The data are still early, but researchers have identified genetic changes in tumors that could be used to predict whether immunotherapy drugs will be effective in a given patient.
"The main clinical implication of this work is being able to predict more accurately patient response to checkpoint inhibitor therapy," lead researcher Kevin Litchfield, PhD, University College London Cancer Institute, London, United Kingdom, told Medscape Medical News. "Our work is still only in a research setting, but we've been able to show with large patient numbers and with separate validation in three independent cohorts that a multivariable predictive test performs better than current clinical tests."
The team found that the number of genetic mutations present in cancer cells was the best predictor of tumor response to immunotherapy with checkpoint inhibitors. The greater the number of mutations, the more likely the patient would respond to therapy.
In addition, expression of CXCL9 was found to be a critical driver of an effective antitumor immune response. Greater expression of CCND1 was linked to tumor resistance, which suggests that other treatments may be more effective for patients with this mutation.
Lichtfield and colleagues point out that although multiple biomarkers have been associated with immune checkpoint inhibitor response, large-scale studies of response with in-depth whole-exome and transcriptome data have been lacking.
"We hope this evidence will pave the way for prospective trials to validate this; we already have work underway to conduct a prospective trial in Denmark," he said.
The study was published online in Cell.
Response and Resistance
Checkpoint inhibitors have demonstrated unprecedented clinical activity across a range of cancers but are effective only in some patients. A study published last year in JAMA Network Open suggests that although approximately 44% of cancer patients may be eligible to be treated with one of the checkpoint inhibitors currently on the market, only about 13% will respond to them. So far, predicting which patients will respond well to the treatment has been challenging.
Litchfield and colleagues hypothesized that a systematic pan-tumor analysis could help tease out the critical features underlying response to immunotherapy. Drawing from 12 studies, they collated exome/transcriptome data from 1008 patients with seven types of cancer that were treated with checkpoint inhibitors. The cancers represented in the analysis were metastatic urothelial cancer (n = 387), malignant melanoma (n = 353), head and neck cancer (n = 107), non–small cell lung cancer (n = 76), renal cell carcinoma (n = 51), colorectal cancer (n = 20), and breast cancer (n = 14]). The patients were treated with three classes of checkpoint inhibitors: anti-CTLA-4 (n = 155), anti-PD-1 (n = 432), and anti-PD-L1 (n = 421).
Overall, the biomarker with the strongest effect size across all 12 studies was clonal tumor mutational burden (TMB) (ie, the number of nonsynonymous mutations estimated to be present in every cancer cell; odds ratio [OR] for complete response/partial response vs stable disease/progressive disease, 1.74), which was closely followed by total TMB (OR, 1.70). For the category of markers of immune infiltration, CXCL9 expression was the predictor with the strongest effect size (OR = 1.67). There were significant associations for CD8A expression (OR = 1.45), the T-cell inflamed gene expression signature (OR = 1.43; P = 2.5), and CD274 (PD-L1) expression level (OR = 1.32).
The researchers note that "CXCL9 is a critical chemokine that binds CXCR3 on T cells, enhancing recruitment of cytotoxic CD8+ T cells into the tumor and promoting the differentiation of inflammatory T-helper type 1 (Th1) and Th17 CD4 T cells."
Additional analyses showed that checkpoint inhibitor outcomes were also influenced by 9q34 (TRAF2) loss associated with response and CCND1 amplification associated with resistance.
"The clinical value from more accurate predictive tests may enable patients not likely to respond to checkpoint inhibitors to be offered, more quickly, alternative treatment or clinical trial options," said Litchfield. "Checkpoint inhibitor response rates are typically less than 50%, so nonresponse is a major challenge. In addition, patients may be spared additional toxicity risk."
Commenting on the research, Michelle Mitchell, chief executive of Cancer Research UK, said in a statement that the "new research has furthered our understanding around these issues, revealing new drug development tactics and approaches to treatment."
The study was partly funded by Cancer Research UK, the Royal Society, the Wellcome Trust, the Medical Research Council, and Rosetrees Trust. Litchfield has a patent on indel burden and CPI response pending and has received the following outside of the submitted work: speaker fees from Roche tissue diagnostics and research funding from CRUK TDL/Ono/LifeArc alliance. She has a consulting role with Monopteros Therapeutics. Several of the coauthors report relationships with industry, as noted in the article.
Cell. Published online January 27, 2021. Abstract
For more from Medscape Oncology, join us on Twitter and Facebook.
Medscape Medical News © 2021
Send news tips to news@medscape.net.
Cite this: Biomarkers Predict Response to Checkpoint Inhibitors - Medscape - Feb 10, 2021.
Comments