3.1 Clinicopathologic and follow‑up data
In total, 730 patients from the SEER database and 79 patients from the China database were eventually enrolled in this study. Patients from SEER were randomly divided into a training set (n = 514) and an internal validation set (n = 216). The cut-off value of age and tumor size in the training set was 68 years and 6.8cm, respectively. Predominantly, patients in the SEER set were ≤ 68 years old and male, consistent with our external validation set. Histologically, pRCC accounted for more than half in both the SEER set (54.9%) and the external validation set (68.4%). Notably, chRCC with TT exhibited the highest survival rate, while Bellini RCC demonstrated the poorest (Fig. 1A-B). The T3b stage was noted for the majority of patients (46% and 40.5%) in both databases. Inferior vena cava TT was observed in 22.1% of the SEER set and 60.8% in the external validation set, possibly due to the latter's more advanced stage (N1: 38% vs. 28.2%; M1: 27.8% vs. 20.3%). Patients who received systemic therapy after surgery were 18.4% in the SEER set and 36.7% in the external validation set. NccRCC-TT tumors generally exhibited size > 6.8 cm and a high Fuhrman grade in both sets. Detailed characteristics are provided in Table 1.
Table 1
Demographic and clinical characteristics of patients with nccRCC-TT.
Variables | SEER set (N = 730) | Training (N = 514) | Internal validation (N = 216) | P-value | External validation set (N = 79) | P-value |
---|
Age (years) | | | | | | |
≤ 68 | 485 (66.4%) | 352 (68.5%) | 133 (61.6%) | 0.086 | 70 (88.6%) | < .001 |
> 68 | 245 (33.6%) | 162 (31.5%) | 83 (38.4%) | | 9 (11.4%) | |
Sex | | | | | | |
Female | 243 (33.3%) | 168 (32.7%) | 75 (34.7%) | 0.655 | 29 (36.7%) | 0.627 |
Male | 487 (66.7%) | 346 (67.3%) | 141 (65.3%) | | 50 (63.3%) | |
Race | | | | | | |
Caucasian | 530 (72.6%) | 375 (73%) | 155 (71.8%) | 0.893 | 0 (0%) | < .001 |
African American | 151 (20.7%) | 104 (20.2%) | 47 (21.8%) | | 0 (0%) | |
Asian and Others | 49 (6.7%) | 35 (6.8%) | 14 (6.5%) | | 79 (100%) | |
Laterality | | | | | | |
Left | 390 (53.4%) | 269 (52.3%) | 121 (56%) | 0.407 | 42 (53.2%) | 1.000 |
Right | 340 (46.6%) | 245 (47.7%) | 95 (44%) | | 37 (46.8%) | |
Histology | | | | | | |
pRCC | 401 (54.9%) | 292 (56.8%) | 109 (50.5%) | 0.479 | 54 (68.4%) | < .001 |
chRCC | 249 (34.1%) | 168 (32.7%) | 81 (37.5%) | | 2 (2.5%) | |
Bellini RCC | 52 (7.1%) | 35 (6.8%) | 17 (7.9%) | | 12 (15.2%) | |
Others | 28 (3.8%) | 19 (3.7%) | 9 (4.2%) | | 11 (13.9%) | |
T stage | | | | | | |
3a | 312 (42.7%) | 215 (41.8%) | 97 (44.9%) | 0.045 | 27 (34.2%) | .006 |
3b | 336 (46%) | 245 (47.7%) | 91 (42.1%) | | 32 (40.5%) | |
3c | 35 (4.8%) | 28 (5.4%) | 7 (3.2%) | | 10 (12.7%) | |
4 | 47 (6.4%) | 26 (5.1%) | 21 (9.7%) | | 10 (12.7%) | |
Thrombus level | | | | | | |
Renal vein | 569 (77.9%) | 401 (78%) | 168 (77.8%) | 1.000 | 31 (39.2%) | < .001 |
Inferior vena cava | 161 (22.1%) | 113 (22%) | 48 (22.2%) | | 48 (60.8%) | |
N stage | | | | | | |
N0/Nx | 524 (71.8%) | 368 (71.6%) | 156 (72.2%) | 0.935 | 49 (62%) | 0.093 |
N1 | 206 (28.2%) | 146 (28.4%) | 60 (27.8%) | | 30 (38%) | |
M stage | | | | | | |
M0 | 582 (79.7%) | 409 (79.6%) | 173 (80.1%) | 0.953 | 57 (72.2%) | 0.154 |
M1 | 148 (20.3%) | 105 (20.4%) | 43 (19.9%) | | 22 (27.8%) | |
Tumour size (cm) | | | | | | |
≤ 6.8 | 226 (31%) | 156 (30.4%) | 70 (32.4%) | 0.645 | 27 (34.2%) | 0.647 |
> 6.8 | 504 (69%) | 358 (69.6%) | 146 (67.6%) | | 52 (65.8%) | |
Systemic therapy | | | | | | |
No | 596 (81.6%) | 416 (80.9%) | 180 (83.3%) | 0.509 | 50 (63.3%) | < .001 |
Yes | 134 (18.4%) | 98 (19.1%) | 36 (16.7%) | | 29 (36.7%) | |
Renal sinus/Perirenal fat invasion | | | | | | |
No | 203 (27.8%) | 136 (26.5%) | 67 (31%) | 0.244 | 30 (38%) | 0.078 |
Yes | 527 (72.2%) | 378 (73.5%) | 149 (69%) | | 49 (62%) | |
Sarcomatoid/Rhabdoid feature | | | | | | |
No | 640 (87.7%) | 447 (87%) | 193 (89.4%) | 0.440 | 55 (69.6%) | < .001 |
Yes | 90 (12.3%) | 67 (13%) | 23 (10.6%) | | 24 (30.4%) | |
Fuhrman grade | | | | | | |
Low (1–2) | 51 (7%) | 40 (7.8%) | 11 (5.1%) | 0.406 | 2 (2.5%) | < .001 |
High (3–4) | 351 (48.1%) | 247 (48.1%) | 104 (48.1%) | | 75 (94.9%) | |
NA | 328 (44.9%) | 227 (44.2%) | 101 (46.8%) | | 2 (2.5%) | |
OS month (Median (95%CI)) | 46.0 (37.8–54.2) | 45.0 (36.1–53.9) | 49.0 (28.8–69.2) | 0.764 | 30.0 (18.5–41.5) | 0.236 |
CSS month (Median (95%CI)) | 61.0 (37.1–84.9) | 59.0 (29.7–88.3) | 73.0 (51.3–94.7) | 0.739 | - | |
Follow-up month OS | 51.0 (45.8–56.2) | 51.0 (43.9–58.1) | 47.0 (37.5–56.5) | 0.201 | 28.0 (21.6–34.4) | < .001 |
Follow-up month CSS | 44.0 (39.1–48.9) | 44.0 (38.1–49.9) | 41.0 (34.2–47.8) | 0.236 | - | |
NOTE. Data are presented as No. (%) unless indicated otherwise. |
IQR, interquartile range; OS, Overall Survival; CSS, Cancer Specific Survival; NA, not available. |
pRCC, papillary renal cell carcinoma; chRCC, chromophobe renal cell carcinoma; Bellini RCC, Bellini renal cell carcinoma/collecting duct carcinoma; Others, Unclassified carcinoma. |
P values were calculated by Chi-square test for categorical variables and Cochran-Mantel-Haenszel(CMH) Chi-square test for ordinal variables. |
The overall cohort has been randomly divided into a training (70%) and a validation (30%) cohort. |
After a median follow-up time of 51 months (95%CI, 45.8–48.9) in the SEER set and 28 months (95%CI, 21.6–34.4) in the external validation set, the median OS time was 46 months (95%CI, 37.8–54.2) and 30 months (95%CI, 18.5–41.5), respectively. The median CSS in the SEER set, after a median follow-up of 44 months (95% CI, 39.1–48.9), was 61 months (95% CI, 37.1–84.9). At the end of follow-up, the 5-year OS rate was 43.1% in the SEER set and 27.4% in the external validation set. The 5-year CSS rate was 50.4% in the SEER set.
3.2 Construction of the nomogram
In the training set, independent risk factors for OS were obtained: age > 68 (HR: 1.88, 95%CI, 1.45–2.45, P < 0.001), laterality (HR: 0.74, 95%CI, 0.57–0.96, P = 0.026), histology (Bellini RCC vs. pRCC; HR: 2.10, 95%CI, 1.38–3.19, P = 0.001), N1 stage (HR: 1.76, 95%CI, 1.28–2.41, P < 0.001), M1 stage (HR: 2.44, 95%CI, 1.78–3.34, P < 0.001), renal sinus/perirenal fat invasion (HR: 1.44, 95%CI, 1.04-2.00, P = 0.027), and sarcomatoid/rhabdoid feature (HR: 1.72, 95%CI, 1.23–2.40, P = 0.001) (Table 2). Similarly, multivariate analysis results suggested that age > 68 (HR: 1.61, 95%CI, 1.19–2.18, P = 0.002), laterality (HR: 0.69, 95%CI, 0.52–0.93, P = 0.014), histology (Bellini RCC vs. pRCC; HR: 2.39, 95%CI, 1.53–3.72, P < 0.001), N1 stage (HR: 1.92, 95%CI, 1.36–2.72, P < 0.001), M1 stage (HR: 2.73, 95%CI, 1.94–3.86, P < 0.001), renal sinus/perirenal fat invasion (HR: 1.50, 95%CI, 1.02–2.19, P < 0.038), and sarcomatoid/rhabdoid feature (HR: 1.97, 95%CI, 1.39–2.79, P < 0.001) were independent risk factors for CSS (Table 2). Nomograms predicting OS and CSS risk for patients at 1-, 3-, or 5-years were then respectively constructed based on identified variables (Fig. 2A-B).
Table 2
Univariate and multivariate Cox analyses on variables for the prediction of OS and CSS of nccRCC-TT patients.
Variables | Overall survival | Cancer-specific survival |
---|
Univariable | Multivariable | Univariable | Multivariable |
---|
HR (95% CI) | P-value | HR (95%CI) | P-value | HR (95% CI) | P-value | HR (95%CI) | P-value |
---|
Age (years) | | | | | | | | |
≤ 68 | Ref. | | Ref. | | Ref. | | Ref. | |
> 68 | 1.76(1.36–2.28) | < .001 | 1.88(1.45–2.45) | < .001 | 1.46(1.09–1.95) | 0.001 | 1.61(1.19–2.18) | 0.002 |
Sex | | | | | | | | |
Female | Ref. | | | | Ref. | | | |
Male | 1.19(0.90–1.56) | 0.219 | | | 1.29(0.95–1.74) | 0.106 | | |
Race | | | | | | | | |
Caucasian | Ref. | | | | Ref. | | Ref. | |
African American | 1.23(0.91–1.66) | 0.183 | | | 1.39(1.01–1.93) | 0.045 | 1.17(0.82–1.65) | 0.381 |
Asian and Others | 1.11(0.67–1.82) | 0.694 | | | 1.32(0.79–2.22) | 0.288 | 1.17(0.69-2.00) | 0.560 |
Laterality | | | | | | | | |
Left | Ref. | | Ref. | | Ref. | | Ref. | |
Right | 0.75(0.58–0.97) | 0.030 | 0.74(0.57–0.96) | 0.026 | 0.70(0.53–0.93) | 0.015 | 0.69(0.52–0.93) | 0.014 |
Histology | | | | | | | | |
pRCC | Ref. | | Ref. | | Ref. | | Ref. | |
chRCC | 0.41(0.29–0.57) | < .001 | 0.71(0.43–1.18) | 0.185 | 0.35(0.24–0.51) | < .001 | 0.86(0.46–1.61) | 0.644 |
Bellini RCC | 2.17(1.44–3.26) | < .001 | 2.10(1.38–3.19) | 0.001 | 2.39(1.56–3.67) | < .001 | 2.39(1.53–3.72) | < .001 |
Others | | 0.835 | 0.97(0.44–2.12) | 0.932 | 1.10(0.51–2.35) | 0.813 | 1.27(0.57–2.83) | 0.563 |
Thrombus level | | | | | | | | |
Renal vein | Ref. | | Ref. | | Ref. | | Ref. | |
Inferior vena cava | 1.62(1.22–2.16) | < .001 | 1.03(0.76–1.39) | 0.864 | 1.73(1.27–2.36) | < .001 | 1.02(0.73–1.42) | 0.925 |
N stage | | | | | | | | |
N0/Nx | Ref. | | Ref. | | Ref. | | Ref. | |
N1 | 2.83(2.18–3.67) | < .001 | 1.76(1.28–2.41) | < .001 | 3.39(2.56–4.50) | < .001 | 1.92(1.36–2.72) | < .001 |
M stage | | | | | | | | |
M0 | Ref. | | Ref. | | Ref. | | Ref. | |
M1 | 3.12(2.37–4.11) | < .001 | 2.44(1.78–3.34) | < .001 | 3.68(2.74–4.93) | < .001 | 2.73(1.94–3.86) | < .001 |
Tumour size (cm) | | | | | | | | |
≤ 6.8 | Ref. | | Ref. | | Ref. | | Ref. | |
> 6.8 | 1.51(1.13–2.03) | 0.006 | 1.32(0.97–1.78) | 0.076 | 1.53(1.10–2.12) | 0.011 | 1.23(0.87–1.73) | 0.243 |
Systemic therapy | | | | | | | | |
No | Ref. | | Ref. | | Ref. | | Ref. | |
Yes | 1.93(1.45–2.57) | < .001 | 0.77(0.54–1.09) | 0.141 | 2.23(1.64–3.02) | < .001 | 0.78(0.53–1.14) | 0.196 |
Renal sinus/Perirenal fat invasion | | | | | | | | |
No | Ref. | | Ref. | | Ref. | | Ref. | |
Yes | 1.91(1.40–2.61) | < .001 | 1.44(1.04-2.00) | 0.027 | 2.17(1.51–3.11) | < .001 | 1.50(1.02–2.19) | 0.038 |
Sarcomatoid/Rhabdoid feature | | | | | | | | |
No | Ref. | | Ref. | | Ref. | | Ref. | |
Yes | 2.44(1.78–3.35) | < .001 | 1.72(1.23–2.40) | 0.001 | 3.06(2.21–4.26) | < .001 | 1.97(1.39–2.79) | < .001 |
Fuhrman grade | | | | | | | | |
Low (1–2) | Ref. | | Ref. | | Ref. | | Ref. | |
High (3–4) | 1.99(1.24–3.21) | 0.005 | 1.41(0.87–2.30) | 0.167 | 1.95(1.18–3.24) | 0.010 | 1.31(0.78–2.22) | 0.306 |
NA | 0.79(0.48–1.32) | 0.370 | 0.96(0.52–1.78) | 0.908 | 0.62(0.35–1.07) | 0.085 | 0.66(0.33–1.33) | 0.244 |
Abbreviations: HR, hazard ratio; CI, confidential interval; NA, not available; Ref, Reference. |
pRCC, papillary renal cell carcinoma; chRCC, chromophobe renal cell carcinoma; Bellini RCC, Bellini renal cell carcinoma/collecting duct carcinoma; Others, Unclassified carcinoma. |
3.3 Validation of the nomogram
The calibration plots in the training, internal validation, and external validation sets for 1-, 3-, and 5-year OS and CSS were depicted in Fig. 3A-E. The nomogram exhibited a C-index for OS prediction of 0.734 (95%CI, 0.703–0.765) in the training set, 0.774 (95%CI, 0.727–0.821) in the internal validation set, and 0.705 (95%CI, 0.600-0.811) in the external validation set. The AUCs for 1-, 3- and 5-year OS prediction were 0.766, 0.787, and 0.818 in the training set, 0.814, 0.866, and 0.815 in the internal validation set, 0.774, 0.668 and 0.601 in the external validation set, respectively (Fig. 4A-C). While the TNM staging system for OS prediction exhibited a lower C-index: 0.675 (95%CI, 0.640–0.710) in the training set, 0.713 (95%CI, 0.660–0.766) in the internal validation set, and 0.649 (95%CI, 0.541–0.757) in the external validation set. Furthermore, the time-dependent AUC curves for OS, including nomogram and stage, demonstrated the nomogram's superior predictive value compared to the commonly used risk factor TNM stage (Supplementary Fig. 1. A-C).
Simultaneously, the C-index of the nomogram for CSS prediction surpassed that of the TNM staging system, with 0.759 (95%CI, 0.726–0.792) vs. 0.709 (95%CI, 0.672–0.746) in the training set and 0.787 (95%CI, 0.736–0.838) vs. 0.730 (95%CI, 0.673–0.787) in the internal validation set. The AUCs for 1-, 3- and 5-year CSS prediction were 0.793, 0.807, and 0.838 in the training set and 0.825, 0.870, and 0.824 in the internal validation set (Fig. 4D-E). Additionally, the time-dependent AUC curves for the CSS nomogram were displayed in Supplementary Fig. 1. D-E, futher indicated the nomogram's stability and discriminability compared to the TNM staging system.
The DCA plots for 1-, 3-, and 5-year rates of OS and CSS in the training and validation sets illustrated that our nomogram for nccRCC-TT achieved positive net clinical benefits across a broad range of threshold probabilities, emphasizing its high clinical utility (Supplementary Fig. 2).
3.4 Risk stratification
Following the acquisition of risk scores for each of the 730 SEER dataset patients through the nomogram, patients with nccRCC-TT were categorized into low-, moderate-, and high-risk groups for OS prediction, with cut-off points at 96 and 193. The Kaplan-Meier OS curves demonstrated significant discrimination among these groups (P < 0.001) (Fig. 5A). The median OS for high-risk, moderate-risk, and low-risk groups were 8 months (95%CI, 0.44–0.65), 24 months (95%CI, 0.46–0.59), and 117 months (95%CI, 0.39–0.62), respectively. The 3-year OS rates for these groups were 6.0% (95%CI, 0.02–0.15), 39.5% (95%CI, 0.33–0.47), and 79.6% (95%CI, 0.75–0.84), respectively.
Similarly, the cut-off points of the nomogram for CSS prediction were 99 and 189. As shown in Fig. 5B, the risk of specific death from nccRCC-TT significantly increased with rising risk levels (P < 0.001). Median CSS for high-risk, moderate-risk, and low-risk groups were 9 months (95%CI, 0.40–0.63), 29 months (95%CI, 0.44–0.58), and not reached (95%CI, NA-NA), respectively. The 3-year CSS rates for these groups were 5.1% (95%CI, 0.02–0.15), 44.5% (95%CI, 0.38–0.52), and 84.1% (95%CI, 0.80–0.89), respectively. The median time, 1-, 3-, and 5-year rates for the three risk groups of OS and CSS were summarized in Table 3.
Table 3
Risk stratification for the nomogram of OS and CSS.
Risk group | | Median time | 1-year rate | 3-year rate | 5-year rate |
---|
| High risk | 8 months (95%CI 0.44–0.65) | 38.3% (95%CI 0.30–0.50) | 6.0% (95%CI 0.02–0.15) | not reached (95%CI NA-NA) |
OS | Moderate risk | 24 months (95%CI 0.46–0.59) | 71.2% (95%CI 0.66–0.77) | 39.5% (95%CI 0.33–0.47) | 25.0% (95%CI 0.19–0.33) |
| Low risk | 117 months (95%CI 0.39–0.62) | 92.2% (95%CI 0.89–0.95) | 79.6% (95%CI 0.75–0.84) | 67.9% (95%CI 0.62–0.74) |
| High risk | 9 months (95%CI 0.40–0.63) | 37.0% (95%CI 0.28–0.49) | 5.1% (95%CI 0.02–0.15) | not reached (95%CI NA-NA) |
CSS | Moderate risk | 29 months (95%CI 0.44–0.58) | 74.0% (95%CI 0.69–0.80) | 44.5% (95%CI 0.38–0.52) | 30.5% (95%CI 0.24–0.39) |
| Low risk | not reached (95%CI NA-NA) | 94.5% (95%CI 0.92–0.97) | 84.1% (95%CI 0.80–0.89) | 76.4% (95%CI 0.71–0.82) |
OS, overall survival; CSS, cancer-specific survival; NA, not available. |
Additionally, two dedicated online tools for OS (https://njmumh1997.shinyapps.io/os-prediction/) and CSS (https://njmumh1997.shinyapps.io/css-prediction/) were built to conveniently calculate nomogram scores, determine risk groups, and predict survival for each patient.