Cryobiopsy-Based Tri-Modality Sampling Using an Ultrathin Bronchoscope for the Diagnosis of Peripheral Lung Lesions: A Prospective Observational Study Open Access

Lung cancer, characterized by high incidence and mortality rates, underscores the critical need for early detection to improve patient outcomes [1]. Systematic screening programs increase the detection of early-stage lung cancers, potentially leading to better prognoses [2‒4]. However, accurate diagnosis of peripheral lung lesions (PLLs) found through these screenings remains a persistent challenge, highlighting the need for improved diagnostic methods.

While transthoracic needle biopsy offers a high diagnostic yield (>90%), it carries a notable risk of complications, including pneumothorax and bleeding [5‒7]. In response to these limitations, a novel approach using a 1.1-mm cryoprobe for transbronchial lung cryobiopsy has shown promising diagnostic yields of up to 90% with acceptable safety profiles [8‒10]. Notably, the recent introduction of the ultrathin bronchoscope (UTB) further improves diagnostic yield by enabling access to more distal peripheral lesions compared to thin bronchoscopes, and the feasibility of performing cryobiopsy with the UTB has also been demonstrated [11, 12].

Despite these advancements, an unmet need remains, with 20–30% of cases still undiagnosed [13‒15]. Currently, no studies have systematically evaluated cryobiopsy-based combination sampling modalities. We hypothesized that combining sampling techniques in the order of the highest reported diagnostic yield, starting with cryobiopsy, followed by forceps biopsy and transbronchial needle aspiration (TBNA), could further improve diagnostic accuracy [8, 16, 17]. In this study, we evaluated the diagnostic yield and safety of this cryobiopsy-based tri-modality sampling approach using UTB.

This prospective observational pilot study was conducted between November 2023 and January 2024 at a university-affiliated tertiary care hospital in Busan, South Korea. The recruitment criteria were as follows: (1) patients aged >18 years with PLL measuring <30 mm in the longest diameter on computed tomography (CT), (2) patients scheduled for transbronchial biopsy to determine the management plan, and (3) patients without a prior diagnosis of PLLs. The exclusion criteria were: (1) patients with visible lesions identifiable through bronchoscopy and (2) patients who declined to participate in the study [18, 19]. The study protocol followed the Standards for Reporting Diagnostic Accuracy Studies guidelines for assessing diagnostic accuracy [20].

Bronchoscopy procedure began with the administration of lidocaine local anesthesia using a spray catheter (PW-5L-1; Olympus, Tokyo, Japan) and conscious sedation with intravenous midazolam, fentanyl, and propofol. Oropharyngeal intubation was performed using an endotracheal tube (inner diameter, 8.0 mm) with a 6.0-mm-diameter therapeutic bronchoscope (TB) (BF-1T290; Olympus, Tokyo, Japan). After inspecting the tracheobronchial tree, the TB was removed and replaced with a 3.0-mm-diameter UTB (BF-MP290F; Olympus, Tokyo, Japan). The UTB was positioned as close to the PLL as possible, guided by CT imaging [21], and virtual bronchoscopic navigation (Lung Point; Broncus Medical, Mountain View, CA) [22]. Subsequently, the 1.4-mm-diameter radial probe endobronchial ultrasound (RP-EBUS) (UM-S20-17; Olympus, Tokyo, Japan) was positioned through the working channel to detect the target lesion. Tri-modality sampling procedures were conducted as follows: the sampling order – starting with cryobiopsy, followed by forceps biopsy, and then TBNA – was predetermined based on our hypothesis that performing the modalities in the order of their historically reported diagnostic yields could maximize cumulative diagnostic accuracy. Cryobiopsy was selected as the initial modality because of its consistently high diagnostic yield in previous studies, followed by forceps biopsy and TBNA [8, 16, 17].

Two transbronchial cryobiopsies used a 1.1-mm-diameter cryoprobe (No. 20402-402; Erbe, Tubingen, Germany): first, the target PLL was frozen for 6 s, then the cryoprobe and the UTB were removed together, and the frozen specimen was thawed in saline solution at room temperature [10].

Five transbronchial forceps biopsies used 1.5-mm-diameter forceps (FB-233D; Olympus, Tokyo, Japan) [23].

Two TBNAs used a 1.5-mm-diameter, 21-gauge PeriView FLEX needle (NA-403D-2021; Olympus, Tokyo, Japan); when the needle punctured the target PLL, then the stylet was removed, and the needle was agitated 20 times in 1-cm strokes while applying vacuum suction with a syringe [24, 25]. For each cryobiopsy and each change in sampling modality, the lesions were precisely relocated using RP-EBUS, and all procedures were conducted under fluoroscopic guidance. If bleeding could not be controlled with a UTB, hemostasis was achieved by switching to a TB. In cases where the risk of bleeding was high, “two-scope technique” was used to prepare for potential bleeding in advance [26, 27]. If the patient exhibits unstable vital signs, uncooperative behavior, uncontrolled bleeding, or coughing, the bronchoscopist might decide to terminate the procedure at any stage of tri-modality sampling. In this study, tri-modality was defined as cases in which cryobiopsy, forceps biopsy, and TBNA were all performed. Bi-modality referred to cases in which cryobiopsy and forceps biopsy were conducted, whereas mono-modality referred to cases in which only cryobiopsy was performed. Chest radiographs were obtained 2 h and 7 days after bronchoscopy to check for pneumothorax or infection. All study participants received oral prophylactic antibiotics such as amoxicillin/clavulanate or respiratory quinolone for 3 days. Antiplatelet agents were withheld for 7 days before the procedure, and anticoagulants were discontinued 1–5 days before the procedure, depending on renal function. All medications were resumed on the day after the procedure. All procedures were conducted on an inpatient basis to allow immediate management of potential complications, including unexpected bleeding. Patients were monitored for at least 2 h after the procedure, and if no bleeding or other adverse events were observed, they were discharged on the same day.

Demographic and clinical data were collected, including age, sex, anatomical location from the hilum of the PLL (peripheral one-third or intermediate one-third), longest diameter of the PLL, morphological classification of the PLL (solid, part-solid, or pure ground-glass nodule), bronchus sign (positive or negative), bronchus level, location of RP-EBUS (concentric, eccentric, adjacent to, or invisible), number of biopsies performed, tissue surface area, total time of bronchoscopy, and bronchoscopic diagnosis. The surface area of the tissue was calculated by multiplying the longest diameter by the vertical diameter of each sample, and the sizes of the largest specimens obtained from each modality were compared.

The final diagnosis was based on pathological reports, microbiological examinations, or follow-up with a CT. Histological examination of bronchoscopic samples was considered nondiagnostic if it revealed the presence of atypical cells, nonspecific inflammation, or suspicious findings without a definitive diagnosis. In such cases, the following procedures were performed: (1) surgical biopsy using thoracoscopy, (2) additional bronchoscopic biopsy, or (3) repeat CT every 3 months. The condition was classified as a benign disease, not otherwise specified, if the size of the PLL did not change over 12 months. Bronchoscopic complications were assessed based on the degree of bleeding, occurrence of pneumothorax, infection, respiratory failure, and other unexpected complications. The degree of bleeding was classified as follows: grade 1, requiring less than 1 min of suctioning or wedging of the bronchoscope; grade 2, requiring >1 min of suctioning, wedging, or instillation of cold saline, diluted vasoactive substances, or thrombin; grade 3, requiring selective intubation using balloon/bronchial blockers for less than 20 min, or premature interruption of the procedure, and grade 4, requiring persistent selective intubation for >20 min, blood transfusion, bronchial artery embolization, or intensive care therapy [28].

The primary outcome was the overall cumulative diagnostic yield of the bronchoscopic biopsy. Diagnostic yield was calculated by dividing the number of bronchoscopy procedures that resulted in a successful pathological diagnosis by the total number of bronchoscopy procedures performed. The secondary outcomes were the diagnostic yield of each modality, additional gain in diagnostic yield from each modality, and incidence of adverse events related to bronchoscopy procedures.

Based on historical study, the expected diagnostic yield for the UTB is 75% [12]. The number of subjects required for this study was calculated using the G*Power software (version 3.1; Heinrich-Heine-Universität Düsseldorf, Düsseldorf, Germany) [29]. In the chi-squared test, to achieve 80% power at a 5% significance level and with an effect size of 30%, the required sample size was 44. Considering a dropout rate of 15%, 51 patients were finally recruited.

Continuous variables are presented as the mean (range or interquartile range) and compared using Student’s t test or the Mann-Whitney U test. Categorical variables are presented as numbers (percentages) and were compared using the chi-square or Fisher’s exact test. To evaluate diagnostic yields, both sensitivity and negative predictive values were calculated. All tests were two-tailed, and p values <0.05 were considered statistically significant. Statistical analyses were performed using the R statistical language (version 4.3.2; R Core Team, 2023).

During the study period, 52 patients with a PLL <30 mm were screened, and 2 patients were excluded. Finally, 50 patients who underwent UTB were selected as study subjects (Fig. 1). All test results were followed up until December 2024. Table 1 lists the baseline characteristics. The median age was 70.0 years, the median longest diameter on CT was 20.3 mm, and the bronchus sign was present in 33 (66%) patients.

Table 2 presents bronchoscopic procedural details. The median procedural time (from the bronchoscope insertion at the vocal folds to its removal) was 26.1 min. The RP-EBUS findings were classified as “concentric,” “eccentric,” and “adjacent to” in 29 (58%), 9 (18%), and 9 (18%) cases, respectively, with 3 (6%) cases classified as invisible. Cryobiopsy was performed in 47 (94%) patients. Among them, 14 patients experienced grade 3 bleeding (eight after the 1st cryobiopsy and six after the 2nd cryobiopsy) that could not be controlled using the UTB. These patients required switching to TB for balloon hemostasis, and the procedures were prematurely terminated owing to safety concerns. In addition, procedures in 3 patients with grade 2 bleeding were terminated early due to severe coughing and poor cooperation during relocation to the target lesion after cryobiopsy. Forceps biopsy was performed in 30 (60%) cases, with 6 cases terminated early due to uncontrolled coughing and poor coordination after forceps biopsy. TBNA was performed on 24 patients (48%). Tri-modality was successfully implemented in 24 (48%) patients. Bi-modality (cryobiopsy and forceps biopsy) was performed in six (12%) cases, and mono-modality (cryobiopsy only) in 17 (34%) cases.

The overall cumulative diagnostic yield was 78% (39/50) (Table 3). The details of the diagnostic yields for each sampling modality are shown in Table 4 and Figure 2a. The diagnostic yields of each modality were as follows: cryobiopsy, 78.7% (37/47) (including the 1st cryobiopsy, 66.0% [31/47] and the 2nd cryobiopsy, 77.1% [27/35]); forceps biopsy, 66.7% (20/30), and TBNA, 54.2% (13/24). The addition of a 2nd cryobiopsy after the 1st cryobiopsy improved the cumulative diagnostic yield by 12%, and forceps biopsy after cryobiopsy increased the yield by 4% (Fig. 2b). TBNA did not result in an additional diagnostic gain. Ultimately, 35 patients were diagnosed with malignant tumors and four with benign disease through bronchoscopic biopsy (Table 3). Among the undiagnosed patients, six had lung cancer following surgical biopsy and one was diagnosed with lung cancer after rebiopsy using UTB. The remaining 4 patients were classified as having benign disease, as no size changes were observed on CT over 1 year. There were no significant differences in malignant diagnostic sensitivity or negative predictive value across the different modalities (Table 4). The addition of a 2nd cryobiopsy after the 1st cryobiopsy improved the malignancy diagnostic sensitivity by 15.4%, whereas the addition of forceps biopsy after cryobiopsy increased the malignancy sensitivity by 5.1% (Fig. 2c).

PLL size significantly influenced the diagnostic yield (62.5% in ≤20 mm vs. 92.3% in >20 mm, p = 0.011). However, no significant differences were found in PLL morphology (80.6% solid vs. 76.9% part-solid, p = 0.781) or RP-EBUS location (81.6% within vs. 88.9% adjacent to, p = 0.600). Nondiagnostic cases from cryobiopsy are listed separately in the online supplementary data (for all online suppl. material, see https://doi.org/10.1159/000546433).

Among the 50 patients, 68% experienced bleeding, predominantly of grade 2 (30%) or grade 3 (28%). Compared with other modalities, cryobiopsy was associated with a higher risk of bleeding (p < 0.001). All grade 3 bleeding cases were controlled using TB and balloon hemostasis. No grade 4 bleeding, pneumothorax, infection, or other life-threatening complications occurred. Among the enrolled patients, 10 received antiplatelet agents, and two were on anticoagulant therapy at enrollment. No thromboembolic complications occurred during this interruption period. Of the 10 patients receiving antiplatelet therapy, 2 experienced grade 3 bleeding, while the remaining 8 experienced grade 1 or 2 bleeding. None of the patients who received anticoagulants experienced grade 3 bleeding. No serious bleeding-related adverse events were observed in any patient.

To our knowledge, this is the first study to investigate the diagnostic yield and safety of a cryobiopsy-based combination sampling modality for PLL using UTB. The technique achieved an overall cumulative diagnostic yield of 78%. Grade 3 bleeding occurred in 28% of the cases, all of which were successfully managed without complications. No cases of pneumothorax, infection, or life-threatening complications were observed. These findings indicate that cryobiopsy-based combination sampling with UTB is both feasible and safe in clinical practice.

In clinical practice, performing multiple sampling modalities is often challenging because of the difficulty in maintaining an optimal sampling path. This challenge arises primarily from bleeding and edematous changes in the bronchial wall after each sampling, which can obscure the pathway to the target lesion and render subsequent biopsies less effective. Therefore, we hypothesized that selecting the most diagnostically efficient method as the initial approach, starting with cryobiopsy, followed by forceps biopsy, and then TBNA, is critical for maximizing the diagnostic yield [8, 16, 17]. The choice of the biopsy sequence can significantly influence both the completion rate of tri-modality sampling and the overall diagnostic yield. Performing TBNA or forceps biopsy before cryobiopsy may reduce the likelihood of procedure discontinuation owing to bleeding or patient intolerance, thereby enabling more patients to complete all three modalities. However, if bleeding or excessive coughing occurs after TBNA or forceps biopsy, the effectiveness of the subsequent cryobiopsy may be compromised by limited visibility or lesion accessibility. These competing considerations underscore the uncertainty surrounding the optimal sequence of biopsy modalities. An ongoing clinical trial (NCT06752395) comparing different biopsy sequences, cryobiopsy following forceps biopsy versus forceps biopsy following cryobiopsy, is expected to provide further insights into this issue. In the present study, adding forceps biopsy to cryobiopsy-based sampling with UTB led to a 4% increase in diagnostic yield, from 74% to 78%. Additionally, forceps biopsy after cryobiopsy increased the sensitivity for malignancy by 5.1%, from 84.6% to 89.7%. However, TBNA did not improve the diagnostic yield or sensitivity for malignancy, as a plateau in the diagnostic yield was reached with cryobiopsy and forceps biopsy. These results were numerically higher than the diagnostic yield and sensitivity for malignancy reported in historical comparisons of transbronchial biopsy with RP-EBUS [30, 31]. These findings underscore the potential advantages of cryobiopsy-based combination sampling using UTB.

Several key factors contributed to the effectiveness of this approach. First, a UTB with a 3.0-mm outer diameter offers enhanced accessibility, bronchial selectivity, and maneuverability in the peripheral bronchi. Compared to a thin bronchoscope with a 4.0-mm outer diameter, the UTB enables access to more peripheral bronchial branches, enhancing positional alignment with the target PLL and improving the diagnostic yield [11, 32]. In this study, lesions were identified in 14 of 17 patients without a bronchus sign on CT, and successful biopsy was achieved with a sensitivity for malignancy of 89.7%. These results highlight ability of UTB to overcome the limitations of traditional bronchoscopic techniques in accessing target PLLs. Second, cryobiopsy offers a larger specimen volume with well-preserved tissue quality for histopathological examination than forceps biopsy [16, 33]. Additionally, cryobiopsy collects tissue by freezing the lateral tissue surrounding the cryoprobe, overcoming the limitations of forceps biopsy, which grasps tissue only in the forward area. In this study, cryobiopsy provided a specimen volume three times larger than that of forceps biopsy and achieved a diagnostic yield of 88.9% adjacent to the lesions, underscoring its ability to overcome the limitations of traditional bronchoscopic biopsy techniques. Finally, the combined sampling modality compensates for the limitations of each biopsy technique. In our study, 2 cases of lung cancer were diagnosed exclusively using forceps biopsy (online suppl. data). In 1 case, a solid lesion located centrally on RP-EBUS suggested a high diagnostic potential (online suppl. Fig. 1). However, owing to the limited flexibility of the cryoprobe, the UTB must bend over 90°, which poses challenges in accessing the lesion optimally. Consequently, cryobiopsy yielded only atypical cells. In contrast, biopsy forceps enabled access to the lesion, resulting in a specimen diagnosed as lung cancer. In the second case, a part-solid lesion was located eccentrically on RP-EBUS (online suppl. Fig. 2). After reevaluation with RP-EBUS before forceps biopsy, it was observed that the positional alignment between the bronchi and target PLL had shifted to a more concentric position. While only atypical cells were obtained from cryobiopsy, lung cancer was confirmed in a subsequent forceps biopsy. This shift in positional alignment may be attributed to the collapse of the bronchial structure after cryobiopsy, potentially creating a new pathway for targeting the PLL. These findings underscore the complementary roles of combination modalities in diagnosis, reinforcing the need for a cryobiopsy-based combination sampling approach using UTB to maximize diagnostic yield.

As a larger amount of tissue is obtained, cryobiopsy carries a higher risk of bleeding [27]. Once bleeding occurs, the use of additional sampling methods may be limited because of the increased risk of rebleeding during subsequent biopsies. To manage this risk, we prepared in advance using the two-scope technique whenever blood vessels around the PLL were identified on CT or RP-EBUS. This approach allowed for better visibility and hemostasis with TB than with UTB, and if bleeding persisted, balloon hemostasis could be readily performed. In this study, grade 3 bleeding occurred in 28% of patients following cryobiopsy, which was higher than the rates reported for conventional transbronchial lung biopsy in previous studies (1.5–11.1%) [16, 34]. However, all grade 3 bleeding events were well controlled using TB with or without balloon hemostasis, and there were no life-threatening cases, suggesting that cryobiopsy is a safe procedure when proper precautions are taken in advance.

Based on the findings of this study, we propose a new biopsy protocol for diagnosing PLLs that prioritizes cryobiopsy-based combination modalities using UTB. This approach is particularly effective for accessing peripheral lesions. However, in cases in which the lesion is centrally located or easily accessible using a thin or conventional bronchoscope, this cryobiopsy-based combination sampling approach can also be applied. In situations with a high risk of bleeding or other safety concerns, bronchoscopists may begin with a forceps biopsy. Although TBNA is not routinely performed when both cryobiopsy and forceps biopsy are conducted, it may be considered in cases where cryobiopsy is not feasible or contraindicated.

This study has some limitations. First, as this was a single-arm, single-center study, the findings may not be generalizable. Second, although the study aimed to investigate the cryobiopsy-based combination sampling modality, 34% of the patients underwent only mono-modality procedures, with bleeding being a significant barrier. Despite the relatively low completion rate of combination sampling (34%), primarily owing to bleeding and poor patient cooperation after cryobiopsy, our study demonstrated a high cumulative diagnostic yield (78%) and sensitivity for malignancy (89.7%). These results suggest that if all modalities can be safely completed, the diagnostic performance of the cryobiopsy-based combination approach can be further improved. Given that bleeding control and patient cooperation are critical factors influencing procedural completion, several strategies may help increase the success rate of multimodal sampling. These include adjusting the biopsy sequence based on the anticipated bleeding risk (e.g., performing forceps biopsy and TBNA before cryobiopsy), preemptively placing bronchial balloons in all cases to manage potential bleeding, and performing procedures under general anesthesia to improve patient cooperation. However, the feasibility and safety of these strategies require further validation in prospective studies. Third, TBNA does not provide any additional diagnostic benefits. When using a cryobiopsy-based sampling approach, combining cryobiopsy with forceps biopsy appears to be sufficient, as the diagnostic benefit seems to plateau with the bi-modality approach. However, previous studies have shown that TBNA can enhance diagnostic yield, particularly for lesions adjacent to RP-EBUS. In this study, cryobiopsy was performed first and likely provided sufficient lateral tissue sampling in these lesions, potentially diminishing the incremental value of TBNA [25]. Additionally, bleeding caused by the preceding modalities may interfere with the TBNA performance. Although TBNA did not demonstrate an added diagnostic benefit in our cohort, we caution against overgeneralization due to the limited sample size. However, further studies are required to evaluate the efficacy of TBNA. Finally, the optimal freezing time for cryobiopsy and the minimum number of procedures required for each modality have not been definitively established. These limitations highlight the need for large-scale multicenter studies. Nevertheless, the feasibility and safety of the cryobiopsy-based combination sampling modality were demonstrated, marking a significant milestone that will likely pave the way for further research.

In conclusion, this study demonstrated the feasibility and safety of a cryobiopsy-based combination sampling approach using UTB for diagnosing PLL, achieving a cumulative diagnostic yield of 78% and a sensitivity for malignancy of 89.7%. This approach leverages the accessibility of UTB, specimen quality and volume benefits of cryobiopsy, and the compensatory advantages of forceps biopsy to overcome the limitations of traditional bronchoscopic techniques. Although bleeding is a common side effect, all cases were successfully managed without severe complications, indicating that the cryobiopsy-based combination modality can be safely performed with appropriate precautions. To refine the clinical practice and improve the diagnostic yield for PLL, further large-scale multicenter randomized controlled trials are necessary to determine the optimal order and number of sampling modalities for this combination approach.

This study was approved by the Institutional Review Board of Pusan National University Hospital (IRB No. 2310-020-132) and registered as a Clinical Trials.Gov (NCT06123312). The study procedures followed the 1975 Declaration of Helsinki and Good Clinical Practice guidelines, and written informed consent was obtained from all study participants.

H.S., J.M., M.K.L., and K.K. have no conflicts of interest to disclose. S.H.K. received speaker fees from Amgen, AstraZeneca, and Yuhan. M.-H.K. received speaker fees from AstraZeneca, Boehringer Ingelheim, and Yuhan, and grants from Yuhan. J.S.E. has received speaker fees from Amgen, AstraZeneca, Boehringer Ingelheim, Boryung, Daiichi Sankyo, Erbe, Lilly, MSD, Mundipharma, Novartis, Olympus, Roche, Takeda, Yuhan, and Merck, and grants from Boehringer Ingelheim, Boryung, and Erbe.

This research was supported by a grant from Erbe. The funder had no role in the study design, data collection, data analysis, and reporting of this study.

J.S.E. is the guarantor of the content of the manuscript, including data and analysis, had full access to all the data in the study, took responsibility for the integrity of the data and the accuracy of the data analysis, including and especially any adverse effects, and contributed to conceptualization, resources, software, and supervision; H.S. contributed to the methodology, data curation, data analysis, and interpretation, and the writing of the manuscript; S.H.K., J.M., M.-H.K., M.K.L., and K.K. contributed substantially to the investigation, study design, and validation.

Trial registration: Clinical Trials.Gov (NCT06123312); URL: www.clinicaltrials.gov

The data that support the findings of this study are not publicly available due to privacy restrictions but are available from the corresponding author upon reasonable request.