Hypometabolic gliomas on FET-PET—is there an inverted U-curve for survival?

See the article by Galldiks et al. in this issue, pp. 1331–1338. The landscape of oncology was permanently changed by [18F]-fluorodeoxyglucose (FDG) positron emission tomography (PET), which enabled clinicians to link a structural abnormality with a metabolic signature. Unfortunately, FDG-PET has not had a major impact in neuro-oncology, largely because the high background glucose uptake in normal brain hinders its sensitivity to assess central nervous system cancers. In an effort to develop a better neuro-oncological PET technique, more than three decades have been devoted to research on amino acid radiotracers. These promise better contrast compared with glucose, as amino acid uptake is much lower in normal brain, yet elevated in neoplasms due to increased transport and utilization.1,2 Several amino acid tracers have been studied with intricate differences in half-life (eg, 11C vs 18F), transport (L-type amino acid transporter 1 and 2, the alanine/serine/cysteine transport system), and metabolic fate (eg, trapping vs efflux).2 Through a concerted effort of primarily European imaging centers, O-(2-[18F]-fluoroethyl)-L-tyrosine (FET) has become one of the most studied to date. FET-PET is now widely considered a clinically viable adjunct to MRI in neuro-oncology,1 although it still has yet to find its way into clinical practice in the United States of America. FET-PET is a highly sensitive technique for brain tumor detection. It allows one to quantify the degree, uptake dynamics, and volume of metabolically active tumor, which can be used to improve clinical planning and treatment monitoring.1–4 It is intuitive to consider higher metabolic activity a feature of more aggressive tumors. This notion was supported by the higher FET uptake found in high-grade gliomas and by the description of increase in FET uptake during malignant transformation of low-grade gliomas.4–6 The picture became more nuanced when overlapping static FET-PET patterns were described in high- and low-grade glioma, or when oligodendroglial tumors were confirmed to have higher amino acid tracer uptake despite their survival advantage.6,7 Thus, the “less uptake is more survival” model had been questioned and refined over the years. While most attention has been directed at high-grade tumor and signal elevations, there is a shortage of studies focusing on the 20–30% of low-grade gliomas with low or no FET uptake at all.5,7,8 Efforts to better understand exceptions such as FET negativity are crucial for the maturation of FET-PET as an imaging technique, and to allow it to gain more traction globally. The featured article from Galldiks, Unterrainer, and colleagues takes a step in this direction. Galldiks et al describe their retrospective analysis of the largest cohort to date of FET-PET negative gliomas.9 They identified 100 treatment-naïve gliomas with isometabolic (“indifferent”) or hypometabolic (“photopenic”) FET signal relative to normal-appearing brain. The photopenic group contained 40 patients with predominantly grade II and III astrocytomas. Interestingly, on quantitative analysis they found photopenic grade III gliomas to have lower uptake compared with grade II, inferring that photopenia may be associated with a more aggressive histopathological phenotype. They then compared progression-free survival (PFS) between photopenic and isometabolic grade II astrocytomas in a group of patients followed without treatment (a common management approach at the time) and then in isocitrate dehydrogenase–mutant tumors regardless of treatment history. PFS was 2-fold shorter in photopenic compared with isometabolic gliomas in both comparisons. Multivariate analyses corroborated photopenia to be a predictor of poorer PFS independent of age, performance status, and measurements by Response Assessment in Neuro-Oncology. The above findings are seemingly in conflict with the present literature’s almost unanimously reported PFS advantage of FET-negative over FET-positive gliomas.5–8 Unfortunately, a retrospective comparison would be very difficult as the published studies contain small sample sizes, inhomogeneous diagnoses and molecular information, heterogeneous treatments, variable criteria for progression, and inconsistent definition of FET negativity. However, it is conceivable that these results fit together. First of all, the FET photopenic group may represent 10% of all gliomas and about a third of those FET negative.7 Thus, if all FET-negative tumors are pooled together as in the other studies, the PFS signal from the photopenic group would likely be masked by those FET isometabolic. The survival discrepancy could be explained with an inverted U-shaped curve, where isometabolic lesions represent the “sweet spot” and both photopenic and FET-positive tumors are on the adverse end of the PFS spectrum (Fig. 1). Unfortunately, none of the published studies cover the entire spectrum and thus the inverted U-curve remains an untested hypothesis. What is known is that those in the FET-negative group, likely dominated by isometabolic lesions, do better than those FET positive,5,6,8 and that isometabolic tumors (at least grade II astrocytomas) do better than those photopenic.9 Malignant transformation is associated with increased signal in FET-positive gliomas.4 This was also confirmed by the longitudinal study of FET-negative low-grade gliomas, where malignant transformation was detected by interval increase in FET uptake in 94% of cases.5 Even if photopenic gliomas had substantially different biology, breakdown of the blood‒brain barrier and subsequent gadolinium-contrast enhancement on MRI could cause increased FET signal as well. Thus, photopenia is suggested to be a rare and transient state. It may represent an early stage prior to the impending transformation of the tumor, similar to when a de novo glioblastoma is diagnosed before it developed radiographic features of malignancy such as contrast enhancement and necrosis. A hypothetical inverse U-curve of PFS and FET uptake. The rainbow scale replicates visual palettes used in PET, where dark blue represents the low end of the uptake spectrum. Overall, the most important aspect of the featured article is to raise awareness that photopenia on FET-PET may suggest a less indolent course for gliomas. It also raises numerous interesting questions, such as what causes the photopenia? Is this phenomenon specific to FET-PET? Can the PFS disadvantage of photopenic grade II astrocytomas be confirmed in the present era of aggressive intervention and does the inverted U-curve exist? But most importantly, it has yet to be learned whether the phenomenon discussed has any effect on overall survival or is predominantly radiographic in nature. Answering these questions will play an important role in determining the clinical utility of amino acid PET as a valuable clinical radiographic biomarker in the management of brain tumors.