Abstract
During brain-tumour surgery, the neurosurgeon must decide
how aggressively to resect long before a final molecular diagnosis is available.
Stimulated Raman Histology (SRH) yields a label-free image in
near real time and a CNN-based classifier turns it into per-entity tumour
scores within minutes after imaging. When can a surgeon act on
this read alone, and when must it defer to the slower molecular result?
As proof of principle, we assembled NanoSRH, a single-centre cohort of
currently 292 CNS specimens with same-sample SRH images and Oxford
Nanopore methylation classification with the crossNN scores aggregated
to the specimen level. SRH-CNN was dependable for morphologically
distinct, high-prevalence entities — glioblastoma and metastasis —
while the molecular distinctions that define tumour subtypes were limited
and were supplied by the methylome; the methylome was likewise
required to separate infiltrative tumour from control brain. The methylome
corrected imaging errors far more often than the reverse (67 vs. 20
specimens). Together these results map precisely where fast image-based
inference already suffices and where it must be backed by methylation.
Building on this, we are now retraining FastGlioma on unmatched SRH
samples and training a purpose-built intraoperative concordance-aware
model that acts on confident image-based calls and defers to the methylome
for molecularly defined decisions. We plan to expand the NanoSRH
dataset to other centres to train and validate on a more diverse sample
set.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/COMPAYL_056.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: https://papers.miccai.org/miccai-2026-sat/supp/COMPAYL_056_supp.pdf
Link to Open Review
BibTex
@InProceedings{StrJak_Act_MICCAISAT2026,
author = { Straehle, Jakob AND Cai, Shifei AND Coubez, Xavier AND Sankowski, Roman},
title = { { Act or defer? Multimodal intraoperative diagnosis of CNS tumours: A proof-of-concept study } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17251},
month = {pending},
page = {pending}
}