Picosecond infrared laser mass spectrometry for 10-second identification of lymphoproliferative imposter tumours in patient-derived xenografts

Patient derived xenografts (PDXs) are widely used in preclinical research. However, lymphoproliferative ‘outgrowths’ at the site of tumour xenoplantation are a common source of failure in the creation of the disease model. In this work, we assessed the performance of 10-second molecular profiling of xenoplanted tissue with picosecond infrared laser mass spectrometry (PIRL-MS) as a new method for rapid identification of lymphoproliferative ‘outgrowths’ in serial passages to streamline the quality control workflow. PIRL-MS can identify ‘imposter’ lymphoproliferative tumours with sensitivity and specificity values of > 99%. This observation is established over n = 258 independent PDX specimens and n = 3,393 ten-second mass spectral data points used for building and validating (blind assessment) a classifier multivariate model to enable discrimination. We first established a classifier model based on principal component analysis coupled with linear discriminant analysis (PCA-LDA) to discriminate between true solid tumour PDXs (of 5 common epithelioid cancer types originating from lung, pancreas, ovarian, colon and head & neck as well as imposter tumours of lymphoproliferative origin. Implementation of the classifier only requires 10 seconds of mass spectral data collection (using a hand-held probe) and less than an additional second for data processing and evaluation against the model towards a classification. In addition, PIRL-MS analysis does not require any tissue preparation before analysis, and from previous research can also be deployed in situ/in vivo to save time. These attributes, coupled with its reported high sensitivity and specificity for identification of imposter lymphoproliferative tumours, position PIRL-MS as a rapid quality control method for fidelity assessment of xenoplanted tissues. These observations motivate follow-on work to reduce the cost and the footprint of the PIRL-MS platform towards lowering the adoption barrier for routine use.