SpiderMass ambient mass spectrometry and multimodal machine learning enable ex vivo ovarian cancer typing and exploratory immunoscoring
Abstract
Complete cytoreductive surgery remains one of the strongest determinants of outcome in ovarian cancer, yet surgeons still lack rapid tissue-assessment tools that can be repeated throughout an operation. Here we evaluated whether SpiderMass ambient mass spectrometry, combined with machine-learning models, could support ex vivo ovarian tissue typing and exploratory immune microenvironment mapping. A total of 128 ovarian specimens, from 119 patients, were analyzed to train subtype-classification models from fresh-frozen and formalin-fixed paraffin-embedded (FFPE) material, and 24 independent tissues (from 16 patients) were reserved for blinded region-level testing. Initial PCA-LDA models were improved by screening 24 classifiers; Ridge models reached up to 97% 5-fold cross-validation accuracy on the combined cohort. In blinded analyses, the mixed Ridge model produced the fewest errors, although misclassification remained concentrated in underrepresented endometrioid regions. A dual-input network combining SpiderMass spectra with digitized histology improved internal performance to 99% in 5-fold cross-validation and 100% on a small, blinded image-spectrum set, outperforming the molecular-only branch. Model explanation followed by MALDI-MSI cross-checking identified 26 subtype-associated lipids. We then trained a LightGBM cell-state model from immune and epithelial cell spectra and applied it to SpiderMass imaging data. Spatial predictions were broadly concordant with multiplex MALDI-IHC and highlighted subtype-specific differences in immune-cell distribution. In an exploratory analysis of eight high-grade serous carcinoma samples obtained before chemotherapy, longer survivalappeared to be associated with higher lymphocyte scores, higher M1-like macrophage scores and a higher M1/M2 ratio, whereas shorter survival appeared to be associated with higher cancer-cell scores. These data support SpiderMass as a promising ex vivo platform for ovarian cancer typing and hypothesis-generating immunoscoring, while underscoring the need for prospective intraoperative studies, orthogonal biomarker validation, validation in larger cohorts for exploratory immunoscoring and multicenter patient-level external validation before clinical implementation.
Related
- Tesorai Search: Cloud-based Database Search Engine Boosts Identifications for Mass Spectrometry Proteomics With a Pretrained Peptide-spectrum Deep-learning Model
- Differentiation of Bacterial Isolates Using Mass Spectrometry and Machine Learning Analysis of Volatile Organic Compounds.
- Multimodal Representation Learning for Exploring Natural Product Chemical Space
- Scaling multimodal materials representation learning with synthetic narratives
- End-to-end multimodal structure elucidation from raw spectra combining contrastive learning and evolutionary algorithms
BibTeX
@article{ledoux2026spidermass,
title = {{SpiderMass ambient mass spectrometry and multimodal machine learning enable ex vivo ovarian cancer typing and exploratory immunoscoring}},
author = {Léa Ledoux and Yanis Zirem and Laurine Lagache and Lucas Roussel and Delphine Hudry and Bertrand Meresse and Marie Delbeke and Eric Leblanc and Yves-Marie Robin and Camille Pasquesoone and Delphine Bertin and Fabrice Narducci and Michel Salzet and Isabelle Fournier},
journal = {Journal of Advanced Research},
year = {2026},
month = sep,
doi = {10.1016/j.jare.2026.07.022},
url = {https://hal.science/hal-05745358v1},
note = {CC BY},
}
RIS
TY - JOUR
TI - SpiderMass ambient mass spectrometry and multimodal machine learning enable ex vivo ovarian cancer typing and exploratory immunoscoring
AU - Léa Ledoux
AU - Yanis Zirem
AU - Laurine Lagache
AU - Lucas Roussel
AU - Delphine Hudry
AU - Bertrand Meresse
AU - Marie Delbeke
AU - Eric Leblanc
AU - Yves-Marie Robin
AU - Camille Pasquesoone
AU - Delphine Bertin
AU - Fabrice Narducci
AU - Michel Salzet
AU - Isabelle Fournier
PY - 2026
DA - 2026/09/10
JO - Journal of Advanced Research
DO - 10.1016/j.jare.2026.07.022
UR - https://hal.science/hal-05745358v1
AB - Complete cytoreductive surgery remains one of the strongest determinants of outcome in ovarian cancer, yet surgeons still lack rapid tissue-assessment tools that can be repeated throughout an operation. Here we evaluated whether SpiderMass ambient mass spectrometry, combined with machine-learning models, could support ex vivo ovarian tissue typing and exploratory immune microenvironment mapping. A total of 128 ovarian specimens, from 119 patients, were analyzed to train subtype-classification models from fresh-frozen and formalin-fixed paraffin-embedded (FFPE) material, and 24 independent tissues (from 16 patients) were reserved for blinded region-level testing. Initial PCA-LDA models were improved by screening 24 classifiers; Ridge models reached up to 97% 5-fold cross-validation accuracy on the combined cohort. In blinded analyses, the mixed Ridge model produced the fewest errors, although misclassification remained concentrated in underrepresented endometrioid regions. A dual-input network combining SpiderMass spectra with digitized histology improved internal performance to 99% in 5-fold cross-validation and 100% on a small, blinded image-spectrum set, outperforming the molecular-only branch. Model explanation followed by MALDI-MSI cross-checking identified 26 subtype-associated lipids. We then trained a LightGBM cell-state model from immune and epithelial cell spectra and applied it to SpiderMass imaging data. Spatial predictions were broadly concordant with multiplex MALDI-IHC and highlighted subtype-specific differences in immune-cell distribution. In an exploratory analysis of eight high-grade serous carcinoma samples obtained before chemotherapy, longer survivalappeared to be associated with higher lymphocyte scores, higher M1-like macrophage scores and a higher M1/M2 ratio, whereas shorter survival appeared to be associated with higher cancer-cell scores. These data support SpiderMass as a promising ex vivo platform for ovarian cancer typing and hypothesis-generating immunoscoring, while underscoring the need for prospective intraoperative studies, orthogonal biomarker validation, validation in larger cohorts for exploratory immunoscoring and multicenter patient-level external validation before clinical implementation.
SN - 2090-1232
ER -