TY - JOUR
T1 - Fully automated segmentation of head CT neuroanatomy using deep learning
AU - Cai, Jason C.
AU - Akkus, Zeynettin
AU - Philbrick, Kenneth A.
AU - Boonrod, Arunnit
AU - Hoodeshenas, Safa
AU - Weston, Alexander D.
AU - Rouzrokh, Pouria
AU - Conte, Gian Marco
AU - Zeinoddini, Atefeh
AU - Vogelsang, David C.
AU - Huang, Qiao
AU - Erickson, Bradley J.
N1 - Publisher Copyright:
© RSNA, 2020.
PY - 2020
Y1 - 2020
N2 - Purpose: To develop a deep learning model that segments intracranial structures on head CT scans. Materials and Methods: In this retrospective study, a primary dataset containing 62 normal noncontrast head CT scans from 62 patients (mean age, 73 years; age range, 27–95 years) acquired between August and December 2018 was used for model development. Eleven intracranial structures were manually annotated on the axial oblique series. The dataset was split into 40 scans for training, 10 for vali-dation, and 12 for testing. After initial training, eight model configurations were evaluated on the validation dataset and the highest performing model was evaluated on the test dataset. Interobserver variability was reported using multirater consensus labels obtained from the test dataset. To ensure that the model learned generalizable features, it was further evaluated on two secondary datasets containing 12 volumes with idiopathic normal pressure hydrocephalus (iNPH) and 30 normal volumes from a publicly available source. Statistical significance was determined using categorical linear regression with P, .05. Results: Overall Dice coefficient on the primary test dataset was 0.84 6 0.05 (standard deviation). Performance ranged from 0.96 6 0.01 (brainstem and cerebrum) to 0.74 6 0.06 (internal capsule). Dice coefficients were comparable to expert annotations and ex-ceeded those of existing segmentation methods. The model remained robust on external CT scans and scans demonstrating ventricular enlargement. The use of within-network normalization and class weighting facilitated learning of underrepresented classes. Conclusion: Automated segmentation of CT neuroanatomy is feasible with a high degree of accuracy. The model generalized to external CT scans as well as scans demonstrating iNPH.
AB - Purpose: To develop a deep learning model that segments intracranial structures on head CT scans. Materials and Methods: In this retrospective study, a primary dataset containing 62 normal noncontrast head CT scans from 62 patients (mean age, 73 years; age range, 27–95 years) acquired between August and December 2018 was used for model development. Eleven intracranial structures were manually annotated on the axial oblique series. The dataset was split into 40 scans for training, 10 for vali-dation, and 12 for testing. After initial training, eight model configurations were evaluated on the validation dataset and the highest performing model was evaluated on the test dataset. Interobserver variability was reported using multirater consensus labels obtained from the test dataset. To ensure that the model learned generalizable features, it was further evaluated on two secondary datasets containing 12 volumes with idiopathic normal pressure hydrocephalus (iNPH) and 30 normal volumes from a publicly available source. Statistical significance was determined using categorical linear regression with P, .05. Results: Overall Dice coefficient on the primary test dataset was 0.84 6 0.05 (standard deviation). Performance ranged from 0.96 6 0.01 (brainstem and cerebrum) to 0.74 6 0.06 (internal capsule). Dice coefficients were comparable to expert annotations and ex-ceeded those of existing segmentation methods. The model remained robust on external CT scans and scans demonstrating ventricular enlargement. The use of within-network normalization and class weighting facilitated learning of underrepresented classes. Conclusion: Automated segmentation of CT neuroanatomy is feasible with a high degree of accuracy. The model generalized to external CT scans as well as scans demonstrating iNPH.
UR - http://www.scopus.com/inward/record.url?scp=85107744115&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85107744115&partnerID=8YFLogxK
U2 - 10.1148/ryai.2020190183
DO - 10.1148/ryai.2020190183
M3 - Article
AN - SCOPUS:85107744115
SN - 2638-6100
VL - 2
SP - 1
EP - 12
JO - Radiology: Artificial Intelligence
JF - Radiology: Artificial Intelligence
IS - 5
M1 - e190183
ER -