Phase II Results of Dovitinib (TKI258) in Patients with Metastatic Renal Cell Cancer Free

Clear-cell renal cell carcinoma (RCC) is commonly associated with mutations, deletions, and modifications of the von Hippel–Lindau (VHL) gene that result in nonfunctional or reduced levels of VHL (1, 2). VHL is involved in degrading hypoxia-inducible factor-α (HIF-α), a transcription factor that mediates expression of VEGF and platelet-derived growth factor (PDGF), which control cell proliferation and angiogenesis (3). Therefore, loss of VHL leads to increased levels of HIF-α, and a subsequent increase in proangiogenic factor levels. This knowledge has led to development of tyrosine kinase inhibitors (TKI) that target the VEGF pathway and mTOR inhibitors, which also target pathways involved in HIF-α expression. Despite the multiple VEGF pathway [VEGF receptor (VEGFR) TKIs and anti-VEGF antibodies] and mTOR inhibitors already approved, these treatments rarely induce complete responses in advanced RCC, and nearly all patients will eventually progress (1).

Antiangiogenic escape, in which tumors develop resistance to VEGF pathway–based antiangiogenic therapies, limits the effectiveness of VEGFR inhibitors in patients with advanced RCC (4). Preclinical studies indicate that signaling through the fibroblast growth factor (FGF) pathway and interleukin-8 (IL-8; ref. 5) may provide possible escape mechanisms after initiation of anti-VEGF therapy (6). Indeed, FGF is another pathway shown to play a role in tumor angiogenesis (7, 8). Highly vascularized tumors often contain high levels of FGFs and FGF receptors (FGFR) after treatment with VEGF pathway inhibitors (9, 10). In addition, dysregulated expression of FGFs or FGFRs has been described in several cancers, including RCC (11).

Targeting the FGF pathway while maintaining inhibition of angiogenesis and tumor growth through inhibition of VEGF signaling may offer additional benefit compared with inhibition of VEGF signaling alone (12). Dovitinib is a TKI that inhibits FGFR, as well as VEGFR and PDGF receptor (PDGFR; refs. 13 and 14). The safety and preliminary activity of dovitinib at 500 mg/day on a 5-days-on/2-days-off schedule were tested in the phase I dose-escalation portion of a phase I/II study (NCT00715182) in patients with advanced or metastatic RCC (15). A 10% response rate and long-lasting disease stabilization were observed in heavily pretreated patients. Here we report on the dose-expansion phase (phase II), which explored dovitinib (500 mg/day on a 5-days-on/2-days-off schedule) in patients with advanced or metastatic RCC enriched for patients previously treated with a VEGFR TKI and an mTOR inhibitor.

Data from the phase II dose-expansion portion of a multicenter, open-label phase I/II study (NCT00715182) are reported here. The phase I dose-escalation portion of the study has been previously reported (15). The primary objective of the dose-expansion phase was to assess the preliminary antitumor activity of dovitinib in patients who had been previously treated with ≥1 VEGFR TKI and ≥1 mTOR inhibitor and received ≥1 dose of dovitinib [per-protocol efficacy set (PPES)]. A multinomial 2-stage design was used to evaluate the primary endpoint of overall response rate (defined as complete response + partial response) and no clinical benefit (NCB) at 8 weeks (16). NCB was defined as no response (partial response or complete response) at 8 weeks posttreatment and no stable disease for ≥8 weeks after start of treatment. Response by 8 weeks was chosen because tumor shrinkage by VEGF-pathway targeting agents is often observed at the first posttreatment scan. Thirty patients were to be enrolled in stage 1 and the trial could stop after the first stage rejecting the inactivity hypothesis if there were: (i) 1 to 2 responses (complete response or partial response) and ≤14 NCB, (ii) 3 responses and ≤15 NCB, (iii) 4 responses and ≤17 NCB, (iv) 5 to 11 responses and ≤18 NCB, or (v) ≥ 12 responses. Based on the analysis performed at the end of stage 1, the per-protocol multinomial 2-stage design stopping criteria for efficacy described above were met. However, the protocol was amended to complete enrollment in stage 2, even if efficacy had been already established at the end of stage 1, to obtain more precise efficacy parameter estimates. Secondary endpoints included determination of progression-free survival (PFS), overall survival (OS), dovitinib pharmacokinetics, and safety.

Dovitinib was administered orally at 500 mg/day on a 5-days-on/2-days-off schedule in 28-day cycles. Patients continued treatment until disease progression, unacceptable toxicity, withdrawal of consent, or death, or at the discretion of the investigator. The protocol and the subsequent amendments were approved by each site's institutional review board/independent ethics committee/research ethics board. All patients provided written informed consent before enrollment. The trial was conducted in accordance with the ethical principles of the Declaration of Helsinki, the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use Harmonised Tripartite Guidelines for Good Clinical Practice, and applicable local regulations.

Adult patients (ages ≥ 18 years) with measurable, histologically or cytologically confirmed progressive metastatic RCC with predominant clear-cell histology (>50%) and an Eastern Cooperative Oncology Group (ECOG) performance status of ≤1 were eligible. The population was enriched for patients previously treated with a VEGFR TKI and an mTOR inhibitor. Originally, the eligibility criteria included enrollment of up to 20 patients who failed standard treatments (e.g., IL-2, IFN-α) to determine the effect of dovitinib in patients who had not received VEGFR TKIs and mTOR inhibitors. However, during the course of the trial, everolimus was approved as a second-line therapy after failure of treatment with the VEGFR TKIs sunitinib and sorafenib. Following this approval, the eligibility criteria were amended to enroll patients who had progressed after treatment with ≥1 VEGFR TKI and ≥1 mTOR inhibitor in addition to patients who previously failed treatment with other therapies; the PPES for efficacy assessment was redefined consistently. To obtain a full pharmacokinetic profile in patients of Asian ethnicity, the enrollment of approximately 30 patients in the Taiwanese sites who were refractory to standard treatments or for whom no standard treatment existed was also planned. The results of the Asian patient pharmacokinetic analysis will be reported in a separate publication. Other key inclusion and exclusion criteria were described in the phase I report (15).

Tumor status was assessed at baseline and every 8 weeks by Response Evaluation Criteria In Solid Tumors v1.0 (17). Both local and central radiologist reviews were used to evaluate tumor response. The study was powered at 90% for the alternative hypothesis (15% response rate and 30% NCB rate) against the null hypothesis (≤5% response rate and ≥65% NCB rate) with a 1-sided type I error rate of 10%. The primary efficacy analysis was performed in the PPES. Supportive analyses were performed on the full analysis set that comprised all patients who received at least one dose of study medication. The Kaplan—Meier method was used to analyze the time-to-event endpoints (PFS and OS). PFS was defined as the time from the start date of study treatment to the date of first documented tumor progression [progressive disease PD)] or death because of any cause, whichever came first; for patients who did not experience PD or death by the cutoff date, the PFS was right-censored on the date of the last adequate tumor assessment. OS was defined as the time from treatment start date to date of death because of any cause; for patients lost to follow-up or who were still alive at the analysis cutoff date, OS was right-censored at the last contact date.

Collection of plasma samples for pharmacodynamic analyses was described previously (15, 18). Briefly, plasma samples were assessed by ELISA for FGF23 (Kainos Laboratories, Inc.) and a panel of markers associated with angiogenesis, including basic FGF (bFGF), placental growth factor (PIGF), soluble VEGFR1/2 (sVEGFR1/2), and VEGF (Meso Scale Discovery). These biomarkers were descriptively analyzed by time point. Data were log2 transformed. Sufficient data from cycle 1 were available to create a 1-way linear mixed effects model with time and/or prior therapy and its interaction with time to determine the postbaseline fold changes on days 15 and 26 of cycle 1. A false discovery rate adjustment was applied to the P values. When available, archival tumors were analyzed for amplification of FGFR1, FGFR2, and FGF3 by quantitative PCR (using a threshold cutoff of ≥6 copies).

Common Terminology Criteria for Adverse Events v3.0 were used to assess adverse events throughout the study until 28 days following the last dose of dovitinib. Hematology (including clotting evaluation), biochemistry, and urine were monitored throughout the study. Other assessments included vital signs, performance status, cardiac assessments [electrocardiography (ECG), blood pressure, cardiac enzymes, and multiple-gated acquisition (MUGA)/echocardiogram (ECHO)], physical condition, and body weight. Collection and pharmacokinetic analysis of serial blood samples were as described previously (12).

A total of 67 patients with advanced RCC were enrolled in the phase II dose-expansion portion of this study from June 2009 to December 2011 and received ≥1 dose of dovitinib. Patient demographics, stage at the time of diagnosis, ECOG performance status, and previous antineoplastic therapies are listed in Table 1. All patients had metastatic disease, the majority of whom presented with metastasis to the lung, lymph nodes, bone, or liver. Patients were heavily pretreated with targeted agents (Table 1). Fifty-five patients (82.1%) were previously treated with VEGFR TKIs and mTOR inhibitors (PPES). Of these 55 patients, 20 were previously treated with 1 VEGFR TKI (sorafenib or sunitinib) and 1 mTOR inhibitor (everolimus or temsirolimus), and 35 were previously treated with ≥2 VEGF pathway inhibitors (axitinib, sunitinib, sorafenib, or bevacizumab) and ≥1 mTOR inhibitor (everolimus or temsirolimus).

As of the March 30, 2012, data cutoff, 3 patients (all in the PPES) were continuing treatment and 64 patients had discontinued, primarily because of disease progression [41 patients (61.2%)] and adverse events irrespective of causality [17 patients (25.4%)].

Of the first 30 evaluable patients that were enrolled in this phase 2 dose-expansion portion of the study, 1 achieved partial response and 14 had NCB at 8 weeks, per central review, thus meeting the prespecified protocol criteria for rejecting the inactivity hypothesis at the end of stage 1. As already specified, the protocol was subsequently amended to enroll additional patients to obtain more precise efficacy parameter estimates. For the final primary efficacy analysis based on central review, the overall response rate at 8 weeks for the 55 patients in the PPES was 1.8% (90% CI, 0.1%–8.3%), which consisted of 1 partial response (Table 2). The NCB rate at 8 weeks in the PPES was 47.3% (90% CI, 35.6%–59.1%). Stable disease at 8 weeks was observed in 28 of the 55 PPES patients (50.9%). Similar results were observed in the entire population (Table 2). Disease control rate at 8 weeks by central radiologist review was 52.7% (90% CI, 40.9%–64.4%) in the PPES and 52.2% (90% CI, 41.5%–62.8%) in all patients.

Throughout the whole study period, central radiologist review identified partial responses in 2 of the 67 patients (3.0%), both of whom were in the 55-patient PPES. One of these partial responses was achieved by 8 weeks—the patient had kidney and pancreas lesions and had PD as a best response following previous treatments with sunitinib and everolimus and unknown responses to prior chemotherapy and hormonal therapy. This patient had a tumor reduction of 33% on day 56 and maintained the tumor response until day 270. This patient discontinued from the study because of hypertriglyceridemia and hyperuricemia 10 months after beginning treatment and died from RCC 19 months later. The second patient to achieve a partial response, also with kidney and pancreas lesions, previously achieved a partial response, partial response, and stable disease following treatments with everolimus, sorafenib, and AS1411 (nucleolin-targeting DNA aptamer), respectively. This patient had a tumor reduction of 31% at day 110 and maintained the tumor response until day 226. This patient discontinued from the study because of hypertriglyceridemia 10 months after beginning treatment and was alive as of the data cutoff date 18 months later. Stable disease was achieved in 29 patients (52.7%) in the PPES and in 35 patients (52.2%) in the entire population (Table 2). Disease control (complete response, partial response, or stable disease) lasted for ≥4 months in 50.9% of patients in the PPES. Rates in all patients were similar, with 49.3% of patients achieving disease control lasting ≥4 months.

The best percentage change from baseline in sum of longest diameters of target lesions as determined by central radiological review in the PPES and full analysis set populations are shown in Fig. 1. The median PFS based on central radiologist review and OS for the PPES were 3.7 months (95% CI, 2.4–5.5) and 8.6 months (95% CI, 7.8–14.5), respectively (Fig. 2A and B). The median PFS (central radiologist review) and OS for the entire population were 3.7 months (95% CI, 3.0–5.6) and 11.8 months (95% CI, 7.9–17.4), respectively (Fig. 2A and B).

Changes in the levels of several plasma biomarkers collected at baseline and on days 15 and 26 of the first cycle of treatment suggest that pharmacologically active concentrations of dovitinib were achieved in patients (Fig. 3). A statistically significant increase from baseline in FGF23, a surrogate pharmacodynamic biomarker for FGFR1 inhibition (18), was observed in patient plasma samples following treatment with dovitinib at cycle 1, day 15 (124% increase; 95% CI, 74%–189%; P < 0.0001) and cycle 1, day 26 (47% increase; 95% CI, 13%–90%; P = 0.0063). Levels of the angiogenesis biomarkers were increased (PIGF, VEGF) or decreased (sVEGFR1, sVEGFR2) following treatment, consistent with the antiangiogenic effect associated with dovitinib-induced inhibition of VEGFR. PlGF increase from baseline was statistically significant at cycle 1, day 26 (90% increase; 95% CI, 64%–119%; P < 0.0001). sVEGFR1 decreases from baseline were statistically significant at cycle 1, day 15 (17% decrease; 95% CI, 27%–69%; P = 0.0057) and cycle 1, day 26 (26% decrease; 95% CI, 35–6%; P < 0.0001). Similarly, sVEGFR2 also exhibited significant decreases from baseline at cycle 1, day 15 (24% decrease; 95% CI, 29%–19%; P < 0.0001) and cycle 1, day 26 (17% decrease; 95% CI, 22%–12%; P < 0.0001). The trend in increased VEGF levels from baseline neared but did not reach statistical significance (20% increase at cycle 1, day 26; 95% CI, 1%–43%; P = 0.0512; Supplementary Fig. S1). No amplification of FGFR1, FGFR2, or FGF3 was identified in any of the archival tumor samples analyzed (n = 12; data not shown).

The role of the FGF signaling pathway in tumorigenesis and antiangiogenic escape in patients with advanced RCC previously treated with VEGF pathway–targeted therapy has been demonstrated by elevation of baseline bFGF levels (6). Data were stratified by prior therapy (VEGFR only, VEGFR and mTOR, or other therapy). Of the 60 patients with evaluable baseline bFGF levels, patients previously treated with VEGFR inhibitors (n = 10) and VEGFR and mTOR inhibitors (n = 49) had baseline bFGF levels of 140.7 pg/mL (95% CI, 63.2–218.2) and 150.4 pg/mL (95% CI, 116.8–184.0), respectively, in comparison with the baseline bFGF level (49.9 pg/mL) of the patient treated with a non–VEGF pathway-targeted therapy. These data were corroborated by the results of a pooled analysis of the 79 evaluable patients from the entire study [phase I dose escalation (previously reported; ref. 15) and phase II dose expansion]. In the pooled dataset, baseline bFGF levels were 177.2 pg/mL (95% CI, 99.2–255.2), 156.0 pg/mL (95% CI, 121.1–191.0), and 107.1 pg/mL (95% CI, 59.3–155.0) for patients previously treated with VEGFR inhibitors (n = 15), VEGFR and mTOR inhibitors (n = 44), or a non–VEGF pathway-targeted therapy (n = 20).

The median duration of treatment was 96 days (range, 4–765 days). The median actual dose intensity was 354.9 mg/day (range, 207.6–500.0). Forty patients (59.7%) had a relative dose intensity of >90%, and 24 patients (35.8%) had a relative dose intensity between 70% and 90%. Twenty-six patients (38.8%) required at least 1 dose change, and 38 patients (56.7%) required at least 1 dose delay or interruption. Adverse events led to dose changes in 23 of 67 patients (34.3%) and to dose delays or interruptions in 34 patients (50.7%).

All patients experienced ≥1 adverse event, and nearly all patients [n = 66 (98.5%)] had adverse events suspected to be related to dovitinib treatment. Treatment-related adverse events led to discontinuation in 14 patients (20.9%), the most common adverse event being metabolism/nutrition disorders and nervous system disorders (4.5% each).

The most common (≥30%) adverse events suspected to be related to dovitinib treatment of any grade were nausea (65.7%; grade 3, 7.5%), diarrhea (62.7%; grade 3, 9.0%), vomiting (61.2%; grade 3, 6.0%), decreased appetite (47.8%; grade 3, 7.5%), and fatigue (32.8%; grade 3, 10.4%; Table 3). Gastrointestinal toxicities were generally managed with antiemetics and antidiarrheal agents. Skin and subcutaneous tissue disorders were reported in 46.3% of patients and were generally grade 1/2, with only 1 grade 3 event (palmar-plantar erythrodysesthesia). Additional grade 3 events suspected to be related to dovitinib occurring in ≥1 patient were hypertension (10.4%), asthenia and γ-glutamyltransferase increase (9.0% each), hypertriglyceridemia (4.5%), and abdominal pain, stomatitis, noncardiac chest pain, pulmonary embolism, hemiparesis, neutropenia, and cerebrovascular accident (3.0% each). Grade 4 events suspected to be related to dovitinib were rare (6 patients [9.0%]) and included hypertriglyceridemia (4 patients [6.0%]), and pulmonary embolism and increased troponin T (1 patient each [1.5%]). Of the patients with grade 4 hypertriglyceridemia, 1 was permanently discontinued from the study as a result of the adverse event and 2 had study drug withheld and resumed treatment as scheduled. The fourth patient experienced grade 4 hypertriglyceridemia on 3 occasions. Study drug was withheld at the first occurrence, but dosing was continued as scheduled during the 2 subsequent occurrences. The patient with pulmonary embolism had doses withheld for 3 weeks and restarted treatment at 500 mg. The patient with increased troponin T level had doses withheld for 3 weeks and restarted at a reduced dose of 400 mg.

Nearly half of all patients [n = 30 (44.8%)] experienced serious adverse events (SAE) suspected to be related to dovitinib. A total of 21 (31.3%) and 4 (6.0%) patients experienced grade 3 and 4 SAEs, respectively. The most frequently reported grade 3 SAEs included nausea, fatigue, and vomiting, which occurred in 3 patients (4.5%) each. Grade 4 SAEs included hypertriglyceridemia in 2 patients (3.0%), pulmonary embolism in 1 patient (1.5%), and increased troponin T level in 1 patient (1.5%). Three patients (4.5%) died within 28 days of the last dose of dovitinib, 2 of whom (3.0%) died from disease progression and 1 (1.5%) who died of dyspnea related to pneumonia. No deaths were considered to be related to dovitinib.

Grade 3 hematologic abnormalities were infrequent and included decreases in absolute lymphocyte count [n = 10 (14.9%)], absolute neutrophil count [n = 2 (3.0%)], hemoglobin level [n = 3 (4.5%)], and partial thromboplastin time [n = 3 (4.5%)]. Only 1 grade 4 hematologic abnormality, a decrease in absolute lymphocyte count, was observed. The majority of biochemistry abnormalities were grade 1 or 2; grade 3/4 abnormalities occurring in ≥10% of patients included triglycerides increase [n = 15 (22.4%)], albumin decrease [n = 12 (17.9%)], and calcium decrease, cholesterol increase, and sodium decrease [n = 8 each (11.9%)]. Hepatic lab abnormalities were generally mild, with no grade 4 abnormalities and no alanine aminotransferase or bilirubin increases above grade 2. Grade 3 aspartate aminotransferase and alkaline phosphatase increases occurred in 2 patients (3.0%) and 7 patients (10.4%), respectively.

Newly occurring ECG abnormalities were reported in 42 of 66 patients (63.6%), the majority of which were related to rhythm [n = 16 (24.2%)] and conduction [n = 18 (27.3%); Supplementary Table S2). QT prolongation events involving QT interval corrected with Fridericia's formula (QTcF) increases >30 ms occurred in 12 of 66 patients (18.2%). Two patients (3.0%) with asymptomatic QTcF increases >60 ms were discontinued from treatment because of PD or pneumonia-related death. The latter case, which correlated with the sole report of QTcF > 500 ms (545 ms), was in an asymptomatic patient with grade 2 hypocalcemia who was hospitalized for ECG monitoring as per protocol and had a dose interruption from day 15 to day 19. After correction of the hypocalcemia, QTcF returned to normal, and this patient was released from the hospital and restarted dovitinib on day 22 at a reduced dose of 450 mg/day. Although this patient's QTcF remained stable, the patient was discontinued from treatment on day 26 because of pneumonia. Myocardial infarctions with ST-segment elevations occurred in 2 patients (3.0%), both grade 3 SAEs suspected to be related to dovitinib. One of these patients discontinued because of the myocardial infarction, whereas the other discontinued because of PD and experienced myocardial infarction 2 days after discontinuation from dovitinib. Both patients recovered from the myocardial infarctions. Eight of 63 patients (12.7%) with normal baseline assessment and postbaseline assessment(s) developed ≥1 episode of depressed ST-segment postbaseline without cardiac-related symptoms reported, including 1 patient with QT prolongation as described above.

ECHO/MUGA scans indicated clinically significant changes in left ventricular ejection fraction in 3 patients, including acute myocardial infarction (n = 1), myocardial infarction with elevated ST-segment and inverted T-waves (n = 1) as described above, and biphasic T-waves (n = 1). The patient with biphasic T-waves had a medical history of cardiac-related events and was discontinued from the study because of pulmonary embolism. This event was not suspected to be related to dovitinib.

Pharmacokinetic parameters were calculated for 65 patients on days 1 and 15 of cycle 1 (Supplementary Table S1). Maximal concentration (C_max) of dovitinib was reached in a median of 6 hours on days 1 and 15 of cycle 1. The mean C_max decreased by approximately 19% on day 15 (263.5 ng/mL) in comparison with day 1 (326.3 ng/mL). The decrease was greater (≈30%) for the mean area under the concentration curve from time zero to the last measurable sampling time (3933.3 h•ng/mL on day 15 vs. 5576.4 h•ng/mL on day 1). The half-life (geometric mean) of dovitinib decreased from approximately 24 hours on day 1 to approximately 11 hours on day 15. As seen previously, these data are supportive of CYP1A1/2 autoinduction by dovitinib (15).

Dovitinib at the maximum tolerated dose of 500 mg/day on a 5-days-on/2-days-off schedule was tolerated and demonstrated antitumor activity in heavily pretreated patients with RCC. Two partial responses (3.6%) and 29 stable diseases (52.7%) were achieved throughout the entire study period in patients previously treated with ≥1 VEGFR TKI and ≥1 mTOR inhibitor (PPES). In this patient population, disease control ≥2 and ≥4 months was achieved in 31 patients (56.4%) and 28 patients (50.9%), respectively. Median PFS and OS in this population were 3.7 and 8.6 months, respectively. Although a limited objective response was observed, clinical benefit was observed with dovitinib treatment in this heavily pretreated patient population.

Current therapies approved for the treatment of advanced or metastatic RCC that target the VEGF and mTOR pathways rarely induce sustained disease remission because of de novo and acquired resistance mechanisms (1, 19). Therefore, patients with advanced RCC previously treated with VEGF pathway and mTOR inhibitors need novel options for targeting antiangiogenic escape, angiogenesis, and tumor growth. FGF pathway signaling, which plays a role in tumor angiogenesis, has been suggested as a possible escape mechanism for VEGF pathway–targeted therapies (6–8). Dovitinib, a TKI that goes beyond VEGFR and PDGFR blockade by also targeting FGFR (13, 14), may overcome the resistance observed following treatment with VEGF pathway inhibitors. Similar to the results observed in the phase I portion of this study (15), there was a trend toward higher bFGF levels at baseline in patients previously treated with VEGF pathway inhibitors compared with patients previously treated with other inhibitors. In addition, a combined analysis of these data showed a similar trend, although the differences did not reach statistical significance. The pharmacodynamic analysis of plasma biomarkers reported here showed that dovitinib treatment inhibited VEGFR (as demonstrated by an increase in PIGF levels and a decrease in sVEGFR2 levels) and FGFR (as demonstrated by an increase in FGF23 levels), consistent with the results from the phase I portion of the study (15).

The most frequently reported, clinically notable adverse events, including diarrhea, fatigue/asthenia, and severe nausea and vomiting, are known adverse effects for VEGFR TKIs (20–24), including dovitinib (18, 25), and were observed at similar levels in the phase I portion of this study (15). Nearly half of all patients experienced an SAE suspected to be related to dovitinib. In contrast, only 15% of patients treated in the phase I portion of the study experienced an SAE (15). This increase in the number of SAEs may be because of the higher percentage of patients treated with multiple VEGFR TKIs and mTOR inhibitors in the phase II portion in comparison with the phase I portion of the study (82% vs. 50%). This may also explain the rate of dose reductions (38.8%) and delays or interruptions (56.7%) in this study (compared with 20.0% and 45.0% rates in the phase 1 portion, respectively), although a larger, controlled study would be needed to determine how the number of prior lines of therapy affects dovitinib tolerability. Hypertension was experienced by 19% of patients treated with dovitinib (18, 25). Although hypertension is a class effect of VEGF pathway–targeted therapies (26, 27), the overall incidence of hypertension observed with dovitinib treatment was lower than that reported for most VEGF pathway inhibitors; however, because of different patient populations and treatment histories, a direct comparison is not possible (28). Of note, the incidence of grade 3 or 4 events (10%) was similar to that reported for other agents targeting the VEGF pathway in RCC (23, 26, 29–31).

The results of this study suggest that dovitinib is an active and tolerable therapy for heavily pretreated patients with advanced or metastatic RCC. In this study, 82% of patients had previously been treated with both a VEGFR TKI and an mTOR inhibitor. Dovitinib showed clinical benefit with disease control in more than half of these heavily pretreated patients. The efficacy results observed in this study trended lower than that observed in the phase I portion of the study (15). The reasons for this are unclear but could reflect the higher percentage of patients treated with multiple VEGFR TKIs and mTOR inhibitors in the phase II portion of the study versus the phase I portion (82% vs. 50%), or the sequencing of these agents. Similarly, poststudy availability of VEGFR TKI agents (including pazopanib and investigational axitinib and tivozanib) could have affected OS.

Several other studies have demonstrated that third-line treatment with VEGF pathway– and mTOR-targeted therapies can provide clinical benefit in patients with metastatic RCC previously treated with VEGF pathway– or mTOR-targeted therapies (32–39). However, most of these studies are small, retrospective, or subgroup analyses and the data are confounded by different patient populations and treatment sequences. Prospective, randomized trials are needed to more clearly elucidate the appropriate sequential therapeutic regimen to be used in the third-line setting. A phase III study (NCT01223027) comparing dovitinib with sorafenib in patients with advanced RCC following failure on ≤1 VEGF pathway–targeted therapy and ≤1 mTOR-targeted therapy was recently completed. However, the preliminary report did not show benefit of dovitinib versus sorafenib in this patient population (40). Standard therapeutic options following disease progression on VEGF pathway–targeted and mTOR inhibitors remain very limited; although, dovitinib shows promise in this heavily pretreated patient population, its potential in RCC has to be determined.

V. Grünwald reports receiving speakers bureau honoraria from Astellas, Bayer, GlaxoSmithKline, Novartis, Pfizer, and Roche, and is a consultant/advisory for Bayer, GlaxoSmithKline, Novartis, Pfizer, and Roche. A. Ravaud is a consultant/advisory board member for Astellas, AVEO, Bayer, GlaxoSmithKline, Novartis, and Pfizer. C.-C. Lin reports receiving speakers bureau honoraria from Bayer, GlaxoSmithKline, Novartis, and Pfizer, and was a consultant/advisory board member for Novartis and Pfizer. J.E. Gschwend reports receiving speakers bureau honoraria from and was a consultant/advisory board member for Astellas, Bayer, GlaxoSmithKline, Janssen, Novartis, Pfizer, and Roche. M.M. Shi is an employee of and has ownership interest (including patents) in Novartis. B. Escudier reports receiving speakers bureau honoraria from AVEO, Bayer, GlaxoSmithKline, Novartis, and Pfizer. S. Beall, N. Pirotta, and M. Squires are employees of Novartis. No potential conflicts of interest were disclosed by the other authors.

Conception and design: B. Escudier, M. Shi

Development of methodology: B. Escudier, M. Shi

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): B. Escudier, V. Grünwald, A. Ravaud, Y.-C. Ou, C.-C. Lin, J.E. Gschwend, A. Harzstark, M. Shi, E. Angevin

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): B. Escudier, V. Grünwald, A. Ravaud, Y.-C. Ou, D. Castellano, A. Harzstark, N. Pirotta, M. Squires, M. Shi

Writing, review, and or revision of the manuscript: B. Escudier, V. Grünwald, A. Ravaud, D. Castellano, C.-C. Lin, J.E. Gschwend, A. Harzstark, S. Beall, N. Pirotta, M. Squires, M. Shi, E. Angevin

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Beall, M. Shi

Study supervision: B. Escudier, D. Castellano, M. Shi

The authors thank the other study principal investigators, W.H.J. Kruit, S. Tykodi, D. George, and Y.-H. Chang. They also thank J. Chang, A. Kay, M. Marker, and P. Sen from Novartis Pharmaceuticals Corporation for data analysis. Medical editorial assistance was provided by J. Brechbiel and P.J. Simon.

This work was supported by Novartis Pharmaceuticals Corporation for clinical studies and medical editorial assistance.