EJNMMI Physics

Table 3 Performance of the amyloid status classification deep learning model

From: Estimation of brain amyloid accumulation using deep learning in clinical [¹¹C]PiB PET imaging

	PET/CT		PET/MRI
	Validation n = 872*	Hold-out test n = 300	Hold-out test n = 100
Cohen’s kappa	0.96	0.97	0.96
Accuracy	98%	99%	98%
Sensitivity	98%	99%	98%
Specificity	97%	97%	98%
F1-score	98%	99%	98%

*The validation subjects represent the sum of all five validation folds (n = 872)

Back to article page