Machine learning is revolutionising medical image analysis, and clearly the future of the field lies in this direction. However, with increasing automation there is a danger of misunderstanding or misinterpreting models. In this paper, we expose an underlying bias in a commonly used publicly available brain tumour MRI dataset. We propose that this is due to implicit radiologist input in the selection of the 2D slices. Through several experiments we show how this bias allows us to achieve a high tumour classification accuracy, even with no information regarding the tumour itself. No other papers that use the dataset mention this bias. These findings demonstrate the importance of understanding machine learning models and their medical context, and the perils of not doing so.

In this paper, we extend the gradient vector flow field to the vector-valued case for robust variational segmentation of 4D images with active surfaces. Instead of only exploiting scalar edge strength in order to identify vector edges, we propagate both directions and amplitudes of vector gradients computed from the analysis of a structure tensor of the vector-valued image. To reduce contributions from noise in the calculation of the structure tensor, image channels are weighted according to a blind estimator of contrast that take profit of the deformable models framework. The proposed 4DGVF vector field is validated on synthetic image datasets and applied to biological volume delineation in dynamic PET imaging.

In this paper, we propose a new method for vector-valued image restoration in order to reduce noise while simultaneously sharpening vector edges. Our approach is a coupled anisotropic diffusion and shock filtering scheme that exploits a new robust 4DGVF vector field tailored for vector-valued images. The proposed scheme sharpens edges in directions diffused from the entire spatio-spectral information available with a more precise and a more stable sharpening effect along the iterative processing. We validate our method on color images as well as on realistic simulations of dynamic PET images.

In this paper, we extend the gradient vector flow field for robust variational segmentation of vector-valued images. Rather than using scalar edge information, we define a vectorial edge map derived from a weighted local structure tensor of the image that enables the diffusion of the gradient vectors in accurate directions through the 4D gradient vector flow equation. To reduce the contribution of noise in the structure tensor, image channels are weighted according to a blind estimator of contrast. The method is applied to biological volume delineation in dynamic PET imaging, and validated on realistic Monte Carlo simulations of numerical phantoms as well as on real images.

A statistical methodology is proposed to rank several estimation methods of a relevant clinical parameter when no gold standard is available. Based on a regression without truth method, the proposed approach was applied to rank eight methods without using any a priori information regarding the reliability of each method and its degree of automation. It was only based on a prior concerning the statistical distribution of the parameter of interest in the database. The ranking of the methods relies on figures of merit derived from the regression and computed using a bootstrap process. The methodology was applied to the estimation of the left ventricular ejection fraction derived from cardiac magnetic resonance images segmented using eight approaches with different degrees of automation: three segmentations were entirely manually performed and the others were variously automated. The ranking of methods was consistent with the expected performance of the estimation methods: the most accurate estimates of the ejection fraction were obtained using manual segmentations. The robustness of the ranking was demonstrated when at least three methods were compared. These results suggest that the proposed statistical approach might be helpful to assess the performance of estimation methods on clinical data for which no gold standard is available.

Two groups of bootstrap methods have been proposed to estimate the statistical properties of positron emission tomography (PET) images by generating multiple statistically equivalent data sets from few data samples. The first group generates resampled data based on a parametric approach assuming that data from which resampling is performed follows a Poisson distribution while the second group consists of nonparametric approaches. These methods either require a unique original sample or a series of statistically equivalent data that can be list-mode files or sinograms. Previous reports regarding these bootstrap approaches suggest different results. This work compares the accuracy of three of these bootstrap methods for 3-D PET imaging based on simulated data. Two methods are based on a unique file, namely a list-mode based nonparametric (LMNP) method and a sinogram based parametric (SP) method. The third method is a sinogram-based nonparametric (SNP) method. Another original method (extended LMNP) was also investigated, which is an extension of the LMNP methods based on deriving a resampled list-mode file by drawings events from multiple original list-mode files. Our comparison is based on the analysis of the statistical moments estimated on the repeated and resampled data. This includes the probability density function and the moments of order 1 and 2. Results show that the two methods based on multiple original data (SNP and extended LMNP) are the only methods that correctly estimate the statistical parameters. Performances of the LMNP and SP methods are variable. Simulated data used in this study were characterized by a high noise level. Differences among the tested strategies might be reduced with clinical data sets with lower noise.

In this study, we evaluated SPECT/MRI registration of ictal data, using similarity based registration methods. An absolute gold standard for registration evaluation was obtained by considering realistic normal and ictal SPECT simulations deduced from a high resolution T1-weighted MRI data set. Those simulations were also used to study the impact of photon attenuation and Compton scatter corrections on registration accuracy. Evaluation of registration was also performed using inconsistency measurements for six patients with temporo-mesial epilepsy. For these data, as no Gold Standard was available, registration accuracy was assessed using inconsistency measurements involving a registration loop between inter-ictal SPECT, ictal SPECT and MRI data. Five registration methods based on statistical similarity measurements were compared, namely: mutual information (MI), normalized mutual information (NMI), L1 and L2 norm-based correlation ratios (CR) and correlation coefficient (CC). It was found that the simulation context had more influence on registration accuracy than the choice of the similarity criterion. Ictal SPECT as well as correction for uniform attenuation clearly decreased SPECT/MRI registration accuracy. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

We present a method to validate MRI/SPECT registration methods based on a set of computer-generated SPECT data. The data set was produced through Monte Carlo simulations from an attenuation map and an activity map derived from a manually labeled T1-weighted MRI data set.Our approach intrinsically provides a gold standard to assess MRI/SPECT registration methods. It was successfully applied to the comparison of four registration methods based on similarity measurements: Mutual Information, Normalised Mutual Information, Correlation Ratio and Woods Criterion.