General characteristics of the study participants
A total of 633 patients who underwent esophagectomy for any reason at our institution between 2018 and 2021 were screened. Out of the 633 screened patients, 563 had ESCC and 329 of them received nCRT followed by esophagectomy. Among these patients, 3 patients underwent R2 resection and 13 patients underwent R1 resection. The R0 resection rate was 95.14%. Excluding 3 patients lost to follow-up, and 1 patient with incomplete pathological data, a total of 309 patients were included in the analysis of the prognostic classification of histopathologic response for OS. Due to missing information on the recurrence status of 5 patients, a total of 304 patients were included in the RFS analysis. A total of 309 patients were included, with 206 patients showing no recurrence and 103 patients experiencing recurrence. The details of recurrence for 5 patients were unknown. In total, 98 patients were included in the recurrence analysis. (Table 1)
Table 1
Variable | | Number 1 (Percent) (n = 309) | Number 2 (Percent) (n = 304) | Number 3 (Percent) (n = 98) |
Sex | Female | 57(18.4%) | 57(18.8%) | 17(17.3%) |
| Male | 252(81.6%) | 247(81.3%) | 81(82.7%) |
Age, years | < 60 | 117(37.9%) | 116(38.2%) | 40(40.8%) |
| ≥ 60 | 192(62.1%) | 188(61.8%) | 58(59.2%) |
PS | 0 | 227(73.5%) | 222(73.0%) | 71(72.4%) |
| 1 | 78(25.2%) | 78(25.7%) | 26(26.5%) |
| 2 | 4(1.3%) | 4(1.3%) | 1(1.0%) |
NRS | 0 | 198(64.1%) | 195(64.1%) | 61(62.2%) |
| > 0 | 111(35.9%) | 109(35.9%) | 37(37.8%) |
Smoking History | NO | 126(40.8%) | 124(40.8%) | 34(34.7%) |
| YES | 138(44.6%) | 137(45.1%) | 47(48.0%) |
| YES but quit | 45(14.6%) | 43(14.1%) | 17(17.3%) |
Alcohol History | NO | 133(43.0%) | 132(43.4%) | 33(33.7%) |
| YES | 151(48.9%) | 149(49.0%) | 58(59.2%) |
| YES but quit | 25(8.1%) | 23(7.6%) | 7(7.1%) |
Tumor Location | Upper | 23(7.4%) | 23(7.6%) | 9(9.2%) |
| Middle | 193(62.5%) | 190(62.5%) | 58(59.2%) |
| Distal | 93(30.1%) | 91(29.9%) | 31(31.6%) |
Barium_Length, cm | | 5.255 ± 1.827 | 5.244 ± 1.842 | 5.319 ± 1.962 |
Radiotherapy Dose, cGy | ≤ 40Gy | 127(41.1%) | 124(40.8%) | 38(38.8%) |
| > 40Gy | 182(58.9%) | 180(59.2%) | 60(61.2%) |
Chemotherapy Regimen | TP | 266(86.1%) | 262(86.2%) | 80(81.6%) |
| Others | 43(13.9%) | 42(13.8%) | 18(18.4%) |
cT Stage | cT1 | 1(0.3%) | 1(0.3%) | 0(0.0%) |
| cT2 | 43(13.9%) | 42(13.8%) | 9(9.2%) |
| cT3 | 202(65.4%) | 198(65.1%) | 68(69.4%) |
| cT4 | 63(20.4%) | 63(20.7%) | 21(21.4%) |
cN Stage | cN0 | 42(13.6%) | 42(13.8%) | 9(9.2%) |
| cN1 | 125(40.4%) | 123(40.5%) | 40(40.8%) |
| cN2 | 117(37.9%) | 115(37.8%) | 40(40.8%) |
| cN3 | 25(8.1%) | 24(7.9%) | 9(9.2%) |
cM Stage | cM0 | 278(90.0%) | 273(89.8%) | 85(86.7%) |
| cM1 | 31(10.0%) | 31(10.2%) | 13(13.3%) |
cTNM Stage | II | 58(18.8%) | 58(19.1%) | 14(14.3%) |
| III | 156(50.5%) | 152(50.0%) | 49(50.0%) |
| IV | 95(30.7%) | 94(30.9%) | 35(35.7%) |
TRG | 0 | 143(46.3%) | 142(46.7%) | 30(30.6%) |
| 1 | 56(18.1%) | 54(17.8%) | 18(18.4%) |
| 2 | 100(32.4%) | 99(32.6%) | 46(46.9%) |
| 3 | 10(3.2%) | 9(3.0%) | 4(4.1%) |
Lymphovascular invasion | NO | 274(88.7%) | 269(88.5%) | 74(75.5%) |
| YES | 23(7.4%) | 23(7.6%) | 17(17.3%) |
| uncertain | 12(3.9%) | 12(3.9%) | 7(7.1%) |
Perineural invasion | NO | 251(81.2%) | 246(80.9%) | 69(70.4%) |
| YES | 51(16.5%) | 51(16.8%) | 24(24.5%) |
| uncertain | 7(2.3%) | 7(2.3%) | 5(5.1%) |
ypT | 0 | 145(46.9%) | 144(47.4%) | 31(31.6%) |
| 1 | 40(12.9%) | 39(12.8%) | 13(13.3%) |
| 2 | 45(14.6%) | 44(14.5%) | 14(14.3%) |
| 3 | 79(25.6%) | 77(25.3%) | 40(40.8%) |
ypN | 0 | 212(68.6%) | 210(69.1%) | 44(44.9%) |
| 1 | 68(22.0%) | 66(21.7%) | 35(35.7%) |
| 2 | 24(7.8%) | 23(7.6%) | 14(14.3%) |
| 3 | 5(1.6%) | 5(1.6%) | 5(5.1%) |
ypM | 0 | 307(99.4%) | 302(99.3%) | 97(99.0%) |
| 1 | 2(0.6%) | 2(0.7%) | 1(1.0%) |
ypTNM | I | 178(57.6%) | 177(58.2%) | 31(31.6%) |
| II | 34(11.0%) | 33(10.9%) | 13(13.3%) |
| III | 90(29.1%) | 87(28.6%) | 48(49.0%) |
| IV | 7(2.3%) | 7(2.3%) | 6(6.1%) |
ypTN (four categories) | T-N- | 125(40.5%) | 124(40.8%) | 21(21.4%) |
| T + N- | 87(28.2%) | 86(28.3%) | 23(23.5%) |
| T-N+ | 20(6.5%) | 20(6.6%) | 10(10.2%) |
| T + N+ | 77(24.9%) | 74(24.3%) | 44(44.9%) |
ypTN (three categories) | T-N- | 125(40.5%) | 124(40.8%) | 21(21.4%) |
| T + N- or T-N+ | 97(34.6%) | 106(34.9%) | 33(33.7%) |
| T + N+ | 77(24.9%) | 74(24.3%) | 44(44.9%) |
Number 1: the overall patients analyzed for overall survival. Number 2: the overall patients analyzed for recurrence-free survival. Number 3: the recurrent patients analyzed for overall survival.
As shown in Table 1, the baseline characteristics of the enrolled patients in the analysis of OS and RFS of overall population, and OS of the recurrence were listed. In all three scenarios, the characteristics of the patients were similar. The majority of patients were male. Over 60 years old accounted for more than half of the patients, and most of them did not experience cancer-related pain. The most common tumor location was the middle esophagus, followed by distant lesions. The study cohort had a higher prevalence of advanced tumors, with cT3, cN1-2, and stage III-IV accounting for the most cases. Only small proportion of patients had supraclavicular lymph node metastasis. The majority of patients received a radiation dose greater than 40Gy and used the TP chemotherapy regimen throughout the nCRT period. The other chemotherapy regimen varied and included switching to different regimens or receiving alternative treatments.
After nCRT, the patients were classified as TRG0, TRG1, TRG2, or TRG3, with TRG0 and TRG2 being the most prevalent categories. Most of the patients had no lymphovascular and perineural invasion. ypT, ypN, ypM, and ypTNM stage were also judged by pathologists according to the 8th edition of the AJCC TNM classification. The distribution of ypTN categories in the patient population was recorded. ypT-N- and ypT + N + were separately counted. If ypT + N- and ypT-N + were counted together, we defined it as ypTN (three categories). If counted separately, we defined it as ypTN (four categories).
Survival analysis of the study participants
All patients were followed up until July 2023 and median follow-up time was 39.2 months (range, 0.6–70.4). The Kaplan-Meier survival curves according to TRG, ypT, ypN, ypTNM, ypTN (four categories), and ypTN (three categories) were plotted in Fig. 1 and Fig. 2, which displayed the prognostic value of the aforementioned classification methods in discriminating and predicting OS and RFS, respectively. All classification methods showed a phased degradation between each group for both OS and RFS, with statistical significance (P < 0.001).
However, further investigation was warranted to identify the best methods for dividing and forecasting OS and RFS. Regarding TRG classification, there were too few cases in the TRG 3 category, which could lead to a significant estimation error. Additionally, when TRG was used to predict RFS, there was no difference between adjacent groups (TRG 1 vs. TRG 2, TRG 2 vs. TRG 3). The same issues applied to ypN and ypTNM in predicting OS and RFS. There were limited examples for ypN3 and ypTNM IV classification. No difference existed in predicting OS and RFS between ypN1 and ypN2 (P = 0.149 for OS, P = 0.194 for RFS). Similar findings were observed for ypTNM II and ypTNM III (P = 0.166 for OS, P = 0.06 for RFS). Although the number of cases was relatively balanced after classification based on ypT, there was no difference between adjacent groups in predicting OS and RFS. In short, TRG, ypT, ypN, and ypTNM were not ideal tools for discriminating and predicting OS and RFS.
There was a need to find a better tool to discriminate and predict OS and RFS. Combining the primary tumor response (TRG) with ypN status may be a direction. For the 4 categories of the ypTN system, the number of cases belonging to the T-N + category was relatively small. In addition, no difference emerged in adjacent groups in survival curves of both OS and RFS. After combining ypT + N- and ypT-N + categories, which had similar survival outcomes, into a single group (3 categories), we were able to have more balanced distribution of cases and achieve statistically significant distinctions between each group, regardless of OS and RFS. Therefore, we considered ypTN system (three categories) to be a relatively effective method for discriminating and predicting OS and RFS in ESCC patients who received nCRT followed by esophagectomy.
Table 2 and Table 3 summarized the results of univariate and multivariate Cox analysis for OS and RFS based ypTN (three categories) system. Sex, smoking and alcohol histories, chemotherapy regimen, lymphovascular and perineural invasion, and ypTN (three categories) system were significantly associated with OS in the univariate analysis. Similarly, we also verified several variables were significantly associated with RFS in the univariate analysis, including alcohol history, chemotherapy regimen, cN, cM, and cTNM stage, lymphovascular and perineural invasion, and ypTN (three categories) system. However, ypTN (three categories) system was defined as the only prognostic factor for OS and RFS after multivariate Cox analysis.
Table 2
Univariate and multivariate analysis of prognostic factors for overall survival in ESCC patients who received nCRT followed by surgery based on ypTN (three categories).
| | Univariate analysis | Multivariate analysis |
Variable | | HR (95%CI) | p value | HR (95%CI) | p value |
Sex | Female | Reference | | 一 | 一 |
| Male | 1.776 (0.971,3.250) | 0.062 | 一 | 一 |
Age, years | < 60 | Reference | | 一 | 一 |
| ≥ 60 | 0.964 (0.642,1.448) | 0.861 | 一 | 一 |
PS | 0 | Reference | | 一 | 一 |
| 1 | 1.258 (0.811,1.952) | 0.306 | 一 | 一 |
| 2 | 1.073 (0.149,7.740) | 0.944 | 一 | 一 |
NRS | 0 | Reference | | 一 | 一 |
| > 0 | 1.247 (0.834,1.866) | 0.284 | 一 | 一 |
Smoking History | NO | Reference | | 一 | 一 |
| YES | 1.565 (0.994,2.466) | 0.053 | 一 | 一 |
| YES but quit | 1.971 (1.109,3.503) | 0.021 | 一 | 一 |
Alcohol History | NO | Reference | | 一 | 一 |
| YES | 1.919 (1.238,2.975) | 0.004 | 一 | 一 |
| YES but quit | 1.519 (0.696,3.315) | 0.293 | 一 | 一 |
Tumor Site | Upper | Reference | | 一 | 一 |
| Middle | 0.718 (0.356,1.447) | 0.354 | 一 | 一 |
| Distal | 0.708 (0.335,1.497) | 0.366 | 一 | 一 |
Barium_Length, cm | | 1.045 (0.938,1.166) | 0.424 | 一 | 一 |
Radiotherapy Dose, cGy | ≤ 40Gy | Reference | | 一 | 一 |
| > 40Gy | 1.259 (0.834,1.899) | 0.273 | 一 | 一 |
Chemotherapy Regimen | TP | Reference | | 一 | 一 |
| Others | 1.864 (1.139,3.049) | 0.013 | 一 | 一 |
cT Stage | cT1 | Reference | | 一 | 一 |
| cT2 | 790.856 (0,3.017E + 053) | 0.911 | 一 | 一 |
| cT3 | 1258.682 (0,4.796E + 053) | 0.904 | 一 | 一 |
| cT4 | 912.039 (0,3.477E + 053) | 0.909 | 一 | 一 |
cN Stage | cN0 | Reference | | 一 | 一 |
| cN1 | 1.230 (0.627,2.411) | 0.547 | 一 | 一 |
| cN2 | 1.533 (0.786,2.991) | 0.210 | 一 | 一 |
| cN3 | 1.957 (0.830,4.611) | 0.125 | 一 | 一 |
cM Stage | cM0 | Reference | | 一 | 一 |
| cM1 | 1.596 (0.890,2.862) | 0.117 | 一 | 一 |
cTNM Stage | II | Reference | | 一 | 一 |
| III | 1.305 (0.743,2.294) | 0.355 | 一 | 一 |
| IV | 1.418 (0.777,2.586) | 0.255 | 一 | 一 |
Lymphovascular invasion | NO | Reference | | 一 | 一 |
| YES | 4.202 (2.469,7.152) | < 0.001 | 一 | 一 |
| uncertain | 2.979 (1.434,6.189) | 0.003 | 一 | 一 |
Perineural invasion | NO | Reference | | 一 | 一 |
| YES | 2.382 (1.513,3.749) | < 0.001 | 一 | 一 |
| uncertain | 3.354 (1.349,8.334) | 0.009 | 一 | 一 |
ypTN (three categories) | T-N- | Reference | | Reference | |
| T + N- or T-N+ | 2.792 (1.541,5.061) | 0.001 | 2.792 (1.541,5.061) | 0.001 |
| T + N+ | 7.257 (4.112,12.808) | < 0.001 | 7.257 (4.112,12.808) | < 0.001 |
ESCC: esophageal squamous cell carcinoma.
Table 3
Univariate and multivariate analysis of prognostic factors for recurrence-free survival in ESCC patients who received nCRT followed by surgery based on ypTN (three categories).
| | Univariate analysis | Multivariate analysis |
Variable | | HR (95%CI) | p value | HR (95%CI) | p value |
Sex | Female | Reference | | 一 | 一 |
| Male | 1.177 (0.698,1.985) | 0.542 | 一 | 一 |
Age, years | < 60 | Reference | | 一 | 一 |
| ≥ 60 | 0.850 (0.568,1.272) | 0.429 | 一 | 一 |
PS | 0 | Reference | | 一 | 一 |
| 1 | 1.073 (0.684,1.681) | 0.759 | 一 | 一 |
| 2 | 0.817 (0.113,5.887) | 0.841 | 一 | 一 |
NRS | 0 | Reference | | 一 | 一 |
| > 0 | 1.110 (0.738,1.671) | 0.616 | 一 | 一 |
Smoking History | NO | Reference | | 一 | 一 |
| YES | 1.345 (0.865,2.091) | 0.188 | 一 | 一 |
| YES but quit | 1.615 (0.902,2.891) | 0.107 | 一 | 一 |
Alcohol History | NO | Reference | | 一 | 一 |
| YES | 1.762 (1.149,2.702) | 0.009 | 一 | 一 |
| YES but quit | 1.267 (0.560,2.865) | 0.570 | 一 | 一 |
Tumor Site | Upper | Reference | | 一 | 一 |
| Middle | 0.695 (0.344,1.403) | 0.310 | 一 | 一 |
| Distal | 0.830 (0.395,1.744) | 0.623 | 一 | 一 |
Barium_Length, cm | | 1.035 (0.928,1.155) | 0.537 | 一 | 一 |
Radiotherapy Dose, cGy | ≤ 40Gy | Reference | | 一 | 一 |
| > 40Gy | 1.126 (0.750,1.690) | 0.568 | 一 | 一 |
Chemotherapy Regimen | TP | Reference | | 一 | 一 |
| Others | 1.620 (0.971,2.704) | 0.065 | 一 | 一 |
cT Stage | cT1 | Reference | | 一 | 一 |
| cT2 | 704.411 (0,3.903E + 049) | 0.905 | 一 | 一 |
| cT3 | 1209.161 (0,6.689E + 049) | 0.897 | 一 | 一 |
| cT4 | 1115.079 (0,6.172E + 049) | 0.898 | 一 | 一 |
cN Stage | cN0 | Reference | | 一 | 一 |
| cN1 | 1.679 (0.815,3.460) | 0.160 | 一 | 一 |
| cN2 | 1.981 (0.961,4.085) | 0.064 | 一 | 一 |
| cN3 | 2.254 (0.894,5.683) | 0.085 | 一 | 一 |
cM Stage | cM0 | Reference | | 一 | 一 |
| cM1 | 1.635 (0.911,2.933) | 0.099 | 一 | 一 |
cTNM Stage | II | Reference | | 一 | 一 |
| III | 1.484 (0.819,2.690) | 0.193 | 一 | 一 |
| IV | 1.832 (0.985,3.408) | 0.056 | 一 | 一 |
Lymphovascular invasion | NO | Reference | | 一 | 一 |
| YES | 5.250 (3.084,8.936) | < 0.001 | 一 | 一 |
| uncertain | 2.763 (1.270,6.012) | 0.010 | 一 | 一 |
Perineural invasion | NO | Reference | | 一 | 一 |
| YES | 2.271 (1.426,3.617) | 0.001 | 一 | 一 |
| uncertain | 3.267 (1.315,8.120) | 0.011 | 一 | 一 |
ypTN (three categories) | T-N- | Reference | | Reference | |
| T + N- or T-N+ | 2.073 (1.199,3.586) | 0.009 | 2.073 (1.199,3.586) | 0.009 |
| T + N+ | 5.930 (3.518,9.998) | < 0.001 | 5.930 (3.518,9.998) | < 0.001 |
ESCC: esophageal squamous cell carcinoma.
Survival analysis of the recurrence
Out of the 309 patients, 206 did not have a recurrence. Among them, 182 survived, 17 died from esophageal cancer, and 7 died from other diseases. The remaining 103 patients experienced recurrence, with 81 dying as a result of recurrence, 19 still alive, and 3 dying from other causes. This highlights the significantly worse prognosis for recurrent patients compared to those without recurrence, emphasizing the need for a separate analysis of the prognosis for recurrent patients.
Figure 3 showed the Kaplan-Meier survival curves of the recurrent population according to TRG, ypT, ypN, ypTNM, ypTN (four categories), and ypTN (three categories). Different from the overall population, the prognostic analysis based on the aforementioned classification methods may be inconsequential for the recurrent population. The values of P were 0.134 and 0.454 for the classification methods of TRG and ypT, respectively. Though the classification methods of ypN and ypTNM showed statistical significance (P ≤ 0.001), the cases of ypN3 and ypTNM IV classification were limited. Excluding the classification of ypN3 and ypTNM IV, there were no differences observed between the other adjacent groups. The discriminate and predictive ability of ypTN system for the prognosis of the recurrent population was at an intermediate level. The P values were 0.052 for ypTN (four categories) and 0.04 for ypTN (three categories). Adjacent groups in ypTN (three categories) showed no significant difference. In short, all the aforementioned classification methods were not ideal tools for the prognosis of the recurrent population.
As was shown in Fig. 3E, there was almost no difference in the prognosis of the recurrence between the ypT-N- and ypT + N- groups, as well as between the ypT-N + and ypT + N + groups. The corresponding P values were 0.931 and 0.991, respectively. Therefore, we divided the recurrent population into ypN- and ypN + based on the status of ypN, and plotted survival curves according to this classification (Fig. 4). The survival curves for OS clearly differentiated based on the status of ypN. The OS of the ypN- group was much longer than that of the ypN + group (HR: 1.819, 95%CI: 1.150–2.878, P = 0.011).
The results of univariate and multivariate Cox analysis for OS of the recurrent population based on ypN status were displayed in Table 4. The univariate analysis identified five variables might associated with OS, that was tumor site, tumor length, chemotherapy regimen, lymphovascular invasion, and ypN status. After multivariate Cox analysis, tumor site and ypN status retained their statistical significance as independent predictors of OS of the recurrent population. (Table 4)
Table 4
Univariate and multivariate analysis of prognostic factors for overall survival in the recurrent population based on dichotomies of ypN status.
| | Univariate analysis | Multivariate analysis |
Variable | | HR (95%CI) | p value | HR (95%CI) | p value |
Sex | Female | Reference | | 一 | 一 |
| Male | 1.317 (0.695,2.498) | 0.398 | 一 | 一 |
Age, years | < 60 | Reference | | 一 | 一 |
| ≥ 60 | 0.771 (0.490,1.212) | 0.260 | 一 | 一 |
PS | 0 | Reference | | 一 | 一 |
| 1 | 0.789 (0.472,1.319) | 0.366 | 一 | 一 |
| 2 | 2.273 (0.310,16.679) | 0.419 | 一 | 一 |
NRS | 0 | Reference | | 一 | 一 |
| > 0 | 1.254 (0.794,1.980) | 0.332 | 一 | 一 |
Smoking History | NO | Reference | | 一 | 一 |
| YES | 1.380 (0.830,2.292) | 0.214 | 一 | 一 |
| YES but quit | 1.124 (0.570,2.218) | 0.736 | 一 | 一 |
Alcohol History | NO | Reference | | 一 | 一 |
| YES | 0.961 (0.590,1.565) | 0.873 | 一 | 一 |
| YES but quit | 0.923 (0.352,2.417) | 0.923 | 一 | 一 |
Tumor Site | Upper | Reference | | Reference | |
| Middle | 0.641 (0.302,1.362) | 0.247 | 0.741 (0.347,1.584) | 0.440 |
| Distal | 0.352 (0.152,0.815) | 0.015 | 0.388 (0.167,0.899) | 0.027 |
Barium_Length, cm | 1.111 (0.986,1.251) | 0.084 | — | — |
Radiotherapy Dose, cGy | ≤ 40Gy | Reference | | 一 | 一 |
| > 40Gy | 1.074 (0.676,1.708) | 0.761 | 一 | 一 |
Chemotherapy Regimen | TP | Reference | | 一 | 一 |
| Others | 1.639 (0.940,2.857) | 0.081 | 一 | 一 |
cT Stage | cT2 | Reference | | 一 | 一 |
| cT3 | 0.866 (0.394,1.907) | 0.722 | 一 | 一 |
| cT4 | 0.931 (0.381,2.272) | 0.875 | 一 | 一 |
cN Stage | cN0 | Reference | | 一 | 一 |
| cN1 | 1.138 (0.541,2.395) | 0.734 | 一 | 一 |
| cN2 | 0.890 (0.420,1.883) | 0.76 | 一 | 一 |
| cN3 | 1.669 (0.641,4.345) | 0.294 | 一 | 一 |
cM Stage | cM0 | Reference | | 一 | 一 |
| cM1 | 1.161 (0.596,2.260) | 0.66 | — | — |
cTNM Stage | II | Reference | | 一 | 一 |
| III | 0.985 (0.522,1.858) | 0.962 | 一 | 一 |
| IV | 1.181 (0.606,2.299) | 0.625 | 一 | 一 |
Lymphovascular invasion | NO | Reference | | 一 | 一 |
| YES | 1.788 (1.004,3.183) | 0.048 | 一 | 一 |
| uncertain | 1.786 (0.803,3.970) | 0.155 | 一 | 一 |
Perineural invasion | NO | Reference | | 一 | 一 |
| YES | 1.136 (0.670,1.924) | 0.636 | 一 | 一 |
| uncertain | 1.931 (0.765,4.878) | 0.164 | 一 | 一 |
ypN status | N- | Reference | | Reference | |
| N+ | 1.822 (1.151,2.883) | 0.010 | 1.822 (1.151,2.883) | 0.010 |