Biom-44.18F-Dopa-Pet Thresholds For Both Newly Diagnosed And Recurrent Astrocytoma Using Pathological Grouping Of Spatially Correlated Biopsies
Deanna Pafundi,Ian Parney,Terry Burns,Caterina Giannini,Aditya Raghunathan,Christopher Hunt,Katharine Dooley,Brian Kabat,Val Lowe,Timothy Kaufmann,Jonathan Morris,Mark Zakhary,Michael Grams,Maasa Seaburg,William Breen,Jann Sarkaria,Derek Johnson,Nadia Laack,Debra Brinkmann
DOI: https://doi.org/10.1093/neuonc/noae165.0117
2024-11-29
Neuro-Oncology
Abstract:BACKGROUNDGlioma radiotherapy planning requires accurate tumor delineation. While T1- contrast-enhanced-MRI (T1-CE-MRI) is the standard, due to blood-brain-barrier breakdown dependencies, T1-CE-MRI fails to reflect the extent of glioma. Studies using amino acid PET tracers report glioma uptake up to 3cm-4cm beyond T1-CE. In this study, amino acid tracer 18 F-fluoro-dihydroxyphenylalanine ( 18 F-DOPA) uptake was correlated with histopathological features to determine tumor thresholds. METHODSUsing registered 18 F-DOPA-PET and T1-CE-MRI images in the Stealth Neuronavigation system targeting concordant and discordant regions, 181 stereotactic biopsy samples (newly-diagnosed=74; recurrent=107) were obtained from 59 enrolled astrocytoma patients across two prospective phase II clinical trials between 2010-2019. An experienced neuropathologist reviewed each sample with hematoxylin and eosin stain and assigned an alphanumeric score: mitotic activity (0=no tumor; 1=no mitosis; 2=few/scattered mitosis; 3=frequent mitosis) and cellular density (A= 75%). Samples assigned 1C-1D, 2B-2D, or 3A-3D were categorized as histopathological high-score (aggressive disease), remaining samples as low-score. Relative mean (rFDOPA_mean) tumor-to-background uptake ratios were computed for 5mm uniformly expanded volumes from each Stealth biopsy coordinate using the crescent and wedge normalization methods. Relative uptake was compared overall and separately within newly-diagnosed/recurrent disease subgroups. An optimal rFDOPA_mean threshold to delineate high- from low-score was determined using the maximal Youden index of ROC curves. RESULTSOverall, high-score samples (n=116) had higher median (IQR) rFDOPA_mean uptake than low-score samples (n=65): high-score rFDOPA:mean=1.70(1.16-2.26), low-score rFDOPA:mean=1.16(0.83-1.49), Kruskal-Wallis p<0.001. The uptake difference between high- vs. low-score samples was observed within newly-diagnosed and recurrent subgroups, but was only significant for newly-diagnosed (p<0.001). For newly-diagnosed, an rFDOPA:mean=1.49 cutoff between high- and low-score groups shows 69% accuracy, 63% sensitivity, 80% specificity (AUC=0.782). Using the wedge normalization method, rFDOPA_mean values were 0.2 higher than crescent method values. CONCLUSIONSThis work presents histologically verified 18 F-DOPA-PET thresholds for aggressive disease in newly-diagnosed astrocytoma patients. Separate thresholds will need to be determined for recurrent glioma patients.
oncology,clinical neurology