Hunt for Reliable Biomarkers for Immunotherapy Continues in Melanoma

Douglas B. Johnson, MD, MSCI

Although immune checkpoint inhibitors have revolutionized the treatment of patients with melanoma, identifying reliable biomarkers of response to this approach has proven to be exceedingly challenging, according to Douglas B. Johnson, MD, MSCI.

Obesity has been linked with improved responses to immune checkpoint inhibitors; however, the question regarding the association between body composition measures, such as muscle and fat, and patient outcomes has remained largely unanswered. 

In an effort to learn more, investigators conducted a body composition analysis to quantify skeletal muscle index, skeletal muscle density, skeletal muscle gauge (SMG), and total adipose tissue index (TATI) of each patient at the third lumbar vertebrae. To do this, Slice-o-matic software was used on pretreatment CT scans. These measures were then correlated to response, progression-free survival (PFS), overall survival (OS), and toxicity.

“We did the body composition analysis on a couple of hundred patients treated with anti–PD-1 [plus or minus anti–CTLA-4]. Surprisingly, we did not see any association with response and elevated body mass index [BMI], obesity, or overweight status,” said Johnson. “This was a surprise for us because we assumed that we would find what some of the other groups had found in terms of obesity and better outcomes; that was not the case.”

Results from univariate analyses showed that patients with sarcopenic obesity experienced inferior PFS (HR, 1.4; P =.04). Moreover, high TATI was also linked with inferior PFS, according to multivariable analyses (HR, 1.7; P =.04). Patients who had intermediate TATI and high SMG were reported to have the best outcomes with this approach. In contrast, those with low SMG and high TATI experienced inferior PFS and OS (P =.02 for both).

In an interview with OncLive, Johnson, an assistant professor of Medicine of Hematology/Oncology at Vanderbilt Institute for Infection, Immunology and Inflammation, of Vanderbilt University Medical Center, further discussed his research examining the impact of body composition on outcomes from anti–PD-1 with or without anti–CTLA-4 therapies in melanoma and other biomarkers under exploration.

OncLive: What biomarkers are currently being studied to better inform treatment with immune checkpoint inhibitors in melanoma?

Johnson: This has been a really challenging area. The initial biomarker of interest was PD-L1 expression; there was some consideration that it may be a good way to stratify patients between receiving single-agent anti–PD-L1 therapy or a combination anti–PD-L1 plus anti–CTLA-4 blockade. However, with longer follow-up, the differences between groups have kind of blurred to some extent, so that's [1 biomarker] that has been less widely used in melanoma compared with other cancers.

[Investigators] have also been interested in tumor mutational burden [TMB]; that also seems to correlate with response, although it's a little unclear how [we will] use it. [We don’t know] whether [TMB will] help to stratify patients between treatments or whether it's a correlate prognostic factor. At this point, many [investigators] are thinking about clinical biomarkers as far as the total tumor bulk. [They consider] whether patients are symptomatic, whether they have cutaneous versus noncutaneous primaries, whether they have brain or liver metastases, and whether they have good performance status. We use [factors] to help guide us in terms of whether we should use a PD-L1 monotherapy or a combination.

Do you have any biomarkers to indicate how to choose between single-agent and combination immunotherapy approaches or would you say that there is still an unmet need?

At this point in time, there are substantial unmet needs. We still lack a convincing biomarker to tell us when to use an anti–PD-L1 monotherapy, when to use a combination, and when neither drug is going to work and we will need to consider something completely different, such as BRAF and MEK inhibitors for patients with mutations or next-generation [therapies,] such as injectable antitumor treatments or tumor-infiltrating lymphocyte therapy. At this point, as far as the biomarker space goes, it's not a question of whether there are unmet needs; unmet needs [are all we have] at this point.

Could you discuss your study examining the impact of body composition on outcomes from anti–PD-1 plus or minus anti-CTLA4 treatment in this disease?

Several studies have suggested that patients who are obese or overweight actually have better outcomes when treated with immune checkpoint inhibitors. Obviously, this is a little bit paradoxical [because] we normally think of obesity as more of a negative prognostic factor; that has certainly been shown in some cancers. However, with several studies, 1 [of which was] led by the University of Texas MD Anderson Cancer Center, have suggested that patients treated with anti–PD-1 monotherapy have better outcomes when they are obese.

We decided to question, “If obesity is associated with better outcomes, can we [find out why]?” We looked at cross-sectional CT images to see if we could tease out the amount of muscle, the amount of fat, the quality of the muscle, and the muscle density to see whether that might help us to predict which patients responded more [to this treatment].

What did your analysis reveal? Were there any features associated with positive outcomes?

We didn't see any association for BMI specifically, but we did still want to dive in and see whether there were any particular associations with body composition. We observed a few things that were somewhat interesting.

We saw that patients who had sarcopenic obesity, those who had very little muscle, but a lot of fat, seemed to do the worst, although perhaps that's not the most surprising finding. [These were] patients who were not as fit or in as good a condition who had muscle wasting in combination with obesity. We did see that the group of patients who seemed to do the best had more of an intermediate BMI phenotype and higher muscle mass; that might be the same group that some of the other [study investigators] picked out as doing particularly well, those with intermediate or high fat, but also high muscle mass. However, I would emphasize that most of these associations were relatively modest. [We most likely won't be] checking for this biomarker in every single patient to help determine treatment.

Another interesting observation was that there did seem to be, as other groups have [observed], a gender interaction. Previously, the MD Anderson [study] had shown that the obesity association was really only present in men. In women, there an association between obesity and bad response [had not been observed]. In general, the women had slightly better outcomes across the board. However, men with obesity had comparable outcomes to women, whereas men with lower BMI had the worst outcomes. 

We found something a little bit different; women who had higher fat did worse, whereas there was no association with men. There seems to be some sort of gender interplay there, but it [might be] a little bit different than [what has been reported in] some of the other papers.

What should be taken away from this work?

We were going in hoping that, between BMI or some of the other body composition markers examined, we would be able to find something to hang our hat on and we could check in patients before starting treatment to help [determine] whether they should receive a single-agent anti–PD-1 or at least if they're going to have a high response rate. We really didn't find that. While we did see some interesting associations, from a clinical utility standpoint, I don't believe that we've really found anything that is going to change practice.

The biology is interesting. Although we did not see [what] other groups [have] in the context of obesity, and I don't have a great explanation for why that is. We had a good representative population of patients who were very similar to those who were treated in other clinical trials. There could be, perhaps, region-specific issues. All our patients were from the southeastern United States and do have a slightly higher incidence of obesity and overweight status compared with [those in] other parts of the country. That could [suggest] some kind of lifestyle interaction factor. 

Beyond your research, are other biomarker research efforts being made that you wanted to highlight?

Several other studies are looking at BMI and body composition and they have actually been mixed, as well. As far as other biomarker studies, a number of [efforts have examined] whole exome type sequencing [and they] have been a little bit disappointing in terms of finding the next specific mutations that correlate with response. However, [these efforts] have reinforced the TMB story, suggesting that it does correlate with response globally. 

We still have a lot of work to do. One biomarker we're looking at is MHC class II expression on tumor cells. We have published some papers on that; it seems to be a promising biomarker that can help us correlate with response. We don't have any data yet as far as whether it correlates with a combination or single agent. It's same challenge [that we are experiencing with] the other biomarkers. 

Where do you feel future research efforts should be focused?

That's an excellent question and one that we have been grappling with for some time. It doesn't seem like the biomarkers we have [examined] so far [have been] a home run. Plenty of biomarkers correlate with response and are somewhat helpful from that standpoint, but many of them seem to correlate with response in both single-agent and combination therapies. As such, that doesn't really help you in terms of picking which approach [to use]. If there is some way to tease that out, I feel as though that would be a really important next step. As to what that biomarker may be, [I'm not sure]. Certainly, that's the key unmet need.

As more effective second- and third-line therapies come down the pike, [we may] figure out which patients won't benefit from immune checkpoint inhibitors. We also may be [able to determine] which patients should receive such therapies up front. It may be that some of these biomarkers, when combined with each other or combined with more clinical factors, do have more [predictive utility].

Is there anything else that you would like to add?

In general, I believe that it's an exciting time in melanoma treatment. I think immune checkpoint inhibitors have revolutionized the way we treat patients. There is definitely a new generation of therapies coming out, but we still have a long way to go in terms of personalizing therapies for patients. Hopefully, that will be the next wave of research to come out within the next few years.

Reference

1. Young AC, Quach HT, Song H, et al. Impact of body composition on outcomes from anti-PD1 +/- anti-CTLA-4 treatment in melanoma. J Immunother Cancer. 2020;8(2):e000821. doi:10.1136/jitc-2020-000821