Monika Engelhardt1,2, Gabriele Ihorst3,4, Martin Schumacher3,5, Michael Rassner6,3, Laura Gengenbach6,3, Mandy Möller6,3, Khalid Shoumariyeh6,3, Jakob Neubauer3,7, Juliane Farthmann3,8, Georg Herget3,9, Ralph Wäsch6,3. 1. University of Freiburg Medical Center, Hematology & Oncology, Faculty of Medicine, Hugstetterstr. 53, 79106, Freiburg, Germany. monika.engelhardt@uniklinik-freiburg.de. 2. Comprehensive Cancer Center Freiburg (CCCF), Freiburg im Breisgau, Germany. monika.engelhardt@uniklinik-freiburg.de. 3. Comprehensive Cancer Center Freiburg (CCCF), Freiburg im Breisgau, Germany. 4. Clinical Trials Unit, Faculty of Medicine, Freiburg im Breisgau, Germany. 5. Medical Biometry and Statistics (IMBI), University of Freiburg, Faculty of Medicine, Freiburg im Breisgau, Germany. 6. University of Freiburg Medical Center, Hematology & Oncology, Faculty of Medicine, Hugstetterstr. 53, 79106, Freiburg, Germany. 7. Department of Radiology, Medical Center-University of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg, Germany. 8. Department of Obstetrics and Gynecology, University Medical Center, Faculty of Medicine, Freiburg im Breisgau, Germany. 9. Department of Orthopaedics and Trauma Surgery, Medical Center, University of Freiburg, Faculty of Medicine, Freiburg, Germany.
Abstract
BACKGROUND: The standard to ensure utmost cancer treatment is a prerequisite in national cancer plans for comprehensive cancer centers (CCCs) and ensured through multidisciplinary tumor boards (MTBs). Despite these being compulsory for CCCs, various analyses on MTBs have been performed, since MTBs are resource-intensive. Outcome measures in these prior analyses had been survival (OS), MTB-adherence and -satisfaction, inclusion of patients into clinical trials and better cancer care. MAIN BODY: A publication from Freytag et al. performed an analysis in multiple tumor entities and assessed the effect of number of MTBs. By matched-pair analysis, they compared response and OS of patients, whose cases were discussed in MTBs vs. those that were not. The analysis included 454 patients and 66 different tumor types. Only patients with > 3 MTBs showed a significantly better OS than patients with no MTB meeting. Response to treatment, relapse free survival and time to progression were not found to be better, nor was there any difference for a specific tumor entity with vs. without MTB discussions. An in-depth discussion of these results, with respect to the literature (PubMed search: "MTBs AND cancer") and within the author group, including statisticians specialized in data analysis of cancer patients and questions addressed in MTBs, was performed to interpret these findings. We conclude that the results by Freytag et al. are deceiving due to an "immortal time bias" that requires more careful data interpretation. CONCLUSIONS: The result of Freytag et al. of a seemingly positive impact of higher number of MTBs needs to be interpreted cautiously: their presumed better OS in patients with > 3 MTB discussions is misleading, due to an immortal time bias. Here patients need to survive long enough to be discussed more often. Therefore, these results should not lead to the conclusion that more MTBs will "automatically" increase cancer patients' OS, rather than that the insightful discussion, at best in MTBs and with statisticians, will generate meaningful advice, that is important for cancer patients.
BACKGROUND: The standard to ensure utmost cancer treatment is a prerequisite in national cancer plans for comprehensive cancer centers (CCCs) and ensured through multidisciplinary tumor boards (MTBs). Despite these being compulsory for CCCs, various analyses on MTBs have been performed, since MTBs are resource-intensive. Outcome measures in these prior analyses had been survival (OS), MTB-adherence and -satisfaction, inclusion of patients into clinical trials and better cancer care. MAIN BODY: A publication from Freytag et al. performed an analysis in multiple tumor entities and assessed the effect of number of MTBs. By matched-pair analysis, they compared response and OS of patients, whose cases were discussed in MTBs vs. those that were not. The analysis included 454 patients and 66 different tumor types. Only patients with > 3 MTBs showed a significantly better OS than patients with no MTB meeting. Response to treatment, relapse free survival and time to progression were not found to be better, nor was there any difference for a specific tumor entity with vs. without MTB discussions. An in-depth discussion of these results, with respect to the literature (PubMed search: "MTBs AND cancer") and within the author group, including statisticians specialized in data analysis of cancer patients and questions addressed in MTBs, was performed to interpret these findings. We conclude that the results by Freytag et al. are deceiving due to an "immortal time bias" that requires more careful data interpretation. CONCLUSIONS: The result of Freytag et al. of a seemingly positive impact of higher number of MTBs needs to be interpreted cautiously: their presumed better OS in patients with > 3 MTB discussions is misleading, due to an immortal time bias. Here patients need to survive long enough to be discussed more often. Therefore, these results should not lead to the conclusion that more MTBs will "automatically" increase cancer patients' OS, rather than that the insightful discussion, at best in MTBs and with statisticians, will generate meaningful advice, that is important for cancer patients.
Authors: Marius Freytag; Ulrich Herrlinger; Stefan Hauser; Franz G Bauernfeind; Maria A Gonzalez-Carmona; Jennifer Landsberg; Jens Buermann; Hartmut Vatter; Tobias Holderried; Thorsten Send; Martin Schumacher; Arne Koscielny; Georg Feldmann; Mario Heine; Dirk Skowasch; Niklas Schäfer; Benjamin Funke; Michael Neumann; Ingo G H Schmidt-Wolf Journal: BMC Cancer Date: 2020-04-28 Impact factor: 4.430
Authors: Francesca Gay; Monika Engelhardt; Evangelos Terpos; Ralph Wäsch; Luisa Giaccone; Holger W Auner; Jo Caers; Martin Gramatzki; Niels van de Donk; Stefania Oliva; Elena Zamagni; Laurent Garderet; Christian Straka; Roman Hajek; Heinz Ludwig; Herman Einsele; Meletios Dimopoulos; Mario Boccadoro; Nicolaus Kröger; Michele Cavo; Hartmut Goldschmidt; Benedetto Bruno; Pieter Sonneveld Journal: Haematologica Date: 2017-12-07 Impact factor: 9.941