Texture analysis and machine learning to predict water T2 and fat fraction from non-quantitative MRI of thigh muscles in Facioscapulohumeral muscular dystrophy

Felisaz, Paolo Florent; Colelli, Giulia; Ballante, Elena; Solazzo, Francesca; Paoletti, Matteo; Germani, Giancarlo; Santini, Francesco; Deligianni, Xeni; Bergsland, Niels; Monforte, Mauro; Tasca, Giorgio; Ricci, Enzo; Bastianello, Stefano; Figini, Silvia; Pichiecchio, Anna

doi:10.1016/j.ejrad.2020.109460

Cited by 16 publications

(18 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…FSHD pathology shares features in common with most muscular dystrophies including muscle fibres that are atrophic and regenerating, that vary in size, exhibit central nucleation and/or are rounded, as well as endomysial fibrosis, fat infiltration and inflammation (Padberg, 1982 ; Wang & Tawil, 2016 ) (Fig 2 ). Endomysial and perivascular inflammation are prominent features of FSHD: detected microscopically via histopathology (Wang & Tawil, 2016 ) (Fig 2 ) and macroscopically via magnetic resonance imaging (MRI) using measures such as short tau inversion recovery (STIR) sequences (Fig 3 ) and decompositions of T2‐weighted signal intensity (Dahlqvist et al , 2020b ; Felisaz et al , 2021 ). MRI measures of macroscopic inflammation are indicators of muscles at greater risk of fatty replacement as detected using T1‐weighted imaging (Fig 3 ), a feature correlated to clinical severity.…”

Section: Pathology In Fshd Musclementioning

confidence: 99%

Pathomechanisms and biomarkers in facioscapulohumeral muscular dystrophy: roles of DUX4 and PAX7

Banerji

Zammit

2021

EMBO Mol Med

View full text Add to dashboard Cite

Facioscapulohumeral muscular dystrophy (FSHD) is characterised by progressive skeletal muscle weakness and wasting. FSHD is linked to epigenetic derepression of the subtelomeric D4Z4 macrosatellite at chromosome 4q35. Epigenetic derepression permits the distal‐most D4Z4 unit to transcribe DUX4, with transcripts stabilised by splicing to a poly(A) signal on permissive 4qA haplotypes. The pioneer transcription factor DUX4 activates target genes that are proposed to drive FSHD pathology. While this toxic gain‐of‐function model is a satisfying “bottom‐up” genotype‐to‐phenotype link, DUX4 is rarely detectable in muscle and DUX4 target gene expression is inconsistent in patients. A reliable biomarker for FSHD is suppression of a target gene score of PAX7, a master regulator of myogenesis. However, it is unclear how this “top‐down” finding links to genomic changes that characterise FSHD and to DUX4. Here, we explore the roles and interactions of DUX4 and PAX7 in FSHD pathology and how the relationship between these two transcription factors deepens understanding via the immune system and muscle regeneration. Considering how FSHD pathomechanisms are represented by “DUX4opathy” models has implications for developing therapies and current clinical trials.

show abstract

Section: Pathology In Fshd Musclementioning

confidence: 99%

Pathomechanisms and biomarkers in facioscapulohumeral muscular dystrophy: roles of DUX4 and PAX7

Banerji

Zammit

2021

EMBO Mol Med

View full text Add to dashboard Cite

show abstract

“…2021) used MSE to estimate the capacity of image analysis of conventional magnetic resonance to predict the results obtained from quantitative MRI. [27][28][29] Values of MSE were calculated for SFC at 80%, 60%, 40%, and 20% for all investigated wavelengths (Figure 5), and the results for all wavelengths were highly accurate. However, the NIR LED (880 nm) produced the best results for SFC 80% and 20%, while the yellow LED (590 nm) had slightly better accuracy than the NIR LED for SFC 60% and 40%, demonstrating a greater capacity to predict SFC in a wider range of temperatures than those with NMR.…”

Section: Resultsmentioning

confidence: 91%

Optical Determination of Solid Fat Content in Fats and Oils: Effects of Wavelength on Estimated Accuracy

Grossi

Valli

Glicerina

et al. 2021

Euro J Lipid Sci & Tech

View full text Add to dashboard Cite

The melting profile of solid fat content (SFC) is a parameter of primary importance for the food industry since it affects many important product characteristics such as stability, physical appearance, spreadability, and sensation in the mouth. Reference techniques to measure SFC in fats and oils include pulsed nuclear magnetic resonance (pNMR) and differential scanning calorimetry (DSC), which are reliable and accurate, but require expensive instrumentation and trained personnel. Herein, the accuracy of a recently proposed optical technique to measure SFC is investigated in terms of peak wavelength of incident radiation. A sensor system featuring an array of seven LEDs with peak wavelength in the visible and NIR range is built, and the results compared with data from DSC. All the wavelengths investigated have high accuracy in SFC estimation, especially at 590 (yellow) and 880 nm (NIR). Practical applications: Quick and easy determination of solid fat content in fats and oils by a simple experimental setup. The technique is based on optical attenuation measurements during a thermal cycle. The technique can be implemented in a measurement instrument for in-situ analysis of solid fat content.

show abstract

“…These maps were derived from CSE-MRI, which delivers sequences with fast acquisition and good reproducibility that can easily be added to routine imaging protocols of the thigh region [19]. Felisaz et al applied TA in thigh musculature on non-quantitative T2-weighted spin-echo images for the purpose of machine-learning-aided prediction of water T2 and fat fraction [28]. However, TA has not been performed before on CSE-MRIderived PDFF maps.…”

Section: Discussionmentioning

confidence: 99%

“…For example, TA based on multiple MRI sequences was shown to differentiate pathological and clinical subtypes of cervical carcinoma [24], and it improved the discrimination of normal and cancerous tissue of the prostate [25]. Regarding musculoskeletal imaging, TA has predominantly been performed on non-quantitative data, such as sonography [26], computed tomography [27], or conventional T2-weighted MRI sequences [28,29]. Furthermore, TA based on CSE-MRI-derived PDFF maps has been used in the quantitative analysis of vertebral bone marrow [30,31].…”

Section: Introductionmentioning

confidence: 99%

Association of Thigh Muscle Strength with Texture Features Based on Proton Density Fat Fraction Maps Derived from Chemical Shift Encoding-Based Water–Fat MRI

et al. 2021

View full text Add to dashboard Cite

: Purpose: Based on conventional and quantitative magnetic resonance imaging (MRI), texture analysis (TA) has shown encouraging results as a biomarker for tissue structure. Chemical shift encoding-based water–fat MRI (CSE-MRI)-derived proton density fat fraction (PDFF) of thigh muscles has been associated with musculoskeletal, metabolic, and neuromuscular disorders and was demonstrated to predict muscle strength. The purpose of this study was to investigate PDFF-based TA of thigh muscles as a predictor of thigh muscle strength in comparison to mean PDFF. Methods: 30 healthy subjects (age = 30 ± 6 years; 15 females) underwent CSE-MRI of the lumbar spine at 3T, using a six-echo 3D spoiled gradient echo sequence. Quadriceps (EXT) and ischiocrural (FLEX) muscles were segmented to extract mean PDFF and texture features. Muscle flexion and extension strength were measured with an isokinetic dynamometer. Results: Of the eleven extracted texture features, Variance(global) showed the highest significant correlation with extension strength (p < 0.001, R2adj = 0.712), and Correlation showed the highest significant correlation with flexion strength (p = 0.016, R2adj = 0.658). Multivariate linear regression models identified Variance(global) and sex, but not PDFF, as significant predictors of extension strength (R2adj = 0.709; p < 0.001), while mean PDFF, sex, and BMI, but none of the texture features, were identified as significant predictors of flexion strength (R2adj = 0.674; p < 0.001). Conclusions: Prediction of quadriceps muscle strength can be improved beyond mean PDFF by means of TA, indicating the capability to quantify muscular fat infiltration patterns.

show abstract

Texture analysis and machine learning to predict water T2 and fat fraction from non-quantitative MRI of thigh muscles in Facioscapulohumeral muscular dystrophy

Cited by 16 publications

References 23 publications

Pathomechanisms and biomarkers in facioscapulohumeral muscular dystrophy: roles of DUX4 and PAX7

Pathomechanisms and biomarkers in facioscapulohumeral muscular dystrophy: roles of DUX4 and PAX7

Optical Determination of Solid Fat Content in Fats and Oils: Effects of Wavelength on Estimated Accuracy

Association of Thigh Muscle Strength with Texture Features Based on Proton Density Fat Fraction Maps Derived from Chemical Shift Encoding-Based Water–Fat MRI

Contact Info

Product

Resources

About