ABSTRACT
Ambient ionization mass spectrometry (MS) requires minimal sample preparation and therefore provides unprecedented throughput compatible with clinical needs and especially with the time restrictions required for intraoperative tumor diagnosis and margin delineation. The ability to reveal molecular changes in heterogeneous tissues prior to visible morphological changes is a clear advantage of label-free molecular imaging techniques in perioperative molecular pathology. The multidisciplinary nature of translational molecular imaging results in different data streams contributed by different imaging experts. Active learning strategies are used to segment MS-based molecular images and can already automatically annotate and classify different cellular phenotypes within a single tissue section taking advantage of available information in other domains. The resulting tissue debris is continuously transported from the surgical site to the mass analyzer by a continuous drain of an aqueous solution. It is evident that personalized medicine based on molecular imaging modalities has entered the era of Big Data.
