Diagnosis of uveal malignant melanoma by a new semiquantitative assessment of N -isopropyl- p -[ 123 I]-iodoamphetamine

Purpose To semiquantify the uptake of N -isopropyl- p -[ 123 I]-iodoamphetamine (I-123 IMP) in patients with uveal malignant melanoma reproducibly and objectively. Methods Fifty-two patients were examined. Twenty-nine patients had malignant melanoma (group A), three were clinically diagnosed with metastatic choroidal tumor, and 20 patients were given either histological or clinical diagnoses of either benign pigmented lesion or tumor (group B). Early and delayed I-123 IMP images were obtained and standardized by Neurostat software. Using fusion software, we applied a three-dimensional region of interest (3D-ROI) template to the standardized I-123 IMP, and calculated the retention index and tumor-to-nontumor (T/N) ratio of the delayed phase using the maximum count for each ROI. Results Sensitivity at the retention index cutoff of 30 was 82.4%, specificity was 85.2%, and accuracy was 83.6%. Sensitivity at the T/N ratio of the delayed phase cutoff of 1.3 was 91.2%, specificity was 77.8%, and accuracy was 85.2%. The positive predictive value of the T/N ratio was better than that of the retention index. The negative predictive value of the retention index was better than that of the T/N ratio. Conclusion This new semiquantitative estimation method is reproducible and objective, especially when the examinations are performed repeatedly for evaluation of both therapy and follow-up.