Machine learning highlights the deficiency of conventional dosimetric constraints for prevention of high-grade radiation esophagitis in non-small cell lung cancer treated with chemoradiation.
Chemoradiation
Intensity-modulated radiation therapy
Machine learning
Non-small cell lung cancer
Proton beam therapy
Radiation esophagitis
Radiation-induced toxicity
Journal
Clinical and translational radiation oncology
ISSN: 2405-6308
Titre abrégé: Clin Transl Radiat Oncol
Pays: Ireland
ID NLM: 101713416
Informations de publication
Date de publication:
May 2020
May 2020
Historique:
received:
16
12
2019
revised:
17
03
2020
accepted:
21
03
2020
entrez:
11
4
2020
pubmed:
11
4
2020
medline:
11
4
2020
Statut:
epublish
Résumé
Radiation esophagitis is a clinically important toxicity seen with treatment for locally-advanced non-small cell lung cancer. There is considerable disagreement among prior studies in identifying predictors of radiation esophagitis. We apply machine learning algorithms to identify factors contributing to the development of radiation esophagitis to uncover previously unidentified criteria and more robust dosimetric factors. We used machine learning approaches to identify predictors of grade ≥ 3 radiation esophagitis in a cohort of 202 consecutive locally-advanced non-small cell lung cancer patients treated with definitive chemoradiation from 2008 to 2016. We evaluated 35 clinical features per patient grouped into risk factors, comorbidities, imaging, stage, histology, radiotherapy, chemotherapy and dosimetry. Univariate and multivariate analyses were performed using a panel of 11 machine learning algorithms combined with predictive power assessments. All patients were treated to a median dose of 66.6 Gy at 1.8 Gy per fraction using photon (89.6%) and proton (10.4%) beam therapy, most often with concurrent chemotherapy (86.6%). 11.4% of patients developed grade ≥ 3 radiation esophagitis. On univariate analysis, no individual feature was found to predict radiation esophagitis (AUC range 0.45-0.55, p ≥ 0.07). In multivariate analysis, all machine learning algorithms exhibited poor predictive performance (AUC range 0.46-0.56, p ≥ 0.07). Contemporary machine learning algorithms applied to our modern, relatively large institutional cohort could not identify any reliable predictors of grade ≥ 3 radiation esophagitis. Additional patients are needed, and novel patient-specific and treatment characteristics should be investigated to develop clinically meaningful methods to mitigate this survival altering toxicity.
Sections du résumé
BACKGROUND AND PURPOSE
OBJECTIVE
Radiation esophagitis is a clinically important toxicity seen with treatment for locally-advanced non-small cell lung cancer. There is considerable disagreement among prior studies in identifying predictors of radiation esophagitis. We apply machine learning algorithms to identify factors contributing to the development of radiation esophagitis to uncover previously unidentified criteria and more robust dosimetric factors.
MATERIALS AND METHODS
METHODS
We used machine learning approaches to identify predictors of grade ≥ 3 radiation esophagitis in a cohort of 202 consecutive locally-advanced non-small cell lung cancer patients treated with definitive chemoradiation from 2008 to 2016. We evaluated 35 clinical features per patient grouped into risk factors, comorbidities, imaging, stage, histology, radiotherapy, chemotherapy and dosimetry. Univariate and multivariate analyses were performed using a panel of 11 machine learning algorithms combined with predictive power assessments.
RESULTS
RESULTS
All patients were treated to a median dose of 66.6 Gy at 1.8 Gy per fraction using photon (89.6%) and proton (10.4%) beam therapy, most often with concurrent chemotherapy (86.6%). 11.4% of patients developed grade ≥ 3 radiation esophagitis. On univariate analysis, no individual feature was found to predict radiation esophagitis (AUC range 0.45-0.55, p ≥ 0.07). In multivariate analysis, all machine learning algorithms exhibited poor predictive performance (AUC range 0.46-0.56, p ≥ 0.07).
CONCLUSIONS
CONCLUSIONS
Contemporary machine learning algorithms applied to our modern, relatively large institutional cohort could not identify any reliable predictors of grade ≥ 3 radiation esophagitis. Additional patients are needed, and novel patient-specific and treatment characteristics should be investigated to develop clinically meaningful methods to mitigate this survival altering toxicity.
Identifiants
pubmed: 32274426
doi: 10.1016/j.ctro.2020.03.007
pii: S2405-6308(20)30020-3
pmc: PMC7132156
doi:
Types de publication
Journal Article
Langues
eng
Pagination
69-75Commentaires et corrections
Type : ErratumIn
Informations de copyright
© 2020 The Author(s).
Références
Int J Radiat Oncol Biol Phys. 2000 Oct 1;48(3):689-96
pubmed: 11020565
Br J Radiol. 2015;88(1056):20150311
pubmed: 26414340
Int J Radiat Oncol Biol Phys. 2004 Mar 15;58(4):1106-13
pubmed: 15001251
Int J Radiat Oncol Biol Phys. 2003 Feb 1;55(2):337-41
pubmed: 12527046
Adv Radiat Oncol. 2016 Nov 16;2(1):37-43
pubmed: 28740914
Anticancer Res. 2019 Jan;39(1):491-497
pubmed: 30591500
J Clin Oncol. 2010 May 1;28(13):2181-90
pubmed: 20351327
JAMA Oncol. 2016 Mar;2(3):359-67
pubmed: 26606200
Radiother Oncol. 2009 Jun;91(3):282-7
pubmed: 18950881
J Appl Clin Med Phys. 2018 Sep;19(5):539-546
pubmed: 29992732
Radiother Oncol. 2019 Apr;133:106-112
pubmed: 30935565
J Chem Inf Comput Sci. 2003 Nov-Dec;43(6):1947-58
pubmed: 14632445
Semin Radiat Oncol. 2004 Oct;14(4):280-6
pubmed: 15558501
Int J Radiat Oncol Biol Phys. 2012 Nov 15;84(4):1010-6
pubmed: 22920974
Int J Radiat Oncol Biol Phys. 2005 Nov 1;63(3):667-71
pubmed: 15927409
Radiother Oncol. 2005 May;75(2):157-64
pubmed: 15890421
Radiother Oncol. 2020 Mar;144:148-151
pubmed: 31812929
Radiother Oncol. 2005 Nov;77(2):176-81
pubmed: 16256230
Semin Radiat Oncol. 2017 Oct;27(4):370-377
pubmed: 28865520
Int J Radiat Oncol Biol Phys. 2013 Nov 15;87(4):690-6
pubmed: 24035329
BMC Med. 2015 Jan 06;13:1
pubmed: 25563062
J Thorac Oncol. 2016 Feb;11(2):213-21
pubmed: 26718880
Semin Oncol. 2005 Apr;32(2 Suppl 3):S60-6
pubmed: 16015537
J Natl Cancer Inst. 2011 Oct 5;103(19):1452-60
pubmed: 21903745
Radiat Oncol. 2013 May 17;8:122
pubmed: 23680396
Radiother Oncol. 2019 Sep;138:45-51
pubmed: 31146070
Cancer. 2017 Nov 1;123(21):4244-4251
pubmed: 28678434
Int J Radiat Oncol Biol Phys. 2005 Feb 1;61(2):335-47
pubmed: 15667951
Transl Oncol. 2018 Feb;11(1):102-108
pubmed: 29220828
Int J Radiat Oncol Biol Phys. 2004 Nov 15;60(4):1106-12
pubmed: 15519781
Radiat Oncol. 2016 May 04;11:66
pubmed: 27142674
Clin Lung Cancer. 2011 Jul;12(4):245-51
pubmed: 21726824
Lancet Oncol. 2015 Feb;16(2):187-99
pubmed: 25601342
Int J Radiat Oncol Biol Phys. 1999 Aug 1;45(1):97-103
pubmed: 10477012