1.Data acquisition
Inclusion criteria were: 1) a histological diagnosis of clear cell renal cell carcinoma (ccRCC); 2) accessible gene expression data and clinicopathological characteristics; and 3) Complete survival records and follow-up periods exceeding 30 days. Obtained from the TCGA-KIRC database, the study included 537 patients diagnosed with ccRCC. The baseline data of all ccRCC patients was meticulously presented in Table 1. And 11 individuals were excluded due to incomplete data. Next, the 526 patients diagnosed with ccRCC were randomly allocated into a training cohort (n = 263) and a testing cohort (n = 263) in a 1:1 ratio. The study flowchart is illustrated in Figure 1. Nineteen PANoptosis-related genes (PRGs) were identified from previous research (27-35), and these genes are detailed in Supplementary Table S1.
2. Prognostic PANoptosis-related gene Signature
Univariate Cox regression analysis was conducted on the training cohort comprising 263 ccRCC samples, utilizing a set of 19 PRGs sourced from previous studies. This initial analysis aimed to identify potential prognostic genes associated with ccRCC. Subsequently, the least absolute shrinkage and selection operator (LASSO) regression method was employed to further refine the candidate gene set. Through LASSO regression, six PRGs were selected based on their predictive power, as illustrated in Figures 2A and 2B. These six PRGs, namely AIM2, TRADD, CASP7, TAB2, TAB3, and TNFAIP3, exhibited significant associations with ccRCC prognosis. Finally, multivariate Cox regression analysis was performed on these six PRGs to assess their independent prognostic value in ccRCC, as depicted in Figure 2C. This comprehensive approach allowed us to identify a refined set of PRGs with strong prognostic potential in ccRCC. A signature of PRGs was formulated (Table 2) using the subsequent risk score equation:
RiskScore=-0.1594*TAB2-0.1061*TAB3-0.0263*TNFAIP3-0.3093*CASP7+0.5010*AIM2-0.0751*TRADD
3. Assessment of a Prognostic PANoptosis-related genes Signature
AIM2 was risk factor for ccRCC, while other five genes (TAB2, TAB3, TNFAIP3, CASP7, and TRADD) were protective factors for ccRCC (Figures 3A–F). The risk score formula was meticulously applied to the gene expression data of 263 patients within the training set, enabling the stratification of individuals into distinct risk categories. With this method, patients were bifurcated into low-risk and high-risk groups, each comprising 131 and 132 individuals, respectively. This stratification, illustrated in Figure 3A, provided a clear visualization of how the risk score influenced patient classification. Furthermore, as depicted in Figure 3B, a noteworthy trend emerged wherein higher risk scores corresponded to progressively shorter survival times and increased mortality rates among patients. These observations underscored the prognostic relevance of the risk score model in predicting patient outcomes. To deepen our understanding of the molecular landscape associated with patient prognosis, heatmaps were generated to visualize the expression levels of OS-related PRGs (Figure 3C). This visualization facilitated the identification of potential correlations between gene expression patterns and patient survival outcomes. Statistical analysis, including the log-rank test, revealed a significant difference in overall survival (OS) between the low-risk and high-risk groups, further validating the prognostic utility of the risk score model (p = 3.7e-06, Figure 3D). Moreover, time-dependent ROC curves were constructed to evaluate the predictive performance of OS-related PRGs across varying time intervals (Figure 3E). Notably, the AUC values indicated the model's ability to accurately predict patient survival at 1, 3, and 5-year timepoints. Following validation in both the testing cohort and the entire cohort (Figures 3F–O), survival analysis consistently demonstrated substantial differences in survival outcomes between the low-risk and high-risk groups. This robust performance was further supported by the AUC values, which underscored the predictive reliability of the six PRGs prognostic signature across different patient cohorts.
4. Clinical Value of a Prognostic PANoptosis-related genes Signature
Univariate Cox regression analysis illuminated the significant prognostic role of the risk score in patients diagnosed with ccRCC (95% confidence interval (CI): 2.494–4.886, p<0.001), where a higher risk score indicated a graver prognosis. Intriguingly, additional clinical parameters such as stage (95% CI: 1.658–2.166, p<0.001), grade (95% CI: 1.881–2.839, p<0.001), and patient age (95% CI:1.015–1.042, p<0.001) were also intimately linked to prognosis (Figure 4A). Impressively, even after adjusting for these clinical characteristics, the risk score retained its independence as a prognostic indicator (hazard ratio = 1.912, 95% CI: 1.311–2.788, p<0.001, Figure 4B). Further exploration revealed intriguing nuances within different patient subgroups. While no significant divergence in overall survival (OS) rates between high-risk and low-risk groups was observed solely in the G1 and G2 subgroups, a striking trend emerged across all age, sex, and other grade subgroups. In these cohorts, ccRCC patients in the high-risk category consistently displayed lower OS rates compared to their low-risk counterparts (Figure 5). These nuanced findings underscore the robust prognostic association between the PRGs signature and ccRCC patient outcomes, enriching our understanding of the intricate interplay between molecular signatures and clinical prognosis in this disease context.
5. Construction and Assessment of Nomogram
The nomogram, a widely used prognostic tool in oncology, provides a visual representation for estimating patient survival by incorporating various prognostic factors into a scoring system. In constructing the nomogram, multivariate Cox regression analysis was employed to identify significant predictors, including age, stage, grade, and risk score. Each predictor was assigned a corresponding score, and the total score was calculated by summing the scores of individual predictors, thus enabling the prediction of survival probabilities at 1, 3, and 5 years (Figure 6A). The nomogram demonstrated good reliability, as indicated by a consistency index of 0.771. Additionally, calibration curves illustrated that the predicted survival probabilities at 1, 3, and 5 years closely aligned with the observed outcomes, showing minimal deviation from the reference line (Figures 6B). In addition, the AUC for the nomogram was 0.86 in 1 year, 0.81 in 3 years and 0.78 in 5 years (Figures 6C). The newly constructed nomogram demonstrated great predictive ability and promising clinical potential.
6. Construction of the Network and the Functional Enrichment Analysis
We conducted functional enrichment analysis on the obtained 6 PRGs. KEGG showed Apoptosis and Necroptosis to be a significantly enriched pathway (Figure 7A; Table 2). The GO analysis showed that it was involved in 177 biological processes, such as regulation of canonical NF-kappaB signal transduction and canonical NF-kappaB signal transduction, constituted 4 cellular components, including tumor necrosis factor receptor superfamily complex, and mediated 27 molecular functions, such as K63-linked polyubiquitin modification-dependent protein binding and polyubiquitin modification-dependent protein binding (Figure 7B; Table 3). In addition, we constructed a network consisting of 6 prognosis-related differentially expressed genes (PRDEGs) and 5 predicted functional partners using STRING website (Figure 8). This may help us identify new genes worthy of further investigation.