PATH-46. COMPUTATIONAL HISTOPATHOLOGY INFORMED RAPID TARGETED NANOPORE SEQUENCING ENABLES AFFORDABLE NEXT DAY REPORTING OF COMPREHENSIVE MOLECULAR MARKERS FOR CNS TUMOUR DIAGNOSTICS

Abstract BACKGROUND Integrative brain tumour diagnostics indisputably requires comprehensive reporting of molecular markers. The 2021 WHO classification central nervous system (CNS) tumours substantially increased the set markers for routine evaluation, with greater significance to DNA methylation analysis in diagnostics. Limited by investment and batching, smaller labs clinics might suffer major delays delivering clinical decisions. To make precision accessible, we introduce an integrated computational histopathology adaptive nanopore sequencing workflow next day CNS METHODS We used CNS-CHiP- a multitask deep transfer learning model predict key alterations from H&E stained slides. For further characterisation subtyping, predictions formulate custom panel each patient. Targeted analyses were performed using Rapid-CNS2- neurooncology pipeline parallel copy-number, mutational that is flexible target selection no additional library preparation can be initiated upon receipt frozen sections. Sequencing was on portable MinION or GridION. RESULTS demonstrate our diagnostic samples received Department Neuropathology, University Hospital Heidelberg. CNS-CHiP predicted multiple pathognomonic (eg. IDH mutation, 7 gain/10 loss) reasonable accuracy. This provided basic information regarding tumor type instantly. Personalised panels enabled small sizes, resulting low time (up 24h) competitive costs. GPU-accelerated bioinformatics reduced > 24h < 3h. CONCLUSIONS Our harnessing histology-based instruct targeted up initial has potential facilitate results sample collection. combined Rapid-CNS2 thus aims affordable accessible hospitals especially low- middle-income countries.