crossref: 90]  [Cited by in RCA: 267]  [Article Influence: 38.1]  [Reference Citation Analysis (0)]
4.  Ng WT, Soong YL, Ahn YC, AlHussain H, Choi HCW, Corry J, Grégoire V, Harrington KJ, Hu CS, Jensen K, Kwong DL, Langendijk JA, Le QT, Lee NY, Lin JC, Lu TX, Mendenhall WM, O'Sullivan B, Ozyar E, Pan JJ, Peters LJ, Poh SS, Rosenthal DI, Sanguineti G, Tao Y, Wee JT, Yom SS, Chua MLK, Lee AWM. International Recommendations on Reirradiation by Intensity Modulated Radiation Therapy for Locally Recurrent Nasopharyngeal Carcinoma. Int J Radiat Oncol Biol Phys. 2021;110:682-695.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Cited by in Crossref: 50]  [Cited by in RCA: 77]  [Article Influence: 15.4]  [Reference Citation Analysis (0)]
5.  Mase S, Shinjo K, Totani H, Katsushima K, Arakawa A, Takahashi S, Lai HC, Lin RI, Chan MWY, Sugiura-Ogasawara M, Kondo Y. ZNF671 DNA methylation as a molecular predictor for the early recurrence of serous ovarian cancer. Cancer Sci. 2019;110:1105-1116.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Full Text (PDF)]  [Cited by in Crossref: 19]  [Cited by in RCA: 32]  [Article Influence: 4.6]  [Reference Citation Analysis (0)]
6.  Lee DY, Abraham J, Ross E, Ridge JA, Lango MN, Liu JC, Bauman JR, Avkshtol V, Galloway TJ. Rapid recurrence in head and neck cancer: Underappreciated problem with poor outcome. Head Neck. 2021;43:212-222.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Cited by in Crossref: 7]  [Cited by in RCA: 16]  [Article Influence: 2.7]  [Reference Citation Analysis (0)]
7.  Tang LL, Chen YP, Mao YP, Wang ZX, Guo R, Chen L, Tian L, Lin AH, Li L, Sun Y, Ma J. Validation of the 8th Edition of the UICC/AJCC Staging System for Nasopharyngeal Carcinoma From Endemic Areas in the Intensity-Modulated Radiotherapy Era. J Natl Compr Canc Netw. 2017;15:913-919.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Cited by in Crossref: 86]  [Cited by in RCA: 155]  [Article Influence: 17.2]  [Reference Citation Analysis (0)]
8.  Lausen B, Schumacher M. Maximally Selected Rank Statistics. Biometrics. 1992;48:73.  [PubMed]  [DOI]  [Full Text]
9.  Li F, Chen FP, Chen YP, Chen Y, He XJ, Huang XD, Zheng ZQ, Zheng WH, Liu X, Sun Y, Zhou GQ. Clinical Characteristics and Prognostic Factors of Early and Late Recurrence After Definitive Radiotherapy for Nasopharyngeal Carcinoma. Front Oncol. 2020;10:1469.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Full Text (PDF)]  [Cited by in Crossref: 3]  [Cited by in RCA: 12]  [Article Influence: 2.0]  [Reference Citation Analysis (0)]
10.  Li Y, Huang Z, Zeng X, Pan Y, Wu L, Wang J, Chen R, Xie Y, Lai J, Lin D, Qiu S. Early recurrence as a pivotal event in nasopharyngeal carcinoma: identifying predictors and key molecular signals for survivors. Head Face Med. 2024;20:55.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Cited by in RCA: 7]  [Reference Citation Analysis (0)]
11.  Zhang LL, Zheng WH, Zhu WJ, Deng QL, Peng JL, Li YY, Sun Y, Lin L. Prognostic models for early and late tumor progression prediction in nasopharyngeal carcinoma: An analysis of 8292 endemic cases. Cancer Med. 2023;12:5384-5396.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Full Text (PDF)]  [Cited by in Crossref: 3]  [Cited by in RCA: 5]  [Article Influence: 1.7]  [Reference Citation Analysis (0)]
12.  Wang J, Huang X, Sun S, Wang K, Qu Y, Chen X, Wu R, Zhang Y, Liu Q, Zhang J, Luo J, Xiao J, Gao L, Xu G, Hu C, Li YX, Yi J. Stage-dependent conditional survival and failure hazard of non-metastatic nasopharyngeal carcinoma after intensity-modulated radiation therapy: Clinical implications for treatment strategies and surveillance. Cancer Med. 2021;10:3613-3621.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Full Text (PDF)]  [Cited by in RCA: 20]  [Reference Citation Analysis (0)]
13.  Chen YP, Chan ATC, Le QT, Blanchard P, Sun Y, Ma J. Nasopharyngeal carcinoma. Lancet. 2019;394:64-80.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Cited by in Crossref: 1109]  [Cited by in RCA: 2408]  [Article Influence: 344.0]  [Reference Citation Analysis (4)]
14.  Zhang LL, Xu F, He WT, Huang MY, Song D, Li YY, Deng QL, Huang YS, Wang T, Shao JY. Development and validation of a prognostic nomogram for the pre-treatment prediction of early metachronous metastasis in endemic nasopharyngeal carcinoma: a big-data intelligence platform-based analysis. Ther Adv Med Oncol. 2020;12:1758835920978132.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Full Text (PDF)]  [Cited by in Crossref: 6]  [Cited by in RCA: 8]  [Article Influence: 1.3]  [Reference Citation Analysis (0)]
15.  Chee J, Ting Y, Ong YK, Chao SS, Loh KS, Lim CM. Relapse status as a prognostic factor in patients receiving salvage surgery for recurrent or residual nasopharyngeal cancer after definitive treatment. Head Neck. 2016;38:1393-1400.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Cited by in Crossref: 11]  [Cited by in RCA: 24]  [Article Influence: 2.4]  [Reference Citation Analysis (0)]
16.  Liu Y, Yan W, Qi X, Zhang Y, Wang K, Qu Y, Chen X, Zhang J, Luo J, Li YX, Huang X, Wu R, Wang J, Yi J. Significance of longitudinal Epstein-Barr virus DNA combined with multipoint tumor response for dynamic risk stratification and treatment adaptation in nasopharyngeal carcinoma. Cancer Lett. 2024;605:217276.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Cited by in Crossref: 3]  [Cited by in RCA: 10]  [Article Influence: 5.0]  [Reference Citation Analysis (0)]
17.  Chen Y, Feng Y, Lin Y, Zhou X, Wang L, Zhou Y, Lin K, Cai L. GSTM3 enhances radiosensitivity of nasopharyngeal carcinoma by promoting radiation-induced ferroptosis through USP14/FASN axis and GPX4. Br J Cancer. 2024;130:755-768.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Full Text (PDF)]  [Cited by in Crossref: 20]  [Cited by in RCA: 43]  [Article Influence: 21.5]  [Reference Citation Analysis (0)]
18.  Li JY, Zhao Y, Gong S, Wang MM, Liu X, He QM, Li YQ, Huang SY, Qiao H, Tan XR, Ye ML, Zhu XH, He SW, Li Q, Liang YL, Chen KL, Huang SW, Li QJ, Ma J, Liu N. TRIM21 inhibits irradiation-induced mitochondrial DNA release and impairs antitumour immunity in nasopharyngeal carcinoma tumour models. Nat Commun. 2023;14:865.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Full Text (PDF)]  [Cited by in RCA: 110]  [Reference Citation Analysis (4)]
19.  Lu Y, Liang Y, Zheng X, Deng X, Huang W, Zhang G. EVI1 promotes epithelial-to-mesenchymal transition, cancer stem cell features and chemo-/radioresistance in nasopharyngeal carcinoma. J Exp Clin Cancer Res. 2019;38:82.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Full Text (PDF)]  [Cited by in Crossref: 22]  [Cited by in RCA: 41]  [Article Influence: 5.9]  [Reference Citation Analysis (0)]
20.  Zhou F, Shayan G, Sun S, Huang X, Chen X, Wang K, Qu Y, Wu R, Zhang Y, Liu Q, Zhang J, Luo J, Shi X, Liu Y, Liang B, Li YX, Wang J, Yi J. Spatial architecture of regulatory T-cells correlates with disease progression in patients with nasopharyngeal cancer. Front Immunol. 2022;13:1015283.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Full Text (PDF)]  [Cited by in RCA: 18]  [Reference Citation Analysis (0)]
21.  McGranahan N, Swanton C. Clonal Heterogeneity and Tumor Evolution: Past, Present, and the Future. Cell. 2017;168:613-628.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Cited by in Crossref: 1799]  [Cited by in RCA: 2090]  [Article Influence: 232.2]  [Reference Citation Analysis (5)]
22.  Dagogo-Jack I, Shaw AT. Tumour heterogeneity and resistance to cancer therapies. Nat Rev Clin Oncol. 2018;15:81-94.  [RCA]  [PubMed]  [DOI]  [Full Text]  [Cited by in Crossref: 3236]  [Cited by in RCA: 2733]  [Article Influence: 341.6]  [Reference Citation Analysis (10)]
Footnotes

Peer review: Externally peer reviewed.

Peer-review model: Single blind

Corresponding Author's Membership in Professional Societies: Member of the Society of Radiation Oncology, Chinese Medical Association; Chair of the Biology Group, Society of Radiation Oncology, Chinese Medical Association.

Specialty type: Oncology

Country of origin: China

Peer-review report’s classification

Scientific quality: Grade B, Grade B, Grade C

Novelty: Grade A, Grade B, Grade B

Creativity or innovation: Grade B, Grade B, Grade C

Scientific significance: Grade B, Grade B, Grade C

P-Reviewer: Lu J, Researcher, China; Xu JY, MD, China S-Editor: Qu XL L-Editor: A P-Editor: Wang WB

Write to the Help Desk