Among 117 patients, breath-hold ventilation did not improve strict diagnostic yield vs conventional (86.8% vs 92.2%) in robotic bronchoscopy.
Standardized breath-hold ventilation did not improve diagnostic yield compared with conventional ventilation during robotic bronchoscopy
Respiratory motion and CT-to-body divergence remain leading causes of localization error during robotic bronchoscopy, and breath-hold ventilation has been proposed to reduce target motion and improve image registration.¹ Prospective randomized evidence supporting the technique has been limited.
“Once you have tool-in-lesion confirmation, the analogy I use is a toothpick on an olive,” said Roshen Mathew, MD, DAABIP, interventional pulmonologist at WVU Medicine Camden Clark Medical Center. “You prick the olive, or the lesion, and it’s there inside, and you verify it in 3 different planes with augmented fluoroscopy and keep on biopsying it. Even if you hold the breath or you move, the olive is going to move along with the toothpick.”
In an interview with HCPLive at CHEST 2026, Mathew said breath-hold and apneic ventilation techniques have been used mainly by thought leaders in robotic bronchoscopy and have not transferred widely into general practice.
The Study of Tool-lesion Alignment and Breath-hold Ventilation in Robotic Bronchoscopy for Peripheral Lung LEsions (STABLE) trial was a prospective randomized study of patients undergoing robotic bronchoscopy with the Noah Galaxy Robotic Bronchoscopy System. All procedures used integrated augmented fluoroscopy (AF) and C-arm-based augmented tomosynthesis (CABT), with real-time tool-position confirmation and correction before tissue sampling.¹
Investigators randomized 117 patients to standardized end-inspiratory breath-hold (n = 53) or conventional ventilation (n = 64) using computer-generated simple randomization. The primary endpoint was strict diagnostic yield according to American Thoracic Society/American College of Chest Physicians (ATS/ACCP) definitions.¹,² Mathew noted the mean lesion size was approximately 1.7 cm.
Overall strict diagnostic yield was 89.7%, with no significant difference between the breath-hold and conventional ventilation arms (86.8% vs 92.2%; P =.338).¹ Successful lesion localization was 98.1% and 95.3%, respectively (P =.625). Tool-in-lesion (TIL) was achieved in 62.3% and 64.1% of procedures (P =.841), and imaging-confirmed TIL or tool-touching-lesion (TTL) was achieved in 97.4%.¹
Independent of ventilation strategy, diagnostic yield followed a graded relationship with tool position.¹ Yield was 0% with no tool-lesion relationship (NTL), 85.0% with TTL, and 95.9% with TIL. TIL showed a strong trend toward greater yield compared with TTL (odds ratio [OR], 4.18; 95% CI, 0.98–17.7; P =.064), and both TIL and TTL were significantly associated with greater yield compared with NTL (P <.001 and P =.007, respectively).¹
Radiation exposure, procedural characteristics, and complication rates were comparable between arms. Pneumothorax requiring chest tube placement occurred in 0% of breath-hold cases and 3.1% of conventional ventilation cases (P =.50).¹ Mathew said tool-in-lesion confirmation in the lower lobes was also comparable between arms.
With > 1 year of ongoing follow-up, overall diagnostic accuracy was 93.2%.¹ Mathew said open questions remain, including whether apneic ventilation, the order of tools, or the number of biopsies affects outcomes, and these will likely require multicenter trials.
“Not just having any robot is not enough. A robot can only get you very close,” Mathew said. “Having just static imaging with a cone beam is not enough; you always should have dynamic imaging in 3 or 4 different planes, continually, with your robot and the integrated imaging you have.”
Watch part 1 of our interview with Matthew here:
Editor’s note: Reported disclosures for Matthew include Noah Medical Corporation & Ethicon.
References
