Introduction

Personalized medicine for lung cancer using molecular-targeted drugs and immune checkpoint inhibitors have become widespread. Still, it is necessary to collect sufficient specimens to determine treatment strategies for each case1,2. Furthermore, new genetic mutations and their corresponding molecular target drugs are expected in the future. Individualized treatment based on programmed death-ligand 1 (PD-L1) expression rate and pathological subtypes are also expected, thereby increasing specimens required for each test3.

Generally, conventional bronchoscopic specimens are small, and it is difficult to prove malignant findings including genetic diagnosis in some cases. With the widespread use of transbronchial lung cryobiopsy (TBLC), the collection of specimens which are several times larger in area are possible4,5. TBLC has been used mostly for interstitial pneumonia, but its usefulness in malignant disease is not yet fully known.

Pathologically, specimens of forceps biopsy are significantly smaller than those collected by TBLC. In addition to the smaller volume of specimens, high quality specimens cannot always be obtained due to crush artifacts by forceps biopsy. In a previous report, the evaluable site and the amount of DNA and RNA of each specimen was significantly larger for frozen specimens, which may be superior to forceps biopsy in the diagnosis of lung cancer histology6,7. However, the diagnostic factors favouring TBLC over forceps biopsy remain unclear except for specimen size. Previous studies have reported that TBLC is a beneficial procedure in cases with adjacent to findings on endobronchial ultrasonography (EBUS)8, and in peripheral lesions with a longest diameter of 22 mm or less on CT findings9, but no consensus has been reached on which cases should be preferred for TBLC.

Additionally, non-small cell lung cancer (NSCLC) may have driver mutations, including EGFR, ALK, ROS1, RET and BRAF10,11 . Furthermore, next-generation sequencing (NGS), such as the Oncomine™ Dx Target Test® (ODxTT), to measure multiple gene mutations simultaneously have been used as a companion diagnostic test, but the test failure rate of ODxTT was reported to be up to 28%12. However, two factors can increase the success rate of ODxTT: (1) volume of specimens and (2) high tumor nuclei content. Therefore, TBLC might be useful for NGS13. For these reasons, TBLC is expected to become more widespread in the future. Furthermore, only few studies have prospectively examined the side effects of TBLC. We compared the diagnostic accuracy between TBLC and forceps biopsy, and examined the incidence of adverse events.

Methods

This prospective, cohort study was performed in single center from January 2020 to September 2021. A total of 217 patients were referred to our hospital to perform bronchoscopy with peripheral lung lesion which were suspected of lung cancer. Of these patients, with no contraindications, underwent a TBLC. The contraindications for TBLC were defined as follows: severe hypoxemia (PaO2 (partial pressure of oxygen) < 60 mmHg) or hypercapnia (PaCO2 (partial pressure of carbon dioxide ) > 50 mmHg), bleeding disorders, large vessels near the tumor on the CT, technically difficult to introduce cryoprobe, excessive bleeding after forceps biopsy, which was needed for extra intervention to stop bleeding.

The trial was approved by the medical ethics committee, clinical research section at St. Marianna University (Case No: 4755) and was registered within UMIN Clinical Trials Registry (UMIN000039618, first posted 27/02/2020). All methods were performed in accordance with the declaration of Helsinki.

The bronchoscopes used in this study were: (1) BF-1T260, a standard bronchoscope with high procedural versatility and image quality, (2) BF-P260F, which is thinner and able to reach peripheral sites, and (3) BF-6C260, a bronchoscope with high image quality. (Olympus Marketing Corp., Tokyo, Japan).

In all patients, tracheal intubation (SASSET tracheal tube 8 mm, ICU Medical, CA, USA) was performed using a bronchoscope with a connector (AQUA + TS, Teleflex Medical Sdn Bhd, Perak, Malaysia), to administer oxygen. Radial EBUS with guide-sheath (K-201, or K-203 guide-sheath kit, Olympus Marketing Corp., Tokyo, Japan) was used for all cases except when the cryoprobe could not be inserted through a guide sheath or when the bronchoscope could reach the lesion directly.

Bronchoscopy was performed under laryngeal anaesthesia with 4% xylocaine, followed by sedation with midazolam and analgesia with fentanyl. After confirming the target lesion using EBUS image and fluoroscopy, we underwent an average of 3 passes for brushing cytology and forceps biopsy in any order, followed by TBLC using a cryoprobe with a diameter of 1.9 mm in most situations (Erbe CRYO2, Erbe Elektromdizin GmbH, Tübingen, Germany). TBLC was usually performed once, and additional biopsies were performed if the specimen volume was insufficient. Prior to the TBLC, a Fogarty balloon catheter (Forgarty catheter, Edwards Lifesciences Corp., Ca, USA) was placed in the access bronchus, and balloon dilation was performed immediately after pulling out the cryoprobe.

After examinations, chest X-ray was taken to confirm complications, such as pneumothorax or bleeding. Severe adverse events were defined as requiring additional treatments or procedures (e.g., tracheal intubation, ventilation, surgical treatment, or embolization)6.

Pathological diagnosis

The collected specimens were cytopathologically analysed by a cytologist and at least two pathologists. The area of the histological specimen by TBLC was measured at maximum split plane by cellSens Standard software (cellSens Standard v.1.7, EVIENT, Tokyo, Japan). Tumor nuclei content was determined graphically by a pathologist, and was evaluated three times, three days apart, with outliers excluded.

In the case of NSCLC, NGS analysis was performed using ODxTT. The success of ODxTT was defined as a positive or negative result for four genetic mutation tests: EGFR, ALK, ROS1 fusion gene, and BRAF, which were performed as a companion diagnosis at the time of this study. Unsuccessful results were defined as invalid or no call. Molecular testing of the single-plex test was performed in a part of cases diagnosed as lung adenocarcinoma by pathological evaluation. The Roche cobas® EGFR mutation test v2 (Roche Molecular Systems, Pleasanton, CA, USA), was used for testing EGFR mutations. Unsuccessful results were defined as invalid or failed. Vysis ALK Break Apart FISH Probe Kit (Abbott Molecular, Des Plaines, IL, USA) was used for testing ALK mutations.

Cytological evaluation was based on Papanicolaou's classification, with Class III or higher judged to be diagnosable. Pathologists evaluated the number of tumor cells and tumor nuclei content. The tumor nuclei content was calculated as the percentage of tumor cell nuclei among the nucleated cells in the tissue section.

Statistical analysis was performed using EZR on R commander ver. 1.56. A comparison of forceps biopsy and TBLC for each condition was performed using the McNemar test or the Pearson’s Chi-squared test.

Ethics approval and consent to participate

The clinical protocol was approved by the institutional review board. All patients provided written informed consent before study entry.

Results

Patients and diagnosis

TBLC were performed in 101 patients. One person could not give consent and was excluded. A total of 100 patients participated in our study (Fig. 1). The median age was 74 years (range; 41–91), and 62 (62%) were male (Table 1). Table 2 shows the diagnostic results of all patients. Bronchoscopy confirmed malignancies in 94 cases (94%), with adenocarcinoma accounting for 59 cases (59%). Forceps biopsy was not performed in two cases due to time restraints.

Figure 1
figure 1

Flow diagram displaying patient selection.

Table 1 Patient and lesion characteristics.
Table 2 Morphological diagnosis from TBLC.

Diagnostic yield

The mean tumor diameter using computed tomography (CT) images was 37 mm (range 11–127 mm), and the average number of biopsies was 2.4 for forceps biopsy and 1.1 for TBLC, activated for 5.0 s (range; 3–7 s) of mean freezing time. The mean sample size of TBLC was 9.77 mm2 (range 0.86–44.31 mm2) (Table 3).

Table 3 Pathological sample characteristics from TBLC.

TBLC diagnosed 86 cases (86%), forceps biopsy diagnosed 81 of 98 cases (82.7%), and cytology diagnosed 82 cases (82%) (Table 4). The number of cases diagnosed by a single procedure alone was 1, 2, and 5 cases for TBLC, forceps, and brushing, respectively. The diagnostic yield increased to 94% when all procedures were combined. Six cases were unable to be diagnosed by all techniques (6.0%). There was no significant difference in the diagnostic yield between TBLC and forceps biopsy (P = 0.206).

Table 4 Diagnostic yield for each condition.

The location of biopsies and the diagnostic yield for each procedure are shown in Table 4. In the upper lobe, the diagnostic yield for brushing cytology and TBLC was 80.0% (44/55, each), while forceps biopsy was 84.9% (45/53). There was no significant difference in the diagnostic yield between forceps biopsy and TBLC. In the middle/lingular lobe, the diagnostic yield for brushing cytology and TBLC was 87.5% (7/8 each), while forceps biopsy was 100% (8/8). In the lower lobe, the diagnostic yield of brushing cytology was 83.8% (31/37), forceps biopsy was 75.7% (28/37), and TBLC was 94.5% (35/37). The diagnostic yield for TBLC was significantly higher than that of forceps biopsy in the lower lobe (P = 0.0082).

Table 4 shows echogenic findings by radial EBUS and the diagnostic yield of each technique. When echo findings were within the lesion, the diagnostic yield for brushing cytology and forceps biopsy were both 83.9% (73/87), while TBLC was 87.4% (76/87). When echo findings were adjacent to the lesion, the diagnostic yield for brushing cytology was 81.8% (9/11), while forceps biopsy was 70.0% (7/10). The diagnostic yield for TBLC was 81.8% (9/11). When echo findings visualized mixed blizzard signs in two cases, the diagnostic yield of brushing cytology was 0% (0/2), forceps biopsy was 100% (1/1), and TBLC was 50.0% (1/2).

Table 4 also shows the diagnostic yield for each procedure and location in cases where GS was used. The total diagnostic yield for brushing cytology with GS was 82.5% (66/80). For the brushing cytology subgroup, the diagnostic yield for the upper lobe, middle/lingular lobe, and lower lobe was 78.6% (33/42), 85.7% (6/7), and 87.1% (27/31), respectively. The total diagnostic yield for forceps biopsy with GS was 84.6% (66/78). For the forceps biopsy subgroup, the diagnostic yield of the upper lobe, middle/lingular lobe, and lower lobe was 87.5% (35/40), 100% (7/7), and 77.4% (24/31), respectively. The total diagnostic yield for TBLC with GS was 88.8% (71/80). For the TBLC subgroup, the diagnostic yield of the upper lobe, middle/lingular lobe, and lower lobe was 83.3% (35/42), 85.7% (6/7), and 96.8% (30/31), respectively.

The diagnostic yield of TBLC without GS was 75% (15/20). There was no significant difference in the diagnostic yield for TBLC with or without GS (P = 0.145).

Adverse events

There were two cases of severe bleeding during TBLC. In one case, a Fogarty balloon catheter was effective but active bleeding continued after deflation of the balloon. An endobronchial Watanabe spigot was placed to stop the persistent bleeding. In the second case, active bleeding with a volume of about 500 ml was suctioned immediately after TBLC due to the balloon catheter displacement. This patient suffered from hypoxemia and hypotension. After haemostasis by wedging the bronchoscope to the access bronchus and changing to the side lying position, the patient recovered from hypoxia and was transferred to the intensive care unit and extubated the following day.

Next-generation sequencing (NGS) and single-plex molecular test

The success rate of NGS for TBLC samples was 100% (26/26). Of the 26 patients who underwent NGS, 18 cases had a tumor cell count of 500 or more, and 8 cases had a cell count of less than 500. The median tumor nuclei content in all TBLC cases was 50% (10–90%). Tumor nuclei content was less than 30% and 20% in 6 (6%) and 3 (3%) cases, respectively. The single-plex test was used in 32 cases of cobas® and 3 cases of FISH, all of which were successfully performed.

Discussion

We revealed that TBLC for peripheral lesions had a diagnostic yield as high as both forceps biopsy and brushing cytology. It should be noted that when all techniques were combined, a diagnostic yield of 94.0% was observed. Whereas the diagnostic yield of each technique alone was around 85% in our study, with no significant difference observed between techniques. A previous study reported that combining all techniques resulted in a high diagnostic yield of 89.9%, but the diagnostic yield of TBLC alone was rather inferior to that of conventional biopsy (74.3% vs. 81.3%)8. Therefore, each technique has its advantages and disadvantages, and it is presumed that combining the procedures complements the diagnostic yield. For this reason, TBLC should be performed in combination with conventional biopsy.

We investigated the characteristics of each lesion, and the efficacy of TBLC was compared with forceps biopsy. Although previous reports have shown that the diagnostic yield of TBLC was higher in patients with EBUS findings adjacent to lesions than forceps biopsy7,8, no significant difference was observed in our study. This might be because there were only 11 cases where EBUS findings were adjacent to the lesion. Although the cryoprobe has the ability to freeze specimens circumferentially, further case accumulation is required to evaluate the usefulness of the cryoprobe for EBUS-adjacent lesions.

In subgroup analysis by lesion site, the diagnostic yield for TBLC was significantly higher than that of forceps biopsy at the lower lobes. However, previous reports stated no difference in diagnostic yield depending on the biopsy location7,14. Empirically, lower lobe lesions are subject to fluctuations in biopsy location due to respiratory variability. However, as the cryoprobe instantly adheres to the lesion by freezing, the displacement risk of the lesion site due to respiratory variability may be reduced.

Generally, TBLC at the upper lobe is considered to be difficult to perform, and various techniques have been reported8,15. In our study, the diagnostic yield of TBLC at the upper lobe was 80.0%, and there was no significant difference between the upper and other lobes. Primary lung cancer is known to frequently affect the upper lobe with previous studies showing a diagnostic yield of 77.8%, which was similar to our study16. However, we found one case of GS displacement by the TBLC probe insertion for a lesion at the lung apex (Fig. 2). In this case, forceps biopsy proved malignant cells, while TBLC showed a negative result, which may indicate a limitation for TBLC with GS. However, as a technique for guiding the cryoprobe more smoothly and accurately, an injection of olive oil into the GS can more easily guide the cryoprobe to the lung periphery16. In the middle/lingular lobe, the diagnostic yield was high for all procedures but the number of cases was low, and further study is needed.

Figure 2
figure 2

Guide sheath displaced due to cryoprobe insertion. Figure 1 shows one case of GS displacement by the TBLC probe insertion for a lesion at the lung apex. White arrows indicate GS tips. The tip of the GS is displaced by the insertion of the cryoprobe and is significantly different from the site where the forceps biopsy was performed. GS guide sheath.

In our study, the diagnostic yield of TBLC with and without GS were 88.8% (71/80) and 75.0% (15/20), respectively. It has been reported that TBLC with GS had a significantly higher diagnostic yield than TBLC without GS17. However, a simple comparison was difficult due to the imbalance of cases in our study.

We described two cases of severe bleeding. The incidence rate of adverse events in our study was similar to previous reports18. A previous study reported that radial EBUS imaging can reduce the risk of bleeding by searching for blood vessels at the biopsy site19. Although we performed the same procedure using the EBUS-GS technique, severe bleeding could not be prevented in all cases. Routine preoperative screening based on medical history and blood sample results are not sensitive enough to detect all factors related to bleeding; especially in rare diseases with primary haemostatic abnormalities such as von Willebrand's disease20. Our two patients with severe bleeding also had no abnormalities on preoperative tests. A previous study reported that histological factors, such as renal cell carcinoma, can predict severe bleeding21. Furthermore, CT findings suggested that lesions in the proximal 1/3 of the hilar region are at risk for severe bleeding22. None of these factors were present for our 2 cases.

The NGS success rate was 100%. Generally, the tumor nuclei content required for NGS is reported to be at least 30%23,24. The median tumor nuclei content in our study was 50%, which was sufficient for NGS. Additionally, tumor cell necrosis may have been a factor for unsuccessful NGS results, but EBUS-GS can be used to avoid biopsies in areas with necrotic lesions13,25. The combination of EBUS in our study may have resulted in the high NGS success rate.

There were some limitations to our study. First, this was a single-centre study and therefore, further study with larger populations are needed to confirm these results. Second, performing TBLC as the final procedure may cause bleeding due to forceps biopsy or other previous procedures. The possibility cannot be ruled out that bleeding from brush or forceps biopsy may have affected the diagnostic rate of TBLC in this study. Furthermore, the number of passes for each procedure and the TBLC freezing time were not equal, which may have affected the diagnostic yield. Finally, it is not clear if our results could apply to thinner, disposable cryoprobes that could be commonly used in the future16. In particular, the disposable cryoprobe's flexibility allows the use of a narrow bronchoscope to approach the apex of the lung, which has a strong curvature, and its combined use with conventional biopsy has been reported to increase the diagnostic yield9,26,27. In addition, it may be useful to combine the peripheral cryoprobe approach with new technologies such as F-ENB and robotic assist, and this should be investigated as well28,29.

Conclusion

TBLC for a peripheral lung lesion is usually combined with forceps biopsy and brushing cytology to improve the diagnostic yield. The diagnostic yield for TBLC in this study was higher than that of forceps biopsy, especially for lower lobe lesions. Furthermore, TBLC was able to provide sufficient quality and quantity of specimens for NGS panel tests.