Results

Clinical characteristics

Between January 2017 and December 2019, 2,059 patients met the study’s inclusion criteria. The median age of the cohort was 49 years (IQR, 41–56 years), including 1504 males and 555 females. The median follow-up duration was 55.57 months (range: 44.83–65.57 months), with 86.79% of patients being monitored for a period exceeding three years. The median of RTT was 46 days (IQR, 45–49 days). Clinical characteristics were stratified by median RTT into two groups: those with RTT longer than 46 days and those with RTT of 46 days or less (see Table 1). Patients exhibiting advanced TNM stages (either advanced T or overall stage) were more prone to experience extended RTT (p < 0.001 for each comparison). Moreover, patients who received 32–36 fractions tended to have a longer RTT (p < 0.001). In addition, significant differences were also observed in treatment approaches (p < 0.001). Another factor demonstrating significant variation was HGB levels (p < 0.05 for all; refer to Table 1).

Characteristic | No. (%) of patients by RTT in days | Total | P
≤46(n = 896) | >46(n = 1163) | (N = 2059)
Gender | – | – | – | 0.147
Male | 640(71.43) | 864(74.29) | 1504(73.05) | –
Female | 256(28.57) | 299(25.71) | 555(26.95) | –
Age, year | – | – | – | 0.766
≤45 | 337(37.61) | 430(36.97) | 767(37.25) | –
>45 | 559(62.39) | 733(63.03) | 1292(62.75) | –
T stage (8th edition) | – | – | – | < 0.001
T1 | 209(23.33) | 169(14.53) | 378(18.36) | –
T2 | 188(20.98) | 137(11.78) | 325(15.78) | –
T3 | 367(40.96) | 506(43.51) | 873(42.40) | –
T4 | 132(14.73) | 351(30.18) | 483(23.46) | –
N stage (8th edition) | – | – | – | 0.394
N0 | 74(8.26) | 87(7.48) | 161(7.82) | –
N1 | 347(38.73) | 432(37.15) | 779(37.83) | –
N2 | 293(32.70) | 422(36.29) | 715(34.73) | –
N3 | 182(20.31) | 222(19.09) | 404(19.62) | –
Overall stage (8th edition) | – | – | – | < 0.001
I | 28(3.13) | 23(1.98) | 51(2.48) | –
II | 144(16.07) | 78(6.71) | 222(10.78) | –
III | 420(46.88) | 518(44.54) | 938(45.56) | –
IV | 304(33.93) | 544(46.78) | 848(41.19) | –
EBV DNA titer, copies | – | – | – | 0.094
negative | 624(69.64) | 849(73.00) | 1473(71.54) | –
positive | 272(30.36) | 314(27.00) | 586(28.46) | –
HGB, g/L | – | – | – | < 0.05
<113 | 224(25.00) | 304(26.14) | 528(25.64) | –
113–151 | 619(69.08) | 822(70.68) | 1441(69.99) | –
>151 | 53(5.92) | 37(3.18) | 90(4.37) | –
Prescribed fractions | – | – | – | < 0.001
28–31 | 254(28.35) | 51( 4.39) | 305( 14.81) | –
32–36 | 642(71.65) | 1112(95.61) | 1754(85.19) | –
Treatment modality | – | – | – | < 0.001
RT alone | 53(5.92) | 52(4.47) | 105(5.10) | –
IC + RT | 185(20.65) | 308(26.48) | 493(23.94) | –
IC + RT + AC | 29(3.24) | 23(1.98) | 52(2.53) | –
CCRT | 81(9.04) | 60(5.16) | 141(6.85) | –
CCRT+AC | 14(1.56) | 11(0.95) | 25(1.21) | –
IC + CCRT | 469(52.34) | 637(54.77) | 1106(53.72) | –
IC+CCRT+AC | 65(7.25) | 72(6.19) | 137(6.65) | –
Residual or not | – | – | – | 0.651
No | 732(81.70) | 941(80.91) | 1673(81.25) | –
Yes | 164(18.30) | 222(19.09) | 386(18.75) | –
NLR | – | – | – | 0.913
Normal | 457(51.00) | 596(51.25) | 1053(51.14) | –
Abnormal | 439(49.00) | 567(48.75) | 1006(48.86) | –
Number of IC sessions | – | – | – | < 0.001
0 | 148(16.52) | 123(10.58) | 271(13.16) | –
≤3 | 448(50.00) | 603(51.85) | 1052(51.04) | –
>3 | 300(33.48) | 437(37.58) | 737(35.79) | –

Improving RTT and survival

Patients with an RTT of 46 days or less had a five-year OS rate of 87.03%, while those with an RTT exceeding 46 days had a rate of 82.69%. Similarly, the five-year progression-free survival (PFS) rate was 86.94% for RTT ≤ 46 days, versus 82.52% for RTT > 46 days. The Kaplan-Meier curve analysis indicated that patients with RTT > 46 days had worse outcomes in OS, PFS, LRFFS, and DMFFS compared to those with RTT ≤ 46 days (p < 0.01, Figure 1; Figure S1). Univariate analyses revealed that an RTT exceeding 46 days was a significant predictor of an increased mortality risk and disease progression risk compared to an RTT of 46 days or less (p < 0.01, Tables S1 and S2).

Figure 1: Kaplan-Meier’s plots showed(A) OS, (B) PFS divided by the RTT of ≤46 days, and >46 days, respectively. Abbreviations: RTT, radiotherapy treatment time; OS, overall survival; PFS, progression-free survival.

To determine the optimal RTT threshold for predicting significant survival differences, RPA was conducted. The optimal threshold identified was 50 days. When RTT was adjusted to this threshold, patients with RTT ≥ 50 days exhibited significantly worse outcomes in OS, PFS, LRFFS, and DMFFS compared to those with RTT < 50 days, as reflected in the Kaplan-Meier curves (Figure 2; Figure S2).

Figure 2: Kaplan-Meier survival curves illustrated(A) OS, (B) PFS categorized by the RTT of < 50 days and ≥50 days. Abbreviations: RTT, radiotherapy treatment time; OS, overall survival; PFS, progression-free survival.

Survival of patients stratified by ΔRTT

The cohort was divided into three groups based on the timely completion of RT: (1) the planned completion group (ΔRTT = 0 days), (2) the early completion group (ΔRTT < 0 days), and (3) the delayed completion group (ΔRTT > 0 days). Subsequent analysis revealed that there were no notable differences in the factors related to the host and treatment among the three groups, with the exception of overall stage, treatment modality, and use of IC (all p < 0.05; Table 2). The planned completion group had a significantly higher proportion of patients receiving the IC + CCRT regimen (p < 0.001) and those with advanced overall stage disease (p < 0.05) compared to the other groups (Table 2). Although Kaplan-Meier curves for OS, PFS, LRFFS, and DMFFS did not show statistical significance among the ΔRTT groups (Figure 3; Figure S3), the planned completion group demonstrated numerically higher cumulative survival rates at equivalent time points compared to other groups across all four endpoints.

 | No. (%) of patients by △RTT status | Total | P
the planned completion group (n = 1036) | the early completion group (n = 131) | the delayed completion group (n = 892) | (N = 2059)
Gender | – | – | – | – | 0.51
Male | 758(73.17) | 101(77.10) | 645(72.31) | 1504(73.05) | –
Female | 278(26.83) | 30(22.90) | 247(27.69) | 555(26.95) | –
Age, year | – | – | – | – | 0.795
≤45 | 392(37.84) | 46(35.11) | 329(36.88) | 767(37.25) | –
>45 | 644(62.16) | 85(64.89) | 563(63.12) | 1292(62.75) | –
T stage (8th edition) | – | – | – | – | 0.13
T1 | 174(16.80) | 20(15.27) | 184(20.63) | 378(18.36) | –
T2 | 169(16.31) | 25(19.08) | 131(14.69) | 325(15.78) | –
T3 | 461(44.50) | 51(38.93) | 361(40.47) | 873(42.40) | –
T4 | 232(22.39) | 35(26.72) | 216(24.22) | 483(23.46) | –
N stage (8th edition) | – | – | – | – | 0.856
N0 | 81(7.82) | 9(6.87) | 71(7.96) | 161(7.82) | –
N1 | 382(36.87) | 55(41.98) | 342(38.34) | 779(37.83) | –
N2 | 373(36.00) | 44(33.59) | 298(33.40) | 715(34.73) | –
N3 | 200(19.31) | 23(17.56) | 181(20.29) | 404(19.62) | –
Overall stage (8th edition) | – | – | – | – | < 0.05
I | 22(2.12) | 1(0.76) | 28(3.14) | 51(2.48) | –
II | 105(10.14) | 24(18.32) | 93(10.43) | 222(10.78) | –
III | 496(47.88) | 52(39.69) | 390(43.72) | 938(45.56) | –
IV | 413(39.86) | 54(41.22) | 381(42.71) | 848(41.19) | –
EBV DNA titer, copies | – | – | – | – | 0.878
negative | 736(71.04) | 94(71.76) | 643(72.09) | 1473(71.54) | –
positive | 300(28.96) | 37(28.24) | 249(27.91) | 586(28.46) | –
HGB, g/L | – | – | – | – | 0.919
<113 | 271(26.16) | 34(25.95) | 223(25.00) | 528(25.64) | –
113–151 | 723(69.79) | 90(68.70) | 628(70.40) | 1441(69.99) | –
>151 | 42(4.05) | 7(5.34) | 41(4.60) | 90(4.37) | –
Prescribed fractions | – | – | – | – | 0.422
28–31 | 143(13.80) | 20(15.27) | 142(15.92) | 305(14.81) | –
32–36 | 893(86.20) | 111(84.73) | 750(84.08) | 1754(85.19) | –
Treatment modality | – | – | – | – | < 0.001
RT alone | 54(5.21) | 6(4.58) | 45(5.04) | 105(5.10) | –
IC + RT | 201(19.40) | 30(22.90) | 262(29.37) | 493(23.94) | –
IC + RT + AC | 23(2.22) | 7(5.34) | 22(2.47) | 52(2.53) | –
CCRT | 54(5.21) | 14(10.69) | 73(8.18) | 141(6.85) | –
CCRT+AC | 13(1.25) | 1(0.76) | 11(1.23) | 25(1.21) | –
IC + CCRT | 612(59.07) | 65(49.62) | 429(48.09) | 1106(53.72) | –
IC+CCRT+AC | 79(7.63) | 8(6.11) | 50(5.61) | 137(6.65) | –
Residual or not | – | – | – | – | 0.931
No | 840(81.08) | 108(82.44) | 725(81.28) | 1673(81.25) | –
Yes | 196(18.92) | 23(17.56) | 167(18.72) | 386(18.75) | –
NLR | – | – | – | – | 0.455
Normal | 540(52.12) | 61(46.56) | 452(50.67) | 1053(51.14) | –
Abnormal | 496(47.88) | 70(53.43) | 440(49.33) | 1006(48.86) | –
Number of IC sessions | – | – | – | – | < 0.05
0 | 121(11.68) | 21(16.03) | 129(14.46) | 271(13.16) | –
≤3 | 563(54.34) | 60(45.80) | 428(47.98) | 1051(51.04) | –
>3 | 352(33.98) | 50(38.17) | 335(37.56) | 737(35.79) | –

Figure 3: Survival outcomes of patients with NPC stratified by RTT completion statusComparison of (A) OS and (B) PFS probability between the planned completion group, the early completion group, and the delayed completion group. Abbreviations: Delayed group: the delayed completion group; planned group: the planned completion group; early group: the early completion group; OS, overall survival; PFS, progression-free survival.

Consequently, RPA was utilized to determine the optimal ΔRTT threshold for predicting substantial differences in survival rates. The optimal cutoff for OS-related ΔRTT was determined to be 7 days. Accordingly, the cohort was divided into high ΔRTT (≥7 days) and low ΔRTT (<7 days) groups for survival analysis. Kaplan-Meier curves demonstrated that patients with ΔRTT ≥ 7 days had worse outcomes in OS, PFS, LRFFS, and DMFFS compared to those with ΔRTT<7 days (p < 0.05, Figure 4; Figure S4). Multivariate analysis, accounting for various prognostic factors, confirmed ΔRTT as an independent adverse prognostic factor for OS, PFS, LRFFS, and DMFFS (all p < 0.05; Tables 3 and 4; Tables S3 and S4).

Figure 4: Kaplan-Meier survival curves for ΔRTT stratified by 7 days(A) OS, (B) PFS. Abbreviations: RTT, radiotherapy treatment time; ΔRTT, the deviation in RTT; OS, overall survival; PFS, progression-free survival.

Characteristic | Univariate analysis | Multivariate analysis
HR (95% CI) | P | HR (95% CI) | P
Gender
Female | Ref. | – | Ref. | –
Male | 1.534 (1.148–2.050) | 0.004 | 1.450 (1.077–1.951) | 0.014
Age, year
>45 | Ref. | – | Ref. | –
≤45 | 0.591 (0.456–0.767) | < 0.001 | 0.614 (0.470–0.802) | < 0.001
Overall stage (8th edition)
I | Ref. | – | Ref. | –
II | 1.376 (0.308–6.146) | 0.676 | 2.072 (0.419–10.255) | 0.372
III | 2.475 (0.609–10.056) | 0.205 | 7.980 (1.783–35.722) | 0.007
IV | 6.530 (1.621–26.302) | 0.008 | 19.915 (4.423–89.670) | < 0.001
EBV DNA titer, copies
negative | Ref. | – | Ref. | –
positive | 2.129 (1.687–2.686) | < 0.001 | 1.906 (1.492–2.434) | < 0.001
HGB, g/L
<113 | Ref. | – | Ref. | –
113-151 | 0.866 (0.669–1.121) | 0.275 | 0.869 (0.661–1.142) | 0.313
>151 | 0.318 (0.129–0.785) | 0.013 | 0.555 (0.211–1.460) | 0.233
Prescribed fractions
28-31 | Ref. | – | Ref. | –
32-36 | 1.668 (1.136–2.448) | 0.009 | 1.024 (0.679–1.546) | 0.909
Treatment modality
RT alone | Ref. | – | Ref. | –
IC + RT | 0.845 (0.511–1.399) | 0.513 | 0.265 (0.149–0.473) | < 0.001
IC + RT + AC | 1.506 (0.744–3.049) | 0.255 | 0.363 (0.169–0.783) | 0.010
CCRT | 0.438 (0.213–0.903) | 0.025 | 0.552 (0.250–1.222) | 0.143
CCRT+AC | 0.414 (0.097–1.779) | 0.236 | 0.801 (0.168–3.832) | 0.781
IC + CCRT | 0.692 (0.429–1.118) | 0.133 | 0.224 (0.130–0.385) | < 0.001
IC+CCRT+AC | 1.078 (0.602–1.931) | 0.800 | 0.308 (0.163–0.584) | < 0.001
Residual or not
Yes | Ref. | – | – | –
No | 0.977 (0.729–1.308) | 0.874 | – | –
NLR
Abnormal | Ref. | – | – | –
Normal | 0.777 (0.616–0.979) | 0.033 | – | –
Number of IC sessions
0 | Ref. | – | – | –
≤3 | 1.192 (0.817–1.740) | 0.363 | – | –
>3 | 1.394 (0.944–2.057) | 0.095 | – | –
ΔRTT, days
<7 | Ref. | – | Ref. | –
≥7 | 1.842 (1.169–2.902) | 0.008 | 1.879 (1.189–2.969) | 0.007
Interruption times
0 | Ref. | – | – | –
1 | 0.894 (0.681–1.172) | 0.417 | – | –
>1 | 1.165 (0.864–1.572) | 0.316 | – | –

Characteristic | Univariate analysis | Multivariate analysis
HR (95% CI) | P | HR (95% CI) | P
Gender
Female | Ref. | – | Ref. | –
Male | 1.554 (1.163–2.077) | 0.003 | 1.449 (1.076–1.951) | 0.015
Age, year
>45 | Ref. | – | Ref. | –
≤45 | 0.621 (0.479–0.806) | < 0.001 | 0.658 (0.504–0.859) | 0.002
Overall stage (8th edition)
I | Ref. | – | Ref. | –
II | 1.385 (0.310–6.187) | 0.670 | 1.997 (0.405–9.852) | 0.396
III | 2.526 (0.622–10.265) | 0.195 | 7.147 (1.597–31.971) | 0.010
IV | 6.820 (1.693–27.471) | 0.007 | 18.075 (4.018–81.301) | < 0.001
EBV DNA titer, copies
negative | Ref. | – | Ref. | –
positive | 2.231 (1.768–2.815) | < 0.001 | 1.988 (1.556–2.540) | < 0.001
HGB, g/L
<113 | Ref. | – | Ref. | –
113-151 | 0.899 (0.695–1.163) | 0.417 | 0.891 (0.678–1.171) | 0.409
>151 | 0.321 (0.130–0.792) | 0.014 | 0.555 (0.211–1.459) | 0.233
Prescribed fractions
28-31 | Ref. | – | Ref. | –
32-36 | 1.688 (1.150–2.479) | 0.007 | 1.003 (0.664–1.514) | 0.990
Treatment modality
RT alone | Ref. | – | Ref. | –
IC + RT | 0.877 (0.530–1.452) | 0.610 | 0.299 (0.167–0.532) | < 0.001
IC + RT + AC | 1.650 (0.815–3.341) | 0.164 | 0.428 (0.198–0.923) | 0.031
CCRT | 0.452 (0.219–0.930) | 0.031 | 0.571 (0.260–1.258) | 0.165
CCRT+AC | 0.421 (0.098–1.809) | 0.245 | 0.811 (0.170–3.875) | 0.793
IC + CCRT | 0.732 (0.453–1.182) | 0.202 | 0.261 (0.152–0.449) | < 0.001
IC+CCRT+AC | 1.246 (0.696–2.232) | 0.459 | 0.385 (0.204–0.728) | 0.003
Residual or not
Yes | Ref. | – | – | –
No | 0.943 (0.705–1.263) | 0.695 | – | –
NLR
Abnormal | Ref. | – | Ref. | –
Normal | 0.778 (0.617–0.982) | 0.034 | 0.820 (0.649–1.036) | 0.096
Number of IC sessions
0 | Ref. | – | – | –
≤3 | 1.256 (0.861–1.833) | 0.237 | – | –
>3 | 1.427 (0.967–2.106) | 0.073 | – | –
ΔRTT, days
<7 | Ref. | – | Ref. | –
≥7 | 1.767 (1.122–2.783) | 0.014 | 1.791 (1.133–2.831) | 0.013
Interruption times
1 | Ref. | – | – | –
0 | 1.141 (0.869–1.497) | 0.343 | – | –
>1 | 1.285 (0.921–1.791) | 0.140 | – | –

Univariate and multivariate analysis

Univariate analysis, detailed in Tables 3 and 4 and Tables S3 and S4, revealed significant differences in OS, PFS, LRFFS, and DMFFS based on factors such as age, sex, overall stage, EBV-DNA (pre-treatment EBV-DNA) levels, HGB (pre-treatment HGB) levels, radiation fractions, treatment modality, NLR, and ΔRTT (p < 0.05). The interruption count reflects the number of treatment suspensions during RT, with 0 times indicating uninterrupted treatment completion, 1 time representing a single interruption event during the course, and >1 times denoting multiple interruptions. Multivariate analysis identified sex, age, stage, treatment modality, EBV-DNA levels, and ΔRTT as independent prognostic variables for OS, PFS, LRFFS, and DMFFS (p < 0.05, Tables 3 and 4; Tables S3 and S4). According to the RPA results, the grouping criteria for ΔRTT were defined as: ΔRTT < 7 days vs. ≥7 days. Patients with ΔRTT ≥ 7 days had a 1.879-fold higher risk of mortality (95% CI 1.189–2.969, p = 0.007) and a 1.791-fold higher risk of disease progression (95% CI 1.133–2.831, p = 0.013) compared to those with ΔRTT < 7 days.

Interaction of ΔRTT with other covariates

Based on RPA results, patients were stratified into ΔRTT < 7 days and ≥ 7 days groups. As shown in Figure S5, the results demonstrate a significant interaction between ΔRTT and interruption frequency. Subgroup analyses demonstrated significantly improved OS in low-risk versus high-risk patients among females, those aged > 45 years, receiving 28–31 RT fractions, with normal NLR levels, undergoing zero cycles of IC, and experiencing a single treatment interruption (Figure S5).

Analysis of survival effects of continuous interruption

Patients were stratified into two cohorts based on the frequency of treatment interruptions: a single interruption group and a multiple interruptions group. An analysis of Kaplan-Meier curves indicated that the OS of patients with ΔRTT > 0 exhibited a statistically significant difference (p < 0.05; Figure 5). Subsequent subgroup analyses were conducted to evaluate the impact of the number of interruptions across various parameters. The results shown in Figure 6 demonstrate a significant interaction between ΔRTT and interruption frequency. The correlation between the number of interruptions and OS remained consistent across stage IV, patients receiving IC ≤ 3, and those without residual focus (p < 0.05, Figure 6).

Figure 5: Kaplan-Meier curves for different populations according to the number of radiotherapy interruptions (interruptions of 1 and more than 1)(A) Total population.(B) Population with ΔRTT = 0 days.(C) Population with ΔRTT <0 days.(D) Population with ΔRTT >0 days. Abbreviations: RTT, radiotherapy treatment time; ΔRTT, the deviation in RTT.

Figure 6: Subgroup analysis according to number of radiotherapy interruptions (1 and more than1interruption)NE, not estimable because of zero events or complete separation.