Computer-aided diagnosis of lung cancer and pulmonary embolism in computed tomography-a review

Improved detection of small pulmonary embolism on unenhanced computed tomography using an artificial intelligence-based algorithm - a single centre retrospective study

To preliminarily verify the feasibility of a deep-learning (DL) artificial intelligence (AI) model to localize pulmonary embolism (PE) on unenhanced chest-CT by comparison with pulmonary artery (PA) CT angiography (CTA). In a monocentric study, we retrospectively reviewed 99 oncological patients (median age in years: 64 (range: 28-92 years); percentage of female: 39.4%) who received unenhanced and contrast-enhanced chest CT examinations in one session between January 2020 and October 2022 and who were diagnosed incidentally with PE. Findings in the unenhanced images were correlated with the contrast-enhanced images, which were considered the gold standard for central, segmental and subsegmental PE. The new algorithm was trained and tested based on the 99 unenhanced chest-CT image data sets. Based on them, candidate boxes, which were output by the model, were post-processed by evaluating whether the predicted box intersects with the patient's lung segmentation at any position. The AI-based algorithm proved to have an overall sensitivity of 54.5% for central, of 81.9% for segmental and 80.0% for subsegmental PE if taking n = 20 candidate boxes into account. Depending on the localization of the pulmonary embolism, the detection rate for only one box was: 18.1% central, 34.7% segmental and 0.0% subsegmental. The median volume of the clots differed significantly between the three subgroups and was 846.5 mm3 (IQR:591.1-964.8) in central, 201.3 mm3 (IQR:98.3-390.9) in segmental and 110.6 mm3 (IQR:94.3-128.0) in subsegmental PA (p < 0.05). The new algorithm proved to have high sensitivity in detecting PE in particular in segmental/subsegmental localization and may guide to decide whether a second contrast enhanced CT is necessary.

Systems Biology in Cancer Diagnosis Integrating Omics Technologies and Artificial Intelligence to Support Physician Decision Making

Cancer is the second major cause of disease-related death worldwide, and its accurate early diagnosis and therapeutic intervention are fundamental for saving the patient's life. Cancer, as a complex and heterogeneous disorder, results from the disruption and alteration of a wide variety of biological entities, including genes, proteins, mRNAs, miRNAs, and metabolites, that eventually emerge as clinical symptoms. Traditionally, diagnosis is based on clinical examination, blood tests for biomarkers, the histopathology of a biopsy, and imaging (MRI, CT, PET, and US). Additionally, omics biotechnologies help to further characterize the genome, metabolome, microbiome traits of the patient that could have an impact on the prognosis and patient's response to the therapy. The integration of all these data relies on gathering of several experts and may require considerable time, and, unfortunately, it is not without the risk of error in the interpretation and therefore in the decision. Systems biology algorithms exploit Artificial Intelligence (AI) combined with omics technologies to perform a rapid and accurate analysis and integration of patient's big data, and support the physician in making diagnosis and tailoring the most appropriate therapeutic intervention. However, AI is not free from possible diagnostic and prognostic errors in the interpretation of images or biochemical-clinical data. Here, we first describe the methods used by systems biology for combining AI with omics and then discuss the potential, challenges, limitations, and critical issues in using AI in cancer research.

Early Triage of Critically Ill Adult Patients With Mushroom Poisoning: Machine Learning Approach

BACKGROUND

Early triage of patients with mushroom poisoning is essential for administering precise treatment and reducing mortality. To our knowledge, there has been no established method to triage patients with mushroom poisoning based on clinical data.

OBJECTIVE

The purpose of this work was to construct a triage system to identify patients with mushroom poisoning based on clinical indicators using several machine learning approaches and to assess the prediction accuracy of these strategies.

METHODS

In all, 567 patients were collected from 5 primary care hospitals and facilities in Enshi, Hubei Province, China, and divided into 2 groups; 322 patients from 2 hospitals were used as the training cohort, and 245 patients from 3 hospitals were used as the test cohort. Four machine learning algorithms were used to construct the triage model for patients with mushroom poisoning. Performance was assessed using the area under the receiver operating characteristic curve (AUC), decision curve, sensitivity, specificity, and other representative statistics. Feature contributions were evaluated using Shapley additive explanations.

RESULTS

Among several machine learning algorithms, extreme gradient boosting (XGBoost) showed the best discriminative ability in 5-fold cross-validation (AUC=0.83, 95% CI 0.77-0.90) and the test set (AUC=0.90, 95% CI 0.83-0.96). In the test set, the XGBoost model had a sensitivity of 0.93 (95% CI 0.81-0.99) and a specificity of 0.79 (95% CI 0.73-0.85), whereas the physicians' assessment had a sensitivity of 0.86 (95% CI 0.72-0.95) and a specificity of 0.66 (95% CI 0.59-0.73).

CONCLUSIONS

The 14-factor XGBoost model for the early triage of mushroom poisoning can rapidly and accurately identify critically ill patients and will possibly serve as an important basis for the selection of treatment options and referral of patients, potentially reducing patient mortality and improving clinical outcomes.

Computed Tomography Features and Tumor Spread Through Air Spaces in Lung Adenocarcinoma: A Meta-analysis

To compare computed tomography (CT)-based radiologic features in patients, who are diagnosed with lung adenocarcinoma with the pathologically detected spread of tumor cells through air spaces (STAS positive [STAS+]) and those with no STAS. PubMed, Embase, and Scopus databases were systematically searched for observational studies (either retrospective or prospective) of patients with lung adenocarcinoma that had compared CT-based features between STAS+ and STAS-negative cases (STAS-). The pooled effect sizes were reported as odds ratio (OR) and weighted mean difference (WMD). STATA software was used for statistical analysis. The meta-analysis included 10 studies. Compared with STAS-, STAS+ adenocarcinoma was associated with increased odds of solid nodule (OR: 3.30, 95% CI: 2.52, 4.31), spiculation (OR: 2.05, 95% CI: 1.36, 3.08), presence of cavitation (OR: 1.49, 95% CI: 1.00, 2.22), presence of clear boundary (OR: 3.01, 95% CI: 1.70, 5.32), lobulation (OR: 1.65, 95% CI: 1.11, 2.47), and pleural indentation (OR: 1.98, 95% CI: 1.41, 2.77). STAS+ tumors had significant association with the presence of pulmonary vessel convergence (OR: 2.15, 95% CI: 1.61, 2.87), mediastinal lymphadenopathy (OR: 2.06, 95% CI: 1.20, 3.56), and pleural thickening (OR: 2.58, 95% CI: 1.73, 3.84). The mean nodule diameter (mm) (WMD: 6.19, 95% CI: 3.71, 8.66) and the mean solid component (%) (WMD: 24.5, 95% CI: 10.5, 38.6) were higher in STAS+ tumors, compared with STAS- ones. The findings suggest a significant association of certain CT-based features with the presence of STAS in patients with lung adenocarcinoma. These features may be important in influencing the nature of surgical management.

Digital Health Applications in Oncology: An Opportunity to Seize

Digital health advances have transformed many clinical areas including psychiatric and cardiovascular care. However, digital health innovation is relatively nascent in cancer care, which represents the fastest growing area of health-care spending. Opportunities for digital health innovation in oncology include patient-facing technologies that improve patient experience, safety, and patient-clinician interactions; clinician-facing technologies that improve their ability to diagnose pathology and predict adverse events; and quality of care and research infrastructure to improve clinical workflows, documentation, decision support, and clinical trial monitoring. The COVID-19 pandemic and associated shifts of care to the home and community dramatically accelerated the integration of digital health technologies into virtually every aspect of oncology care. However, the pandemic has also exposed potential flaws in the digital health ecosystem, namely in clinical integration strategies; data access, quality, and security; and regulatory oversight and reimbursement for digital health technologies. Stemming from the proceedings of a 2020 workshop convened by the National Cancer Policy Forum of the National Academies of Sciences, Engineering, and Medicine, this article summarizes the current state of digital health technologies in medical practice and strategies to improve clinical utility and integration. These recommendations, with calls to action for clinicians, health systems, technology innovators, and policy makers, will facilitate efficient yet safe integration of digital health technologies into cancer care.

Automated detection of pulmonary embolism from CT-angiograms using deep learning

Background

The aim of this study was to develop and evaluate a deep neural network model in the automated detection of pulmonary embolism (PE) from computed tomography pulmonary angiograms (CTPAs) using only weakly labelled training data.

Methods

We developed a deep neural network model consisting of two parts: a convolutional neural network architecture called InceptionResNet V2 and a long-short term memory network to process whole CTPA stacks as sequences of slices. Two versions of the model were created using either chest X-rays (Model A) or natural images (Model B) as pre-training data. We retrospectively collected 600 CTPAs to use in training and validation and 200 CTPAs to use in testing. CTPAs were annotated only with binary labels on both stack- and slice-based levels. Performance of the models was evaluated with ROC and precision–recall curves, specificity, sensitivity, accuracy, as well as positive and negative predictive values.

Results

Both models performed well on both stack- and slice-based levels. On the stack-based level, Model A reached specificity and sensitivity of 93.5% and 86.6%, respectively, outperforming Model B slightly (specificity 90.7% and sensitivity 83.5%). However, the difference between their ROC AUC scores was not statistically significant (0.94 vs 0.91, p = 0.07).

Conclusions

We show that a deep learning model trained with a relatively small, weakly annotated dataset can achieve excellent performance results in detecting PE from CTPAs.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12880-022-00763-z.

Artificial intelligence and imaging: Opportunities in cardio-oncology

Cardiovascular disease is a leading cause of death in cancer survivors. It is critical to apply new predictive and early diagnostic methods in this population, as this can potentially inform cardiovascular treatment and surveillance decision-making. We discuss the application of artificial intelligence (AI) technologies to cardiovascular imaging in cardio-oncology, with a particular emphasis on prevention and targeted treatment of a variety of cardiovascular conditions in cancer patients. Recently, the use of AI-augmented cardiac imaging in cardio-oncology is gaining traction. A large proportion of cardio-oncology patients are screened and followed using left ventricular ejection fraction (LVEF) and global longitudinal strain (GLS), currently obtained using echocardiography. This use will continue to increase with new cardiotoxic cancer treatments. AI is being tested to increase precision, throughput, and accuracy of LVEF and GLS, guide point-of-care image acquisition, and integrate imaging and clinical data to optimize the prediction and detection of cardiac dysfunction. The application of AI to cardiovascular magnetic resonance imaging (CMR), computed tomography (CT; especially coronary artery calcium or CAC scans), single proton emission computed tomography (SPECT) and positron emission tomography (PET) imaging acquisition is also in early stages of analysis for prediction and assessment of cardiac tumors and cardiovascular adverse events in patients treated for childhood or adult cancer. The opportunities for application of AI in cardio-oncology imaging are promising, and if availed, will improve clinical practice and benefit patient care.

Establishing an interdisciplinary research team for cardio-oncology artificial intelligence informatics precision and health equity

Study objective

A multi-institutional interdisciplinary team was created to develop a research group focused on leveraging artificial intelligence and informatics for cardio-oncology patients. Cardio-oncology is an emerging medical field dedicated to prevention, screening, and management of adverse cardiovascular effects of cancer/ cancer therapies. Cardiovascular disease is a leading cause of death in cancer survivors. Cardiovascular risk in these patients is higher than in the general population. However, prediction and prevention of adverse cardiovascular events in individuals with a history of cancer/cancer treatment is challenging. Thus, establishing an interdisciplinary team to create cardiovascular risk stratification clinical decision aids for integration into electronic health records for oncology patients was considered crucial.

Design/setting/participants

Core team members from the Medical College of Wisconsin (MCW), University of Wisconsin-Milwaukee (UWM), and Milwaukee School of Engineering (MSOE), and additional members from Cleveland Clinic, Mayo Clinic, and other institutions have joined forces to apply high-performance computing in cardio-oncology.

Results

The team is comprised of clinicians and researchers from relevant complementary and synergistic fields relevant to this work. The team has built an epidemiological cohort of ~5000 cancer survivors that will serve as a database for interdisciplinary multi-institutional artificial intelligence projects.

Conclusion

Lessons learned from establishing this team, as well as initial findings from the epidemiology cohort, are presented. Barriers have been broken down to form a multi-institutional interdisciplinary team for health informatics research in cardio-oncology. A database of cancer survivors has been created collaboratively by the team and provides initial insight into cardiovascular outcomes and comorbidities in this population.

Performance of a 3D convolutional neural network in the detection of hypoperfusion at CT pulmonary angiography in patients with chronic pulmonary embolism: a feasibility study

BACKGROUND

Chronic pulmonary embolism (CPE) is a life-threatening disease easily misdiagnosed on computed tomography. We investigated a three-dimensional convolutional neural network (CNN) algorithm for detecting hypoperfusion in CPE from computed tomography pulmonary angiography (CTPA).

METHODS

Preoperative CTPA of 25 patients with CPE and 25 without pulmonary embolism were selected. We applied a 48%-12%-40% training-validation-testing split (12 positive and 12 negative CTPA volumes for training, 3 positives and 3 negatives for validation, 10 positives and 10 negatives for testing). The median number of axial images per CTPA was 335 (min-max, 111-570). Expert manual segmentations were used as training and testing targets. The CNN output was compared to a method in which a Hounsfield unit (HU) threshold was used to detect hypoperfusion. Receiver operating characteristic area under the curve (AUC) and Matthew correlation coefficient (MCC) were calculated with their 95% confidence interval (CI).

RESULTS

The predicted segmentations of CNN showed AUC 0.87 (95% CI 0.82-0.91), those of HU-threshold method 0.79 (95% CI 0.74-0.84). The optimal global threshold values were CNN output probability ≥ 0.37 and ≤ -850 HU. Using these values, MCC was 0.46 (95% CI 0.29-0.59) for CNN and 0.35 (95% CI 0.18-0.48) for HU-threshold method (average difference in MCC in the bootstrap samples 0.11 (95% CI 0.05-0.16). A high CNN prediction probability was a strong predictor of CPE.

CONCLUSIONS

We proposed a deep learning method for detecting hypoperfusion in CPE from CTPA. This model may help evaluating disease extent and supporting treatment planning.

Deep learning applications in neuro-oncology

Deep learning (DL) is a relatively newer subdomain of machine learning (ML) with incredible potential for certain applications in the medical field. Given recent advances in its use in neuro-oncology, its role in diagnosing, prognosticating, and managing the care of cancer patients has been the subject of many research studies. The gamut of studies has shown that the landscape of algorithmic methods is constantly improving with each iteration from its inception. With the increase in the availability of high-quality data, more training sets will allow for higher fidelity models. However, logistical and ethical concerns over a prospective trial comparing prognostic abilities of DL and physicians severely limit the ability of this technology to be widely adopted. One of the medical tenets is judgment, a facet of medical decision making in DL that is often missing because of its inherent nature as a "black box." A natural distrust for newer technology, combined with a lack of autonomy that is normally expected in our current medical practices, is just one of several important limitations in implementation. In our review, we will first define and outline the different types of artificial intelligence (AI) as well as the role of AI in the current advances of clinical medicine. We briefly highlight several of the salient studies using different methods of DL in the realm of neuroradiology and summarize the key findings and challenges faced when using this nascent technology, particularly ethical challenges that could be faced by users of DL.

Preliminary study on artificial intelligence diagnosis of pulmonary embolism based on computer in-depth study

Background

Objective to preliminarily verify the feasibility of AI intelligent diagnosis of pulmonary embolism by using a new artificial intelligence (AI) computer-aided diagnosis system (CAD) to localize and quantitatively diagnose pulmonary embolism in pulmonary artery CT angiography (CTA).

Methods

Computed tomography angiography (CTA) data of 85 patients with PE in our hospital from January 2017 to May 2018 were retrospectively collected and randomly allocated to2 groups: computer depth learning group (n=43) and experimental group (n=42). For the training set (13,144 sheets) and the test set (313 sheets), the auxiliary diagnosis method was obtained and applied to the experimental group.

Results

Among the participants, a good sensitivity of 90.9% and an average false positive of 2.0 were obtained by using the deep learning detection method proposed in this paper, and the detection rate was positively correlated with arterial grade.

Conclusions

The computer-aided diagnostic method proposed in this paper can effectively improve the detection rate of PE, especially for the detection of intra-arterial embolism above grade 3. However, because of the high misdetection rate, more in-depth learning datasets are needed for the detection of embolism below grade 3.

Variabilities in Reference Standard by Radiologists and Performance Assessment in Detection of Pulmonary Embolism in CT Pulmonary Angiography

Annotating lesion locations by radiologists' manual marking is a key step to provide reference standard for the training and testing of a computer-aided detection system by supervised machine learning. Inter-reader variability is not uncommon in readings even by expert radiologists. This study evaluated the variability of the radiologist-identified pulmonary emboli (PEs) to demonstrate the importance of improving the reliability of the reference standard by a multi-step process for performance evaluation. In an initial reading of 40 CTPA PE cases, two experienced thoracic radiologists independently marked the PE locations. For markings from the two radiologists that did not agree, each radiologist re-read the cases independently to assess the discordant markings. Finally, for markings that still disagreed after the second reading, the two radiologists read together to reach a consensus. The variability of radiologists was evaluated by analyzing the agreement between two radiologists. For the 40 cases, 475 and 514 PEs were identified by radiologists R1 and R2 in the initial independent readings, respectively. For a total of 545 marks by the two radiologists, 81.5% (444/545) of the marks agreed but 101 marks in 36 cases differed. After consensus, 65 (64.4%) and 36 (35.6%) of the 101 marks were determined to be true PEs and false positives (FPs), respectively. Of these, 48 and 17 were false negatives (FNs) and 14 and 22 were FPs by R1 and R2, respectively. Our study demonstrated that there is substantial variability in reference standards provided by radiologists, which impacts the performance assessment of a lesion detection system. Combination of multiple radiologists' readings and consensus is needed to improve the reliability of a reference standard.

Computer-aided diagnosis in the era of deep learning

Computer-aided diagnosis (CAD) has been a major field of research for the past few decades. CAD uses machine learning methods to analyze imaging and/or nonimaging patient data and makes assessment of the patient's condition, which can then be used to assist clinicians in their decision-making process. The recent success of the deep learning technology in machine learning spurs new research and development efforts to improve CAD performance and to develop CAD for many other complex clinical tasks. In this paper, we discuss the potential and challenges in developing CAD tools using deep learning technology or artificial intelligence (AI) in general, the pitfalls and lessons learned from CAD in screening mammography and considerations needed for future implementation of CAD or AI in clinical use. It is hoped that the past experiences and the deep learning technology will lead to successful advancement and lasting growth in this new era of CAD, thereby enabling CAD to deliver intelligent aids to improve health care.

A dynamic lesion model for differentiation of malignant and benign pathologies

Malignant lesions have a high tendency to invade their surrounding environment compared to benign ones. This paper proposes a dynamic lesion model and explores the 2nd order derivatives at each image voxel, which reflect the rate of change of image intensity, as a quantitative measure of the tendency. The 2nd order derivatives at each image voxel are usually represented by the Hessian matrix, but it is difficult to quantify a matrix field (or image) through the lesion space as a measure of the tendency. We conjecture that the three eigenvalues contain important information of the Hessian matrix and are chosen as the surrogate representation of the Hessian matrix. By treating the three eigenvalues as a vector, called Hessian vector, which is defined in a local coordinate formed by three orthogonal Hessian eigenvectors and further adapting the gray level occurrence computing method to extract the vector texture descriptors (or measures) from the Hessian vector, a quantitative presentation for the dynamic lesion model is completed. The vector texture descriptors were applied to differentiate malignant from benign lesions from two pathologically proven datasets: colon polyps and lung nodules. The classification results not only outperform four state-of-the-art methods but also three radiologist experts.

Application of artificial intelligence in clinical non-small cell lung cancer

Lung cancer is the most common cause of cancer death in the world. Early diagnosis, screening and precise individualized treatment can significantly reduce the death rate of lung cancer. Artificial intelligence (AI) has been shown to be able to help clinicians make more accurate judgments and decisions in many ways. It has been involved in the screening of lung cancer, the judgment of benign and malignant degree of pulmonary nodules, the classification of histological cancer, the differentiation of histological subtypes, the identification of genomics, the judgment of the effectiveness of treatment and even the prognosis. AI has shown that it can be an excellent assistant for clinicians. This paper reviews the application of AI in the field of non-small cell lung cancer and describes the relevant progress. Although most of the studies to evaluate the clinical application of AI in non-small cell lung cancer have not been repeatable and generalizable, the research results highlight the efforts to promote the clinical application of AI technology and influence the future treatment direction.

Assessment of Acute Pulmonary Embolism by Computer-Aided Technique: A Reliability Study

Background

Acute pulmonary embolism is one of the most common cardiovascular diseases. Computer-aided technique is widely used in chest imaging, especially for assessing pulmonary embolism. The reliability and quantitative analyses of computer-aided technique are necessary. This study aimed to evaluate the reliability of geometry-based computer-aided detection and quantification for emboli morphology and severity of acute pulmonary embolism.

Material/Methods

Thirty patients suspected of acute pulmonary embolism were analyzed by both manual and computer-aided interpretation of vascular obstruction index and computer-aided measurements of emboli quantitative parameters. The reliability of Qanadli and Mastora scores was analyzed using computer-aided and manual interpretation.

Results

The time costs of manual and computer-aided interpretation were statistically different (374.90±150.16 versus 121.07±51.76, P<0.001). The difference between the computer-aided and manual interpretation of Qanadli score was 1.83±2.19, and 96.7% (29 out of 30) of the measurements were within 95% confidence interval (intraclass correlation coefficient, ICC=0.998). The difference between the computer-aided and manual interpretation of Mastora score was 1.46±1.62, and 96.7% (29 out of 30) of the measurements were within 95% confidence interval (ICC=0.997). The emboli quantitative parameters were moderately correlated with the Qanadli and Mastora scores (all P<0.001).

Conclusions

Computer-aided technique could reduce the time costs, improve the and reliability of vascular obstruction index and provided additional quantitative parameters for disease assessment.

Predicting lung nodule malignancies by combining deep convolutional neural network and handcrafted features

To predict lung nodule malignancy with a high sensitivity and specificity for low dose CT (LDCT) lung cancer screening, we propose a fusion algorithm that combines handcrafted features (HF) into the features learned at the output layer of a 3D deep convolutional neural network (CNN). First, we extracted twenty-nine HF, including nine intensity features, eight geometric features, and twelve texture features based on grey-level co-occurrence matrix (GLCM). We then trained 3D CNNs modified from three 2D CNN architectures (AlexNet, VGG-16 Net and Multi-crop Net) to extract the CNN features learned at the output layer. For each 3D CNN, the CNN features combined with the 29 HF were used as the input for the support vector machine (SVM) coupled with the sequential forward feature selection (SFS) method to select the optimal feature subset and construct the classifiers. The fusion algorithm takes full advantage of the HF and the highest level CNN features learned at the output layer. It can overcome the disadvantage of the HF that may not fully reflect the unique characteristics of a particular lesion by combining the intrinsic CNN features. Meanwhile, it also alleviates the requirement of a large scale annotated dataset for the CNNs based on the complementary of HF. The patient cohort includes 431 malignant nodules and 795 benign nodules extracted from the LIDC/IDRI database. For each investigated CNN architecture, the proposed fusion algorithm achieved the highest AUC, accuracy, sensitivity, and specificity scores among all competitive classification models.

Cancer as a Tissue Anomaly: Classifying Tumor Transcriptomes Based Only on Healthy Data

Since the turn of the century, researchers have sought to diagnose cancer based on gene expression signatures measured from the blood or biopsy as biomarkers. This task, known as classification, is typically solved using a suite of algorithms that learn a mathematical rule capable of discriminating one group ("cases") from another ("controls"). However, discriminatory methods can only identify cancerous samples that resemble those that the algorithm already saw during training. As such, discriminatory methods may be ill-suited for the classification of cancer: because the possibility space of cancer is definitively large, the existence of a one-of-a-kind gene expression signature is likely. Instead, we propose using an established surveillance method that detects anomalous samples based on their deviation from a learned normal steady-state structure. By transferring this method to transcriptomic data, we can create an anomaly detector for tissue transcriptomes, a "tissue detector," that is capable of identifying cancer without ever seeing a single cancer example. As a proof-of-concept, we train a "tissue detector" on normal GTEx samples that can classify TCGA samples with >90% AUC for 3 out of 6 tissues. Importantly, we find that the classification accuracy is improved simply by adding more healthy samples. We conclude this report by emphasizing the conceptual advantages of anomaly detection and by highlighting future directions for this field of study.

Multi-scale gradual integration CNN for false positive reduction in pulmonary nodule detection

Lung cancer is a global and dangerous disease, and its early detection is crucial for reducing the risks of mortality. In this regard, it has been of great interest in developing a computer-aided system for pulmonary nodules detection as early as possible on thoracic CT scans. In general, a nodule detection system involves two steps: (i) candidate nodule detection at a high sensitivity, which captures many false positives and (ii) false positive reduction from candidates. However, due to the high variation of nodule morphological characteristics and the possibility of mistaking them for neighboring organs, candidate nodule detection remains a challenge. In this study, we propose a novel Multi-scale Gradual Integration Convolutional Neural Network (MGI-CNN), designed with three main strategies: (1) to use multi-scale inputs with different levels of contextual information, (2) to use abstract information inherent in different input scales with gradual integration, and (3) to learn multi-stream feature integration in an end-to-end manner. To verify the efficacy of the proposed network, we conducted exhaustive experiments on the LUNA16 challenge datasets by comparing the performance of the proposed method with state-of-the-art methods in the literature. On two candidate subsets of the LUNA16 dataset, i.e., V1 and V2, our method achieved an average CPM of 0.908 (V1) and 0.942 (V2), outperforming comparable methods by a large margin. Our MGI-CNN is implemented in Python using TensorFlow and the source code is available from https://github.com/ku-milab/MGICNN.

Artificial intelligence in cancer imaging: Clinical challenges and applications

Judgement, as one of the core tenets of medicine, relies upon the integration of multilayered data with nuanced decision making. Cancer offers a unique context for medical decisions given not only its variegated forms with evolution of disease but also the need to take into account the individual condition of patients, their ability to receive treatment, and their responses to treatment. Challenges remain in the accurate detection, characterization, and monitoring of cancers despite improved technologies. Radiographic assessment of disease most commonly relies upon visual evaluations, the interpretations of which may be augmented by advanced computational analyses. In particular, artificial intelligence (AI) promises to make great strides in the qualitative interpretation of cancer imaging by expert clinicians, including volumetric delineation of tumors over time, extrapolation of the tumor genotype and biological course from its radiographic phenotype, prediction of clinical outcome, and assessment of the impact of disease and treatment on adjacent organs. AI may automate processes in the initial interpretation of images and shift the clinical workflow of radiographic detection, management decisions on whether or not to administer an intervention, and subsequent observation to a yet to be envisioned paradigm. Here, the authors review the current state of AI as applied to medical imaging of cancer and describe advances in 4 tumor types (lung, brain, breast, and prostate) to illustrate how common clinical problems are being addressed. Although most studies evaluating AI applications in oncology to date have not been vigorously validated for reproducibility and generalizability, the results do highlight increasingly concerted efforts in pushing AI technology to clinical use and to impact future directions in cancer care.

Using Big Data and Predictive Analytics to Determine Patient Risk in Oncology

Big data and predictive analytics have immense potential to improve risk stratification, particularly in data-rich fields like oncology. This article reviews the literature published on use cases and challenges in applying predictive analytics to improve risk stratification in oncology. We characterized evidence-based use cases of predictive analytics in oncology into three distinct fields: (1) population health management, (2) radiomics, and (3) pathology. We then highlight promising future use cases of predictive analytics in clinical decision support and genomic risk stratification. We conclude by describing challenges in the future applications of big data in oncology, namely (1) difficulties in acquisition of comprehensive data and endpoints, (2) the lack of prospective validation of predictive tools, and (3) the risk of automating bias in observational datasets. If such challenges can be overcome, computational techniques for clinical risk stratification will in short order improve clinical risk stratification for patients with cancer.

eDoctor: machine learning and the future of medicine

Machine learning (ML) is a burgeoning field of medicine with huge resources being applied to fuse computer science and statistics to medical problems. Proponents of ML extol its ability to deal with large, complex and disparate data, often found within medicine and feel that ML is the future for biomedical research, personalized medicine, computer-aided diagnosis to significantly advance global health care. However, the concepts of ML are unfamiliar to many medical professionals and there is untapped potential in the use of ML as a research tool. In this article, we provide an overview of the theory behind ML, explore the common ML algorithms used in medicine including their pitfalls and discuss the potential future of ML in medicine.

Peering Into the Black Box of Artificial Intelligence: Evaluation Metrics of Machine Learning Methods

OBJECTIVE

Machine learning (ML) and artificial intelligence (AI) are rapidly becoming the most talked about and controversial topics in radiology and medicine. Over the past few years, the numbers of ML- or AI-focused studies in the literature have increased almost exponentially, and ML has become a hot topic at academic and industry conferences. However, despite the increased awareness of ML as a tool, many medical professionals have a poor understanding of how ML works and how to critically appraise studies and tools that are presented to us. Thus, we present a brief overview of ML, explain the metrics used in ML and how to interpret them, and explain some of the technical jargon associated with the field so that readers with a medical background and basic knowledge of statistics can feel more comfortable when examining ML applications.

CONCLUSION

Attention to sample size, overfitting, underfitting, cross validation, as well as a broad knowledge of the metrics of machine learning, can help those with little or no technical knowledge begin to assess machine learning studies. However, transparency in methods and sharing of algorithms is vital to allow clinicians to assess these tools themselves.

A new dataset of computed-tomography angiography images for computer-aided detection of pulmonary embolism

The lack of publicly available datasets of computed-tomography angiography (CTA) images for pulmonary embolism (PE) is a problem felt by physicians and researchers. Although a number of computer-aided detection (CAD) systems have been developed for PE diagnosis, their performance is often evaluated using private datasets. In this paper, we introduce a new public dataset called FUMPE (standing for Ferdowsi University of Mashhad's PE dataset) which consists of three-dimensional PE-CTA images of 35 different subjects with 8792 slices in total. For each benchmark image, two expert radiologists provided the ground-truth with the assistance of a semi-automated image processing software tool. FUMPE is a challenging benchmark for CAD methods because of the large number (i.e., 3438) of PE regions and, more especially, because of the location of most of them (i.e., 67%) in lung peripheral arteries. Moreover, due to the reporting of the Qanadli score for each PE-CTA image, FUMPE is the first public dataset which can be used for the analysis of mortality and morbidity risks associated with PE. We also report some complementary prognosis information for each subject.

After Detection: The Improved Accuracy of Lung Cancer Assessment Using Radiologic Computer-aided Diagnosis

RATIONALE AND OBJECTIVES

The aim of this study was to evaluate the improved accuracy of radiologic assessment of lung cancer afforded by computer-aided diagnosis (CADx).

MATERIALS AND METHODS

Inclusion/exclusion criteria were formulated, and a systematic inquiry of research databases was conducted. Following title and abstract review, an in-depth review of 149 surviving articles was performed with accepted articles undergoing a Quality Assessment of Diagnostic Accuracy Studies (QUADAS)-based quality review and data abstraction.

RESULTS

A total of 14 articles, representing 1868 scans, passed the review. Increases in the receiver operating characteristic (ROC) area under the curve of .8 or higher were seen in all nine studies that reported it, except for one that employed subspecialized radiologists.

CONCLUSIONS

This systematic review demonstrated improved accuracy of lung cancer assessment using CADx over manual review, in eight high-quality observer-performance studies. The improved accuracy afforded by radiologic lung-CADx suggests the need to explore its use in screening and regular clinical workflow.

Noninvasive differential diagnosis of dental periapical lesions in cone-beam CT scans

PURPOSE

This paper proposes a novel application of computer-aided diagnosis (CAD) to an everyday clinical dental challenge: the noninvasive differential diagnosis of periapical lesions between periapical cysts and granulomas. A histological biopsy is the most reliable method currently available for this differential diagnosis; however, this invasive procedure prevents the lesions from healing noninvasively despite a report that they may heal without surgical treatment. A CAD using cone-beam computed tomography (CBCT) offers an alternative noninvasive diagnostic tool which helps to avoid potentially unnecessary surgery and to investigate the unknown healing process and rate for the lesions.

METHODS

The proposed semiautomatic solution combines graph-based random walks segmentation with machine learning-based boosted classifiers and offers a robust clinical tool with minimal user interaction. As part of this CAD framework, the authors provide two novel technical contributions: (1) probabilistic extension of the random walks segmentation with likelihood ratio test and (2) LDA-AdaBoost: a new integration of weighted linear discriminant analysis to AdaBoost.

RESULTS

A dataset of 28 CBCT scans is used to validate the approach and compare it with other popular segmentation and classification methods. The results show the effectiveness of the proposed method with 94.1% correct classification rate and an improvement of the performance by comparison with the Simon's state-of-the-art method by 17.6%. The authors also compare classification performances with two independent ground-truth sets from the histopathology and CBCT diagnoses provided by endodontic experts.

CONCLUSIONS

Experimental results of the authors show that the proposed CAD system behaves in clearer agreement with the CBCT ground-truth than with histopathology, supporting the Simon's conjecture that CBCT diagnosis can be as accurate as histopathology for differentiating the periapical lesions.

Optimizing computed tomography pulmonary angiography using right atrium bolus monitoring combined with spontaneous respiration

OBJECTIVES

CT pulmonary angiography (CTPA) aims to provide pulmonary arterial opacification in the absence of significant pulmonary venous filling. This requires accurate timing of the imaging acquisition to ensure synchronization with the peak pulmonary artery contrast concentration. This study was designed to test the utility of right atrium (RA) monitoring in ensuring optimal timing of CTPA acquisition.

METHODS

Sixty patients referred for CTPA were divided into two groups. Group A (n = 30): CTPA was performed using bolus triggering from the pulmonary trunk, suspended respiration and 70 ml of contrast agent (CA). Group B (n = 30): CTPA image acquisition was triggered using RA monitoring with spontaneous respiration and 40 ml of CA. Image quality was compared.

RESULTS

Subjective image quality, average CT values of pulmonary arteries and density difference between artery and vein pairs were significantly higher whereas CT values of pulmonary veins were significantly lower in group B (all P < 0.05). There was no significant difference between the groups in the proportion of subjects where sixth grade pulmonary arteries were opacified (P > 0.05).

CONCLUSIONS

RA monitoring combined with spontaneous respiration to trigger image acquisition in CTPA produces optimal contrast enhancement in pulmonary arterial structures with minimal venous filling even with reduced doses of CA.

KEY POINTS

• Bolus tracking (BT) with pulmonary trunk monitoring is widely used in CTPA. • Pulmonary venous contamination is a disadvantage of BT due to transition delay time. • Right atrium monitoring with spontaneous respiration can optimize CTPA. • It produces optimal contrast enhancement in pulmonary arteries with minimal venous filling. • The contrast dose was significantly reduced.

Computer-aided diagnosis for preoperative invasion depth of gastric cancer with dual-energy spectral CT imaging

RATIONALE AND OBJECTIVES

This study evaluates the accuracy of dual-energy spectral computed tomography (DEsCT) imaging with the aid of computer-aided diagnosis (CAD) system in assessing serosal invasion in patients with gastric cancer.

MATERIALS AND METHODS

Thirty patients with gastric cancer were enrolled in this study. Two types of features (information) were collected with the use of DEsCT imaging: conventional features including patient's clinical information (eg, age, gender) and descriptive characteristics on the CT images (eg, location of the lesion, wall thickness at the gastric cardia) and additional spectral CT features extracted from monochromatic images (eg, 60 keV) and material-decomposition images (eg, iodine- and water-density images). The classification results of the CAD system were compared to pathologic findings. Important features can be found out using support vector machine classification method in combination with feature-selection technique thereby helping the radiologists diagnose better.

RESULTS

Statistical analysis showed that for the collected cases, the feature "long axis" was significantly different between group A (serosa negative) and group B (serosa positive) (P < .05). By adding quantitative spectral features from several regions of interest (ROIs), the total classification accuracy was improved from 83.33% to 90.00%. Two feature ranking algorithms were used in the CAD scheme to derive the top-ranked features. The results demonstrated that low single-energy (approximately 60 keV) CT values, tumor size (long axis and short axis), iodine (water) density, and Effective-Z values of ROIs were important for classification. These findings concurred with the experience of the radiologist.

CONCLUSIONS

The CAD system designed using machine-learning algorithms may be used to improve the identification accuracy in the assessment of serosal invasion in patients of gastric cancer with DEsCT imaging and provide some indicators which may be useful in predicting prognosis.

Evaluation of computer-aided detection and diagnosis systems

Computer-aided detection and diagnosis (CAD) systems are increasingly being used as an aid by clinicians for detection and interpretation of diseases. Computer-aided detection systems mark regions of an image that may reveal specific abnormalities and are used to alert clinicians to these regions during image interpretation. Computer-aided diagnosis systems provide an assessment of a disease using image-based information alone or in combination with other relevant diagnostic data and are used by clinicians as a decision support in developing their diagnoses. While CAD systems are commercially available, standardized approaches for evaluating and reporting their performance have not yet been fully formalized in the literature or in a standardization effort. This deficiency has led to difficulty in the comparison of CAD devices and in understanding how the reported performance might translate into clinical practice. To address these important issues, the American Association of Physicists in Medicine (AAPM) formed the Computer Aided Detection in Diagnostic Imaging Subcommittee (CADSC), in part, to develop recommendations on approaches for assessing CAD system performance. The purpose of this paper is to convey the opinions of the AAPM CADSC members and to stimulate the development of consensus approaches and "best practices" for evaluating CAD systems. Both the assessment of a standalone CAD system and the evaluation of the impact of CAD on end-users are discussed. It is hoped that awareness of these important evaluation elements and the CADSC recommendations will lead to further development of structured guidelines for CAD performance assessment. Proper assessment of CAD system performance is expected to increase the understanding of a CAD system's effectiveness and limitations, which is expected to stimulate further research and development efforts on CAD technologies, reduce problems due to improper use, and eventually improve the utility and efficacy of CAD in clinical practice.

Missed pulmonary emboli on CT angiography: assessment with pulmonary embolism-computer-aided detection

OBJECTIVE

The purpose of this study is to assess the use of a pulmonary embolism (PE)- computer-aided detection (CADx) program in the detection of PE missed in clinical practice.

MATERIALS AND METHODS

Pulmonary CT angiography (CTA) studies (n = 6769) performed between January 2009 and July 2012 were retrospectively assessed by a thoracic radiologist. In studies that were positive for PE, all prior contrast-enhanced pulmonary CTA studies were reviewed. Missed PE was deemed to have occurred if PE was not described in the final interpretation. The presence, proximal extent, and number of PEs were agreed on by three thoracic radiologists. Studies with missed acute PE and available slice thickness of 2 mm or less were assessed with a prototype PE-CADx program. False-positive PE-CADx marks were analyzed. Outcomes of missed acute PEs were assessed in patients with both follow-up imaging and clinical data.

RESULTS

Fifty-three studies with overlooked acute PE met our inclusion criteria for PE-CADx assessment. The PE-CADx program identified at least one PE in 77.4% of instances (41/53). PE-CADx correctly marked at least one PE in 23 of 23 cases (100%) with multiple PEs and 18 of 30 (60%) cases with a solitary PE (p < 0.001). PE-CADx per-study sensitivity was significantly higher for segmental (65.5%) than for subsegmental (91.7%) PEs (p = 0.002). PE-CADx averaged 3.8 false-positive marks per case (range, 0-23 marks). Fourteen patients with missed PE who were not receiving anticoagulation therapy developed new PEs, including nine with an isolated subsegmental PE on the initial CT scan.

CONCLUSION

PE-CADx correctly identified 77.4% of cases of acute PE that were previously missed in clinical practice.

Potential of computer-aided diagnosis to improve CT lung cancer screening

The development of low-dose spiral computed tomography (CT) has rekindled hope that effective lung cancer screening might yet be found. Screening is justified when there is evidence that it will extend lives at reasonable cost and acceptable levels of risk. A screening test should detect all extant cancers while avoiding unnecessary workups. Thus optimal screening modalities have both high sensitivity and specificity. Due to the present state of technology, radiologists must opt to increase sensitivity and rely on follow-up diagnostic procedures to rule out the incurred false positives. There is evidence in published reports that computer-aided diagnosis technology may help radiologists alter the benefit-cost calculus of CT sensitivity and specificity in lung cancer screening protocols. This review will provide insight into the current discussion of the effectiveness of lung cancer screening and assesses the potential of state-of-the-art computer-aided design developments.

Machine learning and radiology

In this paper, we give a short introduction to machine learning and survey its applications in radiology. We focused on six categories of applications in radiology: medical image segmentation, registration, computer aided detection and diagnosis, brain function or activity analysis and neurological disease diagnosis from fMR images, content-based image retrieval systems for CT or MRI images, and text analysis of radiology reports using natural language processing (NLP) and natural language understanding (NLU). This survey shows that machine learning plays a key role in many radiology applications. Machine learning identifies complex patterns automatically and helps radiologists make intelligent decisions on radiology data such as conventional radiographs, CT, MRI, and PET images and radiology reports. In many applications, the performance of machine learning-based automatic detection and diagnosis systems has shown to be comparable to that of a well-trained and experienced radiologist. Technology development in machine learning and radiology will benefit from each other in the long run. Key contributions and common characteristics of machine learning techniques in radiology are discussed. We also discuss the problem of translating machine learning applications to the radiology clinical setting, including advantages and potential barriers.

Optical tomography

The number of applications using optical tomography has significantly increased over the past decade. A literature research providing this term as keyword gives 26 hits for 1990, 719 for 2000, and 9,202 for 2010. With an increasing number of applications, the number of different imaging modalities is also increasing. This review summarizes recent developments in tomographic methods for scattering and nonscattering samples. These two different cases of optical tomography are typically represented by biomedical imaging and atmospheric tomography, representing high- and low-scattering samples, respectively. An essential prerequisite for tomographic analyses is an understanding of light propagation in different media, which allows for the development of specific reconstruction algorithms for the different tomographic tasks.

Image analysis of small pulmonary nodules identified by computed tomography

Detection of small pulmonary nodules has markedly increased as computed tomography (CT) technology has advanced and interpretation evolved from viewing small CT images on film to magnified images on large, high-resolution computer monitors. Despite these advances, determining the etiology of a lung nodule short of major surgery remains problematic. Initial nodule size is a major criterion in evaluating the risk for malignancy, and the majority of CT detected nodules are <10 mm in diameter. Also, the likelihood that the nodule is a lung cancer increases with increasing age and smoking history, and such clinical information needs to be integrated into algorithms that guide the workup of such nodules. Baseline and annual repeat screening results are also very helpful in developing and assessing the usefulness of such algorithms. Based on CT morphology, subtypes of nodules have been identified; today nodules are routinely classified as being solid, part-solid, or nonsolid. It has been shown that part-solid nodules have a higher frequency of being malignant than solid or nonsolid ones. Other nodule characteristics such as spiculation are useful, although granulomas and fibrosis also have such features, so these characteristics have not been as useful as nodule-growth assessment. Depending on the aggressiveness of the lung cancer and the size of the nodule when it is initially seen, a follow-up CT scan 1-3 months after the first CT scan can identify those nodules with growth at a malignant rate. Software has been developed by all CT scanner manufacturers for such growth assessment, but the inherent variability of such assessments needs further development. Nodule-growth assessment based on 2-dimensional approaches is limited; therefore, software has been developed for the 3-dimensional assessment of growth. Different approaches for such growth assessment have been developed, either using automated computer segmentation techniques or hybrid methods that allow the radiologist to adjust such segmentation. There are, however, inherent reasons for variability in such measurements that need to be carefully considered, and this, together with continued technologic advances and integration of the relevant clinical information, will allow for individualization of the algorithms for the workup of small pulmonary nodules.

Computer-aided diagnosis: how to move from the laboratory to the clinic

Computer-aided diagnosis (CAD), encompassing computer-aided detection and quantification, is an established and rapidly growing field of research. In daily practice, however, most radiologists do not yet use CAD routinely. This article discusses how to move CAD from the laboratory to the clinic. The authors review the principles of CAD for lesion detection and for quantification and illustrate the state-of-the-art with various examples. The requirements that radiologists have for CAD are discussed: sufficient performance, no increase in reading time, seamless workflow integration, regulatory approval, and cost efficiency. Performance is still the major bottleneck for many CAD systems. Novel ways of using CAD, extending the traditional paradigm of displaying markers for a second look, may be the key to using the technology effectively. The most promising strategy to improve CAD is the creation of publicly available databases for training and validation. This can identify the most fruitful new research directions, and provide a platform to combine multiple approaches for a single task to create superior algorithms.

Repeatability in computer-aided diagnosis: application to breast cancer diagnosis on sonography

PURPOSE

The aim of this study was to investigate the concept of repeatability in a case-based performance evaluation of two classifiers commonly used in computer-aided diagnosis in the task of distinguishing benign from malignant lesions.

METHODS

The authors performed .632+ bootstrap analyses using a data set of 1251 sonographic lesions of which 212 were malignant. Several analyses were performed investigating the impact of sample size and number of bootstrap iterations. The classifiers investigated were a Bayesian neural net (BNN) with five hidden units and linear discriminant analysis (LDA). Both used the same four input lesion features. While the authors did evaluate classifier performance using receiver operating characteristic (ROC) analysis, the main focus was to investigate case-based performance based on the classifier output for individual cases, i.e., the classifier outputs for each test case measured over the bootstrap iterations. In this case-based analysis, the authors examined the classifier output variability and linked it to the concept of repeatability. Repeatability was assessed on the level of individual cases, overall for all cases in the data set, and regarding its dependence on the case-based classifier output. The impact of repeatability was studied when aiming to operate at a constant sensitivity or specificity and when aiming to operate at a constant threshold value for the classifier output.

RESULTS

The BNN slightly outperformed the LDA with an area under the ROC curve of 0.88 versus 0.85 (p < 0.05). In the repeatability analysis on an individual case basis, it was evident that different cases posed different degrees of difficulty to each classifier as measured by the by-case output variability. When considering the entire data set, however, the overall repeatability of the BNN classifier was lower than for the LDA classifier, i.e., the by-case variability for the BNN was higher. The dependence of the by-case variability on the average by-case classifier output was markedly different for the classifiers. The BNN achieved the lowest variability (best repeatability) when operating at high sensitivity (> 90%) and low specificity (< 66%), while the LDA achieved this at moderate sensitivity (approximately 74%) and specificity (approximately 84%). When operating at constant 90% sensitivity or constant 90% specificity, the width of the 95% confidence intervals for the corresponding classifier output was considerable for both classifiers and increased for smaller sample sizes. When operating at a constant threshold value for the classifier output, the width of the 95% confidence intervals for the corresponding sensitivity and specificity ranged from 9 percentage points (pp) to 30 pp.

CONCLUSIONS

The repeatability of the classifier output can have a substantial effect on the obtained sensitivity and specificity. Knowledge of classifier repeatability, in addition to overall performance level, is important for successful translation and implementation of computer-aided diagnosis in clinical decision making.

A multistage approach to improve performance of computer-aided detection of pulmonary embolisms depicted on CT images: preliminary investigation

This study developed a computer-aided detection (CAD) scheme for pulmonary embolism (PE) detection and investigated several approaches to improve CAD performance. In the study, 20 computed tomography examinations with various lung diseases were selected, which include 44 verified PE lesions. The proposed CAD scheme consists of five basic steps: 1) lung segmentation; 2) PE candidate extraction using an intensity mask and tobogganing region growing; 3) PE candidate feature extraction; 4) false-positive (FP) reduction using an artificial neural network (ANN); and 5) a multifeature-based k-nearest neighbor for positive/negative classification. In this study, we also investigated the following additional methods to improve CAD performance: 1) grouping 2-D detected features into a single 3-D object; 2) selecting features with a genetic algorithm (GA); and 3) limiting the number of allowed suspicious lesions to be cued in one examination. The results showed that 1) CAD scheme using tobogganing, an ANN, and grouping method achieved the maximum detection sensitivity of 79.2%; 2) the maximum scoring method achieved the superior performance over other scoring fusion methods; 3) GA was able to delete "redundant" features and further improve CAD performance; and 4) limiting the maximum number of cued lesions in an examination reduced FP rate by 5.3 times. Combining these approaches, CAD scheme achieved 63.2% detection sensitivity with 18.4 FP lesions per examination. The study suggested that performance of CAD schemes for PE detection depends on many factors that include 1) optimizing the 2-D region grouping and scoring methods; 2) selecting the optimal feature set; and 3) limiting the number of allowed cueing lesions per examination.

Effect of CAD on radiologists' detection of lung nodules on thoracic CT scans: analysis of an observer performance study by nodule size

RATIONALE AND OBJECTIVES

To retrospectively investigate the effect of a computer-aided detection (CAD) system on radiologists' performance for detecting small pulmonary nodules in computed tomography (CT) examinations, with a panel of expert radiologists serving as the reference standard.

MATERIALS AND METHODS

Institutional review board approval was obtained. Our dataset contained 52 CT examinations collected by the Lung Image Database Consortium, and 33 from our institution. All CTs were read by multiple expert thoracic radiologists to identify the reference standard for detection. Six other thoracic radiologists read the CT examinations first without and then with CAD. Performance was evaluated using free-response receiver operating characteristics (FROC) and the jackknife FROC analysis methods (JAFROC) for nodules above different diameter thresholds.

RESULTS

A total of 241 nodules, ranging in size from 3.0 to 18.6 mm (mean, 5.3 mm) were identified as the reference standard. At diameter thresholds of 3, 4, 5, and 6 mm, the CAD system had a sensitivity of 54%, 64%, 68%, and 76%, respectively, with an average of 5.6 false positives (FPs) per scan. Without CAD, the average figures of merit (FOMs) for the six radiologists, obtained from JAFROC analysis, were 0.661, 0.729, 0.793, and 0.838 for the same nodule diameter thresholds, respectively. With CAD, the corresponding average FOMs improved to 0.705, 0.763, 0.810, and 0.862, respectively. The improvement achieved statistical significance for nodules at the 3 and 4 mm thresholds (P = .002 and .020, respectively), and did not achieve significance at 5 and 6 mm (P = .18 and .13, respectively). At a nodule diameter threshold of 3 mm, the radiologists' average sensitivity and FP rate were 0.56 and 0.67, respectively, without CAD, and 0.67 and 0.78 with CAD.

CONCLUSION

CAD improves thoracic radiologists' performance for detecting pulmonary nodules smaller than 5 mm on CT examinations, which are often overlooked by visual inspection alone.

Computer-assisted detection of pulmonary embolism: evaluation of pulmonary CT angiograms performed in an on-call setting

Purpose

The purpose of the study was to assess the stand-alone performance of computer-assisted detection (CAD) for evaluation of pulmonary CT angiograms (CTPA) performed in an on-call setting.

Methods

In this institutional review board-approved study, we retrospectively included 292 consecutive CTPA performed during night shifts and weekends over a period of 16 months. Original reports were compared with a dedicated CAD system for pulmonary emboli (PE). A reference standard for the presence of PE was established using independent evaluation by two readers and consultation of a third experienced radiologist in discordant cases.

Results

Original reports had described 225 negative studies and 67 positive studies for PE. CAD found PE in seven patients originally reported as negative but identified by independent evaluation: emboli were located in segmental (n = 2) and subsegmental arteries (n = 5). The negative predictive value (NPV) of the CAD algorithm was 92% (44/48). On average there were 4.7 false positives (FP) per examination (median 2, range 0–42). In 72% of studies ≤5 FP were found, 13% of studies had ≥10 FP.

Conclusion

CAD identified small emboli originally missed under clinical conditions and found 93% of the isolated subsegmental emboli. On average there were 4.7 FP per examination.

Computer-aided detection of pulmonary embolism in computed tomographic pulmonary angiography (CTPA): performance evaluation with independent data sets

The authors are developing a computer-aided detection system for pulmonary emboli (PE) in computed tomographic pulmonary angiography (CTPA) scans. The pulmonary vessel tree is extracted using a 3D expectation-maximization segmentation method based on the analysis of eigen-values of Hessian matrices at multiple scales. A parallel multiprescreening method is applied to the segmented vessels to identify volume of interests (VOIs) that contained suspicious PE. A linear discriminant analysis (LDA) classifier with feature selection is designed to reduce false positives (FPs). Features that characterize the contrast, gray level, and size of PE are extracted as input predictor variables to the LDA classifier. With the IRB approval, 59 CTPA PE cases were collected retrospectively from the patient files (UM cases). With access permission, 69 CTPA PE cases were randomly selected from the data set of the prospective investigation of pulmonary embolism diagnosis (PIOPED) II clinical trial. Extensive lung parenchymal or pleural diseases were present in 22/59 UM and 26/69 PIOPED cases. Experienced thoracic radiologists manually marked 595 and 800 PE as the reference standards in the UM and PIOPED data sets, respectively. PE occlusion of arteries ranged from 5% to 100%, with PE located from the main pulmonary artery to the subsegmental artery levels. Of the 595 PE identified in the UM cases, 245 and 350 PE were located in the subsegmental arteries and the more proximal arteries, respectively. The detection performance was assessed by free response ROC (FROC) analysis. The FROC analysis indicated that the PE detection system could achieve an overall sensitivity of 80% at 18.9 FPs/case for the PIOPED cases when the LDA classifier was trained with the UM cases. The test sensitivity with the UM cases was 80% at 22.6 FPs/cases when the LDA classifier was trained with the PIOPED cases. The detection performance depended on the arterial level where the PE was located and on the percentage of occlusion. The sensitivity was lower for PE in the subsegmental arteries than in more proximal arteries and was lower for PE with less than 20% occlusion. The results indicate that the PE detection system achieves high sensitivity for PE detection on independent CTPA scans for both the PIOPED and UM data sets and demonstrate the potential that the automated PE detection approach can be generalized to unknown cases.

Computer-aided diagnosis in breast tomosynthesis and chest CT

Computer-aided diagnosis (CAD), in the general sense, includes computer-aided detection and characterization of abnormalities on medical images. The usefulness of CAD for assisting radiologists in detection of breast cancer in screening mammography has been demonstrated by a number of prospective clinical trials in recent years. The development of CAD in other areas is also being actively pursued by researchers. In this talk, the recent work in two areas of CAD, digital breast tomosynthesis (DBT) and chest computed tomography (CT), in the CAD Research Laboratory at the University of Michigan will be reviewed. DBT is a new modality under development for breast imaging. The quasi-3D information in DBT alleviates the problem of overlapping tissue in mammography and holds the promise to improve the sensitivity for cancer detection. DBT image analysis can be performed in the 3D reconstructed volume of the 2D projection view (PV) images. DBT image quality depend on the image acquisition parameters, reconstruction method and parameters. The flexibility in image processing approaches makes CAD development for DBT interesting and challenging. out early experiences in the development of image segmentation and features extraction technique for mass detection and characterization in DBT will be discussed. The performances of the CAD systems using the 2D, 3D, and combined 2D and 3D approaches will be compared. CT has been shown to be superior to chest x-ray in detection of small lung nodules and thus lung cancer screening with CT is still being debated, many research groups are developing CAD methods for detection and characterization of lung nodules in chest CT scans. The specific prescreening, segmentation, and feature extraction techniques designed for our lung nodule detection and characterization systems will be discussed. The effects of CAD on radiologists' accuracy in nodules detection and characterization in CT scans will be demonstrated by results of observer ROC studies. There are similarities in the approaches to developing CAD methods in 3D image volumes such as DBT and CT, these experiences will facilitate the development of CAD systems for other diseases in 3D modalities.

Anniversary paper: History and status of CAD and quantitative image analysis: the role of Medical Physics and AAPM

The roles of physicists in medical imaging have expanded over the years, from the study of imaging systems (sources and detectors) and dose to the assessment of image quality and perception, the development of image processing techniques, and the development of image analysis methods to assist in detection and diagnosis. The latter is a natural extension of medical physicists' goals in developing imaging techniques to help physicians acquire diagnostic information and improve clinical decisions. Studies indicate that radiologists do not detect all abnormalities on images that are visible on retrospective review, and they do not always correctly characterize abnormalities that are found. Since the 1950s, the potential use of computers had been considered for analysis of radiographic abnormalities. In the mid-1980s, however, medical physicists and radiologists began major research efforts for computer-aided detection or computer-aided diagnosis (CAD), that is, using the computer output as an aid to radiologists-as opposed to a completely automatic computer interpretation-focusing initially on methods for the detection of lesions on chest radiographs and mammograms. Since then, extensive investigations of computerized image analysis for detection or diagnosis of abnormalities in a variety of 2D and 3D medical images have been conducted. The growth of CAD over the past 20 years has been tremendous-from the early days of time-consuming film digitization and CPU-intensive computations on a limited number of cases to its current status in which developed CAD approaches are evaluated rigorously on large clinically relevant databases. CAD research by medical physicists includes many aspects-collecting relevant normal and pathological cases; developing computer algorithms appropriate for the medical interpretation task including those for segmentation, feature extraction, and classifier design; developing methodology for assessing CAD performance; validating the algorithms using appropriate cases to measure performance and robustness; conducting observer studies with which to evaluate radiologists in the diagnostic task without and with the use of the computer aid; and ultimately assessing performance with a clinical trial. Medical physicists also have an important role in quantitative imaging, by validating the quantitative integrity of scanners and developing imaging techniques, and image analysis tools that extract quantitative data in a more accurate and automated fashion. As imaging systems become more complex and the need for better quantitative information from images grows, the future includes the combined research efforts from physicists working in CAD with those working on quantitative imaging systems to readily yield information on morphology, function, molecular structure, and more-from animal imaging research to clinical patient care. A historical review of CAD and a discussion of challenges for the future are presented here, along with the extension to quantitative image analysis.