Machine Learning Methods Uncover Radiomorphologic Dose Patterns in Salivary Glands that Predict Xerostomia in Patients with Head and Neck Cancer

Wei Jiang, Pranav Lakshminarayanan, Xuan Hui, Peijin Han, Zhi Cheng, Michael Bowers, Ilya Shpitser, Sauleh Siddiqui, Russell H Taylor, Harry Quon, Todd McNutt

Research output: Contribution to journalArticle

Abstract

Purpose: Patients with head-and-neck cancer (HNC) may experience xerostomia after radiation therapy (RT), which leads to compromised quality of life. The purpose of this study is to explore how the spatial pattern of radiation dose (radiomorphology) in the major salivary glands influences xerostomia in patients with HNC. Methods and materials: A data-driven approach using spatially explicit dosimetric predictors, voxel dose (ie, actual radiation dose in voxels in parotid glands [PG] and submandibular glands [SMG]) was used to predict whether patients would develop xerostomia 3 months after RT. Using planned radiation dose data and other nondose covariates including baseline xerostomia grade of 427 patients with HNC in our database, the machine learning methods were used to investigate the influence of dose patterns across subvolumes in PG and SMG on xerostomia. Results: Of the 3 supervised learning methods studied, ridge logistic regression yielded the best predictive performance. Ridge logistic regression was also preferred to evaluate the influence pattern of highly correlated dose on xerostomia, which showed a discriminative pattern of influence of doses in the PG and SMG on xerostomia. Moreover, the superior–anterior portion of the contralateral PG and medial portion of the ipsilateral PG were determined to be the most influential regions regarding dose effect on xerostomia. The area under the receiver operating characteristic curve from a 10-fold cross-validation was 0.70 ± 0.04. Conclusions: Radiomorphology, combined with machine learning methods, is able to suggest patterns of dose in PG and SMG that are the most influential on xerostomia. The influence pattern identified by this data-driven approach and machine learning methods may help improve RT treatment planning and reduce xerostomia after treatment.

Original languageEnglish (US)
JournalAdvances in Radiation Oncology
DOIs
StateAccepted/In press - Jan 1 2019

Fingerprint

Xerostomia
Head and Neck Neoplasms
Salivary Glands
Parotid Gland
Submandibular Gland
Radiotherapy
Radiation
Logistic Models
Machine Learning
ROC Curve
Quality of Life
Learning
Databases

ASJC Scopus subject areas

  • Oncology
  • Radiology Nuclear Medicine and imaging

Cite this

Machine Learning Methods Uncover Radiomorphologic Dose Patterns in Salivary Glands that Predict Xerostomia in Patients with Head and Neck Cancer. / Jiang, Wei; Lakshminarayanan, Pranav; Hui, Xuan; Han, Peijin; Cheng, Zhi; Bowers, Michael; Shpitser, Ilya; Siddiqui, Sauleh; Taylor, Russell H; Quon, Harry; McNutt, Todd.

In: Advances in Radiation Oncology, 01.01.2019.

Research output: Contribution to journalArticle

@article{ae4a1fd5943c4882979ba1778ec74f9a,
title = "Machine Learning Methods Uncover Radiomorphologic Dose Patterns in Salivary Glands that Predict Xerostomia in Patients with Head and Neck Cancer",
abstract = "Purpose: Patients with head-and-neck cancer (HNC) may experience xerostomia after radiation therapy (RT), which leads to compromised quality of life. The purpose of this study is to explore how the spatial pattern of radiation dose (radiomorphology) in the major salivary glands influences xerostomia in patients with HNC. Methods and materials: A data-driven approach using spatially explicit dosimetric predictors, voxel dose (ie, actual radiation dose in voxels in parotid glands [PG] and submandibular glands [SMG]) was used to predict whether patients would develop xerostomia 3 months after RT. Using planned radiation dose data and other nondose covariates including baseline xerostomia grade of 427 patients with HNC in our database, the machine learning methods were used to investigate the influence of dose patterns across subvolumes in PG and SMG on xerostomia. Results: Of the 3 supervised learning methods studied, ridge logistic regression yielded the best predictive performance. Ridge logistic regression was also preferred to evaluate the influence pattern of highly correlated dose on xerostomia, which showed a discriminative pattern of influence of doses in the PG and SMG on xerostomia. Moreover, the superior–anterior portion of the contralateral PG and medial portion of the ipsilateral PG were determined to be the most influential regions regarding dose effect on xerostomia. The area under the receiver operating characteristic curve from a 10-fold cross-validation was 0.70 ± 0.04. Conclusions: Radiomorphology, combined with machine learning methods, is able to suggest patterns of dose in PG and SMG that are the most influential on xerostomia. The influence pattern identified by this data-driven approach and machine learning methods may help improve RT treatment planning and reduce xerostomia after treatment.",
author = "Wei Jiang and Pranav Lakshminarayanan and Xuan Hui and Peijin Han and Zhi Cheng and Michael Bowers and Ilya Shpitser and Sauleh Siddiqui and Taylor, {Russell H} and Harry Quon and Todd McNutt",
year = "2019",
month = "1",
day = "1",
doi = "10.1016/j.adro.2018.11.008",
language = "English (US)",
journal = "Advances in Radiation Oncology",
issn = "2452-1094",
publisher = "Elsevier Inc.",

}

TY - JOUR

T1 - Machine Learning Methods Uncover Radiomorphologic Dose Patterns in Salivary Glands that Predict Xerostomia in Patients with Head and Neck Cancer

AU - Jiang, Wei

AU - Lakshminarayanan, Pranav

AU - Hui, Xuan

AU - Han, Peijin

AU - Cheng, Zhi

AU - Bowers, Michael

AU - Shpitser, Ilya

AU - Siddiqui, Sauleh

AU - Taylor, Russell H

AU - Quon, Harry

AU - McNutt, Todd

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Purpose: Patients with head-and-neck cancer (HNC) may experience xerostomia after radiation therapy (RT), which leads to compromised quality of life. The purpose of this study is to explore how the spatial pattern of radiation dose (radiomorphology) in the major salivary glands influences xerostomia in patients with HNC. Methods and materials: A data-driven approach using spatially explicit dosimetric predictors, voxel dose (ie, actual radiation dose in voxels in parotid glands [PG] and submandibular glands [SMG]) was used to predict whether patients would develop xerostomia 3 months after RT. Using planned radiation dose data and other nondose covariates including baseline xerostomia grade of 427 patients with HNC in our database, the machine learning methods were used to investigate the influence of dose patterns across subvolumes in PG and SMG on xerostomia. Results: Of the 3 supervised learning methods studied, ridge logistic regression yielded the best predictive performance. Ridge logistic regression was also preferred to evaluate the influence pattern of highly correlated dose on xerostomia, which showed a discriminative pattern of influence of doses in the PG and SMG on xerostomia. Moreover, the superior–anterior portion of the contralateral PG and medial portion of the ipsilateral PG were determined to be the most influential regions regarding dose effect on xerostomia. The area under the receiver operating characteristic curve from a 10-fold cross-validation was 0.70 ± 0.04. Conclusions: Radiomorphology, combined with machine learning methods, is able to suggest patterns of dose in PG and SMG that are the most influential on xerostomia. The influence pattern identified by this data-driven approach and machine learning methods may help improve RT treatment planning and reduce xerostomia after treatment.

AB - Purpose: Patients with head-and-neck cancer (HNC) may experience xerostomia after radiation therapy (RT), which leads to compromised quality of life. The purpose of this study is to explore how the spatial pattern of radiation dose (radiomorphology) in the major salivary glands influences xerostomia in patients with HNC. Methods and materials: A data-driven approach using spatially explicit dosimetric predictors, voxel dose (ie, actual radiation dose in voxels in parotid glands [PG] and submandibular glands [SMG]) was used to predict whether patients would develop xerostomia 3 months after RT. Using planned radiation dose data and other nondose covariates including baseline xerostomia grade of 427 patients with HNC in our database, the machine learning methods were used to investigate the influence of dose patterns across subvolumes in PG and SMG on xerostomia. Results: Of the 3 supervised learning methods studied, ridge logistic regression yielded the best predictive performance. Ridge logistic regression was also preferred to evaluate the influence pattern of highly correlated dose on xerostomia, which showed a discriminative pattern of influence of doses in the PG and SMG on xerostomia. Moreover, the superior–anterior portion of the contralateral PG and medial portion of the ipsilateral PG were determined to be the most influential regions regarding dose effect on xerostomia. The area under the receiver operating characteristic curve from a 10-fold cross-validation was 0.70 ± 0.04. Conclusions: Radiomorphology, combined with machine learning methods, is able to suggest patterns of dose in PG and SMG that are the most influential on xerostomia. The influence pattern identified by this data-driven approach and machine learning methods may help improve RT treatment planning and reduce xerostomia after treatment.

UR - http://www.scopus.com/inward/record.url?scp=85060608692&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85060608692&partnerID=8YFLogxK

U2 - 10.1016/j.adro.2018.11.008

DO - 10.1016/j.adro.2018.11.008

M3 - Article

C2 - 31011686

AN - SCOPUS:85060608692

JO - Advances in Radiation Oncology

JF - Advances in Radiation Oncology

SN - 2452-1094

ER -