Abstract:Objective: The study objective was to investigate if machine learning algorithms can predict whether a lung nodule is benign, adenocarcinoma, or its preinvasive subtype from computed tomography images alone. Methods: A dataset of chest computed tomography scans containing lung nodules was collected with their pathologic diagnosis from several sources. The dataset was split randomly into training (70%), internal validation (15%), and independent test sets (15%) at the patient level. Two machine learning algorithms were developed, trained, and validated. The first algorithm used the support vector machine model, and the second used deep learning technology: a convolutional neural network. Receiver operating characteristic analysis was used to evaluate the performance of the classification on the test dataset. Results: The support vector machine/convolutional neural network & ndash;based models classified nodules into 6 categories resulting in an area under the curve of 0.59/ 0.65 when differentiating atypical adenomatous hyperplasia versus adenocarcinoma in situ, 0.87/0.86 with minimally invasive adenocarcinoma versus invasive adenocarcinoma, 0.76/0.72 atypical adenomatous hyperplasia + adenocarcinoma in situ versus minimally invasive adenocarcinoma, 0.89/0.87 atypical adenomatous hyperplasia + adenocarcinoma in situ versus minimally invasive adenocarcinoma + invasive adenocarcinoma, and 0.93/0.92 atypical adenomatous hyperplasia + adenocarcinoma in situ + minimally invasive adenocarcinoma versus invasive adenocarcinoma. Classifying benign versus atypical adenomatous hyperplasia + adenocarcinoma in situ + minimally invasive adenocarcinoma versus invasive adenocarcinoma resulted in a micro-average area under the curve of 0.93/0.94 for the support vector machine/convolutional neural network models, respectively. The convolutional neural network-based methods had higher sensitivities than the support vector machine-based methods but lower specificities and accuracies. Conclusions: The machine learning algorithms demonstrated reasonable performance in differentiating benign versus preinvasive versus invasive adenocarcinoma from computed tomography images alone. However, the prediction accuracy varies across its subtypes. This holds the potential for improved diagnostic capabilities with less-invasive means.

An adaptive lung nodule detection algorithm

Adaptive Detection Algorithm for Pulmonary Nodules

Lung Nodules Detection System Using Support Vector Machine Classifier Combined Linear Discriminant Analysis-Based Feature Selection with Rule-Based Feature Pruning

A Lung Nodule Detection Algorithm Based on Local Maximum Gray Segmenting and Improved Mahalanobis Distance Classifying

Computer Aided Lung Nodule Detection on CT Data

An Automatic Pulmonary Nodules Detection Method Using 3d Adaptive Template Matching

Study of automatic detection algorithm for lung nodules

Computer-Aided Lung Nodule Detection Based on Ct Images

An Efficient Model for Lungs Nodule Classification Using Supervised Learning Technique

Automated detection of lung nodules in CT scans: Preliminary results

Autonomous Detection of Solitary Pulmonary Nodules on CT Images for Computer-Aided Diagnosis

Evaluation of an Algorithm for the Segmentation of Lung Nodules in Computerized Tomography Images Based on the Automatic Location of a Threshold

An effective neural network model for lung nodule detection in CT images with optimal fuzzy model

Highly accurate model for prediction of lung nodule malignancy with CT scans

A computer-aided diagnosis system for detection of lung nodules in chest radiographs with an evaluation on a public database

Optimized convolutional neural network for automatic lung nodule detection with a new active contour segmentation

Predicting Benign, Preinvasive, and Invasive Lung Nodules on Computed Tomography Scans Using Machine Learning

Computed Tomography Imaging Features of Lung Cancer under Artificial Intelligence Algorithm and Its Correlation with Pathology

Lung Nodule Classification on CT Images Using Deep Convolutional Neural Network Based on Geometric Feature Extraction

Detection and Attention: Diagnosing Pulmonary Lung Cancer from CT by Imitating Physicians

Deep Transfer Convolutional Neural Network and Extreme Learning Machine for lung nodule diagnosis on CT images