Research

Advanced Analysis of Metabolism by combing magnetic resonance with AI

In our projects we combine high-throughput NMR spectroscopy with methods from artificial intelligence to achieve a deeper understanding of human diseases and an improved disease prediction. In the following an overview about some recent projects is given.

In vivo prediction of IDH mutations for improved diagnosis of brain cancers

idh-prediction.spang-lab.de

The enzyme isocitrate dehydrogenase (IDH) affects glioma (a certain type of brain tumor) cell metabolism in multiple ways. Mutation of IDH is not only indicative of the presence of astrocytoma or oligodendroglioma but it also comes with a better prognosis and constitutes a promising therapeutic target. Therefore, determination of IDH mutation status is essential in clinical practice. To this end, we combined non-invasive in vivo magnetic resonance spectroscopy with AI for predicting the IDH mutation status. We made our predictive model available in an easy to use app.

Bumes E., Fellner C., Franz A. Fellner F.A., Fleischander K., Häckl M., Lenz S., Linker R., Mirus T., Oefner P.J., Paar C., Proescholdt M., Riemenschneider M.J., Rosengarth K., Weis S., Wendl C., Wimmer S., Hau P., Gronwald W*. & Hutterer M*. Validation Study for non-invasive Prediction of IDH Mutation Status in Patients with Glioma using In Vivo 1H-Magnetic Resonance Spectroscopy and Machine Learning. Cancers, 14, 2762 (2022). *shared last authorship. Doi: 10.3390/cancers14112762.

Bumes E., Wirtz F.-P., Fellner C., Grosse J., Hellwig D., Oefner P.J., Häckl M., Linker R., Proescholdt M., Schmidt N.O., Riemenschneider M.J., Samol C., Rosengarth K., Wendl C., Hau P., Gronwald W.* & Hutterer M.* Non-Invasive Prediction of IDH Mutation in Patients with Glioma In Vivo 1H-Magnetic Resonance Spectroscopy and Machine Learning. Cancers. 12, 3406, DOI: 10.3390/cancers12113406 (2020). *shared last authorship. Doi: 10.3390/cancers12113406.

Predicting the need for renal replacement therapy in patients with chronic kidney disease

In context of the German Study on Chronic Kidney Disease (GCKD) close to 5,000 human plasma specimens were analyzed by NMR. Based on this data together with a selected set of clinical parameters different AI models for the improved prediction of the need for renal replacement therapy were developed.

Zacharias H.U., Altenbuchinger M., Schultheiss U.T., Raffler J., Kotsis F., Ghasemi S., Ali I., Kollerits B., Metzger M., Steinbrenner I., Sekula P., Massy Z.A., Combe C., Kalra P.A., Kronenberg F., Stengel B., Eckardt K.-U., Köttgen A., Schmid M., Gronwald W. & Oefner P.J. on behalf of the GCKD investigators. A Predictive Model for Progression of CKD to Kidney Failure Based on Routine Laboratory Tests. Am. J. Kidney Dis. 79, 217-230.e.1 (2022). Doi: 10.1053/j.ajkd.2021.05.018.

Zacharias H.U., Altenbuchinger M., Schultheiss U.T., Samol C., Kotsis F., Poguntke I., Sekula        P., Krumsiek J., Köttgen A., Spang R., Oefner P.J. & Gronwald W. A novel metabolic signature to predict the requirement of dialysis or renal transplantation in patients with chronic kidney disease. J. Proteome Res., 18, 1796-1805 (2019). Doi: 10.1021/acs.jproteome.8b00983.

Deconvolution and Integration of 1D NMR Data: MetaboDecon1D

The large number of metabolites present in a biological specimen such as urine or plasma leads to considerable signal overlap in one-dimensional NMR spectra, which in turn hampers both signal identification and quantification. As a consequence, we have developed an easy-to-use R-package that allows the fully automated deconvolution of overlapping signals in the underlying Lorentzian line-shapes.

Häckl M., Tauber P., Schweda F., Zacharias H.U., Altenbuchinger M., Oefner P.J. & Gronwald W. An R-package for the Deconvolution and Integration of 1D NMR data: MetaboDecon1D. Metabolites, 11, 452.  DOI: 10.3390/metabo11070452 (2021). Doi: 10.3390/metabo11070452