Moving Molecular Profiling to Routine Clinical Practice: A Way Forward?

Abstract

Molecular profiling of a patient's tumor to guide targeted treatment selection offers the potential to advance patient care by improving outcomes and minimizing toxicity (by avoiding ineffective treatments). However, current development of molecular profile (MP) panels is often based on applying institution-specific or subjective algorithms to nonrandomized patient cohorts. Consequently, obtaining reliable evidence that molecular profiling is offering clinical benefit and is ready for routine clinical practice is challenging. In particular, we discuss here the problems with interpreting for clinical utility nonrandomized studies that compare outcomes in patients treated based on their MP vs those treated with standard of care, studies that compare the progression-free survival (PFS) seen on a MP-directed treatment to the PFS seen for the same patient on a previous standard treatment (PFS ratio), and multibasket trials that evaluate the response rates of targeted therapies in specific molecularly defined subpopulations (regardless of histology). We also consider some limitations of randomized trial designs. A two-step strategy is proposed in which multiple mutation-agent pairs are tested for activity in one or more multibasket trials in the first step. The results of the first step are then used to identify promising mutation-agent pairs that are combined in a molecular panel that is then tested in the step-two strategy-design randomized clinical trial (the molecular panel-guided treatment for the selected mutations vs standard of care). This two-step strategy should allow rigorous evidence-driven identification of mutation-agent pairs that can be moved into routine clinical practice.

Figure 1.

Progression-free survival data (PFS2 vs PFS1) for patients treated with FOLFIRI followed by FOLFOX6 from the arm of a randomized trial of the sequencing of the two regimens (24). A) All the data. B) Reduced data by omitting patients with PFS1 greater than 1.5. Points above the diagonal lines represent patients with a PFS ratio greater than 1.3.

Figure 2.

Proposed two-step strategy for moving molecular profiling to clinical practice. Step 1: Multi basket trial with some baskets showing no or minimal response rate activity (red Xs) or exceptional activity (green check marks); other mutation-agent baskets require further testing in their own enrichment randomized clinical trial (blue box) or in step 2. Step 2: Patients with mutation-agent pairs identified in step 1 are included in a randomized strategy design using a survival endpoint with a standard-of-care control arm and a molecular profile (MP)-directed treatment experimental arm. RCT = randomized clinical trial.