Background: Implementing a laboratory-developed test sometimes requires incorporating an unconventional device into the laboratory information system (LIS) and customizing an interface to reduce transcription error and improve turnaround time. Such a custom interface is a necessity for complicated high-volume tests such as 25-OH Vitamin D by liquid chromatography-tandem mass spectrometry (LC-MS/MS) when there is no vendor-or LIS-supplied interface available. Here, we describe our work and experience interfacing a API 5000 LC-MS/MS instrument with our newly implemented LIS, Epic Beaker, using a combination of in-house scripting software and a middleware vendor, Data Innovations. Materials and Methods: For input interfacing, custom scripting software was developed to transcribe batched order lists generated by Epic into files usable by the instrument software, Analyst^®. For output interfacing, results from the LC-MS/MS system were fed to a unidirectional instrument driver made by Data Innovations and selected data were transferred to the LIS. Results: Creation and validation of a new driver by Data Innovations took approximately 6 months. The interface was adopted for 25-OH Vitamin D and testosterone testing during periods of increasing test volume (4.5-fold over 8 years and 1.25-fold over 5 years). The amount of time spent reporting 25-OH Vitamin D results decreased 82% per order resulting in a savings of 1370 technician work hours and the amount of time spent reporting testosterone results decreased 75% per order resulting in a savings of 400 technician work hours. Conclusions: A mixed model using custom scripting and curated commercial middleware serve as a durable interface solution for laboratory instrumentation such as an LC-MS/MS and are flexible to future changes in instrument software, networking protocols, and the scope of LISs and work area managers.

Background: The classification of diffuse large B-cell lymphomas into Germinal Center (GCB) and non-GC subtypes defines disease subgroups which are different both in terms of gene expression and prognosis. Given their clinical significance, several classification algorithms have been designed, some by making use of widely availability immunohistochemical techniques. Despite their high concordance with gene expression profiles (GEP) and prognostic value, these algorithms were based on technical and biological assumptions that could be improved in terms of performance for classification. Methods: In order to overcome this limitation, a new algorithm was obtained by analyzing a previously published dataset of 475 patients by using an automatic classification tree method. Results: The resulting algorithm classifies correctly 91.6% of the cases when compared to GEP, displaying a Receiver-Operator Characteristic (ROC) area under the curve of 0.934. Noteworthy features of this algorithm include the capability to classify GEP-unclassifiable cases and a significant prognostic value, both in terms of overall survival (60 months for non-GC vs not reached for GCB, P = 0.007) and progression-free survival (61.9 months vs not reached, P = 0.017). Conclusion: By using a machine learning classification method that avoids most pre-assumptions, the novel algorithm obtained is accurate and maintains relevant features for clinical implementation.