Transcriptomic deconvolution of neuroendocrine neoplasms predicts clinically relevant characteristics

#3146

Introduction: Therapeutic decisions in oncology depend on a precise pathological classification of individual neoplasms. However, a comprehensive training of machine-learning models requires sufficiently large numbers of training samples, which are usually not available for rare cancer types.

Aim(s): Here, we report on a new data-augmentation technique to support the training of machine-learning models on ‘omics’ data from pancreatic (PanNEN) and non-pancreatic neuroendocrine neoplasms (NEN). The approach reconstructed a given transcriptome based on healthy pancreatic cell type signatures and creates a machine-learning model that integrates the observed reconstruction error and predicted cell type proportions as features.

Materials and methods: We deconvolved the transcriptomes from five different NEN and PanNEN studies with machine-learning models, ensured the statistical soundness of the results and analysed whether correlations between the predicted cell type proportions and reconstruction error to clinical characteristics existed.

Conference: 18th Annual ENETS Concerence (2021)

Presenting Author: Otto R

Authors: Otto R, Detjen K, Riemer P, Grötzinger C, Rindi G,

Keywords: bioinformatics, NET, PanNEN, deconvolution, transcriptomics, RNA-seq, neuroendocrine cancer,