Abstract
Objective
To summarize the clinicopathological characteristics and prognostic factors of papillary renal cell carcinoma (pRCC) and to construct clinical and molecular prognostic nomograms using existing databases.
Methods
Clinical prognostic models were developed using the Surveillance, Epidemiology, and End Results (SEER) database, while molecular prognostic models were constructed using The Cancer Genome Atlas (TCGA) database. Cox regression and LASSO regression were employed to identify clinicopathological features and molecular markers related to prognosis. The accuracy of the prognostic models was assessed using ROC curves, C-index, decision curve analysis (DCA) curves, and calibration plots.
Results
In the 2004–2015 SEER cohort, Cox regression analysis revealed that age, grade, AJCC stage, N stage, M stage, and surgery were independent predictors of overall survival (OS) and cancer-specific survival (CSS) in pRCC patients. ROC curves, C-index, and DCA curves indicated that the prognostic nomogram based on clinical independent predictors had better predictive ability than TNM staging and SEER staging. Additionally, in the TCGA cohort, M stage, clinical stage, and the molecular markers IDO1 and PLK1 were identified as independent risk factors. The prognostic nomogram based on molecular independent risk factors effectively predicted the 3-year and 5-year OS and CSS for pRCC patients.
Conclusions
The clinical and molecular nomograms constructed in this study provide robust predictive tools for individualized prognosis in pRCC patients, offering better accuracy than traditional staging systems.
Supplementary Information
The online version contains supplementary material available at 10.1007/s12672-024-01669-8.
Keywords: Papillary renal cell carcinoma, Nomogram, SEER, TCGA
Introduction
Papillary renal cell carcinoma (pRCC), originating from renal tubular epithelial cells, was first described by Mancilla-Jimenez et al. in 1976 [1]. It is a low-grade malignancy, accounting for approximately 15% of all kidney cancer [2]. Traditionally, pRCC was classified into type 1 and type 2 based on cytomorphology. Type 1 tumors consist of single layered small cells, single-layered cells with scant cytoplasm and inconspicuous nucleoli, while type 2 tumors are characterized by larger eosinophilic cells with nuclear pseudostratification [3–5]. It was generally believed that the pathological stage, grade and prognosis of type 1 are better than those of type 2 [6]. However, in clinical practice, tumor tissues often exhibit a mixture of both type 1 and type 2 areas [7, 8]. Furthermore, the latest WHO classification suggests that these may not represent distinct cancer types, but rather a progression from low- to high-grade tumors [6, 9, 10]. Notably, some studies have found that after the adjustment of stage and other variables, there is no significant difference in prognosis between type 1 and type 2 pRCC [11, 12]. pRCC is recognized as a highly heterogeneous disease. Advances in pathology and molecular biology have led to the identification of new tumor entities with papillary features, such as biphasic squamous-alveolar RCC, biphasic hyalinising psammomatous RCC, papillary renal neoplasm with reversed polarity, and Warthin-like pRCC [13–16]. Therefore, it is not convincing to determine the prognosis only by the conventional subtypes of pRCC.
Currently, surgery is the preferred treatment for patients with early-stage pRCC [17]. Despite limited efficacy, cytotoxic chemotherapy and immunotherapy have been widely adopted for advanced-stage patients previously [18]. Historically, the AJCC staging system has been considered the gold standard for predicting survival and prognosis in RCC [19]. However, there are significant differences in survival outcomes among different types of renal tumors [20], and no consensus has been reached regarding the prognostic factors specific to pRCC [11, 21]. In recent years, with in-depth exploration in gene expression profiling and protein microarrays, a series of mutated genes associated with pRCC and biomarkers with potential prognostic value have been identified by taking advantage of sequencing technology. For example, Haake SM et al. [22] found that CDKN2A mutation or promoter hypermethylation was closely associated with aggressive pRCC. However, these biomarkers are not yet widely applied in personalized diagnosis and treatment. In this context, the present study aims to clarify the prognostic factors for pRCC by analyzing data from two large databases. We analyzed the clinicopathological features and survival outcomes, identified genes related to the prognosis of pRCC, and construct and validate clinical and molecular prognostic models through bioinformatics method.
Materials and methods
Data source
Data utilized in this study was extracted from the Surveillance, Epidemiology, and End Results (SEER) and The Cancer Genome Atlas (TCGA) databases. SEER is one of the most representative large-scale oncology registry databases in North America, covering about 28% of the total population in the United States, and it provides relatively extensive and complete cancer data [23]. SEER*Stat 8.3.9 software (https://seer.cancer.gov/data/) was used to retrieve and screen the relevant information of patients with pathologically confirmed pRCC from 2004 to 2015. Funded by National Cancer Institute (NCI) and National Human Genome Research Institute (NHGRI), the TCGA database integrates information of clinical cases, genomic data and biospecimens on a global scale. Transcriptome profiling and prognosis data of pRCC patients were downloaded from TCGA database (https://portal.gdc.cancer.gov/), and the clinical information was matched with genomic information.
Patients
A total of 9,388 patients diagnosed with pRCC between 2004 and 2015 were included according to the International Classification of Diseases for Oncology (ICD-O-3) histology code 8260. The exclusion criteria were as follows: (1) not one primary tumor only (n = 6,472); (2) age < 18 years (n = 15); (3) unknown AJCC stage (n = 371); (4) unknown TNM stage (n = 142); (5) unknown survival months (n = 52); (6) unknown if surgery performed (n = 1) (Figure S1). For the TCGA cohort, the inclusion criteria were: (i) primary tumor site: kidney; (ii) type of disease: papillary carcinoma; (iii) data types: transcriptome data and clinical data. SEER and TCGA are publicly available databases, and informed consent and institutional review board approval were not required.
Variables and endpoints
The following variables were screened from SEER database: gender, age, marital status, race, grade, AJCC stage, TNM stage, SEER stage and treatment regimen (partial nephrectomy or total nephrectomy or no). TCGA database focused on gene expression profiling data, as well as definitions and information on survival time, gender, age, race, TNM stage, clinical stage. OS and CSS were the endpoints of interest in the study. OS was defined as duration from the date of diagnosis to the date of death from any cause or data exclusion. CSS was defined as the duration from the date of diagnosis to the date of death due to pRCC. The optimal cut-off age was determined by X-tile software.
Statistical methods
Data were expressed as percentages for categorical variables. The samples from SEER database were randomly assigned to training cohort (n = 6,571) and validation cohort (n = 2,817) in a 7:3 ratio, and descriptive analyses of demographic, clinical and pathological baseline characteristics of eligible patients were performed. Univariate and multivariate Cox regression analyses identified the independent prognostic factors, and the hazard ratios (HR) and 95% confidence intervals (CI) were calculated. Based on the multivariate Cox regression analysis, a nomogram incorporating all significant variables was constructed to assess the individual prognosis of pRCC patients. The points in the nomogram represent individual scores, reflecting the corresponding score for each variable at different values. Each variable was scored by nomogram and the value was added to obtain the total score, then the 3- and 5-year OS and CSS could be obtained. We also compared the differences in survival prediction between nomogram, TNM stage and SEER stage.
The receiver operating characteristic (ROC) curves, C-index and decision curve analysis (DCA) curves were applied to validate and evaluate the predictive performance of the models. Calibration curves were then employed to estimate the agreement between predictions and observations. ROC was used to assess the overall predictive accuracy of the model, and the area under the curve (AUC) was used for quantification. The C-index measures the model's ability to rank predictions correctly over time and is similar to AUC, but appears more applicable to censored data. The C-index ranges between 0.5 (no discrimination) and 1 (perfect discrimination), with higher values indicating better prognostic models. DCA is a novel method to evaluate the net benefit of a predictive model from a clinical outcomes perspective, assessing whether the model is practically useful in clinical decision-making. DCA analysis determined the clinical utility of the model by quantifying the net benefit at different threshold probabilities in dataset. The nomogram was internally and externally validated in both the training and validation cohorts.
LASSO is a linear regression method, which has the advantages of stability and simplicity in screening of variables, effectiveness in dealing with high-dimensional data of small sample and reliability in solving the problem of the multicollinearity. LASSO regression was used to select prognosis-related genes, and Kaplan–Meier curves were plotted for survival analysis. X-tile software (version 3.6.1, Yale University) is commonly used to determine the optimal cutoff value for continuous variables. It performs statistical tests by grouping the data with different values as cutoff points, and the result with the smallest p-value is considered the optimal cutoff value. SPSS 26.0 and R software (version 3.6.1) were used for statistical analysis and graph drawing. All p values were two-sided, and p < 0.05 was considered statistically significant.
Result
Baseline demographic and clinical characteristics of the SEER cohort
Based on the inclusion and exclusion criteria, a total of 9,388 pRCC patients were finally enrolled from the SEER database. The training cohort (n = 6,571) was utilized to construct the clinical nomogram, while the remaining 2,817 patients formed the validation cohort for external validation. Table 1 demonstrated the demographic and clinicopathological characteristics of the two cohorts. Of the enrolled patients, 6,962 (74.2%) were male and 2,426 (25.8%) were female. X-tile analysis determined the optimal age cut-off points as < 65 years, 65–74 years and > 74 years (Figure S2). The majority of patients were diagnosed at Grade II (41.9%), AJCC stage I (73%), T1 stage (74.7%), N0 stage (94.2%) and M0 stage (95.2%). Over 90% of pRCC patients underwent surgical treatment.
Table 1.
Baseline demographic and clinical characteristics of patients with papillary renal cell carcinoma in the SEER database
| Variables | All patients n (%) | Training set | Validation set |
|---|---|---|---|
| n (%) | n (%) | ||
| Total | 9389 | 6572 (70.0) | 2817 (30.0) |
| Gender | |||
| Male | 6963 (74.2) | 4877 (74.2) | 2086 (74.1) |
| Female | 2426 (25.8) | 1695 (25.8) | 731 (25.9) |
| Age, years | |||
| < 65 | 5690 (60.6) | 3951 (60.1) | 1739 (61.7) |
| 65–74 | 2444 (26.0) | 1717 (26.1) | 727 (25.8) |
| > 74 | 1255 (13.4) | 904 (13.8) | 351 (12.5) |
| Marital status | |||
| Married | 5599 (59.6) | 3931 (59.8) | 1668 (59.2) |
| Unmarried | 3324 (35.4) | 2302 (35.0) | 1022 (36.3) |
| Unknown | 466 (5.0) | 339 (5.2) | 127 (4.5) |
| Race | |||
| White | 6408 (68.3) | 4516 (68.7) | 1892 (67.2) |
| Black | 2557 (27.2) | 1756 (26.7) | 801 (28.4) |
| Others | 424 (4.5) | 300 (4.6) | 124 (4.4) |
| Grade | |||
| Grade I | 911 (9.7) | 619 (9.4) | 292 (10.4) |
| Grade II | 3932 (41.9) | 2779 (42.3) | 1153 (40.9) |
| Grade III | 2454 (26.1) | 1715 (26.1) | 739 (26.2) |
| Grade IV | 270 (2.9) | 191 (2.9) | 79 (2.8) |
| Unknown | 1822 (19.4) | 1268 (19.3) | 554 (19.7) |
| AJCC stage | |||
| I | 6858 (73.0) | 4801 (73.1) | 2057 (73.0) |
| II | 1016 (10.8) | 714 (10.9) | 302 (10.7) |
| III | 923 (9.8) | 636 (9.7) | 287 (10.2) |
| IV | 592 (6.3) | 421 (6.4) | 171 (6.1) |
| T stage | |||
| T1 | 7014 (74.7) | 4909 (74.7) | 2105 (74.7) |
| T2 | 1153 (12.3) | 807 (12.3) | 346 (12.3) |
| T3 | 1148 (12.2) | 797 (12.1) | 351 (12.5) |
| T4 | 74 (0.8) | 59 (0.9) | 15 (0.5) |
| N stage | |||
| N0 | 8847 (94.2) | 6185 (94.1) | 2662 (94.5) |
| N1 | 294 (3.1) | 214 (3.3) | 80 (2.8) |
| N2 | 248 (2.6) | 173 (2.6) | 75 (2.7) |
| M stage | |||
| M0 | 8941 (95.2) | 6249 (95.1) | 2692 (95.6) |
| M1 | 448 (4.8) | 323 (4.9) | 125 (4.4) |
| SEER stage | |||
| Localized | 7874 (83.9) | 5515 (83.9) | 2359 (83.7) |
| Regional | 1061 (11.3) | 729 (11.1) | 332 (11.8) |
| Distant | 454 (4.8) | 328 (5.0) | 126 (4.5) |
| Surgery | |||
| No/Unknown | 380 (4.0) | 266 (4.0) | 114 (4.0) |
| Partial nephrectomy | 4290 (45.7) | 3012 (45.8) | 1278 (45.4) |
| Total nephrectomy | 4719 (50.3) | 3294 (50.1) | 1425 (50.6) |
| Chemotherapy | |||
| Yes | 339 (3.6) | 248 (3.8) | 91 (3.2) |
| No/Unknown | 9050 (96.4) | 6324 (96.2) | 2726 (96.8) |
| Radiotherapy | |||
| Yes | 121 (1.3) | 83 (1.3) | 38 (1.3) |
| No/Unknown | 9268 (98.7) | 6489 (98.7) | 2779 (98.7) |
AJCC American Joint Committee on Cancer, SEER Surveillance, Epidemiology, and End Results
Univariate and multivariate Cox regression analysis of OS and CSS in SEER cohort
Univariate and multivariate Cox regression analyses were performed to assess the impact of clinicopathological parameters on OS and CSS. As shown in Table 2, multivariate Cox regression model identified gender, age, marital status, race, grade, AJCC stage, T stage, N stage, M stage, and surgery as independent predictors of OS, while age, grade, AJCC stage, N stage, M stage and surgery were independent prognostic factors of CSS.
Table 2.
Multivariate analysis of overall survival (OS) and cancer-specific survival (CSS) rates of patients with papillary renal cell carcinoma in the training set
| Variables | Overall survivala | Cancer-specific survivalb | ||
|---|---|---|---|---|
| HR (95% CI) | P value | HR (95% CI) | P value | |
| Gender | ||||
| Male | Reference | |||
| Female | 0.773 (0.683–0.874) | < 0.001 | ||
| Age, years | ||||
| < 65 | Reference | Reference | ||
| 65–74 | 1.709 (1.508–1.936) | < 0.001 | 1.286 (1.066–1.550) | 0.008 |
| > 74 | 2.852 (2.491–3.266) | < 0.001 | 1.442 (1.164–1.787) | 0.001 |
| Marital status | ||||
| Married | Reference | Reference | ||
| Unmarried | 1.252 (1.119–1.400) | < 0.001 | – | 0.484 |
| Unknown | 1.076 (0.820–1.411) | 0.599 | – | 0.909 |
| Race | ||||
| White | Reference | |||
| Black | 1.240 (1.097–1.401) | 0.001 | ||
| Others | 0.860 (0.663–1.114) | 0.253 | ||
| Grade | ||||
| Grade I | Reference | Reference | ||
| Grade II | 1.013 (0.823–1.245) | 0.905 | 1.550 (1.003–2.395) | 0.048 |
| Grade III | 1.179 (0.952–1.460) | 0.131 | 2.400 (1.564–3.682) | < 0.001 |
| Grade IV | 1.851 (1.393–2.460) | < 0.001 | 3.684 (2.301–5.898) | < 0.001 |
| Unknown | 1.198 (0.963–1.491) | 0.105 | 1.873 (1.219–2.877) | 0.004 |
| AJCC stage | ||||
| I | Reference | Reference | ||
| II | 0.797 (0.549–1.156) | 0.232 | 2.749 (2.014–3.753) | < 0.001 |
| III | 1.243 (0.905–1.706) | 0.179 | 6.895 (5.306–8.960) | < 0.001 |
| IV | 2.270 (1.499–3.438) | < 0.001 | 11.933 (7.807–18.241) | < 0.001 |
| T stage | ||||
| T1 | Reference | Reference | ||
| T2 | 1.426 (1.024–1.986) | 0.036 | 0.527 | |
| T3 | 1.467 (1.104–1.950) | 0.008 | 0.590 | |
| T4 | 1.738 (1.198–2.521) | 0.004 | 0.085 | |
| N stage | ||||
| N0 | Reference | Reference | ||
| N1 | 1.737 (1.391–2.169) | < 0.001 | 1.522 (1.194–1.942) | 0.001 |
| N2 | 1.760 (1.339–2.312) | < 0.001 | 1.716 (1.278–2.304) | < 0.001 |
| M stage | ||||
| M0 | Reference | Reference | ||
| M1 | 2.568 (1.886–3.496) | < 0.001 | 3.001 (2.141–4.208) | < 0.001 |
| SEER stage | ||||
| Localized | Reference | Reference | ||
| Regional | – | 0.167 | 0.270 | |
| Distant | – | 0.167 | 0.270 | |
| Surgery | ||||
| No/Unknown | Reference | Reference | ||
| Partial nephrectomy | 0.344 (0.267–0.444) | < 0.001 | 0.169 (0.114–0.252) | < 0.001 |
| Total nephrectomy | 0.693 (0.552–0.869) | 0.001 | 0.420 (0.315–0.559) | < 0.001 |
| Chemotherapy | ||||
| Yes | Reference | Reference | ||
| No/Unknown | 0.801 (0.651–0.985) | 0.035 | – | 0.270 |
| Radiotherapy | ||||
| Yes | Reference | Reference | ||
| No/Unknown | – | 0.236 | – | 0.058 |
OS Overall survival, CSS cancer-specific survival, HR hazard ratio, CI confidence intervals, AJCC American Joint Committee on Cancer, SEER Surveillance Epidemiology and End Results
aModel was adjusted by gender, age, marital status, race, grade, AJCC stage, TNM stage, SEER stage, surgery, chemotherapy and radiotherapy
bModel was adjusted by age, marital status, grade, AJCC stage, TNM stage, SEER stage, surgery, chemotherapy and radiotherapy
Clinical prognostic nomograms development and validation
Nomogram provide a simplified and visual representation of complex regression equations. Using multivariate Cox regression, we integrated all significant variables to develop clinical nomograms for predicting 3- and 5-year OS and CSS in individual patients. As shown in Fig. 1, age, AJCC stage, M stage and surgery were the most influential risk factors for OS in pRCC patients. Similarly, the nomogram predicting CSS highlighted AJCC stage as having the greatest impact on survival, followed by surgery. ROC curves demonstrated that the AUC of 3- and 5-year OS nomograms in the training cohort were 0.819 and 0.794, respectively (Fig. 2A, B), while the AUC for the validation cohort were 0.824 and 0.801, respectively (Fig. 2C, D). For the 3- and 5-year CSS nomogram, the AUC was approximately 0.9 in both the training and validation cohorts (Fig. 2E–H). It indicated that the proposed nomograms have good discriminative ability.
Fig. 1.
The nomogram predicting 3-, and 5-year overall survival (OS) and cancer-specific survival (CSS) rate of papillary renal cell carcinoma patients in the training cohort. A OS nomogram; B CSS nomogram
Fig. 2.
Receiver operating characteristic (ROC) curves detects the predictive value of two nomograms in the training and validation cohorts. A 3-year overall survival (OS) in the training cohort. B 5-year OS in the training cohort. C 3-year OS in the validation cohort. D 5-year OS in the validation cohort. E 3-year cancer-specific survival (CSS) the training cohort. F 5-year CSS the training cohort. G 3-year CSS the validation cohort. H 5-year CSS the validation cohort
Additionally, the nomogram achieved the highest C-indexes values (Table 3) and AUC scores (Fig. 3A–D and Table 4) compared to TNM stage and SEER stage. As shown in Fig. 3E–H, the nomograms could better predict OS and CSS, yielding greater net benefits than TNM staging and SEER-based staging. Subsequently, the calibration curves in both training and validation cohorts further confirmed the high consistency between the predicted and observed survival probability, underscoring the nomograms’ reliability in clinical practice (Fig. 4A–H).
Table 3.
Comparison of C-indexes between the nomogram, TNM and SEER stages in papillary renal cell carcinoma patients
| Characteristics | Training set | Validation set | |||
|---|---|---|---|---|---|
| HR | 95% CI | HR | 95% CI | ||
| OS | Nomogram | 0.789 | 0.776–0.802 | 0.787 | 0.767–0.859 |
| TNM stage | 0.698 | 0.683–0.713 | 0.690 | 0.667–0.713 | |
| SEER stage | 0.681 | 0.667–0.695 | 0.664 | 0.642–0.686 | |
| CSS | Nomogram | 0.894 | 0.880–0.908 | 0.888 | 0.866–0.910 |
| TNM stage | 0.864 | 0847–0.881 | 0.846 | 0.820–0.872 | |
| SEER stage | 0.833 | 0.815–0.851 | 0.809 | 0.781–0.837 | |
HR hazard ratio, CI confidence interval, SEER Surveillance, Epidemiology, and End Results
Fig. 3.
Area under the curve (AUC) and decision curve analysis (DCA) curves between the nomograms, TNM and SEER stage in predicting the OS and CSS of papillary renal cell carcinoma patients. A AUC curves of overall survival (OS) in the training cohort. B AUC curves of cancer-specific survival (CSS) in the training cohort. C AUC curves of OS in the validation cohort. D AUC curves of CSS in the validation cohort. E DCA curves of OS in the training cohort. F DCA curves of CSS in the training cohort. G DCA curves of OS in the validation cohort. H DCA curves of CSS in the validation cohort
Table 4.
Comparison of area under the curve (AUC) between the nomogram, TNM and SEER stages in papillary renal cell carcinoma patients
| Characteristics | Training set | Validation set | |||
|---|---|---|---|---|---|
| AUC | 95% CI | AUC | 95% CI | ||
| OS | Nomogram | 0.793 | 0.780–0.807 | 0.785 | 0.764–0.806 |
| TNM stage | 0.680 | 0.662–0.697 | 0.670 | 0.643–0.697 | |
| SEER stage | 0.657 | 0.640–0.675 | 0.645 | 0.618–0.672 | |
| CSS | Nomogram | 0.888 | 0.872–0.904 | 0.886 | 0.862–0.911 |
| TNM stage | 0.856 | 0.837–0.876 | 0.849 | 0.819–0.878 | |
| SEER stage | 0.821 | 0.799–0.842 | 0.805 | 0.771–0.839 | |
AUC Area under the curve, CI confidence interval, SEER Surveillance, Epidemiology, and End Results
Fig. 4.
Calibration plot of the nomogram for predicting 3-, and 5-year overall survival (OS) and cancer-specific survival (CSS) in the training and validation cohorts. A 3-year OS in the training cohort. B 5-year OS in the training cohort. C 3-year CSS in the training cohort. D 5-year CSS in the training cohort. E 3-year OS in the validation cohort. F 5-year OS in the validation cohort. G 3-year CSS in the validation cohort. H 5-year CSS in the validation cohort
Identification of prognostic-related genes by lasso analysis in TCGA cohort
The abnormal expression of many genes can lead to the progression of malignant tumors. We sought to further explore prognosis-related genes in pRCC patients. We identified 14 OS-related and 420 CSS-related differential genes using screening criteria of HR ≥ 4 and p value < 0.001. Fourteen genes were common between OS and CSS-related differential genes (Fig. 5A). After LASSO regression analysis, we screened 6 key OS-related differential genes and 10 CSS-related differential genes, and KCNG1, PTGIS, IDO1, IGF2BP3, PLK1 and H3C10 were closely associated with pRCC prognosis (Fig. 5B–F). The Kaplan–Meier curves based on above six prognostic-related genes revealed that patients with high gene expression had significantly shorter OS and CSS compared to those with low expression, with differences being statistically significant (p < 0.001) (Fig. 6A–L).
Fig. 5.
Identification of prognostic-related genes by Lasso analysis in TCGA cohort. A Differential prognostic-related genes in overall survival (OS) and cancer-specific survival (CSS). B, C Differential prognostic-related genes of OS Lasso regression analysis. D, E Differential prognostic-related genes of CSS Lasso regression analysis. F Key prognostic differential genes for OS and CSS after Lasso regression analysis
Fig. 6.
Overall survival (OS) and cancer-specific survival (CSS) curves of key prognostic differential genes. A The overall survival (OS) curve of the KCNG1. B The cancer-specific survival (CSS) curve of the KCNG1. C The OS curve of the PTGIS. D The CSS curve of the PTGIS. E The OS curve of the IDO1. F The CSS curve of the IDO1. G The OS curve of the IGF2BP3. H The CSS curve of the IGF2BP3. I The OS curve of the PLK1. J The CSS curve of the PLK1. K The OS curve of the H3C10. L The CSS curve of the H3C10
Univariate and multivariate Cox regression analysis of OS and CSS in TCGA cohort
To comprehensively explore prognostic factors, both clinical parameters and the identified genes were included as variables in the Cox regression analysis using TCGA data. The CI and corresponding p-values of specific variables in the univariate and multivariate analyses of OS and CSS were summarized in Tables 5 and 6, respectively. We found that clinical M stage, clinical stage, PTGIS, IDO1 and PLK1 were independent predictors of OS, while clinical M stage, clinical stage, IDO1 and PLK1 were independent predictors of CSS.
Table 5.
Univariate and multivariate analysis of overall survival (OS) rates of patients with papillary renal cell carcinoma in the TCGA database
| Characteristics | Total | Univariate analysis | Multivariate analysis | |||
|---|---|---|---|---|---|---|
| (N) | HR (95% CI) | P value | HR (95% CI) | P value | ||
| Gender | 290 | |||||
| Male | 213 | Reference | ||||
| Female | 77 | 1.588 (0.824–3.062) | 0.167 | |||
| Age, years | 288 | |||||
| ≤ 60 | 135 | Reference | ||||
| > 60 | 153 | 0.969 (0.533–1.762) | 0.917 | |||
| Race | 274 | |||||
| White | 206 | Reference | ||||
| Black or African American | 62 | 0.929 (0.410–2.105) | 0.860 | |||
| Asian | 6 | 6.009 (0.774–46.647) | 0.086 | |||
| Clinical T stage | 202 | |||||
| T1&T2 | 166 | Reference | Reference | |||
| T3&T4 | 36 | 4.711 (2.303–9.635) | < 0.001 | 0.368 (0.109–1.236) | 0.106 | |
| Clinical N stage | 154 | |||||
| N0 | 133 | Reference | Reference | |||
| N1 | 19 | 12.282 (5.649–26.699) | < 0.001 | 1.885 (0.531–6.689) | 0.326 | |
| N2 | 2 | 4.055 (0.525–31.332) | 0.180 | 4.794 (0.325–70.642) | 0.253 | |
| Clinical M stage | 210 | |||||
| M0 | 200 | Reference | Reference | |||
| M1 | 10 | 21.544 (7.956–58.337) | < 0.001 | 19.445 (4.102–92.177) | < 0.001 | |
| Clinical stage | 199 | |||||
| I&II | 159 | Reference | Reference | |||
| III&IV | 40 | 9.202 (4.458–18.994) | < 0.001 | 21.250 (3.484–129.607) | < 0.001 | |
| KCNG1 | 290 | |||||
| Low | 145 | Reference | Reference | |||
| High | 145 | 5.523 (2.596–11.749) | < 0.001 | 0.238 (0.053–1.069) | 0.061 | |
| PTGIS | 290 | |||||
| Low | 145 | Reference | Reference | |||
| High | 145 | 4.103 (2.025–8.310) | < 0.001 | 4.103 (1.334–12.620) | 0.014 | |
| IDO1 | 290 | |||||
| Low | 144 | Reference | Reference | |||
| High | 146 | 5.053 (2.347–10.879) | < 0.001 | 4.640 (1.337–16.105) | 0.016 | |
| IGF2BP3 | 290 | |||||
| Low | 145 | Reference | Reference | |||
| High | 145 | 5.349 (2.562–11.165) | < 0.001 | 2.141 (0.443–10.344) | 0.343 | |
| PLK1 | 290 | |||||
| Low | 144 | Reference | Reference | |||
| High | 146 | 4.829 (2.313–10.079) | < 0.001 | 9.122 (2.190–38.002) | 0.002 | |
| H3C10 | 290 | |||||
| Low | 144 | Reference | Reference | |||
| High | 146 | 4.273 (2.109–8.656) | < 0.001 | 1.535 (0.438–5.376) | 0.503 | |
Bold indicates a statistical difference
Table 6.
Univariate and multivariate analysis of cancer-specific survival (CSS) rates of patients with papillary renal cell carcinoma in the TCGA database
| Characteristics | Total | Univariate analysis | Multivariate analysis | |||
|---|---|---|---|---|---|---|
| (N) | HR (95% CI) | P value | HR (95% CI) | P value | ||
| Gender | 286 | |||||
| Male | 209 | Reference | ||||
| Female | 77 | 1.815 (0.836–3.940) | 0.132 | |||
| Age | 284 | |||||
| < = 60 | 135 | Reference | Reference | |||
| > 60 | 149 | 0.452 (0.208–0.980) | 0.044 | 2.721 (0.783–9.459) | 0.115 | |
| Race | 270 | |||||
| White | 202 | Reference | ||||
| Black or African American | 62 | 0.928 (0.348–2.470) | 0.881 | |||
| Asian | 6 | 5.903 (0.758–45.991) | 0.090 | |||
| Clinical T stage | 201 | |||||
| T1&T2 | 165 | Reference | Reference | |||
| T3&T4 | 36 | 8.983 (3.831–21.064) | < 0.001 | 0.416 (0.114–1.512) | 0.183 | |
| Clinical N stage | 154 | |||||
| N0 | 133 | Reference | Reference | |||
| N1 | 19 | 21.704 (8.586–54.863) | < 0.001 | 1.641 (0.422–6.384) | 0.475 | |
| N2 | 2 | 8.032 (0.973–66.299) | 0.053 | 5.066 (0.272–94.529) | 0.277 | |
| Clinical M stage | 209 | |||||
| M0 | 199 | Reference | Reference | |||
| M1 | 10 | 23.908 (8.595–66.503) | < 0.001 | 41.918 (5.644–311.332) | < 0.001 | |
| Clinical stage | 198 | |||||
| Stage I&Stage II | 158 | Reference | Reference | |||
| Stage III&Stage IV | 40 | 35.100 (10.369–118.823) | < 0.001 | 68.612 (7.832–601.072) | < 0.001 | |
| KCNG1 | 286 | |||||
| Low | 145 | Reference | Reference | |||
| High | 141 | 10.751 (3.217–35.928) | < 0.001 | 0.624 (0.085–4.595) | 0.643 | |
| PTGIS | 286 | |||||
| Low | 144 | Reference | Reference | |||
| High | 142 | 5.596 (2.125–14.740) | < 0.001 | 2.474 (0.623–9.825) | 0.198 | |
| IDO1 | 286 | |||||
| Low | 142 | Reference | Reference | |||
| High | 144 | 14.717 (3.490–62.060) | < 0.001 | 14.102 (2.152–92.422) | 0.006 | |
| IGF2BP3 | 286 | |||||
| Low | 143 | Reference | Reference | |||
| High | 143 | 10.364 (3.124–34.386) | < 0.001 | 1.271 (0.158–10.257) | 0.822 | |
| PLK1 | 286 | |||||
| Low | 140 | Reference | Reference | |||
| High | 146 | 14.616 (3.467–61.621) | < 0.001 | 13.200 (2.043–85.289) | 0.007 | |
| H3C10 | 286 | |||||
| Low | 144 | Reference | Reference | |||
| High | 142 | 4.650 (1.882–11.485) | < 0.001 | 1.059 (0.176–6.355) | 0.950 | |
Bold indicates a statistical difference
Development and validation of molecular prognostic nomograms
Based on the variables that showed statistical differences in the results of multivariate Cox regression analysis, we developed molecular prognostic nomograms (Fig. 7A, B). In the OS and CSS nomograms, M stage had the greatest influence on survival outcomes, with the expression of related genes also significantly affected the prognosis. Calibration plots indicated that the nomograms were well-calibrated, demonstrating excellent predictive performance (Fig. 7C, D).
Fig. 7.
The nomogram predicting 3-, and 5-year overall survival (OS) and cancer-specific survival (CSS) rate of papillary renal cell carcinoma patients in the TCGA database. A OS nomogram. B CSS nomogram. C Calibration plot of the nomogram for predicting 3-, and 5-year OS. D Calibration plot of the nomogram for predicting 3-, and 5-year CSS
Discussion
Renal cell carcinoma (RCC) is a common urological malignancy, accounting for approximately 90% of all kidney cancers in adults, and its incidence continues to increase [24]. It is estimated to contribute about 4.1% of all new cancer cases in the United States in 2022 [25]. pRCC is the second most common subtype of RCC, differs significantly from clear cell renal cell carcinoma (ccRCC) in terms of clinical presentation, prognosis, and patient outcomes [20]. For instance, research by Patard JJ highlighted distinctions in grade and stage between pRCC and ccRCC, with pRCC often presenting at earlier stages and lower grades [26–28]. However, most studies have focused on the prognostic assessment of RCC or conventional ccRCC, leaving a gap in specific prognostic guidance for pRCC.
In recent years, the World Health Organization (WHO) has emphasized the importance of molecular classification and deeper investigation into the molecular biology of pRCC [10]. This study aims to address this gap by summarizing the clinicopathological features of pRCC and developing clinical and molecular prognostic nomograms that integrate transcriptomic data for the first time.
Age is a significant prognostic factor, with pRCC typically occurring between ages 60 and 80 [29]. Our predictive models for OS and CSS show worsening outcomes with increasing age. Previous studies have also reported a higher incidence of pRCC in males, with a male-to-female ratio of 1.5:1 to 2:1 [30, 31]. Additionally, pRCC is also race-specific, and the OS nomogram in SEER database indicates that black patients have a poorer prognosis compared to other racial groups. Besides objective economic and medical levels, it may be concerned with the relative enrichment of immune system pathways, VEGF pathways and CRYBB2 in the black race [32]. The prognosis of pRCC is closely related to the disease stage. In the subgroup of patients with localized tumors (T1-4, N0, M0), the 5-year OS rate was ranged from 78–79%, and the 5-year CSS was 86–94% [26, 33, 34]. In contrast, patients with metastatic pRCC have far worse outcomes, with a 5-year OS of 15.9% and 5-year CSS of 17.9%) [34]. Since RCC is insensitive to both radiotherapy and chemotherapy, surgery is the primary treatment option for pRCC. Our findings indicate that patients undergoing partial nephrectomy have higher OS rates compared to those receiving radical nephrectomy, likely because the latter group presents with more advanced TNM stages [33]. Although both TNM and SEER staging systems offer some prognostic value, our prognostic nomogram model clearly demonstrated better superior accuracy and reliability in predicting outcomes.
Over time, people realized that tumor development is influenced not only by clinical and pathological factors but also by the underlying molecular and genetic background. Linehan et al. [35] extensively characterized the biologic foundation of pRCC using data from TCGA, providing a broad assessment individual disease occurrence. A deeper understanding of pRCC-related biomarkers is crucial for early diagnosis and treatment. In our molecular prognostic model based on TCGA, IDO1 and PLK1 were independent biological indicators both affecting OS and CSS. Higher expression of these genes was associated with poorer patient outcomes. IDO1 is a rate-limiting metabolic enzyme that contributes to tumor cell proliferation and migration. It plays a crucial role in facilitating tumor immune evasion by suppressing the function of immune effector cells in the local tumor microenvironment [36]. Numerous studies have shown that IDO1 is highly expressed in a variety of cancers [37]. A study by Shu et al. [38] demonstrated that the expression of IDO1 was upregulated during the process of RCC immune escape. Similarly, PLK1 is also closely linked with tumorigenesis, and its expression is increased in a variety of solid tumors such as esophageal cancer, melanoma, breast cancer, colon cancer and RCC [39–41]. As genomics and precision therapies advance, combining clinical research with molecular insights will help identify new therapeutic targets and prognostic markers.
To the best of our knowledge, although the nomograms revealed excellent accuracy in both training and validation cohorts, it should be noted that were still some limitations as follows. First, this study was a retrospective survey, which inevitably leads to unavoidable selection bias. Second, the SEER database failed to provide data on several key variables, such as ECOG score, comorbidities, and other treatment options. Third, while the prognostic models based on TCGA data were developed using bioinformatics approaches, further validation through experimental studies is necessary. Nonetheless, it is undeniable that the nomograms provide a theoretical basis and reference for the prognostic evaluation of pRCC. Patients with higher total scores, indicating a potentially poorer prognosis, may benefit from more aggressive therapeutic interventions and lifestyle modifications.
Conclusion
In conclusion, our study was the first to developed clinical and molecular prognostic nomograms for pRCC patients by leveraging data from the SEER and TCGA databases, and it demonstrated the models’ strong predictive capabilities. Besides predicting OS and CSS, the model may offer new insights and therapeutic targets for future treatment strategies. However, further research is needed to explore the role of the identified genes in pRCC progression and prognosis.
Supplementary Information
Additional file 1: Figure S1. Schematic flowchart of the inclusion and exclusion criteria.
Additional file 2: Figure S2. Using X-tile software to group the age of patients with papillary renal cell carcinoma.
Author contributions
XW collected the data, analyzed and interpreted the data, drafted and revised the manuscript.
Funding
This work was supported by the Agricultural and Social Development Science and Technology Program of Yinzhou District (No. 20231YZQ070032).
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Declarations
Research involving human participants and/or animals
This article does not contain any studies with human participants or animals performed by any of the authors.
Competing interests
he authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Mancilla-Jimenez R, Stanley RJ, Blath RA. Papillary renal cell carcinoma: a clinical, radiologic, and pathologic study of 34 cases. Cancer. 1976;38(6):2469–80. [DOI] [PubMed] [Google Scholar]
- 2.Malouf GG, Su X, Zhang J, Creighton CJ, Ho TH, Lu Y, et al. DNA methylation signature reveals cell ontogeny of renal cell carcinomas. Clin Cancer Res. 2016;22(24):6236–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Moch H, Cubilla AL, Humphrey PA, Reuter VE, Ulbright TM. The 2016 WHO classification of tumours of the urinary system and male genital organs-part a: renal, penile, and testicular tumours. Eur Urol. 2016;70(1):93–105. [DOI] [PubMed] [Google Scholar]
- 4.Delahunt B, Eble JN. Papillary renal cell carcinoma: a clinicopathologic and immunohistochemical study of 105 tumors. Mod Pathol. 1997;10(6):537–44. [PubMed] [Google Scholar]
- 5.Maughan BL, Sirohi D. Papillary renal cell carcinoma: a review of prospective clinical trials. Curr Treat Options Oncol. 2023;24(9):1199–212. [DOI] [PubMed] [Google Scholar]
- 6.Lobo J, Ohashi R, Helmchen BM, Rupp NJ, Rüschoff JH, Moch H. The morphological spectrum of papillary renal cell carcinoma and prevalence of provisional/emerging renal tumor entities with papillary growth. Biomedicines. 2021. 10.3390/biomedicines9101418. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Saleeb RM, Brimo F, Farag M, Rompré-Brodeur A, Rotondo F, Beharry V, et al. Toward biological subtyping of papillary renal cell carcinoma with clinical implications through histologic, immunohistochemical, and molecular analysis. Am J Surg Pathol. 2017;41(12):1618–29. [DOI] [PubMed] [Google Scholar]
- 8.Kovac M, Navas C, Horswell S, Salm M, Bardella C, Rowan A, et al. Recurrent chromosomal gains and heterogeneous driver mutations characterise papillary renal cancer evolution. Nat Commun. 2015;6:6336. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Chevarie-Davis M, Riazalhosseini Y, Arseneault M, Aprikian A, Kassouf W, Tanguay S, et al. The morphologic and immunohistochemical spectrum of papillary renal cell carcinoma: study including 132 cases with pure type 1 and type 2 morphology as well as tumors with overlapping features. Am J Surg Pathol. 2014;38(7):887–94. [DOI] [PubMed] [Google Scholar]
- 10.Lobo J, Ohashi R, Amin MB, Berney DM, Compérat EM, Cree IA, et al. WHO 2022 landscape of papillary and chromophobe renal cell carcinoma. Histopathology. 2022;81(4):426–38. [DOI] [PubMed] [Google Scholar]
- 11.Pan H, Ye L, Zhu Q, Yang Z, Hu M. The effect of the papillary renal cell carcinoma subtype on oncological outcomes. Sci Rep. 2020;10(1):21073. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Le X, Wang XB, Zhao H, Chen RF, Ge P. Comparison of clinicopathologic parameters and oncologic outcomes between type 1 and type 2 papillary renal cell carcinoma. BMC Urol. 2020;20(1):148. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Trpkov K, Williamson SR, Gill AJ, Adeniran AJ, Agaimy A, Alaghehbandan R, et al. Novel, emerging and provisional renal entities: the genitourinary pathology society (GUPS) update on renal neoplasia. Mod Pathol. 2021;34(6):1167–84. [DOI] [PubMed] [Google Scholar]
- 14.Trpkov K, Hes O, Williamson SR, Adeniran AJ, Agaimy A, Alaghehbandan R, et al. New developments in existing WHO entities and evolving molecular concepts: the genitourinary pathology society (GUPS) update on renal neoplasia. Mod Pathol. 2021;34(7):1392–424. [DOI] [PubMed] [Google Scholar]
- 15.Cimadamore A, Cheng L, Scarpelli M, Massari F, Mollica V, Santoni M, et al. Towards a new WHO classification of renal cell tumor: what the clinician needs to know-a narrative review. Transl Androl Urol. 2021;10(3):1506–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Trpkov K, Hes O. New and emerging renal entities: a perspective post-WHO 2016 classification. Histopathology. 2019;74(1):31–59. [DOI] [PubMed] [Google Scholar]
- 17.Courthod G, Tucci M, Di Maio M, Scagliotti GV. Papillary renal cell carcinoma: a review of the current therapeutic landscape. Crit Rev Oncol Hematol. 2015;96(1):100–12. [DOI] [PubMed] [Google Scholar]
- 18.Znaor A, Lortet-Tieulent J, Laversanne M, Jemal A, Bray F. International variations and trends in renal cell carcinoma incidence and mortality. Eur Urol. 2015;67(3):519–30. [DOI] [PubMed] [Google Scholar]
- 19.Gettman MT, Blute ML, Spotts B, Bryant SC, Zincke H. Pathologic staging of renal cell carcinoma: significance of tumor classification with the 1997 TNM staging system. Cancer. 2001;91(2):354–61. [DOI] [PubMed] [Google Scholar]
- 20.Margulis V, Tamboli P, Matin SF, Swanson DA, Wood CG. Analysis of clinicopathologic predictors of oncologic outcome provides insight into the natural history of surgically managed papillary renal cell carcinoma. Cancer. 2008;112(7):1480–8. [DOI] [PubMed] [Google Scholar]
- 21.Sukov WR, Lohse CM, Leibovich BC, Thompson RH, Cheville JC. Clinical and pathological features associated with prognosis in patients with papillary renal cell carcinoma. J Urol. 2012;187(1):54–9. [DOI] [PubMed] [Google Scholar]
- 22.Haake SM, Weyandt JD, Rathmell WK. Insights into the genetic basis of the renal cell carcinomas from the cancer genome atlas. Mol Cancer Res. 2016;14(7):589–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Mao W, Fu Z, Wang K, Wu J, Xu B, Chen M. Prognostic nomogram for patients with lung metastatic renal cell carcinoma: a SEER-based study. Ann Palliat Med. 2021;10(3):2791–804. [DOI] [PubMed] [Google Scholar]
- 24.Mao W, Wang K, Xu B, Zhang H, Sun S, Hu Q, et al. ciRS-7 is a prognostic biomarker and potential gene therapy target for renal cell carcinoma. Mol Cancer. 2021;20(1):142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2021. CA Cancer J Clin. 2021;71(1):7–33. [DOI] [PubMed] [Google Scholar]
- 26.Patard JJ, Leray E, Rioux-Leclercq N, Cindolo L, Ficarra V, Zisman A, et al. Prognostic value of histologic subtypes in renal cell carcinoma: a multicenter experience. J Clin Oncol. 2005;23(12):2763–71. [DOI] [PubMed] [Google Scholar]
- 27.Amin MB, Amin MB, Tamboli P, Javidan J, Stricker H, de Peralta VM, et al. Prognostic impact of histologic subtyping of adult renal epithelial neoplasms: an experience of 405 cases. Am J Surg Pathol. 2002;26(3):281–91. [DOI] [PubMed] [Google Scholar]
- 28.Beck SD, Patel MI, Snyder ME, Kattan MW, Motzer RJ, Reuter VE, et al. Effect of papillary and chromophobe cell type on disease-free survival after nephrectomy for renal cell carcinoma. Ann Surg Oncol. 2004;11(1):71–7. [DOI] [PubMed] [Google Scholar]
- 29.Klatte T, Pantuck AJ, Said JW, Seligson DB, Rao NP, LaRochelle JC, et al. Cytogenetic and molecular tumor profiling for type 1 and type 2 papillary renal cell carcinoma. Clin Cancer Res. 2009;15(4):1162–9. [DOI] [PubMed] [Google Scholar]
- 30.Aron M, Nguyen MM, Stein RJ, Gill IS. Impact of gender in renal cell carcinoma: an analysis of the SEER database. Eur Urol. 2008;54(1):133–40. [DOI] [PubMed] [Google Scholar]
- 31.Qu Y, Chen H, Gu W, Gu C, Zhang H, Xu J, et al. Age-dependent association between sex and renal cell carcinoma mortality: a population-based analysis. Sci Rep. 2015;5:9160. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Chow WH, Shuch B, Linehan WM, Devesa SS. Racial disparity in renal cell carcinoma patient survival according to demographic and clinical characteristics. Cancer. 2013;119(2):388–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Fernandes DS, Lopes JM. Pathology, therapy and prognosis of papillary renal carcinoma. Future Oncol. 2015;11(1):121–32. [DOI] [PubMed] [Google Scholar]
- 34.Steffens S, Janssen M, Roos FC, Becker F, Schumacher S, Seidel C, et al. Incidence and long-term prognosis of papillary compared to clear cell renal cell carcinoma–a multicentre study. Eur J Cancer. 2012;48(15):2347–52. [DOI] [PubMed] [Google Scholar]
- 35.Linehan WM, Spellman PT, Ricketts CJ, Creighton CJ, Fei SS, Davis C, et al. Comprehensive molecular characterization of papillary renal-cell carcinoma. N Engl J Med. 2016;374(2):135–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Hwang SL, Chung NP, Chan JK, Lin CL. Indoleamine 2, 3-dioxygenase (IDO) is essential for dendritic cell activation and chemotactic responsiveness to chemokines. Cell Res. 2005;15(3):167–75. [DOI] [PubMed] [Google Scholar]
- 37.Zhai L, Ladomersky E, Lenzen A, Nguyen B, Patel R, Lauing KL, et al. IDO1 in cancer: a Gemini of immune checkpoints. Cell Mol Immunol. 2018;15(5):447–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Shu G, Chen M, Liao W, Fu L, Lin M, Gui C, et al. PABPC1L induces IDO1 to promote tryptophan metabolism and immune suppression in renal cell carcinoma. Cancer Res. 2024;84(10):1659–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Strebhardt K. Multifaceted polo-like kinases: drug targets and antitargets for cancer therapy. Nat Rev Drug Discov. 2010;9(8):643–60. [DOI] [PubMed] [Google Scholar]
- 40.Kim SA, Kwon SM, Yoon JH, Ahn SG. The antitumor effect of PLK1 and HSF1 double knockdown on human oral carcinoma cells. Int J Oncol. 2010;36(4):867–72. [DOI] [PubMed] [Google Scholar]
- 41.Steinhauser I, Langer K, Strebhardt K, Spänkuch B. Uptake of plasmid-loaded nanoparticles in breast cancer cells and effect on Plk1 expression. J Drug Target. 2009;17(8):627–37. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Additional file 1: Figure S1. Schematic flowchart of the inclusion and exclusion criteria.
Additional file 2: Figure S2. Using X-tile software to group the age of patients with papillary renal cell carcinoma.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.