Abstract
Background
Endoscopic ultrasound features predicting malignant mediastinal lymphadenopathy have been extensively studied. The Canada Lymph Node Score (CLNS) comprises four endoscopic ultrasound characteristics, at least three of which have been shown to predict malignancy in a Western population. Two of the CLNS variables are commonly described in tuberculous mediastinal lymphadenopathy as well, thus posing a diagnostic challenge.
Objectives
To evaluate the diagnostic accuracy of the CLNS in predicting malignant mediastinal involvement (primary objective), and to study other endoscopic ultrasound variables to formulate a new predictive model (secondary objective).
Methods
In this prospective cross-sectional study from North India, consecutive patients undergoing endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) were included. Nodal endoscopic ultrasound features were recorded in real time before needle aspiration, and the CLNS was determined for each lymph node. The primary outcome was diagnostic accuracy of the CLNS, determined by the area under the curve (AUC). The secondary outcome was the identification of additional ultrasound predictors to develop a new model by logistic regression.
Results
A total of 132 nodes from 82 patients were sampled. Characteristics examined were short-axis diameter, central hilar structure, coagulation necrosis sign (CNS), margins, echogenicity, vascular patterns, conglomeration, shape and calcification. The CNS (odds ratio (OR) 4.28, p=0.005), heterogeneous echogenicity (OR 2.45, p=0.036) and round shape (OR 3.02, p=0.03) significantly predicted malignant involvement. The mean (standard deviation) AUC for the CLNS and the new model (comprising the CNS, round shape and heterogeneous echogenicity) was 0.72 (0.04) (95% confidence interval (CI) 0.635 - 0.814) and 0.74 (95% CI 0.635 - 0.844), respectively.
Conclusion
The CLNS predicts malignant mediastinal involvement at a cut-off score of 3 in a tuberculosis-endemic setting. The proposed new model demonstrated comparable diagnostic performance to that of the CLNS for discriminating malignant from non-malignant nodes.
Study synopsis
What the study adds. In a tuberculosis (TB)-endemic setting, the Canada Lymph Node Score (CLNS) maintains reasonable diagnostic accuracy for predicting malignant mediastinal lymph node involvement. A new model, comprising the coagulation necrosis sign, round shape and heterogeneous echotexture, shows equivalent diagnostic performance.
Implications of the findings. The CLNS is a clinically valuable tool for differentiating malignant from non-malignant mediastinal nodes in TB-prevalent regions. However, overlapping endoscopic ultrasound features between granulomatous and malignant conditions may reduce its specificity.
Keywords: Canada Lymph Node Score, diagnostic accuracy, EBUS-TBNA, endoscopic ultrasound, lung cancer, tuberculosis
Background
Mediastinal lymphadenopathy has various causes, which can be benign or malignant.[1] Transbronchial needle aspiration (TBNA) of mediastinal nodes has been done for the past five decades, initially via rigid bronchoscopy and later through flexible bronchoscopy. Since 1992, endobronchial ultrasound (EBUS) has been the preferred modality of pulmonologists and thoracic oncologists for real-time visualisation of a node during needle aspiration.[2] It replaced cervical mediastinoscopy, which was traditionally considered the procedure of choice for evaluating metastatic mediastinal lymphadenopathy. The American College of Chest Physicians’ lung cancer guidelines reported an overall median sensitivity of 89% and a median negative predictive value of 91% for EBUS-TBNA in the mediastinal staging of lung cancer.[3] The corresponding figures reported for traditional mediastinoscopy were 78% and 91%, respectively.[3] Based on these findings, the guidelines recommended ultrasound-guided, needle based sampling techniques over surgical staging as the first step in the mediastinal staging of lung cancer.[3]
Convex-probe EBUS facilitates real-time visualisation of lymph nodes as small as 2 mm along the tracheobronchial tree, which may not even be visible on computed tomography scans.[4] It can assess various endoscopic ultrasound characteristics of the nodes, which helps the operator choose the site and station for obtaining nodal samples. Using four of these lymph node characteristics – namely short-axis diameter (SAD) ≥10 mm, absence of central hilar structure (CHS), presence of the coagulation necrosis sign (CNS), and distinct margins – Hylton et al.[5] proposed the Canada Lymph Node Score (CLNS), which proved useful in predicting malignant mediastinal nodal involvement. Their research was done on patients with suspected thoracic malignancy in a non-tuberculosis (TB) Western setting.
Two of the four CLNS variables – presence of the CNS and distinct margins – are commonly also described in TB mediastinal lymphadenopathy.[6,7] There have been no studies examining the diagnostic usefulness of CLNS in a TB-endemic population. We therefore planned the present study to address this research gap and to evaluate whether other combinations of parameters could better predict malignant involvement in such a population.
Methods
Study design and population
This single-centre, prospective cross-sectional study was conducted at a tertiary care hospital in Northern India between April 2023 and February 2024. Consecutive sampling was used to recruit 82 patients undergoing EBUS-TBNA, with 132 nodes being sampled during this period. Eligible participants included all patients aged >18 years of age who presented with clinically malignant or benign mediastinal lymphadenopathy and consented to undergoing the EBUS-TBNA procedure. Each participant underwent a detailed clinico-radiological assessment before the procedure. Patients with any contraindication to bronchoscopy and lymph nodes that could not be sampled were excluded. The research was conducted in accordance with the Declaration of Helsinki and the protocol was approved by the Institutional Ethics Committee of the All India Institute of Medical Sciences, Rishikesh (India) (ref. no. AIIMS IEC/23/52, dated 10 March 2023). Fully informed written consent was obtained from each participant at the time of enrolment.
EBUS-TBNA procedure
A trained operator performed EBUS-TBNA using the Olympus BF-UC180F echo-bronchoscope (Olympus Medical Systems Corp., Japan) integrated with the EU-ME1 ultrasound processor (Olympus Medical Systems Corp., Japan). A standard protocol for administering premedication, topical anaesthesia and sedatives was followed. The procedure was done via the nasal or oral route based on operator preference, with the patient lying supine and vital parameters being continuously monitored. Lymph nodes were systematically identified according to the 7th edition of the International Association for the Study of Lung Cancer anatomical location.[8] Nodal endoscopic ultrasound characteristics were identified by the operator (first rater) before sampling, and a video was simultaneously recorded. The variables noted were SAD (≥10 mm or <10 mm), CHS (present/ absent), CNS (present/absent), margins (distinct/indistinct), shape (round/oval/triangular/ill-defined), echogenicity (heterogeneous/ homogeneous), conglomeration (present/absent), vascular pattern (central/peripheral), grade of colour Doppler (grade 0/1/2/3), bronchial artery inflow (present/absent), and calcification (absent/ present).[9] Lung cancer patients were staged following the N3→N2→N1 sampling order.[10] In conditions clinically suspected to be benign, one or more nodal stations were sampled based on the operator’s discretion. Mediastinal node fine-needle aspiration was done using a 21-gauge EBUS-TBNA needle. The needle was not changed between two nodes of the same patient in cases where multiple stations were sampled. Smears as well as cell blocks were prepared from the aspirates.
A second, independent rater assessed the recorded video later. Clinical information on the suspected underlying disease was available to both raters before assessing the endoscopic ultrasound characteristics, but both raters were blinded to the final reports. They were experienced in interventional pulmonology and EBUS, having done over 100 EBUS bronchoscopy procedures during the previous year. In the case of any discrepancy between the two raters, the endoscopic ultrasound characteristics noted by the first rater who performed the procedure were taken into account, as the lymph node characteristics were assessed in real time, thereby allowing for more accurate interpretation. Patients were actively monitored during and after the procedure for potential adverse events such as bleeding, respiratory distress or any other untoward clinical concerns. Vital parameters such as oxygen saturation, heart rate and blood pressure were also monitored during the procedure as well as up to 2 hours later, to enable prompt detection of any complications.
The index test used to derive the area under the receiver operating characteristic curve (AUC) was the CLNS. It was considered positive when three or more of the following endoscopic ultrasound characteristics were noted – SAD ≥10 mm, CNS, distinct margins, and absence of CHS. The score was considered negative if fewer than three of these four characteristics were observed.
Pathological reporting
Rapid on-site evaluation was performed either by the pathologist or by the assisting pulmonology fellow (if the pathologist was unavailable), using toluidine blue stain. The smears were examined after staining with May-Grünwald-Giemsa and Papanicolaou stains. Cell blocks were stained using haematoxylin and eosin stain. If needed, pertinent immunohistochemistry (e.g. PanCK (pan cytokeratin), TTF1 (thyroid transcription factor 1), P63 (tumour protein 63), synaptophysin) was performed on the cell block. The pathologist had access to the clinical information during evaluation, but was unaware of the suspected underlying cause according to CLNS results. The cytopathology report based on integration of the EBUS-guided fine-needle aspiration smears and cell blocks was considered the gold standard for interpretation.
Lymph node aspirates were reported in the following categories: (i) inadequate/unsatisfactory (number of cells was too small (<40) or the quality of the smear was unsuitable for opinion); (ii) benign (smear was adequate for interpretation and no malignant cells were seen); (iii) malignant (definite malignant cells were seen, in sufficient quantity to warrant a confident diagnosis); and (iv) indeterminate category (adequate, but neither benign nor malignant).
Primary and secondary objectives
The primary objective was to evaluate the diagnostic accuracy of the CLNS in predicting malignant mediastinal lymph node involvement, using the AUC as the measure of performance. The secondary objective was to assess endoscopic ultrasound features other than the four CLNS variables and to formulate a new model to predict malignant mediastinal lymphadenopathy.
Statistical analysis
The sample size of 132 nodes was sufficient to achieve a confidence interval (CI) of <0.1, assuming an AUC of 0.7. The AUC was calculated using SPSS statistical software, version 26.0 (IBM Corp., USA), to evaluate the diagnostic performance of the model.
The distribution of continuous variables was tested for normality and skewness using the Shapiro-Wilk test. Normally distributed data were expressed as means (standard deviation), whereas skewed data were summarised as medians (interquartile range). Categorical variables were presented as frequencies and percentages. Two independent raters assessed the endoscopic ultrasound characteristics of lymph nodes, and inter-rater agreement was evaluated using Cohen’s kappa statistical measure.
Univariate logistic regression was performed for inferential analysis to calculate the odds ratio (OR) for each endoscopic ultrasound characteristic. The final predictive model was developed based on significant independent predictors identified in the multivariate analysis, and its diagnostic performance was assessed using the AUC.
Results
One hundred and thirty-two lymph nodes were sampled from 82 patients undergoing EBUS-TBNA during the study period. The flow of participants is shown in Fig. 1. The baseline characteristics of the study population and clinical/endoscopic ultrasound features of the sampled lymph nodes are presented in Table 1. Based on initial clinical suspicion, most of the patients were thought to have malignant disease (n=88, 66.7%). The most frequently sampled node was station 7 (n=67, 53.2%), followed by 4R (n=36, 28.6%), 4L and 11L (n=7, 5.6% each), 10R (n=4, 3.2%), and 10L (n=3, 2.4%). Stations 11R and 2R were least frequently sampled (n=1, 0.4% each). The final diagnoses arrived at for the 44 clinically benign nodes were sarcoidosis (n=27), tuberculosis (n=12) and hypersensitivity pneumonitis (n=5). No clinically significant adverse events were encountered during the study.
Fig. 1.
Flow of participants in the study. (CLNS = Canada Lymph Node Score; EBUS-TBNA = endobronchial ultrasound-guided transbronchial needle aspiration.)
Table 1. Baseline clinical and lymph node characteristics.
| Characteristic | n (%)* |
| Clinical characteristics | Study population (n=82) |
| Age (years), mean (SD) | 50.61 (14.54) |
| Male gender | 50 (61.0) |
| Clinical suspicion of malignancy | 88 (66.7) |
| Clinical suspicion of benignity | 44 (33.3) |
| Lymph node characteristics | Lymph nodes sampled (n=132) |
| Thoracic malignancy | 55 (62.5) |
| Extrathoracic malignancy | 33 (37.5) |
| Cytopathology | |
| Negative for malignancy | 89 (67.4) |
| Malignant cells present | 31 (23.5) |
| Inadequate for opinion | 12 (9.1) |
| Short-axis diameter, median (IQR) | 13 (10 - 17.05) |
| Short-axis diameter >10 mm | 92 (69.7) |
| CHS absent | 109 (82.6) |
| CNS present | 25 (18.9) |
| Distinct margin | 108 (81.8) |
| CLNS positive (≥3/4) | 72 (54.5) |
| Heterogeneous echogenicity | 56 (42.4) |
| Central vascular pattern | 32 (24.2) |
| Conglomeration absent | 100 (75.8) |
| Shape | |
| Round | 27 (20.5) |
| Oval | 62 (47) |
| Triangular | 8 (6.1) |
| Ill-defined | 35 (26.5) |
| Calcification present | 6 (4.5) |
| Bronchial artery inflow present | 24 (18.2) |
| Colour Doppler | |
| Grade 0 | 48 (36.4) |
| Grade 1 | 60 (45.5) |
| Grade 2 | 21 (15.9) |
| Grade 3 | 3 (2.3) |
SD = standard deviation; IQR = interquartile range; CHS = central hilar structure; CNS = coagulation necrosis sign; CLNS = Canada Lymph Node Score. *Except where otherwise indicated.
Among the four endoscopic ultrasound characteristics of the CLNS for predicting malignant mediastinal lymphadenopathy, statistical significance was achieved for the CNS (OR 4.28, 95% CI 1.43 - 13.07, p=0.005) (Table 2). For patients with a positive CLNS, the OR was 5.51 (95% CI 1.96 - 18.10, p<0.001). Among endoscopic ultrasound variables other than those of the CLNS, heterogeneous echotexture and round shape were associated with statistically significant prediction of malignant mediastinal lymphadenopathy (Table 2). Other ultrasound features such as SAD, conglomeration, calcification, vascular patterns, bronchial artery inflow and grade of colour Doppler were not associated with significant odds of predicting malignant mediastinal node involvement (Table 2).
Table 2. ORs of various endoscopic ultrasound characteristics in predicting malignant mediastinal lymphadenopathy.
| Characteristic | OR | 95% CI | p-value |
| SAD ≥10 mm | 2.62 | 0.80 - 11.34 | 0.14 |
| CHS: Absent | 3.90 | 0.85 - 36.69 | 0.097 |
| CNS: Present | 4.28 | 1.43 - 13.07 | 0.005* |
| Margins: Distinct | 2.52 | 0.66 - 14.31 | 0.18 |
| CLNS: Positive | 5.51 | 1.96 - 18.10 | <0.001* |
| Echogenicity: Heterogeneous | 2.45 | 0.99 - 6.22 | 0.036* |
| Vascular pattern: Central | 1.06 | 0.36 - 2.91 | 1 |
| Conglomeration: Absent | 3.04 | 0.81 - 17.13 | 0.12 |
| Shape: Round | 3.02 | 1.02 - 8.90 | 0.03* |
| Calcification: Present | 0 | 0.00 - 2.42 | 0.337 |
| Bronchial artery inflow: Present | 0.95 | 0.28 - 2.86 | 1 |
OR = odds ratio; CI = confidence interval; SAD = short-axis diameter; CHS = central hilar structure; CNS = coagulation necrosis sign; CLNS = Canada Lymph Node Score. *Statistically significant (p<0.05)
Table 3 shows the reliability assessment for individual ultrasonographic features. Inter-rater reliability was excellent for calcification; good for shape, conglomeration, CHS, CNS, margins, and colour Doppler; and moderate for echogenicity.
Table 3. Reliability assessment for individual endoscopic ultrasound features.
| Feature | Agreement, % | Cohen’s kappa |
| CHS: Absent | 90.2 | 0.675 |
| CNS: Present | 92.4 | 0.745 |
| Margins: Distinct | 97.7 | 0.727 |
| Echogenicity: Heterogeneous | 75.0 | 0.494 |
| Conglomeration: Absent | 100 | 1 |
| Shape: Round | 86.4 | 0.795 |
| Calcification: Present | 99.2 | 0.919 |
| Colour Doppler | 81.1 | 0.716 |
CHS = central hilar structure; CNS = coagulation necrosis sign.
The diagnostic performance of all endoscopic ultrasound characteristics in predicting malignant mediastinal involvement is shown in Table 4. Of the lymph nodes, 54.5% (n=72) were CLNS positive, while 45.5% (n=60) were CLNS negative. The AUC for the CLNS in predicting malignant mediastinal lymphadenopathy in this study population was 0.724 (95% CI 0.635 - 0.814).
A new score was computed by logistic regression analysis using the CNS, round shape and heterogeneous echogenicity. The AUC for the new model was 0.74 (95% CI 0.635 - 0.844), demonstrating diagnostic performance comparable to that of the CLNS.
Table 4. Diagnostic performance of endoscopic ultrasound characteristics in predicting malignant mediastinal node involvement.
| Variable |
Sensitivity, %
(95% CI) |
Specificity, %
(95% CI) |
PPV, %
(95% CI) |
NPV, %
(95% CI) |
Diagnostic
accuracy, %
(95% CI) |
| SAD >10 mm* | 83.9 (66 - 95) | 37.1 (27 - 48) | 31.7 (22 - 43) | 86.8 (72 - 96) | 49.2 (40 - 58) |
| CHS | 93.5 (79 - 99) | 21.3 (13 - 31) | 29.3 (21 - 39) | 90.5 (70 - 99) | 40.0 (31 - 49) |
| CNS* | 35.5 (19 - 55) | 88.8 (80 - 94) | 52.4 (30 - 74) | 79.8 (71 - 87) | 75.0 (66 - 82) |
| Margins | 90.3 (74 - 98) | 21.3 (13 - 31) | 28.6 (20 - 39) | 86.4 (65 - 97) | 39.2 (30 - 48) |
| Vascular patterns (peripheral, central) | 25.8 (12 - 45) | 75.3 (65 - 84) | 26.7 (12 - 46) | 74.4 (64 - 83) | 62.5 (53 - 71) |
| Echogenicity* | 58.1 (39 - 75) | 64.0 (53 - 74) | 36.0 (23 - 51) | 81.4 (70 - 90) | 62.5 (53 - 71) |
| Shape* | 32.3 (17 - 51) | 86.5 (78 - 93) | 45.5 (24 - 68) | 78.6 (69 - 86) | 72.5 (64 - 80) |
| Conglomeration | 9.7 (2 - 26) | 75.3 (65 - 84) | 12.0 (3 - 31) | 70.5 (60 - 79) | 58.3 (49 - 67) |
| Bronchial artery inflow | 19.4 (7 - 37) | 79.8 (70 - 88) | 25.0 (10 - 47) | 74.0 (64 - 82) | 64.2 (55 - 73) |
| Calcification | 100 (89 - 100) | 6.7 (3 - 14) | 27.2 (19 - 36) | 100 (54 - 100) | 30.8 (23 - 40) |
| Colour Doppler (grade 3, 4) | 3.2 (0 - 17) | 97.8 (92 - 100) | 33.3 (1 - 91) | 74.4 (65 - 82) | 73.3 (64 - 81) |
| CLNS 1 | 100 (89 - 100) | 1.1 (0 - 6) | 26.1 (18 - 35) | 100 (3 - 100) | 26.7 (19 - 36) |
| CLNS 2 | 100 (89 - 100) | 9.0 (4 - 17) | 27.7 (20 - 37) | 100 (63 - 100) | 32.5 (24 - 42) |
| CLNS 3 | 80.6 (63 - 93) | 57.3 (46 - 68) | 39.7 (28 - 53) | 89.5 (78 - 96) | 63.3 (54 - 72) |
| CLNS 4 | 16.1 (5 - 34) | 97.8 (92 - 100) | 71.4 (29 - 96) | 77.0 (68 - 84) | 76.7 (68 - 84) |
CI = confidence interval; PPV = positive predictive value; NPV = negative predictive value; SAD = short-axis diameter; CHS = central hilar structure; CNS = coagulation necrosis sign; CLNS = Canada Lymph Node Score. *Statistical significance (p<0.05) achieved for that endoscopic ultrasound characteristic.
Discussion
This prospective cross-sectional study assessed, for the first time in a TB-endemic setting, the diagnostic performance of the CLNS in predicting malignant mediastinal lymph node involvement. The essential observation of the study was that the CLNS demonstrated reasonable diagnostic accuracy in this population. Use of the CLNS provides the operator with a useful pretest probability of malignant involvement of a node, thus helping them choose the appropriate nodal station and site for sampling. We also assessed endoscopic ultrasound characteristics other than the CLNS. Shape, the CNS and heterogeneous echogenicity were significantly associated with malignant lymphadenopathy. However, SAD, absence of the CHS, distinct margins, absence of conglomeration, lack of calcification and bronchial artery inflow were not significantly associated with malignant mediastinal involvement.
In countries with a high TB burden, tuberculous mediastinal lymphadenopathy is more common than in Western populations where TB is not endemic. Considerable overlap between malignant and tuberculous endoscopic ultrasound characteristics has been noted in previous studies analysing these features. Madan et al.[6] and Prasad et al.[7] compared the predictability of endoscopic ultrasound features between TB and malignancy and noted that the CNS and distinct margins were more commonly associated with a tuberculous cause than with malignancy. Both these endoscopic ultrasound characteristics are integral components of the CLNS, suggesting that its applicability in a high TB burden setting could be limited by this overlap.
Earlier studies have researched other ultrasound features of mediastinal lymph nodes and found varying results.[11,12] Fujiwara et al.[11] evaluated shape, margins, echogenicity and CHS, and found that round shape, distinct margins, heterogeneous echogenicity and absence of CHS were independent predictors of malignant mediastinal involvement. In contrast, Ozgul et al.[12] prospectively evaluated size, shape, margins, echogenicity, granular appearance and CHS, and found that SAD ≥10 mm, round shape, distinct margins, heterogeneous echogenicity and granular appearance were predictive of sarcoidosis. Malignant and benign mediastinal lymphadenopathy can therefore show similar ultrasound features to some extent.
Colour Doppler and vascularity were assessed by Nakajima et al.,[9] who developed a grading system that defined vascularity as grade 0 – no blood flow or a small amount of blood flow; grade 1 – few vessels running towards the centre from the hilum; grade 2 – few punctiform or rod-shaped vessels forming a long strip of the curve; and grade 3 – rich blood flow with more than four vessels that are twisted and helical and have low signal intensity. The sensitivity, specificity and diagnostic accuracy were 87.6%, 69.6% and 78%, respectively, when grades 0 and 1 were defined as benign and grades 2 and 3 were defined as malignant. In our study, statistical significance was not achieved for any of the vascular endoscopic ultrasound characteristics.
The CLNS was originally formulated by Hylton et al.[5] to predict malignant mediastinal lymphadenopathy. Their study was in a Western population that is largely free of TB, and focused on patients with suspected malignant mediastinal lymph nodes in lung or oesophageal cancer. In contrast, we included patients with clinically benign nodes and those with extrathoracic primary lesions, in addition to those with thoracic primary malignancy, and applied the CLNS across all mediastinal lymph nodes. The specificity and positive likelihood ratio reported in a subsequent clinical validation study of the CLNS were 84.4% and 4.0, respectively,[13] while in our study they were 57.3% and 1.89. This difference is likely to be due to differences in study populations as noted above, and to the occurrence of granulomatous lymphadenopathy in our setting, which can show endoscopic ultrasound features similar to malignant lymph nodes.
Additionally, we developed a new score that demonstrated a comparable AUC to the CLNS in predicting malignant nodal involvement. The new score comprised the CNS, round shape and heterogeneous echogenicity, and predicted malignant node involvement with an accuracy of 0.74 (95% CI 0.635 - 0.844) compared with 0.724 (95% CI 0.635 - 0.814) for the CLNS. It is important to note that ultrasound characteristics do not substitute for the need for EBUS sampling. However, the scoring is important in that it will help correlate with the cytopathology report and may be useful in the case of discordant results (e.g. a benign cytopathology report in the presence of a suspicious endoscopic ultrasound score may warrant repeat EBUS-guided lymph node sampling).
This research is notable for several reasons. First, it is a prospective study that objectively evaluates the performance of the CLNS in a TB-endemic setting, where granulomatous lymphadenopathy can mimic malignancy and pose a reasonable diagnostic challenge. Our results demonstrate that a positive CLNS reliably predicts malignant mediastinal lymph node involvement in this complex clinical setting, although with a reduced specificity. The CLNS may assist clinicians and trainees in refining their pre-procedural risk assessment, prioritising nodes for sampling, and potentially improving diagnostic efficiency. Second, the study assesses the usefulness of CLNS in both malignant and benign causes of mediastinal lymphadenopathy, thereby evaluating its broader applicability in a more diverse patient population. Third, the high inter-rater reliability between the two raters who assessed the endoscopic ultrasound characteristics highlights the validity of the findings.
The study has certain limitations. Because it involved a single centre, the performance of the new score is unknown in a wider population. Since this is the first study reporting our new score, validation studies will be required to confirm its diagnostic usefulness. Further multicentre studies exploring the accuracy of the new score would be useful to ensure its generalisability. Second, clinical/radiological information on previous or current TB was not collected for individual patients, and although the study was conducted in a TB-endemic setting, confirmed TB cases were relatively few. Nevertheless, the absence of this information should not affect the primary objective of the study, which was to evaluate the diagnostic performance of the CLNS across a real-world cohort rather than assess TB disease prevalence or clinical profile. Therefore, while it limits our ability to specifically assess the effect of TB on CLNS performance, it does not alter the interpretation of the main outcomes.
Conclusion
The CLNS can reliably predict malignant mediastinal nodal involvement in TB-endemic settings, with a cut-off score of 3 serving as an objective threshold to guide the need for further procedures. We also developed a new predictive model incorporating the CNS, round shape and heterogeneous echogenicity, which demonstrated a comparable AUC for differentiating malignant from non-malignant lymph nodes.
Footnotes
The datasets generated and analysed during the present study are available from the corresponding author (MM) on reasonable written request.
The research for this study was done in partial fulfilment of the requirements for NB’s DM (Pulmonary Medicine) degree at the University of All India Institute of Medical Sciences, Rishikesh.
Conceptualisation: NB, MM; data acquisition: NB, RS; data/statistical analysis: NB, NC, KSK, MM; literature search and preparation of first draft: NB, KSK, RS; manuscript review, editing and approval of final draft: all authors; supervisor and guarantor: MM.
References
- 1.Scano V, Fois AG, Manca A, et al. Role of EBUS-TBNA in non-neoplastic mediastinal lymphadenopathy: Review of literature. Diagnostics (Basel) 2022;12(2):512. doi: 10.3390/diagnostics12020512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Jaliawala HA, Farooqui SM, Harris K, Abdo T, Keddissi JI, Youness HA. Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA): Technical updates and pathological yield. Diagnostics (Basel) 2021;11(12):2331. doi: 10.3390/diagnostics11122331. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Silvestri GA, Gonzalez AV, Jantz MA, et al. Methods for staging non-small cell lung cancer: Diagnosis and management of lung cancer, 3rd ed: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143(5 Suppl):e211S–e250S. doi: 10.1378/chest.12-2355. [DOI] [PubMed] [Google Scholar]
- 4.Yasufuku K, Nakajima T, Chiyo M, Sekine Y, Shibuya K, Fujisawa T. Endobronchial ultrasonography: Current status and future directions. J Thorac Oncol. 2007;2(10):970 979. doi: 10.1097/JTO.0b013e318153fd8d. [DOI] [PubMed] [Google Scholar]
- 5.Hylton DA, Turner S, Kidane B, et al. The Canada Lymph Node Score for prediction of malignancy in mediastinal lymph nodes during endobronchial ultrasound. J Thorac Cardiovasc Surg. 2020;159(6):2499–2507. doi: 10.1016/j.jtcvs.2019.10.205. [DOI] [PubMed] [Google Scholar]
- 6.Madan K, Madan M, Mittal S, et al. The utility of the ultrasonographic characteristics in differentiating between malignant and tuberculous mediastinal lymphadenopathy during EBUS-TBNA. J Bronchology Interv Pulmonol. 2023;30(1):47–53. doi: 10.1097/LBR.0000000000000850. [DOI] [PubMed] [Google Scholar]
- 7.Prasad KT, Muthu V, Sehgal IS, et al. Endosonographic characteristics of mediastinal lymph nodes for predicting malignancy in high tuberculosis burden settings: A study of 774 subjects. Expert Rev Respir Med. 2022;16(9):1011–1015. doi: 10.1080/17476348.2022.2118717. [DOI] [PubMed] [Google Scholar]
- 8.El-Sherief AH, Lau CT, Wu CC, Drake RL, Abbott GF, Rice TW. International Association for the Study of Lung Cancer (IASLC) lymph node map: Radiologic review with CT illustration. Radiographics. 2014;34(6):1680–1691. doi: 10.1148/rg.346130097. [DOI] [PubMed] [Google Scholar]
- 9.Nakajima T, Anayama T, Shingyoji M, Kimura H, Yoshino I, Yasufuku K. Vascular image patterns of lymph nodes for the prediction of metastatic disease during EBUS-TBNA for mediastinal staging of lung cancer. J Thorac Oncol. 2012;7(6):1009–1014. doi: 10.1097/JTO.0b013e31824cbafa. [DOI] [PubMed] [Google Scholar]
- 10.Bennji SM, Sagar D, Jarnagin L, Dairi MS, Sagar AES. Endobronchial ultrasound staging for lung cancer: What we know now and what we need to know. Curr Pulmonol Rep. 2023;12(4):198–209. doi: 10.1007/s13665-023-00326-9. [DOI] [Google Scholar]
- 11.Fujiwara T, Yasufuku K, Nakajima T, et al. The utility of sonographic features during endobronchial ultrasound-guided transbronchial needle aspiration for lymph node staging in patients with lung cancer: A standard endobronchial ultrasound image classification system. Chest. 2010;138(3):641–647. doi: 10.1378/chest.09-2006. [DOI] [PubMed] [Google Scholar]
- 12.Ozgul MA, Cetinkaya E, Kirkil G, et al. Lymph node characteristics of sarcoidosis with endobronchial ultrasound. Endosc Ultrasound. 2014;3(4):232–237. doi: 10.4103/2303-9027.144541. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.He RX, Hylton DA, Bédard ELR, et al. Clinical validation of the Canada Lymph Node Score for endobronchial ultrasound. Ann Thorac Surg. 2023;115(6):1456–1462. doi: 10.1016/j.athoracsur.2021.11.071. [DOI] [PubMed] [Google Scholar]