A Enshaei1, C N Robson1, R J Edmondson2. 1. Medical School, Northern Institute for Cancer Research, University of Newcastle Upon Tyne, Newcastle upon Tyne, UK. 2. Chair of Gynaecological Oncology, Faculty Institute for Cancer Sciences, University of Manchester, Manchester Academic Health Science Centre, St. Mary's Hospital, Manchester, UK. richard.edmondson@manchester.ac.uk.
Abstract
BACKGROUND: The ability to provide accurate prognostic and predictive information to patients is becoming increasingly important as clinicians enter an era of personalized medicine. For a disease as heterogeneous as epithelial ovarian cancer, conventional algorithms become too complex for routine clinical use. This study therefore investigated the potential for an artificial intelligence model to provide this information and compared it with conventional statistical approaches. METHODS: The authors created a database comprising 668 cases of epithelial ovarian cancer during a 10-year period and collected data routinely available in a clinical environment. They also collected survival data for all the patients, then constructed an artificial intelligence model capable of comparing a variety of algorithms and classifiers alongside conventional statistical approaches such as logistic regression. RESULTS: The model was used to predict overall survival and demonstrated that an artificial neural network (ANN) algorithm was capable of predicting survival with high accuracy (93 %) and an area under the curve (AUC) of 0.74 and that this outperformed logistic regression. The model also was used to predict the outcome of surgery and again showed that ANN could predict outcome (complete/optimal cytoreduction vs. suboptimal cytoreduction) with 77 % accuracy and an AUC of 0.73. CONCLUSIONS: These data are encouraging and demonstrate that artificial intelligence systems may have a role in providing prognostic and predictive data for patients. The performance of these systems likely will improve with increasing data set size, and this needs further investigation.
BACKGROUND: The ability to provide accurate prognostic and predictive information to patients is becoming increasingly important as clinicians enter an era of personalized medicine. For a disease as heterogeneous as epithelial ovarian cancer, conventional algorithms become too complex for routine clinical use. This study therefore investigated the potential for an artificial intelligence model to provide this information and compared it with conventional statistical approaches. METHODS: The authors created a database comprising 668 cases of epithelial ovarian cancer during a 10-year period and collected data routinely available in a clinical environment. They also collected survival data for all the patients, then constructed an artificial intelligence model capable of comparing a variety of algorithms and classifiers alongside conventional statistical approaches such as logistic regression. RESULTS: The model was used to predict overall survival and demonstrated that an artificial neural network (ANN) algorithm was capable of predicting survival with high accuracy (93 %) and an area under the curve (AUC) of 0.74 and that this outperformed logistic regression. The model also was used to predict the outcome of surgery and again showed that ANN could predict outcome (complete/optimal cytoreduction vs. suboptimal cytoreduction) with 77 % accuracy and an AUC of 0.73. CONCLUSIONS: These data are encouraging and demonstrate that artificial intelligence systems may have a role in providing prognostic and predictive data for patients. The performance of these systems likely will improve with increasing data set size, and this needs further investigation.
Authors: Adam M Sonabend; Brad E Zacharia; Michael B Cloney; Aarón Sonabend; Christopher Showers; Victoria Ebiana; Matthew Nazarian; Kristin R Swanson; Anne Baldock; Henry Brem; Jeffrey N Bruce; William Butler; Daniel P Cahill; Bob Carter; Daniel A Orringer; David W Roberts; Oren Sagher; Nader Sanai; Theodore H Schwartz; Daniel L Silbergeld; Michael B Sisti; Reid C Thompson; Allen E Waziri; Zoher Ghogawala; Guy McKhann Journal: Neurosurgery Date: 2017-04-01 Impact factor: 4.654
Authors: David D Bowtell; Steffen Böhm; Ahmed A Ahmed; Paul-Joseph Aspuria; Robert C Bast; Valerie Beral; Jonathan S Berek; Michael J Birrer; Sarah Blagden; Michael A Bookman; James D Brenton; Katherine B Chiappinelli; Filipe Correia Martins; George Coukos; Ronny Drapkin; Richard Edmondson; Christina Fotopoulou; Hani Gabra; Jérôme Galon; Charlie Gourley; Valerie Heong; David G Huntsman; Marcin Iwanicki; Beth Y Karlan; Allyson Kaye; Ernst Lengyel; Douglas A Levine; Karen H Lu; Iain A McNeish; Usha Menon; Steven A Narod; Brad H Nelson; Kenneth P Nephew; Paul Pharoah; Daniel J Powell; Pilar Ramos; Iris L Romero; Clare L Scott; Anil K Sood; Euan A Stronach; Frances R Balkwill Journal: Nat Rev Cancer Date: 2015-11 Impact factor: 60.716
Authors: Andrew Wen; Sunyang Fu; Sungrim Moon; Mohamed El Wazir; Andrew Rosenbaum; Vinod C Kaggal; Sijia Liu; Sunghwan Sohn; Hongfang Liu; Jungwei Fan Journal: NPJ Digit Med Date: 2019-12-17
Authors: Alexandros Laios; Raissa Vanessa De Oliveira Silva; Daniel Lucas Dantas De Freitas; Yong Sheng Tan; Gwendolyn Saalmink; Albina Zubayraeva; Racheal Johnson; Angelika Kaufmann; Mohammed Otify; Richard Hutson; Amudha Thangavelu; Tim Broadhead; David Nugent; Georgios Theophilou; Kassio Michell Gomes de Lima; Diederick De Jong Journal: J Clin Med Date: 2021-12-24 Impact factor: 4.241