Automated Storage and Retrieval of Thin-Section CT Images to Assist Diagnosis: System Description and Preliminary Assessment

Aisen, Alex M.; Broderick, Lynn S.; Winer-Muram, Helen T.; Brodley, Carla E.; Kak, Avinash C.; Pavlopoulou, Christina; Dy, Jennifer G.; Shyu, Chi‐Ren; Marchiori, Alan

doi:10.1148/radiol.2281020126

Cited by 145 publications

(78 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Clinical evaluation of the ASSERT CBIR system for high resolution CT lung images [100] showed an improvement in the accuracy of the diagnosis made by physicians [22]. Another study for liver CT concluded that CBIR could be used to provide real-time decision support [25].…”

Section: Medical Content-based Image Retrievalmentioning

confidence: 99%

“…Clinical evaluation has demonstrated that the accuracy of a physician's diagnosis can be improved by using a CBIR system to display images similar to the undiagnosed query; the improvement in diagnostic accuracy was largest for less experienced physicians [22]. CBIR has also been shown to have benefits in radiology education [23] by allowing users to retrieve teaching files that consisted of multiple images of a particular clinical case [24].…”

Section: Content-based Image Retrieval (Cbir) Is An Image Search Techmentioning

confidence: 99%

See 1 more Smart Citation

A graph-based approach for the retrieval of multi-modality medical images

Kumar

Kim

Wen

et al. 2014

Medical Image Analysis

View full text Add to dashboard Cite

Advances in sensors and imaging technologies are contributing to rapidly expanding data repositories that contain interrelated information from different modalities. The extraction and visualisation of knowledge from these repositories is a major challenge in the modern, digital world. In the medical domain, images are routinely acquired for a variety of tasks, including diagnosis and patient monitoring. Advances in imaging technologies have resulted in devices capable of acquiring images in multiple dimensions (volumetric and dynamic) as well as from multiple modalities. One example of a widely used volumetric and multimodality image is combined positron emission tomography and computed tomography (PET-CT), which presents physicians with complementary functional and anatomical features and spatial relationships. In clinical practice, PET-CT imaging has already proven its ability to improve cancer diagnosis, localisation, and staging compared to its single-modality counterparts.The clinical benefits provided by medical imaging have spurred increases in the data volume acquired in clinical environments. As such, massive medical imaging collections offer the opportunity for search-based applications in evidence-based diagnosis, physician training, and biomedical research. However, conventional search techniques that operate upon manually assigned textual annotations are not feasible for the volume of data acquired in modern hospitals. Qualitative text descriptions are also limited in their capacity to quantitatively describe the rich information inherent in medical images. Content-based image retrieval (CBIR) is an image search technique that utilises visual features as search criteria. CBIR has already demonstrated benefits forevidence-based diagnosis, physician training, and biomedical research by allowing clinical staff to consider relevant knowledge from retrieved cases. The majority of medical CBIR research has focused on single modality medical images leaving a clear deficiency in the retrieval of multi-modality images. In particular, images iii like PET-CT offer the ability for retrieval based upon the relationships between regions in different modalities, such as the location of tumour features (from PET) in relation to organ features (from CT). The challenge of multi-modality image retrieval for cancer patients lies in representing these complementary geometric and topologic attributes between tumours and organs. A secondary challenge lies in the human aspect of retrieval -effectively communicating the retrieved results to users and facilitating a better understanding of the similarity between the query and retrieved multi-modality images.As such, in this thesis we propose a new graph representation for multimodality images. Our representation preserves the spatial relationships between modalities by emphasising the inherent characteristics of these images that are used for disease staging and classification. This is done by structurally constraining the graph based on image features, e.g., spatia...

show abstract

Section: Medical Content-based Image Retrievalmentioning

confidence: 99%

Section: Content-based Image Retrieval (Cbir) Is An Image Search Techmentioning

confidence: 99%

A graph-based approach for the retrieval of multi-modality medical images

Kumar

Kim

Wen

et al. 2014

Medical Image Analysis

View full text Add to dashboard Cite

show abstract

“…Content-based image retrieval has been an active area of research in the medical imaging field with many applications [72][73][74] including pathology. 75,76 Toward that goal, images containing certain desired color stains would be used to train the algorithm.…”

Section: Commentmentioning

confidence: 99%

Quantitative Assessment and Classification of Tissue-Based Biomarker Expression With Color Content Analysis

Keller¹,

Chen²,

Gavrielides³

2012

Archives of Pathology &Amp; Laboratory Medicine

View full text Add to dashboard Cite

N Context.-The use of computer aids has been suggested as a way to reduce interobserver variability that is known to exist in the interpretation of immunohistochemical staining in pathology. Such computer aids should be automated in their usage but also they should be trained in an automated and reproducible fashion.Objective.-To present a computer aid for the quantitative analysis of tissue-based biomarkers, based on color content analysis.Design.-The developed system incorporates an automated algorithm to allow retraining based on the color properties of different training sets. The algorithm first generates a color palette containing the colors present in a training subset. Based on the palette, color histograms are derived and are used as feature vectors to a pattern recognition system, which returns an output proportional to biomarker continuous expression or a categorical classification. The method was evaluated on a database of HER2/neu digital breast cancer slides, for which expression scores from a pathologist panel were available. The system was retrained and evaluated on different transformations of the database, including compression, blurring, and changes in illumination, to examine its robustness to different imaging conditions frequently met in digital pathology.Results.-Results showed high agreement between the results of the algorithm and the truth from the pathologist panel as well as robustness to image transformations.Conclusions.-The results of the study are encouraging for the potential of this method as a computer aid to assess biomarker expression in a consistent and reproducible manner.

show abstract

“…Liver lesion diagnosis is particularly challenging, owing to a wide range of appearances of benign and malignant lesions [1]. Studies have shown that medical decision support systems relying on content-based image retrieval (CBIR) may provide improvement in efficiency and accuracy of diagnosis [2]. While CBIR has gained much popularity in non-medical applications [3,4], a great deal of work remains to be done in the medical field.…”

Section: Introductionmentioning

confidence: 99%

Modeling Perceptual Similarity Measures in CT Images of Focal Liver Lesions

et al. 2012

View full text Add to dashboard Cite

Motivation: A gold standard for perceptual similarity in medical images is vital to content-based image retrieval, but inter-reader variability complicates development. Our objective was to develop a statistical model that predicts the number of readers (N) necessary to achieve acceptable levels of variability. Materials and Methods: We collected 3 radiologists' ratings of the perceptual similarity of 171 pairs of CT images of focal liver lesions rated on a 9-point scale. We modeled the readers' scores as bimodal distributions in additive Gaussian noise and estimated the distribution parameters from the scores using an expectation maximization algorithm. We (a) sampled 171 similarity scores to simulate a ground truth and (b) simulated readers by adding noise, with standard deviation between 0 and 5 for each reader. We computed the mean values of 2-50 readers' scores and calculated the agreement (AGT) between these means and the simulated ground truth, and the inter-reader agreement (IRA), using Cohen's Kappa metric. Results: IRA for the empirical data ranged from =0.41 to 0.66. For between 1.5 and 2.5, IRA between three simulated readers was comparable to agreement in the empirical data. For these values , AGT ranged from =0.81 to 0.91. As expected, AGT increased with N, ranging from =0.83 to 0.92 for N = 2 to 50, respectively, with =2. Conclusion: Our simulations demonstrated that for moderate to good IRA, excellent AGT could nonetheless be obtained. This model may be used to predict the required N to accurately evaluate similarity in arbitrary size datasets.

show abstract

Automated Storage and Retrieval of Thin-Section CT Images to Assist Diagnosis: System Description and Preliminary Assessment

Cited by 145 publications

References 27 publications

A graph-based approach for the retrieval of multi-modality medical images

A graph-based approach for the retrieval of multi-modality medical images

Quantitative Assessment and Classification of Tissue-Based Biomarker Expression With Color Content Analysis

Modeling Perceptual Similarity Measures in CT Images of Focal Liver Lesions

Contact Info

Product

Resources

About