Dataset Information

Deep learning with convex probe endobronchial ultrasound multimodal imaging: A validated tool for automated intrathoracic lymph nodes diagnosis.

ABSTRACT:

Background and objectives

Along with the rapid improvement of imaging technology, convex probe endobronchial ultrasound (CP-EBUS) sonographic features play an increasingly important role in the diagnosis of intrathoracic lymph nodes (LNs). Conventional qualitative and quantitative methods for EBUS multimodal imaging are time-consuming and rely heavily on the experience of endoscopists. With the development of deep-learning (DL) models, there is great promise in the diagnostic field of medical imaging.

Materials and methods

We developed DL models to retrospectively analyze CP-EBUS images of 294 LNs from 267 patients collected between July 2018 and May 2019. The DL models were trained on 245 LNs to differentiate benign and malignant LNs using both unimodal and multimodal CP-EBUS images and independently evaluated on the remaining 49 LNs to validate their diagnostic efficiency. The human comparator group consisting of three experts and three trainees reviewed the same test set as the DL models.

Results

The multimodal DL framework achieves an accuracy of 88.57% (95% confidence interval [CI] [86.91%-90.24%]) and area under the curve (AUC) of 0.9547 (95% CI [0.9451-0.9643]) using the three modes of CP-EBUS imaging in comparison to the accuracy of 80.82% (95% CI [77.42%-84.21%]) and AUC of 0.8696 (95% CI [0.8369-0.9023]) by experts. Statistical comparison of their average receiver operating curves shows a statistically significant difference (P < 0.001). Moreover, the multimodal DL framework is more consistent than experts (kappa values 0.7605 vs. 0.5800).

Conclusions

The DL models based on CP-EBUS imaging demonstrated an accurate automated tool for diagnosis of the intrathoracic LNs with higher diagnostic efficiency and consistency compared with experts.

SUBMITTER: Li J

PROVIDER: S-EPMC8544010 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Background and Objectives: Endobronchial ultrasound elastography is a new technique for describing the stiffness of tissue during endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA). The aims of this study were to investigate the diagnostic value of Endobronchial ultrasound (EBUS) elastography for distinguishing the difference between benign and malignant lymph nodes among mediastinal and hilar lymph node. Materials and Methods: From June 2015 to August 2015, 47 patients confirmed of mediastinal and hilar lymph node enlargement through examination of Computed tomography (CT) were enrolled, and a total of 78 lymph nodes were evaluated by endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA). EBUS-guided elastography of lymph nodes was performed prior to EBUS-TBNA. A convex probe EBUS was used with a new EBUS processor to assess elastographic patterns that were classified based on color distribution as follows: Type 1, predominantly non-blue (green, yellow and red); Type 2, part blue, part non-blue (green, yellow and red); Type 3, predominantly blue. Pathological determination of malignant or benign lymph nodes was used as the gold standard for this study. The elastographic patterns were compared with the final pathologic results from EBUS-TBNA. Results: On pathological evaluation of the lymph nodes, 45 were benign and 33 were malignant. The lymph nodes that were classified as Type 1 on endobronchial ultrasound elastography were benign in 26/27 (96.3%) and malignant in 1/27 (3.7%); for Type 2 lymph nodes, 15/20 (75.0%) were benign and 5/20 (25.0%) were malignant; Type 3 lymph nodes were benign in 4/31 (12.9%) and malignant in 27/31 (87.1%). In classifying Type 1 as 'benign' and Type 3 as 'malignant,' the sensitivity, specificity, positive predictive value, negative predictive value and diagnostic accuracy rates were 96.43%, 86.67%, 87.10%, 96.30%, 91.38%, respectively. Conclusion: EBUS elastography of mediastinal and hilar lymph nodes is a noninvasive technique that can be performed reliably and may be helpful in the prediction of benign and malignant lymph nodes among mediastinal and hilar lymph node during EBUS-TBNA.

Project description:ObjectiveAutomated breast ultrasound (ABUS) is a relevant imaging technique for early breast cancer diagnosis and is increasingly being used as a supplementary tool for mammography. This study compared the performance of ABUS and handheld ultrasound (HHUS) in detecting and characterizing the axillary lymph nodes (LNs) in patients with breast cancer.Materials and methodsWe retrospectively reviewed the medical records of women with recently diagnosed early breast cancer (≤ T2) who underwent both ABUS and HHUS examinations for axilla (September 2017-May 2018). ABUS and HHUS findings were compared using pathological outcomes as reference standards. Diagnostic performance in predicting any axillary LN metastasis and heavy nodal-burden metastases (i.e., ≥ 3 LNs) was evaluated. The ABUS-HHUS agreement for visibility and US findings was calculated.ResultsThe study included 377 women (53.1 ± 11.1 years). Among 385 breast cancers in 377 patients, 101 had axillary LN metastases and 30 had heavy nodal burden metastases. ABUS identified benign-looking or suspicious axillary LNs (average, 1.4 ± 0.8) in 246 axillae (63.9%, 246/385). According to the per-breast analysis, the sensitivity, specificity, positive and negative predictive values, and accuracy of ABUS in predicting axillary LN metastases were 43.6% (44/101), 95.1% (270/284), 75.9% (44/58), 82.6% (270/327), and 81.6% (314/385), respectively. The corresponding results for HHUS were 41.6% (42/101), 95.1% (270/284), 75.0% (42/56), 82.1% (270/329), and 81.0% (312/385), respectively, which were not significantly different from those of ABUS (P ≥ 0.53). The performance results for heavy nodal-burden metastases were 70.0% (21/30), 89.6% (318/355), 36.2% (21/58), 97.3% (318/327), and 88.1% (339/385), respectively, for ABUS and 66.7% (20/30), 89.9% (319/355), 35.7% (20/56), 97.0% (319/329), and 88.1% (339/385), respectively, for HHUS, also not showing significant difference (P ≥ 0.57). The ABUS-HHUS agreement was 95.9% (236/246; Cohen's kappa = 0.883).ConclusionAlthough ABUS showed limited sensitivity in diagnosing axillary LN metastasis in early breast cancer, it was still useful as the performance was comparable to that of HHUS.

Project description:ObjectiveEndobronchial ultrasound-guided transbronchial needle aspiration is of diagnostic value in hilar/mediastinal (N1/N2) lymph node staging. We assessed the utility of endobronchial ultrasound-guided transbronchial needle aspiration in lung cancer patients with N1/N2 lymph nodes detected on (18)F-fluorodeoxyglucose positron emission tomography/computed tomography.MethodsFifty lung cancer patients with N1/N2 disease on (18)F-fluorodeoxyglucose positron emission tomography/computed tomography underwent endobronchial ultrasound-guided transbronchial needle aspiration for pathological lymph nodes between November 2012 and April 2015. The diagnostic performance of endobronchial ultrasound-guided transbronchial needle aspiration, lymph node site and size, number of needle passes and complications were evaluated retrospectively from patients' medical records. Malignancy was defined as a maximum standardized uptake value (SUVmax) >2.5.ResultsThe median longest diameter of the 61 lymph nodes (29 subcarinal, 21 right lower paratracheal, 6 left lower paratracheal, 4 right hilar and 1 upper paratracheal) was 23.4 mm (range: 10.4-45.7); the median number of needle passes was 2 (range: 1-5). There were no severe complications. A definitive diagnosis was made by endobronchial ultrasound-guided transbronchial needle aspiration in 39 patients (31 adenocarcinomas, 3 small-cell carcinomas, 2 squamous-cell carcinomas, 3 large-cell neuroendocrine carcinomas). In the remaining 11 patients, the diagnosis was indefinite: insufficient endobronchial ultrasound-guided transbronchial needle aspiration material was collected in two patients and non-specific lymphadenopathy was confirmed by endobronchial ultrasound-guided transbronchial needle aspiration or thoracotomy in the other nine patients. The mean lymph node SUVmax was 7.09 (range: 2.90-26.9) and was significantly higher in true-positive than in false-positive nodes (P < 0.05, t-test). Non-specific lymphadenopathy was diagnosed by expert visual interpretation of (18)F-fluorodeoxyglucose positron emission tomography/computed tomography images in five of the nine patients.ConclusionEndobronchial ultrasound-guided transbronchial needle aspiration accurately diagnoses N1/N2 disease detected on (18)F-fluorodeoxyglucose positron emission tomography/computed tomography.

Project description:BackgroundCurrently there is no consensus on ideal teaching method to train novice trainees in EBUS. Simulation-based procedure training allows direct observation of trainees in a controlled environment without compromising patient safety.ObjectiveWe wanted to develop a comprehensive assessment of endobronchial ultrasound (EBUS) performance of pulmonary fellows and assess the impact of a multimodal simulation-based curriculum for EBUS-guided transbronchial needle aspiration.MethodsPretest assessment of 11 novice pulmonary fellows was performed using a three-part assessment tool, measuring EBUS-related knowledge, self-confidence, and procedural skills. Knowledge was assessed by 20 multiple-choice questions. Self-confidence was measured using the previously validated EBUS-Subjective Assessment Tool. Procedural skills assessment was performed on Simbionix BRONCH Express simulator and was modeled on a previously validated EBUS-Skills and Task Assessment Tool (EBUS-STAT), to create a modified EBUS-STAT based on internal faculty input via the Delphi method. After baseline testing, fellows participated in a structured multimodal curriculum, which included simulator training, small-group didactics, and interactive problem-based learning sessions, followed by individual debriefing sessions. Posttest assessment using the same three-part assessment tool was performed after 3 months, and the results were compared to study the impact of the new curriculum.ResultsThe mean knowledge score improved significantly from baseline to posttest (52.7% vs. 67.7%; P = 0.002). The mean EBUS-Subjective Assessment Tool confidence scores (maximum score, 50) improved significantly from baseline to posttest (26 ± 7.6 vs. 35.2 ± 6.3 points; P < 0.001). The mean modified EBUS-STAT (maximum score, 105) improved significantly from baseline to posttest (44.8 ± 10.6 [42.7%] vs. 65.3 ± 11.4 [62.2%]; P < 0.001). There was a positive correlation (r = 0.81) between the experience of the test participants and the modified EBUS-STAT scores.ConclusionThis study suggests a multimodal simulation-based curriculum can significantly improve EBUS-guided transbronchial needle aspiration-related knowledge, self-confidence, and procedural skills among novice pulmonary fellows. A validation study is needed to determine if skills attained via a simulator can be replicated in a clinical setting.