Deep learning aided preoperative diagnosis of primary central nervous system lymphoma
Paul Vincent Naser,
Miriam Cindy Maurer,
Maximilian Fischer,
Kianush Karimian-Jazi,
Chiraz Ben-Salah,
Awais Akbar Bajwa,
Martin Jakobs,
Christine Jungk,
Jessica Jesser,
Martin Bendszus,
Klaus Maier-Hein,
Sandro M. Krieg,
Peter Neher,
Jan-Oliver Neumann
Affiliations
Paul Vincent Naser
Heidelberg University Hospital, Department of Neurosurgery, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany; Heidelberg University, Medical Faculty, Grabengasse 1, 69117 Heidelberg, Germany; Heidelberg University Hospital, Division of Stereotactic Neurosurgery, Department of Neurosurgery, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany; AI Health Innovation Cluster, German Cancer Research Center (DKFZ), Heidelberg, Germany; Corresponding author
Miriam Cindy Maurer
German Cancer Research Center (DKFZ) Heidelberg, Division of Medical Image Computing, Germany, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany; Department of Medical Informatics, University Medical Center Göttingen, Von-Siebold-Straße 3, 37075 Göttingen, Germany
Maximilian Fischer
Heidelberg University, Medical Faculty, Grabengasse 1, 69117 Heidelberg, Germany; German Cancer Research Center (DKFZ) Heidelberg, Division of Medical Image Computing, Germany, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany; German Cancer Consortium (DKTK), partner site Heidelberg, Heidelberg, Germany
Kianush Karimian-Jazi
Heidelberg University, Medical Faculty, Grabengasse 1, 69117 Heidelberg, Germany; Heidelberg University Hospital, Department of Neuroradiology, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany
Chiraz Ben-Salah
Heidelberg University, Medical Faculty, Grabengasse 1, 69117 Heidelberg, Germany; Department of Diagnostic and Interventional Radiology, University Hospital Heidelberg, Im Neuenheimer Feld 110, 69120 Heidelberg, Germany
Awais Akbar Bajwa
Heidelberg University Hospital, Department of Neurosurgery, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany; Heidelberg University, Medical Faculty, Grabengasse 1, 69117 Heidelberg, Germany
Martin Jakobs
Heidelberg University Hospital, Department of Neurosurgery, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany; Heidelberg University, Medical Faculty, Grabengasse 1, 69117 Heidelberg, Germany; Heidelberg University Hospital, Division of Stereotactic Neurosurgery, Department of Neurosurgery, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany
Christine Jungk
Heidelberg University Hospital, Department of Neurosurgery, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany; Heidelberg University, Medical Faculty, Grabengasse 1, 69117 Heidelberg, Germany
Jessica Jesser
Heidelberg University, Medical Faculty, Grabengasse 1, 69117 Heidelberg, Germany; Heidelberg University Hospital, Department of Neuroradiology, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany
Martin Bendszus
Heidelberg University, Medical Faculty, Grabengasse 1, 69117 Heidelberg, Germany; Heidelberg University Hospital, Department of Neuroradiology, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany
Klaus Maier-Hein
German Cancer Research Center (DKFZ) Heidelberg, Division of Medical Image Computing, Germany, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany; German Cancer Consortium (DKTK), partner site Heidelberg, Heidelberg, Germany; National Center for Tumor Diseases (NCT), NCT Heidelberg, a partnership between DKFZ and the University Medical Center Heidelberg, 69120 Heidelberg, Germany; Pattern Analysis and Learning Group, Department of Radiation Oncology, Heidelberg University Hospital, 69120 Heidelberg, Germany; AI Health Innovation Cluster, German Cancer Research Center (DKFZ), Heidelberg, Germany
Sandro M. Krieg
Heidelberg University Hospital, Department of Neurosurgery, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany; Heidelberg University, Medical Faculty, Grabengasse 1, 69117 Heidelberg, Germany
Peter Neher
German Cancer Research Center (DKFZ) Heidelberg, Division of Medical Image Computing, Germany, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany; German Cancer Consortium (DKTK), partner site Heidelberg, Heidelberg, Germany; Pattern Analysis and Learning Group, Department of Radiation Oncology, Heidelberg University Hospital, 69120 Heidelberg, Germany
Jan-Oliver Neumann
Heidelberg University Hospital, Department of Neurosurgery, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany; Heidelberg University, Medical Faculty, Grabengasse 1, 69117 Heidelberg, Germany; Heidelberg University Hospital, Division of Stereotactic Neurosurgery, Department of Neurosurgery, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany
Summary: The preoperative distinction between glioblastoma (GBM) and primary central nervous system lymphoma (PCNSL) can be difficult, even for experts, but is highly relevant. We aimed to develop an easy-to-use algorithm, based on a convolutional neural network (CNN) to preoperatively discern PCNSL from GBM and systematically compare its performance to experienced neurosurgeons and radiologists. To this end, a CNN-based on DenseNet169 was trained with the magnetic resonance (MR)-imaging data of 68 PCNSL and 69 GBM patients and its performance compared to six trained experts on an external test set of 10 PCNSL and 10 GBM. Our neural network predicted PCNSL with an accuracy of 80% and a negative predictive value (NPV) of 0.8, exceeding the accuracy achieved by clinicians (73%, NPV 0.77). Combining expert rating with automated diagnosis in those cases where experts dissented yielded an accuracy of 95%. Our approach has the potential to significantly augment the preoperative radiological diagnosis of PCNSL.
