A field strength independent MR radiomics model to predict pathological complete response in locally advanced rectal can
- PDF / 1,441,804 Bytes
- 9 Pages / 595.276 x 790.866 pts Page_size
- 77 Downloads / 187 Views
MAGNETIC RESONANCE IMAGING
A field strength independent MR radiomics model to predict pathological complete response in locally advanced rectal cancer Davide Cusumano1 · Gert Meijer2 · Jacopo Lenkowicz3 · Giuditta Chiloiro1 · Luca Boldrini1 · Carlotta Masciocchi3 · Nicola Dinapoli1 · Roberto Gatta3 · Calogero Casà3 · Andrea Damiani3 · Brunella Barbaro1 · Maria Antonietta Gambacorta1 · Luigi Azario1 · Marco De Spirito1 · Martijn Intven2 · Vincenzo Valentini1 Received: 27 February 2020 / Accepted: 12 August 2020 © The Author(s) 2020
Abstract Purpose Aim of this study was to develop a generalised radiomics model for predicting pathological complete response after neoadjuvant chemo-radiotherapy in locally advanced rectal cancer patients using pre-CRT T2-weighted images acquired at a 1.5 T and a 3 T scanner. Methods In two institutions, 195 patients were scanned: 136 patients were scanned on a 1.5 T MR scanner, 59 patients on a 3 T MR scanner. Gross tumour volumes were delineated on the MR images and 496 radiomic features were extracted, applying the intensity-based (IB) filter. Features were standardised with Z-score normalisation and an initial feature selection was carried out using Wilcoxon–Mann–Whitney test: The most significant features at 1.5 T and 3 T were selected as main features. Several logistic regression models combining the main features with a third one selected by those resulting significant were elaborated and evaluated in terms of area under curve (AUC). A tenfold cross-validation was repeated 300 times to evaluate the model robustness. Results Three features were selected: maximum fractal dimension with IB = 0–50, energy and grey-level non-uniformity calculated on the run-length matrix with IB = 0–50. The AUC of the model applied to the whole dataset after cross-validation was 0.72, while values of 0.70 and 0.83 were obtained when 1.5 T and 3 T patients were considered, respectively. Conclusions The model elaborated showed good performance, even when data from patients scanned on 1.5 T and 3 T were merged. This shows that magnetic field intensity variability can be overcome by means of selecting appropriate image features. Keywords Radiomics · Magnetic resonance imaging · Inter-scanner variability · Magnetic field intensity · Rectal cancer Abbreviations AIC Aikake information criteria AUC Area under curve GLNU Grey-level non-uniformity IB Intensity based LARC Locally advanced rectal cancer LE Local excision LOG Laplacian of Gaussian MRI Magnetic resonance imaging nCRT Neoadjuvant chemo-radiotherapy pCR Pathological complete response ROI Region of interest T2-w T2-weighted * Jacopo Lenkowicz [email protected]
TME Total mesorectal excision TRG Tumour regression grade W&W Watch and wait WMW Wilcoxon–Mann–Whitney
Introduction Rectal cancer accounts for one third of colorectal cancers and is to date one of the leading causes of cancer death in the western world [1, 2]. Neoadjuvant chemo-radiotherapy (nCRT) followed by total mesorectal excision (TME) represents the
Data Loading...
