Computer-aided Detection Algorithm Research Articles

S212 Comparing Polyp Detection Rate Between Water Exchange and Air Insufflation by a Computer-aided Detection Algorithm: An Analysis of Withdrawal Phase Videos from a Randomized Controlled Trial

Introduction: Water exchange (WE) improves polyp (PDR) and adenoma detection rate (ADR) compared with air insufflation [GIE 2021;93:1411], but still misses polyps due to human limitations [JCG 2021, in press]. A computer-aided detection (CADe) algorithm can overcome human omissions, but is limited by false positives (FPs) [Lancet Gastroenterol & Hepatol 2019;4:71], which might be reduced by salvage-cleaning effect provided by insertion WE [Tzu-Chi Med J 2021;33:108]. We hypothesized that the strengths of CADe and WE can complement the weaknesses of each other. Their combination can enhance PDR and ADR and reduce false positive rate (FPR). Methods: We used a locally developed convolutional neural network with YOLOv4 to build a CADe model. Images of polyps (n=15961) were divided in 80% and 20% for training and testing. Prior to real-time application, we validated the model in detecting polyps, using edited and coded withdrawal phase videos of the right colon from a RCT comparing right colon adenoma detection between WE and air insufflation. The polyps detected by the CADe were confirmed by the consensus of 2 blinded reviewers (experienced endoscopists). We also developed a computer-aided diagnosis (CADx) model based on YOLOv4 to characterize the detected polyps as adenomas or not. The primary and secondary outcomes were PDR and the FPR, respectively. Results: A total of 245 coded videos of colonoscopies inserted with WE (n=123) or air insufflation (n=122) were analyzed. CADe achieved significantly higher PDR in the WE group (61.8%) than in the air insufflation group (40.2%), P=0.0009. CADe also had significantly higher additional PDR in the WE group (30.1% vs 12.3%, P=0.001). The CADx assessed ADR was significantly higher in the WE group than that in the air insufflation group (46.3% vs 30.3%, P=0.018), after the detected polyps were characterized by the CADx. (Table a) The mean numbers of FPs related to feces [1.78 (1.67) vs 2.09 (2.09), P=0.007] and bubbles [0.53 (0.89) vs 1.25 (2.45), P=0.001] were significantly lower in the WE group than in the air insufflation group. (Table b). Conclusion: The data confirmed the hypothesis that the strengths of CADe and WE complemented the weaknesses of each other, enhanced PDR and ADR and reduced FPR. The locally developed CADe and CADx are validated for polyp detection and adenoma assignment for real-time application. Furthermore, WE provides an optimal platform for the computer-assisted algorithms.Table 1.: ADR, adenoma detection rate; All the percentages calculated according to respective total number of patients for each group; PDR, polyp detection rate; Total numbers of polyps were based on those found by the performing endoscopist plus those detected by the CADe; Proportion of additional polyps found by CADe equals numbers of additional polyp (adenomas) found by CADe divided by total numbers of polyps (adenomas). Others: ileocecal valve, cecum, polypectomy site, suction marks and drug pills. † Mann-Whitney test; ††Fisher exact test.

Present Limitations of Artificial Intelligence in the Emergency Setting - Performance Study of a Commercial, Computer-Aided Detection Algorithm for Pulmonary Embolism.

Since artificial intelligence is transitioning from an experimental stage to clinical implementation, the aim of our study was to evaluate the performance of a commercial, computer-aided detection algorithm of computed tomography pulmonary angiograms regarding the presence of pulmonary embolism in the emergency room. This retrospective study includes all pulmonary computed tomography angiogram studies performed in a large emergency department over a period of 36months that were analyzed by two radiologists experienced in emergency radiology to set a reference standard. Original reports and computer-aided detection results were compared regarding the detection of lobar, segmental, and subsegmental pulmonary embolism. All computer-aided detection findings were analyzed concerning the underlying pathology. False-positive findings were correlated to the contrast-to-noise ratio. Expert reading revealed pulmonary embolism in 182 of 1229 patients (49 % men, 10-97 years) with a total of 504emboli. The computer-aided detection algorithm reported 3331 findings, including 258 (8 %) true-positive findings and 3073 (92 %) false-positive findings. Computer-aided detection analysis showed a sensitivity of 47 % (95 %CI: 33-61 %) on the lobar level and 50 % (95 %CI 43-56 %) on the subsegmental level. On average, there were 2.25 false-positive findings per study (median 2, range 0-25). There was no significant correlation between the number of false-positive findings and the contrast-to-noise ratio (Spearman's Rank Correlation Coefficient = 0.09). Soft tissue (61.0 %) and pulmonary veins (24.1 %) were the most common underlying reasons for false-positive findings. Applied to a population at a large emergency room, the tested commercial computer-aided detection algorithm faced relevant performance challenges that need to be addressed in future development projects. · Computed tomography pulmonary angiograms are frequently acquired in emergency radiology.. · Computer-aided detection algorithms (CADs) can support image analysis.. · CADs face challenges regarding false-positive and false-negative findings.. · Radiologists using CADs need to be aware of these limitations.. · Further software improvements are necessary ahead of implementation in the daily routine.. · Müller-Peltzer K, Kretzschmar L, Negrão de Figueiredo G et al. Present Limitations of Artificial Intelligence in the Emergency Setting - Performance Study of a Commercial, Computer-Aided Detection Algorithm for Pulmonary Embolism. Fortschr Röntgenstr 2021; 193: 1436 - 1444.

End-to-end prostate cancer detection in bpMRI via 3D CNNs: Effects of attention mechanisms, clinical priori and decoupled false positive reduction

We present a multi-stage 3D computer-aided detection and diagnosis (CAD) model22Source code and algorithm have been made publicly available at: https://github.com/DIAGNijmegen/prostateMR_3D-CAD-csPCa https://grand-challenge.org/algorithms/prostate-mri-cad-cspca for automated localization of clinically significant prostate cancer (csPCa) in bi-parametric MR imaging (bpMRI). Deep attention mechanisms drive its detection network, targeting salient structures and highly discriminative feature dimensions across multiple resolutions. Its goal is to accurately identify csPCa lesions from indolent cancer and the wide range of benign pathology that can afflict the prostate gland. Simultaneously, a decoupled residual classifier is used to achieve consistent false positive reduction, without sacrificing high sensitivity or computational efficiency. In order to guide model generalization with domain-specific clinical knowledge, a probabilistic anatomical prior is used to encode the spatial prevalence and zonal distinction of csPCa. Using a large dataset of 1950 prostate bpMRI paired with radiologically-estimated annotations, we hypothesize that such CNN-based models can be trained to detect biopsy-confirmed malignancies in an independent cohort.For 486 institutional testing scans, the 3D CAD system achieves 83.69±5.22% and 93.19±2.96% detection sensitivity at 0.50 and 1.46 false positive(s) per patient, respectively, with 0.882±0.030 AUROC in patient-based diagnosis –significantly outperforming four state-of-the-art baseline architectures (U-SEResNet, UNet++, nnU-Net, Attention U-Net) from recent literature. For 296 external biopsy-confirmed testing scans, the ensembled CAD system shares moderate agreement with a consensus of expert radiologists (76.69%; kappa = 0.51±0.04) and independent pathologists (81.08%; kappa = 0.56±0.06); demonstrating strong generalization to histologically-confirmed csPCa diagnosis.

Computer-Aided Detection False Positives in Colonoscopy.

Randomized control trials and meta-analyses comparing colonoscopies with and without computer-aided detection (CADe) assistance showed significant increases in adenoma detection rates (ADRs) with CADe. A major limitation of CADe is its false positives (FPs), ranked 3rd in importance among 59 research questions in a modified Delphi consensus review. The definition of FPs varies. One commonly used definition defines an FP as an activation of the CADe system, irrespective of the number of frames or duration of time, not due to any polypoid or nonpolypoid lesions. Although only 0.07 to 0.2 FPs were observed per colonoscopy, video analysis studies using FPs as the primary outcome showed much higher numbers of 26 to 27 per colonoscopy. Most FPs were of short duration (91% < 0.5 s). A higher number of FPs was also associated with suboptimal bowel preparation. The appearance of FPs can lead to user fatigue. The polypectomy of FPs results in increased procedure time and added use of resources. Re-training the CADe algorithms is one way to reduce FPs but is not practical in the clinical setting during colonoscopy. Water exchange (WE) is an emerging method that the colonoscopist can use to provide salvage cleaning during insertion. We discuss the potential of WE for reducing FPs as well as the augmentation of ADRs through CADe.

SFCN: Symmetric feature comparison network for detecting ischemic stroke lesions on CT images

Ischemic stroke is the most common stroke and the leading cause of disability and death in the world. Computed tomography (CT) is a popular and economical diagnostic device for the stroke, However the ischemic stroke lesions are not evident on CT images and the diagnostic result relies on the visual observation of neurologists, which may vary from doctor to doctor. To facilitate the treatment, a computer-aided detection algorithm on CT images is proposed to help clinician for the ischemic stroke screening. In order to obtain accurate lesion annotation on CT images, novel automatic algorithms are developed to achieve image pairing, calibration, and registration. Then, a new framework with the symmetric feature extraction and comparison is proposed to identify and locate the ischemic stroke lesion. Experimental results show that this method achieves 75% of DICE in the detection of ischemic stroke lesions, which is higher than other methods by 4%. Its competitive results compared with seven latest methods is shown in terms of extensive qualitative and quantitative evaluation. This method can accurately detect the lesion in the CT images through the comparison of symmetric regional features, which has contributed to the clinical diagnosis of ischemic stroke.

Integrating old and new complexity measures toward automated seizure detection from long-term video EEG recordings

SummaryAutomated seizure detection in long-term video-EEG recordings is far from being integrated into common clinical practice. Here, we leverage classical and state-of-the-art complexity measures to robustly and automatically detect seizures from scalp recordings. Brain activity is scored through eight features, encompassing traditional time domain and novel measures of recurrence. A binary classification algorithm tailored to treat unbalanced dataset is used to determine whether a time window is ictal or non-ictal from its features. The application of the algorithm on a cohort of ten adult patients with focal refractory epilepsy indicates sensitivity, specificity, and accuracy of 90%, along with a true alarm rate of 95% and less than four false alarms per day. The proposed approach emphasizes ictal patterns against noisy background without the need of data preprocessing. Finally, we benchmark our approach against previous studies on two publicly available datasets, demonstrating the good performance of our algorithm.

External Evaluation of 3 Commercial Artificial Intelligence Algorithms for Independent Assessment of Screening Mammograms

A computer algorithm that performs at or above the level of radiologists in mammography screening assessment could improve the effectiveness of breast cancer screening. To perform an external evaluation of 3 commercially available artificial intelligence (AI) computer-aided detection algorithms as independent mammography readers and to assess the screening performance when combined with radiologists. This retrospective case-control study was based on a double-reader population-based mammography screening cohort of women screened at an academic hospital in Stockholm, Sweden, from 2008 to 2015. The study included 8805 women aged 40 to 74 years who underwent mammography screening and who did not have implants or prior breast cancer. The study sample included 739 women who were diagnosed as having breast cancer (positive) and a random sample of 8066 healthy controls (negative for breast cancer). Positive follow-up findings were determined by pathology-verified diagnosis at screening or within 12 months thereafter. Negative follow-up findings were determined by a 2-year cancer-free follow-up. Three AI computer-aided detection algorithms (AI-1, AI-2, and AI-3), sourced from different vendors, yielded a continuous score for the suspicion of cancer in each mammography examination. For a decision of normal or abnormal, the cut point was defined by the mean specificity of the first-reader radiologists (96.6%). The median age of study participants was 60 years (interquartile range, 50-66 years) for 739 women who received a diagnosis of breast cancer and 54 years (interquartile range, 47-63 years) for 8066 healthy controls. The cases positive for cancer comprised 618 (84%) screen detected and 121 (16%) clinically detected within 12 months of the screening examination. The area under the receiver operating curve for cancer detection was 0.956 (95% CI, 0.948-0.965) for AI-1, 0.922 (95% CI, 0.910-0.934) for AI-2, and 0.920 (95% CI, 0.909-0.931) for AI-3. At the specificity of the radiologists, the sensitivities were 81.9% for AI-1, 67.0% for AI-2, 67.4% for AI-3, 77.4% for first-reader radiologist, and 80.1% for second-reader radiologist. Combining AI-1 with first-reader radiologists achieved 88.6% sensitivity at 93.0% specificity (abnormal defined by either of the 2 making an abnormal assessment). No other examined combination of AI algorithms and radiologists surpassed this sensitivity level. To our knowledge, this study is the first independent evaluation of several AI computer-aided detection algorithms for screening mammography. The results of this study indicated that a commercially available AI computer-aided detection algorithm can assess screening mammograms with a sufficient diagnostic performance to be further evaluated as an independent reader in prospective clinical trials. Combining the first readers with the best algorithm identified more cases positive for cancer than combining the first readers with second readers.

Prospective development and validation of a volumetric laser endomicroscopy computer algorithm for detection of Barrett’s neoplasia

Volumetric laser endomicroscopy (VLE) is an advanced imaging modality used to detect Barrett's esophagus (BE) dysplasia. However, real-time interpretation of VLE scans is complex and time-consuming. Computer-aided detection (CAD) may help in the process of VLE image interpretation. Our aim was to train and validate a CAD algorithm for VLE-based detection of BE neoplasia. The multicenter, VLE PREDICT study, prospectively enrolled 47 patients with BE. In total, 229 nondysplastic BE and 89 neoplastic (high-grade dysplasia/esophageal adenocarcinoma) targets were laser marked under VLE guidance and subsequently underwent a biopsy for histologic diagnosis. Deep convolutional neural networks were used to construct a CAD algorithm for differentiation between nondysplastic and neoplastic BE tissue. The CAD algorithm was trained on a set consisting of the first 22 patients (134 nondysplastic BE and 38 neoplastic targets) and validated on a separate test set from patients 23 to 47 (95 nondysplastic BE and 51 neoplastic targets). The performance of the algorithm was benchmarked against the performance of 10 VLE experts. Using the training set to construct the algorithm resulted in an accuracy of 92%, sensitivity of 95%, and specificity of 92%. When performance was assessed on the test set, accuracy, sensitivity, and specificity were 85%, 91%, and 82%, respectively. The algorithm outperformed all 10 VLE experts, who demonstrated an overall accuracy of 77%, sensitivity of 70%, and specificity of 81%. We developed, validated, and benchmarked a VLE CAD algorithm for detection of BE neoplasia using prospectively collected and biopsy-correlated VLE targets. The algorithm detected neoplasia with high accuracy and outperformed 10 VLE experts. (The Netherlands National Trials Registry (NTR) number: NTR 6728.).

Abstract A12: Analysis of advanced quantitative computed tomography imaging features in predicting the surgical resectability of advanced epithelial ovarian cancer

Abstract Background: Ovarian cancer is the fifth leading cause of cancer death in the United States in large part because more than 85% of patients present with metastatic disease. The majority of these are epithelial of high-grade serous (HGSOC) histology where the 5-year overall survival remains poor despite advances in chemotherapy and improved understanding of molecular tumor biology. As a heterogeneous, p53-driven malignancy, HGSOC employs a variety of molecular and metabolomic changes that promote preferential tumor growth within the peritoneal cavity. The miliary spread patterns of ovarian cancers within the peritoneal cavity create distinct clinical challenges, particularly in the preoperative assessment of surgical resectability. Therefore, it is clinically imperative to develop better imaging modalities capable of accurately predicting location, volume, and consistency of tumor implants that will inform treatment planning without exposing the patient to unnecessary harm. To this end, recent research efforts have focused on identifying novel clinical and imaging biomarkers capable of predicting optimal surgical resection, tumor response, and disease specific outcomes. Unfortunately, the clinical utility of such markers in individualizing patient specific treatment strategies remains limited. In the absence of unique molecular biomarkers, the use of advanced quantitative imaging features that utilize novel computer-aided detection (CAD) algorithms and can train artificial intelligence platforms will improve the accuracy with which current imaging modalities predict surgical outcome in HGSOC. Results: Using an established imaging platform established at Tempus Labs, Inc. we evaluated preoperative images of women undergoing cytoreductive surgery for advanced epithelial ovarian cancer. The CT images of forty-five retrospectively collected cases, with complete clinical outcomes data, are undergoing analysis. Lesions have been identified, contoured from CT images, and approved by a trained radiologist. Over 2,000 quantitative imaging (radiomic) features will be extracted from these images, and using the Tempus proprietary imaging platform, feature selection will be performed to identify factors predictive of surgical resectability and compare them to those that do not correlate with surgical outcome. These selected features will be used to model surgical outcomes data. Conclusions: Prior research has demonstrated the utility of quantitative CT image analysis in predicting response to chemotherapy and platinum sensitivity. This same technology, however, has not been applied to the surgical management of women with advanced HGSOC. In this study, we assessed the utility of similar “radiomics algorithms” in order to establish a set of radiomic features that characterize tumors amenable to optimal (R0) surgical resection. The selected features identified in this retrospective training cohort are now be tested in a larger prospective cohort of metastatic HGSOC cancer patients. Citation Format: Paul A. DiSilvestro, Jessica DiSilvestro, Jonathan S. Ball, Stephen Yip, Anna Schwarzbach, Katherine M. Moxley. Analysis of advanced quantitative computed tomography imaging features in predicting the surgical resectability of advanced epithelial ovarian cancer [abstract]. In: Proceedings of the AACR Special Conference on Advances in Ovarian Cancer Research; 2019 Sep 13-16, 2019; Atlanta, GA. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(13_Suppl):Abstract nr A12.

A 2-Phase Merge Filter Approach to Computer-Aided Detection of Breast Tumors on 3-Dimensional Ultrasound Imaging.

The role of image analysis in 3-dimensional (3D) automated breast ultrasound (ABUS) images is increasingly important because of its widespread use as a screening tool in whole-breast examinations. However, reviewing a large number of images acquired from ABUS is time-consuming and sometimes error prone. The aim of this study, therefore, was to develop an efficient computer-aided detection (CADe) algorithm to assist the review process. The proposed CADe algorithm consisted of 4 major steps. First, initial tumor candidates were formed by extracting and merging hypoechoic square cells on 2-dimensional (2D) transverse images. Second, a feature-based classifier was then constructed using 2D features to filter out nontumor candidates. Third, the remaining 2D candidates were merged longitudinally into 3D masses. Finally, a 3D feature-based classifier was used to further filter out nontumor masses to obtain the final detected masses. The proposed method was validated with 176 passes of breast images acquired by an Acuson S2000 automated breast volume scanner (Siemens Medical Solutions USA, Inc., Malvern, PA), including 44 normal passes and 132 abnormal passes containing 162 proven lesions (79 benign and 83 malignant). The proposed CADe system could achieve overall sensitivity of 100% and 90% with 6.71 and 5.14 false-positives (FPs) per pass, respectively. Our results also showed that the average number of FPs per normal pass (7.16) was more than the number of FPs per abnormal pass (6.56) at 100% sensitivity. The proposed CADe system has a great potential for becoming a good companion tool with ABUS imaging by ensuring high sensitivity with a relatively small number of FPs.

Artificial intelligence for breast cancer detection in mammography and digital breast tomosynthesis: State of the art.

Screening for breast cancer with mammography has been introduced in various countries over the last 30 years, initially using analog screen-film-based systems and, over the last 20 years, transitioning to the use of fully digital systems. With the introduction of digitization, the computer interpretation of images has been a subject of intense interest, resulting in the introduction of computer-aided detection (CADe) and diagnosis (CADx) algorithms in the early 2000's. Although they were introduced with high expectations, the potential improvement in the clinical realm failed to materialize, mostly due to the high number of false positive marks per analyzed image. In the last five years, the artificial intelligence (AI) revolution in computing, driven mostly by deep learning and convolutional neural networks, has also pervaded the field of automated breast cancer detection in digital mammography and digital breast tomosynthesis. Research in this area first involved comparison of its capabilities to that of conventional CADe/CADx methods, which quickly demonstrated the potential of this new technology. In the last couple of years, more mature and some commercial products have been developed, and studies of their performance compared to that of experienced breast radiologists are showing that these algorithms are on par with human-performance levels in retrospective data sets. Although additional studies, especially prospective evaluations performed in the real screening environment, are needed, it is becoming clear that AI will have an important role in the future breast cancer screening realm. Exactly how this new player will shape this field remains to be determined, but recent studies are already evaluating different options for implementation of this technology. The aim of this review is to provide an overview of the basic concepts and developments in the field AI for breast cancer detection in digital mammography and digital breast tomosynthesis. The pitfalls of conventional methods, and how these are, for the most part, avoided by this new technology, will be discussed. Importantly, studies that have evaluated the current capabilities of AI and proposals for how these capabilities should be leveraged in the clinical realm will be reviewed, while the questions that need to be answered before this vision becomes a reality are posed.

Combination of generative adversarial network and convolutional neural network for automatic subcentimeter pulmonary adenocarcinoma classification.

The efficient and accurate diagnosis of pulmonary adenocarcinoma before surgery is of considerable significance to clinicians. Although computed tomography (CT) examinations are widely used in practice, it is still challenging and time-consuming for radiologists to distinguish between different types of subcentimeter pulmonary nodules. Although there have been many deep learning algorithms proposed, their performance largely depends on vast amounts of data, which is difficult to collect in the medical imaging area. Therefore, we propose an automatic classification system for subcentimeter pulmonary adenocarcinoma, combining a convolutional neural network (CNN) and a generative adversarial network (GAN) to optimize clinical decision-making and to provide small dataset algorithm design ideas. A total of 206 nodules with postoperative pathological labels were analyzed. Among them were 30 adenocarcinomas in situ (AISs), 119 minimally invasive adenocarcinomas (MIAs), and 57 invasive adenocarcinomas (IACs). Our system consisted of two parts, a GAN-based image synthesis, and a CNN classification. First, several popular existing GAN techniques were employed to augment the datasets, and comprehensive experiments were conducted to evaluate the quality of the GAN synthesis. Additionally, our classification system processes were based on two-dimensional (2D) nodule-centered CT patches without the need of manual labeling information. For GAN-based image synthesis, the visual Turing test showed that even radiologists could not tell the GAN-synthesized from the raw images (accuracy: primary radiologist 56%, senior radiologist 65%). For CNN classification, our progressive growing wGAN improved the performance of CNN most effectively (area under the curve =0.83). The experiments indicated that the proposed GAN augmentation method improved the classification accuracy by 23.5% (from 37.0% to 60.5%) and 7.3% (from 53.2% to 60.5%) in comparison with training methods using raw and common augmented images respectively. The performance of this combined GAN and CNN method (accuracy: 60.5%±2.6%) was comparable to the state-of-the-art methods, and our CNN was also more lightweight. The experiments revealed that GAN synthesis techniques could effectively alleviate the problem of insufficient data in medical imaging. The proposed GAN plus CNN framework can be generalized for use in building other computer-aided detection (CADx) algorithms and thus assist in diagnosis.

Polyp detection algorithm can detect small polyps: Ex vivo reading test compared with endoscopists.

Small polyps are occasionally missed during colonoscopy. This study was conducted to validate the diagnostic performance of a polyp-detection algorithm to alert endoscopists to unrecognized lesions. A computer-aided detection (CADe) algorithm was developed based on convolutional neural networks using training data from 1991 still colonoscopy images from 283 subjects with adenomatous polyps. The CADe algorithm was evaluated on a validation dataset including 50 short videos with 1-2 polyps (3.5±1.5mm, range 2-8mm) and 50 videos without polyps. Two expert colonoscopists and two physicians in training separately read the same videos, blinded to the presence of polyps. The CADe algorithm was also evaluated using eight full videos with polyps and seven full videos without a polyp. The per-video sensitivity of CADe for polyp detection was 88% and the per-frame false-positive rate was 2.8%, with a confidence level of ≥30%. The per-video sensitivity of both experts was 88%, and the sensitivities of the two physicians in training were 84% and 76%. For each reader, the frames with missed polyps appearing on short videos were significantly less than the frames with detected polyps, but no trends were observed regarding polyp size, morphology or color. For full video readings, per-polyp sensitivity was 100% with a per-frame false-positive rate of 1.7%, and per-frame specificity of 98.3%. The sensitivity of CADe to detect small polyps was almost equivalent to experts and superior to physicians in training. A clinical trial using CADe is warranted.

Deep principal dimension encoding for the classification of early neoplasia in Barrett's Esophagus with volumetric laser endomicroscopy

Barrett cancer is a treatable disease when detected at an early stage. However, current screening protocols are often not effective at finding the disease early. Volumetric Laser Endomicroscopy (VLE) is a promising new imaging tool for finding dysplasia in Barrett's esophagus (BE) at an early stage, by acquiring cross-sectional images of the microscopic structure of BE up to 3-mm deep. However, interpretation of VLE scans is difficult for medical doctors due to both the size and subtlety of the gray-scale data. Therefore, algorithms that can accurately find cancerous regions are very valuable for the interpretation of VLE data. In this study, we propose a fully-automatic multi-step Computer-Aided Detection (CAD) algorithm that optimally leverages the effectiveness of deep learning strategies by encoding the principal dimension in VLE data. Additionally, we show that combining the encoded dimensions with conventional machine learning techniques further improves results while maintaining interpretability. Furthermore, we train and validate our algorithm on a new histopathologically validated set of in-vivo VLE snapshots. Additionally, an independent test set is used to assess the performance of the model. Finally, we compare the performance of our algorithm against previous state-of-the-art systems. With the encoded principal dimension, we obtain an Area Under the Curve (AUC) and F1 score of 0.93 and 87.4% on the test set respectively. We show this is a significant improvement compared to the state-of-the-art of 0.89 and 83.1%, respectively, thereby demonstrating the effectiveness of our approach.

Computer-aided polyp detection based on image enhancement and saliency-based selection

This paper presents a computer-aided polyp detection algorithm applicable to both colonoscopy and wireless capsule endoscopy (WCE). The proposed system has three integral parts: image enhancement, saliency map formation and Histogram of gradients (HOG) feature extraction for final classification. We propose a novel and efficient image enhancement algorithm, which enhances the saliency of clinically important features in endoscopic images. A saliency detection method is applied to the enhanced images to highlight the initial polyp candidates. In the classification stage, polyp candidates are selected after performing an image enhancement step and a saliency detection step. Exhaustive experiments have been performed on three publicly available databases: CVC ColonDB, CVC ClinicDB, and ETIS Larib to evaluate the performance of the proposed polyp detection algorithm. Comparison with the state-of-the-art methods shows that the proposed method outperforms the existing ones in terms of recall (=86.33%) and F2 score (=75.51%) for CVC ColonDB and in terms of recall (=74.04%) for the ETIS Larib dataset. With a significantly reduced number of search windows resulting from the saliency-based selection, the proposed scheme ensures a cost-effective and efficient polyp detection algorithm.

Computer-Aided Diagnosis of Pulmonary Fibrosis Using Deep Learning and CT Images.

ObjectivesThe objective of this study is to assess the performance of a computer-aided diagnosis (CAD) system (INTACT system) for the automatic classification of high-resolution computed tomography images into 4 radiological diagnostic categories and to compare this with the performance of radiologists on the same task.Materials and MethodsFor the comparison, a total of 105 cases of pulmonary fibrosis were studied (54 cases of nonspecific interstitial pneumonia and 51 cases of usual interstitial pneumonia). All diagnoses were interstitial lung disease board consensus diagnoses (radiologically or histologically proven cases) and were retrospectively selected from our database. Two subspecialized chest radiologists made a consensual ground truth radiological diagnosis, according to the Fleischner Society recommendations. A comparison analysis was performed between the INTACT system and 2 other radiologists with different years of experience (readers 1 and 2). The INTACT system consists of a sequential pipeline in which first the anatomical structures of the lung are segmented, then the various types of pathological lung tissue are identified and characterized, and this information is then fed to a random forest classifier able to recommend a radiological diagnosis.ResultsReader 1, reader 2, and INTACT achieved similar accuracy for classifying pulmonary fibrosis into the original 4 categories: 0.6, 0.54, and 0.56, respectively, with P > 0.45. The INTACT system achieved an F-score (harmonic mean for precision and recall) of 0.56, whereas the 2 readers, on average, achieved 0.57 (P = 0.991). For the pooled classification (2 groups, with and without the need for biopsy), reader 1, reader 2, and CAD had similar accuracies of 0.81, 0.70, and 0.81, respectively. The F-score was again similar for the CAD system and the radiologists. The CAD system and the average reader reached F-scores of 0.80 and 0.79 (P = 0.898).ConclusionsWe found that a computer-aided detection algorithm based on machine learning was able to classify idiopathic pulmonary fibrosis with similar accuracy to a human reader.

Improved Cancer Detection Using Artificial Intelligence: a Retrospective Evaluation of Missed Cancers on Mammography

To determine whether cmAssist™, an artificial intelligence-based computer-aided detection (AI-CAD) algorithm, can be used to improve radiologists’ sensitivity in breast cancer screening and detection. A blinded retrospective study was performed with a panel of seven radiologists using a cancer-enriched data set from 122 patients that included 90 false-negative mammograms obtained up to 5.8 years prior to diagnosis and 32 BIRADS 1 and 2 patients with a 2-year follow-up of negative diagnosis. The mammograms were performed between February 7, 2008 (earliest) and January 8, 2016 (latest), and were all originally interpreted as negative in conjunction with R2 ImageChecker CAD, version 10.0. In this study, the readers analyzed the 122 studies before and after review of cmAssist™, an AI-CAD software for mammography. The statistical significance of our findings was evaluated using Student’s t test and bootstrap statistical analysis. There was a substantial and significant improvement in radiologist accuracy with use of cmAssist, as demonstrated in the 7.2% increase in the area-under-the-curve (AUC) of the receiver operating characteristic (ROC) curve with two-sided p value < 0.01 for the reader group. All radiologists showed a significant improvement in their cancer detection rate (CDR) with the use of cmAssist (two-sided p value = 0.030, confidence interval = 95%). The readers detected between 25 and 71% (mean 51%) of the early cancers without assistance. With cmAssist, the overall reader CDR was 41 to 76% (mean 62%). The percentage increase in CDR for the reader panel was significant, ranging from 6 to 64% (mean 27%) with the use of cmAssist. There was less than 1% increase in the readers’ false-positive recalls with use of cmAssist. With the use of cmAssist TM, there was a substantial and statistically significant improvement in radiologists’ accuracy and sensitivity for detection of cancers that were originally missed. The percentage increase in CDR for the radiologists in the reader panel ranged from 6 to 64% (mean 27%) with the use of cmAssist, with negligible increase in false-positive recalls.

Visual Interpretation of Convolutional Neural Network Predictions in Classifying Medical Image Modalities.

Deep learning (DL) methods are increasingly being applied for developing reliable computer-aided detection (CADe), diagnosis (CADx), and information retrieval algorithms. However, challenges in interpreting and explaining the learned behavior of the DL models hinders their adoption and use in real-world systems. In this study, we propose a novel method called “Class-selective Relevance Mapping” (CRM) for localizing and visualizing discriminative regions of interest (ROI) within a medical image. Such visualizations offer improved explanation of the convolutional neural network (CNN)-based DL model predictions. We demonstrate CRM effectiveness in classifying medical imaging modalities toward automatically labeling them for visual information retrieval applications. The CRM is based on linear sum of incremental mean squared errors (MSE) calculated at the output layer of the CNN model. It measures both positive and negative contributions of each spatial element in the feature maps produced from the last convolution layer leading to correct classification of an input image. A series of experiments on a “multi-modality” CNN model designed for classifying seven different types of image modalities shows that the proposed method is significantly better in detecting and localizing the discriminative ROIs than other state of the art class-activation methods. Further, to visualize its effectiveness we generate “class-specific” ROI maps by averaging the CRM scores of images in each modality class, and characterize the visual explanation through their different size, shape, and location for our multi-modality CNN model that achieved over 98% performance on a dataset constructed from publicly available images.

Validation of lesion simulations in clinical CT data for anonymized chest and abdominal CT databases.

To make available to the medical imaging community a computed tomography (CT) image database composed of hybrid datasets (patient CT images with digitally inserted anthropomorphic lesions) where lesion ground truth is known a priori. It is envisioned that such a dataset could be a resource for the assessment of CT image quality, machine learning, and imaging technologies [e.g.,computer aided detection (CAD) and segmentation algorithms]. This HIPPA compliant, IRB waiver of approval study consisted of utilizing 120 chest and 100 abdominal clinically acquired adult CT exams. One image series per patient exam was utilized based on coverage of the anatomical region of interest (either the thorax or abdomen). All image series were de-identified. Simulated lesions were derived from a library of anatomically informed digital lesions (93 lung and 50 liver lesions) where six and four digital lesions with nominal diameters ranging from 4 to 20mm were inserted into lung and liver image series, respectively. Locations for lesion insertion were randomly chosen. A previously validated lesion simulation and virtual insertion technique were utilized. The resulting hybrid images were reviewed by three experienced radiologists to assure similarity with routine clinical imaging in a diverse adult population. The database is composed of four datasets that contain 100 patient cases each, for a total of 400 image series accompanied by Matlab.mat tables that provide descriptive information about the virtually inserted lesions (i.e., size, shape, opacity, and insertion location in physical (world) coordinates and voxel indices). All image and metadata are stored in DICOM format on the Quantitative Imaging Data Warehouse (https://qidw.rsna.org/#collection/57d463471cac0a4ec8ff8f46/folder/5b23dceb1cac0a4ec800a770?dialog=login), in two sets: (a) QIBA CT Hybrid Dataset I which contains Lung I and Liver I datasets, and (b) QIBA CT Hybrid Dataset II which contains Lung II and Liver II datasets. The QIDW is supported by the Radiological Society of North America (RSNA). Registration is required upon initial log in. By simulating lesion opacity (full solid, part solid and ground glass), size, and texture, the relationship between lesion morphology and segmentation or CAD algorithm performance can be investigated without the need for repetitive patient exams. This database can also serve as a reference standard for device and reader performance studies.

Deep learning and conditional random fields-based depth estimation and topographical reconstruction from conventional endoscopy

Colorectal cancer is the fourth leading cause of cancer deaths worldwide and the second leading cause in the United States. The risk of colorectal cancer can be mitigated by the identification and removal of premalignant lesions through optical colonoscopy. Unfortunately, conventional colonoscopy misses more than 20% of the polyps that should be removed, due in part to poor contrast of lesion topography. Imaging depth and tissue topography during a colonoscopy is difficult because of the size constraints of the endoscope and the deforming mucosa. Most existing methods make unrealistic assumptions which limits accuracy and sensitivity. In this paper, we present a method that avoids these restrictions, using a joint deep convolutional neural network-conditional random field (CNN-CRF) framework for monocular endoscopy depth estimation. Estimated depth is used to reconstruct the topography of the surface of the colon from a single image. We train the unary and pairwise potential functions of a CRF in a CNN on synthetic data, generated by developing an endoscope camera model and rendering over 200,000 images of an anatomically-realistic colon.We validate our approach with real endoscopy images from a porcine colon, transferred to a synthetic-like domain via adversarial training, with ground truth from registered computed tomography measurements. The CNN-CRF approach estimates depths with a relative error of 0.152 for synthetic endoscopy images and 0.242 for real endoscopy images. We show that the estimated depth maps can be used for reconstructing the topography of the mucosa from conventional colonoscopy images. This approach can easily be integrated into existing endoscopy systems and provides a foundation for improving computer-aided detection algorithms for detection, segmentation and classification of lesions.