{
  "abstract": "Background Immunotherapy, particularly PD-1 inhibitors, have transformed the treatment of advanced cutaneous squamous cell carcinoma (CSCC) not amenable to upfront local therapy. However, responses and long-term outcomes remain highly variable, highlighting the need for more accurate and accessible predictive biomarkers. While high tumor mutational burden (TMB) and UV-induced genetic alterations are suggested as the mechanisms underlying improved PD-1 inhibitor outcomes in CSCC, 1 they are not routinely assessed and are also commonly found in precursor lesions and normal skin, limiting their explanatory value in the pathogenesis of CSCC. Comparing CSCC with head and neck squamous cell carcinoma (HNSCC)—which shares histology but responds poorly to PD-1 inhibitors—may offer new insights into CSCC pathogenesis and immune response.Methods We conducted analysis on CSCC (39 patients) and HNSC (25 patients) cohorts, treated with first line PD-1 inhibitors for metastatic disease, for which H&E slides were available. We used DeepPT, our published algorithm that infers full mRNA profiles directly from H&E slides 2 to calculate activation scores (using GSEA3) for three pathways reflecting immune infiltration and activity. Additionally, we used our ENLIGHT-DP pipeline, which relies on inferred transcriptomics, to calculate individual ENLIGHT-Matching Scores (EMS) for response to PD-1 inhibition.4 We studied: (1) the predictivity of ENLIGHT-DP, comparing it to that of relevant signature scores, and (2) CSCC pathogenesis in comparison to HNSCC.Results We found that the EMS, as well as the activation scores of the three pathways- Interferon-gamma, inflammatory response and signaling in naive CD8 cells - are predictive of response to PD-1 inhibition, with the EMS being the most predictive ( figure 1, ROC AUC of 0.63 and 0.76 for CSCC and HNSCC, respectively). Interestingly, the three immune pathways have significantly higher activation scores in CSCC compared to HNSCC (figure 2), supporting higher immunogenicity of CSCC.Conclusions Our analysis demonstrates ENLIGHT-DP’s ability to predict response to PD-1 inhibition in CSCC and HNSCC directly from H&E slides, providing a practical and easily applicable predictive biomarker. Moreover, we show how DeepPT can be used to study the pathogenesis of tumors directly from H&E slides, shedding light on underlying mechanisms of CSCC immune response. While these signatures can distinguish CSCC from HNSC at the cohort level, they are less predictive of individual response than the EMS. Since only H&E slides are required, this analysis can easily be applied to additional cancers, especially ones where data scarcity is a limitation, like CSCC.References Chang D, Shain AH. The landscape of driver mutations in cutaneous squamous cell carcinoma. npj Genom. Med. 2021;6:61.Hoang DT, Dinstag G, Shulman ED, Hermida LC, Ben-Zvi DS, Elis E, ... Ruppin E. A deep-learning framework to predict cancer treatment response from histopathology images through imputed transcriptomics. Nature Cancer, 2024;5(9):1305-1317.Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, ... Mesirov JP. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proceedings of the National Academy of Sciences, 2005;102(43):15545-15550.Dinstag G, Shulman ED, Elis E, Ben-Zvi DS, Tirosh O, Maimon E, ... Aharonov R. Clinically oriented prediction of patient response to targeted and immunotherapies from the tumor transcriptome. Med, 2023;4(1):15-30.Abstract 60 Figure 1ROC AUC for predicting response to PD1 inhibition. The ‘combined’ cohort takes both CSCC and HNSC togetherAbstract 60 Figure 2The difference between CSCC and HNSC in scores based on 3 hallmark immune pathways and the EMS. p value is for mann-whitney u test",
  "authors": [
    {
      "affiliations": [
        "Pangea Biomed, Tel Aviv, Israel"
      ],
      "name": "Gal Dinstag"
    },
    {
      "affiliations": [
        "Pangea Biomed, Tel Aviv, Israel"
      ],
      "name": "Omer Tirosh"
    },
    {
      "affiliations": [
        "Hebrew University of Jerusalem, Jerusalem, Israel"
      ],
      "name": "Anna Elia"
    },
    {
      "affiliations": [
        "Hadassah-Hebrew University Medical Center, Jerusalem, Israel"
      ],
      "name": "Chayen Bohdana"
    },
    {
      "affiliations": [
        "Pangea Biomed, Tel Aviv, Israel"
      ],
      "name": "Tuvik Beker"
    },
    {
      "affiliations": [
        "Hebrew University-Haddash Medical School, Jerusalem, Israel"
      ],
      "name": "Eli Pikarsky"
    },
    {
      "affiliations": [
        "Hebrew University of Jerusalem, Jerusalem, Israel"
      ],
      "name": "Aron Popovtzer"
    },
    {
      "affiliations": [
        "Pangea Biomed, Tel Aviv, Israel"
      ],
      "name": "Ranit Aharonov"
    },
    {
      "affiliations": [
        "Hadassah-Hebrew University Medical Center, Jerusalem, Israel"
      ],
      "name": "Johnathan Arnon"
    }
  ],
  "title": "60 Predictivity of an H&E-based method for response to PD-1 inhibition and the pathogenesis of CSCC using ENLIGHT-DP",
  "uid": "e10e563f-26df-5021-a0d4-8b9ff90e6b0c"
}