Improved pulmonary nodule classification utilizing quantitative lung parenchyma features

Peri- and intra-nodular radiomic features based on 18F-FDG PET/CT to distinguish lung adenocarcinomas from pulmonary granulomas

Objective

To compare the effectiveness of radiomic features based on 18F-FDG PET/CT images within (intranodular) and around (perinodular) lung nodules/masses in distinguishing between lung adenocarcinoma and pulmonary granulomas.

Methods

For this retrospective study, 18F-FDG PET/CT images were collected for 228 patients. Patients diagnosed with lung adenocarcinoma (n = 156) or granulomas (n = 72) were randomly assigned to a training (n = 159) and validation (n = 69) groups. The volume of interest (VOI) of intranodular, perinodular (1-5 voxels, termed Lesion_margin1 to Lesion_margin5) and total area (intra- plus perinodular region, termed Lesion_total1 to Lesion_total5) on PET/CT images were delineated using PETtumor and Marge tool of segmentation editor. A total of 1,037 radiomic features were extracted separately from PET and CT images, and the optimal features were selected to develop radiomic models. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC).

Results

Good and acceptable performance was, respectively, observed in the training (AUC = 0.868, p < 0.001) and validation (AUC = 0.715, p = 0.004) sets for the intranodular radiomic model. Among the perinodular models, the Lesion_margin2 model demonstrated the highest AUC in both sets (0.883 and 0.616, p < 0.001 and p = 0.122). Similarly, in terms of total models, Lesion_total2 model was found to outperform others in the training (AUC = 0.879, p < 0.001) and validation (AUC = 0.742, p = 0.001) sets, slightly surpassing the intranodular model.

Conclusion

When intra- and perinodular radiomic features extracted from the immediate vicinity of the nodule/mass up to 2 voxels distance on 18F-FDG PET/CT imaging are combined, improved differential diagnostic performance in distinguishing between lung adenocarcinomas and granulomas is achieved compared to the intra- and perinodular radiomic features alone.

A proposed methodology for detecting the malignant potential of pulmonary nodules in sarcoma using computed tomographic imaging and artificial intelligence-based models

The presence of lung metastases in patients with primary malignancies is an important criterion for treatment management and prognostication. Computed tomography (CT) of the chest is the preferred method to detect lung metastasis. However, CT has limited efficacy in differentiating metastatic nodules from benign nodules (e.g., granulomas due to tuberculosis) especially at early stages (<5 mm). There is also a significant subjectivity associated in making this distinction, leading to frequent CT follow-ups and additional radiation exposure along with financial and emotional burden to the patients and family. Even 18F-fluoro-deoxyglucose positron emission technology-computed tomography (18F-FDG PET-CT) is not always confirmatory for this clinical problem. While pathological biopsy is the gold standard to demonstrate malignancy, invasive sampling of small lung nodules is often not clinically feasible. Currently, there is no non-invasive imaging technique that can reliably characterize lung metastases. The lung is one of the favored sites of metastasis in sarcomas. Hence, patients with sarcomas, especially from tuberculosis prevalent developing countries, can provide an ideal platform to develop a model to differentiate lung metastases from benign nodules. To overcome the lack of optimal specificity of CT scan in detecting pulmonary metastasis, a novel artificial intelligence (AI)-based protocol is proposed utilizing a combination of radiological and clinical biomarkers to identify lung nodules and characterize it as benign or metastasis. This protocol includes a retrospective cohort of nearly 2,000-2,250 sample nodules (from at least 450 patients) for training and testing and an ambispective cohort of nearly 500 nodules (from 100 patients; 50 patients each from the retrospective and prospective cohort) for validation. Ground-truth annotation of lung nodules will be performed using an in-house-built segmentation tool. Ground-truth labeling of lung nodules (metastatic/benign) will be performed based on histopathological results or baseline and/or follow-up radiological findings along with clinical outcome of the patient. Optimal methods for data handling and statistical analysis are included to develop a robust protocol for early detection and classification of pulmonary metastasis at baseline and at follow-up and identification of associated potential clinical and radiological markers.

Radiomic biomarkers from chest computed tomography are assistive in immunotherapy response prediction for non-small cell lung cancer

Background

Immunotherapies, such as programmed death 1/programmed death ligand 1 (PD-1/PD-L1) antibodies have been shown to improve overall and progression-free survival (PFS) in patients with locally advanced or metastatic non-small cell lung cancer (NSCLC). However, not all patients derive a meaningful clinical benefit. Additionally, patients receiving anti-PD-1/PD-L1 therapy can experience immune-related adverse events (irAEs). Clinically significant irAEs may require temporary pause or discontinuation of treatment. Having a tool to identify patients who may not benefit and/or are at risk for developing severe irAEs from immunotherapy will aid in an informed decision-making process for the patients and their physicians.

Methods

Computed tomography (CT) scans and clinical data were retrospectively collected for this study to develop three prediction models using (I) radiomic features, (II) clinical features, and (III) radiomic and clinical features combined. Each subject had 6 clinical features and 849 radiomic features extracted. Selected features were run through an artificial neural network (NN) trained on 70% of the cohort, maintaining the case and control ratio. The NN was assessed by calculating the area-under-the-receiver-operating-characteristic curve (AUC-ROC), area-under-the-precision-recall curve (AUC-PR), sensitivity, and specificity.

Results

A cohort of 132 subjects, of which 43 (33%) had a PFS ≤90 days and 89 (67%) of which had a PFS >90 days was used to develop the prediction models. The radiomic model was able to predict progression-free survival with a training AUC-ROC of 87% and testing AUC-ROC, sensitivity, and specificity of 83%, 75%, and 81%, respectively. In this cohort, the clinical and radiomic combined features did add a slight increase in the specificity (85%) but with a decrease in sensitivity (75%) and AUC-ROC (81%).

Conclusions

Whole lung segmentation and feature extraction can identify those that would see a benefit from anti-PD-1/PD-L1 therapy.

Imaging genomics: data fusion in uncovering disease heritability

Sequencing of the human genome in the early 2000s enabled probing of the genetic basis of disease on a scale previously unimaginable. Now, two decades later, after interrogating millions of markers in thousands of individuals, a significant portion of disease heritability still remains hidden. Recent efforts to unravel this 'missing heritability' have focused on garnering new insight from merging different data types, including medical imaging. Imaging offers promising intermediate phenotypes to bridge the gap between genetic variation and disease pathology. In this review we outline this fusion and provide examples of imaging genomics in a range of diseases, from oncology to cardiovascular and neurodegenerative disease. Finally, we discuss how ongoing revolutions in data science and sharing are primed to advance the field.

The value of artificial intelligence in the diagnosis of lung cancer: A systematic review and meta-analysis

Lung cancer is a common malignant tumor disease with high clinical disability and death rates. Currently, lung cancer diagnosis mainly relies on manual pathology section analysis, but the low efficiency and subjective nature of manual film reading can lead to certain misdiagnoses and omissions. With the continuous development of science and technology, artificial intelligence (AI) has been gradually applied to imaging diagnosis. Although there are reports on AI-assisted lung cancer diagnosis, there are still problems such as small sample size and untimely data updates. Therefore, in this study, a large amount of recent data was included, and meta-analysis was used to evaluate the value of AI for lung cancer diagnosis. With the help of STATA16.0, the value of AI-assisted lung cancer diagnosis was assessed by specificity, sensitivity, negative likelihood ratio, positive likelihood ratio, diagnostic ratio, and plotting the working characteristic curves of subjects. Meta-regression and subgroup analysis were used to investigate the value of AI-assisted lung cancer diagnosis. The results of the meta-analysis showed that the combined sensitivity of the AI-aided diagnosis system for lung cancer diagnosis was 0.87 [95% CI (0.82, 0.90)], specificity was 0.87 [95% CI (0.82, 0.91)] (CI stands for confidence interval.), the missed diagnosis rate was 13%, the misdiagnosis rate was 13%, the positive likelihood ratio was 6.5 [95% CI (4.6, 9.3)], the negative likelihood ratio was 0.15 [95% CI (0.11, 0.21)], a diagnostic ratio of 43 [95% CI (24, 76)] and a sum of area under the combined subject operating characteristic (SROC) curve of 0.93 [95% CI (0.91, 0.95)]. Based on the results, the AI-assisted diagnostic system for CT (Computerized Tomography), imaging has considerable diagnostic accuracy for lung cancer diagnosis, which is of significant value for lung cancer diagnosis and has greater feasibility of realizing the extension application in the field of clinical diagnosis.

Self-Supervised Transfer Learning Based on Domain Adaptation for Benign-Malignant Lung Nodule Classification on Thoracic CT

The spatial heterogeneity is an important indicator of the malignancy of lung nodules in lung cancer diagnosis. Compared with 2D nodule CT images, the 3D volumes with entire nodule objects hold richer discriminative information. However, for deep learning methods driven by massive data, effectively capturing the 3D discriminative features of nodules in limited labeled samples is a challenging task. Different from previous models that proposed transfer learning models in a 2D pattern or learning from scratch 3D models, we develop a self-supervised transfer learning based on domain adaptation (SSTL-DA) 3D CNN framework for benign-malignant lung nodule classification. At first, a data pre-processing strategy termed adaptive slice selection (ASS) is developed to eliminate the redundant noise of the input samples with lung nodules. Then, the self-supervised learning network is constructed to learn robust image representation from CT images. Finally, a transfer learning method based on domain adaptation is designed to obtain discriminant features for classification. The proposed SSTL-DA method has been assessed on the LIDC-IDRI benchmark dataset, and it obtains an accuracy of 91.07% and an AUC of 95.84%. These results demonstrate that the SSTL-DA model achieves quite a competitive classification performance compared with some state-of-the-art approaches.

Computed Tomography Features of Lung Structure Have Utility for Differentiating Malignant and Benign Pulmonary Nodules

BACKGROUND

Chronic obstructive pulmonary disease (COPD) is a known comorbidity for lung cancer independent of smoking history. Quantitative computed tomography (qCT) imaging features related to COPD have shown promise in the assessment of lung cancer risk. We hypothesize that qCT features from the lung, lobe, and airway tree related to the location of the pulmonary nodule can be used to provide informative malignancy risk assessment.

METHODS

A total of 183 qCT features were extracted from 278 individuals with a solitary pulmonary nodule of known diagnosis (71 malignant, 207 benign). These included histogram and airway characteristics of the lungs, lobe, and segmental paths. Performances of the least absolute shrinkage and selection operator (LASSO) regression analysis and an ensemble of neural networks (ENN) were compared for feature set selection and classification on a testing cohort of 49 additional individuals (15 malignant, 34 benign).

RESULTS

The LASSO and ENN methods produced different feature sets for classification with LASSO selecting fewer qCT features (7) than the ENN (17). The LASSO model with the highest performing training area under the curve (AUC) (0.80) incorporated automatically extracted features and reader-measured nodule diameter with a testing AUC of 0.62. The ENN model with the highest performing AUC (0.77) also incorporated qCT and reader diameter but maintained higher testing performance AUC (0.79).

CONCLUSIONS

Automatically extracted qCT imaging features of the lung can be informative of the differentiation between individuals with malignant pulmonary nodules and those with benign pulmonary nodules, without requiring nodule segmentation and analysis.

Classification of lung nodules based on the DCA-Xception network

BACKGROUND

Developing deep learning networks to classify between benign and malignant lung nodules usually requires many samples. Due to the precious nature of medical samples, it is difficult to obtain many samples.

OBJECTIVE

To investigate and test a DCA-Xception network combined with a new data enhancement method to improve performance of lung nodule classification.

METHODS

First, the Wasserstein Generative Adversarial Network (WGAN) with conditions and five data enhancement methods such as flipping, rotating, and adding Gaussian noise are used to extend the samples to solve the problems of unbalanced sample classification and the insufficient samples. Then, a DCA-Xception network is designed to classify lung nodules. Using this network, information around the target is obtained by introducing an adaptive dual-channel feature extraction module, and the network learns features more accurately by introducing a convolutional attention module. The network is trained and validated using 274 lung nodules (154 benign and 120 malignant) and tested using 52 lung nodules (23 benign and 29 malignant).

RESULTS

The experiments show that the network has an accuracy of 83.46% and an AUC of 0.929. The features extracted using this network achieve an accuracy of 85.24% on the K-nearest neighbor and random forest classifiers.

CONCLUSION

This study demonstrates that the DCA-Xception network yields higher performance in classification of lung nodules than the performance using the classical classification networks as well as pre-trained networks.

The Role of Radiomics in Lung Cancer: From Screening to Treatment and Follow-Up

PURPOSE

Lung cancer represents the first cause of cancer-related death in the world. Radiomics studies arise rapidly in this late decade. The aim of this review is to identify important recent publications to be synthesized into a comprehensive review of the current status of radiomics in lung cancer at each step of the patients' care.

METHODS

A literature review was conducted using PubMed/Medline for search of relevant peer-reviewed publications from January 2012 to June 2020.

RESULTS

We identified several studies at each point of patient's care: detection and classification of lung nodules (n=16), determination of histology and genomic (n=10) and finally treatment outcomes predictions (=23). We reported the methodology of those studies and their results and discuss the limitations and the progress to be made for clinical routine applications.

CONCLUSION

Promising perspectives arise from machine learning applications and radiomics based models in lung cancers, yet further data are necessary for their implementation in daily care. Multicentric collaboration and attention to quality and reproductivity of radiomics studies should be further consider.

The Effects of Perinodular Features on Solid Lung Nodule Classification

Lung cancer is the most lethal malignant neoplasm worldwide, with an annual estimated rate of 1.8 million deaths. Computed tomography has been widely used to diagnose and detect lung cancer, but its diagnosis remains an intricate and challenging work, even for experienced radiologists. Computer-aided diagnosis tools and radiomics tools have provided support to the radiologist's decision, acting as a second opinion. The main focus of these tools has been to analyze the intranodular zone; nevertheless, recent works indicate that the interaction between the nodule and its surroundings (perinodular zone) could be relevant to the diagnosis process. However, only a few works have investigated the importance of specific attributes of the perinodular zone and have shown how important they are in the classification of lung nodules. In this context, the purpose of this work is to evaluate the impact of using the perinodular zone on the characterization of lung lesions. Motivated by reproducible research, we used a large public dataset of solid lung nodule images and extracted fine-tuned radiomic attributes from the perinodular and intranodular zones. Our best-evaluated model obtained an average AUC of 0.916, an accuracy of 84.26%, a sensitivity of 84.45%, and specificity of 83.84%. The combination of attributes from the perinodular and intranodular zones in the image characterization resulted in an improvement in all the metrics analyzed when compared to intranodular-only characterization. Therefore, our results highlighted the importance of using the perinodular zone in the solid pulmonary nodules classification process.

Contribution of a Blood-Based Protein Biomarker Panel to the Classification of Indeterminate Pulmonary Nodules

RATIONALE

The workup and longitudinal monitoring for subjects presenting with pulmonary nodules is a pressing clinical problem. A blood-based biomarker panel potentially has utility for identifying subjects at higher risk for harboring a malignant nodule for whom additional workup would be indicated or subjects at reduced risk for whom imaging-based follow-up would be indicated.

OBJECTIVES

To assess whether a previously described four-protein biomarker panel, reported to improve assessment of lung cancer risk compared with a smoking-based lung cancer risk model, can provide discrimination between benign and malignant indeterminate pulmonary nodules.

METHODS

A previously validated multiplex enzyme-linked immunoassay was performed on matched case and control samples from each cohort.

MEASUREMENTS

The biomarker panel was tested in two case-control cohorts of patients presenting with indeterminate pulmonary nodules at the University of Pittsburgh Medical Center and the University of Texas Southwestern.

MAIN RESULTS

In both cohorts, the biomarker panel resulted in improved prediction of lung cancer risk over a model on the basis of nodule size alone. Of particular note, the addition of the marker panel to nodule size greatly improved sensitivity at a high specificity in both cohorts.

CONCLUSIONS

A four-marker biomarker panel, previously validated to improve lung cancer risk prediction, was found to also have utility in distinguishing benign from malignant indeterminate pulmonary nodules. Its performance in improving sensitivity at a high specificity indicates potential utility of the marker panel in assessing likelihood of malignancy in otherwise indeterminate nodules.

Texture analysis of iodine maps and conventional images for k-nearest neighbor classification of benign and metastatic lung nodules

BACKGROUND

The purpose of this study was to analyze if the use of texture analysis on spectral detector CT (SDCT)-derived iodine maps (IM) in addition to conventional images (CI) improves lung nodule differentiation, when being applied to a k-nearest neighbor (KNN) classifier.

METHODS

183 cancer patients who underwent contrast-enhanced, venous phase SDCT of the chest were included: 85 patients with 146 benign lung nodules (BLN) confirmed by either prior/follow-up CT or histopathology and 98 patients with 425 lung metastases (LM) verified by histopathology, 18F-FDG-PET-CT or unequivocal change during treatment. Semi-automatic 3D segmentation of BLN/LM was performed, and volumetric HU attenuation and iodine concentration were acquired. For conventional images and iodine maps, average, standard deviation, entropy, kurtosis, mean of the positive pixels (MPP), skewness, uniformity and uniformity of the positive pixels (UPP) within the volumes of interests were calculated. All acquired parameters were transferred to a KNN classifier.

RESULTS

Differentiation between BLN and LM was most accurate, when using all CI-derived features combined with the most significant IM-derived feature, entropy (Accuracy:0.87; F1/Dice:0.92). However, differentiation accuracy based on the 4 most powerful CI-derived features performed only slightly inferior (Accuracy:0.84; F1/Dice:0.89, p=0.125). Mono-parametric lung nodule differentiation based on either feature alone (i.e. attenuation or iodine concentration) was poor (AUC=0.65, 0.58, respectively).

CONCLUSIONS

First-order texture feature analysis of contrast-enhanced staging SDCT scans of the chest yield accurate differentiation between benign and metastatic lung nodules. In our study cohort, the most powerful iodine map-derived feature slightly, yet insignificantly increased classification accuracy  compared to classification based on conventional image features only.

Prediction of tumor doubling time of lung adenocarcinoma using radiomic margin characteristics

Background

Because shape or irregularity along the tumor perimeter can result from interactions between the tumor and the surrounding parenchyma, there could be a difference in tumor growth rate according to tumor margin or shape. However, no attempt has been made to evaluate the correlation between margin or shape features and tumor growth.

Methods

We evaluated 52 lung adenocarcinoma (ADC) patients who had at least two computed tomographic (CT) examinations before curative resection. Volume‐based doubling times (DTs) were calculated based on CT scans, and patients were divided into two groups according to the growth pattern (GP) of their ADCs (gradually growing tumors [GP I] vs. growing tumors with a temporary decrease in DT [GP II]). CT radiomic features reflecting margin characteristics were extracted, and radiomic features reflective of tumor DT were selected.

Results

Among the 52 patients, 41 (78.8%) were assigned to GP I and 11 (21.2%) to GP II. Of the 94 radiomic features extracted, eccentricity, surface‐to‐volume ratio, LoG uniformity (σ = 3.5), and LoG skewness (σ = 0.5) were ultimately selected for tumor DT prediction. Selected radiomic features in GP I were surface‐to‐volume ratio, contrast, LoG uniformity (σ = 3.5), and LoG skewness (σ = 0.5), similar to those for total subjects, whereas the radiomic features in GP II were solidity, energy, and busyness.

Conclusions

This study demonstrated the potential of margin‐related radiomic features to predict tumor DT in lung ADCs.

Key points

Significant findings of the study

We found a relationship between margin‐related radiomic features and tumor doubling time.

What this study adds

Margin‐related radiomic features can potentially be used as noninvasive biomarkers to predict tumor doubling time in lung adenocarcinoma and inform treatment strategies.

Low-Dose CT Screening for Lung Cancer: Evidence from 2 Decades of Study

Lung cancer remains the overwhelmingly greatest cause of cancer death in the United States, accounting for more annual deaths than breast, prostate, and colon cancer combined. Accumulated evidence since the mid to late 1990s, however, indicates that low-dose CT screening of high-risk patients enables detection of lung cancer at an early stage and can reduce the risk of dying from lung cancer. CT screening is now a recommended clinical service in the United States, subject to guidelines and reimbursement requirements intended to standardize practice and optimize the balance of benefits and risks. In this review, the evidence on the effectiveness of CT screening will be summarized and the current guidelines and standards will be described in the context of knowledge gained from lung cancer screening studies. In addition, an overview of the potential advances that may improve CT screening will be presented, and the need to better understand the performance in clinical practice outside of the research trial setting will be discussed. © RSNA, 2020.

Machine Learning and Feature Selection Methods for Disease Classification With Application to Lung Cancer Screening Image Data

As awareness of the habits and risks associated with lung cancer has increased, so has the interest in promoting and improving upon lung cancer screening procedures. Recent research demonstrates the benefits of lung cancer screening; the National Lung Screening Trial (NLST) found as its primary result that preventative screening significantly decreases the death rate for patients battling lung cancer. However, it was also noted that the false positive rate was very high (>94%).In this work, we investigated the ability of various machine learning classifiers to accurately predict lung cancer nodule status while also considering the associated false positive rate. We utilized 416 quantitative imaging biomarkers taken from CT scans of lung nodules from 200 patients, where the nodules had been verified as cancerous or benign. These imaging biomarkers were created from both nodule and parenchymal tissue. A variety of linear, nonlinear, and ensemble predictive classifying models, along with several feature selection methods, were used to classify the binary outcome of malignant or benign status. Elastic net and support vector machine, combined with either a linear combination or correlation feature selection method, were some of the best-performing classifiers (average cross-validation AUC near 0.72 for these models), while random forest and bagged trees were the worst performing classifiers (AUC near 0.60). For the best performing models, the false positive rate was near 30%, notably lower than that reported in the NLST.The use of radiomic biomarkers with machine learning methods are a promising diagnostic tool for tumor classification. The have the potential to provide good classification and simultaneously reduce the false positive rate.

Differentiation of non-small cell lung cancer and histoplasmosis pulmonary nodules: insights from radiomics model performance compared with clinician observers

Background

Histoplasmosis pulmonary nodules often present in computed tomography (CT) imaging with characteristics suspicious for lung cancer. This presents a work-up decision issue for clinicians in regions where histoplasmosis is an endemic fungal infection, when a nodule suspicious for lung cancer is detected. We hypothesize the application of radiomic features extracted from pulmonary nodules and perinodular parenchyma could accurately distinguish between suspicious histoplasmosis lung nodules and non-small cell lung cancer (NSCLC).

Methods

A retrospective clinical cohort of pulmonary nodules with a confirmed diagnosis of histoplasmosis or NSCLC was collected from the University of Iowa Hospitals and Clincs. Radiomic features were extracted describing characteristics of the nodule and perinodular parenchyma regions and used to build a machine learning tool. These cases were assessed by four expert clinicians who gave a blinded risk prediction for NSCLC. Tool and observer performance were assessed by calculating the area under the curve for the receiver operating characteristic (AUC-ROC) and interclass correlation coefficient (ICC).

Results

A cohort of 71 subjects with confirmed histopathology (40 NSCLC, 31 histoplasmosis) were case-matched based on age, sex, and smoking history. Superior performance (AUC-ROC =0.89) was demonstrated using leave-one-subject out validation in the tool that incorporated radiomics from the nodule and perinodular parenchyma region extended to 100% nodule diameter. Observers had perfect intra-repeatability (ICC =1.0) and demonstrated fair inter-reader variability (ICC =0.52).

Conclusions

Radiomics have potential utility in the challenging task of differentiation between lung cancer and histoplasmosis. Expert clinician readers have high intra-repeatability but demonstrated inter-reader variability which could provide context for a supplemental radiomics-based tool.

Machine learning approach for distinguishing malignant and benign lung nodules utilizing standardized perinodular parenchymal features from CT

PURPOSE

Computed tomography (CT) is an effective method for detecting and characterizing lung nodules in vivo. With the growing use of chest CT, the detection frequency of lung nodules is increasing. Noninvasive methods to distinguish malignant from benign nodules have the potential to decrease the clinical burden, risk, and cost involved in follow-up procedures on the large number of false-positive lesions detected. This study examined the benefit of including perinodular parenchymal features in machine learning (ML) tools for pulmonary nodule assessment.

METHODS

Lung nodule cases with pathology confirmed diagnosis (74 malignant, 289 benign) were used to extract quantitative imaging characteristics from computed tomography scans of the nodule and perinodular parenchyma tissue. A ML tool development pipeline was employed using k-medoids clustering and information theory to determine efficient predictor sets for different amounts of parenchyma inclusion and build an artificial neural network classifier. The resulting ML tool was validated using an independent cohort (50 malignant, 50 benign).

RESULTS

The inclusion of parenchymal imaging features improved the performance of the ML tool over exclusively nodular features (P < 0.01). The best performing ML tool included features derived from nodule diameter-based surrounding parenchyma tissue quartile bands. We demonstrate similar high-performance values on the independent validation cohort (AUC-ROC = 0.965). A comparison using the independent validation cohort with the Fleischner pulmonary nodule follow-up guidelines demonstrated a theoretical reduction in recommended follow-up imaging and procedures.

CONCLUSIONS

Radiomic features extracted from the parenchyma surrounding lung nodules contain valid signals with spatial relevance for the task of lung cancer risk classification. Through standardization of feature extraction regions from the parenchyma, ML tool validation performance of 100% sensitivity and 96% specificity was achieved.

An Appraisal of Nodule Diagnosis for Lung Cancer in CT Images

As "the second eyes" of radiologists, computer-aided diagnosis systems play a significant role in nodule detection and diagnosis for lung cancer. In this paper, we aim to provide a systematic survey of state-of-the-art techniques (both traditional techniques and deep learning techniques) for nodule diagnosis from computed tomography images. This review first introduces the current progress and the popular structure used for nodule diagnosis. In particular, we provide a detailed overview of the five major stages in the computer-aided diagnosis systems: data acquisition, nodule segmentation, feature extraction, feature selection and nodule classification. Second, we provide a detailed report of the selected works and make a comprehensive comparison between selected works. The selected papers are from the IEEE Xplore, Science Direct, PubMed, and Web of Science databases up to December 2018. Third, we discuss and summarize the better techniques used in nodule diagnosis and indicate the existing future challenges in this field, such as improving the area under the receiver operating characteristic curve and accuracy, developing new deep learning-based diagnosis techniques, building efficient feature sets (fusing traditional features and deep features), developing high-quality labeled databases with malignant and benign nodules and promoting the cooperation between medical organizations and academic institutions.

Post-imaging pulmonary nodule mathematical prediction models: are they clinically relevant?

OBJECTIVES

Post-imaging mathematical prediction models (MPMs) provide guidance for the management of solid pulmonary nodules by providing a lung cancer risk score from demographic and radiologists-indicated imaging characteristics. We hypothesized calibrating the MPM risk score threshold to a local study cohort would result in improved performance over the original recommended MPM thresholds. We compared the pre- and post-calibration performance of four MPM models and determined if improvement in MPM prediction occurs as nodules are imaged longitudinally.

MATERIALS AND METHODS

A common cohort of 317 individuals with computed tomography-detected, solid nodules (80 malignant, 237 benign) were used to evaluate the MPM performance. We created a web-based application for this study that allows others to easily calibrate thresholds and analyze the performance of MPMs on their local cohort. Thirty patients with repeated imaging were tested for improved performance longitudinally.

RESULTS

Using calibrated thresholds, Mayo Clinic and Brock University (BU) MPMs performed the best (AUC = 0.63, 0.61) compared to the Veteran's Affairs (0.51) and Peking University (0.55). Only BU had consensus with the original MPM threshold; the other calibrated thresholds improved MPM accuracy. No significant improvements in accuracy were found longitudinally between time points.

CONCLUSIONS

Calibration to a common cohort can select the best-performing MPM for your institution. Without calibration, BU has the most stable performance in solid nodules ≥ 8 mm but has only moderate potential to refine subjects into appropriate workup. Application of MPM is recommended only at initial evaluation as no increase in accuracy was achieved over time.

KEY POINTS

• Post-imaging lung cancer risk mathematical predication models (MPMs) perform poorly on local populations without calibration. • An application is provided to facilitate calibration to new study cohorts: the Mayo Clinic model, the U.S. Department of Veteran's Affairs model, the Brock University model, and the Peking University model. • No significant improvement in risk prediction occurred in nodules with repeated imaging sessions, indicating the potential value of risk prediction application is limited to the initial evaluation.

Perinodular and Intranodular Radiomic Features on Lung CT Images Distinguish Adenocarcinomas from Granulomas

Purpose To evaluate ability of radiomic (computer-extracted imaging) features to distinguish non-small cell lung cancer adenocarcinomas from granulomas at noncontrast CT. Materials and Methods For this retrospective study, screening or standard diagnostic noncontrast CT images were collected for 290 patients (mean age, 68 years; range, 18-92 years; 125 men [mean age, 67 years; range, 18-90 years] and 165 women [mean age, 68 years; range, 33-92 years]) from two institutions between 2007 and 2013. Histopathologic analysis was available for one nodule per patient. Corresponding nodule of interest was identified on axial CT images by a radiologist with manual annotation. Nodule shape, wavelet (Gabor), and texture-based (Haralick and Laws energy) features were extracted from intra- and perinodular regions. Features were pruned to train machine learning classifiers with 145 patients. In a test set of 145 patients, classifier results were compared against a convolutional neural network (CNN) and diagnostic readings of two radiologists. Results Support vector machine classifier with intranodular radiomic features achieved an area under the receiver operating characteristic curve (AUC) of 0.75 on the test set. Combining radiomics of intranodular with perinodular regions improved the AUC to 0.80. On the same test set, CNN resulted in an AUC of 0.76. Radiologist readers achieved AUCs of 0.61 and 0.60, respectively. Conclusion Radiomic features from intranodular and perinodular regions of nodules can distinguish non-small cell lung cancer adenocarcinomas from benign granulomas at noncontrast CT. © RSNA, 2018 Online supplemental material is available for this article. See also the editorial by Nishino in this issue.

A Comparative Texture Analysis Based on NECT and CECT Images to Differentiate Lung Adenocarcinoma from Squamous Cell Carcinoma

The purpose of the study was to compare the texture based discriminative performances between non-contrast enhanced computed tomography (NECT) and contrast-enhanced computed tomography (CECT) images in differentiating lung adenocarcinoma (ADC) from squamous cell carcinoma (SCC) patients. Eighty-seven lung cancer subjects were enrolled in the study, including pathologically proved 47 ADC patients and 40 SCC patients, and 261 texture features were extracted from the manually delineated region of interests on CECT and NECT images respectively. Fisher score was then used to select the effective discriminative texture features between groups, and the selected texture features were adopted to differentiate ADC from SCC using Support Vector Machine and Leave-one-out cross-validation. Both NECT and CECT images could achieve the same best classification accuracy of 95.4%, and most of the informative features were from the gray-level co-occurrence matrix. In addition, CECT images were found with enhanced texture features compared with NECT images, and combining texture features of CECT and NECT images together could further improve the prediction accuracy. Besides the texture feature, the tumor location information also contributed to the differential diagnosis between ADC and SCC.

Deep learning enables automated scoring of liver fibrosis stages

Current liver fibrosis scoring by computer-assisted image analytics is not fully automated as it requires manual preprocessing (segmentation and feature extraction) typically based on domain knowledge in liver pathology. Deep learning-based algorithms can potentially classify these images without the need for preprocessing through learning from a large dataset of images. We investigated the performance of classification models built using a deep learning-based algorithm pre-trained using multiple sources of images to score liver fibrosis and compared them against conventional non-deep learning-based algorithms - artificial neural networks (ANN), multinomial logistic regression (MLR), support vector machines (SVM) and random forests (RF). Automated feature classification and fibrosis scoring were achieved by using a transfer learning-based deep learning network, AlexNet-Convolutional Neural Networks (CNN), with balanced area under receiver operating characteristic (AUROC) values of up to 0.85–0.95 versus ANN (AUROC of up to 0.87–1.00), MLR (AUROC of up to 0.73–1.00), SVM (AUROC of up to 0.69–0.99) and RF (AUROC of up to 0.94–0.99). Results indicate that a deep learning-based algorithm with transfer learning enables the construction of a fully automated and accurate prediction model for scoring liver fibrosis stages that is comparable to other conventional non-deep learning-based algorithms that are not fully automated.

Radiomic biomarkers informative of cancerous transformation in neurofibromatosis-1 plexiform tumors

BACKGROUND

This study explores whether objective, quantitative radiomic biomarkers derived from magnetic resonance (MR), positron emission tomography (PET), and computed tomography (CT) may be useful in reliably distinguishing malignant peripheral nerve sheath tumors (MPNST) from benign plexiform neurofibromas (PN).

METHODS

A registration and segmentation pipeline was established using a cohort of NF1 patients with histopathological diagnosis of PN or MPNST, and medical imaging of the PN including MR and PET-CT. The corrected MR datasets were registered to the corresponding PET-CT via landmark-based registration. PET standard-uptake value (SUV) thresholds were used to guide segmentation of volumes of interest: MPNST-associated PET-hot regions (SUV≥3.5) and PN-associated PET-elevated regions (2.0<SUV<3.5). Quantitative imaging features were extracted from the MR, PET, and CT data and compared for statistical differences. Intensity histogram features included (mean, media, maximum, variance, full width at half maximum, entropy, kurtosis, and skewness), while image texture was quantified using Law's texture energy measures, grey-level co-occurrence matrices, and neighborhood grey-tone difference matrices.

RESULTS

For each of the 20 NF1 subjects, a total of 320 features were extracted from the image data. Feature reduction and statistical testing identified 9 independent radiomic biomarkers from the MR data (4 intensity and 5 texture) and 4 PET (2 intensity and 2 texture) were different between the PET-hot versus PET-elevated volumes of interest.

CONCLUSIONS

Our data suggests imaging features can be used to distinguish malignancy in NF1-realted tumors, which could improve MPNST risk assessment and positively impact clinical management of NF1 patients.

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

Computed tomography (CT) examinations are commonly used to predict lung nodule malignancy in patients, which are shown to improve noninvasive early diagnosis of lung cancer. It remains challenging for computational approaches to achieve performance comparable to experienced radiologists. Here we present NoduleX, a systematic approach to predict lung nodule malignancy from CT data, based on deep learning convolutional neural networks (CNN). For training and validation, we analyze >1000 lung nodules in images from the LIDC/IDRI cohort. All nodules were identified and classified by four experienced thoracic radiologists who participated in the LIDC project. NoduleX achieves high accuracy for nodule malignancy classification, with an AUC of ~0.99. This is commensurate with the analysis of the dataset by experienced radiologists. Our approach, NoduleX, provides an effective framework for highly accurate nodule malignancy prediction with the model trained on a large patient population. Our results are replicable with software available at http://bioinformatics.astate.edu/NoduleX .

Clinical applications of textural analysis in non-small cell lung cancer

Lung cancer is the leading cause of cancer mortality worldwide. Treatment pathways include regular cross-sectional imaging, generating large data sets which present intriguing possibilities for exploitation beyond standard visual interpretation. This additional data mining has been termed "radiomics" and includes semantic and agnostic approaches. Textural analysis (TA) is an example of the latter, and uses a range of mathematically derived features to describe an image or region of an image. Often TA is used to describe a suspected or known tumour. TA is an attractive tool as large existing image sets can be submitted to diverse techniques for data processing, presentation, interpretation and hypothesis testing with annotated clinical outcomes. There is a growing anthology of published data using different TA techniques to differentiate between benign and malignant lung nodules, differentiate tissue subtypes of lung cancer, prognosticate and predict outcome and treatment response, as well as predict treatment side effects and potentially aid radiotherapy planning. The aim of this systematic review is to summarize the current published data and understand the potential future role of TA in managing lung cancer.

Feature fusion for lung nodule classification

PURPOSE

This article examines feature-based nodule description for the purpose of nodule classification in chest computed tomography scanning.

METHODS

Three features based on (i) Gabor filter, (ii) multi-resolution local binary pattern (LBP) texture features and (iii) signed distance fused with LBP which generates a combinational shape and texture feature are utilized to provide feature descriptors of malignant and benign nodules and non-nodule regions of interest. Support vector machines (SVMs) and k-nearest neighbor (kNN) classifiers in serial and two-tier cascade frameworks are optimized and analyzed for optimal classification results of nodules.

RESULTS

A total of 1191 nodule and non-nodule samples from the Lung Image Data Consortium database is used for analysis. Classification using SVM and kNN classifiers is examined. The classification results from the two-tier cascade SVM using Gabor features showed overall better results for identifying non-nodules, malignant and benign nodules with average area under the receiver operating characteristics (AUC-ROC) curves of 0.99 and average f1-score of 0.975 over the two tiers.

CONCLUSION

In the results, higher overall AUCs and f1-scores were obtained for the non-nodules cases using any of the three features, showing the greatest distinguishability over nodules (benign/malignant). SVM and kNN classifiers were used for benign, malignant and non-nodule classification, where Gabor proved to be the most effective of the features for classification. The cascaded framework showed the greatest distinguishability between benign and malignant nodules.

Radiologic features of precancerous areas of the lungs in chronic obstructive pulmonary disease

BACKGROUND

Only a few studies have evaluated the radiologic features of pre-existing structural abnormalities where lung cancer may develop. This study aimed to analyze the computed tomography (CT) images of lung areas where new cancer developed in chronic obstructive pulmonary disease (COPD) patients.

PATIENTS AND METHODS

We conducted a multicenter, longitudinal cohort study, called the Keio COPD Comorbidity Research, to assess the incidence of lung cancer. Emphysema and interstitial abnormalities were evaluated in 240 COPD patients who had baseline CT scans applicable for further digital analyses. For patients who developed lung cancer during the 3-year follow-up period, the local spherical lung density of the precancerous area was individually quantified.

RESULTS

Lung cancer was newly diagnosed in 21 participants (2.3% per year). The percent-age of low attenuation area in patients who developed lung cancer was higher than that of the other patients (20.0% vs 10.4%, P=0.014). The presence of emphysema (odds ratio [OR] 4.2, 95% confidence interval [CI] 1.0-29.0, P=0.049) or interstitial lung abnormalities (OR 15.6, 95% CI 4.4-65.4, P<0.0001) independently increased the risk for lung cancer. Compared with the density of the entire lung, the local density of the precancerous area was almost the same in patients with heterogeneous emphysema, but it was higher in most patients with interstitial abnormalities.

CONCLUSION

The presence of emphysema or interstitial abnormalities or a combination of both were independent predictors of lung cancer development in COPD patients. Furthermore, lung cancer most often developed in non-emphysematous areas or in interstitial abnormalities.

Statistical tools for the temporal analysis and classification of lung lesions

BACKGROUND AND OBJECTIVE

Lung cancer remains one of the most common cancers globally. Temporal evaluation is an important tool for analyzing the malignant behavior of lesions during treatment, or of indeterminate lesions that may be benign. This work proposes a methodology for the analysis, quantification, and visualization of small (local) and large (global) changes in lung lesions. In addition, we extract textural features for the classification of lesions as benign or malignant.

METHODS

We employ the statistical concept of uncertainty to associate each voxel of a lesion to a probability that changes occur in the lesion over time. We employ the Jensen divergence and hypothesis test locally to verify voxel-to-voxel changes, and globally to capture changes in lesion volumes.

RESULTS

For the local hypothesis test, we determine that the change in density varies by between 3.84 and 40.01% of the lesion volume in a public database of malignant lesions under treatment, and by between 5.76 and 35.43% in a private database of benign lung nodules. From the texture analysis of regions in which the density changes occur, we are able to discriminate lung lesions with an accuracy of 98.41%, which shows that these changes can indicate the true nature of the lesion.

CONCLUSION

In addition to the visual aspects of the density changes occurring in the lesions over time, we quantify these changes and analyze the entire set using volumetry. In the case of malignant lesions, large b-divergence values are associated with major changes in lesion volume. In addition, this occurs when the change in volume is small but is associated with significant changes in density, as indicated by the histogram divergence. For benign lesions, the methodology shows that even in cases where the change in volume is small, a change of density occurs. This proves that even in lesions that appear stable, a change in density occurs.

Lung nodule malignancy classification using only radiologist-quantified image features as inputs to statistical learning algorithms: probing the Lung Image Database Consortium dataset with two statistical learning methods

In the assessment of nodules in CT scans of the lungs, a number of image-derived features are diagnostically relevant. Currently, many of these features are defined only qualitatively, so they are difficult to quantify from first principles. Nevertheless, these features (through their qualitative definitions and interpretations thereof) are often quantified via a variety of mathematical methods for the purpose of computer-aided diagnosis (CAD). To determine the potential usefulness of quantified diagnostic image features as inputs to a CAD system, we investigate the predictive capability of statistical learning methods for classifying nodule malignancy. We utilize the Lung Image Database Consortium dataset and only employ the radiologist-assigned diagnostic feature values for the lung nodules therein, as well as our derived estimates of the diameter and volume of the nodules from the radiologists' annotations. We calculate theoretical upper bounds on the classification accuracy that are achievable by an ideal classifier that only uses the radiologist-assigned feature values, and we obtain an accuracy of 85.74 [Formula: see text], which is, on average, 4.43% below the theoretical maximum of 90.17%. The corresponding area-under-the-curve (AUC) score is 0.932 ([Formula: see text]), which increases to 0.949 ([Formula: see text]) when diameter and volume features are included and has an accuracy of 88.08 [Formula: see text]. Our results are comparable to those in the literature that use algorithmically derived image-based features, which supports our hypothesis that lung nodules can be classified as malignant or benign using only quantified, diagnostic image features, and indicates the competitiveness of this approach. We also analyze how the classification accuracy depends on specific features and feature subsets, and we rank the features according to their predictive power, statistically demonstrating the top four to be spiculation, lobulation, subtlety, and calcification.

A 3-D Riesz-Covariance Texture Model for Prediction of Nodule Recurrence in Lung CT

This paper proposes a novel imaging biomarker of lung cancer relapse from 3-D texture analysis of CT images. Three-dimensional morphological nodular tissue properties are described in terms of 3-D Riesz-wavelets. The responses of the latter are aggregated within nodular regions by means of feature covariances, which leverage rich intra- and inter-variations of the feature space dimensions. When compared to the classical use of the average for feature aggregation, feature covariances preserve spatial co-variations between features. The obtained Riesz-covariance descriptors lie on a manifold governed by Riemannian geometry allowing geodesic measurements and differentiations. The latter property is incorporated both into a kernel for support vector machines (SVM) and a manifold-aware sparse regularized classifier. The effectiveness of the presented models is evaluated on a dataset of 110 patients with non-small cell lung carcinoma (NSCLC) and cancer recurrence information. Disease recurrence within a timeframe of 12 months could be predicted with an accuracy of 81.3-82.7%. The anatomical location of recurrence could be discriminated between local, regional and distant failure with an accuracy of 78.3-93.3%. The obtained results open novel research perspectives by revealing the importance of the nodular regions used to build the predictive models.

Radiomics of Lung Nodules: A Multi-Institutional Study of Robustness and Agreement of Quantitative Imaging Features

Radiomics is to provide quantitative descriptors of normal and abnormal tissues during classification and prediction tasks in radiology and oncology. Quantitative Imaging Network members are developing radiomic “feature” sets to characterize tumors, in general, the size, shape, texture, intensity, margin, and other aspects of the imaging features of nodules and lesions. Efforts are ongoing for developing an ontology to describe radiomic features for lung nodules, with the main classes consisting of size, local and global shape descriptors, margin, intensity, and texture-based features, which are based on wavelets, Laplacian of Gaussians, Law's features, gray-level co-occurrence matrices, and run-length features. The purpose of this study is to investigate the sensitivity of quantitative descriptors of pulmonary nodules to segmentations and to illustrate comparisons across different feature types and features computed by different implementations of feature extraction algorithms. We calculated the concordance correlation coefficients of the features as a measure of their stability with the underlying segmentation; 68% of the 830 features in this study had a concordance CC of ≥0.75. Pairwise correlation coefficients between pairs of features were used to uncover associations between features, particularly as measured by different participants. A graphical model approach was used to enumerate the number of uncorrelated feature groups at given thresholds of correlation. At a threshold of 0.75 and 0.95, there were 75 and 246 subgroups, respectively, providing a measure for the features' redundancy.

LUNGx Challenge for computerized lung nodule classification

The purpose of this work is to describe the LUNGx Challenge for the computerized classification of lung nodules on diagnostic computed tomography (CT) scans as benign or malignant and report the performance of participants' computerized methods along with that of six radiologists who participated in an observer study performing the same Challenge task on the same dataset. The Challenge provided sets of calibration and testing scans, established a performance assessment process, and created an infrastructure for case dissemination and result submission. Ten groups applied their own methods to 73 lung nodules (37 benign and 36 malignant) that were selected to achieve approximate size matching between the two cohorts. Area under the receiver operating characteristic curve (AUC) values for these methods ranged from 0.50 to 0.68; only three methods performed statistically better than random guessing. The radiologists' AUC values ranged from 0.70 to 0.85; three radiologists performed statistically better than the best-performing computer method. The LUNGx Challenge compared the performance of computerized methods in the task of differentiating benign from malignant lung nodules on CT scans, placed in the context of the performance of radiologists on the same task. The continued public availability of the Challenge cases will provide a valuable resource for the medical imaging research community.

Radiomics: a new application from established techniques

The increasing use of biomarkers in cancer have led to the concept of personalized medicine for patients. Personalized medicine provides better diagnosis and treatment options available to clinicians. Radiological imaging techniques provide an opportunity to deliver unique data on different types of tissue. However, obtaining useful information from all radiological data is challenging in the era of "big data". Recent advances in computational power and the use of genomics have generated a new area of research termed Radiomics. Radiomics is defined as the high throughput extraction of quantitative imaging features or texture (radiomics) from imaging to decode tissue pathology and creating a high dimensional data set for feature extraction. Radiomic features provide information about the gray-scale patterns, inter-pixel relationships. In addition, shape and spectral properties can be extracted within the same regions of interest on radiological images. Moreover, these features can be further used to develop computational models using advanced machine learning algorithms that may serve as a tool for personalized diagnosis and treatment guidance.