Clinical benefit of AI-assisted lung ultrasound in a resource-limited intensive care unit
Nhat PTH., Van Hao N., Tho PV., Kerdegari H., Pisani L., Thu LNM., Phuong LT., Duong HTH., Thuy DB., McBride A., Xochicale M., Schultz MJ., Razavi R., King AP., 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.
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.