Immunotherapy should be given as a tailored, not blanket, treatment approach, especially for patients with lung cancer, according to Giorgio Scagliotti, MD, PhD.
Giorgio V. Scagliotti, MD
The precision medicine movement continues to spread across malignancies with treatments targeted against driver mutations; however, especially in lung cancer, immunotherapy should also be given as a tailored, not blanket, treatment approach.
“I do not believe that any type of immunotherapeutic approach should be considered as a one-size-fits-all approach, but it should be included in a more tailored approach [as a] precision medicine strategy,” said Giorgio Vittorio Scagliotti, MD, PhD, chief of the Medical Oncology Division at the S. Luigi Hospital, Orbassano (Torino), and head of the Department of Oncology at University of Torino, Italy.
Biomarkers, including PD-L1 expression and tumor mutational burden (TMB), continue to be used as a gauge for patient selection for checkpoint inhibition in patients with non—small cell lung cancer (NSCLC). In the phase III KEYNOTE-042 trial,1 for example, patients with NSCLC who were treated with frontline pembrolizumab (Keytruda) lived 4 to 8 months longer than those who received standard chemotherapy, depending on their level of PD-L1 expression.
Data showed that the overall survival (OS) was associated with greater levels of PD-L1 expression. With a tumor proportion score (TPS) ≥50%, the OS was 20 months with pembrolizumab versus 12.2 months with chemotherapy (HR, 0.69; 95% CI, 0.56-0.85; P = .0003). For TPS ≥1%, the median OS was 16.7 versus 12.1 months (HR, 0.81; 95% CI, 0.71-0.93; P = .0018) for pembrolizumab versus chemotherapy, respectively. However, an exploratory analysis showed that in all patients with PD-L1 TPS of 1% to 49%, the median OS was 13.4 months with pembrolizumab versus 12.1 months for chemotherapy (HR, 0.92; 0.77-1.11).
TMB was a focus in data of the CheckMate-227 study, which was presented at the 2018 AACR and ASCO Annual Meetings and published in the New England Journal of Medicine.2,3 Results of the phase III trial, which evaluated the frontline combination of nivolumab (Opdivo) and ipilimumab (Yervoy) in patients with advanced NSCLC, showed a 1-year progression-free survival (PFS) rate of 43% in those with high TMB ≥10 mutations/megabase regardless of PD-L1 status with the combination. This was compared with a PFS rate of 13% in those given chemotherapy. In June 2018, the FDA accepted a supplemental biologics license application for this combination in this setting.
Beyond PD-L1 and TMB, other markers are in the early stages of investigation, such as neoantigens, explained Scagliotti. During the 19th Annual International Lung Cancer Congress, he presented results focused on the available BRAF-mutant treatments as well as the crossroads between immuno-oncology and precision medicine in lung cancer. In an interview with OncLive during the meeting, Scagliotti focused on the latter, highlighting the progress and challenges with PD-L1 and TMB with a forward look to potential biomarkers on the horizon.
OncLive®: Where does the field currently stand in relation to immunotherapy and precision medicine?
Scagliotti: Immunotherapy is one of the emerging pillars in the systemic treatment of NSCLC in general, and also in other types of classic malignancies. This type of approach should be, at least in my view, appropriately considered in the context of precision medicine. In other words, I do not believe that any type of immunotherapeutic approach should be considered as a one-size-fits-all approach, but it should be included in a more tailored approach [as a] precision medicine strategy.
We started looking to one specific biomarker that you can easily assess in clinical practice—that is PD-L1 expression. That is just one of the potential biomarkers. Based on the data that we currently have, it can be considered as an enrichment factor because of the data we got from several trials. It is true that the objective response rates (ORRs) and duration of response are much longer for those tumors that are PD-L1—positive.
One of the issues we have in the PD-L1 assessment is the tumor heterogeneity. [Also], in terms of expression, we don’t know if the time of [assessing] the markers make any difference. [This is affected when] testing PD-L1 expression on tissues in the field of second-line treatment, but even in the context of frontline treatment. We are not assessing the marker at the same time before starting treatment.
Obviously, there are controversial data about this marker. Sometimes [we are] getting a good correlation between the PD-L1 expression and the outcomes, but the next question is to look at the different values that the investigators tested across several trials. There is a general consensus [stemming from] clinical data that led to the registration of pembrolizumab: if [a patient has] more than 50% PD-L1 expression on their tumor cells, [they will have] much more benefit from immunotherapy when compared with chemotherapy.
More recently, it was shown that the tumor mutational load was also correlated with response and activity to immunotherapy. That is a new field; there are different platforms assessing the TMB but there are also different types of cut-offs. Consequently, the results of one study cannot be compared with another study.
In principle, when we are talking about TMB, you need to ask yourself, “When to assess the TMB, how, and where?” “Where” is because it can be in the tissue, but it can also be in the blood, and the cut-offs are totally different when you are testing tumor mutational load in the tumor or in the blood. Also, the specificity and sensitivity could change. For when—it can be at baseline, but you need to also [detect] the mutational load assessment later on, during treatment, because, at least in the melanoma field, there is a clear correlation between the efficacy of immunotherapy and the contraction of TMB over time. In other words, if [a patient has] a reduction in TMB during treatment, you will see much more ORRs and longer durations of response.
In addition to that, when you are talking about immunotherapy, you are also talking about the tumor microenvironment. This is telling you that we are not focusing only on the cancer, but focusing on the interface between the immune system and the tumor. If you are considering this interaction, it is not the only inhibition of the checkpoint inhibitors, it is much more.
The quality and the type of antigens are becoming relevant. That is the reason why a lot of investigators are looking at the amount of neoantigens. There is some correlation between the mutational load and the number or proportion of neoantigens, and that is what the current research is trying to investigate. However, there is no clear or direct evidence of that. However, neoantigens need to be investigated carefully; [these] are the key players in inducing the immune response, which are a different number from the total number of antigens in the context of a tumor.
However, the antigens need to be presented in the context of the human leukocyte antigen (HLA) and the human leucocyte antigen-C (HLC) systems; [these are] quite sophisticated. The way in which the immune system is recognizing the antigen is quite different; this is because if you have a situation of what we call homozygosity or heterozygosity. The condition of homozygosity is leading to a nonperformance in recognizing a [prognostic] number of antigens.
When you have a condition with zygosity, the system is able to recognize a larger number of antigens. When you are combining this type of information, the homozygosity or heterozygosity, with the HLA receptor and TMB—at least in melanoma—you can see a clear difference that is favoring the mutational load and the HLA heterozygosity.
A few years from now we are not going to focus on one single biomarker; it is quite likely that to identify the best subgroup of patients who will have benefit [in] any type of solid tumor, it’s quite likely that it will be [from] a combination of biomarkers.
Several retrospective studies tried to investigate the role of TMB in the context of existing clinical trials. For instance, in a trial of the combination of ipilimumab and nivolumab, the TMB over  mutations/megabase, [it was found] that with this threshold value, the efficacy of the combination was clearly superior to the control arm.
Obviously, on the other side, below this threshold value, there was no efficacy of the immunotherapy cohort versus standard of care. In other retrospective studies, looking at several atezolizumab (Tecentriq) trials in different solid tumors, there was clear evidence in the tumor but also in the blood that about a certain threshold value [demonstrated] efficacy of atezolizumab was superior in the experimental arm, while the same [outcomes] were not seen in the control arm.
In terms of molecular profiling, what are your thoughts on platforms looking to inform treatment decisions for patients with NSCLC—either targeted treatments or immunotherapy?
The systematic assessment through next-generation sequencing (NGS), through different type of available platforms, are more than a clinical exercise. Dealing with lung cancer, there are a few oncogenic drivers that are readily available in the clinical practice. Consequently, you can also consider a limited number of genes to be tested, and this is exactly what the College of American Pathologists (CAP) guidelines are currently recommending. The  recommendations are to test for EGFR mutations, including T790M, ALK translocations, ROS1 fusions, and BRAF mutations. RET fusions are still a matter of research and is an extremely rare type of mutation. Obviously, the CAP guidelines are considering NGS. I’m not saying one system is better than another, but it’s much easier to get the test reimbursed in the United States; it’s less possible in the European setting. Consequently, at least from my European point of view, NGS is still a research tool.
Beyond neoantigens, what other markers are being evaluated that are poised to have an important role?
At the present stage, I believe that the neoantigen story is a pretty early story. Consequently, if we need to move the field in a pretty solid way, I should say that we need to avoid jumping ahead too quickly. If there is a clear research hypothesis, this needs to be tested in the context of clinical trials.
What clinical trials are examples of how PD-L1 or TMB have been effective or even ineffective?