If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Baseline and Dynamic Changes in Hemoglobin Levels Predict Treatment Response and Prognosis in Metastatic Renal Cell Carcinoma: A Multicenter Retrospective Study
Department of Oncology, Netherlands Cancer Institute, Amsterdam, The NetherlandsDepartment of Urology, The Royal Free London NHS Foundation Trust, University College London Division of Surgery and Interventional Science, London, UK
Address for correspondence: Bernadett Szabados, MD, Centre for Experimental Cancer Medicine, Barts Cancer Institute, Queen Mary University of London, Old Anatomy Building, Charterhouse Square, London EC1M 6BQ, UK
Clinical markers of response in metastatic renal cell carcinoma (mRCC) are lacking. Low hemoglobin (Hb) is associated with poor outcomes in the IMDC risk score. This study evaluates the role of Hb as a marker of treatment outcomes in mRCC.
Patients and Methods
This multicenter retrospective study evaluated 276 patients with mRCC treated with frontline immune checkpoint inhibitor (ICI) therapy, ICI and vascular endothelial growth factor (VEGF) inhibitor (VEGFI) combinations (ICI/VEGFI), or VEGFI monotherapy between 2014 and 2021. Hb levels at baseline, week 6 and 12 and at disease progression or death were recorded. Patients were categorized as responders (CR+PR) or nonresponders (SD+PD) using cross-sectional imaging at week 12. The association between baseline and dynamic changes in Hb and oncological outcomes was assessed.
Results
Thirty-seven percent, 40% and 22% of patients received ICIs, ICI/VEGFI and VEGFI respectively. In patients receiving ICIs, there was a significant increase in Hb amongst responders from baseline to week 12 (P= .02). Amongst patients receiving ICI/VEGFI, there was an increase in Hb from baseline to week 12 which was greater in responders (P< .001). In patients receiving VEGFI monotherapy, responders had a higher Hb at baseline (P= .01), week 6 (P= .04), and week 12 (P= .003). An increase in Hb was a significant independent predictor of progression-free survival amongst patients receiving ICIs (HR 0.40, 95%CI, 0.19-0.83, P= .009).
Conclusion
Baseline and dynamic changes in Hb are associated with first-line treatment outcomes in patients with mRCC and represent a pragmatic early serological marker.
Sunitinib for the first-line treatment of advanced and/or metastatic renal cell carcinoma. Technology appraisal guidance [TA169] NICE. Available at: https://www.nice.org.uk/guidance/ta169.Accessed: April 1, 2022.
Pazopanib for the first-line treatment of advanced renal cell carcinoma. Technology appraisal guidance [TA215] NICE. Available at: https://www.nice.org.uk/guidance/ta215.Accessed: April 1, 2022.
have demonstrated the superiority of Ipilimumab and Nivolumab over monotherapy with VEGFIs such as Sunitinib in intermediate and poor-risk disease, as defined by the International mRCC Database Consortium (IMDC) risk model.
External validation and comparison with other models of the international metastatic renal-cell carcinoma database consortium prognostic model: a population-based study.
Novel combinations of ICI and VEGFI (ICI/VEGFI) have also proven effective with higher response rates and improved survival across all IMDC risk groups. Notably, Pembrolizumab/Axitinib, Pembrolizumab/Lenvatinib and Nivolumab/Cabozantinib have shown durable overall survival (OS) benefits and are now licensed in the first-line setting for the treatment of mRCC.
There has therefore been increased interest in developing prognostic markers, in order to identify patients who would benefit most from each treatment modality. Much of this research to date has focused on translational biomarkers, which are costly and often not deliverable in real world clinical practice.
In comparison, there is a relative paucity of established clinical markers, particularly those which can be evaluated at an early stage to predict patient outcome. Serological parameters which appear to correlate with a poorer prognosis in mRCC
Neutrophil-to-lymphocyte ratio, Platelet-to-lymphocyte ratio, and C-reactive protein as new and simple prognostic factors in patients with metastatic renal cell cancer treated with tyrosine kinase inhibitors: a systemic review and meta-analysis.
Pretreatment hemoglobin level as a predictor to evaluate the efficacy of immune checkpoint inhibitors in patients with advanced non-small cell lung cancer.
External validation and comparison with other models of the international metastatic renal-cell carcinoma database consortium prognostic model: a population-based study.
Hb represents an attractive clinical marker of response in mRCC patients as it can be measured readily and applied pragmatically in clinical practice. In this work, we specifically aim to explore the relationship between baseline hemoglobin levels, their dynamic changes and outcomes in patients with mRCC treated with contemporary immune-based therapies.
Patients and Methods
Data on patients with previously untreated mRCC was collected retrospectively from 3 institutions: (1) Barts Health, London, UK; (2) Netherlands Cancer Institute, Amsterdam, the Netherlands; (3) Beatson West of Scotland Cancer Centre, Glasgow, Scotland, in the context of a clinical audit. Patients were categorized based on first-line treatment into 3 groups: ICI combination therapy, ICI in combination with VEGFI (ICI/VEGFI), or VEGFI monotherapy (VEGFI) (Figure 1A).
Figure 1(A) Consort diagram of patients included in the audit. (B) Summary of patient assessment. Hemoglobin levels were collected at the start of treatment (Baseline) and at Week 6 and 12 of treatment. Radiographic assessment was performed at Week 12 and patients were classified as Responders and Nonresponders according to RECIST criteria. Following this, hemoglobin levels were also collected at the time of disease progression or death.
Baseline Hb levels at start of treatment, and serial levels at 6 and 12 weeks of treatment were measured. First radiological disease assessment with computed tomography (CT) was performed at 12 weeks. Radiographic response was assessed according to RECIST 1.1 criteria where available and patients were deemed to have either complete response (CR), partial response (PR), stable disease (SD) or progressive disease (PD). Patients were categorized based on CT findings as responders (CR or PR) or nonresponders (SD or PD) (Figure 1B).
Patients were subsequently followed up for progression-free survival (PFS) and OS. For patients who experienced disease progression or death, Hb levels at the most recent blood test prior to the event were collected. Patients were defined to have an increase or decrease in Hb from baseline if their Hb was higher or lower at the time of disease progression or death than that at start of treatment.
Patients were excluded if they received a blood transfusion within 12 weeks of commencing treatment or had insufficient data for analysis, defined as a lack of 2 or more serial Hb measurements or missing data on radiological assessment at 12 weeks. Other criteria for exclusion were more than 2 blood tests being performed out of schedule from the planned dates at baseline, 6 and 12 weeks, and outlier values following careful statistical analysis to minimize exclusion bias.
A mixed ANOVA model was applied to assess the variance of Hb levels due to the combined effects of time and response to treatment. Significant combined or individual effects were followed up using 1-way ANOVA and pairwise tests. Data was tested for normality using the Shapiro test, homogeneity of variance using the Levene test and homogeneity of covariance using the Box's M test. Nonparametric data was analyzed using robust nonparametric equivalents such as the Wilcoxon signed-rank test.
The Kaplan-Meier method was used to estimate PFS and OS for patients who had an increase vs. decrease from baseline Hb, and the log-rank test used to assess survival differences between the groups. The Cox proportional hazards model was used for univariate and multivariate regression analysis to determine the significance of individual variables on PFS. P-values of less than .05 were considered statistically significant, and where appropriate, the Bonferroni method was used to adjust these to account for multiple hypothesis testing. All analyses were conducted in R version 4.1.1. The WRS2 package was used for robust nonparametric tests.
Data was collected for 276 patients with mRCC who received treatment between January 2014 and July 2021 at 3 tertiary centers: (1) Barts Health (n = 134); (2) Netherlands Cancer Institute (n = 66); (3) Beatson West of Scotland Cancer Centre (n = 76). Thirty-five patients were excluded, resulting in 241 patients being included in the final analysis (Figure 1A).
Baseline clinicopathologic characteristics and follow up data are included in Table 1, and reflect the standard treatment-naïve mRCC population. The median age at start of treatment was 62 years. Most patients were male (n = 184, 76%) and had clear cell RCC (ccRCC) (n = 215, 89%). The majority of patients initially presented with metastatic disease (n = 173, 72%) and had ECOG 0-1 performance status prior to starting treatment (n = 227, 95%). Overall, 90 (37%) patients received Ipilimumab and Nivolumab. Ninety-seven (40%) patients received combined ICI/VEGFI, most of whom were treated with Pembrolizumab and Axitinib (n = 75, 77%) and 54 (22%) patients received VEGF monotherapy, mostly Sunitinib (n = 42, 78%).
Table 1Summary of Baseline Patient Characteristics
The median follow-up was 9.7 months. At the time of analysis, 135 (56%) patients experienced disease progression (60%, 35%, and 87% in the ICI, ICI/VEGFI and VEGFI groups respectively). Twenty-seven (11%) patients died (6%, 22%, and 2% in the ICI, ICI/VEGFI and VEGFI groups respectively). Median PFS for the ICI, ICI/VEGFI and VEGFI cohorts was 10.3 (95%CI, 5.3-13.8), 17.9 (95%CI, 15.2-24.9) and 10.7 (95%CI, 6.2-15.2) months respectively. Median OS was not reached in all 3 cohorts (Supplemental Figure 1).
The majority of blood tests at baseline, week 6 and week 12 were performed within 1 week of the expected blood test date for all groups, with a median deviation of 1 day for all timepoints (Supplemental Table 1).
The mean Hb across all timepoints in the ICI, ICI/VEGFI and VEGFI groups was 91.8, 137.3, and 127.6 g/L respectively. Across all timepoints, 82%, 19%, and 34% of each group had Hb levels below the lower limit of normal at 120 g/L.
Amongst the 90 patients receiving ICI, 32 (36%) responded to treatment at the 12-week disease assessment (CR and PR). Responders had a numerically higher mean baseline Hb than nonresponders, although this difference was not statistically significant (P= 0.08) (Supplemental Figure 2A). After commencing treatment, there was a significant increase in Hb levels in responding patients at 12 weeks (P= .02) after starting therapy (Figure 2A) which resulted in a significant difference in Hb levels between responders and nonresponders at week 6 (P= .01) and week 12 (P= .01) (Supplemental Figure 2A). There was no difference in Hb levels in patients in the nonresponding group (SD and PD) (P= 1) (Figure 2A).
Figure 2Comparison of changes in hemoglobin levels over time (between Baseline, Week 6, and Week 12) amongst responders and nonresponders to treatment with (A) ICI only (B) ICI/VEGFI, (C) VEGFI only. Line plots represent mean and standard error, with interconnecting lines demonstrating directory of change. P-values annotated. ns: nonsignificant; *:< .05; **:< .01, ***:< .001, ****:< .0001. Dotted red line indicates lower limit of normal for hemoglobin levels.
Amongst the 97 patients receiving combination ICI/VEGFI, 55 (57%) responded. At baseline, responders had a lower Hb level than nonresponders although this difference was not significant (P= .3) (Supplemental Figure 2B). There was a trend towards increasing Hb levels from baseline to week 12 for both groups, although a greater increase was observed in responders (P< .001) compared to nonresponders (P= .04) (Figure 2B). Additionally, a small albeit nonsignificant decrease in Hb (P = 1.0) was noted amongst nonresponders between week 6 and 12 which was not seen amongst responders (Figure 2B). Consequently, the mean Hb of responders increased above that of nonresponders at week 6 and week 12, although these differences were not statistically significant (P= .8 and P= .3 at week 6 and 12) (Supplemental Figure 2B).
Amongst the 54 patients receiving VEGFI monotherapy, there were 17 (31%) responders. Responders had a significantly higher Hb level than nonresponders at all three timepoints (Baseline: P= .004; Week 6: P= .01; Week 12: P= .001) (Supplemental Figure 2C). From baseline to week 12, both groups demonstrated a reduction in Hb levels. However, this trend was only significant in nonresponders from baseline to week 12 (P= .03) and from week 6 to 12 (P= .03) (Figure 2C).
We next examined the association between Hb levels and survival. Patients receiving ICI who had an increase in Hb had a significantly longer PFS (mPFS: 10.4 vs. 2.6, P= .02) (Figure 3A) with a reduced risk of disease progression (HR 0.46, 95%CI, 0.24-0.9, P= .01) (Table 2) compared to those with a decrease in Hb. There was no significant difference in OS between patients who had an increase versus decrease in Hb (Median: not reached vs. 21.3, P= .4; Risk of death: HR 0.43, 95%CI, 0.05-3.93, P= .5) (Supplemental Figure 3A). Univariate analysis of PFS was performed on individual variables of age, sex, IMDC prognostic risk, ECOG performance status, prior nephrectomy, initial presentation with upfront metastatic disease, and increase in Hb levels from baseline. Only IMDC prognostic risk (P= .03), ECOG performance status (P< .05) and an increase in Hb levels from baseline were significant determinants of PFS (P= .01). When adjusting for IMDC prognostic risk and ECOG performance status using Cox regression analysis, an increase in Hb remained significant for reduced risk of disease progression (HR 0.40, 95%CI, 0.19-0.83, P= .009) (Table 2).
Figure 3Progression-free survival for patients receiving treatment with (A) ICI only (B) ICI/VEGFI, and (C) VEGFI only, stratified by those who experienced an increase compared to a decrease in hemoglobin levels from baseline. Median survival times indicated by dashed lines. Log-rank P-values annotated.
In contrast, amongst patients receiving ICI/VEGFI and VEGFI monotherapy, an increase in Hb was not noted to be positively prognostic. When comparing patients receiving ICI/VEGFI who experienced an increase versus a decrease in Hb, there was no significant difference in PFS (Median: 5.5 vs. 8.3 months, P= .3; Risk of disease progression: HR 1.40, 95%CI, 0.78-2.5, P= .3) (Figure 3B) or OS (Median: not reached vs. 15.5 months, P= .9; Risk of death: HR 0.92, 95%CI, 0.39-2.2, P= .9) (Supplemental Figure 3B). Similarly, patients receiving VEGFI monotherapy who experienced an increase compared to a decrease in Hb did not have a significant difference in PFS (Median: 11.0 vs. 7.4 months, P= .8; Risk of disease progression: HR 0.92, 95%CI, 0.49-1.70, P= .8) (Figure 3C) or OS (Median: not reached vs. not reached, P= .5; Risk of death not assessable) (Supplemental Figure 3C).
Discussion
In this large international multicenter retrospective clinical audit of 276 patients, we demonstrate that Hb levels are a valuable clinical marker of outcomes in patients receiving first-line therapy for mRCC.
In patients who received front-line combination immune checkpoint inhibitors with Ipilimumab and Nivolumab, responders had a higher baseline Hb (P= .08) and demonstrated a significant increase in Hb levels at week 12 (P= .02), resulting in significantly higher Hb levels than nonresponders at week 6 (P= .01) and 12 (P= .01). An increase in Hb levels was also associated with a significantly longer PFS on multivariate analysis (HR 0.40, 95%CI, 0.19-0.83, P= .009). Taken together, these observations suggest that dynamic changes in Hb levels from as early as 6 weeks can predict treatment response and PFS amongst patients with mRCC treated with first-line Ipilimumab and Nivolumab. This finding is highly relevant as patients receiving ICIs have lower response rates compared to those treated with ICI/VEGFI combination therapies. Therefore, dynamic changes in Hb levels represent an early marker of response, which can be readily applied to everyday clinical practice.
Amongst patients receiving ICI/VEGFI, there were no significant differences in Hb levels at baseline (P= .3) and both responders and nonresponders demonstrated an increase and normalization of Hb levels by week 12 (P< .0001 and P= .04 respectively). Due to the high clinical benefit rate (CR+PR+SD) (n = 85, 88%) of ICI/VEGFI combination therapy and consequent small number of patients with upfront disease progression and reduced Hb levels at week 12 (n = 12, 12%), it is plausible that the overall signal towards reduced Hb levels was therefore not observed amongst nonresponders in this group, in contrast to patients treated with ICIs. There was no association between Hb levels and survival (PFS P= .3, OS P= .9).
In patients treated with VEGFI, baseline Hb levels were significantly higher amongst responders compared to nonresponders (P= .004). Both responders and nonresponders experienced a decrease in Hb levels, a well reported side effect of VEGFI therapy. However, this trend was only significant amongst non-responders with a reduction in Hb from baseline to week 12 (P= .03) and from week 6 to 12 (P= .03). The reduction in Hb was not statistically significant in the responding group (P= .1) which supports previously reported findings.
There were no differences in PFS or OS when stratified by changes in Hb levels.
A key highlight from our study is that amongst patients with mRCC receiving first-line ICI therapy, increasing Hb levels is associated with treatment response and survival. There are several plausible mechanisms for this. Firstly, anemia has been demonstrated to generate a hypoxic tumor microenvironment
patients with lower Hb levels may have higher levels of dysregulated systemic inflammation, a state which has been reported to reduce the effectiveness of immunotherapy.
Neutrophil-to-lymphocyte ratio, Platelet-to-lymphocyte ratio, and C-reactive protein as new and simple prognostic factors in patients with metastatic renal cell cancer treated with tyrosine kinase inhibitors: a systemic review and meta-analysis.
levels have both previously been associated with a poor response to both ICI and VEGFI treatment. However, these serological markers do not appear to differentiate between treatment response to these frontline options. In our study, increases in Hb are associated with treatment response to ICIs but not VEGFI monotherapy, suggesting that Hb may be advantageous as a clinical marker in differentiating between patients who would benefit from ICIs rather than VEGFI.
Finally, our study demonstrates that baseline Hb levels are associated with treatment response to VEGFI monotherapy, as compared to previous studies, which have examined the effect of dynamic changes in Hb rather than baseline levels.
This study has several limitations. Although it demonstrates a clear association between Hb levels and treatment response, due to the observational and retrospective nature, it is not possible to infer causation and temporality between these variables. Additionally, the follow up time was relatively short with a median of 9.7 months for all patients, which may limit the validity of survival analyses. In terms of the patient cohort, both ccRCC and non-ccRCC patients were included. Due to the relatively low number of patients with non-ccRCC no subgroup analyses were performed. As a consequence of regional referral patterns, our cohort includes a high percentage of patients with upfront metastatic disease and clear cell histology. Hence, our results may be more applicable in this patient population. Additionally, patients receiving ICI/VEGFI and VEGFI monotherapy were predominantly treated with Pembrolizumab plus Axitinib and Sunitinib respectively, which may limit the applicability of these findings to other ICI/VEGFI combinations or other VEGFI monotherapies.
Conclusion
A higher baseline Hb prior to therapy and dynamic increases in Hb levels during the first 12 weeks of treatment appear to be early markers of response to ICIs, occurring before formal radiological evaluation. Increasing Hb levels are also associated with longer PFS. These findings are readily applicable to routine clinical practice and may help clinicians deliver more personalized care. Further prospective studies are required to determine the role of Hb as a marker of treatment response and prognosis in mRCC.
Clinical Practice Points
•
In this paper, we report on hemoglobin (Hb) levels as a marker of outcomes in metastatic renal cell carcinoma (mRCC).
•
While mRCC remains largely incurable, treatment outcomes have improved significantly in recent times through the development of novel frontline ICIs and (VEGFI). In parallel, there has been increased interest in prognostic markers to identify patients who would benefit most from each treatment modality. Although tissue-based biomarkers have been extensively studied, there has been much less research into clinical markers of response. Current evidence suggests that Hb levels could predict treatment outcomes in mRCC: low Hb is negatively prognostic in the International mRCC Database Consortium risk score, and has been associated with reduced response rates to ICI therapy in other tumor types.
•
To our knowledge, our research is the first to directly demonstrate a clear association between Hb levels and treatment outcomes in mRCC across 3 established frontline treatment options: ICI therapy, ICI, and VEGFI combination therapy and VEGFI monotherapy. These results are particularly marked amongst patients treated with ICIs, where a higher baseline Hb and increasing Hb during treatment appear to predict response to therapy. Additionally, an increase in Hb amongst these patients seems to be an independent predictor of progression-free survival.
•
Our findings provide a mandate for further prospective validation of Hb as a clinical marker. If established, it represents a readily available and cost-effective parameter that can be integrated into routine clinical practice, helping clinicians to stratify and personalize treatment for mRCC.
Author Contributions
Yu-Hsuen Yang: conceptualization, data collection, data analysis, interpretation of results, original draft preparation, reviewing, and editing. Sonam Ansel: data collection, supervision, reviewing, and editing. Aafke Meerveld-Engink: data collection. Francesca Jackson-Spence: data collection, reviewing, and editing. Kathrine Rallis: data collection, interpretation of results. Paul Brian: data collection. Julia Choy: data collection. Christopher Sng: data collection. Philip Adeniran: data collection. Jubel Amin: data collection. Sarah Galope: data collection. Naomi Anderson: data collection. Axel Bex: conceptualization. Thomas Powles: conceptualization, supervision, reviewing, and editing. Balaji Venugopal: conceptualization, supervision, reviewing, and editing. Bernadett Szabados: conceptualization, supervision, interpretation of results, original draft preparation, reviewing, and editing.
Disclosure
The authors have stated that they have no conflicts of interest.
Acknowledgment
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Previous Publication
The results within this paper have been presented in part at the ESMO Targeted Anticancer Therapies Congress 2022, and appears as a published abstract in the Annals of Oncology.
Yang Y-H, Meerveld-Eggink A, Bex A, et al: 18P Baseline and dynamic changes in hemoglobin levels predict treatment response and disease progression in metastatic renal cell carcinoma. Annals of Oncology 33:S10, 2022.
References
Heng DYC
Choueiri TK
Rini BI
et al.
Outcomes of patients with metastatic renal cell carcinoma that do not meet eligibility criteria for clinical trials.
Sunitinib for the first-line treatment of advanced and/or metastatic renal cell carcinoma. Technology appraisal guidance [TA169] NICE. Available at: https://www.nice.org.uk/guidance/ta169.Accessed: April 1, 2022.
Pazopanib for the first-line treatment of advanced renal cell carcinoma. Technology appraisal guidance [TA215] NICE. Available at: https://www.nice.org.uk/guidance/ta215.Accessed: April 1, 2022.
External validation and comparison with other models of the international metastatic renal-cell carcinoma database consortium prognostic model: a population-based study.
Neutrophil-to-lymphocyte ratio, Platelet-to-lymphocyte ratio, and C-reactive protein as new and simple prognostic factors in patients with metastatic renal cell cancer treated with tyrosine kinase inhibitors: a systemic review and meta-analysis.
Pretreatment hemoglobin level as a predictor to evaluate the efficacy of immune checkpoint inhibitors in patients with advanced non-small cell lung cancer.
00035-6&id=gr1.jpg){kind=link}
00035-6&id=gr2.jpg){kind=link}
00035-6&id=gr3.jpg){kind=link}