Premium
Predicting outcomes of pelvic exenteration using machine learning
Author(s) -
Ivan Dudurych,
Michael E. Kelly,
Arend G. J. Aalbers,
Nauman Aziz,
Nuno Abecasis,
Mirna Abraham-Nordling,
Takashi Akiyoshi,
W Alberda,
Matthew L. Albert,
Mihailo Andric,
Eva Angenete,
Afroditi Antoniou,
R Auer,
Kirk K. S. Austin,
Omer Aziz,
R. P. Baker,
M Bali,
Gediminas Baseckas,
Brendan Bebington,
Matthew Bedford,
Brian Keith Bednarski,
Geerard L. Beets,
P L Berg,
J Bey,
Sebastiano Biondo,
K. Boyle,
L Bordeianou,
A B Bremers,
Markus Brunner,
Pamela Buchwald,
Ai-Tram N. Bui,
Andrea Burgess,
Jacobus W. A. Burger,
D Burling,
Evan Burns,
Nicholas Campain,
Sara Carvalhal,
Luis M. Castro,
Antonio CaycedoMarulanda,
Kkl Chan,
George J. Chang,
Min Hoe Chew,
A. K. Chok,
Peter Siao Tick Chong,
Henrik Christensen,
H Clouston,
Mary Codd,
Danielle Collins,
A J Colquhoun,
Alison Corr,
Maurizio Coscia,
Peter Coyne,
Ben Creavin,
Roland S. Croner,
L Damjanovic,
Ian Daniels,
Mark Davies,
Richard J. Davies,
Clare Delaney,
Johannes H. W. de Wilt,
Quentin Denost,
C Deutsch,
D. Dietz,
Sebastián Domingo,
Eric J. Dozois,
M. J. Duff,
Tim Eglinton,
J M Enrique-Navascues,
Eloy EspínBasany,
Martyn Evans,
Nicola Fearnhead,
Kjersti Flatmark,
Fergal J. Fleming,
Frank A. Frizelle,
Mario Álvarez Gallego,
Eduardo GarcíaGranero,
J.L. García-Sabrido,
Lorenzo Gentilini,
Mark George,
Vijay George,
Laurent Ghouti,
Francisco Giner,
N Ginther,
R Glynn,
Thomas Golda,
Ben Griffiths,
Dean Harris,
J.A.W. Hagemans,
Vishwanath Hanchanale,
Deena Harji,
RM Helewa,
Alexander G. Heriot,
David Hochman,
Werner Hohenberger,
T. Holm,
Roel Hompes,
Jordan Jenkins,
Samuel D. Kaffenberger,
G V Kandaswamy,
Sandeep Kapur,
Yukihide Kanemitsu,
S R Kelley,
Deborah S. Keller,
Mohiuddeen Khan,
Ravi P. Kiran,
H. Kim,
H. J. Kim,
Christopher Koh,
Niels F. M. Kok,
R Kokelaar,
Christos Kontovounisios,
Helle Kristensen,
Hidde M. Kroon,
Miranda Kusters,
Vanessa Carvalho do Lago,
Stein Gunnar Larsen,
D W Larson,
WL Law,
Søren Laurberg,
P. J. Lee,
M Limbert,
Marie Louise Lydrup,
A Lyons,
A. C. Lynch,
Christopher R. Mantyh,
Kellie L. Mathis,
C. F. S. Margues,
Anna Martling,
W. J. H. J. Meijerink,
Susanne Merkel,
A M Mehta,
D. Ray McArthur,
Frank McDermott,
J. S. McGrath,
Sachin Malde,
A. Mirnezami,
J. R. T. Monson,
J R Morton,
Tamara G. Mullaney,
Ionuţ Negoi,
J W M Neto,
Binh P. Nguyen,
M. B. Nielsen,
G. Nieuwenhuijzen,
Per J. Nilsson,
Alex Oliver,
Petra O’Connell,
S T O’Dwyer,
Geraint Palmer,
Emmanouil P. Pappou,
J. Park,
Dimitrios Patsouras,
Gianluca Pellino,
Alexandra Peterson,
Gilberto Poggioli,
David Proud,
Martha Quinn,
Aaron Quyn,
Rami Radwan,
Shahnawaz Rasheed,
Rasmussen Pc,
S E Regenbogen,
Andrew G Renehan,
Rafael Ramos da Rocha,
Mark Rochester,
Jitender Rohila,
Joost Rothbarth,
Matteo Rottoli,
C Roxburgh,
H.J.T. Rutten,
E. James Ryan,
Bashar Safar,
P. M. Sagar,
Amit Sahai,
Avanish Saklani,
Tarik Sammour,
Rameez Sayyed,
Alexis Schizas,
Eugenia Schwarzkopf,
Viorel Scripcariu,
Chelliah Selvasekar,
I. Shaikh,
G. Shellawell,
Dai Shida,
A. Simpson,
Neil Smart,
Philip Smart,
Julia Smith,
A M Solbakken,
Michael J. Solomon,
Mette Møller Sørensen,
Scott R. Steele,
Daniel Steffens,
Karyn B. Stitzenberg,
Luca Stocchi,
Nicholas Stylianides,
Torbjörn Swartling,
H Sumrien,
Paul Sutton,
T Swartking,
Eric J. Tan,
Claire Taylor,
Paris Tekkis,
J Teras,
Ramesh Thurairaja,
EeLin Toh,
Petr Tsarkov,
Yuichiro Tsukada,
Shunsuke Tsukamoto,
J J Tuech,
William H. Turner,
Jurriaan B. Tuynman,
Gabriëlle H. van Ramshorst,
D van Zoggel,
W Vásquez-Jiménez,
Cornelis Verhoef,
Giuseppe Vizzielli,
E L K Voogt,
Kay Uehara,
Chris Wakeman,
Satish Warrier,
Hans H. Wasmuth,
K Weber,
Martin R. Weiser,
James M. Wheeler,
J Wild,
Margaret Wilson,
Albert Wolthuis,
H Yano,
Benjamin Cherng Hann Yip,
Jeremy Yip,
Ri Na Yoo,
Desmond C. Winter
Publication year - 2020
Publication title -
colorectal disease
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.029
H-Index - 89
eISSN - 1463-1318
pISSN - 1462-8910
DOI - 10.1111/codi.15235
Subject(s) - medicine , logistic regression , pelvic exenteration , receiver operating characteristic , machine learning , artificial neural network , artificial intelligence , predictive power , support vector machine , test set , surgery , computer science , philosophy , epistemology
Aim We aim to compare machine learning with neural network performance in predicting R0 resection (R0), length of stay > 14 days (LOS), major complication rates at 30 days postoperatively (COMP) and survival greater than 1 year (SURV) for patients having pelvic exenteration for locally advanced and recurrent rectal cancer. Method A deep learning computer was built and the programming environment was established. The PelvEx Collaborative database was used which contains anonymized data on patients who underwent pelvic exenteration for locally advanced or locally recurrent colorectal cancer between 2004 and 2014. Logistic regression, a support vector machine and an artificial neural network (ANN) were trained. Twenty per cent of the data were used as a test set for calculating prediction accuracy for R0, LOS, COMP and SURV. Model performance was measured by plotting receiver operating characteristic (ROC) curves and calculating the area under the ROC curve (AUROC). Results Machine learning models and ANNs were trained on 1147 cases. The AUROC for all outcome predictions ranged from 0.608 to 0.793 indicating modest to moderate predictive ability. The models performed best at predicting LOS > 14 days with an AUROC of 0.793 using preoperative and operative data. Visualized logistic regression model weights indicate a varying impact of variables on the outcome in question. Conclusion This paper highlights the potential for predictive modelling of large international databases. Current data allow moderate predictive ability of both complex ANNs and more classic methods.