Identifying the Predictive Factors of Response to PD-1 or PD-L1 Antagonists

The study will include 670 patients with melanoma, NSCLC, or HNSCC who are set to receive treatment with a single-agent PD-1 or PD L1 antagonist regimen as indicated in the respective European MA or under the conditions of a TAU and according to the standard practices at the investigational site. Included patients will be followed for a total of 5 years. Prior to initiation of PD-1 or PD-L1 antagonist therapy, included patients will undergo a biopsy of a tumour lesion (unless suitable archived material is available) and provide a blood sample for immunohistochemistry and genomic studies. Patients at selected participating sites will also be asked to provide stool and saliva samples (optional). Additional optional biopsy samples may be collected from consenting patients after 42 (±3) days of PD-1 or PD-L1 antagonist treatment and in the event of disease progression or recurrence. Additional blood samples will also be collected at regular intervals throughout the observation period until disease progression, regardless of whether PD-1 or PD-L1 antagonist treatment is ongoing or has discontinued. Efficacy of treatment will be evaluated using both Response Evaluation Criteria in Solid Tumours (RECIST) and immune-related RECIST (iRECIST). Information regarding the PD-1 or PD-L1 antagonist related toxicities, subsequent antineoplastic treatments, and survival status will also be collected during the trial. An elastic-net approach will be used to identify correlations between different parameters and develop a signature of response to treatment. For each indication, the patients will be separated into two cohorts: a 'training' cohort and a 'validation' cohort. The 'training' cohort will be made up of the first patients included in the indication and will be used to develop a predictive response score. The 'validation' cohort will include all the remaining patients. The performance of the predictive score will be tested in this second cohort.