First in Human Use of a Hybrid Real-Time Ultrasound-Guided Fine Needle Acquisition System for Peripheral Pulmonary Lesions: A Multicenter Pilot Study

Abstract

Background:

The ability to successfully biopsy pulmonary lesions with bronchoscopy varies widely due to anatomic and technological limitations. One major limitation is the lack of the ability to utilize real time guidance during tissue sampling in the periphery. A novel system has been developed enabling real-time visualization and sampling of peripheral lesions by displaying an ultrasound image of the lesion and needle simultaneously.

Methods:

We performed a multicenter prospective pilot in patients with peripheral pulmonary lesions undergoing a clinically indicated bronchoscopy. The purpose of this study was to demonstrate feasibility of visualizing, accessing, and obtaining specimens adequate for cytology of lung lesions when using a novel hybrid real-time ultrasound-guided fine needle aspiration system for peripheral pulmonary lesions.

Results:

Twenty-three patients underwent bronchoscopic sampling of a peripheral pulmonary lesion with the study device. Mean lesion size was 3.6cm (range 1.7–5.7cm). Targeted lesions were located in all lobes of the lung. All lesions were successfully visualized and sampled under real-time visualization with specimens adequate for cytological evaluation. The needle was visualized in all lesions throughout targeting and sampling. There were no pneumothoraces or moderate to severe bleeding encountered.

Conclusions:

In this feasibility study, we report the first in human use of a continuous real-time endobronchial ultrasound guidance system to sample peripheral pulmonary lesions. Future generations of this device may improve usability and further studies are needed to determine the true diagnostic capabilities of this novel technique.

INTRODUCTION

The most common techniques performed for the diagnosis of peripheral lung lesions are bronchoscopy, surgical biopsy and CT-guided transthoracic needle biopsy (CT-TTNA)¹. Though surgical and CT-guided biopsies are often conclusive, they can lead to significant morbidity^2,3. Surgical biopsies have the highest diagnostic yield, but also suffer from higher morbidity than CT-guided biopsies or bronchoscopy^2–4. Although less invasive than surgery, CT-TTNA carries a higher risk of pneumothorax when compared with bronchoscopy⁵.

Less invasive techniques of diagnosing peripheral pulmonary lesion are performed primarily by using bronchoscopy with conventional radial endobronchial ultrasound (CR-EBUS) or electromagnetic navigational bronchoscopy (ENB). CR-EBUS requires the use of an ultrasound catheter, in which the catheter is positioned until visualization of the lesion is confirmed. The ultrasound catheter is subsequently removed and sampling instruments are then advanced blindly without ultrasound/image guidance into the distal airways with an attempt to obtain diagnostic tissue. This procedure is safer when compared to surgery or CT-TTNA, but with a lower diagnostic yield, ranging from 46–86%^6–9. The lower diagnostic yield of guided bronchoscopy may be due to multiple factors including the absence of real time imaging during sample acquisition and respiratory variation with nodule target movement¹⁰.

ENB uses virtual bronchoscopy, in combination with an electromagnetic field to provide guidance during bronchoscopic sampling. Using a pre-procedural CT scan and a software-based tracking system, the bronchoscopist is able to navigate to the peripheral nodule using computer generated airway mapping. The diagnostic yield of ENB is highly variable in the literature ranging from 38.5–94%^8,11,12. Similar to CR-EBUS, there is no real-time visualization of lesion sampling since the image guidance and tracking are computer generated based on pre-procedural CT mapping.

While central nodules are easier to biopsy in real-time using convex endobronchial ultrasound guided transbronchial needle aspiration, we currently lack the technology capable of obtaining tissue of peripheral lesions under direct visualization. We evaluated the safety and early efficacy of a novel radial endobronchial ultrasound prototype with a needle located adjacent to the catheter (Figure 1), which allows for real-time needle positioning and tissue acquisition of peripheral nodules using transbronchial needle aspiration (RT-EBUS-TBNA) (Boston Scientific, Marlborough, MA).

Figure 1.

Real time radial endobronchial ultrasound biopsy tool. The magnified insert illustration shows the proximity of the needle to the ultrasound allowing for real time visualization of lesion sampling.

METHODS

This was a multi-centered, prospective safety and feasibility pilot study conducted at Johns Hopkins University, Washington University School of Medicine, and the Medical University of South Carolina, on the use of a novel radial endobronchial ultrasound with a real-time sampling needle for the biopsy of peripheral pulmonary nodules from January 18^th 2017 through June 8^th 2017. Institutional review board approval was obtained from all institutions (IRB00101395) prior to the study start date. This trial was registered on ClinicalTrials.gov (NCT02832284) and FDA approval under an IDE was obtained to perform this in-human clinical study.

Inclusion criteria included age greater than 18, a predominately solid lung lesion (at least 80%) that is 1cm to 7cm in diameter identified on chest CT, intention to undergo bronchoscopy evaluation under routine clinical care and the decision to pursue biopsy. Participants were excluded if the lesion was greater than 20% ground glass or sub-solid, if the lesion had endobronchial involvement on chest CT, if the participant was unable to undergo bronchoscopy, had a known coagulopathy, were pregnant or nursing mothers, or were currently enrolled in another investigational study that would interfere with the current study.

At the baseline visit, informed consent, medical history and a chest CT image within six weeks of the planned study procedure to confirm the presence of peripheral pulmonary lesions were obtained. All patients initially underwent convex EBUS-TBNA for mediastinal and hilar staging. This was followed by the study procedure where the participant had tissue sampling using the novel RT-EBUS-TBNA system under deep sedation.

The RT-EBUS-TBNA system includes a specially constructed single lumen flexible sheath with a radial ultrasound catheter as well as a specially engineered needle to be passed within the same sheath to establish real-time visualization of lesion sampling (Boston Scientific, Marlborough, MA, USA). The device has an outer diameter of 1.9mm so as to be compatible with a 2.0mm working channel. The sheath is comprised of a disposable 40MHz, 1.1mm ultrasound probe with 25g needle in a single lumen (Boston Scientific, Marlborough, MA, USA).

All participants enrolled underwent flexible bronchoscopy with the investigational RT-EBUS-TBNA biopsy system under fluoroscopic guidance using a 4.2 mm bronchoscope with a 2.0mm working channel (Olympus BF-P190 or MP-160, Olympus America). Nodule localization was accomplished using CT scan anatomic guidance and fluoroscopy. Virtual or electromagnetic navigation platforms were not used in this study although since the outer diameter of the study device is 1.9mm it conceptually could be used with any approximately sized extended working channel greater than 2.0mm. Once the bronchoscope was positioned adjacent to the target, the guidesheath which included the RT-EBUS-TBNA device was passed through the working channel of the bronchoscope and the radial ultrasound probe was advanced into the parenchyma using the same technique as CR-EBUS under fluoroscopic guidance.⁶ No curette device was used during this study. Once an ultrasound view was obtained, the probe was manipulated to obtain the optimal ultrasound image for nodule access. Under fluoroscopic guidance, the TBNA needle was then deployed under direct ultrasound visualization revealing the needle position as either concentric or eccentric to the lesion which was feasible due to the device design which allowed for separation of the radial probe and needle at approximately 10 degrees. (Figure 2) If the needle was concentrically placed, cytologic samples were obtained as described below. If the needle was eccentric to the lesion, the needle was retracted and the entire sheathed device allowed for rotation and repositioning of the device at which point the needle was redeployed in the correct plane and samples were obtained. The needle possesses a stroke limiter on the handle to limit the needle depth to a maximum of 2.5 cm to limit risk over extension of the needle and potential pleural injury.

Figure 2.

Distal tip of radial endobronchial ultrasound biopsy tool with a 10 degree separation between the ultrasound probe and needle.

All samples obtained were assessed using rapid on-site evaluation (ROSE) for specimen adequacy as well as diagnostic yield. After a RT-EBUS-TBNA sample was taken, the aspirate was expressed onto glass slides. The first smear was air-dried and stained immediately with a rapid Giemsa-type stain (Diff-Quick [DQ], Fisher Scientific, Kalamazoo, MI) and the second was fixed in 95% alcohol for permanent cytologic examination using the Papanicolaou stain. Criteria for ROSE sample adequacy was defined as the presence of cellular material which could be stained and evaluated by the onsite cytologist and/or a diagnostic sample (i.e adenocarcinoma). If the sample contained no cellular material (i.e. blood only) it was considered inadequate by ROSE. Diagnostic yield was defined as the fraction of lesion(s) for which a diagnosis could be obtained based on multi-disciplinary evaluation of all information available, including but not limited to analysis of the RT-EBUS-TBNA aspirate. A sample was considered diagnostic during the index procedure if the ROSE evaluation demonstrated a specific malignant or benign diagnosis (i.e granulomatous inflammation consistent with sarcoidosis). The final pathology reports were considered the diagnostic gold standard. If ROSE samples were inadequate for cytological evaluation or the samples were not diagnostic on-site as defined above, additional standard of care procedures were performed using CR-EBUS with TBNA, brush and/or transbronchial biopsy.

The primary endpoint was the ability to acquire adequate specimens of cellular material suitable for cytological evaluation of the targeted lesion under real-time visualization as defined above. Secondary endpoints included the occurrence and severity of adverse events related to the procedure, adverse events related to any subsequent standard of care technique, proportion of lesions visualized using the RT-EBUS-TBNA system, proportion of lesions accessed where RT-EBUS biopsy needles were deployed during the study maneuvers, and the proportion of RT-EBUS maneuvers that acquire specimens of cellular matter for cytology.

Descriptive statistics were used to summarize the end-points of this prospective observational pilot study. Mean (with standard deviation) was used to describe continuous variables with a normal distribution, and median (with interquartile range) was used to describe continuous variables with a skewed distribution. Frequency tables were used to summarize discrete variables, and proportions were used for adverse event data. We conducted statistics using an intention-to-treat (ITT) cohort, consisting of subjects enrolled in the study regardless of whether a specimen was obtained by RT-EBUS. Statistical analysis was conducted using SAS version 9.3 (Cary, NC).

RESULTS

Twenty-three participants were recruited and underwent bronchoscopy. Baseline demographics can be found in Table 1. The mean diameter of the peripheral lung lesions was 3.6cm (range, 1.7–5.7 cm). Targeted lesions were located and accessed in all lobes of the lung. Radiographic specifics and lobar location of the lesions are presented in Table 2.

Table 1.

Baseline demographics

Variable	Patients (N=23)
Demographics
Age (yr±sd)	68.0±10.9
Gender
Female	52.2%
Male	47.8%
Physical Exam
Weight (kg)	78.29±19.87
Height (cm)	171.2±9.4
Body Mass Index (kg/m2)	26.7±6.2
Medical History Smoking status
Never	13.0%
Current	26.1%
Previous	56.5%
Not Disclosed	4.3%
Pack years (±sd)	30.6±25.4

Table 2.

Lesion Characteristics

Variable	Patients (N=23)
Location of lesion
Right upper lobe	45.8%
Right middle lobe	8.3%
Right lower lobe	8.3%
Left upper lobe	16.7%
Left lower lobe	20.8%
Density*
Solid	95.8%
Ground glass component	4.2%
Diameter of lesion (cm ± sd)	3.6±1.3
Positive bronchus sign	95.8%

Primary End-point

The RT-EBUS-TBNA device was successfully inserted into the bronchoscope and advanced to the target lesion in all lobes of the lung. All peripheral lung lesions were able to be visualized using RT-EBUS-TBNA (Figure 3). Adequate cytologic samples were obtained on 100% of these lesions under RT-EBUS-TBNA with the needle visualized within the lesion in all cases. The average time of the RT-EBUS-TBNA procedure defined as sheath insertion into the bronchoscope and subsequent removal was 28.6 minutes (range, 6.0–56.0 minutes) and a total procedure time including linear EBUS, RT-EBUS-TBNA and CR-EBUS of 62.1 minutes (range, 41.0–114.0 minutes)

Figure 3.

Panel A demonstrates the fluoroscopic view of the RT-EBUS-TBNA adjacent to a pulmonary lesion. Panel B demonstrates the RT-EBUS view with the TBNA needle visualized entering the lesion.

Diagnostic yield results are presented in Table 3. RT-EBUS-TBNA was diagnostic in 16/23 cases (70%). 12/16 cases were malignant and the remaining four were ultimately confirmed to be benign after initial cytopathologic findings of inflammation and subsequent complete resolution on follow up CT imaging. Overall, 20/23 cases were performed with both RT-EBUS-TBNA and CR-EBUS. All seven non-diagnostic RT-EBUS-TBNA cases underwent additional standard of care procedures with CR-EBUS (Olympus UM-S20–17S, Olympus America) of which 2/7 (29%) were diagnostic of neoplasm (squamous cell carcinoma and typical carcinoid). The remaining 13 CR-EBUS cases were performed for additional confirmatory tissue immediately after a diagnostic RT-EBUS case as per protocol.

Table 3.

Diagnostic Yield of RT-EBUS-TBNA and CR-EBUS

Case	RT-EBUS-TBNA	CR-EBUS
1	Adenocarcinoma	Adenocarcinoma
2	Adenocarcinoma	Adenocarcinoma
3	Adenocarcinoma	Adenocarcinoma
4	Adenocarcinoma	Not performed
5	Adenocarcinoma	Adenocarcinoma
6	Adenocarcinoma	Adenocarcinoma
7	Squamous Cell Carcinoma	Squamous Cell Carcinoma
8	Nonsmall Cell Carcinoma	Nonsmall Cell Carcinoma
9	Nonsmall Cell Carcinoma	Nonsmall Cell Carcinoma
10	Nonsmall Cell Carcinoma	Not performed
11	Epitheliod Neoplasm	Epitheliod Neoplasm
12	Melanoma	Not performed
13	Acute Inflammation*	Acute Inflammation*
14	Acute Inflammation*	Acute Inflammation*
15	Acute Inflammation*	Acute Inflammation*
16	Acute Inflammation*	Acute Inflammation*
17	Nondiagnostic	Nondiagnostic
18	Nondiagnostic	Nondiagnostic
19	Nondiagnostic	Nondiagnostic
20	Nondiagnostic	Nondiagnostic
21	Nondiagnostic	Nondiagnostic
22	Nondiagnostic	Typical Carcinoid
23	Nondiagnostic	Squamous Cell Carcinoma

^*Acute Inflammation on cytology with CT follow up confirming nodule resolution

The average number of samples taken with RT-EBUS-TBNA was 3.7 (range 2.0, 7.0). Additional standard of care techniques used to collect samples with CR-EBUS included forceps (61.5%), brush (38.5%), lavage (23.1%) and needle biopsy (46.2%). In patients undergoing additional sampling techniques with CR-EBUS there were an average of four different techniques used, and the procedure lasted an additional 10.7 minutes (range, 1.0–32.0 minutes).

Secondary End-points

There were ten adverse events documented in seven individual procedures of which nine were considered minor. All seven procedures with adverse events received RT-EBUS-TBNA and five of the seven also received CR-EBUS. None of the adverse events were related to anesthesia or directly attributed to the study device or CR-EBUS. There were five episodes of mild hemoptysis, two cases of mild sore throat, one participant complained of a mild cough. All minor events were self-limiting and the average time to onset of an adverse event was one day (range, 0–6.0 days). The single participant with a major adverse event initially presented for the procedure with evidence of a post obstructive process and a lavage during the index procedure grew pseudomonas. This patient developed a pseudomonas pneumonia two days after the procedure. All of these AEs were mild, except for the pseudomonas pneumonia. All participants had resolution of their symptoms, and there were no deaths associated with this procedure. There were no pneumothoraces or moderate to severe bleeding encountered.

DISCUSSION

The diagnosis of peripheral pulmonary lesions can be achieved by invasive maneuvers such as surgical biopsy and CT-guided transthoracic needle biopsies, but with significant morbidity in comparison to minimally invasive techniques^13–16. Although conventional radial EBUS allows for real time imaging of the nodule, the currently available platforms do not allow for nodule sampling under direct ultrasound visualization. As a result, diagnostic yields have remained stagnant given the technologic barriers allowing for guided access of eccentric lesions and the lack of ability CR-EBUS to provide anatomic laterality. This device design has offered a novel opportunity to overcome these limitations by its ability to provide needle repositioning under real time ultrasound guidance. In addition, although transbronchial needle sampling has been previously shown to improve diagnostic yield and reduce complications in peripheral bronchoscopy, it remains underutilized given these same issues and a guided sampling approach may improve useability⁸.

In this study, we describe the first real-time visualization and sampling technique for the biopsy of peripheral pulmonary lesions in humans using a novel device. Prior attempts at real time peripheral sampling were limited by the inability to obtain angular separation between the ultrasound and needle.¹⁷ This problem was overcome by designing the sheath to allow separation of the radial probe and needle at approximately 10 degrees thereby permitting direct visualization on the needle under ultrasound. The results of our pilot study using a hybrid radial endobronchial ultrasound system with a needle adjacent to the catheter allowing real-time peripheral pulmonary lesion biopsies demonstrate that this approach is a feasible technique with an acceptable safety profile in this early prototype.

The device was able to be successfully advanced into all lobes of the lung, and the needle was visualized during tissue acquisition in all cases. In addition, because the needle is clearly visualized in a specific position in relation to the target lesion, anatomic localization of a lesion and subsequent accurate repositioning is feasible for the first time. In CR-EBUS, the 360 degree ultrasound image does not provide anatomic landmarks to determine anterior/posterior or medial/lateral positioning. This limits the diagnostic capabilities of this tool, especially in lesions eccentric to the airway. By advancing the needle during RT-EBUS-TBNA the location of the nodule in relation to the needle is clearly defined affording the bronchoscopist the ability to easily retract the needle and reposition the catheter.

There are limitations to this study. This pilot study was observational, with the primary endpoint to identify the feasibility of a prototype RT-EBUS-TBNA. This study was not powered for diagnostic yield endpoints and reporting of yield in this study is to be considered purely observational. Although all lesions were able to be visualized and sampled using the RT-EBUS system, 12 participants required CR-EBUS due to inadequate sampling. The CR-EBUS technique utilized additional tools for sampling and was able to obtain adequate samples from 92.3% of those attempted, however, these participants had nearly double the number of passes to obtain tissue (6.3 vs 3.7). The adequacy estimation by ROSE methodology during RT-EBUS-TBNA was likely underestimated in our cohort since the number of passes was limited to allow for appropriate time for additional tissue sampling to be obtained by standard techniques if needed. Given that the diagnostic yield is a critical component of this device’s capabilities, future generations of this product will warrant further investigation with larger prospective trials investigating diagnostic yield likely concentrating on eccentric lesions which continue to prove challenging given the anatomic challenges discussed earlier. Although all lobes were accessed, the prototype catheter was somewhat stiff which made positional advancements and precise maneuvering challenging and the ultrasound signal of the needle was of variable quality at times. Although the initial results of this pilot study are promising, given that the RT-EBUS-TBNA device is disposable, cost will be an important factor that needs to be further defined before production and adoption is planned. Future generations of this product should focus on catheter maneuverability, needle visualization and human factors to optimize the learning curve and diagnostic yield.

Conclusions

In this multi-centered prospective pilot study, we demonstrate that the use of RT-EBUS-TBNA is feasible with a side effect profile similar to standard CR-EBUS. The RT-EBUS-TBNA device was able to directly visualize pulmonary lesions and sample them under continuous ultrasound visualization for the first time in humans. Future generations of this device may improve usability. Once achieved, adequately powered studies are needed to further validate these findings as well as to determine the true diagnostic yield of this novel technique.

Abbreviations:

CT-TTNA

CT-guided Transthoracic Needle Aspiration

CR-EBUS

Conventional Radial Endobronchial Ultrasound

ENB

Electromagnetic Navigational Bronchoscopy

RT-EBUS-TBNA

Real-Time Endobronchial Ultrasound with Transbronchial Needle Aspiration

ROSE

Rapid On-Site Evaluation

ITT

Intention-To-Treat