Amino acid PET in neuro-oncology: applications in the clinic

KEYWORDS:

• Dynamic FET PET
• glioma
• brain metastasis
• radiomics
• hybrid PET/MR

For more than three decades, radiolabeled amino acids have been used in the field of neuro-oncology. The most experience for this class of positron emission tomography (PET) tracers has been gained with ¹¹C-methyl-l-methionine (MET). MET is labeled with the positron-emitting isotope carbon-11, which has a short half-life of 20 min. Thus, the use of MET is restricted to PET centers with an on-site cyclotron unit. This prompted the development of amino acid tracers labeled with positron emitters that have longer half-lives. More recently, O-(2-[¹⁸F]fluoroethyl)-l-tyrosine (FET) was developed and is a ¹⁸F-labeled amino acid tracer with a half-life of 110 min, resulting in a higher practicability compared to MET. The use of FET has grown rapidly in recent years, especially in Western Europe, and has led to a replacement of MET by the more convenient FET. The improved availability has led to several thousand FET PET scans being performed in some centers. Furthermore, clinical results with PET using FET and MET seem to be comparable [1]. The ¹⁸F-labeled amino acid analogue 3,4-dihydroxy-6-[¹⁸F]-fluoro-l-phenylalanine (FDOPA) – primarily developed to evaluate dopamine synthesis in patients with movement disorders – is also increasingly being used for brain tumor imaging [2]. However, in the USA, the standard tracer for tumor imaging ¹⁸F-2-fluoro-2-deoxy-d-glucose (FDG) PET is still frequently used in brain tumor patients, although the evaluation of brain tumors using FDG is difficult because levels of glucose metabolism in healthy brain parenchyma are usually high. This leads to a poor tumor-to-background contrast compared with amino acid tracers [3].

Structural contrast-enhanced magnetic resonance imaging (MRI) is currently the method of first choice in neuro-oncological practice and is incorporated into resection and radiotherapy planning, evaluation of treatment response and diagnosis of disease progression. Furthermore, this tool has a high availability, lower cost than PET, and an excellent spatial resolution, particularly in the light of commercially available (ultra-) high-field scanners at 3 T or more. However, structural MRI has its limitations and there are several advantages of amino acid PET for brain tumor imaging. In detail, the clearly higher specificity for the detection of neoplastic tissue may help to prevent inconclusive, nondiagnostic biopsies. Furthermore, an improved planning of neurosurgical resection and radiotherapy to the true extent of the tumor may help to spare healthy tissue. Finally, the possibility to diagnose treatment-related changes can avoid a potentially harmful and unnecessary overtreatment, and, an earlier assessment of response to a certain treatment allows preventing unnecessary side effects of an ineffective treatment option.

1. Classical indications

Most common and highly important diagnostic challenges in clinical neuro-oncology are the identification of neoplastic tissue for differential diagnosis, the exact delineation of tumor extent for planning of further diagnostic or therapeutic procedures, and the differentiation of tumor progression from treatment-related changes. In the recent past, these challenges have been evaluated using amino acid PET. It has been demonstrated that this technique adds value to conventional MRI by providing additional diagnostic information in specific situations, including detection of tumor tissue when MRI data are inconclusive [4], delineation of the extent of disease, particularly in non-enhancing gliomas [5], and distinguishing radiation necrosis and pseudoprogression from disease progression [6,7]. Based on these observations, the Response Assessment in Neuro-Oncology (RANO) Working Group has recently analyzed the clinical role of amino acid PET in the diagnostic assessment of brain tumors and strongly recommended the additional use of amino acid PET at every stage of brain tumor management [3]. Additionally, in that guideline, the RANO group advocated PET imaging with superiority of amino acid PET over FDG PET.

2. Newer indications and recent developments

More recently, the additional diagnostic value of amino acid PET has also been successfully described for treatment monitoring of standard treatment options such as chemoradiation [8], alkylating chemotherapy [9], and antiangiogenic therapy [10]. In these studies, predominantly a decrease of the metabolically active volume indicated a response to treatment, and, moreover, a prediction of favorable outcome.

Newer treatment options such as immunotherapy in patients with glioblastoma or brain metastasis by checkpoint inhibitor blockade using drugs such as ipilimumab, nivolumab, or pembrolizumab appear to be promising [11]. Clinical trials are currently ongoing. Importantly, immunotherapy treatment monitoring may be impeded by the phenomenon of pseudoprogression. Recently, the immunotherapy RANO Working Group recommended criteria for problem solution [12]. However, there is still a need for the acquisition of additional diagnostic information. A small pilot study showed for the first time the potential of amino acid PET using FET to identify pseudoprogression in patients with brain metastasis originating from melanoma treated with immune checkpoint modulators [13]. Other newer treatment options are targeted therapies predominantly against brain metastasis and tumor-treating fields used in combination with adjuvant temozolomide chemotherapy after chemoradiation for glioblastoma patients. Up to now, the role of amino acid PET for the latter treatment options as well as for immunotherapy is hitherto still unclear; in order to evaluate its value, further studies are necessary.

Related to improved systemic treatment options for extracranial tumors such as immunotherapy or targeted therapy resulting in an improvement of overall survival, the incidence of brain metastasis is steadily increasing and in parallel the number of amino acid PET scans in this group of patients. Particularly after radiosurgery of brain metastasis, practitioners are often confronted with the diagnostic challenge that structural MRI cannot reliably differentiate radiation-induced changes (e.g. radiation necrosis) from brain metastasis progression. There is a growing body of evidence suggesting that amino acid PET has significant advantages over structural MRI [14–16] and perfusion-weighted MRI in this issue [17].

Furthermore, in the context of brain metastasis heterogeneity, the biological characterization of newly diagnosed and therapy-naive brain metastasis is also of great interest, particularly if the primary tumor has an unknown origin. Unterrainer and colleagues found that brain metastasis originating from non-small cell lung cancer always show FET tracer uptake, even if metastases are very small (diameter, 0.5–1.0 cm) [18]. Furthermore, melanoma metastasis may show an extremely high FET uptake.

Several studies have shown that analysis of amino acid kinetics may provide additional biological information beyond that of static parameters. This has been shown especially for FET and the carbon-11 labeled amino acid α-[¹¹C]methyl-l-tryptophan [6,7,19]. It could be demonstrated that this observation may be especially helpful for prognostication, which is of great importance for patient counseling. Dynamic FET PET studies suggest that newly diagnosed low-grade and high-grade glioma patients with time–activity curves characterized by an early peak of tracer uptake followed by a constant decent in the PET scan prior to therapy may have a poor prognosis [20,21].

From the methodological point of view, recent publications in the field of neuro-oncology highlight the value of radiomics by decoding the tumor phenotype using noninvasive imaging. Radiomics enables the high-throughput extraction of a large amount of quantitative features from medical images of a given modality (usually MRI), providing a comprehensive quantification of the tumor phenotype, based on simple medical imaging at low cost [22]. One concept of radiomics is the use of textural feature analysis as a tool that objectively and quantitatively describes one of the intrinsic properties of cancer, particularly heterogeneity. For FET PET, it could be demonstrated for the first time that radiomic textural feature analysis provides noninvasively quantitative information about brain metastasis heterogeneity and, moreover, may be helpful for the distinction between radiation injury and disease progression [23]. Thus, radiomics provides novel imaging biomarkers that could be potentially helpful in the field of neuro-oncology.

Another methodological innovation is the availability of hybrid PET/MR scanners, allowing the simultaneous acquisition of both imaging modalities. For example, the acquisition of dynamic FET PET, contrast-enhanced structural MRI, MR spectroscopy, and other advanced MRI sequences such as perfusion-weighted MRI in a single session can easily be performed. Moreover, this technology helps besides optimizing co-registration of various imaging modalities to increase convenience for patients, reduces scanning time, and avoids exposition to the additional radiation dose associated with a PET/CT scan. Therefore, this technology appears particularly attractive in children. From the researcher’s point of view, this technology provides the optimal requirements for comparative imaging studies using amino acid PET and advanced MRI, ideally with histological confirmation of imaging findings by stereotactic biopsy. However, the problem of MRI-based attenuation correction is currently not fully solved [24] and the clinical value of adding advanced MRI sequences in multiparametric imaging setting has to be confirmed in further studies [25,26].

3. Conclusion

The present literature provides strong evidence that amino acid PET can be of great value for a broad spectrum of indications in the field of neuro-oncology. The diagnostic improvement probably results in relevant benefits for brain tumor patients and justifies a more widespread use of this diagnostic tool. Furthermore, the necessary PET infrastructure is widely available, and the production of radiolabeled amino acids is well established with comparable costs to FDG. Any additional costs of this diagnostic technique can be potentially saved by the incurred costs of less reliable diagnostic imaging techniques [27].