Intraoperative DNA methylation classification of brain tumors impacts neurosurgical strategy

Luna Djirackor works at the Oslo University Hospital, Norway, where her team see about 300 adults and 30 children with brain tumours every year. A typical case they might see is someone of preteen age coming to the clinic with their parents, who has been developing gradually increasing unsteadiness, which has led to them to stop taking part in sports. Later on, they start to spill when they drink, and develop slurred speech. An MRI scan of the brain reveals a mass, which ‘you can imagine is completely devastating and so emotionally challenging for the family’.

The first thought is, could this be an ependymoma — a tumour seen in children in the brain or spinal cord — of which there are four molecular subgroups based on their genomic/epigenomic alterations. Survival prognosis varies widely for these subgroups. It could alternatively be a medulloblastoma, of which there are also four subgroups, also with differing prognoses. As the outcomes are so variable, it is critical to identify which type of tumour is present. The location of the tumour also needs to be confirmed, as the location of the tumour itself and its potential expansion, and consequences of resection, will vary depending on the location. This can lead to difficult risk vs. benefit scenarios — the growth of the tumour vs. damage from resection could either lead to substantially reduced quality of life.

Cancer is a leading cause of death worldwide, and brain tumours are the most common cause of cancer death in children — ‘we are really passionate about improving these numbers and giving the best chance possible to these families and these patients’.

Luna discussed the current typical workflow for a patient arriving in their neurosurgery ward: the referred patient will have an MRI scan, followed by surgery within a number of days or weeks, to relieve pressure within the brain and debulk the tumour tissue. A biopsy is taken during surgery and sent to histology, from which a diagnosis is arrived at within an hour or so, for the surgeon to then decide whether to do a partial resection or completely remove the tumour. This is based on a clear diagnosis, and not what happens every time — in some instances, after histological analysis, further molecular subtyping can reach a conflicting conclusion. Can we be more specific?

In 2018, a study was published detailing successful, precise DNA methylation-based classification of central nervous system tumours: ‘this study was such a landmark study for the field’, changing diagnostic grading criteria and patient management. However, the turnaround time for such analysis was days to weeks.

At London Calling 2017, Philipp Euskirchen presented a study on how nanopore sequencing showed the potential for same-day, accurate genomic and epigenomic subtyping of brain tumours. ‘That was absolutely amazing’. Luna’s team were incredibly interested in these results and wondered if it could be done even faster, possibly within 2 hours so that it could impact surgery. They therefore set up a nanopore whole-genome sequencing and methylation classification workflow, using the Rapid Barcoding Kit for library preparation and MinION sequencer, barcoding 10-12 samples per run, and sequencing for 24 hours (for retrospective and prospective cases). One intraoperative run was also performed for one individual. The nanoDx pipeline was used for analysis, which contains all the necessary software and outputs a PDF report containing the copy number profile, methylation-based classification results, and a plot of the subtypes.

On research samples from 55 adults and 50 children, the methylation analysis results obtained from a couple of hours of nanopore sequencing matched the pathology results from weeks later. Nanopore DNA methylation analysis (NDMA) also correctly classified tumour tissue that had given inconclusive histology results. This suggested that ‘even on a really tiny little piece of tissue with not too much tumour’ the nanopore sequence data could get the correct result.

Having confirmed that their NDMA workflow was sensitive, they then tested it out on a further 20 intraoperative research biopsy samples, reducing the sequencing time to 30-60 minutes; NDMA analysis took a median of 28 minutes. The first report could be returned within 91-161 minutes. Their sequencing statistics suggested that only 30 min, or sometimes less, was required to detect ~3,500 CpG sites, which usually provided ‘a really solid call’. Out of the 20 cases, the data suggested that nanopore DNA methylation analysis had the potential to impact surgical strategy and management in 12 of those cases.

Luna closed by extending her gratitude to those patients and parents who took part in the study, to her team, to Philipp Euskirchen’s team in Berlin, and to her funders.