Diagnostic Value of Transbronchial Lung Cryobiopsy Using an Ultrathin Cryoprobe and Guide Sheath for Peripheral Pulmonary Lesions

Abstract

Background:

The addition of cryobiopsy to conventional biopsy methods improves the diagnostic yield of peripheral pulmonary lesions. Moreover, cryobiopsy with a guide sheath (GS) provides additional diagnostic benefits. Semi-real-time biopsy can be repeatedly performed using conventional biopsy devices and a GS, and subsequent cryobiopsy can be easily performed at the same location. Recently, a disposable 1.1 mm-diameter ultrathin cryoprobe has been developed and can be used with a 1.95 mm GS in a 2.0 mm working channel. In this study, we evaluated the diagnostic performance of transbronchial lung cryobiopsy (TBLC) with the 1.1 mm cryoprobe and a GS in patients with peripheral pulmonary lesions.

Methods:

We retrospectively reviewed the medical records of patients who underwent endobronchial ultrasound transbronchial lung biopsy with a guide sheath and TBLC from July 23, 2021 to April 30, 2022 at Chungnam National University Hospital.

Results:

Of a consecutive series of 229 patients, 199 were included. The diagnostic yields of forceps biopsy and cryobiopsy were 65.3% (130/199) and 84.4% (168/199), respectively, and the total diagnostic yield was 91.5% (182/199) (P<0.001 vs. forceps biopsy). Multivariate analysis showed that solid lesion morphology [adjusted odds ratio (OR) 3.659, P=0.002] was associated with a significantly greater diagnostic yield of cryobiopsy, whereas a lesion diameter >20 mm (P=0.026; adjusted OR 3.816) and ‘within’ orientation (P=0.004; adjusted OR 6.174) were associated with a significantly greater overall diagnostic yield.

Conclusion:

TBLC using an ultrathin cryoprobe and GS markedly improves the diagnostic yield.

Radial probe endobronchial ultrasound-guided transbronchial lung biopsy (R-EBUS-TBLB) and image-guided transthoracic needle biopsy (TTNB) are mainly used for pathologic diagnosis of peripheral pulmonary lesions (PPLs).¹ R-EBUS-TBLB is associated with few complications, but the sensitivity is only about 72% given the nature of the bronchoscopy approach; accessibility depends on the tumor-bronchus relationship, termed the ‘bronchus sign’.² In lesions negative for the bronchus sign, the aspiration needle may penetrate the bronchial wall adjacent to the lesion, thereby improving the diagnostic yield (70% for a positive bronchus sign vs. 51% when there is no sign).³

TTNB can access a lesion regardless of the tumor-bronchus relationship; the sensitivity is thus excellent (~90%): also, TTNB is inexpensive. However, given the percutaneous approach, pneumothorax develops in 25.3% of patients and hemorrhage in 18.0%, which are difficult to manage.⁴ In contrast, the incidence of pneumothorax during R-EBUS-TBLB is less than 1%. Although hemorrhage occurs in 16% of patients, it can be controlled using a bronchoscope.⁵

Transbronchial lung cryobiopsy (TBLC) is a form of biopsy that uses a cryoprobe with a metallic tip that rapidly freezes tissues. TBLC can obtain large samples with fewer artifacts, and is superior to forceps biopsy for diagnosing interstitial lung diseases.⁶ For central airway lesions, the diagnostic yield of cryobiopsy is higher than that of forceps biopsy.⁷ It has also been reported that the addition of cryobiopsy to conventional methods improves the diagnostic yield of PPLs, with a low incidence of pneumothorax (1.4%). Although mild-to-moderate bleeding complications occurred in 29.3% of patients in a previous study, these were manageable using a bronchoscope.⁸

Recently, a disposable 1.1 mm ultrathin cryoprobe has been developed; it can be used with a 1.95 mm guide sheath (GS) in a 2.0 mm working channel (WC), making it easy to perform cryobiopsy through endobronchial ultrasound transbronchial lung biopsy with a guide sheath (EBUS-GS-TBLB). The combination of the GS and 1.1 mm cryoprobe, which has excellent flexibility, makes most PPLs accessible; it is possible to obtain adjacent samples for most PPLs.⁹

In previous studies, not all patients underwent TBLC with a GS.^10–12 In this study, we sought to evaluate the additional diagnostic yield provided by TBLC using the 1.1 mm cryoprobe after conventional transbronchial forceps biopsy using a GS.

METHODS

Patients

This study was approved by the Institutional Review Board of Chungnam National University Hospital (approval no. CNUH 2022-06-022). All patients (n=229) who underwent EBUS-GS-TBLB and TBLC from July 23, 2021 to April 30, 2022 at Chungnam National University Hospital were retrospectively reviewed. PPLs were defined as lesions located beyond the subsegmental bronchi, and were thus not directly observed by bronchoscopy. Patients who underwent both EBUS-GS-TBLB and TBLC in the same procedure during the study period were included, Supplemental Digital Content 2, http://links.lww.com/LBR/A284.

In our institution, TBLC was introduced on July 23, 2021. The first 15 cases examined in July and August 2021 were excluded from this analysis because our technique was relatively immature at that time. The exclusion criteria were as follows: procedure performed in July or August 2021, lesion invisible on R-EBUS, presence of a diffuse parenchymal lesion, and performance of EBUS-GS-TBLB or TBLC only.

Procedures

When lung biopsy was requested to explore a suspected lung cancer lesion evident on chest computed tomography (CT), we decided whether to use an endobronchial approach (such as bronchoscopy), convex probe EBUS-guided fine needle aspiration, R-EBUS-guided TBLB, or TTNB depending on the location and size of the lesion, presence or absence of the bronchus sign, and locations of blood vessels. In some cases, the clinician’s decision was changed if the patient rejected the recommended biopsy method or preferred a particular method (eg, R-EBUS-guided TBLB may be rejected because of high cost). In cases where R-EBUS was selected, if there was no chest CT sequence at intervals of ≤ 1 mm, thin-section CT (0.8 mm interval) was performed before the procedure. All bronchoscopy procedures were performed by pulmonologists (D.P. and Y.K.), with patients under moderate-to-deep conscious sedation using midazolam, fentanyl, and/or propofol, without an artificial airway. Lidocaine hydrochloride solution (1% w/v) was administered through the WCs of bronchoscopes [BF-P260F with a 2.0 mm WC (thin bronchoscope), BF-260 with a 2.0 mm WC, or BF-1T260 with a 2.8 mm WC (thick bronchoscope); Olympus, Tokyo, Japan] to induce local anesthesia. Each target lesion was approached using a 1.4 mm R-EBUS probe (UM-S20-17S; Olympus) and a 1.95 mm GS setup (K-201; Olympus,) in the absence of virtual bronchoscopy navigation or fluoroscopic guidance (after reviewing the CT images). When the target lesion was reached, the orientation of the lesion with reference to the bronchus (as revealed by R-EBUS) was recorded. The R-EBUS findings were divided into 3 groups: ‘within’, ‘adjacent to’, and ‘invisible’. Then, the R-EBUS was removed and conventional sampling devices (such as forceps and brushes) were introduced through the GS. Lesion orientation was routinely rechecked with R-EBUS after conventional biopsy. Subsequently, the 1.1 mm cryoprobe (Erbe Elektromedizin GmbH, Tübingen, Germany) was inserted through the GS, frozen for 4 to 8 s, and removed with the bronchoscope to obtain the specimen. Cryobiopsy was repeated at the discretion of the operator. If bleeding developed, cold saline or an epinephrine solution was regionally applied if the bleeding did not improve with bronchoscopic suction for 1 minute. A hemostatic balloon catheter (B5-2C/B7-2C; Olympus) was used to treat severe bleeding that did not respond to conservative treatment.

Variables

The diagnostic yield corresponded to the proportion of diagnosed cases among all procedures. As the follow-up period was both short and variable (range: 6 to 12 mo), a ‘diagnostic case’ was defined as strictly as possible to ensure that the yield was no higher than the actual value. ‘Diagnostic’ referred to a malignancy confirmed by pathology; a benign lesion with specific pathologic findings, such as a granuloma with caseous necrosis or parasite eggs; nonspecific benign pathologic findings with supportive tests, such as chronic inflammation with or without a granuloma plus positive culture or a polymerase chain reaction test positive for Mycobacterium tuberculosis/nontuberculous mycobacteria (NTM); eosinophil infiltration with a positive serological test for parasites; inflammation with Legionella polymerase chain reaction -positivity of bronchoalveolar lavage fluid and a positive urinary antigen test (Legionella pneumonia); organizing pneumonia that decreased in size in response to treatment; inflammation that decreased in response to antibiotics (lung abscess/pneumonia); or a radiologically definite benign lesion that did not increase in size on CT scans taken >6 months apart (benign disease, not otherwise specified). All other cases were considered nondiagnostic (eg, a radiologically definite benign lesion but no CT scans >6 months apart/nonspecific benign pathologic findings not ruling out radiological malignancy).

Bleeding severity was defined as follows: mild, suctioning of blood required for <1 minute; moderate, suctioning for >1 minute, repeated wedging of the bronchoscope, instillation of cold saline, diluted vasoactive substances, or thrombin; severe, selective intubation or balloon/bronchial blocking for <20 minute, or premature interruption of the procedure; life-threatening, persistent selective intubation for >20 minute or new admission to the intensive care unit, packed red blood cell transfusion or the need for bronchial artery embolization or resuscitation.¹³

Statistics

Data are presented as number with percentage or median (range). Categorical variables were compared using the χ² or Fisher exact test, and continuous variables using the Mann-Whitney U test. Factors affecting the diagnostic yield were identified using the χ² or Fisher exact test. For multivariate analysis, logistic regression was performed. All P-values were 2-sided and P-values <0.05 were considered statistically significant. All analyses were conducted using SPSS software (version 22.0; IBM Corp., Armonk, NY).

RESULTS

Among all patients (n=229) who underwent EBUS-GS-TBLB and TBLC from July 23, 2021 to April 30, 2022 at Chungnam National University Hospital, 30 were excluded from this study (Fig. 1). Therefore, a total of 199 underwent EBUS-GS-TBLB followed by cryobiopsy using the 1.1 mm cryoprobe through the GS. In terms of conventional methods, forceps biopsy and brushing were performed in 199 (100%) and 76 (38.2%) cases, respectively. No patient underwent brushing only. On average, forceps biopsy was performed 6.54 times (median =6; range: 1 to 13). Cryobiopsy was repeatedly performed at the operator’s discretion (median =2; range: 1 to 8).

FIGURE 1.

Flow diagram showing patient enrollment. EBUS-GS-TBLB indicates endobronchial ultrasound transbronchial lung biopsy with guide sheath; TBLC, transbronchial lung cryobiopsy.

Patient baseline characteristics are summarized in Table 1. The mean age was 70 years, 58.8% of patients were men and 55.8% were ever-smokers. The lesion diameter exceeded 20 mm in 66.3% of cases (mean diameter =26.7 ± 12.4 mm). Of all lesions, 40.7%, 16.6%, and 42.7% were in the right upper lobe (RUL)/left upper segment (LUS), right middle lobe/left lingula, and both lower lobes, respectively. Solid lesions were the most common (75.4%) and the bronchus sign (defined as the presence of a bronchus leading directly to a peripheral lung lesion) positive in 94.5% of cases. The diagnoses are shown in Table 2. Of the 199 patients, 158 and 41 had malignant and benign disease, respectively. Adenocarcinoma was the most common tumor (74/199 patients). Of the benign diseases, organizing pneumonia was the most common (13 patients).

TABLE 1.

Baseline Characteristics

Variable	Number (%)
Age	70 (13–90)
Men	117 (58.8)
Ever smoker	111 (55.8)
Lesion size
<10 mm	8 (4.1)
10-20 mm	59 (29.6)
>20 mm	132 (66.3)
Lobe
RUL/LUS	81 (40.7)
RML/LLS	33 (16.6)
RLL/LLL	85 (42.7)
Morphology
Solid	150 (75.4)
Part-solid	40 (20.1)
Pure ground-class	9 (4.5)
Bronchus sign on CT
Positive	188 (94.5)
Negative	11 (5.5)
Outer diameter of scope
4.0 mm	185 (93.0)
4.9 mm	10 (5.0)
5.9 mm	4 (2.0)

Data are presented as median (range) or number (%).

CT indicates computed tomography; LLL, left lower lobe; LLS, left lingular segment; LUS, left upper segment; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe.

TABLE 2.

Diagnoses and Diagnostic Yield.

	Forceps biopsy	Cryobiopsy	Overall Diagnostic yield
Malignant (n=158)
Adenocarcinoma	74/113	92/113	102/113
MIA	0/2	0/2	0/2
Squamous cell carcinoma	18/22	21/22	22/22
NSCLC	3/6	4/6	5/6
Small cell carcinoma	2/2	2/2	2/2
Metastasis	5/8	7/8	7/8
Pleomorphic carcinoma with AD	0/1	0/1	0/1
Sarcomatoid carcinoma	0/1	1/1	1/1
Lymphoma	1/1	1/1	1/1
Malignancy	0/1	0/1	0/1
Typical carcinoid	0/1	1/1	1/1
	103/158 (65.2%)	129/158 (81.6%)	141/158 (89.2%)
Benign (n=41)
Tuberculosis	2/3	3/3	3/3
NTM	1/4	4/4	4/4
Organizing pneumonia	7/13	12/13	13/13
Lung abscess/pneumonia	2/2	2/2	2/2
Legionella pneumonia	0/1	1/1	1/1
Toxocariasis	1/1	1/1	1/1
Paragonimiasis	0/1	1/1	1/1
Benign disease, NOS	14/16	15/16	16/16
	27/41 (65.9)	39/41 (95.1%)	41/41 (100%)

Data are presented as number (%).

AD indicates adenocarcinoma; MIA, minimally invasive adenocarcinoma; NSCLC, non-small cell lung cancer; NOS, not otherwise specified; NTM, nontuberculous mycobacterial lung disease.

Diagnostic Yield

The diagnostic yields of forceps biopsy and cryobiopsy were 65.3% (130/199) and 84.4% (168/199), respectively, and the total diagnostic yield was 91.5% (182/199) (P<0.001 vs. forceps biopsy). For 12 of the 17 non-diagnosed cases, the possibility of lung cancer could not be ruled out, but there was no significant increase in lesion size; all of these patients are undergoing follow-up CT. Of the remaining five patients, 2 were diagnosed with pleomorphic carcinoma and non-small cell lung carcinoma on diagnostic video-assisted thoracic surgery, 1 with EBUS-TBNA, and 1 with CT-TTNB, and one received stereotactic body radiation therapy diagnosed with recurrent adenocarcinoma without additional chemotherapy due to the high possibility of recurrence on positron emission tomography-CT.

When cryobiopsy was performed after forceps biopsy, the improvement in total diagnostic yield was greatest for the ‘adjacent to’ orientation. (29.9% for ‘adjacent to’ vs. 25.2% for ‘within’) (P=0.003). No case was diagnosed by brushing alone; the diagnostic yield of EBUS-GS-TBLB was thus the same as that of forceps biopsy. A representative case diagnosed by repeated cryobiopsy in the ‘adjacent to’ orientation is shown in Figure 2.

FIGURE 2.

A representative case of a 62-year-old woman diagnosed with adenocarcinoma by repeatedly performing cryobiopsy on an ‘adjacent to’ lesion (hematoxylin-eosin stain, ×200). (A) forceps biopsy specimen performed before cryobiopsy showed normal lung parenchyma with noticeable crush artefacts (asterisks). (B) The first cryobiopsy specimen contains normal lung parenchyma and cartilage (arrow) and bronchial epithelium (arrowhead). (C) The second cryobiopsy specimen contains atypical cells suspicious of adenocarcinoma but not diagnostic with normal lung parenchyma (circle). (D) In the third cryobiopsy specimen, adenocarcinoma is revealed (square).

Univariable and Multivariable Analyses

Univariate analyses revealed that RUL/LUS location, solid morphology, more than one cryobiopsy, and benign pathology were significantly associated with an enhanced diagnostic yield of cryobiopsy. Lesion diameter >20 mm, RUL/LUS location, the ‘within’ orientation, and benign pathology were significantly associated with an enhanced overall diagnostic yield (Table 3). Multivariate analysis showed that solid morphology [P=0.002; adjusted odds ratio (OR), 3.659] was significantly associated with an enhanced diagnostic yield of cryobiopsy. Lesion diameter >20 mm (P=0.026; adjusted OR, 3.816) and the ‘within’ orientation (P=0.004; adjusted OR, 6.174) were significantly associated with an enhanced overall diagnostic yield (Fig. 3 and Fig. 4).

TABLE 3.

Univariable Analyses for Diagnostic Yield.

Variable	Forceps biopsy		Cryobiopsy		Overall diagnostic yield
	Diagnostic yield (%)	P	Diagnostic yield (%)	P	Diagnostic yield (%)	P
Age (y)
≤70	70/105 (66.7)	0.675	90/105 (85.7)	0.742	96/105 (91.4)	0.988
>70	60/94 (63.8)		79/94 (84.0)		86/94 (91.5)
Sex
Male	75/117 (64.1)	0.665	103/117 (88.0)	0.143	108/117 (92.3)	0.608
Female	55/82 (67.1)		66/82 (80.5)		74/82 (90.2)
Smoking history
Never	56/88 (63.6)	0.656	71/88 (80.7)	0.136	80/88 (90.9)	0.805
Ever	74/111 (66.7)		98/111 (88.3)		102/111 (91.9)
Diameter of lesion (mm)
≤20	31/61 (50.8)	0.004	50/61 (82.0)	0.438	52/61 (85.2)	0.037
>20	99/138 (71.7)		119/138 (86.2)		130/138 (94.2)
Diameter of lesion (mm)
<10mm	4/8 (50.0)	0.010	6/8 (75.0)	0.439	6/8 (75.0)	0.012
10-20 mm	30/59 (50.8)		48/59 (81.4)		50/59 (84.7)
>20 mm	96/132 (72.7)		115/132 (87.1)		126/132 (95.5)
Diameter of lesion in lower lobes (mm)
<10mm	4/6 (66.7)	0.220	6/6 (100)	0.350	6/6 (100)	0.250
10-20 mm	16/28 (57.1)		23/28 (82.1)		25/28 (89.3)
>20mm	38/51 (74.5)		46/51 (90.2)		49/51 (96.1)
Location of lesion
RUL and LUS	50/81 (61.7)	0.377	63/81 (77.8)	0.02	69/81 (85.2)	0.009
Others	80/118 (67.8)		106/118 (89.8)		113/118 (95.8)
Morphology
Others	31/49 (63.3)	0.727	34/49 (69.4)	<0.001	42/49 (85.7)	0.098
Solid	99/150 (66.0)		135/150 (90.0)		140/150 (93.3)
CT bronchus sign
Positive	124/188 (66.0)	0.440	161/188 (85.6)	0.219	174/188 (92.6)	0.056
Negative	6/11 (54.5)		8/11 (72.7)		8/11 (72.7)
EBUS orientation
Within	98/139 (70.5)	0.017	112/139 (87.8)	0.199	133/139 (95.7)	0.002
Adjacent	30/57 (52.6)		46/57 (80.7)		47/57 (82.5)
Number of cryobiopsies
1	NA		53/68 (77.9)	0.047	62/68 (91.2)	0.919
>1			116/131 (88.5)		120/131 (91.6)
Pathology
Malignancy	102/158 (64.6)	0.654	130/158 (82.3)	0.049	141/158 (89.2)	0.026
Benign	28/41 (65.3)		39/41 (95.1)		41/41 (100)

Data are presented as number (%).

CT indicates computed tomography; EBUS, endobronchial ultrasound; LUS, left upper segment; RUL, right upper lobe.

FIGURE 3.

The forest plot shows adjusted ORs of each variable in the multivariable analysis of factors affecting diagnostic yield in cryobiopsy. LUS indicates left upper segment; OR, odds ratio; RUL, right upper lobe.

FIGURE 4.

The forest plot shows adjusted ORs of each variable in the multivariable analysis of factors affecting diagnostic yield in combination. LUS indicates left upper segment; OR, odds ratio; RUL, right upper lobe.

Complications

The complications are summarized in Table 4. The most common complication was bleeding, which occurred in 112 (56.3%) patients, but we encountered no cases of severe/life-threatening bleeding. Pneumothorax developed in 11 patients (5.5%) and 2 required chest tube placement. One patient required a systemic steroid because of chronic obstructive pulmonary disease exacerbation. During the procedures, hypoxemia (oxygen saturation <90%) developed in 42 patients (21.1%), all of whom recovered spontaneously.

TABLE 4.

Complications

Adverse event	Number (%)
Bleeding
Mild	80 (40.2)
Moderate	32 (15.0)
Severe/life-threatening	0
Hypoxemia	42 (21.1)
Pneumothorax
Observation	9 (4.5)
Chest tube replacement	2 (1.0)
COPD exacerbation	1 (0.5)

Data are presented as number (%).

COPD indicates chronic obstructive pulmonary disease.

DISCUSSION

To the best of our knowledge, this is the first study to perform cryobiopsy using an ultrathin cryoprobe (1.1 mm) and GS in all patients. The diagnostic yields of EBUS-GS-TBLB (forceps biopsy) and cryobiopsy were 65.3% (130/199) and 84.4% (168/199), respectively. However, the overall diagnostic yield increased to 91.5% when cryobiopsy was also performed.

Recent studies have shown that the addition of cryobiopsy to conventional PPL sampling methods improves the diagnostic yield to about 90% to 94%, which is mainly attributable to the increased accessibility of ‘adjacent to’ lesions.^10–12 Cryobiopsy yields spherical specimens that surround the metal tip; the amount of tissue is about 26-fold that of a forceps biopsy specimen (standard oval forceps, FB-231D; Olympus) when using a 1.9 mm cryoprobe (CRYO2; Erbe Elektromedizin GmbH).¹⁴ Cryobiopsy destroys benign tissue; it is thus possible to remove normal structures (such as bronchial walls) between the R-EBUS probe and target lesion in the ‘adjacent to’ orientation (Fig. 2).^15,16 When cryobiopsy was performed after forceps biopsy, we found that the improvement in total diagnostic yield was greatest in the ‘adjacent to’ orientation (29.9% for ‘adjacent to’ vs. 25.2% for ‘within’) (P=0.003).

Cryobiopsy provided an additional diagnostic gain when a GS was used. Semi-real-time biopsy can be repeatedly performed using forceps, aspiration needles, and brushes through the GS, followed by cryobiopsy at the same location. Nasu et al¹⁶ reported that GS use increased the diagnostic yield; in all cases that were positive on forceps biopsy but negative on cryobiopsy, a GS was not used. As the 1.1 mm ultrathin cryoprobe was not available at that time, cryobiopsy was performed using a 2.55 mm GS (K-203/204; Olympus). When using a non-ultrathin cryoprobe, it is difficult to direct the cryoprobe toward the targeted peripheral lesion; special techniques including the ‘crawling up’ method and cryoprobe bending are needed.^10,17 As we used the 1.1 mm cryoprobe and a GS to evaluate all patients, we cannot comment on whether this affected the diagnostic yield, but the superior accessibility afforded by an ultrathin cryoprobe may have contributed to the high yield observed.

Multivariate analysis showed that a solid morphology (P=0.002; adjusted OR, 3.659) was associated with a significantly greater cryobiopsy diagnostic yield. Our cohort included more patients with benign diseases than previous studies [20.6% (41/199) vs. 8.6% (Nakai and colleagues), 10.4% (Matsumoto and colleagues) and 4.0% (Nasu and colleagues)], and all benign lesions had a solid morphology.^10,16 Therefore, the excellent diagnostic yield of cryobiopsy when used to evaluate benign diseases may reflect the enhancement provided by such morphology (65.9% on forceps biopsy vs. 95.1% on cryobiopsy). With the exception of some lymphomas, carcinomas can be diagnosed when only a few malignant cells are present, whereas benign diseases are often diagnosed only on the basis of typical findings (granulomas or eggs). The large sample size of cryobiopsy aids recognition of such findings. Three tuberculosis cases (3/4), 5 organizing pneumonia cases (5/13), and 1 paragonimiasis case (1/1) were diagnosed by cryobiopsy only.

Multivariate analysis revealed that a lesion diameter >20 mm (P=0.026; adjusted OR, 3.816) and the ‘within’ orientation (P=0.004; adjusted OR, 6.174) were associated with a significantly greater overall diagnostic yield. A diameter larger than 20 mm and within orientation were known as favorable prognostic factors for diagnostic yield of forceps biopsy.^3,18 Also in this study, a lesion diameter >20 mm and within orientation improved the diagnostic yield by 20.9% (P=0.004) and 17.9% (P=0.017), respectively, compared with lesions ˂20 mm in a diameter and ‘adjacent to’ orientation. The results of multivariate analysis may have been affected by the great improvement of the diagnostic yield in forceps biopsy.

Unlike previous studies, we found that the bronchus sign was not associated with a greater diagnostic yield. In our study, the frequency of the bronchus sign (94.5%) was higher than in other studies [74.8% (104/139, Nakai and colleagues.), 74.7% (192/257, Matsumoto and colleagues), 92.5% (49/53, Nasu and colleagues)]. The bronchus sign is somewhat subjective, and there is no standard minimum size for a bronchus. Therefore, the frequency probably increases as the CT resolution improves or more rigorous reviews are performed. It is possible that some challenging peripheral lesions with ambiguous, thin bronchus connections were included in this study. Minezawa and colleagues classified a prominent bronchus sign as type A, an invisible sign as type C, and all other signs as type B. The diagnostic yield varied according to the prominence of the sign, and was significantly higher in lesions with the type A sign than the ambiguous type B sign.¹⁹

In this study, we did not use fluoroscopic guidance. The bleeding rate was comparable to those of previous studies, but pneumothorax developed in 11 patients (5.5%) and 2 required chest tube placement.^10,11 As both procedures were performed in the same session, the complications cannot be definitively attributed to either forceps biopsy or TBLC alone. Hong et al¹⁸ reported that pneumothorax occurred in 2% of patients when EBUS-GS-TBLB was performed without fluoroscopic guidance. No study has reported the complications of cryobiopsy in the absence of fluoroscopic guidance. However, as the sample size of cryobiopsy is much larger than that of forceps biopsy, the incidence of severe bleeding, pneumothorax, or fistula may increase if performed without fluoroscopy guidance.

This study had some limitations. First, it was retrospective in nature and performed in a single institution. In general, it is difficult to precisely evaluate the diagnostic performance of a specific biopsy method because of selection bias. In addition, in previous retrospective studies, cryobiopsy was performed for only a few patients.^10,11,20 However, in this study, in all except 15 patients who underwent R-EBUS-guided biopsy of peripheral lung lesions, cryobiopsy was performed simultaneously, and only 20 patients underwent TTNB (which is the only non-surgical option other than bronchoscopy in our hospital) (Supplement 1: TTNB cases during the study period, Supplemental Digital Content 1, http://links.lww.com/LBR/A283). Therefore, selection bias was minimized. Second, there is a possibility that the diagnostic yield was underestimated, and the complication rate overestimated, because all cases (except the first 15) were enrolled when fluoroscopy was not performed. The learning curve of TBLC and effects of fluoroscopy are not well known. However, these factors may have contributed to the high incidence of pneumothorax observed in this study. Third, we did not perform needle aspiration, which improves the diagnostic yield by collecting tissues adjacent to lesions. Therefore, we could not evaluate the effect of needle aspiration on subsequent cryobiopsy. Fourth, as cryobiopsy was always performed after forceps biopsy, an effect of the latter on the former cannot be excluded. It is possible that initial forceps biopsy created a ‘track’ for TBLC. Finally, the follow-up period was short, at 6 to 12 months. However, we set very strict diagnostic criteria to prevent overestimation of the diagnostic yield. For example, if CT revealed a possible adenocarcinoma but the result was benign, this was classified as a nondiagnostic case. In cases of adenocarcinoma in situ and minimally invasive adenocarcinoma, tumor size often shows no change even after >1 year. In other studies, even if the biopsy result was benign, such cases may have been considered diagnostic.²¹

In conclusion, transbronchial lung cryobiopsy using an ultrathin cryoprobe and guide sheath markedly improves the diagnostic yield. For peripheral pulmonary lesions, cryobiopsy, which has relatively fewer complications than transthoracic needle biopsy, can be considered preferentially.

Supplementary Material

lbr-31-13-s001.pdf ^{(190.1KB, pdf)}

lbr-31-13-s002.xlsx ^{(11.6KB, xlsx)}