Automatic arterial input function selection in CT and MR perfusion datasets using deep convolutional neural networks

TY - JOUR

T1 - Automatic arterial input function selection in CT and MR perfusion datasets using deep convolutional neural networks

AU - Winder, Anthony

AU - d’Esterre, Christopher D.

AU - Menon, Bijoy K.

AU - Fiehler, Jens

AU - Forkert, Nils D.

N1 - © 2020 American Association of Physicists in Medicine.

PY - 2020/9/1

Y1 - 2020/9/1

N2 - Purpose: The computation of perfusion parameter images requires knowledge of the arterial blood flow in the form of an arterial input function (AIF). This work proposes a novel method to automatically identify AIFs in computed tomography perfusion (CTP) and dynamic susceptibility contrast perfusion-weighted MRI (PWI) datasets using a deep convolutional neural network (CNN). Methods: One-hundred CTP and 100 PWI datasets of acute ischemic stroke patients were available for model development and evaluation. For each modality, 50 datasets were used for CNN training and 20 for validation using manually selected AIFs and non-arterial tissue concentration time curves. Model evaluation was performed using the remaining 30 independent validation datasets from each modality with manual AIF selections provided by two experts as ground truth. Additionally, AIFs were also extracted using an established automatic shape-based algorithm for comparison purposes. The extracted AIFs were compared using normalized cross-correlation and shape features as well as using the Dice similarity metric and volume of the corresponding hypoperfusion (Tmax ' 6 s) lesions. Results: The cross-correlation values comparing the manual AIFs and those extracted by the proposed CNN method were significantly greater than those comparing the manual AIFs to the shape-based comparison method. Likewise, hypoperfusion lesions generated using the manually selected AIFs and CNN-based AIFs showed higher Dice values compared to hypoperfusion lesions generated using the comparison AIF extraction method. Shape features for AIFs generated by the proposed method did not differ significantly from the manual AIFs, with the exception that the CNN-derived AIFs for the PWI datasets showed marginally greater peak heights. Conclusion: Deep convolutional neural network models are viable for the automatic extraction of the AIF from CTP and PWI datasets.

AB - Purpose: The computation of perfusion parameter images requires knowledge of the arterial blood flow in the form of an arterial input function (AIF). This work proposes a novel method to automatically identify AIFs in computed tomography perfusion (CTP) and dynamic susceptibility contrast perfusion-weighted MRI (PWI) datasets using a deep convolutional neural network (CNN). Methods: One-hundred CTP and 100 PWI datasets of acute ischemic stroke patients were available for model development and evaluation. For each modality, 50 datasets were used for CNN training and 20 for validation using manually selected AIFs and non-arterial tissue concentration time curves. Model evaluation was performed using the remaining 30 independent validation datasets from each modality with manual AIF selections provided by two experts as ground truth. Additionally, AIFs were also extracted using an established automatic shape-based algorithm for comparison purposes. The extracted AIFs were compared using normalized cross-correlation and shape features as well as using the Dice similarity metric and volume of the corresponding hypoperfusion (Tmax ' 6 s) lesions. Results: The cross-correlation values comparing the manual AIFs and those extracted by the proposed CNN method were significantly greater than those comparing the manual AIFs to the shape-based comparison method. Likewise, hypoperfusion lesions generated using the manually selected AIFs and CNN-based AIFs showed higher Dice values compared to hypoperfusion lesions generated using the comparison AIF extraction method. Shape features for AIFs generated by the proposed method did not differ significantly from the manual AIFs, with the exception that the CNN-derived AIFs for the PWI datasets showed marginally greater peak heights. Conclusion: Deep convolutional neural network models are viable for the automatic extraction of the AIF from CTP and PWI datasets.

KW - ischemic stroke

KW - machine learning

KW - perfusion imaging

KW - Magnetic Resonance Imaging

KW - Neural Networks, Computer

KW - Brain Ischemia

KW - Algorithms

KW - Perfusion

KW - Humans

KW - Stroke/diagnostic imaging

KW - Tomography, X-Ray Computed

UR - http://www.scopus.com/inward/record.url?scp=85088094484&partnerID=8YFLogxK

U2 - 10.1002/mp.14351

DO - 10.1002/mp.14351

M3 - SCORING: Journal article

C2 - 32583617

AN - SCOPUS:85088094484

VL - 47

SP - 4199

EP - 4211

JO - MED PHYS

JF - MED PHYS

SN - 0094-2405

IS - 9

ER -