Elevated baseline C-reactive protein predicts poorer survival in lung cancer: a 22-year retrospective cohort study

Introduction

Lung cancer is the leading cause of incidence and mortality related to malignant tumor worldwide, resulting in over 2.48 million new cases and 1.8 million deaths per year (1,2). In China, lung cancer has also emerged as the top formidable health challenge and it is estimated that there are 1.06 million new cases and 0.73 million people died from lung cancer annually (3). Nowadays, despite major advances in screening, molecularly targeted therapy, and immunotherapy, lung cancer remains the leading cause of cancer-related death worldwide (4). Five-year survival rates are still unsatisfactory, particularly for lung cancer patients diagnosed at advanced stages, and substantial heterogeneity in outcomes persists even among those with similar stage and treatment (4,5). Although many potential biomarkers have been proposed and evaluated, most of them provide very limited prognostic value beyond stage at diagnosis. Thus, finding a simple, inexpensive and widely applicable prognostic and predictive biomarkers to refine risk stratification and guide individualized treatment decisions in lung cancer, has become an urgent issue.

Inflammation is a notable hallmark of malignancy with strong links to tumor initiation, progression and response to therapy (6,7). C-reactive protein (CRP), an acute-phase reactant synthesized by hepatocytes in response to interleukin-6 and other pro-inflammatory cytokines, is a well-established marker of systemic inflammation and tissue injury (8,9). Circulating CRP is a well-established and highly sensitive biomarker of systemic inflammation that has been shown to be associated with survival of patients with various diseases including cancer (10-12). Given that CRP is routinely measured in clinical practice and is inexpensive and reproducible, it is an attractive candidate biomarker for real-world prognostication. Previous studies reveal that elevated serum levels of CRP among lung cancer patients are linked to poor prognosis (11,13,14), it might be associated with a complex interplay among tumor burden, host immune response and comorbid inflammatory conditions. Moreover, multiple studies and meta-analyses have reported that higher pretreatment CRP levels are associated with worse Overall survival (OS) and progression-free survival (PFS) in both early-stage and advanced non-small cell lung cancer (NSCLC), as well as among patients treated with immune checkpoint inhibitors (ICIs) (15-17). However, previous studies were limited by modest sample sizes, restriction to specific clinical scenarios, heterogeneous CRP cutoff value and incomplete adjustment for established prognostic variables. Furthermore, few studies have examined how baseline CRP relates both to detailed clinicopathologic characteristics and to real-world treatment efficacy in lung cancer patients.

The aims of our study were to explore the potential value of baseline CRP in clinical outcome in patients with lung cancer, and explore the association among the baseline CRP elevation and an adverse clinicopathologic profile and inferior treatment outcomes in lung cancer patients. Our study describes the distribution of baseline CRP and its associations with clinical and pathologic features and determines whether baseline CRP is independently associated with PFS and OS. By establishing these evidence-based parameters, our findings may inform the prognostic utility of CRP as a surrogate marker for efficacy in treatment and lung cancer survival. We present this article in accordance with the STROBE reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0118/rc).

Methods

Study design and patient population

We conducted a single-center retrospective cohort study of patients treated at the Lung Cancer Center of Peking Union Medical College Hospital (PUMCH). The study protocol was approved by the Ethical Committee of Peking Union Medical College Hospital (No. K7024-K24C3469) and was conducted in accordance with the Declaration of Helsinki (18) and its subsequent amendments. All patients have signed informed consent forms. Eligible patients were enrolled at first diagnosis between October 2000 and March 2022 and followed longitudinally for clinical outcomes.

In this study, the inclusion criteria were as follows: (I) patients aged >18 years old; (II) histologically or cytologically confirmed primary lung cancer [including NSCLC and small cell lung cancer (SCLC)]; (III) initiation of first-line systemic anti-tumor therapy (chemotherapy, targeted therapy, immunotherapy, or their combinations) at Lung Cancer Center of PUMCH during the study period; (IV) availability of at least one serum CRP measurement performed within 30 days before the start of first-line systemic therapy; (V) adequate clinical, pathologic, treatment and follow-up data to determine response and survival outcomes. The exclusion criteria were: (I) evidence of active infection, autoimmune disease flare, or other acute inflammatory condition at the time of baseline CRP assessment, when this information was available in the medical record; (II) history of another malignancy; (III) missing key data on baseline CRP or survival endpoints; (IV) receipt of prior systemic anti-tumor therapy for lung cancer before the baseline CRP measurement used in this analysis.

Data collection

From electronic medical records, we abstracted patient demographics (age at diagnosis, sex), lifestyle factors (smoking status and alcohol consumption) and comorbid conditions (chronic diseases such as hypertension, diabetes and coronary heart disease). Body mass index (BMI) at diagnosis was classified as underweight (<18.5 kg/m²), normal (18.5–24.9 kg/m²), overweight (25.0–29.9 kg/m²) or obese (≥30 kg/m²). Clinical and tumor-related variables included histologic subtype (NSCLC vs. SCLC), tumor-node-metastasis (TNM) stage at presentation [American Joint Committee on Cancer (AJCC) stage I–II vs. III–IV], Eastern Cooperative Oncology Group (ECOG) performance status (0–1 vs. 2–4), presence of distant metastases and driver gene mutation status. Tumor programmed death ligand 1 (PD-L1) expression was assessed by immunohistochemistry when available; PD-L1 was defined as positive if the tumor proportion score (TPS) was ≥1% (or according to assay-specific cutoffs). Treatment variables (chemotherapy, targeted therapy and ICIs) were recorded, immunotherapy was treated as an observational covariate and interpreted cautiously in multivariable analyses.

CRP measurement

Serum CRP was measured via standard turbidimetric or immunonephelometric assays in the hospital laboratory. Consistent with clinical conventions and prior literature, we defined elevated CRP as ≥10 mg/L, using 10 mg/L as a cutoff to distinguish between low CRP (normal or mildly elevated, <10 mg/L) and high CRP (markedly elevated inflammation) groups (19). CRP values were treated both as a continuous variable (for descriptive statistics) and categorically for group comparisons. Given that CRP values were right-skewed, they are summarized as median and interquartile range (IQR).

Statistical analysis

Baseline characteristics were summarized using median (IQR) for continuous variables and n (%) for categorical variables. Differences in categorical variables by CRP group were compared using the Chi-squared test (or Fisher’s exact test when appropriate). Given the right-skewed distribution of CRP, subgroup CRP levels were compared using non-parametric tests (Mann-Whitney U test for two groups or Kruskal-Wallis test for ≥3 groups).

OS was defined as the time from date of diagnosis to death from any cause; patients alive at last follow-up were censored. PFS was defined as the time from initiation of first-line systemic therapy to radiographic or clinical progression or death, whichever occurred first; patients without an event were censored at the last disease assessment. Median follow-up time was estimated using the reverse Kaplan-Meier method.

Kaplan-Meier curves were generated for OS and PFS stratified by CRP group (CRP <10 vs. ≥10 mg/L) and compared using the log-rank test. Cox proportional hazards regression was used to estimate hazard ratios (HRs) with 95% confidence intervals (CIs). Multivariable models were prespecified to include baseline CRP group and clinically relevant covariates (age, sex, BMI, smoking, alcohol use, comorbidities, histology, stage, ECOG performance status, driver mutation status, PD-L1 status, immunotherapy and distant metastasis). Missing values were handled using complete-case analysis for the primary models, with sensitivity analyses recommended using an “unknown” category or multiple imputation for variables with substantial missingness (e.g., PD-L1). The proportional hazards assumption should be evaluated using Schoenfeld residuals or time-varying covariate tests.

All tests were two-sided, with a significance threshold of P<0.05. Statistical analyses were conducted using SPSS (version 26) and GraphPad Prism (version 9.0) software.

Results

Patient characteristics

A total of 1,339 lung cancer patients met the inclusion criteria (Figure 1 shows the details of patient selection criteria). The cohort included 1,108 patients (82.7%) with NSCLC and 231 (17.3%) with SCLC. Among NSCLC cases, the major subtypes were adenocarcinoma (56.3%), squamous cell carcinoma (23.8%) and others (2.6%). The overall median age was 62 years (IQR, 56–68 years), with 40.1% of patients aged ≥65 years. There was a male predominance (68.3% male). Most patients (79.7%) had ECOG score 0–1 at diagnosis, while 20.3% had ECOG score 2–4. The majority (89.2%) presented with advanced stage III–IV disease and 63.3% had documented distant metastases at diagnosis. Driver oncogenic mutations (such as EGFR, ALK, ROS1, etc.) were identified in 36.5% of tested patients. PD-L1 expression data were available for 382 NSCLC patients in the later years; 28.5% of those had PD-L1-positive tumors (TPS ≥1%). During the study period, 14.8% of patients received ICI therapy at some point (mostly after 2015, reflecting evolving treatment patterns). Baseline CRP values ranged widely (median 5.4 mg/L, IQR, 1.7–27.3 mg/L). Overall, 549 patients (41.0%) had CRP ≥10 mg/L at baseline, while 790 (59.0%) had CRP <10 mg/L. Complete demographic and clinical characteristics are presented in Table S1, Tables 1,2.

Figure 1 Study participant selection flow diagram. CRP, C-reactive protein.

Associations of CRP with clinical and pathological variables

We observed significant associations between elevated CRP and various clinical characteristics (Tables 1,2). Patients ≥65 years old had substantially higher CRP on average than those <65 years (median 7.2 vs. 4.6 mg/L), and older patients (≥65 years) were more likely to have high CRP (43.4% vs. 37.8% of low-CRP patients, P=0.043). Male had substantially higher CRP on average than women (median 7.6 vs. 3.2 mg/L), and 76.3% of the high-CRP group were male (vs. 62.8% of low-CRP group; P<0.001). Lifestyle factors showed strong correlations with CRP. Current or former smokers comprised 60.7% of the cohort; importantly, among patients with CRP ≥10, 67.2% were smokers, compared to 56.2% in the CRP <10 group (P<0.001). Similarly, alcohol use was more frequent in the high-CRP group (50.3% vs. 39.9%; P<0.001). These findings suggest that smoking and alcohol-related inflammation may contribute to elevated CRP.

Performance status and comorbidities were related to CRP as well. Patients with poor ECOG score 2–4 had a much higher prevalence of CRP ≥10 (29.9% vs. 13.7%; P<0.001) and median CRP in ECOG score 2–4 was 20.9 vs. 4.6 mg/L in ECOG score 0–1. Interestingly, the presence of chronic co-morbid diseases (e.g., hypertension, diabetes, coronary heart disease and so on) did not significantly differ by CRP status (P=0.83), indicating that it was primarily cancer-related factors rather than chronic comorbidity driving CRP in this cohort.

Tumor characteristics demonstrated notable links with elevated CRP (Table 2). SCLC patients tended to have slightly lower CRP than NSCLC (median 4.2 vs. 5.5 mg/L) and a smaller fraction of SCLC patients had CRP ≥10 mg/L (15.1% of high-CRP group were SCLC vs. 18.7% of low-CRP group), but this difference was not statistically significant (P=0.09). By contrast, advanced tumor stage was strongly associated with elevated CRP. Among stage III–IV patients, 95.3% had high CRP, compared to only 85.1% in the CRP <10 mg/L (P<0.001), median CRP was 6.4 vs. 1.7 mg/L for advanced vs. early stage (P<0.001). Nearly all patients in the high-CRP group had advanced disease (95.3%), reflecting that early-stage lung cancers rarely exhibited CRP ≥10 mg/L.

Elevated baseline CRP was also linked to tumor molecular and immune markers. Patients whose tumors lacked a driver gene mutation were more likely to have high CRP than those with EGFR/ALK or other mutations (P=0.005). Specifically, 67.9% of CRP-high patients were in the mutation-negative group, whereas mutation-positive cases made up 39.6% of the CRP-low group, suggesting that oncogene-driven tumors (which often occur in never-smokers) had lower systemic inflammation. Furthermore, PD-L1 status data revealed a striking pattern: 38.6% of PD-L1-positive tumors were in patients with high CRP, compared to only 21.5% in the low-CRP group (P<0.001). Among PD-L1 positive cases, the median CRP was 13.0 mg/L, significantly higher than 3.8 mg/L for PD-L1-negative cases. This aligns with prior findings that CRP levels correlate with PD-L1 expression (12). High-CRP patients were also more likely to have been treated with immunotherapy (20.0% received ICI vs. 11.1% of low-CRP patients, P<0.001), presumably because they had advanced, PD-L1-positive disease requiring ICI, or perhaps reflecting an “inflamed” phenotype leading clinicians to choose immunotherapy.

Finally, the presence of distant metastases at diagnosis was associated with higher CRP: 70.1% of metastatic patients fell in the high-CRP group vs. 58.5% in low-CRP (P<0.001). Median CRP among those with metastases was 7.2 (IQR, 2.1–33.3) compared to 3.4 (IQR, 1.1–17.9) mg/L in non-metastatic patients. In summary, baseline CRP ≥10 mg/L was a marker of a more unfavorable patient profile, such as older, male, smokers, alcohol use with worse performance status, and more aggressive tumor features, such as advanced stage, metastasis, absence of targetable mutations and positive PD-L1 status.

Survival outcomes by CRP level

At the data cutoff (September 2025), the median follow-up was 84 months (reverse Kaplan-Meier). A total of 1,094 deaths occurred (81.7% of the cohort), and 1,272 progression events were recorded for the PFS analysis (95.0%). Kaplan-Meier survival curves revealed significant differences in both OS and PFS according to baseline CRP status.

As shown in Figure 2, patients with CRP <10 mg/L had a markedly longer OS compared to those with CRP ≥10 mg/L. The median OS for the low-CRP group was 44.0 months, whereas median OS in the high-CRP group was 19.0 months. This more than two-fold difference in median survival was statistically significant (log-rank P<0.001). The survival divergence appeared early and widened over time: at 1 year, OS rates were 84.9% in the low-CRP group vs. 66.8% in the high-CRP group; at 3 years, 54.5% vs. 26.8%; and at 5 years, 37.2% vs. 15.8%, respectively (log-rank P<0.001).

Figure 2 Overall survival in the low and high CRP group. CRP, C-reactive protein.

The hazard of death was clearly higher in the elevated CRP group throughout the follow-up period. Notably, even among patients with similar stages, those with high CRP tended to fare worse, suggesting CRP provides additional prognostic stratification beyond stage alone. For example, in subgroup analysis of stage IV lung cancer patients, median OS was 29.0 vs. 16.0 months for low- vs. high-CRP (log-rank P<0.001) (Figure S1). A similar trend was observed in limited-stage (I–III) disease, median OS was 65.0 vs. 27.0 months for low- vs. high-CRP (log-rank P<0.001) (Figure S2). Furthermore, in subgroup analysis of NSCLC patients, median OS was 50.0 vs. 26.0 months for low- vs. high-CRP (log-rank P<0.001) (Figure S3). Similar trend was also observed in SCLC patients, median OS was 19.0 vs. 13.0 months for low- vs. high-CRP (log-rank P<0.01) (Figure S4). These findings reinforce that baseline systemic inflammation as indicated by CRP is inversely related to survival probability.

Figure 3 demonstrates the impact of CRP on progression-free survival. Patients with low CRP enjoyed a longer duration of disease control from first-line therapy. Median PFS in the CRP <10 mg/L group was 12.0 months, compared to 8.0 months in the CRP ≥10 mg/L group. The difference was highly significant (log-rank P<0.001). At 1 year from treatment initiation, an estimated 45.8% of low-CRP patients were progression-free, vs. 28.4% of high-CRP patients. By 2 years, PFS rates were 23.1% vs. 12.2% for low vs. high CRP, respectively (log-rank P<0.001). In subgroup analysis of NSCLC patients, median PFS was 13.0 vs. 9.0 months for low- vs. high-CRP (log-rank P<0.001) (Figure S5). Similar results were also observed in SCLC patients, median PFS was 9.0 vs. 6.0 months for low- vs. high-CRP (log-rank P<0.01) (Figure S6).

Figure 3 Progression-free survival in the low and high CRP group. CRP, C-reactive protein.

The curves indicate that high baseline CRP was associated with a faster time to progression or relapse, consistent with more aggressive disease biology or possibly reduced treatment efficacy in those patients. These trends suggest that elevated CRP might be indicative of primary resistance to systemic therapies.

Univariate and multivariate Cox regression

The prognostic importance of CRP was further examined in Cox regression models (Tables 3,4). On univariate analysis, our study found that a number of factors were associated with shorter OS, such as age ≥65 years (HR 1.16, P=0.02), male gender (HR 1.27, P<0.001), underweight BMI (P=0.01), current smoking (HR 1.29, P<0.001), heavy alcohol use (HR 1.13, P=0.04), presence of chronic comorbidity (HR 1.14, P=0.03), SCLC histology (HR 1.58, P<0.001), advanced stage III–IV (HR 2.76, P<0.001), ECOG PS 2–4 (HR 1.73, P<0.001), negative driver mutation status (HR 1.24, P=0.001), presence of metastasis (HR 1.95, P<0.001) and CRP ≥10 mg/L (HR 1.64, 95% CI: 1.45–1.85, P<0.001). Many of these factors were interrelated (e.g., older patients more likely to have comorbidities and high CRP; smokers more likely male, etc.). We therefore constructed a multivariate Cox model to identify independent predictors of OS. In the multivariable analysis (Table 3), CRP ≥10 mg/L remained a significant independent prognostic factor for worse OS with an adjusted HR of 1.43 (P<0.001) after controlling for other covariates. Other independent OS predictors included: SCLC histology (HR ~1.54, P<0.001), stage III–IV (HR 1.85, P<0.001), ECOG 2–4 (HR 1.51, P<0.001), presence of metastases (HR 1.61, P<0.001), absence of driver mutation (HR 1.16, P=0.02) and low BMI (obese patients actually had slightly improved OS vs. underweight, HR ~0.63, P=0.001). Notably, factors like age, sex, smoking and PD-L1 status did not retain significance in the multivariate model once the major disease-related variables were included. This suggests that the effect of male sex or smoking on survival may be mediated through more aggressive disease (e.g., higher stage, more inflammation) rather than being independent risk factors. The C-statistic of the multivariate OS model improved marginally when CRP was added, indicating CRP contributes incremental prognostic value.

For progression-free survival, univariate Cox results were consistent with OS, male gender (HR 1.17, P=0.009), smoking (HR 1.19, P=0.002), SCLC histology (HR 1.41, P<0.001), advanced stage (HR 2.48, P<0.001), poor ECOG score (HR 1.56, P<0.001), metastasis (HR 1.99, P<0.001) and CRP ≥10 mg/L (HR 1.46, P<0.001) were associated with shorter PFS. In the multivariate PFS model (Table 4), CRP ≥10 mg/L remained an independent predictor of shorter PFS (adjusted HR 1.28, P<0.001). Other independent PFS predictors were SCLC histology (HR 1.38, P<0.001), stage III–IV (HR 1.66, P<0.001), ECOG 2–4 (HR 1.37, P<0.001) and presence of metastasis (HR 1.66, P<0.001). Interestingly, unlike OS, the presence of a driver mutation was not significantly associated with PFS on multivariate analysis (likely because many mutation-positive patients eventually progressed on targeted therapy despite longer OS due to subsequent lines of treatment). PD-L1 status and immunotherapy use did not show a significant independent effect on PFS in the overall analysis, possibly due to confounding by indication and limited follow-up for those treated with ICIs. Crucially, however, the persistence of CRP’s prognostic effect in both OS and PFS models underscores that CRP level is capturing a facet of disease aggressiveness not fully accounted for by traditional variables. Even among patients of the same stage and treatment, those with systemic inflammation (high CRP) had poorer outcomes.

In summary, our multivariate findings confirm that baseline CRP ≥10 mg/L is an independent adverse prognostic factor in lung cancer, associated with approximately 1.4-fold higher mortality risk and 1.3-fold higher risk of disease progression, controlling for other prognostic indicators.

Discussion

In this large single-institution cohort spanning over two decades, we found that an elevated baseline CRP level is strongly associated with more advanced and aggressive lung cancer and is an independent predictor of worse survival outcomes. To our knowledge, this study, with 1,339 patients, represents one of the most comprehensive analyses of CRP in a broad lung cancer population (including NSCLC and SCLC, various stages) in an Asian medical center. Our findings reinforce and extend the growing body of evidence that systemic inflammation, as quantified by CRP, plays a pivotal role in lung cancer progression and can serve as a clinically useful prognostic biomarker.

Consistent with prior reports, high CRP was linked to adverse baseline factors. High-CRP patients were older, predominantly male, and more likely to be smokers or alcohol users—demographics typically associated with inflammation and lung cancer risk. Notably, smoking is known to induce a chronic inflammatory state in the lung; our data show smokers had significantly higher CRP, which is in line with Koch study who also identified smoking status and CRP as independent prognostic factors in advanced NSCLC (20). We similarly observed that even after adjusting for smoking and other factors, CRP remained independently prognostic, highlighting that CRP captures more than just smoking-related risk.

Importantly, our study illustrates that CRP elevation aligns with key tumor characteristics: advanced stage, metastasis and lack of targetable gene mutations. Patients with high CRP were far more likely to have stage III–IV disease at diagnosis (95.3% vs. 85.1%), suggesting that systemic inflammation intensifies as the tumor burden increases. Indeed, the median CRP in stage I–II patients was only ~1.7 mg/L, near normal levels, whereas in stage IV it was ~6.4 mg/L. This gradient supports the concept that tumor load and invasion drive systemic inflammatory responses (via cytokines like IL-6, IL-1, TNF-α from tumor or host cells). Additionally, CRP was higher in PD-L1 positive tumors and SCLC (although the latter not significantly so), which are often biologically more aggressive. he observed association between CRP and PD-L1 positivity is intriguing. It echoes the findings of Akamine et al., who reported that high CRP and smoking history were jointly predictive the high expression of PD-L1 in NSCLC (21). One interpretation is that tumors with heavy immune infiltration or “inflamed” phenotype (which tend to express PD-L1) also stimulate systemic inflammation. Conversely, an alternative explanation could be that underlying inflammation and smoking exposure create a carcinogenic milieu resulting in tumors that are both immunogenic (triggering PD-L1 upregulation) and fast-growing, thereby elevating CRP. This interaction between host inflammation and tumor immune environment merits further mechanistic study.

The crux of our results is the significant survival disadvantage conferred by elevated CRP. We found an over 24-month decrement in median OS associated with CRP ≥10 mg/L. Even after accounting for stage, performance status, and other factors, high CRP carried a 43% higher risk of death. These findings corroborate numerous smaller studies and meta-analyses highlighting CRP as a robust prognostic factor. For instance, a meta-analysis by Tong study focusing on NSCLC patients on immunotherapy found an HR of ~1.9 for OS when comparing high vs. low CRP (22), which is comparable to our unadjusted HR of 1.64 in an unselected cohort. Our study extends those observations beyond the ICI setting to all-comers with lung cancer, indicating the prognostic value of CRP is broad and not treatment-specific. The consistency of CRP’s effect in both OS and PFS suggests it is reflecting an intrinsic disease aggressiveness (or perhaps general health state) that influences both how rapidly patients progress and how long they ultimately survive.

The relationship between CRP and treatment efficacy is clinically important. Although objective response was not consistently available for all patients in this retrospective dataset, the shorter PFS observed in the high-CRP group suggests faster disease progression and/or reduced durability of disease control. Prior studies have similarly shown that elevated CRP is associated with inferior outcomes across multiple systemic therapies, including ICIs (23). Future analyses that incorporate line-specific treatment regimens and objective response criteria (RECIST) could clarify whether CRP is a predictive marker of treatment benefit or primarily a general prognostic marker.

Our findings have practical clinical implications. CRP is a cheap and widely available blood test. Given how strongly it stratified survival in our study, clinicians could use CRP alongside traditional factors to refine prognostic assessments. For example, two patients with the same stage and performance status might have very different outcomes if one has CRP 2 mg/L and the other CRP 50 mg/L. Recognizing the high-CRP patient as higher-risk could prompt closer monitoring, early palliative care integration, or consideration of clinical trials for novel therapies. Moreover, CRP could be incorporated into prognostic scoring systems. The modified Glasgow Prognostic Score (mGPS), which combines CRP and albumin, is one such tool; high mGPS (CRP >10 mg/L + low albumin) has shown prognostic value in lung cancer (24-26). In our dataset, due to substantial missingness of albumin, a full-cohort comparison with mGPS was not feasible. While we did not directly evaluate mGPS here, the high prevalence of hypoalbuminemia in advanced cancer suggests that many high-CRP patients would also have a high mGPS, portending poor outcomes. Future work could validate combined indices in our population.

Limitations of this study include its retrospective single-center design, potential residual confounding (e.g., occult infection or other inflammatory conditions at baseline), and substantial treatment heterogeneity over the 22-year study period. PD-L1 testing and immunotherapy were introduced later in the study era, resulting in the high rate of missing data for PD-L1, particularly in the earlier years of the cohort, limits the generalizability of exploratory analyses. Results related to PD-L1 should be interpreted cautiously, and prospective validation in cohorts with systematic testing is needed. In addition, defining immunotherapy exposure as “ever received” may introduce immortal time bias; if treatment line and timing data are available, immunotherapy should be modeled as a baseline first-line variable or as a time-dependent covariate in sensitivity analyses. These issues should be considered when interpreting treatment-related associations, whereas the consistent prognostic value of CRP across OS and PFS supports its role as a robust risk marker.

Furthermore, our present study evaluates the prognostic value of baseline CRP rather than the full dynamic behavior of CRP over the treatment course. We acknowledge that a single pretreatment measurement cannot capture the trajectory, early on-treatment changes, variability over time, or treatment-related inflammatory dynamics—all of which may carry additional prognostic or predictive value. We agree that serial CRP analyses would be highly informative; however, repeated measurements were not available in a sufficiently standardized manner across the entire 22-year cohort to support a robust longitudinal analysis without introducing substantial selection bias. Future studies should prioritize serial CRP trajectory analysis as an important direction for prospective research.

Lastly, as a single-center study conducted at a national tertiary referral institution, this study is subject to unavoidable referral and case-mix bias. Our cohort likely over-represents patients with advanced disease, complex management needs, and access to specialized diagnostics and therapies, which may limit generalizability to community hospitals, primary care settings, and non-Chinese populations. Therefore, external validation in multicenter cohorts and other healthcare settings is necessary before the observed CRP effect size can be generalized.

Another area of interest is whether interventions to reduce inflammation could improve outcomes. Our study design cannot answer that, but it sets the stage for hypothesis generation. If CRP is not merely a bystander but a contributor to tumor progression (e.g., via IL-6/STAT3 pathways), then anti-inflammatory treatments like nonsteroidal anti-inflammatory drugs (NSAIDs), cytokine inhibitors (e.g., tocilizumab against IL-6) or lifestyle modifications (diet, exercise) might conceivably impact cancer outcomes. In fact, in 2023, the conclusion of the Christopher Azzoli study hinted at “inflammation-suppressing treatments” as a possible avenue to explore in NSCLC management (27). Prospective trials would be needed to test if lowering CRP (or targeting its upstream mediators) can translate into survival benefit.

Conclusions

Elevated baseline C-reactive protein is a clear indicator of adverse prognosis in lung cancer patients. In this 22-year retrospective cohort, CRP ≥10 mg/L was associated with high-risk clinical features and independently predicted shorter progression-free and OS. These findings underscore the value of CRP as a readily available biomarker that reflects the interplay between tumor biology and host inflammatory response. Incorporating CRP into routine risk stratification could help identify patients with aggressive disease who might benefit from intensified treatment or closer follow-up. Moreover, our results provide a rationale for further research into anti-inflammatory interventions as a component of lung cancer therapy. Ultimately, measuring an inflammatory marker like CRP moves us toward a more holistic approach in oncology, one that acknowledges the host-tumor interaction as a determinant of outcome and a potential therapeutic target. Future prospective studies and clinical trials should validate the prognostic utility of CRP and explore whether modulating the systemic inflammatory milieu can improve the trajectory of patients with lung cancer.

Acknowledgments

The authors would like to thank all medical staff at the Lung Cancer Center of Peking Union Medical College Hospital.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0118/rc

Data Sharing Statement: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0118/dss

Peer Review File: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0118/prf

Funding: This research was supported by the CSCO-Hengrui Oncology Research Fund Project (No. Y-HR2019) and Bethune Charitable Foundation Advanced Solid Tumor Research Funding (No. STLKY0056).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2026-1-0118/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethical Committee of Peking Union Medical College Hospital (No. K7024-K24C3469). Informed consent was obtained from the participants, explicitly granting permission for the publication of their anonymized data in this paper.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Jokhadze N, Das A, Dizon DS. Global cancer statistics: A healthy population relies on population health. CA Cancer J Clin 2024;74:224-6. [Crossref] [PubMed]
2. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
3. Han B, Zheng R, Zeng H, et al. Cancer incidence and mortality in China, 2022. J Natl Cancer Cent 2024;4:47-53. [Crossref] [PubMed]
4. Vicidomini G. Current Challenges and Future Advances in Lung Cancer: Genetics, Instrumental Diagnosis and Treatment. Cancers (Basel) 2023;15:3710. [Crossref] [PubMed]
5. Rami-Porta R, Nishimura KK, Giroux DJ, et al. The International Association for the Study of Lung Cancer Lung Cancer Staging Project: Proposals for Revision of the TNM Stage Groups in the Forthcoming (Ninth) Edition of the TNM Classification for Lung Cancer. J Thorac Oncol 2024;19:1007-27. [Crossref] [PubMed]
6. Balkwill F, Mantovani A. Inflammation and cancer: back to Virchow? Lancet 2001;357:539-45. [Crossref] [PubMed]
7. Pęczek P, Gajda M, Rutkowski K, et al. Cancer-associated inflammation: pathophysiology and clinical significance. J Cancer Res Clin Oncol 2023;149:2657-72. [Crossref] [PubMed]
8. Favaro E, Amadori A, Indraccolo S. Cellular interactions in the vascular niche: implications in the regulation of tumor dormancy. APMIS 2008;116:648-59. [Crossref] [PubMed]
9. Marnell L, Mold C, Du Clos TW. C-reactive protein: ligands, receptors and role in inflammation. Clin Immunol 2005;117:104-11. [Crossref] [PubMed]
10. Gwenzi T, Schrotz-King P, Anker SC, et al. Post-operative C-reactive protein as a strong independent predictor of long-term colorectal cancer outcomes: consistent findings from two large patient cohorts. ESMO Open 2024;9:102982. [Crossref] [PubMed]
11. Buğday İ, İnanç M, Özkan M, et al. Prognostic impact of elevated C-reactive protein and procalcitonin in patients with extensive-stage small cell lung cancer. Anticancer Drugs 2025;36:135-9. [Crossref] [PubMed]
12. Villaseñor A, Flatt SW, Marinac C, et al. Postdiagnosis C-reactive protein and breast cancer survivorship: findings from the WHEL study. Cancer Epidemiol Biomarkers Prev 2014;23:189-99. [Crossref] [PubMed]
13. Lopez-Pastorini A, Tatli Z, von Bargen A, et al. The Prognostic Value of Preoperative C-Reactive Protein Levels in Resected Early-Stage Lung Cancer. J Surg Res 2025;305:85-92. [Crossref] [PubMed]
14. Wang J, Zhang X, Gao Q, et al. Early C-reactive protein reduction predicts survival in real-world extensive-stage small cell lung cancer treated with first-line adebrelimab-based immunotherapy. Front Oncol 2025;15:1709336. [Crossref] [PubMed]
15. Wang H, Yang R, Zhou K, et al. Association between pretreatment C-reactive protein level and survival in non-small cell lung cancer patients treated with immune checkpoint inhibitors: A meta-analysis. Int Immunopharmacol 2023;124:110937. [Crossref] [PubMed]
16. Leuzzi G, Galeone C, Gisabella M, et al. Baseline C-reactive protein level predicts survival of early-stage lung cancer: evidence from a systematic review and meta-analysis. Tumori 2016;102:441-9. [Crossref] [PubMed]
17. Onodera R, Chiba S, Nihei S, et al. High level of C-reactive protein as a predictive factor for immune-related adverse events of immune checkpoint inhibitors in non-small cell lung cancer: a retrospective study. J Thorac Dis 2023;15:4237-47. [Crossref] [PubMed]
18. Issue Information-Declaration of Helsinki. J Bone Miner Res 2019;34. BMi-BMii.
19. Iivanainen S, Ahvonen J, Knuuttila A, et al. Elevated CRP levels indicate poor progression-free and overall survival on cancer patients treated with PD-1 inhibitors. ESMO Open 2019;4:e000531. [Crossref] [PubMed]
20. Koch A, Fohlin H, Sörenson S. Prognostic significance of C-reactive protein and smoking in patients with advanced non-small cell lung cancer treated with first-line palliative chemotherapy. J Thorac Oncol 2009;4:326-32. [Crossref] [PubMed]
21. Akamine T, Takada K, Toyokawa G, et al. Association of preoperative serum CRP with PD-L1 expression in 508 patients with non-small cell lung cancer: A comprehensive analysis of systemic inflammatory markers. Surg Oncol 2018;27:88-94. [Crossref] [PubMed]
22. Tong W, Xu H, Tang J, et al. Inflammatory factors are associated with prognosis of non-small cell lung cancer patients receiving immunotherapy: a meta-analysis. Sci Rep 2024;14:26102. [Crossref] [PubMed]
23. Kuusisalo S, Tikkanen A, Lappi-Blanco E, et al. The prognostic and predictive roles of plasma C-reactive protein and PD-L1 in non-small cell lung cancer. Cancer Med 2023;12:16087-97. [Crossref] [PubMed]
24. Saal J, Bald T, Eckstein M, et al. Integration of on-treatment modified Glasgow prognostic score (mGPS) to improve imaging-based prediction of outcomes in patients with non-small cell lung cancer on immune checkpoint inhibition. Lung Cancer 2024;189:107505. [Crossref] [PubMed]
25. Fan H, Shao ZY, Xiao YY, et al. Comparison of the Glasgow Prognostic Score (GPS) and the modified Glasgow Prognostic Score (mGPS) in evaluating the prognosis of patients with operable and inoperable non-small cell lung cancer. J Cancer Res Clin Oncol 2016;142:1285-97. [Crossref] [PubMed]
26. Abbass T, Dolan RD, MacLeod N, et al. Comparison of the prognostic value of MUST, ECOG-PS, mGPS and CT derived body composition analysis in patients with advanced lung cancer. Clin Nutr ESPEN 2020;40:349-56. [Crossref] [PubMed]
27. Azzoli C, Huynh L, Yi D, et al. Retrospective Study to Examine Prognostic Value of C-Reactive Protein in Patients With Surgically Resectable Non-Small-Cell Lung Cancer. Clin Lung Cancer 2023;24:329-38. [Crossref] [PubMed]