CPG used headspace solid-phase microextraction gas chromatography-mass spectrometry to generate unique, chromatographic fingerprints from the urine of cancer patients relative to biopsy-negative controls. Researchers from MIT and John Hopkins then used the HS SPME GC-MS information in combination with data from trained cancer detection dogs to train an artificial neural network (ANN) to emulate canine olfactory diagnostics. Findings published in PLOS ONE.