144P - Analysis of a novel predictive marker of immune checkpoint response in head and neck cancer, calculated from histopathological slides through inferred transcriptomics

Background

Immune checkpoint blockers (ICB) are in the forefront of contemporary clinical oncology and have become an integral part of treatment of many malignancies, including metastatic head and neck cancers (HNC). Nevertheless, tumor response to ICB varies widely, with known predictive biomarkers, such as combined PDL score (CPS) showing limited predictive value. We present results of a blind retrospective analysis of a novel predictive digital pathology biomarker of ICB in HNC.

Methods

We obtained high resolution Hematoxylin and Eosin (H&E) slides from tumor-tissue samples of 26 cases of metastatic HNC patients treated with first-line PD-1 inhibitors, all with CPS>1%. We applied our ENLIGHT-DP pipeline to generate, in a blinded manner, individual response scores to PD-1 inhibition for each slide. ENLIGHT-DP has two main steps: (i) prediction of mRNA expression directly from an H&E slide using DeepPT, our digital-pathology based algorithm; (ii) use of these values as input to ENLIGHT, our transcriptome-based precision oncology platform, which generates a score that predicts response to targeted therapies and ICB (based on a 10-gene signature in this case, composed of the key genetically interacting genes of PD-1). We then unblinded the clinical response (RECIST1.1), and evaluated ENLIGHT-DP’s performance.

Results

ENLIGHT-DP’s score is predictive of response in this cohort, which had an overall response rate of 54% (14/26), with ROC AUC = 0.65. Using a predefined threshold for binary classification of response derived from independent data, 6 of 8 patients that were predicted to respond by ENLIGHT-DP indeed responded (75% PPV, 43% sensitivity, Odds Ratio of 3.75). In comparison, stratification by CPS exhibits no predictive power (PPV of 50% and 58%, respectively, for CPS > 20% and CPS 1-19%, and AUC of 0.46 for CPS as a continuous score).

Conclusions

ENLIGHT-DP demonstrates high predictive power for response to ICB in first-line HNC, outperforming the commonly used CPS marker. Importantly, our approach does not require training on prior treatment outcomes, and can therefore be generalized to drugs for which such data is unavailable or scarce.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

O. Tirosh, L. Gugel, G. Dinstag, Y. Kinar, T. Gottlieb, R. Aharonov: Financial Interests, Personal, Full or part-time Employment: Pangea Biomed. All other authors have declared no conflicts of interest.