Deep learning-enabled coronary CT angiography for plaque and stenosis quantification and cardiac risk prediction: an international multicentre study

Andrew  Lin; Nipun  Manral; Priscilla  McElhinney; Aditya  Killekar; Hidenari  Matsumoto; Jacek Kwiecinski; Konrad  Pieszko; Aryabod  Razipour; Kajetan  Grodecki; Caroline  Park; Yuka Otaki; Mhairi Doris; Alan C. Kwan; Donghee Han; Keiichiro  Kuronuma; Guadalupe Flores  Tomasino; Evangelos Tzolos; Aakash  Shanbhag; Markus Goeller; Mohamed  Marwan; Heidi Gransar; Balaji K Tamarappoo; Sebastien Cadet; Stephan Achenbach; Stephen J Nicholls; Dennis T  Wong; Daniel S. Berman; Marc R Dweck; David E Newby; Michelle C Williams; Piotr J Slomka; Damini Dey

doi:0.1016/S2589-7500(22)00022-X

Deep learning-enabled coronary CT angiography for plaque and stenosis quantification and cardiac risk prediction: an international multicentre study

Andrew Lin, Nipun Manral, Priscilla McElhinney , Aditya Killekar, Hidenari Matsumoto, Jacek Kwiecinski, Konrad Pieszko, Aryabod Razipour, Kajetan Grodecki, Caroline Park, Yuka Otaki, Mhairi Doris, Alan C. Kwan, Donghee Han, Keiichiro Kuronuma, Guadalupe Flores Tomasino, Evangelos Tzolos, Aakash Shanbhag, Markus Goeller, Mohamed MarwanHeidi Gransar, Balaji K Tamarappoo, Sebastien Cadet, Stephan Achenbach, Stephen J Nicholls, Dennis T Wong, Daniel S. Berman, Marc R Dweck, David E Newby, Michelle C Williams, Piotr J Slomka, Damini Dey

Research output: Contribution to journal › Article › peer-review

Abstract / Description of output

Background: Atherosclerotic plaque quantification from coronary computed tomography angiography (CCTA) enables accurate assessment of coronary artery disease burden and prognosis. We sought to develop and externally validate a deep learning (DL) system for CCTA- derived measures of plaque volume and stenosis severity.
Methods: This was an international multicentre study of 1196 patients undergoing CCTA at 11 sites, with external evaluation of prognostic risk in 1,611 patients from the prospective SCOT- HEART trial. A novel DL convolutional neural network was trained to segment coronary plaque in 921 patients (5,045 lesions). The DL network was then applied to an independent test set, which included an external validation cohort of 175 patients (1,081 lesions) as well 50 patients (84 lesions) assessed by intravascular ultrasound (IVUS) within one month of CCTA. Thereafter, we evaluated the prognostic value of DL-based plaque measurements for fatal or nonfatal myocardial infarction (MI) in 1,611 patients from the prospective SCOT-HEART trial.
Findings: In the overall test set, there was excellent agreement between DL and expert reader measurements of total plaque volume (intraclass correlation coefficient [ICC] 0·964) and percent diameter stenosis (ICC 0·879; both p<0.0001). When compared with IVUS, there was excellent agreement for DL total plaque volume (ICC 0·949) and minimal luminal area (ICC 0·904). The average per-patient DL plaque analysis time was 5.7 seconds versus 25-30 minutes taken by experts. Over a median follow-up of 4·7 years, MI occurred in 41 patients (2·5%) from the SCOT-HEART trial. DL-based total plaque volume ≥238·5mm3 was associated with an increased risk of MI (HR 5·36, 95% CI 1·70-16·86; p=0·0042) after adjustment for the presence of DL-based obstructive stenosis (HR 2·49, 95% CI 1·07-5·50; p=0·0089) and the cardiovascular risk score (HR 1·01, 95% CI 0·99-1·04; p=0·35).
Interpretation: A novel externally validated DL system provides rapid measurements of plaque volume and stenosis severity from CCTA which agree closely with expert readers and IVUS and carry prognostic value for future MI.

Original language	English
Journal	The Lancet Digital Health
Early online date	22 Mar 2022
DOIs	https://doi.org/0.1016/S2589-7500(22)00022-X
Publication status	E-pub ahead of print - 22 Mar 2022

Access to Document

0.1016/S2589-7500(22)00022-X

Deep Learning Plaque - Lancet DH FINALAccepted author manuscript, 1.72 MB
1-s2.0-S258975002200022X-mainFinal published version, 2.44 MBLicence: Creative Commons: Attribution-NonCommercial-NoDerivatives (CC BY-NC-ND)

Cite this

Lin, A., Manral, N., McElhinney , P., Killekar, A., Matsumoto, H., Kwiecinski, J., Pieszko, K., Razipour, A., Grodecki, K., Park, C., Otaki, Y., Doris, M., Kwan, A. C., Han, D., Kuronuma, K., Tomasino, G. F., Tzolos, E., Shanbhag, A., Goeller, M., ... Dey, D. (2022). Deep learning-enabled coronary CT angiography for plaque and stenosis quantification and cardiac risk prediction: an international multicentre study. The Lancet Digital Health. https://doi.org/0.1016/S2589-7500(22)00022-X

@article{bb1898b92efc4c4ea759cdd10938234a,

title = "Deep learning-enabled coronary CT angiography for plaque and stenosis quantification and cardiac risk prediction: an international multicentre study",

abstract = "Background: Atherosclerotic plaque quantification from coronary computed tomography angiography (CCTA) enables accurate assessment of coronary artery disease burden and prognosis. We sought to develop and externally validate a deep learning (DL) system for CCTA- derived measures of plaque volume and stenosis severity. Methods: This was an international multicentre study of 1196 patients undergoing CCTA at 11 sites, with external evaluation of prognostic risk in 1,611 patients from the prospective SCOT- HEART trial. A novel DL convolutional neural network was trained to segment coronary plaque in 921 patients (5,045 lesions). The DL network was then applied to an independent test set, which included an external validation cohort of 175 patients (1,081 lesions) as well 50 patients (84 lesions) assessed by intravascular ultrasound (IVUS) within one month of CCTA. Thereafter, we evaluated the prognostic value of DL-based plaque measurements for fatal or nonfatal myocardial infarction (MI) in 1,611 patients from the prospective SCOT-HEART trial. Findings: In the overall test set, there was excellent agreement between DL and expert reader measurements of total plaque volume (intraclass correlation coefficient [ICC] 0·964) and percent diameter stenosis (ICC 0·879; both p<0.0001). When compared with IVUS, there was excellent agreement for DL total plaque volume (ICC 0·949) and minimal luminal area (ICC 0·904). The average per-patient DL plaque analysis time was 5.7 seconds versus 25-30 minutes taken by experts. Over a median follow-up of 4·7 years, MI occurred in 41 patients (2·5%) from the SCOT-HEART trial. DL-based total plaque volume ≥238·5mm3 was associated with an increased risk of MI (HR 5·36, 95% CI 1·70-16·86; p=0·0042) after adjustment for the presence of DL-based obstructive stenosis (HR 2·49, 95% CI 1·07-5·50; p=0·0089) and the cardiovascular risk score (HR 1·01, 95% CI 0·99-1·04; p=0·35). Interpretation: A novel externally validated DL system provides rapid measurements of plaque volume and stenosis severity from CCTA which agree closely with expert readers and IVUS and carry prognostic value for future MI. ",

author = "Andrew Lin and Nipun Manral and Priscilla McElhinney and Aditya Killekar and Hidenari Matsumoto and Jacek Kwiecinski and Konrad Pieszko and Aryabod Razipour and Kajetan Grodecki and Caroline Park and Yuka Otaki and Mhairi Doris and Kwan, {Alan C.} and Donghee Han and Keiichiro Kuronuma and Tomasino, {Guadalupe Flores} and Evangelos Tzolos and Aakash Shanbhag and Markus Goeller and Mohamed Marwan and Heidi Gransar and Tamarappoo, {Balaji K} and Sebastien Cadet and Stephan Achenbach and Nicholls, {Stephen J} and Wong, {Dennis T} and Berman, {Daniel S.} and Dweck, {Marc R} and Newby, {David E} and Williams, {Michelle C} and Slomka, {Piotr J} and Damini Dey",

year = "2022",

month = mar,

day = "22",

doi = "0.1016/S2589-7500(22)00022-X",

language = "English",

journal = "The Lancet Digital Health",

issn = "2589-7500",

publisher = "Elsevier",

}

Lin, A, Manral, N, McElhinney , P, Killekar, A, Matsumoto, H, Kwiecinski, J, Pieszko, K, Razipour, A, Grodecki, K, Park, C, Otaki, Y, Doris, M, Kwan, AC, Han, D, Kuronuma, K, Tomasino, GF, Tzolos, E, Shanbhag, A, Goeller, M, Marwan, M, Gransar, H, Tamarappoo, BK, Cadet, S, Achenbach, S, Nicholls, SJ, Wong, DT, Berman, DS, Dweck, MR , Newby, DE , Williams, MC, Slomka, PJ & Dey, D 2022, 'Deep learning-enabled coronary CT angiography for plaque and stenosis quantification and cardiac risk prediction: an international multicentre study', The Lancet Digital Health. https://doi.org/0.1016/S2589-7500(22)00022-X

TY - JOUR

T1 - Deep learning-enabled coronary CT angiography for plaque and stenosis quantification and cardiac risk prediction: an international multicentre study

AU - Lin, Andrew

AU - Manral, Nipun

AU - McElhinney , Priscilla

AU - Killekar, Aditya

AU - Matsumoto, Hidenari

AU - Kwiecinski, Jacek

AU - Pieszko, Konrad

AU - Razipour, Aryabod

AU - Grodecki, Kajetan

AU - Park, Caroline

AU - Otaki, Yuka

AU - Doris, Mhairi

AU - Kwan, Alan C.

AU - Han, Donghee

AU - Kuronuma, Keiichiro

AU - Tomasino, Guadalupe Flores

AU - Tzolos, Evangelos

AU - Shanbhag, Aakash

AU - Goeller, Markus

AU - Marwan, Mohamed

AU - Gransar, Heidi

AU - Tamarappoo, Balaji K

AU - Cadet, Sebastien

AU - Achenbach, Stephan

AU - Nicholls, Stephen J

AU - Wong, Dennis T

AU - Berman, Daniel S.

AU - Dweck, Marc R

AU - Newby, David E

AU - Williams, Michelle C

AU - Slomka, Piotr J

AU - Dey, Damini

PY - 2022/3/22

Y1 - 2022/3/22

N2 - Background: Atherosclerotic plaque quantification from coronary computed tomography angiography (CCTA) enables accurate assessment of coronary artery disease burden and prognosis. We sought to develop and externally validate a deep learning (DL) system for CCTA- derived measures of plaque volume and stenosis severity. Methods: This was an international multicentre study of 1196 patients undergoing CCTA at 11 sites, with external evaluation of prognostic risk in 1,611 patients from the prospective SCOT- HEART trial. A novel DL convolutional neural network was trained to segment coronary plaque in 921 patients (5,045 lesions). The DL network was then applied to an independent test set, which included an external validation cohort of 175 patients (1,081 lesions) as well 50 patients (84 lesions) assessed by intravascular ultrasound (IVUS) within one month of CCTA. Thereafter, we evaluated the prognostic value of DL-based plaque measurements for fatal or nonfatal myocardial infarction (MI) in 1,611 patients from the prospective SCOT-HEART trial. Findings: In the overall test set, there was excellent agreement between DL and expert reader measurements of total plaque volume (intraclass correlation coefficient [ICC] 0·964) and percent diameter stenosis (ICC 0·879; both p<0.0001). When compared with IVUS, there was excellent agreement for DL total plaque volume (ICC 0·949) and minimal luminal area (ICC 0·904). The average per-patient DL plaque analysis time was 5.7 seconds versus 25-30 minutes taken by experts. Over a median follow-up of 4·7 years, MI occurred in 41 patients (2·5%) from the SCOT-HEART trial. DL-based total plaque volume ≥238·5mm3 was associated with an increased risk of MI (HR 5·36, 95% CI 1·70-16·86; p=0·0042) after adjustment for the presence of DL-based obstructive stenosis (HR 2·49, 95% CI 1·07-5·50; p=0·0089) and the cardiovascular risk score (HR 1·01, 95% CI 0·99-1·04; p=0·35). Interpretation: A novel externally validated DL system provides rapid measurements of plaque volume and stenosis severity from CCTA which agree closely with expert readers and IVUS and carry prognostic value for future MI.

AB - Background: Atherosclerotic plaque quantification from coronary computed tomography angiography (CCTA) enables accurate assessment of coronary artery disease burden and prognosis. We sought to develop and externally validate a deep learning (DL) system for CCTA- derived measures of plaque volume and stenosis severity. Methods: This was an international multicentre study of 1196 patients undergoing CCTA at 11 sites, with external evaluation of prognostic risk in 1,611 patients from the prospective SCOT- HEART trial. A novel DL convolutional neural network was trained to segment coronary plaque in 921 patients (5,045 lesions). The DL network was then applied to an independent test set, which included an external validation cohort of 175 patients (1,081 lesions) as well 50 patients (84 lesions) assessed by intravascular ultrasound (IVUS) within one month of CCTA. Thereafter, we evaluated the prognostic value of DL-based plaque measurements for fatal or nonfatal myocardial infarction (MI) in 1,611 patients from the prospective SCOT-HEART trial. Findings: In the overall test set, there was excellent agreement between DL and expert reader measurements of total plaque volume (intraclass correlation coefficient [ICC] 0·964) and percent diameter stenosis (ICC 0·879; both p<0.0001). When compared with IVUS, there was excellent agreement for DL total plaque volume (ICC 0·949) and minimal luminal area (ICC 0·904). The average per-patient DL plaque analysis time was 5.7 seconds versus 25-30 minutes taken by experts. Over a median follow-up of 4·7 years, MI occurred in 41 patients (2·5%) from the SCOT-HEART trial. DL-based total plaque volume ≥238·5mm3 was associated with an increased risk of MI (HR 5·36, 95% CI 1·70-16·86; p=0·0042) after adjustment for the presence of DL-based obstructive stenosis (HR 2·49, 95% CI 1·07-5·50; p=0·0089) and the cardiovascular risk score (HR 1·01, 95% CI 0·99-1·04; p=0·35). Interpretation: A novel externally validated DL system provides rapid measurements of plaque volume and stenosis severity from CCTA which agree closely with expert readers and IVUS and carry prognostic value for future MI.

U2 - 0.1016/S2589-7500(22)00022-X

DO - 0.1016/S2589-7500(22)00022-X

M3 - Article

SN - 2589-7500

JO - The Lancet Digital Health

JF - The Lancet Digital Health

ER -

Deep learning-enabled coronary CT angiography for plaque and stenosis quantification and cardiac risk prediction: an international multicentre study

Abstract / Description of output

Access to Document

Fingerprint

Cite this