| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 350.87 KB | Adobe PDF |
Advisor(s)
Abstract(s)
Healthy controls reference values (RV) of DatScanTM and its semi-quantification (SQ) adapted to Infanta Cristina’s Hospital in Badajoz were created. It was used with a sample of 120 DaTScanTM tests divided into 4 groups with n = 30. The first one of healthy controls (GI) was used for calculation of RV. It was applied a semiautomatic method for segmentation and posterior calculi of the specific uptake ratios (SUR), respectively: Left and Right Caudate Nucleus/Occipital (A) and (B); Left and Right Putamen/Occipital (C) and (D); Striatum/Occipital (E); Left and Right Striatum/Occipital (F) and (G); Putamen/Caudate Nucleus (H). The second and third groups (healthy group (GIIH) and pathological group (GIIP)) were compared to GI for validation proposes. Control chards and ROC curves (ROCC) were obtained. The fourth group of randomized patients was used to evaluate the ability of SQ classification and it was compared with the physician evaluation. Sensibility (S), Specificity (SP), Positive (PPV) and Negative (NPV) predictive values were calculated. RV ( ̅ ): A (2.60±0.40); B (2.57±0.36); C (2.29±0.36); D (2.31±0.35); E (2.44±0.35); F (2,44±0,37); G (2,44±0,34); H (0.89±0.07). The GIIH values are above the RV and the GIIP group below the RV. The AUC, S and SP values for each SUR obtained by ROCC were: A (0.805; 0.765; 0.846); B (0.787; 0.711; 0.864); C (0.907; 0.853; 0.962); D (0.933; 0.930; 0.933); E (0.884; 0.871; 0.897); F (0.860; 0.800; 0.920); G (0.868; 0.844; 0.893); H (0.866; 0.732; 1.00). Were obtained results for PPV=0.466; NPV=1; S=1 and SP=0.65. The SQ combined with the visual analysis is a good method to detect healthy patients.
Description
Este trabalho foi financiado pelo Concurso Anual para Projetos de Investigação, Desenvolvimento, Inovação e Criação Artística (IDI&CA) 2016 do Instituto Politécnico de Lisboa. Código de referência: IPL/2016/CardiaCor_ESTeSL
Keywords
DaTScan TM Semi-quantification Reference values Accuracy Precision Parkinson disease IPL/2016/CardiaCor_ESTeSL
Citation
Queiroga M, Elias M, Rayo JI, Carolino E, Vieira L, Sousa E, et al. Influence of semiquantification in DATSCANTM studies for diagnosis of Parkinsonian syndromes. In: SYMCOMP 2017, 3rd International Conference on Numerical and Symbolic Computation: developments and applications. APMTAC – Associação Portuguesa de Mecânica Teórica, Aplicada e Computacional; 2017. p. 271-81.