The American Society of Clinical Oncology has hailed immunotherapy, also known as immuno-oncology, as one of the more transformative advances in recent years. But with the inherent novelty of immunotherapies comes unique challenges in their development. Many traditional clinical trial parameters common to chemotherapy – or even the more recent targeted antibodies or kinase inhibitors – must be revisited for their application to examining the safety and effectiveness of immunotherapies.

Determining the ideal dosing protocol for the best patient response can be novel territory for immunotherapy developers. Large pharmaceutical companies can obtain experience through multiple advisors, trials and resources that permit revisiting the drawing board to make such determinations. In contrast, smaller sponsors usually only have budget for one trial.

In part two of this two-part Q&A series (part one here), we spoke with Andrew Zupnick of Novella Clinical to provide insights into dosing and measuring response in immunotherapy clinical trials that are of particular interest to small and mid-sized oncology companies.

Dosing and Measuring Response

Contract Pharma: Has there been discussion among thought leaders on how best to measure disease progression with these new immunotherapy products?

Andrew Zupnick: Yes. With our improved understanding of the process of an immune response, especially in solid tumors, we have variations on standard measurement of response that require serial radiological assessment of tumors to confirm if the tumor is, indeed, growing in size, or if the increase in size is the first step in an immune response within the tumor, known as a “flare effect.” Immune-related response criteria (irRC) were introduced to help address these questions, based on World Health Organization (WHO) criteria, and has evolved to a new standard based on response evaluation criteria in solid tumors (RECIST), known as irRECIST. irRECIST criteria are based on irRC criteria, adapted for unidimensional measurements. These criteria are also hoped to improve reproducibility and inter-rater reliability.  

CP: Have you seen any situations where total tumor burden being assessed volumetrically should be considered?

AZ: Baseline tumor volume has been shown to be a good prognostic factor for survival. It is now routinely being used as a quantitative measure of treatment response. The current standard for treatment response involves both clinical assessment and cross-sectional, two-dimensional imaging. Although CT is effective in identifying and measuring tumor response, defining tumor boundaries can be imprecise. MR imaging improves the accuracy but is less routinely used. Three-dimensional assessment of tumor volume using MR has been shown to have an acceptable rate of inter-rater reliability, and may evolve into a more common standard for the assessment of tumor response.  

CP: What is the trending alternative to the classic 3+3 design?

AZ: Though 3+3 designs are still the most frequently used strategies, particularly for first-in-human (FIH) studies in oncology patients, variations to this model are becoming common which seek to minimize the number of patients treated at sub-optimal doses. Accelerated titration designs combine features from traditional 3+3 and model-based design, where one patient is enrolled per dose, beginning with the lowest dose. The dose is titrated up until a patient experiences a Grade 2 adverse event, where at that point, traditional 3+3 rules apply. Another design is pharmacologically guided dose escalation (PGDE), where plasma drug concentrations are used to direct dose escalation. When the target area under the curve (AUC) is reached, or if dose-limiting toxicities (DLTs) occur, titration reverts to 3+3. A rolling six design is often used in pediatric studies, allowing for the enrollment of two to six patients concurrently, based on the numbers of patients currently enrolled and evaluable for safety assessment. This is permitted, primarily, because more dosing data are available from adult studies.  Continual reassessment models apply Bayesian methods to assess safety and drive dose escalation, where the estimation of the probability of a DLT is updated as each new patient enters the study. Stopping rules are established, such as reaching a maximum enrollment at a dose level or the incidence of DLTs.