Evaluation of multiple-atlas-based strategies for segmentation of the thyroid gland in head and neck CT images for IMRT

Abstract

Segmenting the thyroid gland in head and neck CT images is of vital clinical significance in designing intensity-modulated radiation therapy (IMRT) treatment plans. In this work, we evaluate and compare several multiple-atlas-based methods to segment this structure. Using the most robust method, we generate automatic segmentations for the thyroid gland and study their clinical applicability. The various methods we evaluate range from selecting a single atlas based on one of three similarity measures, to combining the segmentation results obtained with several atlases and weighting their contribution using techniques including a simple majority vote rule, a technique called STAPLE that is widely used in the medical imaging literature, and the similarity between the atlas and the volume to be segmented. We show that the best results are obtained when several atlases are combined and their contributions are weighted with a measure of similarity between each atlas and the volume to be segmented. We also show that with our data set, STAPLE does not always lead to the best results. Automatic segmentations generated by the combination method using the correlation coefficient (CC) between the deformed atlas and the patient volume, which is the most accurate and robust method we evaluated, are presented to a physician as 2D contours and modified to meet clinical requirements. It is shown that about 40% of the contours of the left thyroid and about 42% of the right thyroid can be used directly. An additional 21% on the left and 24% on the right require only minimal modification. The amount and the location of the modifications are qualitatively and quantitatively assessed. We demonstrate that, although challenged by large inter-subject anatomical discrepancy, atlas-based segmentation of the thyroid gland in IMRT CT images is feasible by involving multiple atlases. The results show that a weighted combination of segmentations by atlases using the CC as the similarity measure slightly outperforms standard combination methods, e.g. the majority vote rule and STAPLE, as well as methods selecting a single most similar atlas. The results we have obtained suggest that using our contours as initial contours to be edited has clinical value.

Figure 1

Flow charts illustrating the registration process. Top panel: Registration of patient image with the average image volume at full scale. Bottom panel: Registration in the bounding box containing the thyroid between the patient image and the atlases.

Figure 2

Boxplots showing the sample minimum, Q1, Q2, Q3, and the sample maximum of volume DSC’s obtained using the most similar atlas selected by maximum CC (CC_max), minimum MAX_df (MAX_df_min) , and minimum AVG_df (AVG_df_min), the combination of all segmentation weighted by CC (CC_weighted), MAX_df (MAX_df_weighted), AVG_df (AVG_df_weighted), the average of all segmentations (avg_all), and the combination of all segmentations by STAPLE. Left panel: Left thyroids. Right panel: Right thyroids.

Figure 3

Boxplots showing the sample minimum, Q1, Q2, Q3, and the sample maximum of the averages of slice DSC’s obtained using CC_max, CC_weighted, avg_all, and STAPLE. Left panel: Left thyroids. Right panel: Right thyroids.

Figure 4

Boxplots showing the sample minimum, Q1, Q2, Q3, and the sample maximum of the averages of Hausdorff distance in mm on 2D slices obtained using CC_max, CC_weighted, avg_all, and STAPLE. Left panel: Left thyroids. Right panel: Right thyroids.

Figure 5

Segmentations obtained using the four representative methods shown in contours with dotted lines compared with the manual segmentation shown in solid lines. For each row, from the left to the right: The original patient image, images with contours obtained using CC_max, CC_weighted, avg_all, and STAPLE compared with the manual contours. Rows from top to bottom: Left thyroids for patients 3, 8, and 16, right thyroids for patients 3, 5, and 11.

Figure 6

3D surfaces of the modified segmentations, with blue color representing zero or little distance to the surface of the original automatic segmentation obtained using CC_weighted, and red color representing large distance. Columns from left to right: Left thyroids for patients 3, 8, and 16, and right thyroids for patients 3, 5, and 11. For each column, the top and bottom rows show the same surface viewed from two different angles.