Useful on-site macroscopic quantitative evaluation of EBUS-TBNA samples

Abstract

Background

Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is a minimally invasive diagnostic test with a high diagnostic yield. Optimal specimen acquisition is of a key importance; however, rapid on-site cytology examination (ROSE) during EBUS-TBNA has limited access. In an attempt to provide immediate feedback to the operator, the intervention pulmonology team started to classify macroscopically every lymph node sample in terms of abundance. The main objective of this study was to clarify whether this on-site macroscopic evaluation of EBUS sample was related to histology results and therefore any helpful for clinical practice.

Results

From a total of 288 lymph node stations sampled, adequate/diagnostic histological results were present in 11.1% of “insufficient” samples, 40.9% “sufficient” samples, 76.4% “good” samples and 88.1% of “excellent” samples. This was statistically significant with p-value under 0.0001. Also, there was a statistical agreement between samples that were macroscopically “insufficient” or “sufficient” with “inadequate” and “dubious” histological results and also between macroscopically “good” and “excellent” samples and histological results that were adequate/diagnostic. And that statistical agreement was of moderate magnitude according to the “Landis and Koch” classification (k = 0.386).

Conclusions

The positive relationship between the on-site macroscopic evaluation of the EBUS sample as “excellent” and the “adequate/diagnostic” histological result is of high clinical importance since it represents a fast, cheap and easy method that provides immediate feedback to the operator influencing his conduct; by obtaining additional passes to achieve an “excellent” will ensure a higher percentage of samples with a diagnosis, accelerating the patients treatment.

Background

Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is a minimally invasive diagnostic test with a high diagnostic yield. In this procedure, abundance of material and optimal specimen acquisition is crucial for obtaining adequate tissue for diagnosis. In an attempt to increase diagnostic yield [1], there have been studies associating rapid on-site cytology examination (ROSE) during EBUS-TBNA, which might be useful, however with limited access [2, 3]. Based on a similar concept to ROSE of providing immediate feedback to the operator and therefore increasing diagnostic yield and for not having access to it in the institution, the endoscopic team started to classify macroscopically every lymph node sample in terms of abundance. To the authors’ knowledge, to date, there is no quantitative measure of EBUS-TBNA samples. The main objective of this study was to clarify whether this on-site macroscopic evaluation of EBUS sample abundance was related to histology results, namely lymph node representativeness or disease diagnosis and therefore any helpful for clinical practice.

Methods

Retrospective study of patients undergoing EBUS-TBNA in the year of 2018. All procedures were performed with an EBUS bronchoscope (BF-UC180F Olympus) and Olympus ViziShot 22-g needle. The procedures were monitored and performed under general anaesthesia or mild sedation after patients signed a written informed consent. The lymph nodes to be sampled were left up to the discretion of each operator, who made a decision based on image, association with higher cancer staging and/or ease of sampling. At least three needle passes were conducted for each location as recommended [4]. Fragments of tissue were removed gently and put in a formalin solution, and assessment was made on cell block. The fragments were evaluated as “insufficient”, “sufficient”, “good” and “excellent” according to material abundance as exemplified in Table 1.

Table 1 Macroscopic evaluation of EBUS-TBNA samples

Patient demographic variables, procedure variables and histologic reports were analysed. Because there is no “gold standard” for EBUS specimen adequacy and being aware of cellular “contaminants” that are acquired when the needle passes through the tracheal wall en route to the targeted lesion, authors classified histological results as “inadequate”, “dubious” or “adequate/diagnostic” as described in Table 2.

Table 2 Histological classification of samples

Statistical analyses were made by SPSS; comparisons were performed with chi-square test and agreement by Cohen’s K-test. Agreement’s magnitude was classified according to Landis and Koch classification. All tests were two sided, and a p-value < 0.05 was considered to be statistically significant.

Results

One-hundred thirty-four EBUS-TBNA were conducted in the year of 2018; 13 exams were excluded for corresponding solely to biopsies of structures other than lymph nodes or exams in which biopsy was not possible as described in Table 3. A total of 121 procedures were included, corresponding to 117 patients.

Table 3 Excluded EBUS procedures

Of the 117 patients included, 68.37% (n = 80) were male and 31.62% were female (n = 37). The mean age was 64.1 years (SD 11.1 years).

The indications for EBUS-TBNA conducted in 2018 were lung micronodules/nodules and adenopathies in 24 exams (19.84%), lung mass and adenopathies in 26 exams (21.49%), isolated adenopathies in 22 exams (18.18%), cancer staging in 28 exams (23.14%) and suspicion of cancer recurrence in 21 exams (17.35%) (Table 4). Only 7 procedures (5.7%) were made under mild sedation with midazolam and 114 (94,3%) under general anaesthesia. During these procedures, a total of 288 lymph node stations were sampled. The median number of lymph node stations biopsied per procedure was 2.1. The median number of lymph node stations biopsied per procedure under mild sedation with midazolam was 1.5 and 2.18 with general anaesthesia. There was at least 1 station punctured and a maximum of 5 stations punctured in the procedures analysed. The mean number that each lymph node station was punctured was of 3.81 times, and there was no significant difference in the number of punctures of lymph nodes’ samples classified as “insufficient” (3.33 times), “sufficient” (3.63 times), “good” (3.56 times) or “excellent”(3.59 times).

Table 4 Motives for EBUS execution

Location of the lymph nodes biopsied according to the IASLC classification [5] is described in Table 5.

Table 5 Location of the lymph nodes biopsied according to the IASLC classification

The most common lymph node station biopsied was station 7 (34.03%), followed by station 4R (26.02%) and station 11R (18.06%). No statistical difference was found between the station number and histological adequacy, since some stations were rarely punctured.

By cross tabulation, macroscopically “insufficient” samples were mostly inadequate (n = 9/88.9%), and only 1 sample was adequate/diagnostic (1.1%) in the histological classification. Of the 22 macroscopically “sufficient” samples, 3.8% were inadequate, 27.5% were dubious and 40.9% were adequate/diagnostic by histological classification. From the 55 macroscopically “good” samples, 20.0% were inadequate and 76.4% adequate. And finally, from the 178 macroscopically “excellent” samples, 88.1% were adequate/diagnostic, only 6.4% were dubious, and 5.4% were inadequate by histological classification. This was statistically significant with p-value under 0.0001. This is described in Table 6.

Table 6 Macroscopic evaluation versus histologic evaluation cross tabulation

Moreover, there was a statistical concordance between samples that were macroscopically “insufficient” or “sufficient” with “inadequate” and “dubious” histological results. Also, between macroscopically “good” and “excellent” samples and histological results that were adequate/diagnostic. And that statistic concordance (k = 0.386) was of moderate magnitude according to the “Landis and Koch” classification.

Discussion

Macroscopic quantitative evaluation of EBUS-TBNA samples is statistically related to histological results, being the macroscopically “excellent” samples more often (88.2%) representative of lymph node or diagnostic.

From the 202 samples classified as “excellent”, 11 were considered “inadequare”, and 13 were considered “dubious” by final cytologic diagnosis, totalling 11.8%, somewhat higher than other published studies, where samples considered adequate/satisfatory by ROSE [6,7,8,9] resulted in about 5% of inconclusive/indeterminate samples by final cytologic diagnosis. This probably means that the histological material was made by cellular “contaminants”, acquired when the needle passed through the tracheal wall en route to the targeted lesion, extensive necrosis, blood or insufficient lymphoid or neoplastic cells, therefore not meeting the histological criteria for adequate/positive.

Moreover, when samples were classified as “insufficient” or “sufficient” in our study, 88.9% and 59.1%, were considered “inadequare” or “dubious” by cytology respectively. These percentages are very satisfactory since in a study using ROSE [6], 50.6% of samples classified as inadequate remained inadequate on final cytologic diagnosis.

This study had some limitations. It was conducted in a single centre, and the macroscopic evaluation might be considered too subjective. Also, because ROSE is not available in our institution, there was no direct comparison/agreement between ROSE and our macroscopic evaluation of the same sample.

On the other hand, the proposed macroscopic classification in this study had positive statistical results, not too far from ROSE, and does not require any further economic or human resourses, that is, a pathologist readily available at the time of the procedure, which most institutions do not dispose. It is a practical, easy, cheap, fast and on-site classification that provides immediate feedback to the operator, potencially shortening the duration of the procedure and/or ensuring a higher percentage of samples with a diagnosis.

Conclusions

Although it is a subjective way of quantifying samples, authors consider this findings of high clinical importance since this represents a fast, cheap and easy method that evaluates the adequacy and representativeness of the material obtained for final interpretation by cytology in the absence of ROSE. This immediate macroscopic evaluation should influence the procedure, namely, obtaining additional passes trying to achieve an “excellent” sample and thereby potentially lowering nondiagnostic sampling.