Comparison of machine learning algorithms and oversampling techniques for urinary toxicity prediction after prostate cancer radiotherapy

被引:2
作者
Mylona, Eugenia [1 ]
Lebreton, Clement [1 ]
Fontaine, Pierre [2 ]
Supiot, Stephane [3 ]
Magne, Nicolas [4 ]
Crehange, Gilles [5 ]
de Crevoisier, Renaud [6 ]
Acosta, Oscar [1 ]
机构
[1] Univ Rennes, UMR 1099, LTSI, INSERM, Rennes, France
[2] Univ Rennes, UMR 1099, LTSI, INSERM,HES SO, Rennes, France
[3] Ctr Georges Franois Leclerc, Dept Radiat Oncol, Dijon, France
[4] Lucien Neuwirth Canc Inst, Dept Radiotherapy, St Priest En Jarez, France
[5] Inst Cancrol Ouest, Dept Med Phys, St Herblain, France
[6] Univ Rennes, INSERM, LTSI UMR 1099, CLCC E Marquis, Rennes, France
来源
2019 IEEE 19TH INTERNATIONAL CONFERENCE ON BIOINFORMATICS AND BIOENGINEERING (BIBE) | 2019年
关键词
Prostate cancer radiotherapy; Imbalanced data; Machine Learning; Radiotherapy; Urinary toxicity; MODELS; SMOTE;
D O I
10.1109/BIBE.2019.00180
中图分类号
R318 [生物医学工程];
学科分类号
0831 ;
摘要
Prostate cancer radiotherapy unavoidably involves the irradiation not only of the target volume, but also of healthy organs-at-risk, neighboring the prostate, likely causing adverse, toxicity-related side-effects. Specifically, in the case of urinary toxicity, these side effects might be associated with a variety of dosimetric, clinical and genetic factors, making its prediction particularly challenging. Given the inconsistency of available data concerning radiation-induced toxicity, it is crucial to develop robust models with superior predictive performance in order to perform tailored treatments. Machine Learning techniques emerge as appealing in this context, nevertheless without any consensus on the best algorithms to be used. This work proposes a comparison of several machine-learning strategies together with different minority class oversampling techniques for prediction of urinary toxicity following prostate cancer radiotherapy using dosimetric and clinical data. The performance of these classifiers was evaluated on the original dataset and using four different synthetic oversampling techniques. The area under the ROC curve (AUC) and the F-measure were employed to evaluate their performance. Results suggest that, regardless of the technique, oversampling always increases the prediction performance of the models (p=0.004). Overall, oversampling with Synthetic Minority Oversampling Technique (SMOTE) followed by Edited Nearest Neighbour algorithm (ENN) together with Regularized Discriminant Analysis (RDA) classifier provide the best performance (AUC=0.71).
引用
收藏
页码:964 / 971
页数:8
相关论文
共 27 条
  • [1] [Anonymous], 1996, REGRESSION SHRINKAGE
  • [2] 70 GY VERSUS 80 GY IN LOCALIZED PROSTATE CANCER: 5-YEAR RESULTS OF GETUG 06 RANDOMIZED TRIAL
    Beckendorf, Veronique
    Guerif, Stephane
    Le Prise, Elisabeth
    Cosset, Jean-Marc
    Bougnoux, Agnes
    Chauvet, Bruno
    Salem, Naji
    Chapet, Olivier
    Bourdain, Sylvain
    Bachaud, Jean-Marc
    Maingon, Philippe
    Hannoun-Levi, Jean-Michel
    Malissard, Luc
    Simon, Jean-Marc
    Pommier, Pascal
    Hay, Men
    Dubray, Bernard
    Lagrange, Jean-Leon
    Luporsi, Elisabeth
    Bey, Pierre
    [J]. INTERNATIONAL JOURNAL OF RADIATION ONCOLOGY BIOLOGY PHYSICS, 2011, 80 (04): : 1056 - 1063
  • [3] SMOTE: Synthetic minority over-sampling technique
    Chawla, Nitesh V.
    Bowyer, Kevin W.
    Hall, Lawrence O.
    Kegelmeyer, W. Philip
    [J]. 2002, American Association for Artificial Intelligence (16)
  • [4] Integrated models for the prediction of late genitourinary complaints after high-dose intensity modulated radiotherapy for prostate cancer: Making informed decisions
    De Langhe, Sofie
    De Meerleer, Gert
    De Ruyck, Kim
    Ost, Piet
    Fonteyne, Valerie
    De Neve, Wilfried
    Thierens, Hubert
    [J]. RADIOTHERAPY AND ONCOLOGY, 2014, 112 (01) : 95 - 99
  • [5] Escalated-dose versus control-dose conformal radiotherapy for prostate cancer: long-term results from the MRC RT01 randomised controlled trial
    Dearnaley, David P.
    Jovic, Gordana
    Syndikus, Isabel
    Khoo, Vincent
    Cowan, Richard A.
    Graham, John D.
    Aird, Edwin G.
    Bottomley, David
    Huddart, Robert A.
    Jose, Chakiath C.
    Matthews, John H. L.
    Millar, Jeremy L.
    Murphy, Claire
    Russell, J. Martin
    Scrase, Christopher D.
    Parmar, Mahesh K. B.
    Sydes, Matthew R.
    [J]. LANCET ONCOLOGY, 2014, 15 (04) : 464 - 473
  • [6] Dumancas Gerard, 2015, COMP MACHINE LEARNIN
  • [7] Bayes' Theorem in the 21st Century
    Efron, Bradley
    [J]. SCIENCE, 2013, 340 (6137) : 1177 - 1178
  • [8] Ferri C., 2009, PATTERN RECOGNITION
  • [9] Predictive Models of Toxicity in External Radiotherapy Dosimetric Issues
    Fiorino, Claudio
    Rancati, Tiziana
    Valdagni, Riccardo
    [J]. CANCER, 2009, 115 (13) : 3135 - 3140
  • [10] Fordellone M., 2018, Partial least squares discriminant analysis: A dimensionality reduction method to classify hyperspectral data