Electromagnetic Navigation Bronchoscopy and Transthoracic Sampling of Peripheral Pulmonary Nodules: One Step Back, One Leap Forward for the Evaluation of Technology Targeting the Lung Periphery

The accurate diagnosis of peripheral lung lesions is arguably more important now, in the era of lung cancer screening and increasing numbers of incidentally detected nodules, than ever before. There have been broad efforts to develop novel guided bronchoscopy techniques targeting the lung periphery and to establish the success rate of these techniques. Accomplishing the latter goal has proven challenging. The guided bronchoscopy literature is plagued with inconsistent definitions of diagnostic yield and methodological weaknesses, including too many retrospective and single-center series (1, 2). One theme that has emerged is that prospectively collected data using strict definitions of diagnostic yield consistently show less impressive results (1, 3–6). Further complicating matters, variable quality of reporting has occurred alongside the rapid innovation in guided bronchoscopy and eager uptake on the basis of promise rather than proof.

In this issue of the Journal, Thiboutot and colleagues (pp. 837–845) report a single-arm, multicenter prospective trial evaluating the diagnostic yield of relatively older electromagnetic navigation (EMN) technology (7). The investigators evaluated the performance of a combined staged procedure for the workup of peripheral pulmonary nodules (<3 cm). Each of the 160 enrolled subjects first underwent linear endobronchial ultrasound (EBUS) for nodal staging, followed by an EMN-guided bronchoscopy (ENB) guided by a computed tomography (CT) scan acquired the same day Physicians proceeded with ENB if there was: 1) no resolution on same-day CT and 2) no rapid on-site evaluation–confirmed malignancy by linear EBUS. For participants with a lung nodule localized in the outer third of the lung, an EMN-guided transthoracic needle aspiration and biopsy (EMN-TTNA) was performed following ENB if on-site cytology did not detect malignancy from peripheral needle aspirations and/or bronchial brushings. It is worth noting that the EMN-TTNA (SPiNPerc Kit, Veran Inc.) was performed using the same device tower as the ENB (SPiN Thoracic Navigation System, Veran Inc.), allowing both biopsy techniques to be performed in a single procedural encounter. Given the prospective design and inclusion criteria, patient selection enrollment was open to all patients undergoing lung nodule biopsy, ostensibly without further refinement of patient selection based on anatomy or predictors of success. The procedure did not include commonly used guidance tools such as two-dimensional fluoroscopy and/or radial EBUS. Importantly, the authors used a strict definition of diagnostic yield based on pathological findings of malignancy or a specific benign diagnosis, and all cases underwent central pathology review.

At first glance, the results reported by Thiboutot and colleagues are disappointing. Only 58.8% of subjects were definitively diagnosed by the staged procedure in a study population with a mean nodule size of 1.8 cm and a cancer prevalence of 67.3%. After accounting for the diagnoses made by same-day CT (nodule regression seen in 2.5%) and EBUS-guided lymph node sampling (7.1%), ENB alone had a diagnostic yield of 49.3%. When performed, EMN-TTNA had an even lower diagnostic yield of 26.7%. The combined approach was associated with a pneumothorax rate of 10%, with pneumothorax occurring in 11 of 30 EMN-TTNA procedures (37%). It is unfortunately not the first time that disappointing results have been reported for advanced diagnostic bronchoscopy procedures. These results are similar to prior prospective evaluations of ENB using a strict definition of diagnostic yield (4). The 49% diagnostic yield is significantly worse (>10% lower), however, compared with prior ENB studies that used less strict definitions of diagnostic yield, enrolled participants with larger lung nodules on average and a higher prevalence of malignancy, and allowed for the addition of fluoroscopy and radial EBUS to ENB (8–11).

The use of a strict definition of diagnostic yield based exclusively on pathology findings from the procedural encounter is an important methodological strength of the study. The inclusion of clinical follow-up into the definition of what constitutes a diagnostic procedure results in erroneously optimistic estimates of diagnostic yield (2). Clinicians and patients are interested in knowing the likelihood that a bronchoscopy will provide information that guides clinical care immediately after the procedure, not 12–24 months later. As the authors point out, the study represents a benchmark for the performance of ENB alone in a population of smaller nodules (mean nodule size <2 cm) at intermediate to high risk of lung cancer. There are certain weaknesses born out of the decision to study ENB alone. One could argue that a benchmark for ENB combined with widely available, relatively inexpensive additive technologies such as standard fluoroscopy and radial EBUS would serve as a more clinically relevant standard. As noted by the authors, there was no attempt to compare biopsy techniques given the single-arm study design.

How do we place this study in the context of the aforementioned literature? Does this mark the end of ENB alone as a clinically acceptable procedure? It appears that the expectations of an ENB procedure should indeed be guarded. Furthermore, most would agree that EMN-TTNA has an unacceptable risk/benefit profile in its present form. That being said, ENB alone may still be acceptable for the biopsy of larger lesions with a bronchus sign (11, 12). The study also provides insight for those planning a prospective study involving an emerging technology, for which the diagnostic yield of ENB alone may serve as a reference point. Finally, the fact that 2.5% of participants enrolled had nodule regression on their day-of-procedure CT scan is important to keep in mind. Repeat CT scanning may be a reasonable preprocedural step, particularly in those with older initial CT scans or low- or intermediate-risk nodules, but should be balanced against the added expense and radiation exposure.

Overall, the study by Thiboutot and colleagues (7) should act as a springboard for future investigation of guided bronchoscopy. High-quality prospective data with strict definitions of diagnostic yield are needed more than ever as we look to delineate the value of robotic bronchoscopy, augmented fluoroscopy, intraprocedural advanced imaging, novel forms of on-site tissue evaluation, and new biopsy tools. Comparative studies, preferably randomized, are also needed, with attention to the incorporation of patient-centered and patient-reported outcome measures.