Predicting respiratory tumor motion with multi-dimensional adaptive filters and support vector regression

Nadeem Riaz, Piyush Shanker, Rodney Wiersma, Olafur Gudmundsson, Weihua Mao, Bernard Widrow, Lei Xing

Research output: Contribution to journalArticle

51 Citations (Scopus)

Abstract

Intra-fraction tumor tracking methods can improve radiation delivery during radiotherapy sessions. Image acquisition for tumor tracking and subsequent adjustment of the treatment beam with gating or beam tracking introduces time latency and necessitates predicting the future position of the tumor. This study evaluates the use of multi-dimensional linear adaptive filters and support vector regression to predict the motion of lung tumors tracked at 30 Hz. We expand on the prior work of other groups who have looked at adaptive filters by using a general framework of a multiple-input single-output (MISO) adaptive system that uses multiple correlated signals to predict the motion of a tumor. We compare the performance of these two novel methods to conventional methods like linear regression and single-input, single-output adaptive filters. At 400 ms latency the average root-mean-square-errors (RMSEs) for the 14 treatment sessions studied using no prediction, linear regression, single-output adaptive filter, MISO and support vector regression are 2.58, 1.60, 1.58, 1.71 and 1.26 mm, respectively. At 1 s, the RMSEs are 4.40, 2.61, 3.34, 2.66 and 1.93 mm, respectively. We find that support vector regression most accurately predicts the future tumor position of the methods studied and can provide a RMSE of less than 2 mm at 1 s latency. Also, a multi-dimensional adaptive filter framework provides improved performance over single-dimension adaptive filters. Work is underway to combine these two frameworks to improve performance.

Original languageEnglish (US)
Pages (from-to)5735-5748
Number of pages14
JournalPhysics in Medicine and Biology
Volume54
Issue number19
DOIs
StatePublished - 2009

Fingerprint

Neoplasms
Linear Models
Radiotherapy
Radiation
Lung
Therapeutics

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Radiological and Ultrasound Technology

Cite this

Predicting respiratory tumor motion with multi-dimensional adaptive filters and support vector regression. / Riaz, Nadeem; Shanker, Piyush; Wiersma, Rodney; Gudmundsson, Olafur; Mao, Weihua; Widrow, Bernard; Xing, Lei.

In: Physics in Medicine and Biology, Vol. 54, No. 19, 2009, p. 5735-5748.

Research output: Contribution to journalArticle

Riaz, N, Shanker, P, Wiersma, R, Gudmundsson, O, Mao, W, Widrow, B & Xing, L 2009, 'Predicting respiratory tumor motion with multi-dimensional adaptive filters and support vector regression', Physics in Medicine and Biology, vol. 54, no. 19, pp. 5735-5748. https://doi.org/10.1088/0031-9155/54/19/005
Riaz, Nadeem ; Shanker, Piyush ; Wiersma, Rodney ; Gudmundsson, Olafur ; Mao, Weihua ; Widrow, Bernard ; Xing, Lei. / Predicting respiratory tumor motion with multi-dimensional adaptive filters and support vector regression. In: Physics in Medicine and Biology. 2009 ; Vol. 54, No. 19. pp. 5735-5748.
@article{1f560c6dddef4505ab2d9dabace9889b,
title = "Predicting respiratory tumor motion with multi-dimensional adaptive filters and support vector regression",
abstract = "Intra-fraction tumor tracking methods can improve radiation delivery during radiotherapy sessions. Image acquisition for tumor tracking and subsequent adjustment of the treatment beam with gating or beam tracking introduces time latency and necessitates predicting the future position of the tumor. This study evaluates the use of multi-dimensional linear adaptive filters and support vector regression to predict the motion of lung tumors tracked at 30 Hz. We expand on the prior work of other groups who have looked at adaptive filters by using a general framework of a multiple-input single-output (MISO) adaptive system that uses multiple correlated signals to predict the motion of a tumor. We compare the performance of these two novel methods to conventional methods like linear regression and single-input, single-output adaptive filters. At 400 ms latency the average root-mean-square-errors (RMSEs) for the 14 treatment sessions studied using no prediction, linear regression, single-output adaptive filter, MISO and support vector regression are 2.58, 1.60, 1.58, 1.71 and 1.26 mm, respectively. At 1 s, the RMSEs are 4.40, 2.61, 3.34, 2.66 and 1.93 mm, respectively. We find that support vector regression most accurately predicts the future tumor position of the methods studied and can provide a RMSE of less than 2 mm at 1 s latency. Also, a multi-dimensional adaptive filter framework provides improved performance over single-dimension adaptive filters. Work is underway to combine these two frameworks to improve performance.",
author = "Nadeem Riaz and Piyush Shanker and Rodney Wiersma and Olafur Gudmundsson and Weihua Mao and Bernard Widrow and Lei Xing",
year = "2009",
doi = "10.1088/0031-9155/54/19/005",
language = "English (US)",
volume = "54",
pages = "5735--5748",
journal = "Physics in Medicine and Biology",
issn = "0031-9155",
publisher = "IOP Publishing Ltd.",
number = "19",

}

TY - JOUR

T1 - Predicting respiratory tumor motion with multi-dimensional adaptive filters and support vector regression

AU - Riaz, Nadeem

AU - Shanker, Piyush

AU - Wiersma, Rodney

AU - Gudmundsson, Olafur

AU - Mao, Weihua

AU - Widrow, Bernard

AU - Xing, Lei

PY - 2009

Y1 - 2009

N2 - Intra-fraction tumor tracking methods can improve radiation delivery during radiotherapy sessions. Image acquisition for tumor tracking and subsequent adjustment of the treatment beam with gating or beam tracking introduces time latency and necessitates predicting the future position of the tumor. This study evaluates the use of multi-dimensional linear adaptive filters and support vector regression to predict the motion of lung tumors tracked at 30 Hz. We expand on the prior work of other groups who have looked at adaptive filters by using a general framework of a multiple-input single-output (MISO) adaptive system that uses multiple correlated signals to predict the motion of a tumor. We compare the performance of these two novel methods to conventional methods like linear regression and single-input, single-output adaptive filters. At 400 ms latency the average root-mean-square-errors (RMSEs) for the 14 treatment sessions studied using no prediction, linear regression, single-output adaptive filter, MISO and support vector regression are 2.58, 1.60, 1.58, 1.71 and 1.26 mm, respectively. At 1 s, the RMSEs are 4.40, 2.61, 3.34, 2.66 and 1.93 mm, respectively. We find that support vector regression most accurately predicts the future tumor position of the methods studied and can provide a RMSE of less than 2 mm at 1 s latency. Also, a multi-dimensional adaptive filter framework provides improved performance over single-dimension adaptive filters. Work is underway to combine these two frameworks to improve performance.

AB - Intra-fraction tumor tracking methods can improve radiation delivery during radiotherapy sessions. Image acquisition for tumor tracking and subsequent adjustment of the treatment beam with gating or beam tracking introduces time latency and necessitates predicting the future position of the tumor. This study evaluates the use of multi-dimensional linear adaptive filters and support vector regression to predict the motion of lung tumors tracked at 30 Hz. We expand on the prior work of other groups who have looked at adaptive filters by using a general framework of a multiple-input single-output (MISO) adaptive system that uses multiple correlated signals to predict the motion of a tumor. We compare the performance of these two novel methods to conventional methods like linear regression and single-input, single-output adaptive filters. At 400 ms latency the average root-mean-square-errors (RMSEs) for the 14 treatment sessions studied using no prediction, linear regression, single-output adaptive filter, MISO and support vector regression are 2.58, 1.60, 1.58, 1.71 and 1.26 mm, respectively. At 1 s, the RMSEs are 4.40, 2.61, 3.34, 2.66 and 1.93 mm, respectively. We find that support vector regression most accurately predicts the future tumor position of the methods studied and can provide a RMSE of less than 2 mm at 1 s latency. Also, a multi-dimensional adaptive filter framework provides improved performance over single-dimension adaptive filters. Work is underway to combine these two frameworks to improve performance.

UR - http://www.scopus.com/inward/record.url?scp=70350649031&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70350649031&partnerID=8YFLogxK

U2 - 10.1088/0031-9155/54/19/005

DO - 10.1088/0031-9155/54/19/005

M3 - Article

C2 - 19729711

AN - SCOPUS:70350649031

VL - 54

SP - 5735

EP - 5748

JO - Physics in Medicine and Biology

JF - Physics in Medicine and Biology

SN - 0031-9155

IS - 19

ER -