Prediction of lung cancer incidence on the low-dose computed tomography arm of the National Lung Screening Trial: A dynamic Bayesian network

Abstract

Introduction Identifying high-risk lung cancer individuals at an early disease stage is the most effective way of improving survival. The landmark National Lung Screening Trial (NLST) demonstrated the utility of low-dose computed tomography (LDCT) imaging to reduce mortality (relative to x-ray screening). As a result of the NLST and other studies, imaging-based lung cancer screening programs are now being implemented. However, LDCT interpretation results in a high number of false positives. A set of dynamic Bayesian networks (DBN) were designed and evaluated to provide insight into how longitudinal data can be used to help inform lung cancer screening decisions. Methods The LDCT arm of the NLST dataset was used to build and explore five DBNs for high-risk individuals. Three of these DBNs were built using a backward construction process, and two using structure learning methods. All models employ demographic, smoking status, cancer history, family lung cancer history, exposure risk factors, comorbidities related to lung cancer, and LDCT screening outcome information. Given the uncertainty arising from lung cancer screening, a cancer state-space model based on lung cancer staging was utilized to characterize the cancer status of an individual over time. The models were evaluated on balanced training and test sets of cancer and non-cancer cases to deal with data imbalance and overfitting. Results Results were comparable to expert decisions. The average area under the curve (AUC) of the receiver operating characteristic (ROC) for the three intervention points of the NLST trial was higher than 0.75 for all models. Evaluation of the models on the complete LDCT arm of the NLST dataset (N = 25, 486) demonstrated satisfactory generalization. Consensus of predictions over similar cases is reported in concordance statistics between the models’ and the physicians’ predictions. The models’ predictive ability with respect to missing data was also evaluated with the sample of cases that missed the second screening exam of the trial (N = 417). The DBNs outperformed comparison models such as logistic regression and naïve Bayes. Conclusion The lung cancer screening DBNs demonstrated high discrimination and predictive power with the majority of cancer and non-cancer cases.

Using Sequential Decision Making to Improve Lung Cancer Screening Performance

“…Figure 2(a) illustrates the lung cancer screening POMDP state space, observations, and potential state transitions. We adopted a state space used in our earlier work [16]. This state space consists of three states defined around the true cancer state of each subject after each screening.…”

Estimating survival benefit of adjuvant therapy based on a Bayesian network prediction model in curatively resected advanced gallbladder adenocarcinoma

“…Bayesian network (BN) is a directed acyclic graph for probabilistic reasoning[14]. Combined with a machine learning algorithm, BNs have great advantages in exploring the unknown probability of variables from the known probability knowledge, which has been applied widely in the field of medicine[15,16]. We have previously applied a BN model and importance measures to identify the significant factors of survival after surgery for patients with GBC[17].…”

Sensor, Signal, and Imaging Informatics

“…Temporal information also plays a role in imaging-based decision making. Petousis et al applied a dynamic Bayesian network (DBN), a model that repeats the static interactions of a conventional Bayesian network over time, to data from the National Lung Screening Trial to improve the prediction of lung cancer using longitudinal data [9].…”