ionScell enables microscopy-independent single-cell spectral extraction from MALDI mass spectrometry imaging.

Reliable single-cell spectrum extraction from MALDI mass spectrometry imaging (MSI) is constrained by pixel-level signal mixing and by the lack of automated segmentation tools that operate without microscopy. We present ionScell, an open-source pipeline that derives cell boundaries directly from the Total Ion Current (TIC) image using adaptive thresholding, watershed segmentation and six-metric spectral quality control, without co-registered microscopy, fluorescent labels or manual ROI annotation. Across four cancer cell lines and two ionization modes (eight independent acquisition settings), ionScell generated thousands of QC-retained single-cell spectra per acquisition, discriminated binary NCH82/AU565 (AUC = 0.987; 96% accuracy) and ternary breast-cancer mixtures (macro-AUC = 0.926; 86% accuracy), and resolved NCH82 intra-line molecular heterogeneity supported by orthogonal single-cell proteomics. Cross-platform operation was demonstrated on a public MALDI-Orbitrap SpaceM HeLa/NIH3T3 co-culture dataset. On the same NCH82 dataset, ionScell recovered 2.3–3.1× more individually contoured cells than MSI Parser and provided outputs not available from the comparator, including per-cell ion maps, six-metric QC and probabilistic molecular-subpopulation assignments. ionScell is available as a Python package, Jupyter notebook and standalone GUI, with user documentation and example data.