gms | German Medical Science

67th Annual Meeting of the German Society of Neurosurgery (DGNC)
Joint Meeting with the Korean Neurosurgical Society (KNS)

German Society of Neurosurgery (DGNC)

12 - 15 June 2016, Frankfurt am Main

Accuracy and reliability of MRI interpretation in patients with CNS lymphoma and glioma

Meeting Abstract

  • Witold H. Polanski - Klinik und Poliklinik für Neurochirurgie, Uniklinikum Dresden, Germany
  • Georgina Kusnick - Klinik und Poliklinik für Neurochirurgie, Uniklinikum Dresden, Germany
  • Johannes Gerber - Abteilung Neuroradiologie, Institut für Radiologische Diagnostik, Uniklinikum Dresden, Germany
  • Gabriele Schackert - Klinik und Poliklinik für Neurochirurgie, Uniklinikum Dresden, Germany
  • Stephan B. Sobottka - Klinik und Poliklinik für Neurochirurgie, Uniklinikum Dresden, Germany

Deutsche Gesellschaft für Neurochirurgie. 67. Jahrestagung der Deutschen Gesellschaft für Neurochirurgie (DGNC), 1. Joint Meeting mit der Koreanischen Gesellschaft für Neurochirurgie (KNS). Frankfurt am Main, 12.-15.06.2016. Düsseldorf: German Medical Science GMS Publishing House; 2016. DocP 021

doi: 10.3205/16dgnc396, urn:nbn:de:0183-16dgnc3961

Published: June 8, 2016

© 2016 Polanski et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License. See license information at http://creativecommons.org/licenses/by/4.0/.


Outline

Text

Objective: In patients with inoperable brain tumors located in eloquent brain regions histological diagnosis is a prerequisite for the determination of the adjuvant treatment. However, in certain cases brain biopsy may be risky and/or the patient may refuse the procedure and the diagnosis can only be based on MRI findings. Unfortunately, the exact MRI dependent diagnosis accuracy of certain brain tumors is rather unknown. We assessed and correlated radiological and histological diagnosis in brain tumor patients having undergone brain biopsy.

Method: From 2006 to 2013, 330 patients with inoperable brain lesions underwent tissue biopsies. In all patients the radiological suspected diagnosis and the histological diagnosis were prospectively analyzed and compared. The descriptive statistics were calculated for the most common suspected diagnosis.

Results: Only gliomas and lymphomas were analyzed, because they represented the most common diagnoses with 76.2% of all cases. In 146 of 330 patients a glioma was suspected and was confirmed histologically in 128 cases. This revealed a positive predictive value of 87.7%. The negative predictive value for gliomas was 76.6%. 18 suspected gliomas were other lesions including lymphomas (33.3%), metastases (27.8%), plexus papillomas (5.6%), encephalitis (5.6%) and other none neoplastic lesions (27.8%). In 103 of 330 patients the MRI dependent diagnosis revealed a lymphoma and was confirmed in 66 cases with a positive predictive value of 64.1% and a negative predictive value of 95.6%. In 37 patients the MRI-diagnosis of lymphoma histologically turned out to be glioma (51.4%), unknown demyelination (27%), encephalitis (18.9%), and metastasis (2,7%).

Conclusions: It is fact that the suspected diagnosis of brain tumors based on MRI interpretation may be incongruent to the histological results, but till now it is rather unknown which exact error rates exist. In this study, we can show the actual positive and negative predictive value of MRI determined diagnosis “glioma” and “lymphoma” according to current clinical standards. In conclusion, we can exclude with a high probability a lymphoma, but the accuracy of the right diagnosis is low. In more than half of these cases a glioma is found. In patients with the suspected diagnosis of a glioma, it is vice versa and was most commonly confused with lymphomas or brain metastases.