Jan C Peeken1,2, Josefine Hesse3, Bernhard Haller4, Kerstin A Kessel3,4,5, Fridtjof Nüsslin3, Stephanie E Combs3,6,5. 1. Department of Radiation Oncology, Klinikum rechts der Isar, Technical University Munich (TUM), Ismaninger Straße 22, 81675, Munich, Germany. jan.peeken@tum.de. 2. Deutsches Konsortium für Translationale Krebsforschung (DKTK), Partner Site Munich, Munich, Germany. jan.peeken@tum.de. 3. Department of Radiation Oncology, Klinikum rechts der Isar, Technical University Munich (TUM), Ismaninger Straße 22, 81675, Munich, Germany. 4. Institut for Medical Statistics and Epidemiology, Technical University Munich (TUM), Ismaninger Straße 22, 81675, Munich, Germany. 5. Deutsches Konsortium für Translationale Krebsforschung (DKTK), Partner Site Munich, Munich, Germany. 6. Institute of Innovative Radiotherapy (iRT), Department of Radiation Sciences (DRS), Helmholtz Zentrum München, Ingolstaedter Landstraße 1, 85764, Neuherberg, Germany.
Abstract
BACKGROUND: For glioblastoma (GBM), multiple prognostic factors have been identified. Semantic imaging features were shown to be predictive for survival prediction. No similar data have been generated for the prediction of progression. The aim of this study was to assess the predictive value of the semantic visually accessable REMBRANDT [repository for molecular brain neoplasia data] images (VASARI) imaging feature set for progression and survival, and the creation of joint prognostic models in combination with clinical and pathological information. METHODS: 189 patients were retrospectively analyzed. Age, Karnofsky performance status, gender, and MGMT promoter methylation and IDH mutation status were assessed. VASARI features were determined on pre- and postoperative MRIs. Predictive potential was assessed with univariate analyses and Kaplan-Meier survival curves. Following variable selection and resampling, multivariate Cox regression models were created. Predictive performance was tested on patient test sets and compared between groups. The frequency of selection for single variables and variable pairs was determined. RESULTS: For progression free survival (PFS) and overall survival (OS), univariate significant associations were shown for 9 and 10 VASARI features, respectively. Multivariate models yielded concordance indices significantly different from random for the clinical, imaging, combined, and combined + MGMT models of 0.657, 0.636, 0.694, and 0.716 for OS, and 0.602, 0.604, 0.633, and 0.643 for PFS. "Multilocality," "deep white-matter invasion," "satellites," and "ependymal invasion" were over proportionally selected for multivariate model generation, underlining their importance. CONCLUSIONS: We demonstrated a predictive value of several qualitative imaging features for progression and survival. The performance of prognostic models was increased by combining clinical, pathological, and imaging features.
BACKGROUND: For glioblastoma (GBM), multiple prognostic factors have been identified. Semantic imaging features were shown to be predictive for survival prediction. No similar data have been generated for the prediction of progression. The aim of this study was to assess the predictive value of the semantic visually accessable REMBRANDT [repository for molecular brain neoplasia data] images (VASARI) imaging feature set for progression and survival, and the creation of joint prognostic models in combination with clinical and pathological information. METHODS: 189 patients were retrospectively analyzed. Age, Karnofsky performance status, gender, and MGMT promoter methylation and IDH mutation status were assessed. VASARI features were determined on pre- and postoperative MRIs. Predictive potential was assessed with univariate analyses and Kaplan-Meier survival curves. Following variable selection and resampling, multivariate Cox regression models were created. Predictive performance was tested on patient test sets and compared between groups. The frequency of selection for single variables and variable pairs was determined. RESULTS: For progression free survival (PFS) and overall survival (OS), univariate significant associations were shown for 9 and 10 VASARI features, respectively. Multivariate models yielded concordance indices significantly different from random for the clinical, imaging, combined, and combined + MGMT models of 0.657, 0.636, 0.694, and 0.716 for OS, and 0.602, 0.604, 0.633, and 0.643 for PFS. "Multilocality," "deep white-matter invasion," "satellites," and "ependymal invasion" were over proportionally selected for multivariate model generation, underlining their importance. CONCLUSIONS: We demonstrated a predictive value of several qualitative imaging features for progression and survival. The performance of prognostic models was increased by combining clinical, pathological, and imaging features.
Authors: W J Curran; C B Scott; J Horton; J S Nelson; A S Weinstein; A J Fischbach; C H Chang; M Rotman; S O Asbell; R E Krisch Journal: J Natl Cancer Inst Date: 1993-05-05 Impact factor: 13.506
Authors: Roger Stupp; Monika E Hegi; Warren P Mason; Martin J van den Bent; Martin J B Taphoorn; Robert C Janzer; Samuel K Ludwin; Anouk Allgeier; Barbara Fisher; Karl Belanger; Peter Hau; Alba A Brandes; Johanna Gijtenbeek; Christine Marosi; Charles J Vecht; Karima Mokhtari; Pieter Wesseling; Salvador Villa; Elizabeth Eisenhauer; Thierry Gorlia; Michael Weller; Denis Lacombe; J Gregory Cairncross; René-Olivier Mirimanoff Journal: Lancet Oncol Date: 2009-03-09 Impact factor: 41.316
Authors: Maximilian Diehn; Christine Nardini; David S Wang; Susan McGovern; Mahesh Jayaraman; Yu Liang; Kenneth Aldape; Soonmee Cha; Michael D Kuo Journal: Proc Natl Acad Sci U S A Date: 2008-03-24 Impact factor: 11.205
Authors: David A Gutman; Lee A D Cooper; Scott N Hwang; Chad A Holder; Jingjing Gao; Tarun D Aurora; William D Dunn; Lisa Scarpace; Tom Mikkelsen; Rajan Jain; Max Wintermark; Manal Jilwan; Prashant Raghavan; Erich Huang; Robert J Clifford; Pattanasak Mongkolwat; Vladimir Kleper; John Freymann; Justin Kirby; Pascal O Zinn; Carlos S Moreno; Carl Jaffe; Rivka Colen; Daniel L Rubin; Joel Saltz; Adam Flanders; Daniel J Brat Journal: Radiology Date: 2013-02-07 Impact factor: 11.105
Authors: Michael Weller; Jörg Felsberg; Christian Hartmann; Hilmar Berger; Joachim P Steinbach; Johannes Schramm; Manfred Westphal; Gabriele Schackert; Matthias Simon; Jörg C Tonn; Oliver Heese; Dietmar Krex; Guido Nikkhah; Torsten Pietsch; Otmar Wiestler; Guido Reifenberger; Andreas von Deimling; Markus Loeffler Journal: J Clin Oncol Date: 2009-10-05 Impact factor: 44.544
Authors: Jan C Peeken; Matthew B Spraker; Carolin Knebel; Hendrik Dapper; Daniela Pfeiffer; Michal Devecka; Ahmed Thamer; Mohamed A Shouman; Armin Ott; Rüdiger von Eisenhart-Rothe; Fridtjof Nüsslin; Nina A Mayr; Matthew J Nyflot; Stephanie E Combs Journal: EBioMedicine Date: 2019-09-12 Impact factor: 8.143