Clinical correlation of HNSCC tissue cultures in the ex-vivo model to predict response to checkpoint inhibitor therapy

Checkpoint inhibitors (CPIs) are increasingly used in treating head and neck squamous cell carcinoma (HNSCC). However, they have a response rate of only 20%. Therefore, reliable test systems are needed to identify patients who could benefit from CPI before the start of therapy. The patient-derived ex-vivo 3D-HNSCC model established in our working group could be helpful in personalized treatment decisions. Seven vital 3D-HNSCC tissue models were treated ex-vivo with Cisplatin 80 μmol/ml or Pembrolizumab 5 mg/ml on days 1, 3, and 7 (of 10 days of culture). Taking into account the 3R principles (refine, reduce, replace), in addition to fetal bovine serum (FCS), xeno-free serum substitutes were used. The PD-L1 and Ki-67 status was determined immunohistochemically using TPS (Tumor Proportion Score) and proliferation index. The staining results were correlated with clinical patient data (e.g., TNM, recurrences/metastases, noxae). (ethics vote 2019- 528N) Two samples showed ex-vivo Cisplatin-mediated induction of PD-L1 (ex-vivo nr. 2 TPS from 5% to 15%, ex-vivo nr. 8 TPS from 5% to 25%). These donor patients had advanced lymph nodal metastasis with a pN3b status and tumor stages pT1 and pT2 without distant metastasis (cM0). A higher T-classification (pT3) without lymph node metastasis did not affect PD-L1 expression (pN0, ex-vivo nr. 4 TPS 35%). In summary, there is a heterogeneous response to the experimental treatment and the therapeutic response of the examined tissues and patients. Limitations of HNSCC primary cultures are the relatively short culture time and the risk of contamination. However, implementing 3D tissue culture models in the clinical routine could mean a real opportunity for improving the therapy response by selecting optimal treatment strategies.