ICHNO-ECHNO 2022 - Abstract Book

S2

ICHNO-ECHNO 2022

feasibility of ART has been assessed by exploratory studies. ART is not routinely used, as a consequence of limited resources, its time consuming character, and unreliable automatic contouring. 2 - Normal Tissue Complication Probability Assessment NTCP models are usually based on dose–volume histograms (DVHs), which are not ideal representations of 3D doses and assume that all the regions of an OAR have an equal functional importance, thus discarding organ specific spatial information. Models taking into account more data such as 3D dose distribution in OARs, dependencies between the dose delivered at other OARs may enhance toxicity prediction. Models using machine learning might be well-suited since the bigger amount of interdependent data to take into account. 3 - Tumor Control Probability Assessment Improving the tumor control probability assessment before treatment is a promising way to adapt the treatment strategy. Head and Neck tumor treatment strategy mainly depend on the TNM stage. AI and radiomics could also be used to improve prognosis classification, determine HPV Status and extra nodal extension.

Proffered papers: Innovative highlights 1

OC-0006 Early detection of head and neck squamous cell carcinoma using RNA-seq on tumor-educated platelets

N. Wondergem 1 , J.B. Poell 1 , S.G. in 't Veld 2 , E. Post 2 , S.W. Mes 1 , M.G. Best 2,3 , E. Bloemena 3,4 , C.R. Leemans 1 , T. Würdinger 2 , R.H. Brakenhoff 1 1 Amsterdam UMC, Vrije Universiteit Amsterdam, Cancer Center Amsterdam, Otolaryngology / Head & Neck Surgery, Amsterdam, The Netherlands; 2 Amsterdam UMC, Vrije Universiteit Amsterdam, Cancer Center Amsterdam, Neurosurgery, Amsterdam, The Netherlands; 3 Amsterdam UMC, Vrije Universiteit Amsterdam, Cancer Center Amsterdam, Pathology, Amsterdam, The Netherlands; 4 Academic Centre for Dentistry Amsterdam (ACTA), Maxillofacial Surgery/Oral Pathology, Amsterdam, The Netherlands Purpose or Objective Head and neck squamous cell carcinomas (HNSCC) arise in the mucosal lining of the upper aerodigestive tract, and are caused by carcinogen exposure or infection with human papillomavirus (HPV). The major determinant of prognosis is the occurrence of locoregional recurrence. Recurrences can be treated curatively, but late detection hampers clinical management. Tumor-educated platelets (TEPs) have been shown to hold promise as a diagnostic tool for several malignancies. The objective of this study is to provide proof of concept that TEPs might be well suited for early detection of (recurrent) HNSCC. Materials and Methods Platelet mRNA was isolated and sequenced from 101 HNSCC patients and 101 non-cancer controls. Non-cancer controls were matched for age, tobacco smoking and alcohol use. Patients and matched control pairs were randomly divided over training, evaluation and validation series (40/30/30%, respectively) to ensure equal distribution of matched variables among series. A particle swarm optimized (PSO) support vector machine (SVM) learning classifier was trained and re-trained iteratively using the training series to optimize classification performance of the evaluation series. Performance of the best classifier was subsequently evaluated using the independent validation series. Parallel classic statistical appoaches were employed for data analysis. Results 5,464 spliced mRNAs were detected. Those most contributive to the SVM algorithm were selected by ANOVA differential expression analysis, using PSO to determine the optimal FDR threshold. The resulting 200 gene classifier reached an AUC of 0.85 (95% CI, 0.75-0.95) and accuracy of 80% in the independent validation set for the detection of HNSCC, irrespective of HPV-status. Conclusion We show that a classifier of 200 TEP derived differentially expressed genes is able to identify HNSCC patients with an AUC of 0.85 and accuracy of 80%.

OC-0007 Intra-tumoral genetic heterogeneity in head and neck cancer

A. Pierik 1 , A. Brink 1 , J. Poell 1 , T. Poli 2 , R. Leemans 1 , R. Brakenhoff 1

1 Amsterdam UMC, Otolaryngology/Head and Neck Surgery, Amsterdam, The Netherlands; 2 University Hospital of Parma, Maxillofacial Surgery Unit, Parma, Italy Purpose or Objective Advanced stage head and neck squamous cell carcinomas (HNSCC) outside the oral cavity are often treated by definitive chemoradiotherapy, but locoregional recurrences still occur in 30–40%. In a previous study, low-coverage whole genome sequencing (lcWGS) for copy number alterations (CNAs) and target-enrichment deep sequencing for mutations were applied to paired tumors and local recurrences (LRs). Remarkably, half of the LRs did not show any identical mutation in the important head and neck cancer driver genes when compared to the corresponding primary tumors. This might be explained by intratumor genetic heterogeneity and treatment resistance of minor subclones. Here we analyzed the occurrence of intratumor genetic heterogeneity by analyzing CNAs and mutations in multiple biopsies of resected specimen.