Systematic oxidative stress indices predicts prognosis in patients with urothelial carcinoma of the upper urinary tract after radical nephroureterectomy

This retrospective analysis evaluated data from patients without evidence of distant metastases who underwent RNU between March 1996 and June 2021 at Beijing Hospital, National Center of Gerontology, Institute of the Geriatric Medicine, Chinese Academy of Medical Sciences, Beijing. Patients who had not undergone RNU, had evidence of metastatic disease at the time of surgery, or had incomplete preoperative medical information on hematologic indicators; patients who were lost to follow-up; patients with autoimmune diseases; and patients who had received preoperative adjuvant chemotherapy, radiotherapy, or any other antitumor therapy were excluded from the study. Finally, we assessed data for a final total of 483 patients who underwent open or laparoscopic RNU.

Postoperative follow-up included routine urine tests, urine pathology chest radiography, computed tomography, and cystoscopy. Patients were generally assessed postoperatively every 3–4 months in the first year after RNU, every 6 months from the second year to the fifth year, and annually thereafter. PFS (measured from RNU until the date of last follow-up or date of disease progression (including local recurrence or distant metastasis or death) and OS (measured from RNU until the date of last follow-up or date of death from any cause) were selected as primary endpoints.

Comparisons between the clinicopathological characteristics of the patients were performed using the chi-square test and Mann–Whitney U test, as appropriate. Seven plasma biomarkers associated with increased oxidative stress, including Fib, CRE, GGT, ALB, LDH, ALP, and BUN, were identified for calculating SOSIs, which were determined by the lowest Akaike information criterion (AIC) value. The value was defined as 1 if the plasma biomarker value was above the cutoff level; the value was defined as 0 if the plasma biomarkers value was below the cutoff level. Survival curves were generated using the Kaplan–Meier method with the log-rank test. Univariate and multivariate Cox proportional hazard models were conducted to evaluate the impact of variables on OS and PFS after RNU. Time-dependent receiver operating characteristic (ROC) analysis and decision curve analysis (DCA) were used to assess the prognostic capacity of the nomogram. A P value lower than 0.05 was considered statistically significant. Data analysis was completed with R software, version 4.2.1.

The clinical and pathologic characteristics of the patients are shown in Table 1. In this cohort, males accounted for 51.1% (247 patients) and females for 48.9% (236 patients), with a median age of 70 years (interquartile range: 62–76). The cutoff values of Fib, CRE, GGT, ALB, BUN, ALP, and LDH derived from the ROC curves were 3.978 g/L, 83.5 μmol/L, 43.5 U/L, 39.5 g/L,9.07 mg/dL, 83.5 UL/L, and 170.5 U/L, respectively. During the median follow-up of 36.8 (interquartile range [IQR]: 22.0–68.5) months, a total of 182 (37.7%) patients died, and 224 (46.4%) had progressive disease.

First, we divided the seven plasma biomarkers into 2 groups using the cutoff value. Values of “0” and “1” were used for scoring according to the algorithm mentioned above. Univariate and multivariate Cox regression analyses were used to explore whether the above seven plasma biomarkers are independent prognostic factors for UTUC. According to univariate Cox regression analysis, we found that Fib, GGT, CRE, BUN, LDH, and ALB correlated significantly with overall survival (OS) (all p < 0.05) (Fig. 1a). Multivariate Cox regression analysis indicated that Fib, GGT, CRE, LDH, and ALB were associated with OS (Fig. 1b). Then, the SOSIs prognostic model was generated based on the lowest AIC value: SOSIs = Fib*0.3141 + GGT*0.6059 + CRE*0.3582 + LDH*0.7149–ALB*0.3600. The cutoff value of SOSIs based on the median value was set as 0.3582 and used to divide the patients into high-risk (272 with SOSIs of 0.3582 or above) and low-risk groups (211 with a SOSIs level less 0.3582). The clinical characteristics of all the patients grouped by SOSIs are presented in Table 2. Patients in the high-risk group had higher incidences of ≥ pT3 stage and pN + . In the high-risk patient group, tumor size, hydronephrosis, and positive urine pathology were significantly higher than those in the low-risk patient group. There was an obvious difference between the two groups in terms of age, tumor site, and surgical approach. However, no significant difference was found for sex, BMI, tumor side, previous ureteroscopic surgical margin status, multifocality, presence of lymphovascular invasion, chemotherapy, lymph node stage, presence of CIS in surgical specimens, or tumor grade.

The Kaplan–Meier curve showed that the patients in the high-risk group had worse overall survival (OS) (Fig. 2a) and progression-free survival (PFS) (Fig. 2b) than those in the low-risk group. We also found that OS was worse in the high-risk group than in the low-risk group in subgroup analysis performed by age (Fig. 2c) and T stage (Fig. 2d). In the size ≤ 5 and pN0&Nx groups, the high-risk group had worse prognosis, but there was no significant difference in the size > 5 and pN + groups between the two groups (Fig. 2e, f). Subgroup analyses suggested a significant difference in PFS between the two risk groups for the two T-stage groups (Fig. 2g). In size ≤ 5 (Fig. 2h), age ≥ 65 (Fig. 2i) and pN0&Nx (Fig. 2j) subgroups, survival analysis showed that patients with low-risk had significantly favorable PFS compared with patients with high-risk.

Univariate and multivariate analyses were performed to determine predictors of OS (Table 3). In univariate analyses,we found that age (HR = 1.681, p = 0.003), tumor site at the ureter (HR = 1.413, p = 0.029), use of ureteroscopy (HR = 1.620, p = 0.014), hydronephrosis (HR = 1.904, p = 0.001), size ≥ 5 (HR = 1.744, p = 0.005), LVI (HR = 3.766, p < 0.001), T stage (T2, HR = 2.507, p < 0.001; T3, HR = 3.404, p < 0.001; T4, HR = 12.446, p < 0.001), margin (HR = 5.062, p < 0.001), N stage (HR = 5.344, p < 0.001), grade (HR = 2.013, p = 0.001), chemotherapy (HR = 2.241, p < 0.001), and SOSIs (HR = 2.217, p < 0.001) were significantly associated with OS. Those significant and potential risk factors were evaluated in multivariate analyses, in which OS (all p < 0.05)was independently predicted by age, LVI, T stage, N stage, margin, and SOSIs.

Interestingly, sex (HR = 1.335, p = 0.030), age (HR = 1.435, p = 0.018), tumor site at the ureter (HR = 1.617, p = 0.001), use of ureteroscopy (HR = 1.489, p = 0.025), hydronephrosis (HR = 1.544, p = 0.010), LVI (HR = 2.467, p < 0.001), T stage (T2, HR = 1.832, p = 0.002; T3, HR = 2.098, p < 0.001; T4, HR = 5.538, p < 0.001), margin (HR = 2.308, p = 0.021), N stage (HR = 4.302, p < 0.001), grade (HR = 1.809, p = 0.001), and SOSI (HR = 1.721, p < 0.001) were all highly significantly different in univariate PFS analysis. Next, age, tumor site at the ureter, LVI, T stage, N stage, and SOSIs were all significantly identified in PFS multivariate analysis (Table 4).

The independent predictors were used to construct a nomogram (Fig. 3). The areas under the curve (AUCs) of the nomogram for predicting 1-year, 3-year, and 5-year survival rates were 0.77, 0.78, and 0.81, respectively (Fig. 4a). The time‐dependent AUC was > 0.7 for prediction of OS within 10 years, indicating favorable discrimination by the nomogram (Fig. 4b). As shown in Fig. 4c, calibration curves indicated good agreement between the predicted and observed probabilities. Furthermore, in decision curve analysis, the nomogram consistently achieved a greater net benefit than traditional prognostic indicators (Fig. 4d). Collectively, these results indicate that the nomogram has good concordance and accuracy.