ESMO 2022: 동반진단 Poster submission 

<Introduction>

Background

Immune checkpoint inhibitor (ICI) has achieved a great success as a

promising regime for the treatment of patients with many types of solid

malignancies, associated with predictive biomarker of PD-L1 expression.

However, generally a low rate of ICI therapy response remains a critical

hurdle to overcome for expanding its versatile therapeutic efficacy, and

necessitates importance of developing a biomarker better predicting

response outcome after ICI.

Method

Peripheral blood CD8+ T cell compartment from patients with stage IV of nonsmall

cell lung cancer (n=119 (Atezolizumab) + 19 (Pembrolizumab)) before

ICI treatment targeting programmed cell death 1 (PD-1) or its ligand 1 (PDL1)

was analyzed and correlated with patients’ ICI treatment outcome.

Results

Strong correlation between the patients' response outcome after ICI and the

proportion of two distinct subsets of blood CD8+ T cells, namely CD27+

CD28+ CD45RA- CCR7- and CD27+ CD28+ CD45RA+ CCR7- cells was observed.

Using these two cellular parameters combined with machine learning based

probability graph, we found that both initial discovery (n=70) and later

validation cohorts of patients (n=49) showed a power of predicting

approximately ~57.14% of ICI responders (95% CI, 20.5~93.8%; PR based on

the RECISTv1.1 criteria) compared to that of ~17.1% (95% CI, 8.3~26.0%; no

biomarker included) and of ~25.0% (95% CI, -5~55.0%; biomarker based on

tumor PD-L1 expression). Mechanistically, the observed strong correlation

was due to the basal functional fitness and reactive capacity of responding

CD8+ T cells, which was characterized by lower initial levels of perforin and

granzyme B expression in these cells before ICI treatment. As such, the

patients with enhanced proportion of this subset in their bloods showed

greater capacity to enhance ICI-driven upregulation of cytotoxic molecules

and accordingly better ICI response outcomes.

Conclusion

These observations are in line with current notion that a relatively less

differentiated subset of CD8+ T cells would be a major target for ICI and

provide a potential of developing this non-invasive blood-based approach as

an ICI response predictor for patients with cancer.