Happy to announce the official release of MassVision extension v 1.0, now available through the 3D Slicer Extension Manager (compatible with the latest stable release, v5.8.1).
MassVision is an end-to-end platform for AI-driven exploration and analysis of mass spectrometry imaging (MSI) data.
Key Features:
- Data compatibility: imzML, structured CSV, hierarchical HDF5, DESI TXT images
- Visualization: targeted (single-ion heatmap, multi-ion colormap), untargeted (global/local contrast PCA, UMAP, t-SNE), multi-pixel spectrum plotting
- Exploration: clustering, segmentation, correlation analysis
- Dataset generation: spatial co-localization with pathology annotations, pathology-guided spatial annotation, spectrum labeling and extraction
- Multi-slide merge: feature alignment, peak matching
- Preprocessing: normalization (TIC, TSC, single-ion, mean, median, RMS), spectral filtering, spatial pixel aggregation
- Statistical analysis: ANOVA, volcano plots, boxplots
- AI model training/validation:
- Feature ranking (PLS-DA, Linear SVC, LDA)
- Feature selection (automated, manual)
- Data stratification (random, patient-based, cross-validation)
- Data balancing (oversampling, undersampling)
- Model training (PCA-LDA, Linear SVM, Random Forest, PLS-DA)
- Whole-slide AI deployment: feature matching, global or masked deployment
Useful Links:
Demonstrations:
Visualization and exploration
Spatial colocalization
Pathology guided ROI dataset curation
Statustucal analysis
Citation
Please use the following citations if you use MassVision in your research
A Jamzad, J Warren, A Syeda, M Kaufmann, N Iaboni, C JB Nicol, J F Rudan, K YM Ren, D Hurlbut, S Varma, G Fichtinger, and P Mousavi; "MassVision: An Open-Source End-to-End Platform for AI-Driven Mass Spectrometry Imaging Analysis, Analytical Chemistry 2025, DOI: 10.1021/acs.analchem.5c04018.