Impact of early trials in molecularly-characterized patients (pts) with head and neck cancer (HNC).

Abstract

6031 Background: Multiple genomic alterations were described in HNC, including squamous cell carcinomas (SCC), salivary gland (SG) and nasopharyngeal (NF) tumors. Tumor molecular profiling (TMP) may increase therapeutic alternatives in early trials for pts with refractory metastatic (met) HNC. We evaluate the impact of matched/unmatched therapy (mT/uT) in HNC with potentially targetable alterations. Methods: From 2010-16, 47 met HNC pts were treated in 57 early trials after TMP. Clinical benefit was measured by: time to progression (TTP); clinical benefit rate (CBR: complete response [CR], partial response [PR] and stable disease [SD] > 4months [m]); progression-free survival [PFS] ratio≥1.3 (PFS under molecular therapeutics/PFS upon last prior chemotherapy [pT]). Results: Median age was 51 years; median number of pT lines was 1 (0-5). In total, 26 SCC, 11 SG, 8 NF and 2 nasosinusal pts (mostly with lymph nodes and lung met) were treated with small kinases (SK) inhibitors (inh) (50%; main targets PI3K/HER/FGFR), immune-oncology (IO) drugs (40%, PD-1/PD-L1), angiogenesis inh (5%) or chaperone, cytidine analog, RNA polymerase (5%). 14/57 trials were mT including CDKinh (1 CDKN2A mutation [mut]), PI3Kinh (5 PIK3CA mut, 2 PTEN mut), PI3K/MEKinh (1 NRAS mut), HERinh (1 ERBB3mut), FGFRinh (2 FGFR1mRNA high, 1 FGF3/4/19 ligand amplified) and porcupine inh (1 ZNFR3mut). Distribution by tumor: IO (16 SCC/5 NF), SKmT (8SG/4SCC/2NF) and SKuT (9SCC/4SG). Responses: 1CR in SCC (IO), 5 PR in SCC (4 IO, 1 FGFRinh), 3 PR in NF (1 cytidin analog, 1 IO, 1 PI3K/MEKinh). Benefit: median TTP 9.33m (CI95% 7-20) with upward trend in NF and SG vs SCC (HR 0.6;p = 0.28), without differences according to target therapies (HR 0.8; p = 0.9); CBR of 58%, without differences by tumor type (p = 0.42) or therapy (SKmT, SKuT, IO; p = 0.5); 60% with PFS ratio≥1.3, significantly higher in SCC and SG (p = 0.016), particularly with IO drug and SKmT, respectively. Conclusions: Considering our preselected fit population as bias selection, individualized treatment selection with novel therapeutics based on TMP, especially in NF and SG irrespective of targets (SKmT, SKuT, IO), and in SCC with IO drugs, seems to confer substantial clinical benefit.