Automated T1 and T2 mapping segmentation on cardiovascular magnetic resonance imaging using deep learning

Figures

• 

  Figure 1. The proportion of healthy volunteers, athletes and pathologies in our dataset was reported using cardiovascular magnetic resonance (CMR) imaging. “No myocardial alterations” were reported, when the patient was referred to CMR due to a specific diagnostic question, but no pathology was found.

• 

  Figure 2. Deep neural network architecture: U-Net with ResNet-based encoder and convolutional decoder.

• 

  Figure 3. Learning rate schedule used throughout the training as proposed by Loshchilov and Hutter (28).

• 

  Figure 4. (A) Illustration of our algorithm that fits epi- and endocardial contours to segmentation masks. (B) Segmentation mask with the same resolution as the input image. (C) Contour points fitted on the low-resolution segmentation mask using a dominant point detection algorithm by C-H Teh and Roland T Chin (29). (D) Illustration of a case where correcting the contour using its convex hull improves the validity and accuracy of the contour. (E) Result the interpolation step which is performed by first fitting a smooth spline on the fitted contour points (30), then sampling it between the initial points. This step improves the numeric accuracy of contour-based metric computation (e.g. Hausdorff distance) and also enforces a smooth property on the fitted contours, similarly as the ones of expert annotators. On subfigure (A), steps indicated with dashed outline are only performed if necessary: - Filtering of contours is only necessary if the segmentation network incorrectly predicts multiple disjoint epi- or endocardial regions. We select the contour with the largest area and an approximately round shape (have a less than 1:3 with-height ratio) - Correcting a contour using its convex hull…

• 

  Figure 5. Violin plots showing the results of pairwise comparison of our automated method (pink halves) to expert annotators (green halves). Horizontal dashed lines show median and Q1, Q3 quartiles. Violin plots are cut at the bounds of underlying data. Note the logarithmic y-axis on the Hausdorff Distance. We plot epi- and endocardial results separately to allow for accurate reporting of Dice score results (plotting the mean of epi- and endocardial dice score would hide the difference in these two metrics).

• 

  Figure 6. Predicted and ground truth contours with low, intermediate and high Hausdorff distance results, selected separately for apical, midventricular and basal T1 maping images.

• 

  Figure 7. Example contours from two experts with different levels of agreement (measured by Hausdorff distance), plotted for two observers as a reference. All subfigures show apical T1 maps.

• 

  Figure 8. Per-slice violin plots comparing the agreement of our model with experts to agreement among experts. Table 1 shows the number of mapping images we used for training and evaluation (testing). Violin plots were generated on the test set.

• 

  Figure 9. Distribution of achievable best Hausdorff distance values. We measure achievable lowest contour accuracy for each mapping by converting ground truth contours to mask, fitting contours to these, then computing Hausdorff distance between these and the initial ground truth contour.

• 

  Figure A1. Ablation results showing the effect of each step in the proposed contour fitting pipeline. The figure shows the distribution of Hausdorff distance results over all samples for the same model, with the indicated post-processing steps added to the pipeline.

• 

  Figure A2. Predicted and ground truth contours with low, intermediate and high Hausdorff distance results, selected separately for apical, midventricular and basal T2 maping images.

• 

  Figure A3. Example contours from two experts with different levels of agreement (measured by Hausdorff distance), plotted for two observers as comparison baseline to Figure A2. All subfigures show apical T2 maps.