Survival of patients receiving systematic therapy for metachronous or synchronous metastatic renal cell carcinoma: a retrospective analysis

Background The differences in progression-free survival (PFS) and cancer-specific survival (CSS) of metastatic renal cell carcinoma (mRCC) patients according to treatment, type of metastasis, and Heng criteria risk are unclear. In this study, we compared survival according to various such parameters. Methods Between 2000 to 2014, 214 mRCC patients, of whom 171 (79.9%) were intermediate-risk and 43 (20.1%) were poor-risk, were retrospectively selected; 126 (58.9%) patients were treated with immunotherapy (IT) and 88 (41.1%) with targeted therapy (TT). Moreover, 144 patients had synchronous mRCCs (67.3%, SM) and 70 had metachronous mRCCs (32.7%, MM). The Kaplan-Meier method and log-rank test were used to compare progression-free survival (PFS) and CSS. Results During a median 4.2 (1.0–70.4) months of systemic treatment and 98.3 (4.8–147.6) months of follow-up, the median PFS and CSS were 4.7 (95% confidence interval [CI]: 3.8–5.5) and 13.8 (95% CI, 9.8–18.3) months, respectively. The PFS and CSS were significantly better in the MM (5.9 and 21.3 months) and intermediate-risk groups (5.2 and 18.3 months) than those in the SM (4.4 and 9.6 months) and poor-risk groups (2.7 and 5.8 months), respectively (p < 0.05). Further stratification showed that TT produced significantly better PFS than IT in intermediate-risk patients with SM and a treatment-free interval (TFI) < 1 year, and in those with MM with a TFI ≥1 year (p < 0.05). There were no differences in survival outcomes according to various other subgroup stratifications (p > 0.05). Conclusion Dividing patients into specific subcategories helps to better predict therapeutic outcomes. Electronic supplementary material The online version of this article (10.1186/s12885-019-5900-1) contains supplementary material, which is available to authorized users.


Background
The standard systemic therapy for metastatic renal cell carcinoma (mRCC) has recently been changed to targeted therapy (TT). TT produces improved prognosis, with an observed median cancer-specific survival (CSS) of 29.5 months, and has markedly extended progressionfree survival (PFS) intervals [1]. Furthermore, TT is relatively well tolerated compared to immunotherapy (IT) with cytokines [2], which was the mainstay systemic treatment in the past decades [3]. Despite poor 5-year survival rates of approximately 10%, IT had an advantage of significantly improving overall survival (OS), with some patients achieving 10-year persistent complete remission rates of 7-10% [4,5]. Conversely, whether TT prolongs OS has not yet been determined [4]. The heterogeneous characteristics of mRCC patients, as well as the heterotrophic characteristics of their tumors, makes the prediction of therapeutic outcomes challenging. Therefore, dividing patients into subcategories ought to be considered more thoroughly to derive more reliable predictive prognostic factors. mRCCs are categorized into synchronous mRCC (SM), which involves approximately one-third of newly diagnosed patients; and metachronous mRCC, (MM), that has a prevalence rate of 30-40% among localized RCC patients and involves disease progression to metastasis after a certain interval of time has elapsed since curative surgery. MM and SM have different pathophysiologies and metabolisms; furthermore, the heterotrophic and pleomorphic characteristics of RCC result in unpredictable and diverse responses to systemic therapy. Moreover, the microenvironments and tumor activities of primary vs. metastatic tumor lesions, as well as their underlying tumor burdens, are different [3,6]. However, the TT treatment guidelines for MM and SM are similar, and not many studies have addressed the differential prognoses and comparative responses between TT and IT according to the tumors' metastatic types and prognostic risk groups. Such clinical data regarding survival outcomes are required for clinicians to understand the patients' prognoses and to devise effective treatment strategies. Therefore, this retrospective study analyzed the PFS and CSS according to first-line systemic therapy, as well as the survival rates of patients with MM or SM mRCCs treated with either TT or IT at a single institution. Furthermore, prognostic outcomes according to the types of metastases were compared with respect to the Heng risk model and treatment-free intervals (TFIs).

Ethics statement
Following approval of this retrospective study by the Institutional Review Board of the National Cancer Center (IRB No. NCC2016-0263), the IRB waived the written informed consent requirement. All patient data were anonymized and de-identified prior to our analysis. All study protocols were performed in accordance with the ethical tenets of the Declaration of Helsinki.

Patients' criteria and evaluating tools
The medical records of 214 mRCC patients treated between 2000 and 2014, including 171 (79.9%) intermediate-risk and 43 (20.1%) poor-risk patients according to the Heng criteria, were retrospectively reviewed. Patients with the following characteristics were excluded from the analysis: those with incomplete follow-up medical records (records beyond the prospectively recorded National Center Center RCC registry), patients who refused systemic therapies after receiving an explanation of the possible adverse events, patients who stopped receiving medications because of the associated financial expenses, patients for whom no therapeutic effects were expected because of poor clinical statuses involving multiple underlying diseases and severe tumor burdens, patients with missing data for all of the risk factors included in the analysis, patients with missing treatment records, patients under 20 years of age, patients belonging to the favorable risk group according to the Heng risk criteria, and patients who developed progression within 1 month of treatment (Additional file 1: Figure S1). There were 144 SM patients (67.3%) and 70 MM patients (32.7%); 126 patients were treated with IT and 88 were treated with TT. All mRCC patients underwent a complete evaluation after every 1-4 cycles (6-12 weeks) of IT and every 2 cycles of TT (12 weeks). The follow-up protocol that included laboratory and imaging evaluations was previously described in detail [7]. Treatment continued until disease progression was detected. Patients were further stratified into TFI < 1 year vs. TFI ≥1 year groups. The TFI was defined as the time from the diagnosis of disease to the start of systemic treatment [8][9][10] The International Metastatic Renal Cell Carcinoma Database Consortium risk criteria (also known as the Heng criteria) [8] for prognostic risk stratification, the Response Evaluation Criteria in Solid Tumors v1.1 for therapeutic responsive evaluation to systemic therapy [11], and the Fuhrman nuclear grade [12] and TNM stages for pathological RCC evaluation [13] were used.

Statistical analysis
The baseline characteristics are summarized as frequency with percentage for categorical variables and median with range (min-max) for continuous variables. The differences between SM and MM were assessed using the Wilcoxon rank-sum test and chi-square test (or Fisher's exact test). The date of the most recent hospital visit was used as the date of last follow-up. The PFS duration was defined as the time between first-line systemic therapy and disease progression. The CSS duration was defined as the time between first-line systemic therapy and death from cancer or alive. Follow-up durations were estimated using the reverse Kaplan-Meier method, in which being alive is treated as the event of interest and deaths are censored. The estimated survival curves were computed using the Kaplan-Meier method. Survival curves according to systemic treatment and Heng risk groups were compared using the log-rank test. Cox proportional hazards models were used to analyze subgroups of the patients with mRCC. To identify factors associated with prognosis, univariable Cox models of PFS and CSS were applied to the variables included in the Heng risk model. Variables that had significant (p < 0.05) associations with PFS or CSS in the univariable analyses were included in multivariable Cox models. A backward variable selection method was then applied to the included variables with an elimination p-value criterion of 0.05. P-values less than 0.05 were considered statistically significant. All statistical analyses were performed using SAS (version 9.3; SAS Institute Inc., Cary, NC, USA) and R (version 3.3.2) software.

Results
Following a median 4.2 (range: 1.0-70.4) months of systemic treatment and 98.3 (range: 4.8-147.6) months of follow-up, only 30 patients (14.0%) were alive at the end of the study. The median PFS and CSS were 4.7 (95% confidence interval: 3.8-5.5) months and 13.8 (95% confidence interval: 9.8-18.3) months, respectively. The patients' characteristics are summarized in Table 1. The gender ratio, nephrectomy rates, TFIs, Heng risk groups, clinical T and N stages, clear cell and nonclear cell histologies, follow-up durations, PFS, and CSS were significantly different between the SM and MM groups (p < 0.05, Table 2).
Significantly better PFS and CSS rates in the MM group (PFS: 5.9 months; CSS: 21.3 months) and in the intermediate-risk group (PFS: 5.2 months; CSS: 18.3 months) were observed than in the SM (PFS: 4.4 months; CSS: 9.6 months) and poor-risk groups (PFS: 2.7 month; CSS: 5.8 months) (p < 0.05; Additional file 1: Figure S2). In the intermediate-risk groups, patients with MM showed longer PFS and CSS rates than those with SM (PFS: 6.2 vs. 5.1 months; CSS: 25.2 vs. 13.9 months, respectively). Furthermore, the MM group showed longer PFS and CSS than the SM group (3.7 vs. 2.7 months and 10.2 vs. 5.6 months, respectively) in poor-risk patients (Additional file 1: Figure S2); statistical significance was only achieved for CSS in the intermediate-risk group (p = 0.01) (Additional file 1: Figure S3).
As for the stratified Heng risk groups, only the intermediate-risk group treated with TT had a significantly longer PFS (6.1 months) than that of the IT group (2.6 months) among SM patients (p < 0.001, Fig. 2a When stratifying patients according to TFI of < 1 year vs. ≥1 year, intermediate-risk patients treated with TT had significantly better PFS than those treated with IT (6.1 months vs. 2.6 months, respectively) among SM patients with TFI < 1 year (p < 0.001, Fig. 3a). However, no other significant differences in PFS or CSS were observed between IT-and TT-treated patients with SM or MM and TFI < 1 year (p > 0.05, Fig. 3). Among patients with mRCC and TFIs ≥1 year, only intermediate-risk MM patients treated with TT had significantly better PFS (10.5 months) than those treated with IT (4.0 months) (p = 0.008; Fig. 4c). The poor-risk patients with TFIs ≥1 year could not be assessed because of their small sample size.
The results of univariable and multivariable Cox proportional hazards analyses of PFS and CSS are summarized in Additional file 1:

Discussion
TT has recently replaced IT as the mainstay therapy for mRCC, resulting in modest survival benefits with improved PFS despite failing to provide measurable benefits to CSS except in poor-risk patients administered temsirolimus [4], as in the present study, where the CSS were the same between poor-and intermediate-risk mRCC. Our previously published study also showed that TT improved prognoses of mRCC patients compared to  The results of our study showed that the intermediaterisk group had significantly better PFS and CSS than the poor-risk group. We also observed that, as compared with IT, TT significantly lengthened PFS, but not CSS (Fig. 1). In particular, a significant difference in CSS was observed in neither the SM subgroup (TT vs. IT: 9.6 vs.  Fig. 1b and d). Nonetheless, our results provide some statistically nonsignificant indications that IT may be associated with better CSS than TT: in both the SM and MM subgroups ( Fig. 1b and d), comparison of the tails of the IT and TT curves for CSS show that the former had longer durable responses without cancer-specific deaths. Overall, our results suggest that TT was significantly better for PFS, while IT might be more suitable for extending CSS. However, the survival benefit from IT should be considered carefully, because the IT group included patients who benefited from TT (9.5%, N = 12), and their follow-up periods were In previous large-scale IT and TT studies, Motzer et al. [16] and Naito et al. [17] showed median CSS of 10    respectively [18]. Our study had similar median CSS results, with respectively 13.9 and 25.2 months for the intermediate-risk and 5.6 and 10.2 months for the poorrisk SM and MM groups. Some differences were expected owing to the enrolled patients' ethnic group, rate of nephrectomy (Motzer: 80.5%, Naito: 55%, and our study: 56.5%), and the different cytokine-treatment regimens and primary and metastatic tumor burdens.  The outcomes of TT in our study were similar to, or slightly poorer than, previous phase 3 trials [7,14]. The different outcomes between this study and previous studies were attributed to the different characteristics of the enrolled patients, including the absence of favorablerisk groups in our study. To our knowledge, no studies have directly compared prognostic outcomes according to metastasis type outside of metastatic site-specific comparison studies and case reports [19][20][21][22][23][24]. Although tumor cells' pathophysiologies, activities, and burdens differ between SM and MM, the systemic treatment protocol for these cancers remain the same. Previous studies and ours showed different pathophysiologic activities of tumor cells as identified by immunohistochemical staining of multiple tissue markers that are closely related to the TT and immune responses [15,19]. The absence of a primary renal tumor burden post-nephrectomy influences prognoses following both IT and TT [25,26].
Our study classified mRCC patients according to the Heng risk criteria, TFI, metastatic types, and systemic therapies to determine the influence of each on prognoses. Prognostic comparisons according to the Heng risk groups showed a significant difference in CSS between SM (13.9 months) and MM (25.2 months) in the intermediate-risk group (p = 0.010), but not in PFS. The poor-risk group also showed no prognostic differences according to metastatic types. CSS rates were better in MM than in SM patients because the metastatic tumor burden was less in the former, and the tumor metabolic activity was different from those with SM. When systemic therapies were compared, only the PFS in the intermediate-risk group showed a significant improvement with TT over IT in both the SM (6.1 vs. 2.6 months) and MM (10.2 vs. 4.1 months) groups, respectively. One interesting finding was that receipt of IT was associated with nonsignificantly better PFS and CSS rates for poor-risk patients in both the SM and MM subgroups (p > 0.05, Fig. 1c and d). When considering TFI, TT significantly improved PFS in SM group patients with TFIs < 1 year and in MM group patients with TFIs ≥1 year; these findings were true regardless of metastatic types. In addition, because of multicollinearity between TFI and metastatic type (SM vs. MM), metastatic type was not a significant prognostic factor in the multivariable analysis, even though it had been significantly prognostic in the univariable analysis.
This study had some inherent limitations owing to its retrospective design, the small number of poor-risk patients, the exclusion of patients with medical records that were incomplete because of loss to follow-up, and the different follow-up periods in the compared subgroups. Most metastatic lesions that were diagnosed pathologically might not accurately represent the whole disease, especially when comparing the primary and various metastatic lesions. Further studies that include tissue analysis, gene sequencing, and clinicopathological data are warranted to identify the therapeutic benefit of systemic therapies on MM and SM patients. Prospective data collection and different statistical analyses may help to address the possibility that selection bias and other biases could have affected the comparisons that were made in the present study. For example, the use of propensity score matching might allow more rigorous comparisons of outcomes after IT and TT, as well as of the prognostic implications of SM and MM mRCC. There was a difficulty in applying propensity score matching in this study. Since some of the variables were unknown, the size of the after matching set will be smaller. It is difficult to analyze each sub-groups (SM/MM, risk group, Treatment interval) after matching because IT and TT survival is identified in all subgroups of patients. Our study is only the second to compare prognoses of SM and MM patients treated with systemic TTs.

Conclusions
Our study showed that dividing patients into additional subcategories improved the prediction of therapeutic outcomes. We found that patients with intermediate-risk mRCC treated with TT had significantly better PFS rates than those treated with IT among patients with SM and TFIs < 1 year and patients with MM with TFIs ≥1 year.

Additional files
Additional file 1: Table S1. The Cox proportional hazards model of progression-free survival and cancer-specific survival of prognostic risk factors. Figure S1. Flow-chart of patients included in analysis. Figure S2. Comparison of Kaplan-Meier curves of progression-free survival and cancer-specific survival of metastatic renal cell carcinoma patients according to the metastatic types and risks according to the Heng criteria.