Kyeong-Min Kim

Optimal Clustering of Kinetic Patterns on Malignant Breast Lesions: Comparison between K-means Clustering and Three-time-points Method in Dynamic Contrast-enhanced MRI

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is useful for breast cancer diagnosis and treatment planning. Nevertheless, due to the multi-temporal nature of DCE-MRI data, the assessment of early stage breast cancer is a challenging task. In this study, we applied an unsupervised clustering approach and cluster validation technique to the analysis of malignant intral-tumoral kinetic curves in DCE-MRI. K-means cluster analysis was performed from real world malignant tumor cases and the data were transformed into an optimal number of reference patterns representative each cluster. The optimal number of clusters was estimated by a cluster validation index, which was calculated with the ratio of inter-class scatter to intra-class scatter. This technique then classifies tumor specific patterns from a given MRI data by measuring the vector distances from the reference pattern set, and compared the result from the k- means clustering with that from three-time-points (3TP) method, which represents a clinical standard protocol for analysis of tumor kinetics. The evaluation of twenty five cases indicates that optimal k-means clustering reflects partitioning intra-tumoral kinetic patterns better than the 3TP technique. This method will greatly enhance the capability of radiologists to identify and characterize internal kinetic heterogeneity and vascular change of a tumor in breast DCE-MRI.

Coregistration of small animal PET and autoradiography for in vivo-ex vivo comparison

Non-invasive imaging of small animal and its quantification techniques are needed to be evaluated by comparison with ex vivo image. To overcome the existing method, hand-operated comparison with the unaided eye observation, we present an algorithm that matches the target area in PET scanned in vivo with an autoradiography image acquired ex vivo. We applied a coregistration algorithm that uses voxel similarity to find the corresponding slices and to make quantitative measurements. Automatic matching of in vivo and ex vivo images is novel, and can provide better validation than manual matching techniques.

Tumor detection from small animal PET using clustering based on intensity and connectivity

We present an efficient clustering method for detecting the tumor in positron emission tomography(PET) of the tumor bearing small animal. We used iterative threshold method to remove the background noise and then we applied two clustering procedures in order. The one is clustering method based on intensity to segment the tumor region and the other is clustering based on connectivity to remove false positive region from the segmented region. The tumor tissue looks bright in the image compared to surrounding normal tissue because of glucose uptake. Therefore, based on volume intensity, we divided all elements of the image into several clusters, the tumor, living bodies, background using improved fuzzy c-means clustering(FCM). Using FCM with the sorted initial mean of each cluster gets out of the wrong optimization and reduces the amount of time-consumed. However, not only the tumor tissue, but also the other organs like heart, bladder can also have high intensity value because of glucose metabolism. So in order to separate the tumor and false positive region, we applied geometric clustering based on connectivity. Proposed segmentation method can lead a robust analysis of the tumor growth with the aid of the quantitative measurements such the tumor size or volume.