Keyes, Colleen

Abstract INTRODUCTION Guidelines recommend non-surgical biopsy for indeterminate pulmonary lesions > 8 mm. Shape-sensing robotic-assisted bronchoscopy (ssRAB) is growing in use to biopsy small pulmonary nodules; however current literature is limited to single center studies with limited follow-up. This study, PRECIsE, assessed safety and performance of the first ssRAB iteration across multiple sites and nascent users. METHODS Prospective, multicenter, observational study evaluating ssRAB without cone-beam CT guidance in patients with nodules 10–30 mm located in or beyond the sub-segmental airways. Patients were followed for 2 years; the primary endpoint was sensitivity for malignancy with diagnostic yield, and safety as secondary endpoints. Multilevel logistic regression models were used to control for factors associated with sensitivity and diagnostic yield. RESULTS A total of 305 procedures were performed across 6 centers. Median nodule size was 17.0 mm (IQR: 14.0–23.0) with bronchus sign present in 37% of cases. Sensitivity for malignancy through 2 years was 81.3% (95% CI: 75.7, 86.1). Multi-level modelling demonstrated female sex, smaller nodule size, lower lobe location, semi-solid density, and higher body-mass-index were associated with lower sensitivity. Diagnostic yield was 74.1% (95% CI: 68.8, 78.9) according to the ATS/ACCP criteria and 77.5% [95% CI: 72.4, 81.8] according to the intermediate criteria. Multi-level modeling demonstrated a non-significant site/center level effect on ATS/ACCP diagnostic yield (P = .36). Pneumothorax requiring intervention was 1.6% (5/305); bleeding was 1.0% (3/305) with two (2) Nashville grade 2 and one (1) Nashville grade 3 events. CONCLUSIONS The first iteration of ssRAB demonstrated encouraging performance and a strong safety profile among nascent users for the biopsy of small peripheral nodules.

Abstract Introduction: Endobronchial optical coherence tomography (EB-OCT) imaging has unique potential to diagnose and quantify early, microscopic disease changes and assess therapeutic responsivity in interstitial lung disease (ILD). However, training and time requirements for interpretation and quantitative evaluation of large volumetric EB-OCT datasets need to be optimized to facilitate clinical adoption. We develop and validate a computationally-efficient, rapid EB-OCT quantitative image analysis framework using artificial intelligence (AI) with deep learning architecture for 2D feature segmentation, 3D feature burden and spatial distribution mapping, and computer-aided diagnosis in early ILD subjects. Methods: Volumetric EB-OCT datasets, consisting of 2D cross-sectional images, were acquired from multiple locations in the bilateral lungs of ILD and healthy control subjects. EB-OCT features were manually segmented from cross-sectional images by an expert EB-OCT reader, including fibrosis, microscopic honeycombing, traction bronchiolectasis, normal parenchyma, and emphysema. EB-OCT images with corresponding feature segmentations were used to train and validate a multiclass convolutional neural network. The 2D AI-based segmentation model was subsequently evaluated against manual ‘ground truth’ segmentations for each feature in a de novo ‘test’ dataset of independent subjects using dice similarity score, precision, recall/sensitivity, specificity, and balanced accuracy. To generate 3D AI-based quantitative feature burden and spatial distribution maps, AI-based feature segmentation was performed on each EB-OCT 2D cross-sectional image within a volumetric imaging site. Three external pathologist readers independently interpreted 3D AI-based features maps for each EB-OCT imaging site acquired for a subject (without access to EB-OCT images) and provided a single diagnosis of usual interstitial pneumonia (UIP) or non-UIP ILD. Performance of computer-aided ILD diagnosis was compared against independently evaluated histopathology, clinical diagnosis, and prior EB-OCT image interpretation. Results: EB-OCT was acquired in 51 subjects: 30 in vivo early ILD, 5 in vivo non-ILD controls, and 16 ex vivo end-stage ILD. Overall pixel similarity between 2D manual and AI-based segmentation demonstrated high accuracy for each ILD feature (mean balanced accuracy 0.9). 3D quantitative AI-based feature maps demonstrated strong agreement (Spearman ρ0.87, p<0.001) between manual and AI-based segmentation. AI-based computer-aided diagnosis demonstrated 100% sensitivity/specificity for histologic UIP and clinical IPF for all pathologist readers, with improved diagnostic accuracy and significantly reduced training (45-minute training, 75% reduction) and interpretation time (<1 minute per subject for computer-aided diagnosis, 88% reduction) compared to EB-OCT image interpretation. Conclusion: The AI-based EB-OCT image analysis framework could expedite computer-aided diagnosis of early ILD and robust, quantitative assessment of microscopic disease burden, progression, and therapeutic responsivity.

AbstractBackground and ObjectiveShape‐sensing robotic‐assisted bronchoscopy (ssRAB) has expanded as an important diagnostic tool for peripheral pulmonary nodules (PPNs), with diagnostic yields ranging from 60% to 88%. However, sampling and diagnosing PPN less than 2 cm in size has historically been challenging. Mobile cone‐beam computed tomography (mCBCT) has been recently integrated into ssRAB to improve diagnostic accuracy, but its added value remains uncertain. We aim to describe the role of mCBCT and determine if it provides any diagnostic advantage.MethodsA multicentre, retrospective study on the use of ssRAB and mCBCT in two tertiary care institutions: Mayo Clinic Florida and Massachusetts General Hospital. The primary outcome was diagnostic yield and sensitivity for malignancy of ssRAB complemented with mCBCT, compared to ssRAB with the standard 2D fluoroscopy.ResultsA total of 192 nodules were biopsied from 173 patients. mCBCT was used in 117 (60.9%) nodules. The overall diagnostic yield was 85.4%. Diagnostic yield between subgroups with and without mCBCT was 83.8% and 88% (p = 0.417), respectively. The mCBCT group had fewer solid nodules (65.8% vs. 81.3%, p = 0.020) and a higher number of ground‐glass nodules (10.3% vs. 1.3%, p = 0.016).ConclusionOverall, diagnostic yield between subgroups with and without mCBCT was similar. The complementary use of mCBCT to ssRAB allows proceduralists to target more complex and subsolid PPNs with a diagnostic yield comparable to simple solid PPNs while maintaining an excellent safety profile.