Predicting response to neoadjuvant chemoradiotherapy in esophageal cancer with textural features derived from pre-treatment 18F-FDG PET/CT imaging

Roelof J. Beukinga, Jan B. Hulshoff, Lisanne V. van Dijk, Christina T. Muijs, Johannes G.M. Burgerhof, Gursah Kats-Ugurlu, Riemer H.J.A. Slart, Cornelis H. Slump, Véronique E.M. Mul, John Th.M. Plukker

Research output: Contribution to journalArticleAcademicpeer-review

18 Citations (Scopus)

Abstract

Adequate prediction of tumor response to neoadjuvant chemoradiotherapy (nCRT) in esophageal cancer (EC) patients is important in a more personalized treatment. The current best clinical method to predict pathologic complete response is SUVmax in 18F-FDG PET/CT imaging. To improve the prediction of response, we constructed a model to predict complete response to nCRT in EC based on pretreatment clinical parameters and 18F-FDG PET/CT–derived textural features.

Methods: From a prospectively maintained single-institution database, we reviewed 97 consecutive patients with locally advanced EC and a pretreatment 18F-FDG PET/CT scan between 2009 and 2015. All patients were treated with nCRT (carboplatin/paclitaxel/41.4 Gy) followed by esophagectomy. We analyzed clinical, geometric, and pretreatment textural features extracted from both 18F-FDG PET and CT. The current most accurate prediction model with SUVmax as a predictor variable was compared with 6 different response prediction models constructed using least absolute shrinkage and selection operator regularized logistic regression. Internal validation was performed to estimate the model’s performances. Pathologic response was defined as complete versus incomplete response (Mandard tumor regression grade system 1 vs. 2–5).

Results: Pathologic examination revealed 19 (19.6%) complete and 78 (80.4%) incomplete responders. Least absolute shrinkage and selection operator regularization selected the clinical parameters: histologic type and clinical T stage, the 18F-FDG PET–derived textural feature long run low gray level emphasis, and the CT-derived textural feature run percentage. Introducing these variables to a logistic regression analysis showed areas under the receiver-operating-characteristic curve (AUCs) of 0.78 compared with 0.58 in the SUVmax model. The discrimination slopes were 0.17 compared with 0.01, respectively. After internal validation, the AUCs decreased to 0.74 and 0.54, respectively.

Conclusion: The predictive values of the constructed models were superior to the standard method (SUVmax). These results can be considered as an initial step in predicting tumor response to nCRT in locally advanced EC. Further research in refining the predictive value of these models is needed to justify omission of surgery.
Original languageEnglish
Pages (from-to)723-729
JournalJournal of nuclear medicine
Volume58
Issue number5
DOIs
Publication statusPublished - 1 May 2017

Fingerprint

Fluorodeoxyglucose F18
Chemoradiotherapy
Esophageal Neoplasms
Area Under Curve
Logistic Models
Therapeutics
Neoplasms
Esophagectomy
Carboplatin
Paclitaxel
ROC Curve
Regression Analysis
Databases
Research

Keywords

  • METIS-320193

Cite this

Beukinga, R. J., Hulshoff, J. B., van Dijk, L. V., Muijs, C. T., Burgerhof, J. G. M., Kats-Ugurlu, G., ... Plukker, J. T. M. (2017). Predicting response to neoadjuvant chemoradiotherapy in esophageal cancer with textural features derived from pre-treatment 18F-FDG PET/CT imaging. Journal of nuclear medicine, 58(5), 723-729. https://doi.org/10.2967/jnumed.116.180299
Beukinga, Roelof J. ; Hulshoff, Jan B. ; van Dijk, Lisanne V. ; Muijs, Christina T. ; Burgerhof, Johannes G.M. ; Kats-Ugurlu, Gursah ; Slart, Riemer H.J.A. ; Slump, Cornelis H. ; Mul, Véronique E.M. ; Plukker, John Th.M. / Predicting response to neoadjuvant chemoradiotherapy in esophageal cancer with textural features derived from pre-treatment 18F-FDG PET/CT imaging. In: Journal of nuclear medicine. 2017 ; Vol. 58, No. 5. pp. 723-729.
@article{15f8f83861a84bc6ac5c3d5ef670f65e,
title = "Predicting response to neoadjuvant chemoradiotherapy in esophageal cancer with textural features derived from pre-treatment 18F-FDG PET/CT imaging",
abstract = "Adequate prediction of tumor response to neoadjuvant chemoradiotherapy (nCRT) in esophageal cancer (EC) patients is important in a more personalized treatment. The current best clinical method to predict pathologic complete response is SUVmax in 18F-FDG PET/CT imaging. To improve the prediction of response, we constructed a model to predict complete response to nCRT in EC based on pretreatment clinical parameters and 18F-FDG PET/CT–derived textural features. Methods: From a prospectively maintained single-institution database, we reviewed 97 consecutive patients with locally advanced EC and a pretreatment 18F-FDG PET/CT scan between 2009 and 2015. All patients were treated with nCRT (carboplatin/paclitaxel/41.4 Gy) followed by esophagectomy. We analyzed clinical, geometric, and pretreatment textural features extracted from both 18F-FDG PET and CT. The current most accurate prediction model with SUVmax as a predictor variable was compared with 6 different response prediction models constructed using least absolute shrinkage and selection operator regularized logistic regression. Internal validation was performed to estimate the model’s performances. Pathologic response was defined as complete versus incomplete response (Mandard tumor regression grade system 1 vs. 2–5).Results: Pathologic examination revealed 19 (19.6{\%}) complete and 78 (80.4{\%}) incomplete responders. Least absolute shrinkage and selection operator regularization selected the clinical parameters: histologic type and clinical T stage, the 18F-FDG PET–derived textural feature long run low gray level emphasis, and the CT-derived textural feature run percentage. Introducing these variables to a logistic regression analysis showed areas under the receiver-operating-characteristic curve (AUCs) of 0.78 compared with 0.58 in the SUVmax model. The discrimination slopes were 0.17 compared with 0.01, respectively. After internal validation, the AUCs decreased to 0.74 and 0.54, respectively. Conclusion: The predictive values of the constructed models were superior to the standard method (SUVmax). These results can be considered as an initial step in predicting tumor response to nCRT in locally advanced EC. Further research in refining the predictive value of these models is needed to justify omission of surgery.",
keywords = "METIS-320193",
author = "Beukinga, {Roelof J.} and Hulshoff, {Jan B.} and {van Dijk}, {Lisanne V.} and Muijs, {Christina T.} and Burgerhof, {Johannes G.M.} and Gursah Kats-Ugurlu and Slart, {Riemer H.J.A.} and Slump, {Cornelis H.} and Mul, {V{\'e}ronique E.M.} and Plukker, {John Th.M.}",
note = "Online first",
year = "2017",
month = "5",
day = "1",
doi = "10.2967/jnumed.116.180299",
language = "English",
volume = "58",
pages = "723--729",
journal = "Journal of nuclear medicine",
issn = "0161-5505",
publisher = "Society of Nuclear Medicine Inc.",
number = "5",

}

Predicting response to neoadjuvant chemoradiotherapy in esophageal cancer with textural features derived from pre-treatment 18F-FDG PET/CT imaging. / Beukinga, Roelof J.; Hulshoff, Jan B.; van Dijk, Lisanne V.; Muijs, Christina T.; Burgerhof, Johannes G.M.; Kats-Ugurlu, Gursah; Slart, Riemer H.J.A.; Slump, Cornelis H.; Mul, Véronique E.M.; Plukker, John Th.M.

In: Journal of nuclear medicine, Vol. 58, No. 5, 01.05.2017, p. 723-729.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Predicting response to neoadjuvant chemoradiotherapy in esophageal cancer with textural features derived from pre-treatment 18F-FDG PET/CT imaging

AU - Beukinga, Roelof J.

AU - Hulshoff, Jan B.

AU - van Dijk, Lisanne V.

AU - Muijs, Christina T.

AU - Burgerhof, Johannes G.M.

AU - Kats-Ugurlu, Gursah

AU - Slart, Riemer H.J.A.

AU - Slump, Cornelis H.

AU - Mul, Véronique E.M.

AU - Plukker, John Th.M.

N1 - Online first

PY - 2017/5/1

Y1 - 2017/5/1

N2 - Adequate prediction of tumor response to neoadjuvant chemoradiotherapy (nCRT) in esophageal cancer (EC) patients is important in a more personalized treatment. The current best clinical method to predict pathologic complete response is SUVmax in 18F-FDG PET/CT imaging. To improve the prediction of response, we constructed a model to predict complete response to nCRT in EC based on pretreatment clinical parameters and 18F-FDG PET/CT–derived textural features. Methods: From a prospectively maintained single-institution database, we reviewed 97 consecutive patients with locally advanced EC and a pretreatment 18F-FDG PET/CT scan between 2009 and 2015. All patients were treated with nCRT (carboplatin/paclitaxel/41.4 Gy) followed by esophagectomy. We analyzed clinical, geometric, and pretreatment textural features extracted from both 18F-FDG PET and CT. The current most accurate prediction model with SUVmax as a predictor variable was compared with 6 different response prediction models constructed using least absolute shrinkage and selection operator regularized logistic regression. Internal validation was performed to estimate the model’s performances. Pathologic response was defined as complete versus incomplete response (Mandard tumor regression grade system 1 vs. 2–5).Results: Pathologic examination revealed 19 (19.6%) complete and 78 (80.4%) incomplete responders. Least absolute shrinkage and selection operator regularization selected the clinical parameters: histologic type and clinical T stage, the 18F-FDG PET–derived textural feature long run low gray level emphasis, and the CT-derived textural feature run percentage. Introducing these variables to a logistic regression analysis showed areas under the receiver-operating-characteristic curve (AUCs) of 0.78 compared with 0.58 in the SUVmax model. The discrimination slopes were 0.17 compared with 0.01, respectively. After internal validation, the AUCs decreased to 0.74 and 0.54, respectively. Conclusion: The predictive values of the constructed models were superior to the standard method (SUVmax). These results can be considered as an initial step in predicting tumor response to nCRT in locally advanced EC. Further research in refining the predictive value of these models is needed to justify omission of surgery.

AB - Adequate prediction of tumor response to neoadjuvant chemoradiotherapy (nCRT) in esophageal cancer (EC) patients is important in a more personalized treatment. The current best clinical method to predict pathologic complete response is SUVmax in 18F-FDG PET/CT imaging. To improve the prediction of response, we constructed a model to predict complete response to nCRT in EC based on pretreatment clinical parameters and 18F-FDG PET/CT–derived textural features. Methods: From a prospectively maintained single-institution database, we reviewed 97 consecutive patients with locally advanced EC and a pretreatment 18F-FDG PET/CT scan between 2009 and 2015. All patients were treated with nCRT (carboplatin/paclitaxel/41.4 Gy) followed by esophagectomy. We analyzed clinical, geometric, and pretreatment textural features extracted from both 18F-FDG PET and CT. The current most accurate prediction model with SUVmax as a predictor variable was compared with 6 different response prediction models constructed using least absolute shrinkage and selection operator regularized logistic regression. Internal validation was performed to estimate the model’s performances. Pathologic response was defined as complete versus incomplete response (Mandard tumor regression grade system 1 vs. 2–5).Results: Pathologic examination revealed 19 (19.6%) complete and 78 (80.4%) incomplete responders. Least absolute shrinkage and selection operator regularization selected the clinical parameters: histologic type and clinical T stage, the 18F-FDG PET–derived textural feature long run low gray level emphasis, and the CT-derived textural feature run percentage. Introducing these variables to a logistic regression analysis showed areas under the receiver-operating-characteristic curve (AUCs) of 0.78 compared with 0.58 in the SUVmax model. The discrimination slopes were 0.17 compared with 0.01, respectively. After internal validation, the AUCs decreased to 0.74 and 0.54, respectively. Conclusion: The predictive values of the constructed models were superior to the standard method (SUVmax). These results can be considered as an initial step in predicting tumor response to nCRT in locally advanced EC. Further research in refining the predictive value of these models is needed to justify omission of surgery.

KW - METIS-320193

U2 - 10.2967/jnumed.116.180299

DO - 10.2967/jnumed.116.180299

M3 - Article

VL - 58

SP - 723

EP - 729

JO - Journal of nuclear medicine

JF - Journal of nuclear medicine

SN - 0161-5505

IS - 5

ER -