Differential protein expression patterns obtained by mass spectrometry can aid in the diagnosis of Hodgkin's disease.

More than 90% of patients with cancer, if diagnosed early, can be promptly treated; however diagnosis usually occurs after cancer cells have metastasized. Recent technological advances in mass spectrometry challenges the field of machine learning to model such high dimensional datasets for clinical diagnosis and prognosis. Here we use support vector machines recursive feature elimination to hunt for protein expression patterns in the serum mass spectra of Hodgkin's disease (HD) patients and control subjects (CS) that could aid in diagnosing-the disease. Based on eight selected features, support vector machines was able to correctly classify among all CS and HD patients based on the leave-one-out. We also correctly classified an independent dataset, acquired from the same samples, with the previously generated SVM model.