3D Ultrasound Spine Image Selection Using Convolution Learning-to-Rank Algorithm.

3D Ultrasound imaging has become an important means of scoliosis assessment as it is a real-time, cost-effective and radiation-free imaging technique. However, the coronal images from different depths of a 3D ultrasound image have different imaging definitions. So there is a need to select the coronal image that would give the best image definition. Also, manual selection of coronal images is time-consuming and limited to the discretion and capability of the assessor. Therefore, in this paper, we have developed a convolution learning-to-rank algorithm to select the best ultrasound images automatically using raw ultrasound images. The ranking is done based on the curve angle of the spinal cord. Firstly, we approached the image selection problem as a ranking problem; ranked based on probability of an image to be a good image. Here, we use the RankNet, a pairwise learning-to-rank method, to rank the images automatically. Secondly, we replaced the backbone of the RankNet, which is the traditional artificial neural network (ANN), with convolution neural network (CNN) to improve the feature extracting ability for the successive iterations. The experimental result shows that the proposed convolutional RankNet achieves the perfect accuracy of 100% while conventional DenseNet achieved 35% only. This proves that the convolutional RankNet is more suitable to highlight the best quality of ultrasound image from multiple mediocre ones.

Leveraging prior quantitative knowledge to guide drug development decisions and regulatory science recommendations: impact of FDA pharmacometrics during 2004-2006.

The End-of-Phase 2A meetings are proposed to identify opportunities to make innovative medical products available sooner and to increase the quality of drug applications through early meetings between sponsors and the FDA. This article summarizes the overall experience across 11 pilot End-of-Phase 2A meetings since 2004. Four case studies are presented in more detail to demonstrate the various issues and methods encountered at these meetings. Overall, industry and FDA scientists ranked these meetings to be "very helpful" (average score of 4 on a scale of 1 to 5). In almost all the instances the sponsors changed their drug development plans subsequent to these extensive quantitative analyses-based meetings. A draft Guidance is being developed to be issued in 2008, and we hope this initiative will be resourced by then.

Accumulated chromosomal instability in murine bone marrow mesenchymal stem cells leads to malignant transformation.

Despite recent emerging evidence suggesting that cancer stem cells subsist in a variety of tumors, it is not yet fully elucidated whether postnatal stem cells are directly involved in tumorigenesis. We used murine bone marrow-derived mesenchymal stem cells (BMMSCs) as a model to test a hypothesis that tumorigenesis may originate from spontaneous mutation of stem cells. In this study, we demonstrated that murine BMMSCs, after numerous passages, obtained unlimited population doublings and proceeded to a malignant transformation state, resulting in fibrosarcoma formation in vivo. Transformed BMMSCs colonized to multiple organs when delivered systemically through the tail vein. Fibrosarcoma cells formed by transformed BMMSCs contained cancer progenitors, which were capable of generating colony clusters in vitro and fibrosarcoma in vivo by the second administration. The mechanism by which BMMSCs transformed to malignant cells was associated with accumulated chromosomal abnormalities, gradual elevation in telomerase activity, and increased c-myc expression. Moreover, BMMSCs and their transformed counterpart, fibrosarcoma-forming cells, demonstrated different sensitivity to anti-cancer drugs. BMMSCs/fibrosarcoma transformation system may provide an ideal system to elucidate the mechanism of how stem cells become cancer cells and to screen anti-sarcoma drugs.