Purpose: Reliable differential diagnosis between tumour recurrence and treatment-induced lesions is required to take advantage of new therapeutic approaches to recurrent gliomas. Structural imaging methods offer a high sensitivity but a low specificity, which might be improved by neurofunctional imaging. This study aimed to test the hypothesis that incorporation of 18-fluoro-deoxy-glucose positron emission tomography (FDG-PET) increases the accuracy of this differential diagnosis obtained with 201Tl chloride-single-photon emission computed tomography (201Tl-SPECT).
Materials and methods: Seventy-six patients (mean age 47.72 +/- 16.19 years) under suspicion of glioma recurrence, 42% with low-grade and 58% with high-grade lesions, were studied by (201)Tl-SPECT and FDG-PET, reporting results under blinded conditions using visual analysis. Tumour was confirmed by histological confirmation (23 patients) or clinical and structural neuroimaging follow-up (mean of 2.6 years).
Results: This population had a high disease prevalence (72%). Globally, highest sensitivity was obtained using 201Tl-SPECT assessed with MRI (96%) and highest specificity using FDG-PET + MRI (95%). FDG-PET appeared slightly better for confirming tumour recurrence, whereas 201Tl-SPECT was superior for ruling out possible recurrence (disease present in 38% of FDG-PET negative explorations). In the high-grade subgroup, there were no false-positive examinations (specificity: 100%), but sensitivity differed among techniques (201Tl-SPECT : 94%; 201Tl-SPECT + MRI: 97%; FDG-PET + MRI: 83%). In the low-grade subgroup, 201Tl-SPECT + MRI showed highest sensitivity (95%) and lowest posttest negative probability (9%); FDG-PET + MRI offered highest specificity (92%) with a posttest negative probability of 35%.
Conclusions: FDG-PET does not clearly improve the diagnostic accuracy of (201)Tl-SPECT, which appears to be a more appropriate examination for the diagnosis of possible brain tumour recurrence, especially for ruling it out.