Krüger, Julia
Opfer, Roland
Spies, Lothar
Hedderich, Dennis
Buchert, Ralph http://orcid.org/0000-0002-0945-0724
Funding for this research was provided by:
Universitätsklinikum Hamburg-Eppendorf (UKE)
Article History
Received: 28 January 2023
Revised: 19 August 2023
Accepted: 26 August 2023
First Online: 9 November 2023
Declarations
:
: The scientific guarantor of this publication is Ralph Buchert.
: The authors of this manuscript declare relationships with the following companies: Julia Krüger, Roland Opfer, and Lothar Spies are employees of jung diagnostics GmbH, Germany (). There is no actual or potential conflict of interest for the other authors. The non-employee authors had control of the data and information that might present a conflict of interest for the employee authors.
: No complex statistical methods were necessary for this paper.
: Written informed consent was waived by the ethics review board of the General Medical Council of the state of Hamburg, Germany, and by the ethics committee of the Technical University of Munich.
: The MRI data of the training set and of the multiple-scanner normal database had been transferred to jung diagnostics GmbH under the terms and conditions of the European General Data Protection Regulation for remote image analysis. Subsequently, the data had been anonymized. The need for written informed consent for the retrospective use of the anonymized data in the present study was waived by the ethics review board of the General Medical Council of the state of Hamburg, Germany.The use of the test set and its normal database for retrospective research was approved by the ethics committee of the Technical University of Munich (reference number 622/20S).
: The test dataset comprising T1w-MRI from 118 subjects was reported previously []. This prior article dealt with the utility of conventional voxel-based morphometry with reference to a scanner- and sequence-specific normal database to support the differential diagnosis of dementing neurodegenerative diseases whereas the current manuscript deals with the design, training, and testing of a convolutional neural network for voxel-based morphometry without a normal database. There is no further overlap with previously published studies or work currently undergoing review or in press at a journal. Neither the training dataset of 8945 MRI scans nor the convolutional neural network was described previously.
: • retrospective• diagnostic or prognostic study• multi-center study