NIMG-46. IMPACT OF FET PET ON MULTIDISCIPLINARY NEUROONCOLOGICAL TUMOR BOARD DECISIONS IN PATIENTS WITH BRAIN TUMORS

Abstract

BACKGROUND: Following neurooncological treatment of brain tumors, neurooncologists are frequently confronted with equivocal MRI findings (e.g., treatment-related changes, nonmeasurable (speckled) contrast-enhancing lesions, increase of T2/FLAIR signal alterations, pseudoresponse). Especially in Europe, amino acid PET is increasingly being integrated into multidisciplinary neurooncological tumor boards (MNTB) to overcome these diagnostic uncertainties as well as to improve patient management. We here evaluated the correctness of MNTB decisions, in which amino acid PET findings were taken into account. METHODS: In a single university center, we retrospectively evaluated 114 MNTB decisions concerning 99 patients with malignant glioma (n=81) (glioblastoma, n=54; anaplastic glioma, n=26; gliosarcoma, n=1) or brain metastases (n=18) secondary to NSCLC, melanoma, breast cancer, or colorectral cancer, presenting with equivocal MRI findings following neurooncological treatment. All patients underwent amino acid PET imaging using O-(2-[18F]-fluoroethyl)-L-tyrosine (FET) as an adjunct for decision-making. Additionally, the patients' clinical status, pretreatment, and multimodal MRI findings were considered for decision-making. Presence of neoplastic tissue in PET was considered if the mean FET uptake as assessed by tumor-to-brain ratios was >2. The decisions' diagnostic performance was evaluated by 2x2 contingency tables using the neuropathological results or clinicoradiological follow-up. RESULTS: In the majority of MNTB decisions (n=102; 89%), FET PET results were integrated into the decision-making with considerable impact on the clinical management. In particular, 85% of MNTB decisions (n=87) prompted a treatment change (i.e., resection, radiotherapy, chemotherapy, or combinations thereof, as well as palliative therapy), or, in the case of suspected treatment-related changes, the continuation of the initial treatment regimen (15%; n=15). The MNTB decisions were validated using neuropathological data in 38% (n=39) or clinicoradiological information in 62% (n=63) and yielded a diagnostic accuracy of 88% (sensitivity, 89%; specificity, 75%; P=0.008). CONCLUSIONS: Our results suggest that the integration of FET PET derived information significantly aids MNTB decisions.