Summary
A bioRxiv preprint describes CIDER, a whole-genome sequencing score based on reconstructed T-cell receptor sequences. In patient cohorts, higher scores were associated with better outcomes among people treated with immune checkpoint inhibitors.
Researchers have developed CIDER, a score that estimates T-cell infiltration in tumours from whole-genome sequencing (WGS). In a bioRxiv preprint, higher scores were associated with better response and overall survival among patients treated with immune checkpoint inhibitors (ICIs). The analysis also found that these associations were weaker or absent in a comparison group that did not receive ICIs.
The findings suggest that DNA sequencing data could add an immune-related measure to genomic profiling for cancer. They come from analyses of patient cohorts, rather than a prospective trial in which treatment decisions were guided by the score.
Reconstructing a T-cell signal from tumour DNA
T cells are immune cells involved in recognising and attacking abnormal cells. Each T cell carries a receptor whose variable regions help determine what it can recognise. The study uses reconstructed sequences from one variable region, the complementarity-determining region 3 (CDR3), to estimate how many distinct T-cell receptor sequences are present in a tumour sample.
CIDER’s score is based on the number of unique reconstructed CDR3 sequences obtained from WGS data. The researchers characterised it in 8,275 metastatic tumours from the Hartwig Medical Foundation cohort. For 3,572 patients with both WGS and RNA sequencing, the score correlated with an RNA-based T-cell infiltration signature (Pearson’s r = 0.80). It also correlated with ImmuneLENS, a separate DNA-based lymphocyte-infiltration measure (r = 0.92).
The score had little correlation with tumour mutational burden (TMB), a measure of mutations in a tumour’s DNA (r = 0.035). This suggests CIDER captures a different feature from TMB. The authors also report that combining CIDER with TMB improved patient stratification, although the abstract does not give a numerical estimate of that improvement.
Associations in patients treated with checkpoint inhibitors
The researchers assessed clinical outcomes in an independent cohort of 647 patients treated with ICIs across six cancer types. The median CIDER score was 20 among responders and 11 among non-responders (P = 5.7 × 10⁻¹¹). Higher scores were also associated with better overall survival (hazard ratio 0.56, P < 4.3 × 10⁻⁸). The score remained associated with response and survival after covariate adjustment, according to the preprint.
To help distinguish a treatment-related signal from a general association with cancer outcomes, the team also examined 2,432 patients who had not received ICIs. The CIDER associations were attenuated or absent in that group. This pattern supports the authors’ interpretation that the score may be predictive of ICI outcomes, rather than simply marking patients with better outcomes regardless of treatment.
CIDER could offer an additional way to assess the tumour immune environment using sequencing data that also supports other genomic analyses. The results are cohort-based associations reported in a preprint; prospective evaluation would be needed before using the score to guide treatment choices.