Artikel
Textural features of high-grade glioma in pretherapeutic [18F]-FET-PET are correlated with tumor grade and survival
Suche in Medline nach
Autoren
-
Thomas Pyka
- Nuklearmedizinische Klinik, Klinikum rechts der Isar, Technische Universität München
-
Jens Gempt
- Neurochirurgische Klinik und Poliklinik, Klinikum rechts der Isar, Technische Universität München
-
Florian Ringel
- Neurochirurgische Klinik und Poliklinik, Klinikum rechts der Isar, Technische Universität München
-
Hans-Jürgen Wester
- Nuklearmedizinische Klinik, Klinikum rechts der Isar, Technische Universität München
-
Markus Schwaiger
- Nuklearmedizinische Klinik, Klinikum rechts der Isar, Technische Universität München
-
Stefan Förster
- Nuklearmedizinische Klinik, Klinikum rechts der Isar, Technische Universität München
| Veröffentlicht: | 2. Juni 2015 |
|---|
Gliederung
Text
Objective: [18F]-fluoroethyl-L-tyrosine (FET) is a well-established radiotracer in the therapy planning and therapy monitoring of malignant brain tumors. FET-PET analysis using tumor-to-background ratios (TBR) has been shown to be of high value for the detection of viable brain tumor tissue, however it is limited regarding the grading of brain tumors. Recently, textural features in FDG-PET have been proposed as a method to quantify heterogeneity of tracer uptake in a variety of tumor entities. With the present study we evaluate the value of textural FET-PET features for sub-grading and prognostication of survival in patients with high-grade gliomas.
Method: 113 patients (70 m; 43 f) with high-grade gliomas were included in this study. All patients received a FET-PET scan prior to first line therapy. Static FET-PET images were evaluated using Matlab (MathWorks, Inc.; Image Processing Toolbox and own code). Textural FET uptake parameters based on gray level co-occurence matrices and gray level neighborhood difference matrices were derived. Student’s t-test and ROC analysis were used for assessing the usefulness for tumor grading; Univariate and multivariate cox regression were employed for survival analysis.
Results: Textural features of FET uptake were correlated significantly with tumor grade (III, IV; p<0.01). ROC analysis showed an area under the curve of 0.80 (p<0.001) for differentiating grade III from grade IV tumors, while standard analysis based on TBRmax and metabolic tumor volume (MTV) showed no statistically significant results. Furthermore, a correlation of FET-PET texture patterns with survival in all patients and in the Glioblastoma subgroup was revealed (p=0.001; HR 5.3, CI 2.0 - 14.0), which also proved to be statistical significant in multivariate analysis.
Conclusions: Measurement of tracer uptake heterogeneity in pretherapeutic FET-PET using textural features proved valuable for the sub-grading of high-grade glioma as well as prediction of patient survival. In contrast, standard FET-PET analysis parameters such as TBR and tumor volume were not useful for this purpose in patients with gliomas.
