Background: Intratumoral immunotherapies are gaining interest in oncology, particularly in melanoma. These therapies, however, have faced some issues. For instance, standard response criteria do not accurately describe tumor burden, and responses may differ for injected/non injected lesions. Besides, target lesions may become non evaluable. Biomarkers provide interesting information for these therapies. In addition, some radiomic signatures have been associated with CD-8 infiltration. BO-112 is a double stranded synthetic RNA formulated with polyethyleneimine (PEI) that mimics a viral infection, mobilizing the immune system and changing tumor microenvironment. Clinical data are available from a first-in-human study, which showed ORR of 11% and DCR of 46% in patients who had developed progressive disease on immunotherapy. In patients with melanoma, this ORR was 20%. A phase 2 clinical study of BO-112 with pembrolizumab in patients with liver metastases from digestive tumors is ongoing. Both studies brought up data regarding how some biomarkers are increased after a single dose of BO-112 and correlated with responses. In this phase II study in patients with pretreated melanoma (NCT04570332), we will prospectively assess CD-8 and PD-L1 by immunohistochemistry, which will be compared with multi-parametric radiologic findings and correlated with clinical benefit. In addition, retrospective DNA sequencing will be performed. This kind of exploratory analysis in intratumoral immunotherapies might be key to identify predictive and prognostic factors. Methods: Phase 2, single arm, open label study of BO-112 with pembrolizumab in patients with advanced melanoma. BO-112 is administered once weekly (QW) in 1 to 8 tumor lesions, total dose 1-2 mg (depending on the number of injected lesions), for the first 7 weeks and then once every three weeks (Q3W); pembrolizumab 200 mg will be administered Q3W. Key eligibility criteria: advanced cutaneous or mucosal melanoma; patients must have progressed on or after treatment with an antiPD-1/L1 mAb; at least one measurable lesion amenable for weekly IT injection. Primary efficacy variable is ORR by RECIST 1.1, assessed by independent central radiologist (QUIBIM Precision platform). A 1-sided alpha of 4.19% and power of 81.8% are used. If less than 8 patients out of 40 have ORR, the study will not meet the statistical bar. Secondary endpoints include clinical activity by RECIST1.1 and iRECIST, overall survival, safety and PKs. Exploratory objectives include itRECIST and evaluation of CD-8 and PD-L1 expression by immunohistochemistry (Pangaea laboratory), which will be correlated with radiomic signatures (first order and second order) from standard-of-care computed tomography (CT) images. Enrollment is open and 1 of planned 40 patients has been enrolled. Nineteen sites are planned to participate. Clinical trial information: NCT04570332