Goldilocks Dosing of TKIs: A Dose that Is Just Right Leads to Optimal Outcomes

In this issue of Clinical Cancer Research, Sternberg and colleagues report the results of a retrospective analysis of patients with resected renal cell carcinoma (RCC) enrolled on a phase III study of adjuvant pazopanib, a VEGF receptor TKI, versus placebo (1). Overall, this trial failed to show a disease-free survival (DFS) benefit (2). Notably, however, the starting dose of pazopanib was reduced from the standard 800 mg daily in an initial cohort of patients to 600 mg daily due to a perceived high discontinuation rate from toxicity. The cohort of patients starting at 800 mg showed a DFS benefit, generating a hypothesis that the overall negative trial result may have been in part due to suboptimal drug exposure after the starting dose was reduced to 600 mg. The authors' analyses complement data established in the metastatic RCC setting that patients able to achieve and tolerate higher drug exposure have better clinical outcomes (3).

Three large phase III trials have been completed examining the DFS benefit of adjuvant VEGF receptor TKIs compared with placebo after resection of high-risk, localized RCC. The first to be reported, ASSURE, failed to show a benefit to 1 year of either sunitinib or sorafenib versus placebo (4). Notably, toxicity and discontinuations in the early part of the trial led to adopting a lower starting dose (37.5 mg instead of 50 mg), and dose reductions to 25 mg were allowed. The second trial to report, S-TRAC, did show a benefit in DFS with 1 year of sunitinib versus placebo [HR = 0.76; 95% confidence interval (CI), 0.59–0.98; P = 0.03; ref. 5]. This trial started all patients at 50 mg and only allowed dose reduction to 37.5 mg. Among the differences between these trials including enrollment of higher risk patients in S-TRAC and restriction to clear cell histology, dosing differences have been hypothesized to explain the discrepant outcomes. The PROTECT trial of adjuvant pazopanib versus placebo reported most recently and in many ways, supports the results of both previous trials (2). As noted, the PROTECT trial initially enrolled patients at the 800-mg dose, which is standard in the metastatic setting. After 403 patients, the starting dose was lowered to 600 mg due to a higher than expected rate of toxicity. The option to dose increase to 800 mg was given, but only 21% of patients who started at 600 mg increased to 800 mg. In the patients who started at 600 mg and the overall population, no DFS advantage was seen, supporting the ASSURE results. In the subset of patients who started at 800 mg, a significant advantage in DFS was observed (HR = 0.69; 95% CI, 0.51–0.94; P = 0.02). These data support the S-TRAC result in that patients with adequate TKI exposure in the adjuvant RCC setting can have a DFS benefit compared with placebo, balanced against toxicity.

Sternberg and colleagues provide steady-state blood trough concentrations on a subset of patients, and the data give important insight into optimal dosing of TKIs. Patients with higher early (week 3 or 5) or late (week 16 or 20) C_trough levels achieved a longer DFS (HR = 0.58; 95% CI, 0.42–0.82; P = 0.002) in a multivariate Cox regression analysis. Both early and late C_trough were significant covariates for DFS, aligning with the notion that sustained optimal drug exposure is required for clinical benefit. Notably, dose intensity did not correlate to DFS. This critical point highlights that numerical dose in mg is not important per se, but rather the drug exposure of each individual patient is a determinant of clincial outcome. Further complicating this is the fact that there is not one threshold of exposure required for clinical benefit in all patients. Thus, titration to an optimal tolerated dose based on clinical parameters (i.e., toxicity and patient quality of life) is required in the adjuvant and metastatic RCC settings to optimize exposure for each individual patient and to maximize outcomes (Fig. 1). Also notable is that the proportion of adverse event (AE)–related drug discontinuations and grade 3/4 AEs (except hypertension) were not correlated to C_trough levels. Hypertension, an on-target effect, is related to exposure but generally is not dose limiting and does not reduce patient quality of life. This again highlights that an individual patient's dose or drug level cannot be used to predict clinically relevant toxicity but rather supports clinical titration to maximize efficacy and minimize toxicity. This concept is also supported by the fact that the cohort of patients started at 600 mg did not substantially differ from the 800-mg cohort in discontinuation rates due to AEs (35% vs. 39%). This clinical titration requires frequent interaction with the patient, especially during the first several weeks of dosing, so that each patient achieves optimal exposure.

In summary, the data from Sternberg and colleagues provide critical insight into the discrepant results of RCC adjuvant clinical trials. Clinical efficacy is linked to optimal drug exposure. Optimal drug exposure requires an adequate starting dose, but more importantly, titration of individual patient dose to a level that can have antitumor effect with acceptable toxicity. Just as Goldilocks required trial and error to find the right chair, porridge, and bed, so does each patient require TKI dose trial and error, including sometimes off-label doses and breaks from dosing, to find a dose that is just right.

No potential conflicts of interest were disclosed.