Dr. Steven M. Sorscher
Journal: JCO Precision Oncology
Metastatic colorectal cancer (mCRC) remains the second most common cause of cancer death in the United States, and therapeutic options are limited. Recently, the checkpoint inhibitor pembrolizumab was given the U.S. Food and Drug Administration (FDA) breakthrough therapy designation for the treatment of patients with mCRC whose tumors demonstrate deficient mismatch repair gene (dMMR) expression (as evidenced by microsatellite instability–high [MSI-H]). The designation was based on a phase II study showing that in patients with dMMR, an objective response rate of 40% was seen, whereas in patients with proficient mismatch repair gene mCRCs, the response rate was 0%. To our knowledge, this is the first case of a patient with a proficient mismatch repair gene mCRC whose tumor demonstrated a dramatic response to a checkpoint inhibitor. Because this patient’s tumor harbored amplification of both the PD-L1 and PD-L2 genes, the observed response was consistent with the presumed mechanism of action of checkpoint inhibitors. Checkpoint inhibitors are thought to activate a cytotoxic immune response that has been inhibited through tumor expression of PD-L1 and PD-L2. Given this result, dMMR in mCRC may not be the only predictor of responsiveness to checkpoint inhibition. As in non–small cell lung cancer (NSCLC), PD-L1 or PD-L2 expression (or perhaps gene amplification) may also be predictors of checkpoint inhibitor efficacy.
Steven M. Sorscher, MD, Wake Forest Baptist Health
Q: How does your case report inform future approaches to treating patients with mCRC?
Dr. Sorscher: Based on our case report, clinicians can now consider next-generation sequencing of tissue or cell-free DNA from patients whose tumors have progressed through standard therapy. Precision oncology relies on identifying markers predictive of potential benefit from a particular (often targeted) therapy. Once identified, clinical trial results are critical to expanding or contracting the definition of a marker. Our case report shows that once a predictive marker of efficacy is identified (e.g., dMMR), it does not rule out the possibility of finding other independent markers of efficacy (PD-L1/PD-1 gene amplification in proficient mismatch repair [pMMR] colorectal cancers [CRCs]) or other markers potentially shown to limit the targeted therapy’s efficacy.
For example, months after the FDA granted accelerated approval for crizotinib for treating patients with ALK-rearranged NSCLC, it also approved crizotinib for treating patients with ROS-1-altered NSCLC after ROS-1 alteration in NSCLC was shown to be a predictive marker of benefit from crizotinib.1 Years after the EGFR-targeting antibodies cetuximab and panitumumab were endorsed for patients with mCRCs, it was shown that these therapies are potentially effective only in patients with CRCs not harboring a RAS mutation.
Recently, the FDA granted accelerated approval for the checkpoint inhibitor pembrolizumab2 for patients with dMMR/MSI-H solid tumors. This approval was a landmark in that it represented the first-ever cancer FDA approval not based on the cancer’s tissue of origin. Yet it remains largely unknown whether dMMR colorectal or other dMMR solid tumors harboring a somatic BRAF mutation have a higher or lower response rate than the roughly 40% response rate cited as the approval’s basis. Further studies of dMMR and BRAF-mutated CRCs will clarify whether pembrolizumab benefits these patients.
The robust response to the checkpoint inhibitor in the patient we described suggests that PD-L1/PD-1 gene amplification might represent another predictive marker of benefit in CRCs. As more tumors are evaluated using next-generation sequencing, hopefully other markers of benefit from checkpoint inhibitors will be proposed and confirmed.
Identifying PD-L1 gene amplification as a potential marker of benefit for patients with mCRC from checkpoint inhibitor therapy underscores the importance of pursuing studies investigating other possible predictive markers of benefit (predictive markers of predictive markers) and suggests the opportunity to study whether some dMMR tumors might harbor other molecular markers imparting a lack of benefit, despite those tumors being characterized as dMMR.