Dual Laser and Desorption Electrospray Ionization Mass Spectrometry Imaging Using the Same Interface

For a more comprehensive
characterization of molecular heterogeneities
of matter, multimodal mass spectrometry imaging must be developed
to take advantage of the complementarity of information available
through different ionization mechanisms. We report the design, implementation,
and performance validation of a laser desorption imaging interface
composed of add-on components that adapt a commercial Desorption Electrospray
Ionization Mass Spectrometry (DESI–MS) imaging interface for
dual imaging of Picosecond Infrared Laser Mass Spectrometry (PIRL–MS)
with DESI–MS. The interface utilizes hardware elements and
data analysis pipelines already established for DESI–MS imaging,
and was further validated in cancer margin assessments using human
medulloblastoma cancers. The PIRL–MS images were robust and
reproducible across multiple experimental runs on independently prepared
xenograft tumors, and could be segmented into cancer and healthy regions
in concordance with pathology using a variety of supervised and unsupervised
clustering methods. The spectral quality and complexity obtained with
this interface were examined with infiltrating and noninfiltrating
tumors, and were comparable to other mass spectrometry analysis interfaces.
The average PIRL–MS spectra from spatially resolved images
could be used for robust cancer m/z model building to classify medulloblastoma cancer from healthy tissue
without any misclassifications, an observation that held true over
close to 70 sampling data points. While the unsupervised spectral
analysis methods suggested a slight suppression of signal in the phospholipid
range compared to the hand-held configuration, these changes were
insufficient to hamper utility in cancer margin assessment with spatially
resolved data obtained with our interface. Dual PIRL–MS and
DESI–MS imaging of consecutive sections, as suggested by multivariate
loading plots, revealed highly complementary molecular information
with m/z values identifiable with
one desorption method sufficient to reveal cancer regions being absent
in another, further emphasizing the need for effective hardware and
software interfaces for dual mass spectrometry imaging.