We are in the era of real-time sequencing, but no protocol reliant on that technology has yet been implemented for diagnostic testing in clinical medicine. In a quarter of children with cancer, the disease has already metastasized by the time it is diagnosed, meaning that time is a prognostic factor for survival. Nanopore sequencing applied to diagnosis and classification of leukemias would enable medical teams to circumvent the latency period between the appearance of initial symptoms and the start of personalized treatment.
Eight-year-old Nathan arrives in hospital, referred by his family doctor because of symptoms including pale skin and extreme fatigue. Blood tests orient the initial diagnosis toward suspected leukemia. He is then scheduled for an operating-room procedure to collect a sample of his bone marrow, on which a battery of tests is conducted. The precise diagnosis is rendered two weeks later: Nathan has an acute lymphoblastic leukemia, and the disease is already in an advanced stage. He begins chemotherapy immediately. Based on further specific tests, the medical team designs a treatment adapted to Nathan’s profile.
DNA and RNA sequencing is gradually being incorporated into the care protocols for people with leukemia. Until now, however, the technique has been reserved for diagnosis and analysis of the most complex cases, in part because of its high cost (thousands of dollars per patient) but also the bulky equipment required. In 2016, however, Oxford Nanopore Technologies, an English biotech startup spun out from Oxford University, commercialized the world’s first portable sequencing technology, which fits on a USB drive and is dubbed MinION. The device, in addition to being very compact, easily transportable and inexpensive, is extremely sensitive and enables real-time data analysis. Using this tool to diagnose pediatric leukemias would provide medical teams with the information they need to validate the subtype of leukemia and choose a treatment method adapted to the child’s profile within a few hours, or even minutes, at a cost of about $100.
The goal of my project is to incorporate nanopore sequencing technology into the standard diagnostic pathway for pediatric leukemia at Sainte-Justine university hospital centre in Montréal, to enable ultrarapid and precise classification and therapeutic direction at low cost. Studies conducted using the previous generation of sequencers have highlighted the molecular rearrangements characteristic of leukemias and demonstrated similarities of profiles in patients with the same subtypes. The technologies currently in use, however, do not allow for analysis of the raw sequencing data in real time. The in-clinic protocols used require between 30 and 50 days on average, and are therefore reserved for the most complex cases, involving particular rearrangements or uncommon subtypes.
The “real-time” dimension of third-generation sequencing, of which MinION is an example, changes everything and opens the door to widespread use in clinical settings. These sequencers can also detect epigenetic changes, which are modifications occurring in nucleotides, the building blocks of life, and which change the configuration of molecules. Some of those modifications are well known and are considered as prognostic markers for treatment responses.
This epigenetic analysis, which so far has been very seldom conducted in clinical settings, is now accessible, which widens the scope of possibilities for identification of new markers characteristic of certain leukemia subtypes that could serve as new therapeutic targets.
The race to provide real-time diagnostics is therefore under way. The key is the use of new machine-learning algorithms applied to the raw electric signal generated by the MinION device, which allows for identification of a high number of specific characteristics in record time. Our preliminary results show that pediatric leukemias can be classified with less than 10 minutes of sequencing. Many parameters still need to be optimized, however, such as the choice of algorithms and hyperparameters, and the choice of biological medium. Some studies show that it is possible to establish a complete profile of the patient using only a peripheral blood sample, if it meets certain criteria. Using blood instead of bone marrow would avoid the need for a painful surgical procedure and do away with the latency period between the initial visit and the start of the adapted treatment. This way, in our example, Nathan could begin his personalized course of treatment the same day as his blood work results are received, which would significantly boost his chances of remission.
Application of this protocol as part of the standardized in-clinic patient intake procedure will open the door a little wider to real-time personalized medicine, while placing priority on comfort of patients, family caregivers and the care delivery team. Eventually, a similar protocol could be used for various diseases with distinguishable profiles, such as lymphomas, inflammatory diseases and many others.
This article was produced by Mélanie Sagniez, PhD student in bioinformatics in the Department of Biochemistry and Molecular Medicine (Université de Montréal), with the guidance of Marie-Paule Primeau, science communication advisor, as part of our “My research project in 800 words” initiative.