A state-based probabilistic model for tumor respiratory motion prediction

Alan Kalet, George Sandison, Huanmei Wu, Ruth Schmitz

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

This work proposes a new probabilistic mathematical model for predicting tumormotion and position based on a finite state representation using the natural breathing states of exhale, inhale and end of exhale. Tumor motion was broken down into linear breathing states and sequences of states. Breathing state sequences and the observables representing those sequences were analyzed using a hidden Markov model (HMM) to predict the future sequences and new observables. Velocities and other parameters were clustered using a k-means clustering algorithm to associate each state with a set of observables such that a prediction of state also enables a prediction of tumor velocity. A time average model with predictions based on average past state lengths was also computed. State sequences which are known a priori to fit the data were fed into the HMM algorithm to set a theoretical limit of the predictive power of the model. The effectiveness of the presented probabilistic model has been evaluated for gated radiation therapy based on previously tracked tumor motion in four lung cancer patients. Positional prediction accuracy is compared with actual position in terms of the overall RMS errors. Various system delays, ranging from 33 to 1000 ms, were tested. Previous studies have shown duty cycles for latencies of 33 and 200 ms at around 90% and 80%, respectively, for linear, no prediction, Kalman filter and ANN methods as averaged over multiple patients. At 1000 ms, the previously reported duty cycles range from approximately 62% (ANN) down to 34% (no prediction). Average duty cycle for the HMMmethod was found to be 100% and 91±3% for 33 and 200 ms latency and around 40% for 1000 ms latency in three out of four breathing motion traces. RMS errors were found to be lower than linear and no prediction methods at latencies of 1000 ms. The results show that for system latencies longer than 400 ms, the time average HMM prediction outperforms linear, no prediction, and the more general HMM-type predictive models. RMS errors for the time average model approach the theoretical limit of the HMM, and predicted state sequences are well correlated with sequences known to fit the data.

Original languageEnglish
Pages (from-to)7615-7631
Number of pages17
JournalPhysics in Medicine and Biology
Volume55
Issue number24
DOIs
StatePublished - Dec 21 2010

Fingerprint

Statistical Models
Respiration
Neoplasms
Theoretical Models
Cluster Analysis
Lung Neoplasms
Radiotherapy

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Radiological and Ultrasound Technology

Cite this

A state-based probabilistic model for tumor respiratory motion prediction. / Kalet, Alan; Sandison, George; Wu, Huanmei; Schmitz, Ruth.

In: Physics in Medicine and Biology, Vol. 55, No. 24, 21.12.2010, p. 7615-7631.

Research output: Contribution to journalArticle

Kalet, Alan ; Sandison, George ; Wu, Huanmei ; Schmitz, Ruth. / A state-based probabilistic model for tumor respiratory motion prediction. In: Physics in Medicine and Biology. 2010 ; Vol. 55, No. 24. pp. 7615-7631.
@article{b34cdde61fff41389d6f62855aef60c2,
title = "A state-based probabilistic model for tumor respiratory motion prediction",
abstract = "This work proposes a new probabilistic mathematical model for predicting tumormotion and position based on a finite state representation using the natural breathing states of exhale, inhale and end of exhale. Tumor motion was broken down into linear breathing states and sequences of states. Breathing state sequences and the observables representing those sequences were analyzed using a hidden Markov model (HMM) to predict the future sequences and new observables. Velocities and other parameters were clustered using a k-means clustering algorithm to associate each state with a set of observables such that a prediction of state also enables a prediction of tumor velocity. A time average model with predictions based on average past state lengths was also computed. State sequences which are known a priori to fit the data were fed into the HMM algorithm to set a theoretical limit of the predictive power of the model. The effectiveness of the presented probabilistic model has been evaluated for gated radiation therapy based on previously tracked tumor motion in four lung cancer patients. Positional prediction accuracy is compared with actual position in terms of the overall RMS errors. Various system delays, ranging from 33 to 1000 ms, were tested. Previous studies have shown duty cycles for latencies of 33 and 200 ms at around 90{\%} and 80{\%}, respectively, for linear, no prediction, Kalman filter and ANN methods as averaged over multiple patients. At 1000 ms, the previously reported duty cycles range from approximately 62{\%} (ANN) down to 34{\%} (no prediction). Average duty cycle for the HMMmethod was found to be 100{\%} and 91±3{\%} for 33 and 200 ms latency and around 40{\%} for 1000 ms latency in three out of four breathing motion traces. RMS errors were found to be lower than linear and no prediction methods at latencies of 1000 ms. The results show that for system latencies longer than 400 ms, the time average HMM prediction outperforms linear, no prediction, and the more general HMM-type predictive models. RMS errors for the time average model approach the theoretical limit of the HMM, and predicted state sequences are well correlated with sequences known to fit the data.",
author = "Alan Kalet and George Sandison and Huanmei Wu and Ruth Schmitz",
year = "2010",
month = "12",
day = "21",
doi = "10.1088/0031-9155/55/24/015",
language = "English",
volume = "55",
pages = "7615--7631",
journal = "Physics in Medicine and Biology",
issn = "0031-9155",
publisher = "IOP Publishing Ltd.",
number = "24",

}

TY - JOUR

T1 - A state-based probabilistic model for tumor respiratory motion prediction

AU - Kalet, Alan

AU - Sandison, George

AU - Wu, Huanmei

AU - Schmitz, Ruth

PY - 2010/12/21

Y1 - 2010/12/21

N2 - This work proposes a new probabilistic mathematical model for predicting tumormotion and position based on a finite state representation using the natural breathing states of exhale, inhale and end of exhale. Tumor motion was broken down into linear breathing states and sequences of states. Breathing state sequences and the observables representing those sequences were analyzed using a hidden Markov model (HMM) to predict the future sequences and new observables. Velocities and other parameters were clustered using a k-means clustering algorithm to associate each state with a set of observables such that a prediction of state also enables a prediction of tumor velocity. A time average model with predictions based on average past state lengths was also computed. State sequences which are known a priori to fit the data were fed into the HMM algorithm to set a theoretical limit of the predictive power of the model. The effectiveness of the presented probabilistic model has been evaluated for gated radiation therapy based on previously tracked tumor motion in four lung cancer patients. Positional prediction accuracy is compared with actual position in terms of the overall RMS errors. Various system delays, ranging from 33 to 1000 ms, were tested. Previous studies have shown duty cycles for latencies of 33 and 200 ms at around 90% and 80%, respectively, for linear, no prediction, Kalman filter and ANN methods as averaged over multiple patients. At 1000 ms, the previously reported duty cycles range from approximately 62% (ANN) down to 34% (no prediction). Average duty cycle for the HMMmethod was found to be 100% and 91±3% for 33 and 200 ms latency and around 40% for 1000 ms latency in three out of four breathing motion traces. RMS errors were found to be lower than linear and no prediction methods at latencies of 1000 ms. The results show that for system latencies longer than 400 ms, the time average HMM prediction outperforms linear, no prediction, and the more general HMM-type predictive models. RMS errors for the time average model approach the theoretical limit of the HMM, and predicted state sequences are well correlated with sequences known to fit the data.

AB - This work proposes a new probabilistic mathematical model for predicting tumormotion and position based on a finite state representation using the natural breathing states of exhale, inhale and end of exhale. Tumor motion was broken down into linear breathing states and sequences of states. Breathing state sequences and the observables representing those sequences were analyzed using a hidden Markov model (HMM) to predict the future sequences and new observables. Velocities and other parameters were clustered using a k-means clustering algorithm to associate each state with a set of observables such that a prediction of state also enables a prediction of tumor velocity. A time average model with predictions based on average past state lengths was also computed. State sequences which are known a priori to fit the data were fed into the HMM algorithm to set a theoretical limit of the predictive power of the model. The effectiveness of the presented probabilistic model has been evaluated for gated radiation therapy based on previously tracked tumor motion in four lung cancer patients. Positional prediction accuracy is compared with actual position in terms of the overall RMS errors. Various system delays, ranging from 33 to 1000 ms, were tested. Previous studies have shown duty cycles for latencies of 33 and 200 ms at around 90% and 80%, respectively, for linear, no prediction, Kalman filter and ANN methods as averaged over multiple patients. At 1000 ms, the previously reported duty cycles range from approximately 62% (ANN) down to 34% (no prediction). Average duty cycle for the HMMmethod was found to be 100% and 91±3% for 33 and 200 ms latency and around 40% for 1000 ms latency in three out of four breathing motion traces. RMS errors were found to be lower than linear and no prediction methods at latencies of 1000 ms. The results show that for system latencies longer than 400 ms, the time average HMM prediction outperforms linear, no prediction, and the more general HMM-type predictive models. RMS errors for the time average model approach the theoretical limit of the HMM, and predicted state sequences are well correlated with sequences known to fit the data.

UR - http://www.scopus.com/inward/record.url?scp=78650042812&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78650042812&partnerID=8YFLogxK

U2 - 10.1088/0031-9155/55/24/015

DO - 10.1088/0031-9155/55/24/015

M3 - Article

VL - 55

SP - 7615

EP - 7631

JO - Physics in Medicine and Biology

JF - Physics in Medicine and Biology

SN - 0031-9155

IS - 24

ER -