Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
July 8, 2026Cardiovascular UltrasoundOpen Access

Machine learning LVOTd estimation matches expert accuracy while boosting stroke volume calculation yield to 92%.

View Full Paper
Ask AI
Bookmark
Share

Why the study?

The accuracy of machine learning computer model versus human expert measurement of LVOTd in estimating stroke volume relative to a pulmonary artery catheter was unknown.

Does a machine learning computer model for estimating LVOTd improve the accuracy and yield of stroke volume measurement by echocardiography compared to human expert measurement?

Comparison

Echo SV using LVOTd CM vs LVOTd HEM, both compared to PAC

Design

Prospective observational study

Key result

A machine learning computer model for estimating left ventricular outflow tract diameter yielded a mean bias of 1.75 compared to pulmonary artery catheter, similar to human expert measurement (bias 3.1), while increasing the yield of stroke volume calculation from 70% to 92%.

Authors

SMSarah Balderston MurthiSYShiming YangPOPeter P. Oliveri

Discussion

Loading...

Member takes

Overview

CM-derived LVOTd raises SV measurement yield to 92% with bias comparable to expert LVOTd versus PAC; leaves open adoption pending prospective validation.

Key Points

  • This research aims to evaluate the accuracy of a machine learning model for estimating stroke volume compared to human expert measurement in patients with pulmonary artery catheters.
  • Conducted a prospective observational study over 20 months in patients with a pulmonary artery catheter (N=84).
  • Measured stroke volume using echocardiography with both human expert measurement and machine learning model outputs.
  • Utilized Bland-Altman analysis to assess bias and agreement between stroke volume measurements and pulmonary artery catheter results.
  • With human expert measurement, stroke volume was calculable in 59 (70%) of patients; with machine learning, it was calculable in 78 (92%).
  • Bland-Altman analysis showed a mean bias of 3.1 for human expert measurements and 1.75 for machine learning, with narrow limits of agreement.
  • The machine learning model increased measurement yield by 22%, suggesting it enhances echocardiographic assessment in critically ill patients.

Study Design

Type

Observational (n=84)

Multicenter

No

Structured PICO

Does a machine learning computer model for estimating LVOTd improve the accuracy and yield of stroke volume measurement by echocardiography compared to human expert measurement?

P
Population
84 adult patients with a pulmonary artery catheter placed for clinical indications who underwent concurrent echocardiography to compare stroke volume measurement methods.
E
Exposure
Stroke volume calculation using a machine learning computer model (CM) to estimate left ventricular outflow tract diameter (LVOTd) combined with velocity time integral (VTI)
C
Comparator
Stroke volume calculation using human expert measurement (HEM) of LVOTd combined with VTI
O
Outcome
Bias and agreement of stroke volume measurement compared to pulmonary artery catheter (PAC) reference standardsurrogate

Main Result

Mean Difference: 1.75 (95% CI -29–32.5)

Absolute Event Rate: 1.75% vs 3.1%

A machine learning model for estimating LVOTd allows accurate calculation of stroke volume using only VTI, simplifying the assessment and increasing the measurement yield by 22% compared to human experts.

Limitations

  • Convenience sample obtained during daylight hours at a single institution
  • Clinical data was obtained prospectively and analyzed retrospectively
  • Considerable time lag between data collection and analysis
  • Images were obtained by fellows in training, potentially limiting generalizability

Cite This Study

Murthi et al. (2026) conducted an observational in Critically ill patients requiring hemodynamic monitoring (n=84). Machine learning computer model for left ventricular outflow tract diameter (LVOTdCM) vs. Human expert measurement (LVOTdHEM) was evaluated on Mean bias in stroke volume measurement compared to pulmonary artery catheter (PAC) (mean bias 1.75, 95% CI -29 to 32.5). A machine learning computer model for estimating left ventricular outflow tract diameter yielded a mean bias of 1.75 compared to pulmonary artery catheter, similar to human expert measurement (bias 3.1), while increasing the yield of stroke volume calculation from 70% to 92%.

synapsesocial.com/papers/6a4de7d8d2ea289ef6282d63https://doi.org/10.1186/s12947-026-00373-7
View Full Paper
Ask AI
Bookmark
Share