Bernard Escudier, Institut Gustave Roussy, 114 rue Edouard Vaillant, 94805 Villejuif, France. e-mail: email@example.com
Study Type – Prognosis (cohort series)
Level of Evidence 2b
What's known on the subject? and What does the study add?
In the literature, few studies have evaluated the role of tumour burden (TB) in metastatic real cell carcinoma (mRCC), even though it has been considered as important in localized tumours. In metastatic patients the role of TB is uncertain because it was analyzed in chemotherapy treated patients or using a partial evaluation of TB.
This study, first reports the independent prognostic and predictive role of TB in mRCC patients treated with targeted agents in prospective clinical trials. TB is able to predict prognosis independently to localization of metastases and prognostic class defined by MSKCC criteria, moreover it is strictly related to patient's performance status.
• To investigate the possible prognostic role of baseline tumour burden (TB) in terms of progression-free survival (PFS) and overall survival (OS), in patients with metastatic renal cell carcinoma (mRCC).
PATIENTS AND METHODS
• A homogenous group of patients with mRCC enrolled in second-line trials post-cytokine treatment were selected for the present analysis.
• The Response Evaluation Criteria in Solid Tumors (the sum of the longest unidimensional diameter of each target lesion) were used to assess TB.
• The PFS and OS rates were estimated using the Kaplan–Meier method and compared across the groups using the log-rank test.
• The association between TB and PFS or OS was evaluated using a Cox proportional hazards model. Multivariable analyses were adjusted for other prognostic variables: the Memorial Sloan Kettering Cancer Centre (MSKCC) risk class and treatment.
• A total of 124 patients were included in the final analysis. Of these, 66% received sorafenib or sunitinib and 34% received placebo. The median follow-up was 80.1 month.
• TB was directly related to PFS and OS and these associations remained significant after adjusting for modified MSKCC risk class and treatment,.
• Each 1-cm increase in TB increased the risk of progression by 4.5% (hazard ratio [HR]: 1.05; 95% confidence interval [CI] 1.02–1.07; P < 0.001) and the risk of death by 5% (HR: 1.05; 95% CI 1.03–1.08; P < 0.001).
• TB is easy to calculate from standard computed tomography and significantly relates to OS in patients with mRCC.
• We report for the first time the independent prognostic role of baseline TB in multivariate analysis.
• We believe that this information could be translated into clinical practice.
In developed countries RCC is the tenth greatest cancer-related cause of death, with an estimated 60 920 new cases expected in the USA in 2011 and 46 151 new cases in the European Union in 2008 [1,2]. Generally, 30% of new cases are diagnosed at an advanced stage and 20% become metastatic during follow-up [3,4].
Although the prognosis of patients with metastatic RCC (mRCC) has been dramatically improved since targeted therapies have been introduced into the clinical armamentarium, the median overall survival (OS) is still limited, with very few patients reaching cure and around 20% getting no benefit from available therapies [5,6].
Prognostic models have been described to help to stratify the patient risk of cancer progression or death. These prognostic models are very useful for the selection of patients for clinical trials to avoid selection bias and erroneous interpretation of results. Ideally, they should also help to better select which patients will benefit from a given treatment and which drug to use in an individual patient.
During the last 10 years, several classifications have been proposed to assess the prognosis of mRCC. The most commonly used has been the Memorial Sloan Kettering Cancer Centre (MSKCC) classification , which has been used to stratify patients in the majority of the registration trials. This model stratifies patients using biological variables, such as haemoglobin, corrected calcium and lactate dehydrogenase, rather than using nephrectomy and performance status. Other models, such as the French classification, have been further developed in the cytokine era and then validated in patients treated with target agents .
Recently, a new model has been built for patients receiving target agents. In this new model, baseline values of neutrophils and platelets count have been added to MSKCC variables .
So far, tumour burden (TB) has not been considered in any of these models as a possible prognostic factor. Although the number of metastatic sites, reported as an independent prognostic factor by Mekhail et al. , may be considered a surrogate of TB, this is limited by the major differences in TB that could be present in the same metastatic site of different patients.
The aim of the present study was to investigate the possible prognostic role of baseline TB in a homogeneous group of patients treated with anti-angiogenic drugs within clinical trials.
MATERIALS AND METHODS
Patients eligible for this analysis were patients with mRCC enrolled at the Institut Gustave Roussy Villejuif, France in two prospective trials of targeted agents (sorafenib or sunitinib) in second-line post-cytokine treatments [11,12]. In the first trial, patients received either sorafenib (400 mg twice daily) or placebo, while in the second trial all patients received sunitinib at the continuous daily dose of 37.5 mg.
The inclusion criteria for these two studies were similar: age >18 years with histologically proven clear-cell mRCC and evidence of measurable disease (≥1 cm), based on Response Evaluation Criteria in Solid Tumors (RECIST v 1.0) . All patients had failed one previous cytokine-based regimen in the metastatic setting, had Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1, and adequate haematological, hepatic and renal function.
In both trials, for each patient the prognostic risk group was defined before the start of the anti-angiogenic therapy or placebo, according to the modified MSKCC score , using baseline values of serum lactate dehydrogenase (LDH), haemoglobin level, corrected serum calcium level, as well as time from initial diagnosis of RCC to randomization, Karnofsky performance score and number of metastatic sites.
Baseline CT was performed within 28 days before the start of therapy. For each patient the TB was defined as the sum of the longest unidimensional diameter of each target lesion (in cm) and restricted to axial CT imaging, with a maximum of 10 lesions, according to RECIST 1.0. The measures of targeted lesions were performed prospectively during the studies and retrospectively related with survival and other prognostic factors in the current analysis.
Baseline values were expressed as median values with interquartile ranges (IQRs). Baseline was defined as the date of start of experimental treatment: sorafenib, sunitinib or placebo. Progression-free survival (PFS) was defined as the time from treatment to the first documentation of disease progression or to death from any cause, whichever occurred first. Disease progression was defined as ≥20% increase of sum of the longest diameter (SLD) as for RECIST 1.0. OS was defined as the time from treatment to death or last follow-up. The PFS and OS were estimated using the Kaplan–Meier method with Rothman's 95% CIs and compared across the groups using the log-rank test.
The association of TB as a continuous variable with progression and death was evaluated using the Cox proportional hazards model. The association between the number of metastatic sites and survival was also evaluated. Multivariable analyses, adjusted for age, sex, MSKCC risk groups (low- or intermediate- vs high-risk) and active therapy (yes if sorafenib or sunitinib, no if placebo) and type of study, were performed using a stepwise selection approach with a type I error of 0.05 for model entry and 0.10 for elimination. Additional elimination was applied to identify significant variables at the level of P < 0.05. The correlations between TB and MSKCC risk groups or ECOG performance status were evaluated by the non-parametric Spearman-rank test.
Two exploratory analyses were performed: in the first, patients were divided into two groups based on value of TB above or under the median value; in the second one, patients were divided into three groups based on tertiles of TB values at baseline; for each group, OS and PFS were evaluated and compared as described above. As a sensitivity analysis, the PFS rates in the groups were evaluated and compared in patients receiving active therapy. A chi-squared test was used to assess if the distribution of observed frequencies for baseline characteristics differed between populations enrolled in the two trials and included in the final analysis. We used Predictive Analytics Software (PASW v 18; IBM, SPSS).
Overall, 141 patients were screened in the two trials; 16 patients did not match the inclusion criteria and one patient had insufficient information regarding survival (Fig. 1). A total of 124 patients were included in the final analysis: 85 patients were included in the phase III TARGET trial and 39 patients were included in the phase II sunitinib continuous dose trial (SCDT); no differences were found between patients' baseline characteristics (Table 1[11,12]).
Table 1. Differences between the populations enrolled in two trials
No significant differences were present between patients enrolled in phase III sorafenib TARGET trial  and in phase II SCDT . The difference in treatment was not applicable because of the different study design.
Median (IQR) age, years
56.1 (51.0– 64.0)
Median (IQR) TB, cm
ECOG performance status, %
Motzer risk class, %
The characteristics of patients included in the final analysis are shown in Table 2: 74.2% of the patients were males with a median (IQR) age of 59 (52–65) years. All patients had clear-cell histology, and all had previously undergone nephrectomy and received cytokines as a previous first-line treatment, the median (IQR) number of metastatic sites was 3 (2–4). ECOG performance status was 0 in 64% and 1 in 36% of patients. A total of 66% of patients received an anti-angiogenic treatment with sorafenib or sunitinib and 34% received placebo; the median (IQR) follow-up was 80.1 (68.8–88.0) months.
Table 2. Baseline characteristics of 124 patients included in the final analysis
Overall (n= 124)
Male, n (%)
Median (IQR) age, years
Previous nephrectomy, n (%)
Median (IQR) TB, cm
MSKCC risk class, n (%)
ECOG performance status, n (%)
Treatment, n (%)
Modified MSKCC risk score classified the patients as low risk in 68%, intermediate risk in 23% and high risk in 9% of cases (Table 1). The median (IQR) baseline TB of 124 patients was 12.8 (8.2–22.0) cm and was higher in the MSKCC risk groups with poorer prognosis (Table 3). The TB was positively related to ECOG performance status (Spearman's correlation: rs= 0.357; P < 0.001) and to MSKCC risk score (rs= 0.478; P < 0.001 [Table 3]).
Table 3. Baseline TB by MSKCC risk class and ECOG performance status
Median (IQR) TB, cm
Spearman correlation (two-tailed)
Baseline TB was divided into tertiles: the first tertile ranged from 1.3 to 9.4 cm; the second from 9.5 to 19.0 cm and the third from 19.1 to 47.3 cm.
The median PFS for the 124 patients was 5.6 months (95% CI: 4.3–6.9). The PFS was longer in patients receiving sunitinib or sorafenib, than in those receiving placebo (8.4 vs 5.6 vs 2.8 months; P < 0.001). The MSKCC risk group was strongly associated with PFS: 5.6 months (95% CI: 4.2–7.0) for the low-, 6.5 months (95% CI: 4.0–9.0) for the intermediate- and 1.6 months (95% CI: 1.4–1.8) for the high-risk group (P < 0.001), but no difference was found between low- and intermediate-risk groups (P= 0.2). PFS differences remained when adjusting for active therapy (P < 0.001). No significant relationship was found between the number of metastatic sites and PFS (HR = 1.16; 95% CI: 0.99–1.34; P= 0.065).
Tumour burden was a prognosis factor for PFS: each 1-cm increase in TB increased the risk of progression by 5% (HR: 1.05; 95% CI: 1.02–1.07; P < 0.001).
When adjusting for age, gender, MSKCC risk group (high- vs low-/intermediate-risk group), active therapy (yes vs no) and type of study, the increase in TB remained associated with a higher risk of progression (HR: 1.024; 95% CI: 0.001–1.048; P= 0.011 [Table 4]). The significant value for the type of study, probably reflects the increase in PFS reached with sunitinib compared with sorafenib or placebo.
Table 4. Crude estimates and adjusted estimated ratio for PFS and OS
Significant values on multivariate analysis are underlined.
Gender (female vs male)
MSKCC risk group (high vs low/intermediate)
Therapy (yes vs placebo)
TB (1 cm increase)
Type of study (SCDT vs TARGET)
HR (95% CI)
Gender (female vs male)
MSKCC risk group
Therapy (yes vs placebo)
TB (1 cm increase)
Type of study (SCDT vs TARGET)
The exploratory analyses confirms that the increase in TB was associated with a decrease in PFS: the median PFS value for TB above the median was 4.2 months (95% CI: 3.2–5.1), compared with 5.6 months (95% CI: 3.3–7.9), for patients with TB below the median (P= 0.012), but differences were not significant when adjusted for treatment. The median values for the first, second and third tertile were: 5.6 months (95% CI: 2.2–9.0), 6.5 months (95% CI: 4.1–9.0), and 4.2 months (95% CI: 3.6–4.8), respectively (P= 0.012, Fig. 2), whereas the pairwise analysis showed no differences between the first and second tertiles. In the sensitivity analysis restricted to patients exposed to active therapy, the group with higher TB remained associated with poorer survival (P= 0.04).
The median OS was 20.4 months (95% CI: 16.6–24.3) and the higher the MSKCC risk score, the shorter the OS: 26.3 months (95% CI: 19.3–33.2) for low-, 10.7 months (95% CI: 5.0–16.3) for intermediate- and 5.6 months (95% CI: 4.1–7.3) for high-risk group (P < 0.001). No difference in survival was found between patients who received sorafenib, sunitinib or placebo (21.4 vs 21.3 vs 16.4 months, respectively; P= 0.58). A significant relationship was also found between the absolute number of metastatic sites and OS (HR: 1.26; 95% CI: 1.08–1.48; P= 0.005).
Tumour burden was a predictive factor for survival – each 1 cm increase in TB increased the death risk of 5% (HR: 1.05; 95% CI: 1.03–1.08; P < 0.001) – as was the MSKCC risk group. In multivariable analysis adjusted for this factor, the TB remained an independent prognostic factor for death (Table 4).
Exploratory analyses confirmed that the increase in TB was associated with the decrease in OS: the median OS value for TB above the median was 16.4 months (95% CI: 9.4–23.5) compared with 27.4 months (95% CI: 16.6–38.2) for patients with TB below the median (P= 0.001); the differences were maintained when adjusted for treatment. The analysis of tertiles confirms that OS decreased with the increase of TB: the first, second and third tertile median OS values were 42.1 month (95% CI: 24.0–60.2), 20.1 month (95% CI: 14.5–25.7), and 11.2 months (95% CI: 4.8–17.6), respectively (P < 0.001, Fig. 2).
Tumour burden has been shown to be a prognostic factor in several tumour types; in diffuse large cell lymphoma, the independent predictive role for survival and for achievement of remission as well as relapse-free survival was reported over 25 years ago . In Hodgkin lymphomas, the initial TB was predictive for response to two different chemotherapy regimens  and increases the half life of rituximab in treated patients . The negative impact of high tumour volume at baseline on response to treatment was also reported in neoadjuvant treatment for breast cancer . Recently, the metabolic volume assessed by postitron emission tomography has also been reported to have a prognostic role in lymphoma , such as in advanced solid tumours .
With regard to mRCC, few studies have evaluated the role of TB, even though it has been considered to be an important factor in localized disease. For example, the risk of malignancy and the tumour grade increase with tumour's size in patients with suspected renal lesions [20,21]. It was calculated that each 1-cm increase in tumour size, leads to an increase of 16% in the odds of malignancy .
In localized RCC, the UCLA Integrated Staging System divides patients into three prognostic groups based on Fuhrman grade, ECOG performance status and tumour size. Among patients with high-risk non-metastatic disease, the study found a shorter survival in those with the greatest primary lesions (T4N0M0) . The prognostic relevance of primary tumour extension was confirmed by the Mayo Clinic SSIGN algorithm that reported, in univariate analysis, a direct relationship between the tumour size and the risk of death. These data were confirmed in multivariate analysis showing a doubled risk of death when tumour was >5 cm (risk ratio: 1.99; 95% CI: 1.47–2.70; P < 0.001) .
In metastatic patients, three studies have retrospectively analysed the impact on survival of baseline TB. Stein et al.  described the role of TB in 89 patients with mRCC, treated with ixabepilone in two prospective trials [25,26]. As with the present study, TB was defined as the sum of the longest diameter of all measurable lesions (>1.0 cm), although limited to five lesions. The study reported a median TB value of 21.7 cm and showed a negative correlation between TB and survival (Pearson r=−0.49; P < 0.001) as well as with the tumour growth rate .
In patients with mRCC treated with targeted agents, two retrospective studies analysed the role of TB as a prognostic factor. In the first study, Basappa et al.  reviewed 69 patients treated with sunitinib after failure of different therapies with cytokine or other antiangiogenic agents. In their analysis, the median value of baseline TB (14.0 cm; range 3.0–42.2 cm) was used as the threshold for assessing the impact on PFS and OS; other variables were also analysed, such as soft tissue TB present above and below the diaphragm and their ratio, the number of metastatic sites involved and the total number of metastases . Among variables related to tumour extension, a TB < 13 cm and a TB above the diaphragm ≤6.5 cm were prognostic factors for both PFS and OS in univariate analysis. In multivariate analysis, a TB above the diaphragm ≤6.5 cm, in addition to <10 metastases and the Cleveland Clinic Foundation risk classification were found to be independent prognostic factors .
The second retrospective study investigated the prognostic role of TB in 42 patients enrolled in the RECORD-1 and REACT expanded access programmes, testing everolimus after sorafenib, sunitinib or both. The results of the univariate analysis confirmed that patients with a TB below the median value (15.5 cm) have a longer OS (25.2 vs 11.8 months; P= 0.01) but not better PFS compared with patients with greater TB; whether this is an independent prognostic factor in multivariate analysis is not reported .
In the present series, we found a direct relationship between the absolute baseline value of TB and OS, and we report for the first time, its prognostic role in multivariate analysis independently of the site of metastases (above or below the diaphragm) and the MSKCC risk score. Notably, we have used the modified MSKCC score, which already accounts for the number of metastatic sites, for the present analysis. We believe that this information could be translated in clinical practice, by concluding that patients with the same MSKCC risk group may have a different survival based on initial tumour extension. In clinical practice, a greater TB may be associated with a symptomatic disease, a worse ECOG performance status and a worse prognostic score; however, it remains controversial whether active treatment can be delayed in asymptomatic patients.
The present data might be useful in patient's treatment decision-making to schedule the beginning of an active treatment with anti-angiogenic drugs. The evidence from randomized phase III trials comparing sorafenib or everolimus with placebo suggest that patients who received active drugs after placebo have no difference in PFS, explaining the lack of difference in OS in both studies [29,30]; however, the present data suggest that there may be a threshold when active treatment should be started. This hypothesis needs to be confirmed in future prospective trials.
Limited results are available in the literature regarding the role of TB in predicting the efficacy of anti-angiogenic drugs. In patients with mRCC treated with sunitinib and sorafenib, Yuasa et al.  reported greater tumour shrinkage in smaller tumours (<2.4 cm), but this tumour shrinkage did not correlate with survival. Furthermore, in patients with mRCC who receive sunitinib before undergoing nephrectomy, a tumour shrinkage >10% has been found able to determine an improvement of 2 years in OS (78.8 vs 27.3%; P= 0.01) . The same threshold of tumour shrinkage has been reported in metastatic disease as sufficient to determine a statistically significant improvement from 5.6 to 11.1 month in terms of PFS , and from 15.8 to 32.5 months in terms of OS . Although these studies report a relationship between the TB and response to therapy and between response and prognosis, they fail to report a direct relationship between baseline TB and OS.
The present study better analyses this aspect, relating the tumour extension to PFS and OS and showing that a smaller TB leads to a better OS, and adding predictive information to traditional risk classification to increase the possibility of identifying patients who may have a greater benefit from a treatment with targeted agents. The major limits of the present study are its retrospective nature and the timing of treatment, with targeted agents used after previous cytokine treatment. In fact, even if cytokines are still recommended for selected cases by current guidelines , they do not represent current clinical practice. Despite this, in the present series of patients, the relevance of the prognostic role of baseline TB may be only marginally influenced by previous treatment, considering the low response rate to interferon and interleukin reported in the medical literature . Furthermore, this analysis includes all patients treated in the two trials with targeted agents at our institution without selection for previous tumour response to cytokines.
To our knowledge, the present study includes the largest and most homogeneous population in which the prognostic role of baseline TB has been reported. It may be considered as a solid background to explore these data prospectively in patients with untreated mRCC.
In conclusion, TB is easy to calculate from standard CT and significantly relates to PFS and strongly relates to OS in patients with clear-cell mRCC treated with anti-angiogenic agents after cytokines. Prospective studies are needed to determine whether there is a well-defined threshold for TB, which could help to determine when to start treatment in asymptomatic patients.