Evaluating the efficiency of endobronchial ultrasound for diagnosis of patients with suspected lung cancer

Journal of military pharmaco-medicine n 0 1-2020 188 EVALUATING THE EFFICIENCY OF ENDOBRONCHIAL ULTRASOUND FOR DIAGNOSIS OF PATIENTS WITH SUSPECTED LUNG CANCER Tran Tan Cuong1; Mai Xuan Khan1; Do Quyet2 SUMMARY Objectives: To evaluate the efficiency of endobronchial ultrasound for diagnosis of patients with suspected lung cancer. Subjects and methods: 82 patients with suspected primary lung cancer, treated at 103 Military Hospital and K3 Hospital located at Tan Trieu, Thanh Tri, Hanoi from January, 2013 to October, 2018. Equipment and tools: Endoscope system OLYMPUS BF-UC180F. Results: The tumor detection efficiency of endobronchial ultrasound compared to surgery had the sensitivity of 72.97%, the specificity of 50.0%. The sensitivity and specificity of endobronchial ultrasound biopsy compared to the postoperative tumor pathology results were 89.4% and 100%, respectively. The effectiveness of identifying mediastinal lymph node of endobronchial ultrasound compared to surgery has sensitivity of 82.85%, specificity of 77.78%. Compared with the results of postoperative pathology, the endobronchial ultrasound lymphnode biopsy results had the sensitivity of 50.0% and the specificity of 100%. Conclusion: Endobronchial ultrasound technique was more effective than bronchoscopy in terms of sensitivity and specificity for diagnosis of primary lung cancer. * Keywords: Primary lung cancer; Endobronchial ultrasound; Efficiency. INTRODUCTION Primary lung cancer (PLC) is the most common with the highest death-rate in the world. According to the World Health Organization (WHO), every year 886,000 males and 330,000 females die from PLC globally. In Vietnam, PLC ranks the first among cancers in males and the third in females [1, 8]. Currently, there have been many advances in the diagnosis and early diagnosis of PLC such as screening techniques (CT, magnetic resonance, virtual endoscopy, positron emission tomography); endoscopy and biopsy techniques (ultrasound endoscopy, fluorescent endoscopy, magnetic anchor guided endoscopic...) [2]. Endobronchial ultrasound (EBUS) and endoscopic ultrasound guided biopsy such as transbronchial fine needle aspiration (TBNA) and fine needle aspiration (FNA) play an important role in the diagnosis of PLC, stage diagnosis and evaluation of treatment results. EBUS has advantages over conventional endoscopy, fluorescence guided endoscopy, fluorescence endoscopy and mediastoscopy, because this is a minimally invasive technique, but still allows investigating the internal lesions of bronchi and bronchial walls, bronchial adjacent lesions, peripheral lung lesions and 1. 103 Military Hospita 2. Vietnam Military Medical University Corresponding author: Tran Tan Cuong (trancuongqy1@gmail.com) Date received: 18/12/2019 Date accepted: 7/1/2020 Journal of military pharmaco-medicine n 0 1-2020 189 especially mediastinal lesions. The techniques of EBUS-guided aspiration biopsy and TBNA biopsy are more effective than conventional endoscopy: the sensitivity and specificity for diagnosis of primary lung cancer were 73% and 100%. In the diagnosis of mediastinal lymphadenopathy, sensitivity, specificity and accuracy were 95%, 100% and 92%, respectively [9]. In our country, in 2011, EBUS was initially applied in the Department of Tuberculosis, 103 Military Hospital and it has just developed more applications in some other hospitals. However, there have been very few reports of effectiveness on Vietnamese patients. Thus, we conducted this study: To evaluate the effectiveness of EBUS for diagnosing patients with suspected lung cancer. SUBJECTS AND METHODS 1. Subjects. 82 patients with suspected PLC, treated at 103 Military Hospital and K3 Tan Trieu Hospital, Thanh Tri, Hanoi from January 2013 to October 2018. * Inclusion criteria: - Clinical symptoms or not. - Prolonged cough or hemoptysis. - Normal chest X-ray and/or thoracic CT: Single circular, multi-bowed, ragged translucent white mass. Attached images: pneumothorax, pneumonia, pleural effusion. There may be lung hilum or mediastinal lymph nodes. - Agreed to participate in the research. * Exclusion criteria: - Contraindications to bronchoscopy. - Diagnosed with other cancers of PLC and HIV infection. - Did not agree to participate in the research. 2. Methods. * Research design: Prospective clinical research, cross- sectional description, convenient sampling. * Methods: - All patients underwent EBUS, assessed lesions, performed EBUS-guided aspiration biopsies or TBNA biopsy when indicated at endoscopy room, Respirator Department of 103 Military Hospital and Department of Endoscopy K3 Tan Trieu Hospital. - Equipment and tools: Endoscope OLYMPUS BF-UC180F. * Data analysis: Inputting, managing and analyzing data through STATA 14.2 software. Describe the qualitative variables by frequency and rate (%). The variables are quantitative via parameters: mean (X) and standard deviation (SD). Comparing the mean number difference and the difference in distribution ratio by the tests: χ2 test, Fisher test... Calculate sensitivity (Se), specificity (Sp) (according to the formula of Goldman L (1998). Journal of military pharmaco-medicine n 0 1-2020 190 RESULTS 1. Tumor and lymph node determination by EBUS compared with CT. Table 1: Determination of tumor position by EBUS and CT (n = 82). EBUS CT Position n % n % p* Trachea 2 2.44 0 - Right main bronchus 3 3.66 0 - Right upper lobe bronchus 13 15.85 15 18.29 Right middle lobe bronchus 5 6.1 8 9.76 Right lower lobe bronchus 5 6.1 17 20.73 Right upper and middle lobe bronchia 2 2.44 0 - Left main bronchus 2 2.44 0 - Left upper lobe bronchus 8 9.76 20 24.39 Left lower lobe bronchus 9 10.98 13 15.85 No tumor 33 40.24 9 10.98 0.000 (*: Fisher’s exact) The total number of tumors detected on CT was greater than that on EBUS (73 and 49). Among them, the highest number of detectable cases on CT was in the left upper lobe (20 cases), accounting for 24.39%; the lower right lung was detected in 17 cases, accounting for 20.73%. EBUS found most cases in the right upper lobe (13 cases) accounting for 15.85% and the left lower lobe was detected 9 cases, accounting for 10.98%. However, on CT, it was impossible to diagnose the tumors in trachea and main bronchia like NSPQSA. The difference between tumor position detection on CT and EBUS was statistically significant with p < 0.05. Table 2: Compare the ability to detect tumors between CT and EBUS. CT EBUS Tumor No tumor KAPPA (p) p* Tumor 47 (64.38%) 2 (22.22%) No tumor 26 (35.62%) 7 (77.78%) 0.194 (0.008) 0.027 (*: Fisher’s exact) Comparing the diagnosis of tumor between EBUS and CT showed 64.38% of inter-rater reliability in tumors and 77.78% of inter-rater reliability in no tumors. The difference was only statistically significant with p < 0.05. Journal of military pharmaco-medicine n 0 1-2020 191 Table 3: Compare the diagnosis of mediastinal lymph node between EBUS and CT (n = 82). EBUS CT Location (group) n % n % p* 2R 4 4.88 7 8.54 0.305 4R 11 13.41 6 7.32 0.029 4L 2 2.44 5 6.1 1.00 7 13 15.85 11 13.41 0.068 10R 15 18.29 5 6.1 0.040 10L 10 12.20 6 7.32 0.002 11R 13 16.67 2 2.56 1.00 11L 8 9.76 1 1.22 0.098 (*: Fisher’s exact) EBUS had more pro-ability to diagnose lymph node group 4R, 7, 10R, 10L, 11R and 11L, but less in 2R and 4L than CT. However, the differences in lymph node group 2R, 4L, 7, 11R, 11L were not statistically significant with p > 0.05. Table 4: Comparing lymph node detection between EBUS and CT (n = 82). EBUS Methods Node No node Total p* Node 27 (32.93%) 9 (10.97%) 36 (43.9%) CT No node 18 (21.95%) 28 (34.15%) 46 (56.1%) Total 45 (54.88) 37 (45.12%) 82 (100.0) 0.001 KAPPA 0.349 (p = 0.0006) (*: Pearson Chi2 test = 10.493) In the detection of lymph node, EBUS diagnosed with lymphadenopathy in 45 patients, accounting for 54.88%, whereas the CT diagnosed with lymphadenopathy in 36 patients, accounting for 43.9%. The difference was statistically significant with p < 0.05. The inter-rater reliability was 0.349 (p < 0.05). Journal of military pharmaco-medicine n 0 1-2020 192 2. Comparison of diagnosis between EBUS biopsy and surgery. * The effectiveness of EBUSA's tumor and lymph node detection compared with surgery: Table 5: Comparing tumor detection effectiveness between EBUS and surgery (n = 43). Surgery EBUS Tumor No tumor Sensitivity/ specificity p * Tumor 10 (90.91%) 3 (60.0%) Central tumor (n = 16) No tumor 1 (0.09%) 2 (40.0%) Se = 90.91% Sp = 40% 0.214 Tumor 17 (65.38%) 0 Peripheral tumor (n = 27) No tumor 9 (3.62%) 1 (100.0%) Se = 65.38% Sp = 100% 0.370 Tumor 27 (72.97%) 3 (50.0%) Total (n = 43) No tumor 10 (27.03%) 3 (50.0%) Se = 72.97% Sp = 50.0% 0.345 (*: Fisher’s exact) 43 patients were determined tumor after surgery, EBUS detected 27 cases, accounting for 72.97%. EBUS’s detection accuracy was 90.91% for central tumors and 65.38% for peripheral tumors. The sensitivity and specificity for EBUS’s tumor detection were 72.97% and 50.0%, respectively. Table 6: Comparing lymph node detection effectiveness between EBUS and surgery (n = 43). Surgery Method Node No node Total Sensitivity/ specificity, p* Node 28 (82.35%) 2 (22.22%) 30 (69.77%) EBUS No node 6 (17.65%) 7 (77.78%) 13 (30.23%) Total 34 (100.0%) 9 (100.0%) 43 (100.0%) Se = 82.35 Sp = 77.78 p = 0.001 (*: Fisher’s exact) Comparing lymph node detection with EBUS and surgery, the results showed 82.35% of sensitivity and 77.78% of specificity with p < 0.05 Table 7: Comparing lymph node diagnosis between EBUS and surgery (n = 43). EBUS Surgery Location n % n % p 2R 3 6.98 2 4.65 0.136 2L 0 - 0 - - 4R 9 20.93 6 13.95 0.000 4L 1 2.33 2 4.65 0.047 Journal of military pharmaco-medicine n 0 1-2020 193 7 8 18.6 8 18.6 0.000 10R 11 25.58 11 25.58 0.000 10L 7 16.28 8 18.60 0.001 11R 12 27.91 14 32.56 0.000 11L 5 11.63 7 16.28 0.001 5 0 - 6 13.95 - 6 0 - 1 2.33 - 8 0 - 1 2.33 - 9 0 - 10 23.26 - 12R 0 - 2 4.65 - 12L 0 - 3 6.98 - (*: Fisher’s exact) Compared to the postoperative diagnosis, EBUS mainly detected lymph node groups 2R, 4R, 7, 10, 11 and did not detect lymph nodes in other groups such as 5, 6, 8, 9, 12. The rate of lymph nodes 4L detection was lower, whereas the 2R, 4R groups was higher in EBUS; the difference was statistically significant with p < 0.05. 3. Other results. Table 8: Sampling times of EBUS biopsy (n = 82). Linear probe (n = 50) Radial probe (n = 32) Total (n = 82) Sampling times per tumor/node n % n % n % 2 6 19.35 0 0 6 18.18 3 3 9.68 0 0 3 9.09 4 22 70.97 2 100.0 24 72,73 Node sampling times Total 31 100.0 2 100.0 33 100.0 2 2 8.33 0 0 2 3.13 3 2 8.33 1 10.0 3 6.25 4 20 83.33 9 90.0 29 90.63 Tumor sampling times Total 24 100.0 10 100.0 34 100.0 For each lesion, there should be 2 to 4 needle aspirations to get enough samples. With lymph nodes, conducted 4 aspiration times (72.73%), 2 times (18.18%) and 3 times (9.09%). Biopsy of tumor conducted 4 times to get enough samples (90.63%), and 2 times (3.13%). Journal of military pharmaco-medicine n 0 1-2020 194 * Complications of EBUS biopsy (n = 82): There was a small number of complications after performing EBUS biopsy, hemorrhage (8 patients = 9.76%), hemoptysis (2 patients = 2.44%) and fever (4 patients = 4.88%). DISCUSSION Endobronchial ultrasound is one of the lastest advances in respiratory endoscopy. Studies showed that this technique was remarkably safe and cost-effective. In recent years, EBUS has expanded its range of applications due to new support tools such as 19G needles and micro-clamps. Along with the breakthrough achievements in the recent treatment of lung cancer, EBUS is considered as a safe and effective diagnostic tool [7, 11, 12]. The research results showed that the sampling rate was higher than other bronchoscopy techniques in some studies in our country. Specifically, according to a study by Dong Duc Hung (2014) in 125 patients with lung tumors, only 25 patients (20%) obtained tumor samples using EBUS technique [2]. By rigid bronchoscope, Hoang Hong Thai (2009) also collected only 76 samples out of 142 patients (53.52%). The reasons for not obtaining the sample were: difficulty in sampling from smooth mucosa, and some tumors outside the biopsied bronchus, so it cannot be conducted [4]. Overally, the rate of sampling through normal bronchoscopic techniques was not high. The pathological results from samples obtained through EBUS showed that the rate of cancer detections for tumor samples was quite high (79.41%). This showed that under ultrasound guidance, tumor lesions were located and biopsied more accurately. This result was consistent with the bronchoscopy method by Hoang Hong Thai et al (2009) [5]. For lymph nodes, the biopsy results under the guidance of EBUS identified 5 patients with lymphoma metastasis (19.23%), 21 patients with inflammatory. According to a study by Hazem El-Osta (2017), the rate of metastatic lymph node detection accounted for 12.8% and corresponded to a specificity of 93.0% [13]. The determination of lymph node metastasis in lung cancer was one of the significant factors to determine surgical indication. According to Nguyen Huy Luc et al (2013), the rate of mediastinal lymph node and adjacent bronchial lymph node dectections through EBUS were lower than that through CT. However, by EBUS, we can perform transbronchial aspiration to take a sample and identify the nature of the lymph node [6]. For tumor and lymph node specimens obtained from EBUS, the incidence of cancer cells was found to be lower than for those collected from surgery. The reason was that after removing the lobes and lymph nodes, the tumor and the lymph nodes were taken for independent biopsies, so the accuracy was higher. For lymph node sample, we found that many lymph nodes were located far from the ultrasound probe, some of them were difficult to biopsy leading to lower sensitivity. Tian et al (2013) showed the sensitivity, specificity, negative predictive value and accuracy in the diagnosis of lung and mediastinum cancer of TBNA under the guidance of the EBUS were 96%, 100%, Journal of military pharmaco-medicine n 0 1-2020 195 33% and 96%, respectively [14]. Several studies suggested that the TBNA biopsy under EBUS guidance was highly effective for central and mediastinal lung cancers, less effective for peripheral lung cancers [15]. When comparing the results of EBUS with CT, the ultrasound detected only 49 cases of tumors, while the CT detected 73 cases. Although the number of patients found to be less than tumors with CT, EBUS was able to identify tumors position where CT could not be identified as in trachea, right main bronchus, left main bronchus. The similarity of the two methods in identifying node groups was not high. Specifically, in 82 EBUS patients detected 75 abnormal lymph nodes, while CT only detected 43 abnormal lymph nodes. EBUS detected more lymph node groups 4R, 7, 10R, 10L, 11R and 11L, but less lymph node groups 2R and 4L than CT. Some documents suggested that using EBUS to guide TBNA biopsy, the accurate detection rate of mediastinal lymph nodes may be up to 92%, higher than that of blind TBNA biopsy (86%) [7]. Research by Yasufuku T et al (2006), for the diagnosis of mediastinal and lung hilum lymph nodes, EBUS had high sensitivity and specificity compared with PET and CT [16]. Tran Van Ngoc (2014) pointed out that EBUS was able to detect 25% of metastatic mediastinal lymph nodes and prevented 12% of unnecessary surgery due to the N2/N3 stage cancers [7]. Besides the robust advantages, EBUS also has some limitations. It is not possible to investigate the entire mediastinum, especially anterior mediastinum periphery metastasal tumors. In addition, this technique also has difficulty in examining lung tumors in some areas such as the upper lobe or angled bronchial areas. TBNA biopsy under EBUS guidance is also difficult in the elderly due to the narrow cartilage space and calcification of bronchial cartilage. In addition, the cough reaction during process could also affect the success of the technique [17]. Therefore, this technique often applied to lesions in the center of the lung, near the bronchi and tumors in the postero-inferior mediastinum. CONCLUSION According to this research on 82 patients with suspected lung cancer, from January 2013 to October 2018, the results showed that: - Endobronchial ultrasound biopsy obtained samples of tumor and lymph node with a high rate. Among them, probability of collecting 4 samples from tumor biopsy accounted for 90.63%, from lymph nodes accounted for 72.73%; the percentage of malignant samples in tumors was 74.41%, in lymph nodes was 19.23%. - The tumor detection effectiveness of EBUS compared to surgery had a sensitivity of 72.97%, specificity of 50.0%. EBUS biopsy effectiveness compared to the postoperative pathological results was 89.4% sensitivity and 100% specificity. - The mediastinal lymph nodes detection effectiveness of EBUS compared with surgery had sensitivity of 82.85%, specificity of 77.78%. Compared with postoperative pathological results, EBUS biopsy was performed with 50.0% sensitivity and 100% specificity. Journal of military pharmaco-medicine n 0 1-2020 196 - Low rate of complications: hemorrhage 9.76%; hemoptysis after biopsy 2.44%; fever 4.88%. REFERENCES 1. Đức N.B, Thưởng L.Đ và CS. Ghi nhận tình hình ung thư giai đoạn 2001 - 2004 tại 5 tỉnh và thành phố Việt Nam.Tạp chí Y học Thực hành. 2006, tr.9-17. 2. Châu N.Q. Ung thư phổi. Nhà xuất bản Y học. Hà Nội. 2008, tr.15. 3. Đồng Đức Hưng. Nghiên cứu phương pháp xuyên thành ngực trong chẩn đoán tổn thương phổi. Trường Đại học Y Dược TP. Hồ Chí Minh. 2014, tr.155. 4. Hoàng Hồng Thái. Nội soi phế quản ống cứng chẩn đoán ung thư phế quản tại Khoa Hô hấp - Bệnh viện Bạch Mai. Tạp chí Y học Thực hành, 2009. 667 (7), tr.61-64. 5. Hoàng Hồng Thái. Nghiên cứu hiệu quả của phương pháp chọc hút bằng kim Wang 319 trong nội soi phế quản chẩn đoán ung thư phế quản tại Khoa Hô hấp, Bệnh viện Bạch Mai. Tạp chí Y học Lâm sàng. 2009, 42 (7), tr.41-44. 6. Nguyễn Huy Lực và CS. Hiệu quả bước đầu của nội soi siêu âm phế quản - sinh thiết trong chẩn đoán ung thư phế quản. Tạp chí Y - Dược học Quân sự. 2013, số 3, tr.1-4. 7. Ngọc T.V. Chẩn đoán sớm và điều trị ung thư phổi qua nội soi phế quản. Tạp chí Y học Thành phố Hồ Chí Minh. 2014, 18 (Phụ bản số 1), tr.1-4. 8. Alberg A.J, J.G. Ford, J.M. Samet. Epidemiology of lung cancer: ACCP evidence- based clinical practice guidelines. Chest Journal. 2007, 132 (3 suppl), pp.29S-55S. 9. T. B, H. F J. Endobronchial ultrasound: A new innovation in bronchoscopy. Lung India: Official organ of Indian Chest Society. 2009, 26 (1), pp.17-21. 10. He T, A.C.Mehta. Linear endobronchial ultrasound: What's new? Semin Respir Crit Care Med. 2018, 39 (6), pp.649-660. 11. Canadian Agency for D and. H. Technologies in, Endobronchial ultrasound for lung cancer diagnosis and staging: A review of the clinical and cost-effectiveness. CADTH Technology Overviews. 2010, 1 (2), pp.e0115-e0115. 12. Sampsonas F, L. Kakoullis. EBUS: Faster, cheaper and most effective in lung cancer staging. 2018, 72 (2). 13. El-Osta H, Jani P, Mansour A, Rascoe P, Jafri S. Endobronchial ultrasound for nodal staging of patients with non-small-cell lung cancer with radiologically normal mediastinum. A meta-analysis. Ann Am Thorac Soc. 2018, Jul, 15 (7), pp.864-874. doi: 10.1513/AnnalsATS. 201711-863OC. 14. Tian Q et al. The reasons of false negative results of endobronchial ultrasound- guided transbronchial needle aspiration in the diagnosis of intrapulmonary and mediastinal malignancy. Thoracic Cancer. 2013, 4 (2), pp.186-190. 15. Sanz-Santos J et al. The role of endobronchial ultrasound in central early lung cancer. Thoracic Cancer. 2012, 3 (2), pp.139-144. 16. Yasufuku K et al. Comparison of endobronchial ultrasound, positron emission tomography, and CT for lymph node staging of lung cancer. Chest. 2006, 130 (3), pp.710-718. 17. Jalil B.A, K. Yasufuku, A.M. Khan. Uses, limitations, and complications of endobronchial ultrasound. Proceedings (Baylor University. Medical Center). 2015, 28 (3), pp.325-330.