Introduction

Diagnosis of peripheral pulmonary lesions (PPLs) is challenging in clinical practice. The incidence of pulmonary abnormal lesions has been increasing owing to the widespread adoption of chest CT screening [1, 2]. Conventional bronchoscopic techniques have been used in lung cancer diagnosis for decades. However, PPLs beyond the segmental bronchus are difficult to sample under direct visualization with a conventional bronchoscope as its outer diameter (5–6 mm) is too large to easily access beyond the third or fourth generation bronchi. When sampling such peripheral PPLs, several methods including transbronchial biopsy (TBB), transbronchial needle aspiration (TBNA), transbronchial brushing, and bronchial washing can be performed sequentially. Even with fluoroscopic guidance, the diagnostic yield of conventional bronchoscopy is poor. According to analysis of the AQuIRE (ACCP Quality Improvement Registry, Evaluation, and Education) registry, the sensitivity of conventional bronchoscopy without advanced navigation assistance [e.g. radial probe endobronchial ultrasound (RP-EBUS) or electromagnetic navigation bronchoscopy (ENB)] was 54% for TBB with single-plane fluoroscopy, 65% for TBNA with fluoroscopic guidance, 54% for transbronchial brushing and 53% for bronchial washing [3]. Conventional bronchoscopy does not have sufficient sensitivity for reliable diagnosis of PPLs due to poor access to target lesions.

Various navigational technologies and techniques have been developed to overcome the limitations of conventional bronchoscopy. The objective of this review is to provide an overview of the recent development in bronchoscopic navigation for sampling and diagnosis of PPLs. This will include a review of key bronchoscopic technologies that are complementary to navigation bronchoscopy platforms, such as thin/ultrathin bronchoscopy and RP-EBUS.

Complementary technologies for navigational bronchoscopy

Thin/ultrathin bronchoscopy

Approaching PPLs by conventional bronchoscopy is limited by its large outer diameter and restricted range of motion. Thin and ultrathin bronchoscopes have smaller outer diameters and greater range of motion to improve access to subsegmental bronchi, including those in the upper lobes. Although the exact definition of conventional and thin/ultrathin bronchoscopes does not exist, the former usually have an OD of ≥ 5.0 mm, and ultrathin bronchoscopes are considered to have an OD of ≤ 3.0 mm. While conventional bronchoscopes can visualize the third or fourth generation bronchi at most, ultrathin bronchoscopes can reach up to the ninth bronchial generation [4].

Thin/ultrathin bronchoscopy is often performed combining other image-guided techniques (CT guidance, virtual navigation [VBN], RP-EBUS), which can improve routing and/or target localization. One single-center study demonstrated that ultrathin bronchoscopy combined with CT guidance and VBN diagnosed small PPLs (<20 mm) with a sensitivity of 65% [5]. A prospective multicenter study of ultrathin bronchoscopy combined with VBN showed in its subgroup analysis higher yields for lesions in the right upper lobe, peripheral third of the lung, or that are invisible on posterior–anterior radiographs compared to the non-VBN-assisted group [4]. Another prospective study found the diagnostic yield of a combined thin bronchoscope and RP-EBUS approach was as high as 80% for malignant peripheral lesions [6]. One randomized prospective multicenter study compared an ultrathin bronchoscope (OD: 3.0 mm) to a thin bronchoscope (OD: 4.0 mm) for diagnosis in PPLs less than 3 cm in size [7]. Guidance with fluoroscopy, VBN and/or RP-EBUS were allowed in both groups. Despite the 1 mm difference in diameter, yield with ultrathin bronchoscopy was superior to thin bronchoscopy: 81% vs. 70% for malignant lesions, and 42% vs. 36% for benign lesions. This study suggested that a smaller bronchoscope supported with advanced navigational support improves diagnosis of PPLs. This was validated by a recent study from the same group [8], which also showed that ultrathin bronchoscopy combined with multimodal navigation resulted in significantly shorter procedure time compared to thin bronchoscopy.

Major complications of thin/ultrathin bronchoscopy were not reported in most clinical studies. However, pneumothoraces due to visceral pleural perforation caused by an ultrathin bronchoscope have been reported [9]. The ultrathin bronchoscope should be manipulated with care, especially in the peripheral lung.

Radial probe endobronchial ultrasound (RP-EBUS)

RP-EBUS is a flexible catheter with an ultrasound probe mounted on the end, which generates a 360-degree radial ultrasound field of view (Fig. 1). RP-EBUS is now widely used as an effective tool for localization of PPLs before biopsy.

Fig. 1
figure 1

Radial probe endobronchial ultrasound (RP-EBUS). a The appearance of RP-EBUS (UM-S20-17S, Olympus) with a guide sheath. b The fluoroscopic image of approaching the target using RP-EBUS. c The RP-EBUS image of a peripheral pulmonary lesion as a hypoechoic lesion

Full size image

The RP-EBUS can be passed through the working channel of a bronchoscope because of its small OD (1.4 mm), allowing it to reach distal bronchi where the bronchoscope cannot reach. The probe can characterize the densities of surrounding tissues in distal airways. Normal lung parenchyma appears as a “snowstorm” pattern due to the air-filled alveolar tissue. PPLs and masses appear as hypoechoic lesions with hyperechoic lines demarcating the lesions from normal lung parenchyma. Pure ground-glass opacities can also be detected by RP-EBUS, which is described as a “blizzard” pattern due to the increased echogenicity with enlarged acoustic shadows [10]. After the target lesion is identified by RP-EBUS, the radial probe is withdrawn and replaced by a sampling device such as biopsy forceps or a brush. Key to this technique it to ensure the bronchoscope tip remains in position, as this allows the sampling device to be deployed in the same location as where the radial probe identified the target lesion. A guide sheath (GS) was developed to assist in this task, which fits over the radial probe. The radial probe with GS is advanced together into the target bronchus through the working channel of the bronchoscope, leaving the distal tip of the RP-EBUS exposed. Once the location of the target lesion is confirmed by RP-EBUS, the radial probe is removed while the GS stays in place; the GS serves as a path for the sampling devices.

A meta-analysis of 16 studies evaluating RP-EBUS for diagnosis of PPLs in a total of 1420 patients reported that a pooled sensitivity of 73% (range 49–88%) [11]. A larger meta-analysis of navigational bronchoscopy including RP-EBUS with and without GS, ultrathin bronchoscopy, VBN, and ENB including a total of 3052 lesions across 39 studies reported a pooled diagnostic yield of RP-EBUS of 73% and 71% with and without GS, respectively [12]. This is higher than the yield for conventional TBB. Several factors that affect the diagnostic yield of RP-EBUS have been identified: the size of the lesions, the location of the lesions, and the location of the radial probe in relation to the lesions. A multivariate analysis found that larger lesions (≥20 mm) and shorter distance from the hilum (≤50 mm) was associated with improved diagnostic yield [13]. The relative location of the radial probe to the target lesion has also been shown to be a significant factor for diagnostic yield [14,15,16]; ‘concentric’ lesions that surround the probe have a significantly higher diagnostic yield (87%) than ‘eccentric’ lesions that are off-axis from the probe (42%) [14]. The complication rates of RP-EBUS are low; pooled rates of pneumothoraces are 1% and chest tube placement are 0.4% [11].

Navigational bronchoscopy technologies and techniques

Virtual bronchoscopic navigation (VBN)

One of the challenges in bronchoscopy for small peripheral lesions is selecting the correct bronchus to reach the target. VBN is a navigational technique that uses imaging data from CT or MRI to construct a virtual bronchial tree and endoscopic view for routing to the target lesion. A pre-procedure chest scan is transferred to a computer workstation which then automatically generates the virtual reconstruction with or without manual modification. Some platforms can synchronize the virtual endoscopic view with the real-time endoscopic view to provide a route to the target.

One randomized controlled study including 199 patients with small PPLs < 30 mm in diameter found that VBN had significantly higher diagnostic yield than control (80% vs. 67%) when diagnostic bronchoscopy was performed with RP-EBUS [17]. Another prospective randomized multicenter trial compared VBN to control for PPLs < 30 mm using an ultrathin bronchoscope and fluoroscopic guidance without RP-EBUS [4]. Although there was no significant difference in diagnostic yield, the group with VBN had a trend towards significance (67% vs. 60%). According to a systematic review, the average diagnostic yield using VBN was 74% in all lesions and 67% in small lesions < 20 mm, although the studies varied in adjunct navigational modalities used [18].

Several points should be kept in mind when employing VBN. Firstly, VBN is a planning tool, not a real-time imaging technique. VBN image is created from imaging data prior to bronchoscopy, and the reconstruction’s quality depends on the image source data. Second, there is a potential risk for an inappropriate route to be mapped without careful review by a skilled operator.

Electromagnetic navigation bronchoscopy (ENB)

ENB is a navigational technique that generates a virtual 3D reconstruction of the bronchial tree from CT scan data and superimposes the real-time position of the bronchoscope instruments using electromagnetic sensors. These sensors are located on the tip of ENB instruments that are inserted through the bronchoscope working channel. An electromagnetic field generator outside the patient’s body tracks the position of the sensors.

There are currently two commercially available ENB systems: superDimension (Medtronic, Minneapolis, Minnesota, USA) and SPiNDrive (Veran Medical Technologies, Inc, St Louis, Missouri, USA) (Fig. 2). The electromagnetic field generator is either a board under the patient (superDimension) or a square pad for the SPiNDrive. The electromagnetic sensor in the superDimension is placed within the tip of a steerable guide (called “the locatable guide”) which is used housed in a larger guide sheath (“the extended working channel”). Once the tip of the locatable guide reaches the target lesion, the locatable guide is removed while the extended working channel is locked onto the bronchoscope. Sampling instruments are inserted into the extended working channel. In contrast, SPiNDrive utilizes a continuous tracking system; the electromagnetic sensors are embedded in the tip of the sampling instruments, which allows continuous electromagnetic tracking even during sampling.

Fig. 2
figure 2

Electromagnetic navigation bronchoscopy. A screen image of the a superDimension system, and b SPiNDrive system

Full size image

The pooled diagnostic yield of ENB was about 65% according to a meta-analysis published in 2014 that included 15 studies [19]. A recently published prospective multicenter trial evaluating superDimension, part of the larger NAVIGATE study [20], showed a 73% diagnostic yield for 1157 lung lesions with an average size of 20 mm [21]. Several factors associated with higher diagnostic yield have been identified: larger lesion size [22, 23], and a positive bronchus sign [22]. According to a meta-analysis, the complications of ENB for peripheral lesions included pneumothoraces in 3.1%, with chest tube drainage needed in 1.6% of all cases, and minor or moderate bleeding in 0.9% [19]. This is similar to conventional bronchoscopy.

Bronchoscopic transparenchymal nodule access (BTPNA) / transbronchial access tool (TBAT)

Bronchoscopic transparenchymal nodule access (BTPNA) and transbronchial access tool (TBAT) are emerging techniques for overcoming the limitations of previous navigational techniques. The presence of a bronchus leading to the target is a strong predictive factor for successful diagnosis of peripheral lesions with techniques such as RP-EBUS and ENB. To better approach lesions without a bronchus sign, BTPNA or TBAT has been developed.

BTPNA generates a route to the PPL through the lung parenchyma, integrated as part of the larger Archimedes platform (Broncus Medical, Inc, San Jose, California, USA). Archimedes is a VBN system, integrating bronchoscopic images, CT data and fused fluoroscopic images to provide a 3D reconstructed real-time airway image. This includes highlighting a point-of-entry in the central airway and a lung parenchymal route that avoids blood vessels. Based on this virtual guidance, a coring needle punctures the planned point-of-entry, which is dilated by a balloon catheter. A sheath with a blunt dissection stylet is then advanced to the target.

The first report of BTPNA in humans by Herth and colleagues was published in 2015 [24]. This study showed that adequate sampling was achieved in 10 out of 12 patients (83%) using Archimedes and fluoroscopy guidance, though nodules were not visible by fluoroscopy in seven patients. Although no severe postprocedural complications in BTPNA have been reported thus far [24, 25], further investigation on safety is necessary. A large multicenter international trial to assess the utility of BTPNA is currently recruiting patients, which will likely provide more clear data on the technique’s safety (EAST-2 trial: ClinicalTrial.gov registration no. NCT02867371).

Another technique for approaching lesions without a bronchus sign is TBAT, which is part of the CrossCountry system (Medtronic, Minneapolis, Minnesota, USA). This platform integrates with the superDimension system described in the ENB section. According to a virtual navigation pathway generated by the superDimension system, a guidewire needle punctures the bronchial wall creating an entry point to the lung parenchyma. A cone-shaped dilator catheter is advanced over the guidewire toward the target, followed by placement of the extended working channel over the dilator. After removal of the guidewire and the dilator, biopsy instruments are deployed through the extended working channel. Several case series are available on the use of TBAT [26,27,28]. Most cases underwent TBAT under general anesthesia with cone-beam CT confirmation of the sampling location, though one case series demonstrated TBAT’s use under conscious sedation [26].

These new techniques are promising for sampling lesions without a bronchus sign. Despite its invasiveness, no severe adverse events have been reported so far. However, it is too early to verify the safety of these techniques given because of the scare evidence. Further studies are needed to confirm safety.

Robotic bronchoscopy

Robotic-assisted technology has become increasingly common in surgery [29, 30] and endoscopy [31, 32]. It has been similarly applied to interventional bronchoscopy in hopes of overcoming limitations of conventional bronchoscopy and current navigation technologies. There are currently two robotic bronchoscopy platforms cleared by the FDA; the Monarch by Auris Health Inc. (Redwood City, California, USA), and the Ion by Intuitive Surgical Inc. (Sunnyvale, California, USA) (Fig. 3).

Fig. 3
figure 3

Robotic bronchoscopy. a–d The Monarch platform (Auris Health Inc.). e The Ion platform (Intuitive Surgical Inc.)

Full size image

The Monarch platform was the first FDA-cleared robotic bronchoscopy platform. It consists of an inner bronchoscope and an outer sheath (OD 4.2 mm and 6.0 mm, respectively) that are robotically propelled with 4-way steering via a hand-held controller; these can be locked once the optimal location is reached [33, 34]. The Monarch platform uses EM navigation for guidance. The inner bronchoscope includes a camera, light source, and a 2.1 mm working channel [34]. A report in human cadaveric lungs found that robotic bronchoscopy was able to be advanced further into the peripheral bronchus compared to a thin bronchoscope, despite the fact that the OD of the inner bronchoscope is same as the thin bronchoscope [35]. This was attributed to the Monarch’s outer sheath providing the inner bronchoscope additional structural support. Furthermore, wedging by the outer sheath is thought to potentially reduce complications by tamponading biopsy-related bleeding [34]. A feasibility study of the Monarch platform in 15 patients acquired tissue under direct visualization in 14 patients (93%), with no severe adverse events.

The Ion platform differs from the Monarch through the use of a trackball controller. The bronchoscope catheter has a 3.5 mm OD with a 2.0 mm working channel and is flexible enough to bend almost 180 degrees. A camera-probe is used for live visualization while advancing the catheter. This is supplemented by virtual navigation generated from a patient’s CT images. Once the bronchoscope catheter advances close to the target, the camera-probe is removed, and sampling instruments are inserted through the catheter. While switching the instruments, the catheter can hold its position. Shape sensors embedded in the catheter wall monitor the shape of the catheter and provide locational information during navigation and biopsies, which allows continuous stability of the tip to maintain the sampling position. A first report of the Ion platform in 29 patients with small nodules (mean largest diameter 14.8 mm) demonstrated successful target access in 28 patients (97%) as confirmed by RP-EBUS. Diagnostic yield for malignancy was 88% through 6-month follow-up, despite the fact that CT bronchus sign was absent in more than 40% of lesions and only half of the lesions were ‘concentric’ lesions on RP-EBUS.

The strengths of robotic bronchoscopy may include the flexibility to advance into more peripheral airways with continuous visualization and the stability of the bronchoscope to maintain a static position for biopsy. Although reports about the utility and the safety are slowly accumulating, the evidence as a whole is still preliminary. Several ongoing multicenter studies (ClinicalTrial.gov registration no. NCT03727425, NCT03893539) may more clearly demonstrate the potential advantages, diagnostic yield, and safety of these new robotic bronchoscopy platforms.

Obtaining a tissue diagnosis in peripheral pulmonary lesions

Just as methods to improve navigation to the target have advanced, so too have the sampling methods themselves (i.e. beyond standard TBB, TBNA, brushing, and washing).

To acquire larger specimens with better quality for histopathological assessment, a cryobiopsy technique has been developed for bronchoscopy. A cryoprobe is passed through the working channel of the bronchoscope and placed near to the target. The probe is then rapidly cooled to − 79 °C (− 110°K) with carbon dioxide or − 89 °C (− 128°K) with nitrogen oxide, thus freezing the tissue of the target in contact with the probe followed by a tissue removal. The use of cryoprobes have been reported for airway recanalization [36], interstitial lung disease [37, 38], sampling endobronchial lesions [39] and PPLs [40]. In comparison with forceps biopsy, larger and better-preserved samples can be obtained.

Conclusion

Innovative bronchoscopic techniques and technologies have been developed in recent years that allow precise approach to peripheral lung lesions with improved diagnostic yield. Greatest success is achieved when each technique is not used in isolation, but as adjuncts for one another. For example, the successful integration of robotic bronchoscopy, ENB, and RP-EBUS, as reported as others. It is important for physicians to understand the strengths and limitations of each of these technologies, and to select the combination of modalities that best address the needs of a given patient.