TY - JOUR
T1 - Ultra-fast deep-learned CNS tumour classification during surgery
AU - Vermeulen, C.
AU - Pagès-Gallego, M.
AU - Kester, L.
AU - Kranendonk, M. E. G.
AU - Wesseling, P.
AU - Verburg, N.
AU - de Witt Hamer, P.
AU - Kooi, E. J.
AU - Dankmeijer, L.
AU - van der Lugt, J.
AU - van Baarsen, K.
AU - Hoving, E. W.
AU - Tops, B. B. J.
AU - de Ridder, J.
N1 - Funding Information: We acknowledge the Utrecht Sequencing Facility (USEQ) for providing sequencing service and data. USEQ is subsidized by the University Medical Center Utrecht and The Netherlands X-omics Initiative (NWO project 184.034.019). We thank the Princess Máxima Center Biobanking facility and Big Data Core for storing and providing samples and data. We also thank the UMC Utrecht Bioinformatics Expertise Core (www.ubec.nl) for the help in applying research data management according to the FAIR Principles. We thank W. de Leng and E. van der Biezen for retrieving data. We thank Cyclomics for allowing us to use their GridION device. this project was partially funded by the Oncode Institute technology development fund and a stichting Kinderen Kanker Vrij (KIKA) pilot project grant. Funding Information: We acknowledge the Utrecht Sequencing Facility (USEQ) for providing sequencing service and data. USEQ is subsidized by the University Medical Center Utrecht and The Netherlands X-omics Initiative (NWO project 184.034.019). We thank the Princess Máxima Center Biobanking facility and Big Data Core for storing and providing samples and data. We also thank the UMC Utrecht Bioinformatics Expertise Core ( www.ubec.nl ) for the help in applying research data management according to the FAIR Principles. We thank W. de Leng and E. van der Biezen for retrieving data. We thank Cyclomics for allowing us to use their GridION device. this project was partially funded by the Oncode Institute technology development fund and a stichting Kinderen Kanker Vrij (KIKA) pilot project grant. Publisher Copyright: © 2023, The Author(s).
PY - 2023/10/26
Y1 - 2023/10/26
N2 - Central nervous system tumours represent one of the most lethal cancer types, particularly among children1. Primary treatment includes neurosurgical resection of the tumour, in which a delicate balance must be struck between maximizing the extent of resection and minimizing risk of neurological damage and comorbidity2,3. However, surgeons have limited knowledge of the precise tumour type prior to surgery. Current standard practice relies on preoperative imaging and intraoperative histological analysis, but these are not always conclusive and occasionally wrong. Using rapid nanopore sequencing, a sparse methylation profile can be obtained during surgery4. Here we developed Sturgeon, a patient-agnostic transfer-learned neural network, to enable molecular subclassification of central nervous system tumours based on such sparse profiles. Sturgeon delivered an accurate diagnosis within 40 minutes after starting sequencing in 45 out of 50 retrospectively sequenced samples (abstaining from diagnosis of the other 5 samples). Furthermore, we demonstrated its applicability in real time during 25 surgeries, achieving a diagnostic turnaround time of less than 90 min. Of these, 18 (72%) diagnoses were correct and 7 did not reach the required confidence threshold. We conclude that machine-learned diagnosis based on low-cost intraoperative sequencing can assist neurosurgical decision-making, potentially preventing neurological comorbidity and avoiding additional surgeries.
AB - Central nervous system tumours represent one of the most lethal cancer types, particularly among children1. Primary treatment includes neurosurgical resection of the tumour, in which a delicate balance must be struck between maximizing the extent of resection and minimizing risk of neurological damage and comorbidity2,3. However, surgeons have limited knowledge of the precise tumour type prior to surgery. Current standard practice relies on preoperative imaging and intraoperative histological analysis, but these are not always conclusive and occasionally wrong. Using rapid nanopore sequencing, a sparse methylation profile can be obtained during surgery4. Here we developed Sturgeon, a patient-agnostic transfer-learned neural network, to enable molecular subclassification of central nervous system tumours based on such sparse profiles. Sturgeon delivered an accurate diagnosis within 40 minutes after starting sequencing in 45 out of 50 retrospectively sequenced samples (abstaining from diagnosis of the other 5 samples). Furthermore, we demonstrated its applicability in real time during 25 surgeries, achieving a diagnostic turnaround time of less than 90 min. Of these, 18 (72%) diagnoses were correct and 7 did not reach the required confidence threshold. We conclude that machine-learned diagnosis based on low-cost intraoperative sequencing can assist neurosurgical decision-making, potentially preventing neurological comorbidity and avoiding additional surgeries.
UR - http://www.scopus.com/inward/record.url?scp=85173857767&partnerID=8YFLogxK
U2 - https://doi.org/10.1038/s41586-023-06615-2
DO - https://doi.org/10.1038/s41586-023-06615-2
M3 - Article
C2 - 37821699
SN - 0028-0836
VL - 622
SP - 842
EP - 849
JO - NATURE
JF - NATURE
IS - 7984
ER -