Abstract

Poster Presentations

Day 1: Wednesday, May 14  Poster Room(Gekko)

Development and Application of Diagnosis System Using Probe Electrospray Ionization and Dual Penalized Logistic Regression Machine

(1Univ. Yamanashi, 2Waseda Univ., 3Univ. Yamanashi, 4Univ. Yamanashi, 5Shimadzu, 6Univ. Yamanashi)
oKentaro Yoshimura1, Kunio Tanabe1,2, Lee Chuin Chen4, Satoshi Ninomiya4, Hori Hirokazu3, Hideaki Izumi5, Kenzo Hiraoka6, Sen Takeda1

Probe electrospray ionization (PESI) is a modified method of electrospray ionization, and is one of the best solutions to realize the analysis of biological molecules in living animals. We are now trying to establish a diagnosis system based on the PESI-2020 combined with dual penalized logistic regression machine (PESI-2020+dPLRM). To evaluate the applicability of PESI-2020+dPLRM to the cancer diagnosis, we have analyzed human tissues. Total of 22,770 spectra were used as a training dataset of dPLRM, to which randomly selected test data were fed into the dPLRM. We obtained over 95 % coincidence to those diagnosed by the conventional pathohistological method. These data demonstrate that PESI-2020+dPLRM is a promising strategy in cancer diagnosis. For further diversify the application of PESI-2020, we optimized the ion inlet by using long tube made of stainless steel (2φ, length 200 mm), which is connected to the mass spectrometer. This modification enables us to place the orifice very near to the specimens even if they were too large to the original system. We are now evaluating the applicability of modified PESI-2020 system to various situations to open up new applications of this gadgetry.