Eike Steidl1,2,3, Mathias Müller4, Andreas Müller4, Ulrich Herrlinger5, Elke Hattingen6,4,7. 1. Institute of Neuroradiology, University Hospital Frankfurt, Schleusenweg 2-16, 60528, Frankfurt, Germany. eike.steidl@kgu.de. 2. Department of Radiology, Neuroradiology, University Hospital Bonn, Sigmund-Freud-Str. 25, 53127, Bonn, Germany. eike.steidl@kgu.de. 3. German Cancer Consortium (DKTK) and German Cancer Research Center (DKFZ), Heidelberg, Germany. eike.steidl@kgu.de. 4. Department of Radiology, Neuroradiology, University Hospital Bonn, Sigmund-Freud-Str. 25, 53127, Bonn, Germany. 5. Division of Clinical Neurooncology, Department of Neurology, University Hospital Bonn, Sigmund-Freud-Str. 25, 53127, Bonn, Germany. 6. Institute of Neuroradiology, University Hospital Frankfurt, Schleusenweg 2-16, 60528, Frankfurt, Germany. 7. German Cancer Consortium (DKTK) and German Cancer Research Center (DKFZ), Heidelberg, Germany.
Abstract
PURPOSE: Dynamic susceptibility contrast (DSC) MR-perfusion is becoming a standard of care for the monitoring of glioblastoma. Yet, technical standards are lacking and measurements without leakage correction are still common. Also, data on leakage corrected measurements during stable disease is scarce. In this study we hypothesized that basic leakage correction would significantly enhance data quality during stable disease and improve progress detection. We furthermore investigated whether longitudinal data could increase diagnostic performance. METHODS: Patients with histologically proven glioblastoma undergoing first-line therapy were prospectively recruited. We conducted DSC perfusion measurements without prebolus administration in 6-week intervals from the end of radiotherapy until progression. Maximum relative cerebral volume values (rCBVmax) with and without leakage correction were calculated using Philips IntelliSpace®. RESULTS: We recruited 16 patients and conducted 82 MRI scans with a mean follow up of 7.2 month. During stable disease, corrected rCBVmax was significantly more stable than uncorrected rCBVmax. Detection of progression with a rCBVmax cutoff was better for corrected (specificity 86%) than for uncorrected rCBVmax (specificity 41%). Interestingly, the increase of corrected rCBVmax upon progression also had a good diagnostic performance with a combination of both cutoffs delivering the best result (sensitivity/specificity 89%/93%). CONCLUSION: Corrected rCBVmax supports the imaging finding of a stable disease and large increases during longitudinal observation support the diagnosis of tumor progression. rCBV values without prebolus or leakage correction are not reliable to monitor glioblastomas. Further studies to investigate the value of longitudinal rCBV dynamics for the differentiation of real tumor progression from pseudoprogression are warranted.
PURPOSE: Dynamic susceptibility contrast (DSC) MR-perfusion is becoming a standard of care for the monitoring of glioblastoma. Yet, technical standards are lacking and measurements without leakage correction are still common. Also, data on leakage corrected measurements during stable disease is scarce. In this study we hypothesized that basic leakage correction would significantly enhance data quality during stable disease and improve progress detection. We furthermore investigated whether longitudinal data could increase diagnostic performance. METHODS:Patients with histologically proven glioblastoma undergoing first-line therapy were prospectively recruited. We conducted DSC perfusion measurements without prebolus administration in 6-week intervals from the end of radiotherapy until progression. Maximum relative cerebral volume values (rCBVmax) with and without leakage correction were calculated using Philips IntelliSpace®. RESULTS: We recruited 16 patients and conducted 82 MRI scans with a mean follow up of 7.2 month. During stable disease, corrected rCBVmax was significantly more stable than uncorrected rCBVmax. Detection of progression with a rCBVmax cutoff was better for corrected (specificity 86%) than for uncorrected rCBVmax (specificity 41%). Interestingly, the increase of corrected rCBVmax upon progression also had a good diagnostic performance with a combination of both cutoffs delivering the best result (sensitivity/specificity 89%/93%). CONCLUSION: Corrected rCBVmax supports the imaging finding of a stable disease and large increases during longitudinal observation support the diagnosis of tumor progression. rCBV values without prebolus or leakage correction are not reliable to monitor glioblastomas. Further studies to investigate the value of longitudinal rCBV dynamics for the differentiation of real tumor progression from pseudoprogression are warranted.