Early-phase clinical drug development of novel agents: a changing paradigm.

Abstract

The mitogen-activated protein kinase (MAPK) cell signaling pathway has a significant role in tumoral proliferation and survival and is dysregulated in approximately one-third of cancers [1.Tolcher A.W. Peng W. Calvo E. Rational approaches for combination therapy strategies targeting the MAP kinase pathway in solid tumors.Mol Cancer Ther. 2018; 17: 3-16Crossref PubMed Scopus (66) Google Scholar]. In fact, dual vertical inhibition with BRAF and MEK inhibitors (MEKi) is already established as adjuvant and first-line therapy for melanoma and initial treatment of advanced non-small-cell lung cancer harboring BRAFV600 mutations [2.Robert C. Karaszewska B. Schachter J. Two year estimate of overall survival in COMBI-v, a randomized, open-label, phase III study comparing the combination of dabrafenib (D) and trametinib (T) with vemurafenib (vem) as first-line therapy in patients (pts) with unresectable or metastatic BRAF V600E/K mutation-positive cutaneous melanoma.Ann Oncol. 2015; 26 (Abstract): 3301Google Scholar, 3.Ascierto P.A. McArthur G.A. Dréno B. et al.Cobimetinib combined with vemurafenib in advanced BRAF(V600)-mutant melanoma (coBRIM): updated efficacy results from a randomised, double- blind, phase 3 trial.Lancet Oncol. 2016; 17: 1248-1260Abstract Full Text Full Text PDF PubMed Scopus (649) Google Scholar, 4.Flaherty K. Davies M.A. Grob J.J. et al.Genomic analysis and 3-y efficacy and safety update of COMBI-d: a phase 3 study of dabrafenib (d) + trametinib (T) vs monotherapy in patients (pts) with unresectable or metastatic BRAF V600E/K-mutant cutaneous melanoma.J Clin Oncol. 2016; 34 (Abstr 9502)PubMed Google Scholar, 5.Dummer R. Ascierto P.A. Gogas H.J. et al.Overall survival in patients with BRAFmutant melanoma receiving encorafenib plusbinimetinib versus vemurafenib or encorafenib (COLUMBUS): a multicentre, open-label, randomised, phase 3 trial.Lancet Oncol. 2018; 19: 1315-1327Abstract Full Text Full Text PDF PubMed Scopus (334) Google Scholar, 6.Planchard D. Smit E.F. Groen H.J.M. et al.Dabrafenib plus trametinib in patients with previously untreated BRAFV600E-mutant metastatic non-small-cell lung cancer: an open-label, phase 2 trial.Lancet Oncol. 2017; 18: 1307-1316Abstract Full Text Full Text PDF PubMed Scopus (545) Google Scholar]. In addition, the MAPK pathway is also involved in immune signaling, as the RAF/MEK/ERK pathway is crucial for T-cell activation, development and differentiation [7.Chen D. Heath V. O'Garra A. et al.Sustained activation of the Raf-MEK-ERK pathway elicits cytokine unresponsiveness in T cells.J Immunol. 1999; 163: 5796-5805PubMed Google Scholar, 8.Ebert P.J. Cheung J. Yang Y. et al.MAP kinase inhibition promotes T cell and anti-tumor activity in combination with PD-L1 checkpoint blockade.Immunity. 2016; 44: 609-621Abstract Full Text Full Text PDF PubMed Scopus (464) Google Scholar]. Preclinical models have shown that MEKi upregulate MHC I expression, increase tumor-infiltrating lymphocytes and enhance PD-1 inhibitor activity [8.Ebert P.J. Cheung J. Yang Y. et al.MAP kinase inhibition promotes T cell and anti-tumor activity in combination with PD-L1 checkpoint blockade.Immunity. 2016; 44: 609-621Abstract Full Text Full Text PDF PubMed Scopus (464) Google Scholar]. Therefore, the combination of MEKi with immune checkpoint inhibitors (ICIs) have a strong rationale, especially in tumor types such as colorectal cancer (CRC) where the MAPK pathway has a relevant role [9.Fang J.Y. Richardson B.C. The MAPK signalling pathways and colorectal cancer.Lancet Oncol. 2005; 6: 322-327Abstract Full Text Full Text PDF PubMed Scopus (707) Google Scholar] and ICIs, in general, have shown very modest activity [10.Modest D.P. Pant S. Sartore-Bianchi A. Treatment sequencing in metastatic colorectal cancer.Eur J Cancer. 2019; 109: 70-83Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar]. In this setting, the phase Ib NCT01988896 trial reported in this issue of Annals studied the combination of fixed bi-weekly doses of atezolizumab (anti-PDL-1) with uptitrating doses of cobimetinib (MEK1/2 inhibitor) in solid tumors [11.Hellman M.D. Kim T.W. Lee C.B. et al.Phase Ib study of atezolizumab combined with cobimetinib in patients with solid tumors.Ann Oncol. 2019; 30 (doi.org/101093/annonc/mdz113)Abstract Full Text Full Text PDF Scopus (74) Google Scholar]. Despite the fact that some exciting initial results of the combination were seen in this study, with preliminary antitumor activity and adequate tolerance profile at the dose escalation phase according to the protocol definition of DLTs, the overall results of the study did not confirm these encouraging findings. Overall, the combination was hard to tolerate at the given dose and schedule. In fact, the same treatment combination has recently failed to improve survival in the IMblaze370 study, a phase III trial of this combination versus regorafenib or atezolizumab in 363 heavily pretreated patients with mCRC. Overall survival, the primary end point, was 8.9 months for the combination compared with 8.5 months for regorafenib [HR 1.00 (0.73–1.38), P = 0.99] and 7.1 months for atezolizumab in the intention-to-treat population. Disappointing results were also obtained with regard to median PFS (2 months). Grade 3–4 adverse events (AEs) occurred at similar rates in the combination and regorafenib arms (45% and 49%, respectively, with 56% patients with any grade of diarrhea) [12.Bendell J. Ciardiello F. Tabernero J. et al.Efficacy and safety results from IMblaze370, a randomised phase III study comparing atezolizumab+cobimetinib and atezolizumab monotherapy vs regorafenib in chemotherapy-refractory metastatic colorectal cancer.Ann Oncol. 2018; 29 (Abstr LBA-004): v123Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar], which is considerable for this fragile population. It is critical for effective clinical drug development to be able to distinguish as soon as possible whether a given drug or combination is not providing our patients with clinical benefit. Also, optimizing the schedule and dose, especially in combinations, is a must to assess a therapeutic regimen at its best potential. Here, we identify three different aspects that, potentially, might have predicted the outcomes of this rationally sound combination and might have led to the ability of making it ‘fail fast’ before launching late phase studies. (1) Tolerability profile and alternative definitions for dose-limiting toxicity (DLT). One of the key concerns with MEKi combinations is its tolerability profile, as previously seen (Table 1). For example, the double combination of cobimetinib–vemurafenib was associated with 65% G3 AEs, with gastrointestinal events, photosensitivity, and transaminitis as the most frequent toxicities [22.Larkin J. Ascierto P.A. Dréno B. et al.Combined vemurafenib and cobimetinib in BRAF-mutated melanoma.N Engl J Med. 2014; 371: 1867-1876Crossref PubMed Scopus (1497) Google Scholar]. Unacceptable toxicity with immunotherapy and targeted therapies has also been described with the triple combination of dabrafenib–trametinib–pembrolizumab in melanoma, which led to a significant 58% rate of G3–5 toxicities in the phase II trial [23.Ascierto PA, Dummer R et al. KEYNOTE-022 Part 3: phase 2 randomized study of 1L dabrafenib (D) and trametinib (T) plus pembrolizumab (Pembro) or placebo (PBO) for BRAF-mutant advanced melanoma. Presented at ESMO 2018 Congress, Munich, Germany, 19–23 October 2018; Abstract 1244O.Google Scholar]. Also, dabrafenib–trametinib–ipilimumab [24.Minor D.R. Puzanov I. Callahan M.K. et al.Severe gastrointestinal toxicity with administration of trametinib in combination with dabrafenib and ipilimumab.Pigment Cell Melanoma Res. 2015; 28: 611-612Crossref PubMed Scopus (109) Google Scholar] led to a high frequency of severe side-effects, where hepatoxicity was significant and, sometimes, unpredictable.Table 1Combinations of immune checkpoints inhibitors plus targeted drugs involved in the MAPK pathwayRef.Trial codeTumour typeTrial designDosing schedulesDLTG3/4 TRAEComments[13.Gettinger S. Hellmann M.D. Chow L.Q.M. et al.Nivolumab plus erlotinib in patients with EGFR-mutant advanced NSCLC.J Thorac Oncol. 2018; 13: 1363-1372Abstract Full Text Full Text PDF PubMed Scopus (117) Google Scholar]NCT01454102NSCLC EGFR+I (safety lead-in)Nivolumab 3 mg/kg Q2W and erlotinib 150 mg daily–24%Tolerable. 15% ORR[14.Oxnard G.R. Ramalingam S.S. Ahn M.-J. et al.Preliminary results of TATTON, a multi-arm phase Ib trial of AZD9291 combined with MEDI4736, AZD6094 or selumetinib in EGFR-mutant lung cancer.J Clin Oncol. 2015; 33 (2509): 2509Crossref PubMed Google Scholar]NCT02143466NSCLC EGFR+Ib (rolling 6)Osirmetinib 80 mg QD and durvalumab 10 mg/kg Q2WNo DLTs–3 PR[15.Chia S.K.L. Bedard P.L. Hilton J. et al.A phase I study of a PD-L1 antibody (Durvalumab) in combination with trastuzumab in HER-2 positive metastatic breast cancer (MBC) progressing on prior anti HER-2 therapies (CCTG IND.229)[NCT02649686].J Clin Oncol. 2018; 36: 1029Crossref Google Scholar]NCT02649686HER-2 positive mBCI (safety lead-in)Durvalumab (1125 mg day 1) and trastuzumab (8 mg/kg loading then 6 mg/kg day 1) Q3WNo DLTs (0/6)6.6%No significant clinical activity[16.Boland P.M. Hutson A. Maguire O. et al.A phase Ib/II study of cetuximab and pembrolizumab in RAS-wt mCRC.J Clin Oncol. 2018; 36: 834Crossref Google Scholar]NCT02713373RAS-wt mCRCIb (safety lead-in with a de-escalation design)Pembrolizumab 200 mg Q3W and cetuximab 400 mg/m2 loading dose, followed by 250 mg/m2 weeklyNo DLTs88% (TRAE and not related)66% patients achieved SD lasting ≥16 weeks[17.Ribas A. Hodi F.S. Lawrence D. 1216O - KEYNOTE-022 update: phase 1 study of first-line pembrolizumab (pembro) plus dabrafenib (D) and trametinib (T) for BRAF-mutant advanced melanoma.Ann Oncol. 2017; 28: v428Google Scholar]NCT02130466BRAFV600E/K mutant melanoma1/2 (safety lead-in with a de-escalation design)Pembrolizumab 2 mg/kg Q3W + dabrafenib 150 mg b.i.d.+trametinib 2 mg QD20%67%67% ORR[18.Hwu P. Hamid O. Gonzalez R. 1109PD - preliminary safety and clinical activity of atezolizumab combined with cobimetinib and vemurafenib in BRAF V600-mutant metastatic melanoma.Ann Oncol. 2016; 27: 379-400PubMed Google Scholar]NCT01656642BRAFV600-mutant melanomaIb (safety lead-in)Atezolizumab 800 mg Q2W, cobimetinib 60 mg QD 21/28 days and vemurafenib 720 mg b.i.d. (after a run-in period with cobimetinib–vemurafenib)–35.7%93% ORR[19.Ribas A. Butler M. Lutzky J. et al.Phase I study combining anti-PD-L1 (MEDI4736) with BRAF (dabrafenib) and/or MEK (trametinib) inhibitors in advanced melanoma.J Clin Oncol. 2015; 33: 3003Crossref Google Scholar]NCT02027961MelanomaI (3+3)Cohort A (durvalumab 10 mg/kg Q2W, dafrafenib 150 mg BD and Trametinib 2 mg OD) Cohort B (durvalumab 10 mg/kg Q2W, trametinib 2 mg OD) Cohort C (sequential trametinib 2 mg OD → durvalumab 10 mg/kg Q2W)4%Cohort A: 33% Cohort B: 30% Cohort C: 17%–[20.Ribas A. Hodi F.S. Callahan M. et al.Hepatotoxicity with combination of vemurafenib and ipilimumab.N Engl J Med. 2013; 368: 1365-1366Crossref PubMed Scopus (576) Google Scholar]NCT01400451BRAFV600 mutant melanoma1 (two escalation cohorts planned)Ipilimumab 3 mg/kg Q3W and vemurafenib explored at two different doses (960 mg BD and 720 mg BD)66%Significant severe hepatotoxicity–[21.Puzanov. Combining targeted and immunotherapy: BRAF inhibitor dabrafenib (D) ± the MEK inhibitor trametinib (T) in combination with ipilimumab (Ipi) for V600E/K mutation-positive unresectable or metastatic melanoma (MM).J Transl Med. 2015; 13: K8Crossref PubMed Scopus (0) Google Scholar]NCT01767454BRAFV600 mutant melanoma1 (one dose level explored)Cohort 1: dabrafenib 150 mg b.i.d.+ipilimumab 3 mg/kg Q3W × 4 dosesCohort 2: dabrafenib 100 mg b.i.d., trametinib 1 mg QD and ipilimumab 3 mg/kg Q3W × 4 doses–Cohort 1: 0 Cohort 2: 42.9%Two colon perforation in the triplet, no further development.Ref., reference; DLT, dose-limiting toxicity; TRAE, treatment-related adverse event; OD, once per day; BD, twice per day; NSCLC, non-small-cell lung cancer; ORR, overall response rate; mBC, metastatic breast cancer; RAS-wt, RAS wild-type; mCRC, metastatic colorectal cancer; SD, stable disease; PR, partial response. Open table in a new tab Ref., reference; DLT, dose-limiting toxicity; TRAE, treatment-related adverse event; OD, once per day; BD, twice per day; NSCLC, non-small-cell lung cancer; ORR, overall response rate; mBC, metastatic breast cancer; RAS-wt, RAS wild-type; mCRC, metastatic colorectal cancer; SD, stable disease; PR, partial response. In combination with atezolizumab, cobimetinib was escalated in a classic 3 + 3 design for only 3 different dosing schedules (20, 40 and 60 mg QD, 21 days on/7 days off), with just 14 patients included in the escalation part and, of them, only 6 at the recommended dose (RD) of 60 mg. Eight patients (57%) developed AEs leading to dose modifications or interruptions, five of them at the lower dose levels. Despite this, there were no DLTs reported within the 28-day DLT period, and the maximum tolerated dose (MTD) was not reached, as per protocol. In spite of these red warning flags, without further fine tuning of the optimal schedule and dose, the expansion part was run then at the 60 mg dose level in five different cohorts: mCRC (59 pts), NSCLC (20 pts), melanoma (20 pts) and two serial biopsy cohorts (solid tumors 16 pts and CRC 21 pts, these on a modified schedule). Not surprising, at the RD of the dose escalation part, the rate of patients with toxicity limiting further administration at same dose (DLTs, by definition) was unacceptably high (78%). Interestingly, while the general rate for discontinuation or dose reduction in the expansion part (all patients) was ∼70%, in the serial biopsy mCRC cohort (the only one with a different dosing schedule of 14 days on/14 off) this rate decreased to 57%. Typically, ICI toxicity may occur within weeks to months after the initiation of treatment, where skin and liver toxicities occur before gastrointestinal or endocrine AEs [25.Haanen J.B.A.G. Carbonnel F. Robert C. et al.Management of toxicities from immunotherapy: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up.Ann Oncol. 2017; 28: iv119-iv142Abstract Full Text Full Text PDF PubMed Scopus (1304) Google Scholar]. Because of this, studies involving ICIs may not achieve the true MTD, as the DLT period normally includes just the first cycle and the relevant ICI toxicities occur often afterwards [26.Brahmer J.R. Drake C.G. Wollner I. et al.Phase I study of single-agent anti-programmed death-1 (MDX-1106) in refractory solid tumors: safety, clinical activity, pharmacodynamics, and immunologic correlates.J Clin Oncol. 2010; 28: 3167-3175Crossref PubMed Scopus (2311) Google Scholar, 27.Patnaik A. Kang S.P. Tolcher A.W. Phase I study of MK-3475 (anti-PD-1 monoclonal antibody) in patients with advanced solid tumors.J Clin Oncol. 2012; 30: 2512Crossref Google Scholar]. Similarly, it has been reported that moderate to severe toxicities may appear after the first cycle in up to 50% of patients treated with targeted agents [28.Postel-Vinay S. Gomez-Roca C. Molife L.R. et al.Phase I trials of molecularly targeted agents: should we pay more attention to late toxicities?.J Clin Oncol. 2011; 29: 1728-1735Crossref PubMed Scopus (104) Google Scholar]. Therefore, late-onset toxicities occur in a significant subset of patients and those might be not properly recorded or taken into consideration as DLTs. But it is not just a matter of time to onset, it is also about the grade of toxicity interfering with patient’s quality of life, as cumulative diverse but milder toxicities can drive decreased dose intensity. Durable or repeated grade 2 toxicities might lead to ‘limitations of the dose’ of treatments that are intended to be maintained for a prolonged time. Perhaps, for immune-related and target-based drugs, a more appropriate concept would be the ‘treatment-limiting toxicity’ rather than the classic DLT, taking into account late onset AEs, which drive dose reductions even though they are not classified as severe, as well as acute severe infusion-related reactions typical of some immuno-oncology (IO) agents. In the cobimetinib–atezolizumab trial, it would have been interesting to study if fewer dose reductions or discontinuations could have modified outcomes and maximized its potential. (2) More appropriate use of PK/PD surrogate information to optimize dosing schedules. The optimal bio-immunological dose (OBD) might be defined as the dose with the highest desirability in the risk–benefit trade-off [29.Liu S. Guo B. Yuan Y. A Bayesian phase I/II trial design for immunotherapy.J Am Stat Assoc. 2018; 113: 1016-1027Crossref PubMed Scopus (21) Google Scholar], which clearly applies for the mechanism of action of the two combined drugs. A proper selection of the OBD based on surrogate information was done, for example, during the early-phase studies of the dual MEK-RAF inhibitor RO5126766. A previous phase I study explored high-dose schedules finding a 40% shrinkage but with a challenging toxicity profile [30.Martinez-Garcia M. Banerji U. Albanell J. et al.First-in-human, phase I dose-escalation study of the safety, pharmacokinetics, and pharmacodynamics of RO5126766, a first-in-class dual MEK/RAF inhibitor in patients with solid tumors.Clin Cancer Res. 2012; 18: 4806-4819Crossref PubMed Scopus (116) Google Scholar]. Further PK studies showed that less intense, intermittent doses could improve its toxicity profile, and this was confirmed in a redesigned trial where encouraging clinical activity was reported with no severe AEs [31.de Miguel-Luken M. Roda D. Perez R. A pharmacokinetic (PK) and pharmacodynamic (PD) biomarker-driven phase I study of intermittent, low dose intensity schedules of the dual MEK/RAF inhibitor, RO5126766 (RO) in patients (pts) with advanced solid tumors.J Clin Oncol. 2015; 33: 104Crossref Google Scholar]. Some recent data in patient-derived tumor xenografts indicated that intermittent RAF, MEK and ERK inhibition resulted in inhibition of tumor growth without apparent toxicity in mice [32.Xue Y. Martelotto L. Baslan T. et al.An approach to suppress the evolution of resistance in BRAFV600E-mutant cancer.Nat Med. 2017; 23: 929-937Crossref PubMed Scopus (102) Google Scholar]. Also, MEKi pulsatile dosing may maintain inhibited ERK signaling sufficiently in the tumor while allowing normal tissue recovery. This approach could perhaps even increase administered doses of the MEKi [32.Xue Y. Martelotto L. Baslan T. et al.An approach to suppress the evolution of resistance in BRAFV600E-mutant cancer.Nat Med. 2017; 23: 929-937Crossref PubMed Scopus (102) Google Scholar]. Moreover, other mouse models suggest that intermittent RAF inhibition may extend the period of response and even prevent drug resistance [33.Yaeger R. Yao Z. Chapman P.B. Intermittent treatment with MEK inhibitors in patients with oncogenic RAF-driven tumors: a potential strategy to manage toxicity and maintain efficacy.JCO Precis Oncol. 2018; 2: 1-2PubMed Google Scholar]. On the other hand, dose and effectiveness do not seem to follow a proportional correlation with ICIs. In different comparative pembrolizumab trials, 10 mg/kg Q3W showed very similar response rates than the same dose given Q2W or even 2 mg/kg Q3W [34.Robert C. Ribas A. Wolchok J.D. et al.Anti-programmed-death-receptor-1 treatment with pembrolizumab in ipilimumab-refractory advanced melanoma: a randomised dose-comparison cohort of a phase 1 trial.Lancet. 2014; 384: 1109-1117Abstract Full Text Full Text PDF PubMed Scopus (1432) Google Scholar, 35.Ribas A. Hamid O. Daud A. et al.Association of pembrolizumab with tumor response and survival among patients with advanced melanoma.JAMA. 2016; 315: 1600-1609Crossref PubMed Scopus (726) Google Scholar, 36.Freshwater T. Kondic A. Ahamadi M. et al.Evaluation of dosing strategy for pembrolizumab for oncology indications.J Immunother Cancer. 2017; 5: 43Crossref PubMed Scopus (156) Google Scholar]. With nivolumab, similar results were observed, as response rate did not significantly differ from doses ranging from 0.1 to 10 mg/kg Q2W [34.Robert C. Ribas A. Wolchok J.D. et al.Anti-programmed-death-receptor-1 treatment with pembrolizumab in ipilimumab-refractory advanced melanoma: a randomised dose-comparison cohort of a phase 1 trial.Lancet. 2014; 384: 1109-1117Abstract Full Text Full Text PDF PubMed Scopus (1432) Google Scholar, 35.Ribas A. Hamid O. Daud A. et al.Association of pembrolizumab with tumor response and survival among patients with advanced melanoma.JAMA. 2016; 315: 1600-1609Crossref PubMed Scopus (726) Google Scholar, 36.Freshwater T. Kondic A. Ahamadi M. et al.Evaluation of dosing strategy for pembrolizumab for oncology indications.J Immunother Cancer. 2017; 5: 43Crossref PubMed Scopus (156) Google Scholar, 37.Topalian S.L. Sznol M. McDermott D.F. et al.Survival, durable tumor remission, and long-term safety in patients with advanced melanoma receiving nivolumab.J Clin Oncol. 2014; 32: 1020-1030Crossref PubMed Scopus (1771) Google Scholar]. These early-phase studies of ICIs might be better named as ‘dose-ranging’ studies instead of ‘dose-finding’ studies. Altogether, exploring less intense and dense schedule options than the one used in the atezolizumab–cobimetinib trial might have given some more insight about the combination. (3) Classical versus modern early-clinical trial designs for innovative oncology drugs. Mouse models need improvement with IO drugs, which is an unmet need in this setting. As a consequence, the initial patients in the early-phase studies with these agents definitely need to be highly informative in terms of pharmacological (PK/PD) and clinical (activity and toxicity) information. The escalation part of the cobimetinib–atezolizumab trial included few patients, not particularly information rich, and after treating only 6 of them at the RD of 60 mg in a 3 + 3 design, the study was expanded to treat 136 additional patients at that dose and schedule. Currently, the trend with ICI development is to use other dose-estimation designs, such as Bayesian seamless models, that incorporate unique features of immunotherapy and all possible information gained from the initial patients by considering three outcomes at any time of each patient’s treatment: immune response, toxicity and efficacy. Accumulation of data allows for adaptive dosing of patients based on what the updated model estimates [29.Liu S. Guo B. Yuan Y. A Bayesian phase I/II trial design for immunotherapy.J Am Stat Assoc. 2018; 113: 1016-1027Crossref PubMed Scopus (21) Google Scholar]. In those models, the escalation defines a range of safe doses that are tested in a second stage that also includes response rates and survival [38.Guo B. Li D. Yuan Y. SPIRIT: a seamless phase I/II randomized design for immunotherapy trials.Pharmaceutical Statistics. 2018; 17: 527-540PubMed Google Scholar]. Bayesian designs allow efficient real-time ‘go/no-go’ interim decision-making in the presence of late-onset responses or toxicities by using all available data and maximizes statistical power for detecting effective treatments [39.Lin R, Coleman RL, Yuan Y. TOP: time-to-event Bayesian optimal phase II trial design for cancer immunotherapy. J Natl Cancer Inst, djz049, https://doi.org/10.1093/jnci/djz049.Google Scholar]. Perhaps a Bayesian model would have helped in the cobimetinib–atezolizumab trial to determine different dosing and schedules with less toxicity. We now consider this type of approach to be mandatory for this type of studies. In conclusion, current clinical drug development with IO agents needs to redefine classical clinical trial parameters, including the concept of DLT or MTD, and to move to adaptive trial designs in order that patients get the most benefit from new therapies. In this exciting era, with a plethora of new drugs coming to clinic, we have to adjust and upgrade our old ways of performing clinical studies to the unique attributes of these novel agents.