The use of predictive biomarkers in cancer medicine may allow oncologists to target interventions to populations with greater response rates, affect sizes, and benefit-risk ratios.
Suzanne Topalian, MD
The use of predictive biomarkers in cancer medicine may allow oncologists to target interventions to populations with greater response rates, affect sizes, and benefit-risk ratios. Incorporating appropriate biomarkers into the design of clinical studies may also allow the use of smaller, less costly trials during drug development that focus on the patient subsets that are most likely to benefit. Ongoing clinical trials are testing drugs that target the immune checkpoints, programmed cell death 1 (PD-1) and its ligand, PD-L1 (B7H1, CD274). Several of those trials are also evaluating the potential of PD-L1 expression as a predictive biomarker of drug response.
Programmed cell death 1 is a membrane receptor expressed on T cells and other types of immune cells; PD-L1 is normally expressed on antigen-presenting cells (APCs). The interaction between PD-1 and PD-L1 inhibits T-cell activation and reduces collateral damage to healthy tissue during inflammation.
Various types of epithelial tumors have been found to exhibit aberrant expression of PD-L1, including renal cell carcinoma (RCC), melanoma, and lung cancer, among others. These PD-L1-positive tumors are able to escape immune surveillance and are associated with advanced stage, worse prognosis, and rapid metastatic progression.1Drugs that disrupt PD-1/PD-L1 signaling aim to restore normal antitumor immune activity, and PD-L1 expression is seen as a potential biomarker of whether or not a given tumor will respond to such an intervention.
In 2012, Suzanne L. Topalian, MD, director, melanoma program, Johns Hopkins University School of Medicine, Baltimore, MD, and her colleagues published the results of a phase I study (NCT00730639) of the PD-1 antibody, nivolumab (BMS-936558).2The study population included patients with melanoma, non-small cell lung cancer (NSCLC), castration-resistant prostate cancer, RCC, and colorectal cancer (CRC). In a preliminary finding, 9 of 25 patients with PD-L1-positive tumor biopsies (36%) had an objective response (OR) to nivolumab. Conversely, of the 17 patients with tumors negative for PD-L1, none responded to the treatment.
More recently, at the 2013 World Conference on Lung Cancer, Scott J. Antonia, MD, PhD, H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, reported preliminary data from another phase I clinical trial (NCT00730639) of nivolumab.3Antonia and his colleagues examined the relationship between the expression of PD-L1 and the response to nivolumab in patients with NSCLC. They documented a 16.1% response rate (5/31) in patients with PD-L1-positive tumors, and a 12.5% rate in patients with PD-L1-negative tumors (4/32).Another presenter at the meeting, Edward B. Garon, the David Geffen School of Medicine at UCLA, discussed data from a study of the safety and efficacy of the PD-1 antibody, MK-3475, in previously treated patients with NSCLC.5PD-L1 expression was assessed prior to treatment and all of the patients with a confirmed OR had PD-L1-positive tumors.
Roy S. Herbst, MD, PhD
Clinical studies examining the correlation of PD-L1 expression and response to antibodies against PD-L1 have produced results similar to those found with PD-1 inhibitors.Roy S. Herbst, MD, PhD, Yale University School of Medicine, New Haven, CT, presented preliminary results from a phase I clinical trial (NCT01375842) of the PD-L1 antibody, MPDL3280A.4The patients enrolled had locally advanced or metastatic solid tumors, and an OR was observed in 25 of 122 (21%) patients, including those with NSCLC, RCC, melanoma, CRC, and gastric cancer. PD-L1 expression was determined using archival tumor samples; the study found a positive correlation between PD-L1 status and OR, with 13 of 33 (39%) of patients with PD-L1-positive tumors responding. However 8 of 61 (13%) patients with PD-L1-negative tumors also responded.
“There is now preliminary data from multiple studies showing an association between high-level PD-L1 staining and response to inhibitors of the PD-1/PD-L1 axis,” Garon explained. “However, the correlation is not perfect and the number of specimens assessed to date remains small,” he said, adding, “The degree to which PD-L1 expression predicts response is still not entirely clear.”Although these and other studies have produced results suggesting that PD-L1 may be useful as a predictive biomarker for drugs targeting the PD-1/PD-L1 axis, questions still remain about its utility. Although PD-L1 expression is significantly correlated with a higher likelihood of response to PD-1/PD-L1 interventions across several studies, a number of responses have been observed in patients with PD-L1-negative tumors. This raises the concern that using PD-L1 as a biomarker to select patients for treatment with immunotherapeutic agents could result in the exclusion of some who would benefit. Because methods of determining PD-L1 expression differed across trials, a direct comparison of the specific protocols and antibodies used, as well as testing a greater number of specimens, would help to determine the true relationship between the marker and responder status.
Questions also remain about the dynamics of PD-L1 expression, according to Suresh S. Ramalingam, MD, director of the division of medical oncology, Emory University School of Medicine, Atlanta, GA. “It is not clear if PD-L1 expression is altered by the use of prior anticancer therapies and exposure to radiotherapy. The differences in PD-L1 expression between primary tumors and metastatic site, and the change in expression status as a function of time, are important questions that are now being studied,” Ramalingam said.