Multimodal large language models demonstrate consistently low accuracy for 12-lead electrocardiogram interpretation: a 599-case benchmark study

H. Kamalanathan; P. Sahai; D. Downes; K. Kaleeny; A. Wilson-Smith; M. Mo; T. Chauraya; K. Rajesh; S. Christoforou; W. Ho; J. Kaleeny; D. Ng; J. Millhouse

doi:10.1016/j.hlc.2026.07.622

Multimodal large language models demonstrate consistently low accuracy for 12-lead electrocardiogram interpretation: a 599-case benchmark study

H. Kamalanathan, P. Sahai, D. Downes, K. Kaleeny, A. Wilson-Smith, M. Mo, T. Chauraya, K. Rajesh, S. Christoforou, W. Ho, J. Kaleeny, D. Ng, J. Millhouse

Research output: Contribution to journal › Meeting abstract › peer-review

Original language	English
Article number	0615
Pages (from-to)	S417
Number of pages	1
Journal	Heart Lung and Circulation
Volume	35
Issue number	Supplement 3
DOIs	https://doi.org/10.1016/j.hlc.2026.07.622
Publication status	Published - Aug 2026
Externally published	Yes
Event	74th Annual Scientific Meeting of the Cardiac Society of Australia and New Zealand - Sydney, Australia Duration: 6 Aug 2026 → 9 Aug 2026

Access to Document

10.1016/j.hlc.2026.07.622

Cite this

Kamalanathan, H., Sahai, P., Downes, D., Kaleeny, K., Wilson-Smith, A., Mo, M., Chauraya, T., Rajesh, K., Christoforou, S., Ho, W., Kaleeny, J., Ng, D., & Millhouse, J. (2026). Multimodal large language models demonstrate consistently low accuracy for 12-lead electrocardiogram interpretation: a 599-case benchmark study. Heart Lung and Circulation, 35(Supplement 3), S417. Article 0615. https://doi.org/10.1016/j.hlc.2026.07.622

@article{aff689e4e411475f94b513e2138a0a57,

title = "Multimodal large language models demonstrate consistently low accuracy for 12-lead electrocardiogram interpretation: a 599-case benchmark study",

author = "H. Kamalanathan and P. Sahai and D. Downes and K. Kaleeny and A. Wilson-Smith and M. Mo and T. Chauraya and K. Rajesh and S. Christoforou and W. Ho and J. Kaleeny and D. Ng and J. Millhouse",

year = "2026",

month = aug,

doi = "10.1016/j.hlc.2026.07.622",

language = "English",

volume = "35",

pages = "S417",

journal = "Heart Lung and Circulation",

issn = "1443-9506",

publisher = "ACADEMIC PRESS LTD- ELSEVIER SCIENCE LTD",

number = "Supplement 3",

note = "74th Annual Scientific Meeting of the Cardiac Society of Australia and New Zealand ; Conference date: 06-08-2026 Through 09-08-2026",

}

Kamalanathan, H, Sahai, P, Downes, D, Kaleeny, K, Wilson-Smith, A, Mo, M, Chauraya, T, Rajesh, K, Christoforou, S, Ho, W, Kaleeny, J, Ng, D & Millhouse, J 2026, 'Multimodal large language models demonstrate consistently low accuracy for 12-lead electrocardiogram interpretation: a 599-case benchmark study', Heart Lung and Circulation, vol. 35, no. Supplement 3, 0615, pp. S417. https://doi.org/10.1016/j.hlc.2026.07.622

TY - JOUR

T1 - Multimodal large language models demonstrate consistently low accuracy for 12-lead electrocardiogram interpretation

T2 - 74th Annual Scientific Meeting of the Cardiac Society of Australia and New Zealand

AU - Kamalanathan, H.

AU - Sahai, P.

AU - Downes, D.

AU - Kaleeny, K.

AU - Wilson-Smith, A.

AU - Mo, M.

AU - Chauraya, T.

AU - Rajesh, K.

AU - Christoforou, S.

AU - Ho, W.

AU - Kaleeny, J.

AU - Ng, D.

AU - Millhouse, J.

PY - 2026/8

Y1 - 2026/8

U2 - 10.1016/j.hlc.2026.07.622

DO - 10.1016/j.hlc.2026.07.622

M3 - Meeting abstract

SN - 1443-9506

VL - 35

SP - S417

JO - Heart Lung and Circulation

JF - Heart Lung and Circulation

IS - Supplement 3

M1 - 0615

Y2 - 6 August 2026 through 9 August 2026

ER -