Arian Lasocki MBBS, FRANZCR, Presenter: Nothing to Disclose

Alpha Tsui, Abstract Co-Author: Nothing to Disclose

Mark Tacey, Abstract Co-Author: Nothing to Disclose

Kate Drummond, Abstract Co-Author: Nothing to Disclose

Kathryn Field, Abstract Co-Author: Nothing to Disclose

Frank Gaillard MBBS, Abstract Co-Author: Founder, Radiopaedia.org CEO, Radiopaedia.org Editor, Radiopaedia.org Support, UBM Medica LLC Support, Toshiba Corporation

Grading of intracranial astrocytomas using histopathology alone is affected by sampling error and inter- and intra-observer variability. Under-grading can result in less aggressive therapy and potentially a worse prognosis. We propose that incorporating MRI into grading will predict patient survival better than the current gold standard of histopathology alone.

Patients with a new diagnosis of a WHO grade II-IV fibrillary astrocytoma or oligoastrocytoma were identified through the ACCORD neuro-oncology database of The Royal Melbourne Hospital. Pre-operative MRIs performed between September 2007 and December 2010 were independently reviewed on PACS by two readers, blinded to the histological grade, and an MRI grade was given. The grade was assigned primarily on the basis of the post-contrast appearances, with supplementary information from both standard and advanced sequences. The MRI and histopathological grades were compared against patient survival, adjusted for patient age.

A total of 245 patients met the inclusion criteria. Correlation between the two MRI readers was high, at 95% (kappa 0.87). Correlation between the MRI consensus grade and the histological grade was moderate, at 82% (kappa 0.58). Patients with MRI appearances consistent with a grade IV tumour but lower grade (II or III) histology had significantly worse survival than patients with the same histology but lower grade MRI appearances (p = 0.001 for grade II histology and p = 0.013 for grade III). Taken as a group, the survival of all these patients up-graded from lower grade histology to grade IV based on MRI was equivalent to those patients with grade IV tumours on both histology and MRI (no significant difference, p = 0.896). Therefore, the tumours up-graded to grade IV based on MRI behave as grade IV tumours, and at least some may truly be grade IV tumours under-graded by histology.

MRI is a better predictor of survival than histopathology for high grade gliomas, with high inter-observer agreement. Incorporating MRI into grading can therefore decrease the risk of under-grading. This has the potential to guide optimal therapy and thus substantially improve patient survival.

MRI is currently under-utilised in the management of intracranial astrocytomas. Adding MRI information to the current histopathological grading system allows more accurate grading of astrocytomas.

Lasocki, A, Tsui, A, Tacey, M, Drummond, K, Field, K, Gaillard, F, MRI Grading versus Histology: Predicting Survival of WHO Grade II-IV Astrocytomas.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13015077.html