Originally identified as an inhibitor of Raf kinase, Sorafenib, during its pre-clinical development, proved to be endowed with a significant antiangiogenic activity, characterized by the ability to inhibit, at pharmacological concentrations, all three VEGF receptors (VEGFR-1, -2 and -3), the Platelet-Derived Growth Factor Receptor (PDGFR)-α and -β, in addition to a number of other kinases [34]. Depending on the experimental models considered, Sorafenib would act more as an anti-angiogenic (most of the time) or as antiproliferative/pro-apoptotic, agent [35].

Four phase I studies contributed to identify an effective dose to be used in later stages of development, i.e. 400 mg twice daily, continuous dosing; from these studies the capability of sorafenib to induce long-lasting disease stabilizations, more than well defined objective responses, clearly emerged [36].

Consequently, a randomization-discontinuation phase II trial (RDT) was performed, aimed at confirming such a putative cytostatic activity of the drug, as well as ruling out a possible indolent growth of the tumor itself [37]. In this study, all patients (mostly suffering from RCC) were initially treated with sorafenib in a run-in period of 12 weeks; following this period, patients were evaluated according to WHO criteria, and those who progressed dropped out of the study, those in response continued the treatment in an open-label fashion, while those with stable disease were randomized to receive either sorafenib or placebo. A significant improvement in progression-free survival (PFS) was then observed in patients randomized to receive sorafenib, compared with those randomized to receive placebo (24 vs. 6 weeks) [37].

This study confirmed the antitumor activity of Sorafenib and represented the rational basis for the subsequent conduct of the TARGET pivotal trial, conducted in pre-treated RCC patients (mainly with cytokines) [38].

In this randomized phase III trial, conducted globally, 903 patients were randomized to receive, in a double-blind fashion, either sorafenib or corresponding placebo. Main inclusion criteria were: histologic diagnosis of renal clear cell carcinoma, and a previous first-line systemic treatment. The primary efficacy endpoint of the study was from overall survival (OS), while PFS was among the secondary end-points; however, in anticipation of the potential confounding effect of a possible cross-over to active drug for patients initially randomized to placebo, a preplanned analysis that excluded these patients was included into the study design.

In January 2005, an independent evaluation of disease status showed a mean PFS of 5.5 months for Sorafenib-treated patients, compared to 2.8 months for patients treated with placebo, the difference being statistically significant and equivalent to a reduction in the risk of progression of 56 % [38].

On the basis of these results, it was allowed to cross-over to the active drug of those patients still receiving placebo, and these data were then sufficient to lead to the registration of Sorafenib by regulatory authorities.

As expected, the number of objective responses induced by Sorafenib was low [38], which was compatible with the now known cytostatic activity of the drug; in contrast, there was a high disease control rate (DCR), represented by the sum of objective responses with stabilization of disease.

Furthermore, the TARGET study confirmed the manageable safety profile of Sorafenib [38]; indeed, among the most frequent adverse events observed in patients treated with Sorafenib there were; diarrhea, skin rash, fatigue, hand-foot syndrome, as well as hypertension, while among the abnormalities in blood chemistry lymphopenia, hypophosphatemia, hyperlipasemia (without evidence of associated pancreatitis) and hypothyroidism, were recorded.

Regarding OS, the preplanned analysis that excluded patients treated with the active drug after the cross-over from the placebo arm, showed a statistically significant difference in favor of Sorafenib [39].

The use of Sorafenib in two Expanded Access Programs (EAPs), conducted in Europe and the United States, i.e. in a setting similar to that of everyday clinical practice, has allowed us to confirm the activity and tolerability of Sorafenib also in the subgroup of patients quite different from those usually considered for clinical trials, such as the elderly, those with brain metastases, or those with non-clear cell histologies [40, 41].

Following the publication of the first results of the TARGET study, conducted in pre-treated patients, a randomized phase II trial in which Sorafenib was compared with IFN-α in a pure first-line setting, was designed and conducted [42]. Surprisingly, the PFS of the patients from this study was not statistically different between the two treatment arms (5.7 months for patients treated with Sorafenib and 5.6 months for patients treated with IFN-α) [42].

Obviously, the lack of superiority of Sorafenib over IFN-α in this randomized, phase II, study has been interpreted as a sign of ineffectiveness of Sorafenib in the first line setting. This study, however, had a number of significant methodological flaws, that call into question these conclusions.

Indeed, in subsequent studies in which Sorafenib has been used in first-line, PFS values of around 9 months were observed, shortening the efficacy gap between Sorafenib and the other tyrosine kinase inhibitors used in the first line [43, 44].

Sunitinib is an oral multikinase inhibitor selectively directed against all three VEGF receptors (VEGFR-1, -2 and -3), against the PDGFR-α and -β, against the Fibroblast Growth Factor Receptor-1 (FGFR-1) as well as against a range of other kinases [45].

From phase I studies, the dose of 50 mg per day within a 4 weeks on, 2 weeks off, scheduled emerged as the one to be used in later stages of development [45].

Two phase II studies conducted in patients with RCC and refractory to cytokines, not only clearly showed an extremely high rate of objective responses (40 % and 39 %, respectively), but also yielded a unprecedentedly long time-to-progression (TTP), i.e., 8.7 months, as well as an intriguing OS of 16.4 months [46, 47]. These striking results have not only led to an accelerated approval by the US Food and Drug Administration (FDA), but they also represented the rational basis for the subsequent conduct of a pivotal, registrative, trial.

In this randomized phase III trial, conducted globally, 750 patients not previously treated for their metastatic disease were randomized to receive either Sunitinib, or IFN-α (given s.c. at a dose of 9 MU three times week) [48].

The primary endpoint of the study was PFS, while OS was among the secondary end-points.

The average PFS in the group of patients treated with Sunitinib was significantly longer than that of patients treated with IFN-α (11 vs. 5 months), corresponding to a HR 0.42 [48]. The advantage in terms of PFS in favor of Sunitinib was then maintained in all three prognostic groups according to the classification of Motzer.

As expected based on the results of previous phase I and II studies, Sunitinib has been shown to induce objective responses in a high percentage of patients (31 %), in contrast to an overall response rate of only 6 % for IFN-α [48]. Regarding tolerability, patients treated with Sunitinib showed a higher incidence of diarrhea, vomiting, hypertension, hand-foot syndrome, and neutropenia [48]. Overall, a better quality of life was observed in patients treated with sunitinib, compared to what was observed in those treated with IFN-α [49].

Regarding overall survival, although it was higher in patients treated with Sunitinib compared with those treated with IFN-α (26.4 vs 21.8 months, respectively), this difference did not reach statistical significance [50]. However, since the primary endpoint of the study was PFS (and not OS), it is obvious that the study was simply underpowered to show a significant benefit in terms of OS.

As with Sorafenib, the use of Sunitinib in an unselected patient population as the one enrolled into its EAP, which was conducted on a global scale, allowed to confirm the activity of this drug in a general patients’ population, as well as in specific subpopulations of patients (e.g., elderly, patients with metastatic brain disease, patients with non clear-cell histotypes, etc. …) [51].

Subsequently, a randomized phase II trial [52] compared the traditional schedule of Sunitinib (50 mg daily, for 4 weeks every 6) with a reduced (37.5 mg per day), but continuous, dose; from a certain viewpoint surprisingly, the alternative schedule, not only proved to be less active, but also was not better tolerated – as initially expected, thus confirming pharmacokinetic data suggesting the existence of a close relationship between the AUC of Sunitinib and its activity [53].

The recombinant humanized monoclonal antibody directed against VEGF, Bevacizumab is able to selectively bind and neutralize all active isoforms of VEGF (also known as VEGF-A), but not other members of the family of VEGF, i.e. VEGF-B, -C and -D [54].

The activity of Bevacizumab against metastatic RCC was initially evaluated in a randomized phase II trial, in which 116 patients with advanced RCC refractory to a previous treatment were randomized to receive placebo or low-dose (3 mg/kg) Bevacizumab, or high-dose Bevacizumab (10 mg/kg), every 2 weeks, intravenously [55].

The TTP observed in the group treated at a dose of 10 mg/kg (4.8 months) was significantly longer than that observed in the placebo group (2.5 months), while the observed difference between the group treated with the low dose and the group treated with placebo was borderline; the dosage of 10 mg/kg and allowed also to achieve an objective response rate of 10 %, some kind of tumor shrinkage having been observed in the majority of patients [55].

The subsequent development of Bevacizumab in RCC continued with the combination with IFN-α, and this combination – even in the absence of a clear pre-clinical rationale – was thus evaluated within two randomized phase III, very similar (but not equal) between them: the pivotal AVOREN study [56] and the American CALGB 90206 study [57].

In the AVOREN study, 649 patients with clear cell RCC (or a mixed histology comprising at least 50 % of clear cells) were randomized to receive, until progression of disease, a combination of IFN-α (administered s.c. at the dose of 9 MU three times a week, with a possible dose reductions for toxicity) plus Bevacizumab (10 mg/kg every 2 weeks) or IFN-α plus placebo. The primary endpoint of the study was OS, while secondary endpoints were PFS, objective response rate, as well as tolerability profile [56].

A statistically significant advantage in terms of PFS in favor of the Bevacizumab-containing arm (median 10.2 vs. 5.4 months, HR = 0.63) was documented, even though this advantage was observed only in patients favorable (median: 12.9 vs. 7.6 months, HR = 0.60) and intermediate prognosis (median: 10.2 vs. 4.5 months, HR = 0.55), according to Motzer’s criteria, but not in those with a poor prognosis, the benefit in terms of PFS being indeed lost in this latter subgroup (median 2.2 vs 2.1 months) [56].

The objective response rate observed in the Bevacizumab-containing arm was 31 %, as compared to 13 % obtained from the IFN-α plus placebo arm, while a DCR was obtained in 70 % of the patients treated with the combination of bevacizumab and IFN-α; finally, the average duration of responses and stabilization of disease was 13 and 10 months in the Bevacizumab- and placebo-containing arm, respectively [56].

As far as tolerability, the experimental treatment proved to be very well tolerated; in particular, the combination of Bevacizumab and IFN-α induced grade 3–4 proteinuria in 6.5 % of patients, together with as a modest, although not significant, increase in the incidence of haemorrhage, hypertension, thromboembolic events, and gastrointestinal perforations, with respect to the combination of IFN-α and placebo [56].

Protocol-driven IFN-α dose reductions in the event of toxicity did not lead to any loss of efficacy of the combination arm, and indeed, the group of 131 patients who received a reduced dose of IFN-α not only had a smaller number of severe adverse events, but also experienced a median PFS greater than that of the intention-to-treat analysis (12.4 vs 10.2 months) [58].

This figure has been somehow confirmed by a subsequent study of the association between Bevacizumab and low-dose IFN-α (the BEVLIN study) [59], which yielded extremely intriguing PFS, OS and overall response rates (15.3 months [95 % confidence interval: 11.7–18.0], 30.7 months [95 % confidence interval: 25.7-not reached], and 28.8 %).

With regard to OS, the primary endpoint of the AVOREN study, the statistical design assumed that the experimental arm could achieve an OS advantage of about 4 months compared to the control arm, corresponding to a reduction in the risk of death from any cause of 34 %. Surprisingly, the final analysis, performed at a median follow-up of 22 months [60], did not show any statistically significant difference between the two treatment arms (median overall survival being 23.3 vs. 21.3 months).

As for the American CALGB 90206 study, which was similar, but not identical, to the AVOREN study, the experimental arm (i.e., Bevacizumab + IFN-α) yielded a superior median PFS (8.4 vs 4.9 months, HR: 0.71), an almost double overall response rate, and an expected higher percentage of severe adverse events, compared to the control arm (i.e., IFN-α alone) [57].

Also with regard to OS, the American study proved to be in line – with absolute values also in this case lower – with the AVOREN study. OS between the two treatment arms, infact, did not reach the statistical significance (18.3 months for the Bevacizumab plus IFN-α arm vs. 17.4 for IFN-α alone arm), even in the presence of a risk reduction of death of 14 %, exactly the same observed in the AVOREN study, but lower than expected from the original statistical design of the two studies [57].

Failure to achieve, for both studies, the primary endpoint (i.e., OS), could however be explained on the basis of two considerations. First, an unrealistic estimate of the activity of IFN-α, from which the statistical design of the two studies was built.

Furthermore, beyond a certain number of patients who crossed-over to Bevacizumab plus IFN-α at progression on IFN-α alone, even more relevant appears the problem of subsequent active treatment received by the patients. In fact, in the AVOREN study, 55 % of patients treated with Bevacizumab plus IFN-α, and 63 % of those treated with IFN-α plus placebo, have subsequently received one or more active treatments [61].

Despite the unexpected lack of significance in terms of overall survival for both, largely justified on the basis of the above considerations, the AVOREN and CALGB 90206 studies have confirmed the importance of VEGF as a therapeutic target in RCC, as well as the substantial activity and excellent tolerability of Bevacizumab plus IFN-α in first-line treatment of this cancer.

Pazopanib is another oral multi-kinase inhibitor capable of inhibiting the activation of different tyrosine kinases heavily implicated in the mechanisms of angiogenesis (VEGFR-1, -2 and -3, PDGFR-α and -β, etc. …) [62].

The recommended dose resulting from a phase I study, which showed a correlation between plasma concentrations of pazopanib and development of hypertension in patients treated, was equal to 800 mg/day [63].

The first demonstration of activity of pazopanib in RCC came from a randomized discontinuation phase II study, with an overall PFS of 52 weeks (95 % CI: 44–60), an overall response rate of 34.7 % and a DCR of 79.5 % [64].

Based on the results of this study, a pivotal phase III trial was designed, in which 435 patients with locally advanced or metastatic RCC were randomized 2:1, in a double-blind fashion, to receive either pazopanib or placebo. Patients could be treatment- naive or pre-treated with a line of immunotherapy, its primary endpoint being PFS [65].

A significant benefit in terms of PFS in favor of Pazopanib was observed in both groups of patients, with a median PFS in of 11.1 months in treatment-naive patients (vs. 2.8 months for placebo-treated subjects, HR: 0.4), and 7.4 months (vs. 4.2, HR: 0.54) in cytokine pre-treated patients [65]. An objective response was then observed in 30 % of patients treated with Pazopanib, with a median duration of responses equal to 58.7 weeks; as far as OS, its assessment was flawed by the very high percentage of patients who have crossed-over from the placebo to the active treatment arm [65].

The most common adverse events attributable to Pazopanib, still mostly of grade 1 and 2, included: diarrhea, hypertension and fatigue, while the most frequent laboratory abnormality was transaminases elevation, an event seen in more 50 % of patients; in particular, ALT increase proved to be the commonest Pazopanib-related adverse event of grade 3 or 4 [65].

Recently, the results of two studies (PISCES and COMPARZ studies), directly comparing pazopanib and sunitinib, were presented.

In the COMPARZ study, 1,110 patients with clear-cell, metastatic renal-cell carcinoma, were randomized 1:1, to receive Pazopanib (given at the standard dose of 800 mg once daily, continuous dosing) or sunitinib (50 mg once daily for 4 weeks, followed by 2 weeks’ rest), its primary end-point being PFS; the study was powered to show the noninferiority of Pazopanib versus Sunitinib.

Pazopanib proved to be not inferior to Sunitinib with respect to PFS (HR = 1.05; 95 % confidence interval [CI]: 0.90–1.22), meeting the predefined non-inferiority margin (upper bound of the 95 % CI: <1.25); also OS was similar (HR = 0.91; 95 % CI, 0.76–1.08). Furthermore, 11 of 14 health-related quality-of-life (QoL) domains favored Pazopanib, when it came to QoL [66].

Differently from COMPARZ (and almost uniquely), the PISCES study had as its primary end-point preference of patients. In this innovative study, patients with metastatic RCC were randomized to pazopanib (800 mg/day) for 10 weeks, a 2-week washout, and then Sunitinib (50 mg/day, 4 weeks on/2 week off) for other 10 weeks, or the reverse sequence. The primary endpoint, patient preference for a specific treatment, was assessed by questionnaire at the end of the two treatment periods. Other endpoints and analyses included reasons for preference, and HRQoL. Significantly more patients preferred pazopanib (70 %) over sunitinib (22 %), whilst 8 % expressed no preference (P < 0.001) with all the preplanned sensitivity analyses, including the intent-to-treat population, which statistically favored Pazopanib [67]. Less fatigue and better overall QoL were the main reasons for preferring Pazopanib, with less diarrhea was the main reason of their choice for those patients who preferred Sunitinib. Again, adverse events were consistent with each drug’s known profile, but Pazopanib proved to be superior to sunitinib in terms of QoL, thus corroborating the QoL results of the COMPARZ study [67].

Even though methodologically not faultless, these two important studies have clearly confirmed the role of Pazopanib as a credible alternative to Sunitinib for the treatment of patients with RCC in first-line treatment.

Axitinib is a so-called third-generation VEGFRs TKI [68], characterized by a particular selectivity of action (for all three VEGF receptors) and a high power.

Axitinib pivotal phase III trial [69], the AXIS study, was conducted in a second-line setting, in patients pre-treated with a variety of first-line treatment, and was the very first study in RCC which compared head-to-head two active drugs, Sorafenib having been chosen as the control arm.

In this study, Axitinib proved to be superior in terms of PFS (primary endpoint of the study) to Sorafenib (which however proved to be active), but not in terms of OS, which did not differ between the two treatment arms.

Indeed, median PFS was 6.7 months with Axitinib compared to 4.7 months with Sorafenib (HR = 0.665; 95 % CI: 0.544–0.812; p < 0.0001), the biggest advantage in favor of Axitinib having been observed in patients pre-treated with cytokines [69]. As far as OS, it was 20.1 months (95 % CI: 16.7–23.4) with Axitinib and 19.2 months (17.5–22.3) with Sorafenib (HR = 0.969, 95 % CI: 0.800–1.174; p = 0.3744) [70].

In a subsequent randomised, open-label, phase III trial, patients with treatment-naive, clear-cell metastatic RCC were randomly assigned (in a 2:1 fashion) to receive Axitinib 5 mg twice daily, or Sorafenib 400 mg twice daily. The primary endpoint of this first-line study was PFS, assessed by centralized independent review [71].

One hundred ninety two patients were randomized into the Axitinib arm, while 96 other patients received Sorafenib. There was no significant difference in median PFS between patients treated with the two durgs, even though a clinically relevant advantage was recorded in patients treated with Axitinib (10.1 months [95 % CI: 7.2–12.1] vs 6.5 months [4.7–8.3], respectively; HR = 0.77, 95 % CI: 0.56–1.05) [71].

This discrepancy between the lack of statistically significant difference in terms of PFS, and the absolute gain achieved by Axitinib-treated patients was mainly due to an overestimation of the superiority of Axitinib over Sorafenib at the time of study design.