Outcome prediction of head and neck squamous cell carcinoma by MRI radiomic signatures

Steven W Mes; Floris H P van Velden; Boris Peltenburg; Carel F W Peeters; Dennis E Te Beest; Mark A van de Wiel; Joost Mekke; Doriene C Mulder; Roland M Martens; Jonas A Castelijns; Frank A Pameijer; Remco de Bree; Ronald Boellaard; C René Leemans; Ruud H Brakenhoff; Pim de Graaf

doi:https://doi.org/10.1007/s00330-020-06962-y

Outcome prediction of head and neck squamous cell carcinoma by MRI radiomic signatures

Steven W Mes, Floris H P van Velden, Boris Peltenburg, Carel F W Peeters, Dennis E Te Beest, Mark A van de Wiel, Joost Mekke, Doriene C Mulder, Roland M Martens, Jonas A Castelijns, Frank A Pameijer, Remco de Bree, Ronald Boellaard, C René Leemans, Ruud H Brakenhoff, Pim de Graaf

Research output: Contribution to journal › Article › Academic › peer-review

51 Citations (Scopus)

Abstract

OBJECTIVES: Head and neck squamous cell carcinoma (HNSCC) shows a remarkable heterogeneity between tumors, which may be captured by a variety of quantitative features extracted from diagnostic images, termed radiomics. The aim of this study was to develop and validate MRI-based radiomic prognostic models in oral and oropharyngeal cancer.

MATERIALS AND METHODS: Native T1-weighted images of four independent, retrospective (2005-2013), patient cohorts (n = 102, n = 76, n = 89, and n = 56) were used to delineate primary tumors, and to extract 545 quantitative features from. Subsequently, redundancy filtering and factor analysis were performed to handle collinearity in the data. Next, radiomic prognostic models were trained and validated to predict overall survival (OS) and relapse-free survival (RFS). Radiomic features were compared to and combined with prognostic models based on standard clinical parameters. Performance was assessed by integrated area under the curve (iAUC).

RESULTS: In oral cancer, the radiomic model showed an iAUC of 0.69 (OS) and 0.70 (RFS) in the validation cohort, whereas the iAUC in the oropharyngeal cancer validation cohort was 0.71 (OS) and 0.74 (RFS). By integration of radiomic and clinical variables, the most accurate models were defined (iAUC oral cavity, 0.72 (OS) and 0.74 (RFS); iAUC oropharynx, 0.81 (OS) and 0.78 (RFS)), and these combined models outperformed prognostic models based on standard clinical variables only (p < 0.001).

CONCLUSIONS: MRI radiomics is feasible in HNSCC despite the known variability in MRI vendors and acquisition protocols, and radiomic features added information to prognostic models based on clinical parameters.

KEY POINTS: • MRI radiomics can predict overall survival and relapse-free survival in oral and HPV-negative oropharyngeal cancer. • MRI radiomics provides additional prognostic information to known clinical variables, with the best performance of the combined models. • Variation in MRI vendors and acquisition protocols did not influence performance of radiomic prognostic models.

Original language	English
Pages (from-to)	6311-6321
Number of pages	11
Journal	European Radiology
Volume	30
Issue number	11
Early online date	4 Jun 2020
DOIs	https://doi.org/10.1007/s00330-020-06962-y
Publication status	Published - 1 Nov 2020

Keywords

Factor analysis
Head and neck neoplasms
Magnetic resonance imaging
Prognosis

Access to Document

https://doi.org/10.1007/s00330-020-06962-y

Cite this

Mes, S. W., van Velden, F. H. P., Peltenburg, B., Peeters, C. F. W., Te Beest, D. E., van de Wiel, M. A., Mekke, J., Mulder, D. C., Martens, R. M., Castelijns, J. A., Pameijer, F. A., de Bree, R., Boellaard, R., Leemans, C. R., Brakenhoff, R. H., & de Graaf, P. (2020). Outcome prediction of head and neck squamous cell carcinoma by MRI radiomic signatures. European Radiology, 30(11), 6311-6321. https://doi.org/10.1007/s00330-020-06962-y

@article{38975d9eb2de4e53b64849008b008828,

title = "Outcome prediction of head and neck squamous cell carcinoma by MRI radiomic signatures",

abstract = "OBJECTIVES: Head and neck squamous cell carcinoma (HNSCC) shows a remarkable heterogeneity between tumors, which may be captured by a variety of quantitative features extracted from diagnostic images, termed radiomics. The aim of this study was to develop and validate MRI-based radiomic prognostic models in oral and oropharyngeal cancer.MATERIALS AND METHODS: Native T1-weighted images of four independent, retrospective (2005-2013), patient cohorts (n = 102, n = 76, n = 89, and n = 56) were used to delineate primary tumors, and to extract 545 quantitative features from. Subsequently, redundancy filtering and factor analysis were performed to handle collinearity in the data. Next, radiomic prognostic models were trained and validated to predict overall survival (OS) and relapse-free survival (RFS). Radiomic features were compared to and combined with prognostic models based on standard clinical parameters. Performance was assessed by integrated area under the curve (iAUC).RESULTS: In oral cancer, the radiomic model showed an iAUC of 0.69 (OS) and 0.70 (RFS) in the validation cohort, whereas the iAUC in the oropharyngeal cancer validation cohort was 0.71 (OS) and 0.74 (RFS). By integration of radiomic and clinical variables, the most accurate models were defined (iAUC oral cavity, 0.72 (OS) and 0.74 (RFS); iAUC oropharynx, 0.81 (OS) and 0.78 (RFS)), and these combined models outperformed prognostic models based on standard clinical variables only (p < 0.001).CONCLUSIONS: MRI radiomics is feasible in HNSCC despite the known variability in MRI vendors and acquisition protocols, and radiomic features added information to prognostic models based on clinical parameters.KEY POINTS: • MRI radiomics can predict overall survival and relapse-free survival in oral and HPV-negative oropharyngeal cancer. • MRI radiomics provides additional prognostic information to known clinical variables, with the best performance of the combined models. • Variation in MRI vendors and acquisition protocols did not influence performance of radiomic prognostic models.",

keywords = "Factor analysis, Head and neck neoplasms, Magnetic resonance imaging, Prognosis",

author = "Mes, {Steven W} and {van Velden}, {Floris H P} and Boris Peltenburg and Peeters, {Carel F W} and {Te Beest}, {Dennis E} and {van de Wiel}, {Mark A} and Joost Mekke and Mulder, {Doriene C} and Martens, {Roland M} and Castelijns, {Jonas A} and Pameijer, {Frank A} and {de Bree}, Remco and Ronald Boellaard and Leemans, {C Ren{\'e}} and Brakenhoff, {Ruud H} and {de Graaf}, Pim",

year = "2020",

month = nov,

day = "1",

doi = "https://doi.org/10.1007/s00330-020-06962-y",

language = "English",

volume = "30",

pages = "6311--6321",

journal = "European Radiology",

issn = "0938-7994",

publisher = "Springer Verlag",

number = "11",

}

Mes, SW, van Velden, FHP, Peltenburg, B, Peeters, CFW, Te Beest, DE, van de Wiel, MA, Mekke, J, Mulder, DC, Martens, RM, Castelijns, JA, Pameijer, FA, de Bree, R, Boellaard, R , Leemans, CR , Brakenhoff, RH & de Graaf, P 2020, 'Outcome prediction of head and neck squamous cell carcinoma by MRI radiomic signatures', European Radiology, vol. 30, no. 11, pp. 6311-6321. https://doi.org/10.1007/s00330-020-06962-y

TY - JOUR

T1 - Outcome prediction of head and neck squamous cell carcinoma by MRI radiomic signatures

AU - Mes, Steven W

AU - van Velden, Floris H P

AU - Peltenburg, Boris

AU - Peeters, Carel F W

AU - Te Beest, Dennis E

AU - van de Wiel, Mark A

AU - Mekke, Joost

AU - Mulder, Doriene C

AU - Martens, Roland M

AU - Castelijns, Jonas A

AU - Pameijer, Frank A

AU - de Bree, Remco

AU - Boellaard, Ronald

AU - Leemans, C René

AU - Brakenhoff, Ruud H

AU - de Graaf, Pim

PY - 2020/11/1

Y1 - 2020/11/1

N2 - OBJECTIVES: Head and neck squamous cell carcinoma (HNSCC) shows a remarkable heterogeneity between tumors, which may be captured by a variety of quantitative features extracted from diagnostic images, termed radiomics. The aim of this study was to develop and validate MRI-based radiomic prognostic models in oral and oropharyngeal cancer.MATERIALS AND METHODS: Native T1-weighted images of four independent, retrospective (2005-2013), patient cohorts (n = 102, n = 76, n = 89, and n = 56) were used to delineate primary tumors, and to extract 545 quantitative features from. Subsequently, redundancy filtering and factor analysis were performed to handle collinearity in the data. Next, radiomic prognostic models were trained and validated to predict overall survival (OS) and relapse-free survival (RFS). Radiomic features were compared to and combined with prognostic models based on standard clinical parameters. Performance was assessed by integrated area under the curve (iAUC).RESULTS: In oral cancer, the radiomic model showed an iAUC of 0.69 (OS) and 0.70 (RFS) in the validation cohort, whereas the iAUC in the oropharyngeal cancer validation cohort was 0.71 (OS) and 0.74 (RFS). By integration of radiomic and clinical variables, the most accurate models were defined (iAUC oral cavity, 0.72 (OS) and 0.74 (RFS); iAUC oropharynx, 0.81 (OS) and 0.78 (RFS)), and these combined models outperformed prognostic models based on standard clinical variables only (p < 0.001).CONCLUSIONS: MRI radiomics is feasible in HNSCC despite the known variability in MRI vendors and acquisition protocols, and radiomic features added information to prognostic models based on clinical parameters.KEY POINTS: • MRI radiomics can predict overall survival and relapse-free survival in oral and HPV-negative oropharyngeal cancer. • MRI radiomics provides additional prognostic information to known clinical variables, with the best performance of the combined models. • Variation in MRI vendors and acquisition protocols did not influence performance of radiomic prognostic models.

AB - OBJECTIVES: Head and neck squamous cell carcinoma (HNSCC) shows a remarkable heterogeneity between tumors, which may be captured by a variety of quantitative features extracted from diagnostic images, termed radiomics. The aim of this study was to develop and validate MRI-based radiomic prognostic models in oral and oropharyngeal cancer.MATERIALS AND METHODS: Native T1-weighted images of four independent, retrospective (2005-2013), patient cohorts (n = 102, n = 76, n = 89, and n = 56) were used to delineate primary tumors, and to extract 545 quantitative features from. Subsequently, redundancy filtering and factor analysis were performed to handle collinearity in the data. Next, radiomic prognostic models were trained and validated to predict overall survival (OS) and relapse-free survival (RFS). Radiomic features were compared to and combined with prognostic models based on standard clinical parameters. Performance was assessed by integrated area under the curve (iAUC).RESULTS: In oral cancer, the radiomic model showed an iAUC of 0.69 (OS) and 0.70 (RFS) in the validation cohort, whereas the iAUC in the oropharyngeal cancer validation cohort was 0.71 (OS) and 0.74 (RFS). By integration of radiomic and clinical variables, the most accurate models were defined (iAUC oral cavity, 0.72 (OS) and 0.74 (RFS); iAUC oropharynx, 0.81 (OS) and 0.78 (RFS)), and these combined models outperformed prognostic models based on standard clinical variables only (p < 0.001).CONCLUSIONS: MRI radiomics is feasible in HNSCC despite the known variability in MRI vendors and acquisition protocols, and radiomic features added information to prognostic models based on clinical parameters.KEY POINTS: • MRI radiomics can predict overall survival and relapse-free survival in oral and HPV-negative oropharyngeal cancer. • MRI radiomics provides additional prognostic information to known clinical variables, with the best performance of the combined models. • Variation in MRI vendors and acquisition protocols did not influence performance of radiomic prognostic models.

KW - Factor analysis

KW - Head and neck neoplasms

KW - Magnetic resonance imaging

KW - Prognosis

UR - http://www.scopus.com/inward/record.url?scp=85086000214&partnerID=8YFLogxK

U2 - https://doi.org/10.1007/s00330-020-06962-y

DO - https://doi.org/10.1007/s00330-020-06962-y

M3 - Article

C2 - 32500196

SN - 0938-7994

VL - 30

SP - 6311

EP - 6321

JO - European Radiology

JF - European Radiology

IS - 11

ER -

Outcome prediction of head and neck squamous cell carcinoma by MRI radiomic signatures

Abstract

Keywords

Access to Document

Other files and links

Cite this