Multi-pAradigM voxel-Based Analysis: a computational cookbot
The Matlab toolbox MAMBA is designed for the flexible application of statistical voxel-based (VB) analysis in different scenarios in medical imaging and radiation oncology. It provides open-source functions to compute VB statistical models of the input data, according to a great variety of regression schemes, and to derive VB maps of the observed significance level, performing a non-parametric permutation inference. The toolbox allows for including VB and global outcomes, as well as an arbitrary amount of VB and global explanatory variables. In addition, the Matlab Parallel Computing Toolbox is exploited to take advantage of the perfect parallelizability of most workloads.
MAMBA is an open-source toolbox, freely available for academic and non-commercial purposes. It is designed to make state-of-the-art VB analysis accessible to research scientists without the programming resources needed to build from scratch their own software solutions. At the same time, the source code is handed out for more experienced users to complement their own tools, also customizing user-defined models .
Current version of MAMBA represents the computational companion to an initial published VB analysis methodological cookbook where the key elements and possible options for accurate VB statistical inference on dose maps were discussed, and the underlying mathematics was theoretically elucidated [1]. In the computational cookbot introduced with the current version, several novel strategies were implemented to address problems encountered in many different clinical studies performed in the last years. In this way, MAMBA is intended to make the statistical segment of VB analysis accessible to radiation oncology community and to enable investigation on complex radiobiological mechanisms. At the same time, thanks to its high flexibility, MAMBA makes the methods originally developed for specific needs of the RT field available for the medical imaging community [2].
Features
The MAMBA toolbox consists of four main functions:
- image_read;
- image_clin_merge;
- VBmodel;
- perm_test.
They are meant to be directly called by the user, as opposed to a set of auxiliary functions (stored in the subfolders Engine/ and External/), which are required by the main functions to perform internal steps and should not be considered of direct interest to the end-user performing a VB analysis. The tasks performed by the four main functions along with their interaction and usage are illustrated in Fig. 1) [2].
The detailed description of the toolbox functions can be found in the Chapter “Reference manual” of the User’s Guide.
Documentation (includes the extensive reference manual and several examples that the user can fully work out on a synthetic dataset, as well as the toolbox configurations that led to clinical results previously published in the literature)
References
Users are encouraged to freely adapt MAMBA according to their needs, and assume all responsibility and risk with respect to their use of the toolbox, which is provided “AS IS”. In addition, users are welcome to cite the following references, anywhere they use MAMBA:
[1] Palma, G., Monti, S. and Cella, L., 2020. Voxel-based analysis in radiation oncology: A methodological cookbook. Physica Medica, 69, pp.192-204. DOI: 10.1016/j.ejmp.2019.12.013
[2] Palma, G., Cella, L. and Monti, S., 2023. MAMBA—Multi‐pAradigM voxel‐Based Analysis: A computational cookbot. Medical Physics. DOI: 10.1002/mp.16260
NMMItools 