Clinical benefit of AI-assisted lung ultrasound in a resource-limited intensive care unit

Phung tran huy Nhat; Nguyen Van hao; Phan vinh Tho; Hamideh Kerdegari; Luigi Pisani; Le ngoc minh Thu; Le thanh Phuong; Ha thi hai Duong; Duong bich Thuy; Angela Mcbride; Miguel Xochicale; Marcus Schultz; Reza Razavi; Andrew King; Louise Thwaites; Nguyen Van vinh chau; Sophie Yacoub; Dang phuong Thao; Dang trung Kien; Doan bui xuan Thy; Dong huu khanh Trinh; Du hong Duc; Ronald Geskus; Ho bich Hai; Ho quang Chanh; Ho Van hien; Huynh trung Trieu; Evelyne Kestelyn; Lam minh Yen; Le dinh Van khoa; Le thanh Phuong; Le thuy thuy Khanh; Luu hoai bao Tran; Luu phuoc An; Angela Mcbride; Nguyen lam Vuong; Nguyen quang Huy; Nguyen than ha Quyen; Nguyen thanh Ngoc; Nguyen thi Giang; Nguyen thi diem Trinh; Nguyen thi Le thanh; Nguyen thi phuong Dung; Nguyen thi phuong Thao; Ninh thi thanh Van; Pham tieu Kieu; Phan nguyen quoc Khanh; Phung khanh Lam; Phung tran huy Nhat; Guy Thwaites; Louise Thwaites; Tran minh Duc; Trinh manh Hung; Hugo Turner; Jennifer ilo Van nuil; Vo tan Hoang; Vu ngo thanh Huyen; Sophie Yacoub; Cao thi Tam; Duong bich Thuy; Ha thi hai Duong; Ho dang trung Nghia; Le buu Chau; Le mau Toan; Le ngoc minh Thu; Le thi mai Thao; Luong thi hue Tai; Nguyen hoan Phu; Nguyen quoc Viet; Nguyen thanh Dung; Nguyen thanh Nguyen; Nguyen thanh Phong; Nguyen thi kim Anh; Nguyen Van hao; Nguyen Van thanh duoc; Pham kieu nguyet Oanh; Phan thi hong Van; Phan tu Qui; Phan vinh Tho; Truong thi phuong Thao; Natasha Ali; David Clifton; Mike English; Jannis Hagenah; Ping Lu; Jacob Mcknight; Chris Paton; Tingting Zhu; Pantelis Georgiou; Bernard hernandez Perez; Kerri Hill-Cawthorne; Alison Holmes; Stefan Karolcik; Damien Ming; Nicolas Moser; Jesus rodriguez Manzano; Liane Canas; Alberto Gomez; Hamideh Kerdegari; Andrew King; Marc Modat; Reza Razavi; Miguel Xochicale; Walter Karlen; Linda Denehy; Thomas Rollinson; Luigi Pisani; Marcus Schultz; Alberto Gomez

doi:10.1186/s13054-023-04548-w

Clinical benefit of AI-assisted lung ultrasound in a resource-limited intensive care unit

Phung tran huy Nhat, Nguyen Van hao, Phan vinh Tho, Hamideh Kerdegari, Luigi Pisani, Le ngoc minh Thu, Le thanh Phuong, Ha thi hai Duong, Duong bich Thuy, Angela Mcbride, Miguel Xochicale, Marcus Schultz, Reza Razavi, Andrew King, Louise Thwaites, Nguyen Van vinh chau, Sophie Yacoub, Dang phuong Thao, Dang trung Kien, Doan bui xuan ThyDong huu khanh Trinh, Du hong Duc, Ronald Geskus, Ho bich Hai, Ho quang Chanh, Ho Van hien, Huynh trung Trieu, Evelyne Kestelyn, Lam minh Yen, Le dinh Van khoa, Le thanh Phuong, Le thuy thuy Khanh, Luu hoai bao Tran, Luu phuoc An, Angela Mcbride, Nguyen lam Vuong, Nguyen quang Huy, Nguyen than ha Quyen, Nguyen thanh Ngoc, Nguyen thi Giang, Nguyen thi diem Trinh, Nguyen thi Le thanh, Nguyen thi phuong Dung, Nguyen thi phuong Thao, Ninh thi thanh Van, Pham tieu Kieu, Phan nguyen quoc Khanh, Phung khanh Lam, Phung tran huy Nhat, Guy Thwaites, Louise Thwaites, Tran minh Duc, Trinh manh Hung, Hugo Turner, Jennifer ilo Van nuil, Vo tan Hoang, Vu ngo thanh Huyen, Sophie Yacoub, Cao thi Tam, Duong bich Thuy, Ha thi hai Duong, Ho dang trung Nghia, Le buu Chau, Le mau Toan, Le ngoc minh Thu, Le thi mai Thao, Luong thi hue Tai, Nguyen hoan Phu, Nguyen quoc Viet, Nguyen thanh Dung, Nguyen thanh Nguyen, Nguyen thanh Phong, Nguyen thi kim Anh, Nguyen Van hao, Nguyen Van thanh duoc, Pham kieu nguyet Oanh, Phan thi hong Van, Phan tu Qui, Phan vinh Tho, Truong thi phuong Thao, Natasha Ali, David Clifton, Mike English, Jannis Hagenah, Ping Lu, Jacob Mcknight, Chris Paton, Tingting Zhu, Pantelis Georgiou, Bernard hernandez Perez, Kerri Hill-Cawthorne, Alison Holmes, Stefan Karolcik, Damien Ming, Nicolas Moser, Jesus rodriguez Manzano, Liane Canas, Alberto Gomez, Hamideh Kerdegari, Andrew King, Marc Modat, Reza Razavi, Miguel Xochicale, Walter Karlen, Linda Denehy, Thomas Rollinson, Luigi Pisani, Marcus Schultz, Alberto Gomez

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

16 Citations (Scopus)

Abstract

Background: Interpreting point-of-care lung ultrasound (LUS) images from intensive care unit (ICU) patients can be challenging, especially in low- and middle- income countries (LMICs) where there is limited training available. Despite recent advances in the use of Artificial Intelligence (AI) to automate many ultrasound imaging analysis tasks, no AI-enabled LUS solutions have been proven to be clinically useful in ICUs, and specifically in LMICs. Therefore, we developed an AI solution that assists LUS practitioners and assessed its usefulness in a low resource ICU. Methods: This was a three-phase prospective study. In the first phase, the performance of four different clinical user groups in interpreting LUS clips was assessed. In the second phase, the performance of 57 non-expert clinicians with and without the aid of a bespoke AI tool for LUS interpretation was assessed in retrospective offline clips. In the third phase, we conducted a prospective study in the ICU where 14 clinicians were asked to carry out LUS examinations in 7 patients with and without our AI tool and we interviewed the clinicians regarding the usability of the AI tool. Results: The average accuracy of beginners’ LUS interpretation was 68.7% [95% CI 66.8–70.7%] compared to 72.2% [95% CI 70.0–75.6%] in intermediate, and 73.4% [95% CI 62.2–87.8%] in advanced users. Experts had an average accuracy of 95.0% [95% CI 88.2–100.0%], which was significantly better than beginners, intermediate and advanced users (p < 0.001). When supported by our AI tool for interpreting retrospectively acquired clips, the non-expert clinicians improved their performance from an average of 68.9% [95% CI 65.6–73.9%] to 82.9% [95% CI 79.1–86.7%], (p < 0.001). In prospective real-time testing, non-expert clinicians improved their baseline performance from 68.1% [95% CI 57.9–78.2%] to 93.4% [95% CI 89.0–97.8%], (p < 0.001) when using our AI tool. The time-to-interpret clips improved from a median of 12.1 s (IQR 8.5–20.6) to 5.0 s (IQR 3.5–8.8), (p < 0.001) and clinicians’ median confidence level improved from 3 out of 4 to 4 out of 4 when using our AI tool. Conclusions: AI-assisted LUS can help non-expert clinicians in an LMIC ICU improve their performance in interpreting LUS features more accurately, more quickly and more confidently.

Original language	English
Article number	257
Journal	CRITICAL CARE
Volume	27
Issue number	1
DOIs	https://doi.org/10.1186/s13054-023-04548-w
Publication status	Published - 1 Jul 2023

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1186/s13054-023-04548-wLicence: CC BY

https://ccforum.biomedcentral.com/articles/10.1186/s13054-023-04548-w

Cite this

Nhat, P. T. H., Van hao, N., Tho, P. V., Kerdegari, H., Pisani, L., Thu, L. N. M., Phuong, L. T., Duong, H. T. H., Thuy, D. B., Mcbride, A., Xochicale, M., Schultz, M., Razavi, R., King, A., Thwaites, L., Van vinh chau, N., Yacoub, S., Thao, D. P., Kien, D. T., ... Gomez, A. (2023). Clinical benefit of AI-assisted lung ultrasound in a resource-limited intensive care unit. CRITICAL CARE, 27(1), Article 257 . https://doi.org/10.1186/s13054-023-04548-w

@article{6d3ce36cf0194a619efe54eeee002571,

title = "Clinical benefit of AI-assisted lung ultrasound in a resource-limited intensive care unit",

abstract = "Background: Interpreting point-of-care lung ultrasound (LUS) images from intensive care unit (ICU) patients can be challenging, especially in low- and middle- income countries (LMICs) where there is limited training available. Despite recent advances in the use of Artificial Intelligence (AI) to automate many ultrasound imaging analysis tasks, no AI-enabled LUS solutions have been proven to be clinically useful in ICUs, and specifically in LMICs. Therefore, we developed an AI solution that assists LUS practitioners and assessed its usefulness in a low resource ICU. Methods: This was a three-phase prospective study. In the first phase, the performance of four different clinical user groups in interpreting LUS clips was assessed. In the second phase, the performance of 57 non-expert clinicians with and without the aid of a bespoke AI tool for LUS interpretation was assessed in retrospective offline clips. In the third phase, we conducted a prospective study in the ICU where 14 clinicians were asked to carry out LUS examinations in 7 patients with and without our AI tool and we interviewed the clinicians regarding the usability of the AI tool. Results: The average accuracy of beginners{\textquoteright} LUS interpretation was 68.7% [95% CI 66.8–70.7%] compared to 72.2% [95% CI 70.0–75.6%] in intermediate, and 73.4% [95% CI 62.2–87.8%] in advanced users. Experts had an average accuracy of 95.0% [95% CI 88.2–100.0%], which was significantly better than beginners, intermediate and advanced users (p < 0.001). When supported by our AI tool for interpreting retrospectively acquired clips, the non-expert clinicians improved their performance from an average of 68.9% [95% CI 65.6–73.9%] to 82.9% [95% CI 79.1–86.7%], (p < 0.001). In prospective real-time testing, non-expert clinicians improved their baseline performance from 68.1% [95% CI 57.9–78.2%] to 93.4% [95% CI 89.0–97.8%], (p < 0.001) when using our AI tool. The time-to-interpret clips improved from a median of 12.1 s (IQR 8.5–20.6) to 5.0 s (IQR 3.5–8.8), (p < 0.001) and clinicians{\textquoteright} median confidence level improved from 3 out of 4 to 4 out of 4 when using our AI tool. Conclusions: AI-assisted LUS can help non-expert clinicians in an LMIC ICU improve their performance in interpreting LUS features more accurately, more quickly and more confidently.",

author = "Nhat, {Phung tran huy} and {Van hao}, Nguyen and Tho, {Phan vinh} and Hamideh Kerdegari and Luigi Pisani and Thu, {Le ngoc minh} and Phuong, {Le thanh} and Duong, {Ha thi hai} and Thuy, {Duong bich} and Angela Mcbride and Miguel Xochicale and Marcus Schultz and Reza Razavi and Andrew King and Louise Thwaites and {Van vinh chau}, Nguyen and Sophie Yacoub and Thao, {Dang phuong} and Kien, {Dang trung} and Thy, {Doan bui xuan} and Trinh, {Dong huu khanh} and Duc, {Du hong} and Ronald Geskus and Hai, {Ho bich} and Chanh, {Ho quang} and {Van hien}, Ho and Trieu, {Huynh trung} and Evelyne Kestelyn and Yen, {Lam minh} and {Van khoa}, {Le dinh} and Phuong, {Le thanh} and Khanh, {Le thuy thuy} and Tran, {Luu hoai bao} and An, {Luu phuoc} and Angela Mcbride and Vuong, {Nguyen lam} and Huy, {Nguyen quang} and Quyen, {Nguyen than ha} and Ngoc, {Nguyen thanh} and Giang, {Nguyen thi} and Trinh, {Nguyen thi diem} and {Le thanh}, {Nguyen thi} and Dung, {Nguyen thi phuong} and Thao, {Nguyen thi phuong} and Van, {Ninh thi thanh} and Kieu, {Pham tieu} and Khanh, {Phan nguyen quoc} and Lam, {Phung khanh} and Nhat, {Phung tran huy} and Guy Thwaites and Louise Thwaites and Duc, {Tran minh} and Hung, {Trinh manh} and Hugo Turner and {Van nuil}, {Jennifer ilo} and Hoang, {Vo tan} and Huyen, {Vu ngo thanh} and Sophie Yacoub and Tam, {Cao thi} and Thuy, {Duong bich} and Duong, {Ha thi hai} and Nghia, {Ho dang trung} and Chau, {Le buu} and Toan, {Le mau} and Thu, {Le ngoc minh} and Thao, {Le thi mai} and Tai, {Luong thi hue} and Phu, {Nguyen hoan} and Viet, {Nguyen quoc} and Dung, {Nguyen thanh} and Nguyen, {Nguyen thanh} and Phong, {Nguyen thanh} and Anh, {Nguyen thi kim} and {Van hao}, Nguyen and {Van thanh duoc}, Nguyen and Oanh, {Pham kieu nguyet} and Van, {Phan thi hong} and Qui, {Phan tu} and Tho, {Phan vinh} and Thao, {Truong thi phuong} and Natasha Ali and David Clifton and Mike English and Jannis Hagenah and Ping Lu and Jacob Mcknight and Chris Paton and Tingting Zhu and Pantelis Georgiou and Perez, {Bernard hernandez} and Kerri Hill-Cawthorne and Alison Holmes and Stefan Karolcik and Damien Ming and Nicolas Moser and Manzano, {Jesus rodriguez} and Liane Canas and Alberto Gomez and Hamideh Kerdegari and Andrew King and Marc Modat and Reza Razavi and Miguel Xochicale and Walter Karlen and Linda Denehy and Thomas Rollinson and Luigi Pisani and Marcus Schultz and Alberto Gomez",

note = "Funding Information: This work was supported by the Wellcome Trust under grant 217650/Z/19/Z. Funding Information: We thank all the patients, nurses and clinicians who generously donated their time during the study. Dang Phuong Thao1, Dang Trung Kien1, Doan Bui Xuan Thy1, Dong Huu Khanh Trinh1,5, Du Hong Duc1, Ronald Geskus1, Ho Bich Hai1, Ho Quang Chanh1, Ho Van Hien1, Huynh Trung Trieu1, Evelyne Kestelyn1, Lam Minh Yen1, Le Dinh Van Khoa1, Le Thanh Phuong1, Le Thuy Thuy Khanh1, Luu Hoai Bao Tran1, Luu Phuoc An1, Angela Mcbride1, Nguyen Lam Vuong1, Nguyen Quang Huy1, Nguyen Than Ha Quyen1, Nguyen Thanh Ngoc1, Nguyen Thi Giang1, Nguyen Thi Diem Trinh1, Nguyen Thi Le Thanh1, Nguyen Thi Phuong Dung1, Nguyen Thi Phuong Thao1, Ninh Thi Thanh Van1, Pham Tieu Kieu1, Phan Nguyen Quoc Khanh1, Phung Khanh Lam1, Phung Tran Huy Nhat1,5, Guy Thwaites1,3, Louise Thwaites1,3, Tran Minh Duc1, Trinh Manh Hung1, Hugo Turner1, Jennifer Ilo Van Nuil1, Vo Tan Hoang1, Vu Ngo Thanh Huyen1, Sophie Yacoub1,3, Cao Thi Tam2, Duong Bich Thuy2, Ha Thi Hai Duong2, Ho Dang Trung Nghia2, Le Buu Chau2, Le Mau Toan2, Le Ngoc Minh Thu2, Le Thi Mai Thao2, Luong Thi Hue Tai2, Nguyen Hoan Phu2, Nguyen Quoc Viet2, Nguyen Thanh Dung2, Nguyen Thanh Nguyen2, Nguyen Thanh Phong2, Nguyen Thi Kim Anh2, Nguyen Van Hao2, Nguyen Van Thanh Duoc2, Pham Kieu Nguyet Oanh2, Phan Thi Hong Van2, Phan Tu Qui2, Phan Vinh Tho2, Truong Thi Phuong Thao2, Natasha Ali3, David Clifton3, Mike English3, Jannis Hagenah3, Ping Lu3, Jacob McKnight3, Chris Paton3, Tingting Zhu3, Pantelis Georgiou4, Bernard Hernandez Perez4, Kerri Hill-Cawthorne4, Alison Holmes4, Stefan Karolcik4, Damien Ming4, Nicolas Moser4, Jesus Rodriguez Manzano4, Liane Canas5, Alberto Gomez5, Hamideh Kerdegari5, Andrew King5, Marc Modat5, Reza Razavi5, Miguel Xochicale5, Walter Karlen6, Linda Denehy7, Thomas Rollinson7, Luigi Pisani8, Marcus Schultz8.1Oxford University Clinical Research Unit;2Hospital for Tropical Diseases, Ho Chi Minh City;3University of Oxford;4Imperial College London;5King{\textquoteright}s College London;6University of Ulm;7The University of Melbourne;8Mahidol Oxford Tropical Medicine Research Unit. Publisher Copyright: {\textcopyright} 2023, The Author(s).",

year = "2023",

month = jul,

day = "1",

doi = "10.1186/s13054-023-04548-w",

language = "English",

volume = "27",

journal = "CRITICAL CARE",

issn = "1364-8535",

publisher = "BIOMED CENTRAL LTD",

number = "1",

}

Nhat, PTH, Van hao, N, Tho, PV, Kerdegari, H, Pisani, L, Thu, LNM, Phuong, LT, Duong, HTH, Thuy, DB, Mcbride, A, Xochicale, M, Schultz, M, Razavi, R , King, A, Thwaites, L, Van vinh chau, N, Yacoub, S, Thao, DP, Kien, DT, Thy, DBX, Trinh, DHK, Duc, DH, Geskus, R, Hai, HB, Chanh, HQ, Van hien, H, Trieu, HT, Kestelyn, E, Yen, LM, Van khoa, LD, Phuong, LT, Khanh, LTT, Tran, LHB, An, LP, Mcbride, A, Vuong, NL, Huy, NQ, Quyen, NTH, Ngoc, NT, Giang, NT, Trinh, NTD, Le thanh, NT, Dung, NTP, Thao, NTP, Van, NTT, Kieu, PT, Khanh, PNQ, Lam, PK, Nhat, PTH, Thwaites, G, Thwaites, L, Duc, TM, Hung, TM, Turner, H, Van nuil, JI, Hoang, VT, Huyen, VNT, Yacoub, S, Tam, CT, Thuy, DB, Duong, HTH, Nghia, HDT, Chau, LB, Toan, LM, Thu, LNM, Thao, LTM, Tai, LTH, Phu, NH, Viet, NQ, Dung, NT, Nguyen, NT, Phong, NT, Anh, NTK, Van hao, N, Van thanh duoc, N, Oanh, PKN, Van, PTH, Qui, PT, Tho, PV, Thao, TTP, Ali, N, Clifton, D, English, M, Hagenah, J, Lu, P, Mcknight, J, Paton, C, Zhu, T, Georgiou, P, Perez, BH, Hill-Cawthorne, K, Holmes, A, Karolcik, S, Ming, D, Moser, N, Manzano, JR, Canas, L, Gomez, A , Kerdegari, H , King, A , Modat, M , Razavi, R, Xochicale, M, Karlen, W, Denehy, L, Rollinson, T, Pisani, L, Schultz, M & Gomez, A 2023, 'Clinical benefit of AI-assisted lung ultrasound in a resource-limited intensive care unit', CRITICAL CARE, vol. 27, no. 1, 257 . https://doi.org/10.1186/s13054-023-04548-w

TY - JOUR

T1 - Clinical benefit of AI-assisted lung ultrasound in a resource-limited intensive care unit

AU - Nhat, Phung tran huy

AU - Van hao, Nguyen

AU - Tho, Phan vinh

AU - Kerdegari, Hamideh

AU - Pisani, Luigi

AU - Thu, Le ngoc minh

AU - Phuong, Le thanh

AU - Duong, Ha thi hai

AU - Thuy, Duong bich

AU - Mcbride, Angela

AU - Xochicale, Miguel

AU - Schultz, Marcus

AU - Razavi, Reza

AU - King, Andrew

AU - Thwaites, Louise

AU - Van vinh chau, Nguyen

AU - Yacoub, Sophie

AU - Thao, Dang phuong

AU - Kien, Dang trung

AU - Thy, Doan bui xuan

AU - Trinh, Dong huu khanh

AU - Duc, Du hong

AU - Geskus, Ronald

AU - Hai, Ho bich

AU - Chanh, Ho quang

AU - Van hien, Ho

AU - Trieu, Huynh trung

AU - Kestelyn, Evelyne

AU - Yen, Lam minh

AU - Van khoa, Le dinh

AU - Phuong, Le thanh

AU - Khanh, Le thuy thuy

AU - Tran, Luu hoai bao

AU - An, Luu phuoc

AU - Mcbride, Angela

AU - Vuong, Nguyen lam

AU - Huy, Nguyen quang

AU - Quyen, Nguyen than ha

AU - Ngoc, Nguyen thanh

AU - Giang, Nguyen thi

AU - Trinh, Nguyen thi diem

AU - Le thanh, Nguyen thi

AU - Dung, Nguyen thi phuong

AU - Thao, Nguyen thi phuong

AU - Van, Ninh thi thanh

AU - Kieu, Pham tieu

AU - Khanh, Phan nguyen quoc

AU - Lam, Phung khanh

AU - Nhat, Phung tran huy

AU - Thwaites, Guy

AU - Thwaites, Louise

AU - Duc, Tran minh

AU - Hung, Trinh manh

AU - Turner, Hugo

AU - Van nuil, Jennifer ilo

AU - Hoang, Vo tan

AU - Huyen, Vu ngo thanh

AU - Yacoub, Sophie

AU - Tam, Cao thi

AU - Thuy, Duong bich

AU - Duong, Ha thi hai

AU - Nghia, Ho dang trung

AU - Chau, Le buu

AU - Toan, Le mau

AU - Thu, Le ngoc minh

AU - Thao, Le thi mai

AU - Tai, Luong thi hue

AU - Phu, Nguyen hoan

AU - Viet, Nguyen quoc

AU - Dung, Nguyen thanh

AU - Nguyen, Nguyen thanh

AU - Phong, Nguyen thanh

AU - Anh, Nguyen thi kim

AU - Van hao, Nguyen

AU - Van thanh duoc, Nguyen

AU - Oanh, Pham kieu nguyet

AU - Van, Phan thi hong

AU - Qui, Phan tu

AU - Tho, Phan vinh

AU - Thao, Truong thi phuong

AU - Ali, Natasha

AU - Clifton, David

AU - English, Mike

AU - Hagenah, Jannis

AU - Lu, Ping

AU - Mcknight, Jacob

AU - Paton, Chris

AU - Zhu, Tingting

AU - Georgiou, Pantelis

AU - Perez, Bernard hernandez

AU - Hill-Cawthorne, Kerri

AU - Holmes, Alison

AU - Karolcik, Stefan

AU - Ming, Damien

AU - Moser, Nicolas

AU - Manzano, Jesus rodriguez

AU - Canas, Liane

AU - Gomez, Alberto

AU - Kerdegari, Hamideh

AU - King, Andrew

AU - Modat, Marc

AU - Razavi, Reza

AU - Xochicale, Miguel

AU - Karlen, Walter

AU - Denehy, Linda

AU - Rollinson, Thomas

AU - Pisani, Luigi

AU - Schultz, Marcus

AU - Gomez, Alberto

N1 - Funding Information: This work was supported by the Wellcome Trust under grant 217650/Z/19/Z. Funding Information: We thank all the patients, nurses and clinicians who generously donated their time during the study. Dang Phuong Thao1, Dang Trung Kien1, Doan Bui Xuan Thy1, Dong Huu Khanh Trinh1,5, Du Hong Duc1, Ronald Geskus1, Ho Bich Hai1, Ho Quang Chanh1, Ho Van Hien1, Huynh Trung Trieu1, Evelyne Kestelyn1, Lam Minh Yen1, Le Dinh Van Khoa1, Le Thanh Phuong1, Le Thuy Thuy Khanh1, Luu Hoai Bao Tran1, Luu Phuoc An1, Angela Mcbride1, Nguyen Lam Vuong1, Nguyen Quang Huy1, Nguyen Than Ha Quyen1, Nguyen Thanh Ngoc1, Nguyen Thi Giang1, Nguyen Thi Diem Trinh1, Nguyen Thi Le Thanh1, Nguyen Thi Phuong Dung1, Nguyen Thi Phuong Thao1, Ninh Thi Thanh Van1, Pham Tieu Kieu1, Phan Nguyen Quoc Khanh1, Phung Khanh Lam1, Phung Tran Huy Nhat1,5, Guy Thwaites1,3, Louise Thwaites1,3, Tran Minh Duc1, Trinh Manh Hung1, Hugo Turner1, Jennifer Ilo Van Nuil1, Vo Tan Hoang1, Vu Ngo Thanh Huyen1, Sophie Yacoub1,3, Cao Thi Tam2, Duong Bich Thuy2, Ha Thi Hai Duong2, Ho Dang Trung Nghia2, Le Buu Chau2, Le Mau Toan2, Le Ngoc Minh Thu2, Le Thi Mai Thao2, Luong Thi Hue Tai2, Nguyen Hoan Phu2, Nguyen Quoc Viet2, Nguyen Thanh Dung2, Nguyen Thanh Nguyen2, Nguyen Thanh Phong2, Nguyen Thi Kim Anh2, Nguyen Van Hao2, Nguyen Van Thanh Duoc2, Pham Kieu Nguyet Oanh2, Phan Thi Hong Van2, Phan Tu Qui2, Phan Vinh Tho2, Truong Thi Phuong Thao2, Natasha Ali3, David Clifton3, Mike English3, Jannis Hagenah3, Ping Lu3, Jacob McKnight3, Chris Paton3, Tingting Zhu3, Pantelis Georgiou4, Bernard Hernandez Perez4, Kerri Hill-Cawthorne4, Alison Holmes4, Stefan Karolcik4, Damien Ming4, Nicolas Moser4, Jesus Rodriguez Manzano4, Liane Canas5, Alberto Gomez5, Hamideh Kerdegari5, Andrew King5, Marc Modat5, Reza Razavi5, Miguel Xochicale5, Walter Karlen6, Linda Denehy7, Thomas Rollinson7, Luigi Pisani8, Marcus Schultz8.1Oxford University Clinical Research Unit;2Hospital for Tropical Diseases, Ho Chi Minh City;3University of Oxford;4Imperial College London;5King’s College London;6University of Ulm;7The University of Melbourne;8Mahidol Oxford Tropical Medicine Research Unit. Publisher Copyright: © 2023, The Author(s).

PY - 2023/7/1

Y1 - 2023/7/1

N2 - Background: Interpreting point-of-care lung ultrasound (LUS) images from intensive care unit (ICU) patients can be challenging, especially in low- and middle- income countries (LMICs) where there is limited training available. Despite recent advances in the use of Artificial Intelligence (AI) to automate many ultrasound imaging analysis tasks, no AI-enabled LUS solutions have been proven to be clinically useful in ICUs, and specifically in LMICs. Therefore, we developed an AI solution that assists LUS practitioners and assessed its usefulness in a low resource ICU. Methods: This was a three-phase prospective study. In the first phase, the performance of four different clinical user groups in interpreting LUS clips was assessed. In the second phase, the performance of 57 non-expert clinicians with and without the aid of a bespoke AI tool for LUS interpretation was assessed in retrospective offline clips. In the third phase, we conducted a prospective study in the ICU where 14 clinicians were asked to carry out LUS examinations in 7 patients with and without our AI tool and we interviewed the clinicians regarding the usability of the AI tool. Results: The average accuracy of beginners’ LUS interpretation was 68.7% [95% CI 66.8–70.7%] compared to 72.2% [95% CI 70.0–75.6%] in intermediate, and 73.4% [95% CI 62.2–87.8%] in advanced users. Experts had an average accuracy of 95.0% [95% CI 88.2–100.0%], which was significantly better than beginners, intermediate and advanced users (p < 0.001). When supported by our AI tool for interpreting retrospectively acquired clips, the non-expert clinicians improved their performance from an average of 68.9% [95% CI 65.6–73.9%] to 82.9% [95% CI 79.1–86.7%], (p < 0.001). In prospective real-time testing, non-expert clinicians improved their baseline performance from 68.1% [95% CI 57.9–78.2%] to 93.4% [95% CI 89.0–97.8%], (p < 0.001) when using our AI tool. The time-to-interpret clips improved from a median of 12.1 s (IQR 8.5–20.6) to 5.0 s (IQR 3.5–8.8), (p < 0.001) and clinicians’ median confidence level improved from 3 out of 4 to 4 out of 4 when using our AI tool. Conclusions: AI-assisted LUS can help non-expert clinicians in an LMIC ICU improve their performance in interpreting LUS features more accurately, more quickly and more confidently.

AB - Background: Interpreting point-of-care lung ultrasound (LUS) images from intensive care unit (ICU) patients can be challenging, especially in low- and middle- income countries (LMICs) where there is limited training available. Despite recent advances in the use of Artificial Intelligence (AI) to automate many ultrasound imaging analysis tasks, no AI-enabled LUS solutions have been proven to be clinically useful in ICUs, and specifically in LMICs. Therefore, we developed an AI solution that assists LUS practitioners and assessed its usefulness in a low resource ICU. Methods: This was a three-phase prospective study. In the first phase, the performance of four different clinical user groups in interpreting LUS clips was assessed. In the second phase, the performance of 57 non-expert clinicians with and without the aid of a bespoke AI tool for LUS interpretation was assessed in retrospective offline clips. In the third phase, we conducted a prospective study in the ICU where 14 clinicians were asked to carry out LUS examinations in 7 patients with and without our AI tool and we interviewed the clinicians regarding the usability of the AI tool. Results: The average accuracy of beginners’ LUS interpretation was 68.7% [95% CI 66.8–70.7%] compared to 72.2% [95% CI 70.0–75.6%] in intermediate, and 73.4% [95% CI 62.2–87.8%] in advanced users. Experts had an average accuracy of 95.0% [95% CI 88.2–100.0%], which was significantly better than beginners, intermediate and advanced users (p < 0.001). When supported by our AI tool for interpreting retrospectively acquired clips, the non-expert clinicians improved their performance from an average of 68.9% [95% CI 65.6–73.9%] to 82.9% [95% CI 79.1–86.7%], (p < 0.001). In prospective real-time testing, non-expert clinicians improved their baseline performance from 68.1% [95% CI 57.9–78.2%] to 93.4% [95% CI 89.0–97.8%], (p < 0.001) when using our AI tool. The time-to-interpret clips improved from a median of 12.1 s (IQR 8.5–20.6) to 5.0 s (IQR 3.5–8.8), (p < 0.001) and clinicians’ median confidence level improved from 3 out of 4 to 4 out of 4 when using our AI tool. Conclusions: AI-assisted LUS can help non-expert clinicians in an LMIC ICU improve their performance in interpreting LUS features more accurately, more quickly and more confidently.

UR - http://www.scopus.com/inward/record.url?scp=85164150325&partnerID=8YFLogxK

U2 - 10.1186/s13054-023-04548-w

DO - 10.1186/s13054-023-04548-w

M3 - Article

SN - 1364-8535

VL - 27

JO - CRITICAL CARE

JF - CRITICAL CARE

IS - 1

M1 - 257

ER -

Clinical benefit of AI-assisted lung ultrasound in a resource-limited intensive care unit

Abstract

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this