PubMed Health⌕ Search

Biomedical subjects

Sanjay Ranka

Publications and source records attributed to Sanjay Ranka.

8 recordsLinked to original sources

Markers improve clustering of CGH data.

MOTIVATION: We consider the problem of clustering a population of Comparative Genomic Hybridization (CGH) data samples using similarity based clustering methods. A key requirement for clustering is to avoid using the noisy aberrations in the CGH samples. RESULTS: We develop a dynamic programming algorithm to identify a small set of important genomic intervals called markers. The advantage of using these markers is that the potentially noisy genomic intervals are excluded during the clustering process. We also develop two clustering strategies using these markers. The first one, prototype-based approach, maximizes the support for the markers. The second one, similarity-based approach, develops a new similarity measure called RSim and refines clusters with the aim of maximizing the RSim measure between the samples in the same cluster. Our results demonstrate that the markers we found represent the aberration patterns of cancer types well and they improve the quality of clustering significantly. AVAILABILITY: All software developed in this paper and all the datasets used are available from the authors upon request.

Biomarkers, Tumor↗

Distance-based clustering of CGH data.

MOTIVATION: We consider the problem of clustering a population of Comparative Genomic Hybridization (CGH) data samples. The goal is to develop a systematic way of placing patients with similar CGH imbalance profiles into the same cluster. Our expectation is that patients with the same cancer types will generally belong to the same cluster as their underlying CGH profiles will be similar. RESULTS: We focus on distance-based clustering strategies. We do this in two steps. (1) Distances of all pairs of CGH samples are computed. (2) CGH samples are clustered based on this distance. We develop three pairwise distance/similarity measures, namely raw, cosine and sim. Raw measure disregards correlation between contiguous genomic intervals. It compares the aberrations in each genomic interval separately. The remaining measures assume that consecutive genomic intervals may be correlated. Cosine maps pairs of CGH samples into vectors in a high-dimensional space and measures the angle between them. Sim measures the number of independent common aberrations. We test our distance/similarity measures on three well known clustering algorithms, bottom-up, top-down and k-means with and without centroid shrinking. Our results show that sim consistently performs better than the remaining measures. This indicates that the correlation of neighboring genomic intervals should be considered in the structural analysis of CGH datasets. The combination of sim with top-down clustering emerged as the best approach. AVAILABILITY: All software developed in this article and all the datasets are available from the authors upon request. CONTACT: juliu@cise.ufl.edu.

Algorithms↗

A comparison of step-and-shoot leaf sequencing algorithms that eliminate tongue-and-groove effects.

The performances of three recently published leaf sequencing algorithms for step-and-shoot intensity-modulated radiation therapy delivery that eliminates tongue-and-groove underdosage are evaluated. Proofs are given to show that the algorithm of Que et al (2004 Phys. Med. Biol. 49 399-405) generates leaf sequences free of tongue-and-groove underdosage and interdigitation. However, the total beam-on times could be up to n times those of the sequences generated by the algorithms of Kamath et al (2004 Phys. Med. Biol. 49 N7-N19), which are optimal in beam-on time for unidirectional leaf movement under the same constraints, where n is the total number of involved leaf pairs. Using 19 clinical fluence matrices and 100000 randomly generated 15 x 15 matrices, the average monitor units and number of segments of the leaf sequences generated using the algorithm of Que et al are about two to four times those generated by the algorithm of Kamath et al.

Algorithms↗

Optimal leaf sequencing with elimination of tongue-and-groove underdosage.

The individual leaves of a multileaf collimator (MLC) have a tongue-and-groove or stepped-edge design to minimize leakage radiation between adjacent leaves. This design element has a drawback in that it creates areas of underdosages in intensity-modulated photon beams unless a leaf trajectory is specifically designed such that for any two adjacent leaf pairs, the direct exposure under the tongue-and-groove is equal to the lower of the direct exposures of the leaf pairs. In this work, we present a systematic study of the optimization of a leaf sequencing algorithm for segmental multileaf collimator beam delivery that completely eliminates areas of underdosages due to tongue-and-groove or stepped-edge design of the MLC. Simultaneous elimination of tongue-and-groove effect and leaf interdigitation is also studied. This is an extension of our previous work (Kamath et al 2003a Phys. Med. Biol. 48 307) in which we described a leaf sequencing algorithm that is optimal for monitor unit (MU) efficiency under most common leaf movement constraints that include minimum leaf separation. Compared to our previously published algorithm (without constraints), the new algorithms increase the number of sub-fields by approximately 21% and 25%, respectively, but are optimal in MU efficiency for unidirectional schedules.

Algorithms↗

Algorithms for optimal sequencing of dynamic multileaf collimators.

Dynamic multileaf collimator (DMLC) intensity modulated radiation therapy (IMRT) is used to deliver intensity modulated beams using a multileaf collimator (MLC), with the leaves in motion. DMLC-IMRT requires the conversion of a radiation intensity map into a leaf sequence file that controls the movement of the MLC while the beam is on. It is imperative that the intensity map delivered using the leaf sequence file be as close as possible to the intensity map generated by the dose optimization algorithm, while satisfying hardware constraints of the delivery system. Optimization of the leaf-sequencing algorithm has been the subject of several recent investigations. In this work, we present a systematic study of the optimization of leaf-sequencing algorithms for dynamic multileaf collimator beam delivery and provide rigorous mathematical proofs of optimized leaf sequence settings in terms of monitor unit (MU) efficiency under the most common leaf movement constraints that include leaf interdigitation constraint. Our analytical analysis shows that leaf sequencing based on unidirectional movement of the MLC leaves is as MU efficient as bi-directional movement of the MLC leaves.

Algorithms↗

Optimal field splitting for large intensity-modulated fields.

The multileaf travel range limitations on some linear accelerators require the splitting of a large intensity-modulated field into two or more adjacent abutting intensity-modulated subfields. The abutting subfields are then delivered as separate treatment fields. This workaround not only increases the treatment delivery time but it also increases the total monitor units (MU) delivered to the patient for a given prescribed dose. It is imperative that the cumulative intensity map of the subfields is exactly the same as the intensity map of the large field generated by the dose optimization algorithm, while satisfying hardware constraints of the delivery system. In this work, we describe field splitting algorithms that split a large intensity-modulated field into two or more intensity-modulated subfields with and without feathering, with optimal MU efficiency while satisfying the hardware constraints. Compared to a field splitting technique (without feathering) used in a commercial planning system, our field splitting algorithm (without feathering) shows a decrease in total MU of up to 26% on clinical cases and up to 63% on synthetic cases.

Algorithms↗

Leaf sequencing algorithms for segmented multileaf collimation.

The delivery of intensity-modulated radiation therapy (IMRT) with a multileaf collimator (MLC) requires the conversion of a radiation fluence map into a leaf sequence file that controls the movement of the MLC during radiation delivery. It is imperative that the fluence map delivered using the leaf sequence file is as close as possible to the fluence map generated by the dose optimization algorithm, while satisfying hardware constraints of the delivery system. Optimization of the leaf sequencing algorithm has been the subject of several recent investigations. In this work, we present a systematic study of the optimization of leaf sequencing algorithms for segmental multileaf collimator beam delivery and provide rigorous mathematical proofs of optimized leaf sequence settings in terms of monitor unit (MU) efficiency under most common leaf movement constraints that include minimum leaf separation constraint and leaf interdigitation constraint. Our analytical analysis shows that leaf sequencing based on unidirectional movement of the MLC leaves is as MU efficient as bidirectional movement of the MLC leaves.

Motion↗

Characterization of a Class of Sigmoid Functions with Applications to Neural Networks.

We study two classes of sigmoids: the simple sigmoids, defined to be odd, asymptotically bounded, completely monotone functions in one variable, and the hyperbolic sigmoids, a proper subset of simple sigmoids and a natural generalization of the hyperbolic tangent. We obtain a complete characterization for the inverses of hyperbolic sigmoids using Euler's incomplete beta functions, and describe composition rules that illustrate how such functions may be synthesized from others. These results are applied to two problems. First we show that with respect to simple sigmoids the continuous Cohen-Grossberg-Hopfield model can be reduced to the (associated) Legendre differential equations. Second, we show that the effect of using simple sigmoids as node transfer functions in a one-hidden layer feedforward network with one summing output may be interpreted as representing the output function as a Fourier series sine transform evaluated at the hidden layer node inputs, thus extending and complementing earlier results in this area. Copyright 1996 Elsevier Science Ltd

Journal Article↗