Traditional histopathology involves time-consuming procedures and requires pathologists' availability for cancer diagnosis. With the increase in global cancer burden, the current health system could be overwhelmed, resulting in patients' delayed treatment.

Hence, we developed a rapid, on-site, cost-effective cancer diagnosis alternative that can be easily operated by any personnel to assist medical personnel with routine workloads and intraoperative decision-making.

A commercialized miniature mass spectrometer is utilized in this screening method to reduce overall costs, space requirements and difficulty of routine maintenance. Besides, its small size enables easy placement at the desired site for analysis. By coupling the instrument with an ambient ionization technique known as paper spray ionization, specimens commonly used
for histopathological examinations can be analyzed with our platform with minimal sample preparation.

To further boost its adaptability into clinical settings, our sample collection protocols do not interfere with the current standard operating procedures for cancer diagnosis. It merely requires transfer of specimens onto a filter paper. The sample will then be placed onto a self-designed 3D-printed interface. Mass spectrum of the specimen can then be acquired and pre-processed for prediction using our self-constructed database. The database consists of over 1000 mass spectra of various specimens and cancers, intended for different applications. Using the appropriate machine learning model, prediction of the specimen can be to provide a second, objective opinion in cancer diagnosis for routine applications or intraoperative decision making within 10 minutes at a ~90% accuracy. As acquired mass spectra contains rich metabolite information on the specimens, our platform can also be utilized for research on cancer biomarker discovery.