Predicting radiation-induced valvular heart damage

Laura Cella, Jung Hun Oh, Joseph O. Deasy, Giuseppe Palma, Raffaele Liuzzi, Vittoria D'Avino, Manuel Conson, Marco Picardi, Marco Salvatore, Roberto Pacelli

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

Purpose. To develop a predictive multivariate normal tissue complication probability (NTCP) model for radiation-induced heart valvular damage (RVD). The influence of combined heart-lung irradiation on RVD development was included.Material and methods. Multivariate logistic regression modeling with the least absolute shrinkage and selection operator (LASSO) was used to build an NTCP model to predict RVD based on a cohort of 90 Hodgkin lymphoma patients treated with sequential chemo-radiation therapy. In addition to heart irradiation factors, clinical variables, along with left and right lung dose-volume histogram statistics, were included in the analysis. To avoid overfitting, 10-fold cross-validation (CV) was used for LASSO logistic regression modeling, with 50 reshuffled cycles. Model performance was assessed using the area under the receiver operating characteristic (ROC) curve (AUC) and Spearman's correlation coefficient (Rs).Results. At a median follow-up time of 55 months (range 12-92 months) after the end of radiation treatment, 27 of 90 patients (30%) manifested at least one kind of RVD (mild or moderate), with a higher incidence of left-sided valve defects (64%). Fourteen prognostic factors were frequently selected (more than 100/500 model fits) by LASSO, which included mainly heart and left lung dosimetric variables along with their volume variables. The averaged cross-validated performance was AUC-CV = 0.685 and Rs = 0.293. The overall performance of a final NTCP model for RVD obtained applying LASSO logistic regression to the full dataset was satisfactory (AUC = 0.84, Rs = 0.55, p <0.001).Conclusion. LASSO proved to be an improved and flexible modeling method for variable selection. Applying LASSO, we showed, for the first time, the importance of jointly considering left lung irradiation and left lung volume size in the prediction of subclinical radiation-related heart disease resulting in RVD.

Original languageEnglish
Pages (from-to)1796-1804
Number of pages9
JournalActa Oncologica
Volume54
Issue number10
DOIs
Publication statusPublished - Nov 26 2015

Fingerprint

Radiation
Lung
Area Under Curve
Logistic Models
Hodgkin Disease
ROC Curve
Heart Diseases
Radiotherapy
Incidence

ASJC Scopus subject areas

  • Oncology
  • Radiology Nuclear Medicine and imaging
  • Hematology

Cite this

Cella, L., Oh, J. H., Deasy, J. O., Palma, G., Liuzzi, R., D'Avino, V., ... Pacelli, R. (2015). Predicting radiation-induced valvular heart damage. Acta Oncologica, 54(10), 1796-1804. https://doi.org/10.3109/0284186X.2015.1016624

Predicting radiation-induced valvular heart damage. / Cella, Laura; Oh, Jung Hun; Deasy, Joseph O.; Palma, Giuseppe; Liuzzi, Raffaele; D'Avino, Vittoria; Conson, Manuel; Picardi, Marco; Salvatore, Marco; Pacelli, Roberto.

In: Acta Oncologica, Vol. 54, No. 10, 26.11.2015, p. 1796-1804.

Research output: Contribution to journalArticle

Cella, L, Oh, JH, Deasy, JO, Palma, G, Liuzzi, R, D'Avino, V, Conson, M, Picardi, M, Salvatore, M & Pacelli, R 2015, 'Predicting radiation-induced valvular heart damage', Acta Oncologica, vol. 54, no. 10, pp. 1796-1804. https://doi.org/10.3109/0284186X.2015.1016624
Cella L, Oh JH, Deasy JO, Palma G, Liuzzi R, D'Avino V et al. Predicting radiation-induced valvular heart damage. Acta Oncologica. 2015 Nov 26;54(10):1796-1804. https://doi.org/10.3109/0284186X.2015.1016624
Cella, Laura ; Oh, Jung Hun ; Deasy, Joseph O. ; Palma, Giuseppe ; Liuzzi, Raffaele ; D'Avino, Vittoria ; Conson, Manuel ; Picardi, Marco ; Salvatore, Marco ; Pacelli, Roberto. / Predicting radiation-induced valvular heart damage. In: Acta Oncologica. 2015 ; Vol. 54, No. 10. pp. 1796-1804.
@article{e8b4c15b59654719831a2019a96f0d93,
title = "Predicting radiation-induced valvular heart damage",
abstract = "Purpose. To develop a predictive multivariate normal tissue complication probability (NTCP) model for radiation-induced heart valvular damage (RVD). The influence of combined heart-lung irradiation on RVD development was included.Material and methods. Multivariate logistic regression modeling with the least absolute shrinkage and selection operator (LASSO) was used to build an NTCP model to predict RVD based on a cohort of 90 Hodgkin lymphoma patients treated with sequential chemo-radiation therapy. In addition to heart irradiation factors, clinical variables, along with left and right lung dose-volume histogram statistics, were included in the analysis. To avoid overfitting, 10-fold cross-validation (CV) was used for LASSO logistic regression modeling, with 50 reshuffled cycles. Model performance was assessed using the area under the receiver operating characteristic (ROC) curve (AUC) and Spearman's correlation coefficient (Rs).Results. At a median follow-up time of 55 months (range 12-92 months) after the end of radiation treatment, 27 of 90 patients (30{\%}) manifested at least one kind of RVD (mild or moderate), with a higher incidence of left-sided valve defects (64{\%}). Fourteen prognostic factors were frequently selected (more than 100/500 model fits) by LASSO, which included mainly heart and left lung dosimetric variables along with their volume variables. The averaged cross-validated performance was AUC-CV = 0.685 and Rs = 0.293. The overall performance of a final NTCP model for RVD obtained applying LASSO logistic regression to the full dataset was satisfactory (AUC = 0.84, Rs = 0.55, p <0.001).Conclusion. LASSO proved to be an improved and flexible modeling method for variable selection. Applying LASSO, we showed, for the first time, the importance of jointly considering left lung irradiation and left lung volume size in the prediction of subclinical radiation-related heart disease resulting in RVD.",
author = "Laura Cella and Oh, {Jung Hun} and Deasy, {Joseph O.} and Giuseppe Palma and Raffaele Liuzzi and Vittoria D'Avino and Manuel Conson and Marco Picardi and Marco Salvatore and Roberto Pacelli",
year = "2015",
month = "11",
day = "26",
doi = "10.3109/0284186X.2015.1016624",
language = "English",
volume = "54",
pages = "1796--1804",
journal = "Acta Oncologica",
issn = "0284-186X",
publisher = "Informa Healthcare",
number = "10",

}

TY - JOUR

T1 - Predicting radiation-induced valvular heart damage

AU - Cella, Laura

AU - Oh, Jung Hun

AU - Deasy, Joseph O.

AU - Palma, Giuseppe

AU - Liuzzi, Raffaele

AU - D'Avino, Vittoria

AU - Conson, Manuel

AU - Picardi, Marco

AU - Salvatore, Marco

AU - Pacelli, Roberto

PY - 2015/11/26

Y1 - 2015/11/26

N2 - Purpose. To develop a predictive multivariate normal tissue complication probability (NTCP) model for radiation-induced heart valvular damage (RVD). The influence of combined heart-lung irradiation on RVD development was included.Material and methods. Multivariate logistic regression modeling with the least absolute shrinkage and selection operator (LASSO) was used to build an NTCP model to predict RVD based on a cohort of 90 Hodgkin lymphoma patients treated with sequential chemo-radiation therapy. In addition to heart irradiation factors, clinical variables, along with left and right lung dose-volume histogram statistics, were included in the analysis. To avoid overfitting, 10-fold cross-validation (CV) was used for LASSO logistic regression modeling, with 50 reshuffled cycles. Model performance was assessed using the area under the receiver operating characteristic (ROC) curve (AUC) and Spearman's correlation coefficient (Rs).Results. At a median follow-up time of 55 months (range 12-92 months) after the end of radiation treatment, 27 of 90 patients (30%) manifested at least one kind of RVD (mild or moderate), with a higher incidence of left-sided valve defects (64%). Fourteen prognostic factors were frequently selected (more than 100/500 model fits) by LASSO, which included mainly heart and left lung dosimetric variables along with their volume variables. The averaged cross-validated performance was AUC-CV = 0.685 and Rs = 0.293. The overall performance of a final NTCP model for RVD obtained applying LASSO logistic regression to the full dataset was satisfactory (AUC = 0.84, Rs = 0.55, p <0.001).Conclusion. LASSO proved to be an improved and flexible modeling method for variable selection. Applying LASSO, we showed, for the first time, the importance of jointly considering left lung irradiation and left lung volume size in the prediction of subclinical radiation-related heart disease resulting in RVD.

AB - Purpose. To develop a predictive multivariate normal tissue complication probability (NTCP) model for radiation-induced heart valvular damage (RVD). The influence of combined heart-lung irradiation on RVD development was included.Material and methods. Multivariate logistic regression modeling with the least absolute shrinkage and selection operator (LASSO) was used to build an NTCP model to predict RVD based on a cohort of 90 Hodgkin lymphoma patients treated with sequential chemo-radiation therapy. In addition to heart irradiation factors, clinical variables, along with left and right lung dose-volume histogram statistics, were included in the analysis. To avoid overfitting, 10-fold cross-validation (CV) was used for LASSO logistic regression modeling, with 50 reshuffled cycles. Model performance was assessed using the area under the receiver operating characteristic (ROC) curve (AUC) and Spearman's correlation coefficient (Rs).Results. At a median follow-up time of 55 months (range 12-92 months) after the end of radiation treatment, 27 of 90 patients (30%) manifested at least one kind of RVD (mild or moderate), with a higher incidence of left-sided valve defects (64%). Fourteen prognostic factors were frequently selected (more than 100/500 model fits) by LASSO, which included mainly heart and left lung dosimetric variables along with their volume variables. The averaged cross-validated performance was AUC-CV = 0.685 and Rs = 0.293. The overall performance of a final NTCP model for RVD obtained applying LASSO logistic regression to the full dataset was satisfactory (AUC = 0.84, Rs = 0.55, p <0.001).Conclusion. LASSO proved to be an improved and flexible modeling method for variable selection. Applying LASSO, we showed, for the first time, the importance of jointly considering left lung irradiation and left lung volume size in the prediction of subclinical radiation-related heart disease resulting in RVD.

UR - http://www.scopus.com/inward/record.url?scp=84946231187&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84946231187&partnerID=8YFLogxK

U2 - 10.3109/0284186X.2015.1016624

DO - 10.3109/0284186X.2015.1016624

M3 - Article

AN - SCOPUS:84946231187

VL - 54

SP - 1796

EP - 1804

JO - Acta Oncologica

JF - Acta Oncologica

SN - 0284-186X

IS - 10

ER -