Optimization of diagnosis and follow-up of differentiated thyroid cancer

Abstract

We studied the value of thyroid nodule (TN) elastography in determining whether a TN is malignant. The results of twenty studies, compromising 3973 nodules, were included. The results showed a pooled sensitivity and specificity of 85% and 80%, respectively. The pooled NPV and PPV were 97% and 40%. From this we concluded that if a nodule appeared soft, the chance that it would be malignant are negligible and FNA can be omitted. The use of elastography has been studied as well by others in patients that underwent FNA of which the result was indeterminate. This analysis resulted in a sensitivity of 69%, a specificity of 75%, and a NPV and PPV of 82 and 63%. Potential causes of the difference with the results in are the low number of included studies and nodules and heterogeneity of the studies. If a mutation in the BRAF gene is found, the nodule is malignant. We showed that BRAF analysis on nodules with a Bethesda 3 or 4 FNA result is cost-effective in the Netherlands. We determined that Rapid Onsite Adequacy Assessment and multiple needle passes reduce the percentage of nondiagnostic FNA results. Part 2 The combination of 124I and 18F-FDG PET/CT is promising in determining which patients suspected of DTC recurrence are likely to benefit from 131I therapy, which was the basis of the Thyropet study. For this study 124I PET/CT scanning had to be standardized among participating centers. We calibrated 18 participating scanners. A phantom containing vials with increasing activities of 124I was scanned. After calibration the relative errors of the calibrated activities was smaller than for the measured activities. This showed that our procedure improved quantification per scanner and made the cohort of scanners more accurate. The Thyropet study was preliminary terminated after inclusion of 17 patients the study, based on the predefined stopping rule. Of the nine positive post-therapy WBS, five pre-therapeutic 124I PET/CT scans were disconcordant. Implementation of 124I PET in this setting would thus have led to 47% (8/17) less futile 131I treatments, but 29% of patients (5/17) would have been denied the only potential curative therapy. A potential cause of the false negative 124I scans is the difference in administered activity of 124I (usually around 74 MBq) versus the administered therapeutic activity 131I (usually around 7400 MBq). On the other hand, PET has a higher sensitivity in comparison to SPECT. We showed with a phantom study that the relatively low activity used for 124I PET/CT scanning, around 1% of 131I activity, is sufficient for small spheres in optimal PET reconstruction settings. However, false-negative 124I PET/CT as compared with the post-therapy 131I SPECT/CT scans in larger lesions (>10 mm) and for no-PSF and no TOF PET might be caused by a difference in detectability.