Easily accessible information from hematoxylin and eosin (H&E)-stained tumor tissue on microscopy slides can predict which patients with metastatic clear cell renal cell carcinoma (mccRCC) are most likely to respond to immunotherapy, with the prediction further refined by adding PBRM1 mutation status, research suggests.
The findings, from a retrospective study led by investigators at the Johns Hopkins Kimmel Cancer Center and its Bloomberg–Kimmel Institute for Cancer Immunotherapy, could be used to help preselect patients for the most appropriate therapy.
The authors write in Cell Reports Medicine that the immunotherapeutic landscape for mccRCC is rapidly evolving, but the increasing number of treatment options means that there is now “an unmet need for biomarkers that can help match a patient to the regimen most likely to result in clinical benefit.”
To address this, they investigated whether information from H&E-stained slides from tumor biopsies could predict outcomes. The staining modality has been in clinical use for well over a century and is a part of routine surgical pathology workflows worldwide.
Yet, it has largely been overlooked as a possible source of biomarker information for immune checkpoint blockade even though features easily observed with H&E staining include the presence of immune cell populations such as tumor-infiltrating lymphocytes (TILs), macrophages, plasma cells, and neutrophils. Necrosis is also readily identified on H&E slides.
The researchers used light microscopy to examine H&E slides from 136 pretreatment metastatic tumor samples from patients with mccRCC who took part in one of three clinical trials—CheckMate 009, CheckMate 025, and CheckMate 214.
Specifically, there were 63 patients given the immune checkpoint inhibitor nivolumab monotherapy as a first- or later-line treatment in CheckMate 009, 58 patients receiving later-line nivolumab (n=19) or everolimus (n=39) in CheckMate 025; and 15 treatment-naive patients receiving nivolumab plus ipilimumab in CheckMate 214.
Overall, 70 patients had evidence of mononuclear immune infiltrate, including tumor infiltrating lymphocytes, macrophages, plasma cells, and other associated immune cells (termed TILplus), that also involved the tumor and were given a TILplus score of 1. The remaining 66 patients had mononuclear immune infiltrate at the interface with the tumor and were given a TILplus score of 0.
The investigators report that, in a meta analysis of all three cohorts, individuals with a TILplus score of 1 had a significant 58% lower risk for death during 5 years of follow-up than those with a TILplus score of 0. Median overall survival (OS) was 47.9 months for the patients with TILs and 16.0 months for those without, while median progression-free survival was a respective 7.5 and 2.7 months.
The associations were similar in each of the three cohorts among the patients who received immune checkpoint inhibition, but was there was notably no link between the presence of a pre-existing immune infiltrate and either OS or objective response among the patients who received everolimus chemotherapy. This suggests that the findings are specific to patients treated with immune checkpoint inhibitors, say the researchers.
Next, they hypothesized that the presence of necrosis could modify the beneficial effects of TILs, and found this to be true. Specifically, patients whose tumors had substantial necrosis (greater than 10% surface area) had lower OS than their counterparts with the same TILplus score but whose tumors lacked necrosis. For example, in the CheckMate 009 cohort, median OS was not reached among participants with a TILplus score of 1 and no necrosis but was approximately 34 months among those with necrosis.
“This is important, because traditionally, areas of necrosis are often excluded from biomarker studies because it can’t be used for genomic or transcriptomic studies since the tissue is dead,” said lead study author Julie Stein Deutsch, M.D., a clinical fellow in dermatopathology at the Johns Hopkins University School of Medicine, in a press statement.
“We show there is important information in that necrotic area that’s conferring some sort of negative disadvantage to patients. It’s important not to overlook these areas when you’re investigating biomarkers that predict how patients are going to do.”
Finally, investigators looked at loss of function mutations in the PBRM1 gene and how they impact the other factors. These mutations have previously been associated with immune checkpoint inhibition response and correlated with OS in the current analysis. There was no association, however, with TILplus score.
Despite this, a statistical analysis found that the combination of of TILplus score, necrosis, and PBRM1 mutation status stratified patients into three groups. Patients who had all three positive factors—a TILplus score of 1, less than 10% necrosis, and a PBRM1 mutation—had the best OS rates at 5 years. Median OS was not reached among the patients with all three features, 33.3 months for the with two features, 23.4 months for patients with only one of the three features.
Furthermore, the risk of death during follow-up was a significant 82% lower among patients with all three features present relative to those with only one feature present, which the team says shows the benefit of the combinatorial biomarker over H&E scores or PBRM1 mutation status alone.
When investigators performed a literature search, they were unable to find other studies that used H&E features as part of tumor characterization using multimodality approaches. “This demonstrates the underutilization of these insights in biomarker discovery for immunotherapies,” Deutsch said.
She added: “The ability to use information from an H&E slide, pretreatment, to predict overall survival of patients receiving this therapy is extremely powerful and is something that can be used in resource-poor settings as well.”
Next steps for the researchers will include validating the findings in larger groups of patients and potentially prospectively in a clinical trial.