Literature Spotlight | A New Breakthrough in Bedside Monitoring: Real-Time Visualization of Pulmonary Perfusion Without the Need for Breath-Holding


08/14

2026

PART 01 Background

 

Pulmonary perfusion assessment is critical for diagnosing pulmonary embolism and optimizing ventilation strategies in patients with acute respiratory distress syndrome (ARDS). An ideal technique should be bedside‑accessible, radiation‑free, and capable of dynamic, repeated measurements. Electrical impedance tomography (EIT), which generates images by detecting changes in thoracic electrical impedance, has been well established for evaluating lung ventilation. When applied to perfusion assessment, conventional approaches require breath‑holding (end‑expiratory pause) during contrast injection to capture blood‑flow signals under static conditions. This “breath‑holding method” demonstrates good reliability in hemodynamically stable patients. However, for critically ill patients who are hemodynamically unstable or unable to tolerate breath‑holding, this maneuver is limited. Consequently, the development of an EIT‑based perfusion‑assessment technique that does not rely on patient cooperation and can be performed during continuous ventilation has become an essential step toward expanding its clinical utility.

 

PART 02 Methods

 

The study employed a prospective, randomized crossover design and enrolled 20 ICU patients who were under deep sedation and receiving controlled ventilation. Each patient received, in random order, two bolus intravenous injections of 10 mL of 10% sodium chloride solution:

Traditional breath-holding technique: Injection is administered during the end-expiratory pause lasting ≥8 seconds.

Screen-free new method: administered during ongoing mechanical ventilation.

 

Time‑range selection: (A) A representative EIT image shows the selected pixel (ROI 1). The global and pixel‑level (ROI 1) impedance curves illustrate the chosen perfusion time window for two methods—apneic (B) and non‑apneic (C). The time range is defined as the interval from the onset of the pixel impedance decline to the trough of the global impedance curve. For the non‑apneic method, the starting point is set at the moment when the impedance curve falls below the end‑expiratory lung impedance (EELI) line (the blue dashed line in subfigure C). Signal processing: Core innovation: signal separation. For the breath‑hold‑free method, a dual‑cutoff low‑pass filter was employed: 0.67 Hz for analyzing ventilation and 0.17 Hz specifically to extract the perfusion signal. This frequency setting aims to suppress respiratory artifacts (0.2–0.33 Hz) while preserving perfusion information.

The impedance–time curves and pulmonary perfusion images of a representative patient with a respiratory rate of 18 breaths per minute (0.3 Hz) are shown, corresponding to different low‑pass filter cutoff frequencies (0.08 Hz, 0.17 Hz, and 0.25 Hz). Analysis procedure: The pixel exhibiting the greatest tidal impedance change in the dorsal region of the right lung was selected; the time window from the onset of impedance decline to the lowest point on the whole‑lung impedance curve was defined as the perfusion analysis period, enabling precise temporal alignment.

 

PART 03

Results: Consistency between the two methods

 

1. High degree of correspondence with perfusion imaging

The median correlation coefficient for pixel-level perfusion was r = 0.94 (IQR: 0.90–0.97).

Eighty percent of the patients (16/20) exhibited a high degree of correlation (r ≥ 0.9).

Results: In one patient, the perfusion images obtained with and without breath-holding (G and H, respectively) showed a high pixel-level correlation (r = 0.98). The scatter plot (I) clearly demonstrates a strong correlation between the pixel-level perfusion values derived from the two methods. Bland–Altman analysis: Bland–Altman plots for regional perfusion distribution (%) (J) and regional ventilation/perfusion matching (%) (K).

Bland–Altman analysis demonstrated good agreement between the two methods in terms of regional perfusion distribution and V/Q matching.

2. Equivalent core functional parameters

Table 1. Comparison of patient characteristics, pulmonary perfusion parameters, and V/Q imaging patterns for the consistency between apnea‑detecting and non‑apnea‑detecting methods.

There were no statistically significant differences between the two methods in global V/Q matching, dead space ratio, or shunt fraction (p > 0.05), demonstrating equivalence of the functional assessments.

A finding revealing physiological conditions: the duration of impedance decline during breath-holding was significantly shorter (5.3 seconds vs. 6.1 seconds, p = 0.008). This suggests that continuous respiratory motion may accelerate contrast agent clearance, yielding results that more closely reflect physiological states.

Insights from a Few Discordant Cases: Four patients exhibited low correlation (r < 0.9).

Supplementary Figure 2. Analysis of perfusion images and pixel-level correlation maps under conditions of low correlation indicates that the primary cause is respiratory motion, which complicates the selection of the temporal window for perfusion analysis and results in some images being contaminated by cardiac signals. This highlights a clear direction for technical optimization.

 

PART 04

Discussion: Exploring the Expansion of the Audience

This study demonstrates that even a standard linear low-pass filter (0.17 Hz) can effectively remove respiratory artifacts in the majority of patients. This simplicity represents its greatest advantage, facilitating its clinical translation and widespread adoption.

The Value of Breath-Holding Technology

Crowd expansion: Enables critically ill patients who cannot tolerate breath-holding—such as those with severe hypoxia, intracranial hypertension, or hemodynamic instability—to undergo assessment.

Safety Enhancement: It avoids the hemodynamic fluctuations and pulmonary injury risks that may be triggered by breath-holding.

More realistic clinical practice: Monitoring during ongoing treatment may more directly reflect the actual pulmonary perfusion under the current ventilation settings, thereby guiding adjustments to the therapeutic regimen.