Table 6.

Potential predictive biomarkers for immunotherapy with immune CPIs.

	Rationale	Open problems
PD-L1	Hypothesis that high levels of PD-L1 in tumor and/or immunological cells in tumor microenvironment may predict clinical response to CPIs with good evidence of correlation in NSCLC, melanoma, renal cell carcinoma	• Discordant results across different trials • Different IHC platforms, detection antibodies, cell types evaluated, and scores for defining positivity • Dynamic marker (variable over time and space) • No evaluation of microenvironment • Low predictive negative value
TMB (tumor mutational burden)	Tumors with a higher TMB seem more likely to express neoantigens, inducing a more robust response if treated with CPIs	• Discordant results across different trials • Analysis considered expensive, time-consuming and misleading if performed with an unsuitable NGS panel • No evaluation of microenvironment • Low predictive negative value
Immune cell gene expression profiling	It is considered a comprehensive biomarker that can enable to assess tumor microenvironment and its inflammatory status to distinguish hot tumors from cold ones	• Lack of standardized commercially available gene panel • Expensive • Uncertain negative predictive value of the various gene panel
Granzyme B	It acts as a mediator of target cell apoptosis induced by immune effectors and might be used as a surrogate marker of CD8+ cells activation	• No standardized method of evaluation (levels of soluble marker rather than double staining for CD8+ cells and granzyme B) • Lack of solid data (tried to correlate with response only in a few trials)
DNA damage response (DDR) genes alterations	Association with better response to neoadjuvant chemotherapy and higher TMB and copy number alteration. Plausible a good relation also to CPIs response	• Lack of solid data
Retinoblastoma 1 (RB1) gene alterations	In addition to being fundamental in cell cycle regulation, it has been discovered to be involved in immune function	• Lack of solid data
Epithelial-mesenchymal transition (EMT) markers	In some studies higher EMT-related gene expression was linked to a major benefit from immune checkpoint blockade	• Lack of solid data and contradictory correlations with CPIs response in different studies
TGF-β pathway	It acts as a key factor in cancer development and progression. In some studies high levels of expression were related with resistance to CPIs	• Lack of solid data
Molecular subtyping	Heterogeneous tumors may be grouped by molecular features in several subtypes, different for treatment response and prognosis	• Need for a consensus classification • Need for prospectical trial to validate the retrospective findings about correlation between specific subtypes and different therapy responses