Search

Modern medicine is supported by increasingly sophisticated algorithms. In diagnostics or screening, statistical models are commonly used to assess the risk of disease development, the severity of its course, and expected treatment outcome. The growing availability of very detailed data and increased interest in personalized medicine are leading to the development of effective but complex machine learning models. For these models to be trusted, their predictions must be understandable to both the physician and the patient, hence the growing interest in the area of Explainable Artificial Intelligence (XAI). In this paper, we present selected methods from the XAI field in the example of models applied to assess lung cancer risk in lung cancer screening through low-dose computed tomography. The use of these techniques provides a better understanding of the similarities and differences between three commonly used models in lung cancer screening, i.e., BACH, PLCOm2012, and LCART. For the presentation of the results, we used data from the Domestic Lung Cancer Database. The XAI techniques help to better understand (1) which variables are most important in which model, (2) how they are transformed into model predictions, and facilitate (3) the explanation of model predictions for a particular screenee.

Objective: Mediastinoscopy as diagnostic procedure for evaluation of mediastinum in patients with non-small-cell lung cancer has long been considered the reference standard. However, less invasive method has occurred. Endobronchial ultrasound–guided transbronchial needle aspiration came into widespread use and has resulted in controversy as to whether it is a good replacement for mediastinoscopy. We chose to demonstrate the usefulness of endobronchial ultrasound–guided transbronchial needle aspiration in evaluating the mediastinum in patients with non-small-cell lung cancer. Material and methods: Over a 48-month period, 1841 patients underwent endobronchial ultrasound–guided transbronchial needle aspiration at our healthcare centre. In all patients, 2964 biopsies from the lymph node group N2 and 783 from group N1 were taken. The mean short axis of the lymph nodes biopsied was 2.0 (range: 0.6–2.6). The mean number of lymph node stations biopsied per patient was 2.6. Patients with a negative result of endobronchial ultrasound–guided transbronchial needle aspiration underwent mediastinoscopy. All patients with a negative result in endobronchial ultrasound–guided transbronchial needle aspiration and mediastinoscopy underwent surgical resection with lymph node sampling. Results: The metastases to lymph nodes N2/N3 and N1 were found in 1111 (60.3%) and 199 (9.3%), respectively. Mediastinoscopy was performed in 730 patients with a positive result in 83 (11.4%) patients. In the group of operated patients, metastatic N1 disease was found in 264 (14.1%). In the group of the operated patients, mediastinal involvement of disease (N2) was found in 30 patients (4.5%). The sensitivity, negative predictive value and diagnostic accuracy for hilar lymph node staging for endobronchial ultrasound–guided transbronchial needle aspiration were 57%, 96% and 96%, respectively. The sensitivity, negative predictive value and diagnostic accuracy per patient for mediastinal lymph node staging for endobronchial ultrasound–guided transbronchial needle aspiration and mediastinoscopy were 91%, 85%, 93% and 73%, 95.5%, 97%, respectively. The specificity and positive predictive value of both tests were 100%. Conclusion: The clinical usefulness of endobronchial ultrasound–guided transbronchial needle aspiration is undeniable according to diagnostic performance data. Endobronchial ultrasound–guided transbronchial needle aspiration should be considered complementary to mediastinoscopy in the evaluation of patients with radiographically abnormal mediastinum.

Since the introduction of anatomic lung resection by video-assisted thoracoscopic surgery (VATS) 20 years ago, VATS has experienced major advances in both equipment and technique, introducing a technical challenge in the surgical treatment of both benign and malignant lung disease. The demonstrated safety, decreased morbidity, and equivalent efficacy of this minimally invasive technique have led to the acceptance of VATS as a standard surgical modality for early-stage lung cancer and increasing application to more advanced disease. Formerly there was much debate about the feasibility of the technique in cancer surgery and proper lymph node handling. Although there is a lack of proper randomized studies, it is now generally accepted that the outcome of a VATS procedure is at least not inferior to a resection via a traditional thoracotomy.