Deep learning vs. atlas-based models for fast auto-segmentation of the masticatory muscles on head and neck CT images

Abstract

Background: Impaired function of masticatory muscles will lead to trismus. Routine delineation of these muscles during planning may improve dose tracking and facilitate dose reduction resulting in decreased radiation-related trismus. This study aimed to compare a deep learning model with a commercial atlas-based model for fast autosegmentation of the masticatory muscles on head and neck computed tomography (CT) images. Material and methods: Paired masseter (M), temporalis (T), medial and lateral pterygoid (MP, LP) muscles were manually segmented on 56 CT images. CT images were randomly divided into training (n = 27) and validation (n = 29) cohorts. Two methods were used for automatic delineation of masticatory muscles (MMs): Deep learning autosegmentation (DLAS) and atlas-based auto-segmentation (ABAS). The automatic algorithms were evaluated using Dice similarity coefficient (DSC), recall, precision, Hausdorff distance (HD), HD95, and mean surface distance (MSD). A consolidated score was calculated by normalizing the metrics against interobserver variability and averaging over all patients. Differences in dose (ΔDose) to MMs for DLAS and ABAS segmentations were assessed. A paired t-test was used to compare the geometric and dosimetric difference between DLAS and ABAS methods. Results: DLAS outperformed ABAS in delineating all MMs (p < 0.05). The DLAS mean DSC for M, T, MP, and LP ranged from 0.83 ± 0.03 to 0.89 ± 0.02, the ABAS mean DSC ranged from 0.79 ± 0.05 to 0.85 ± 0.04. The mean value for recall, HD, HD95, MSD also improved with DLAS for auto-segmentation. Interobserver variation revealed the highest variability in DSC and MSD for both T and MP, and the highest scores were achieved for T by both automatic algorithms. With few exceptions, the mean ΔD98%, ΔD95%, ΔD50%, and ΔD2% for all structures were below 10% for DLAS and ABAS and had no detectable statistical difference (P > 0.05). DLAS based contours had dose endpoints more closely matched with that of the manually segmented when compared with ABAS. Conclusions: DLAS auto-segmentation of masticatory muscles for the head and neck radiotherapy had improved segmentation accuracy compared with ABAS with no qualitative difference in dosimetric endpoints compared to manually segmented contours.

Barriers and facilitators to clinical implementation of radiotherapy treatment planning automation: A survey study of medical dosimetrists

“…It is noteworthy that a minority of respondents (6/34) reported use of deep-learning-based AC, which has been shown in the literature to be significantly more accurate than other methods. 11,38,39 Therefore, a broader availability of deep-learning-based tools could facilitate a broader use of AC in the clinic. A strong majority of respondents thought that it was difficult to modify the output of an automated planning algorithm and this limited the algorithms' usefulness.…”

Impact of Using Unedited CT-Based DIR-Propagated Autocontours on Online ART for Pancreatic SBRT

“…Several studies have shown improvements in different treatment sites (e.g. head-and-neck ( 38 – 40 ), prostate ( 39 , 41 ), rectum ( 42 ), whole body ( 43 )). However, abdominal organs present additional challenges including strong interpatient variability, bowel loop displacements and hollow organs, which causes AI studies still report similar results to those achieved in our current study ( 10 , 44 – 46 ).…”

Investigation of Autosegmentation Techniques on T2-Weighted MRI for Off-line Dose Reconstruction in MR-Linac Adapt to Position Workflow for Head and Neck Cancers

“…Various autosegmentation methods that have historically been applied to computed tomography (CT) images have shown promise on MRI [6], including deformable image registration (DIR)based structure propagation [4,7], atlas-based autosegmentation [8][9][10], and deep learning [11][12][13]. While several studies have shown that deep learning can improve organ-at-risk (OAR) segmentation accuracy compared to atlas-based autosegmentation on CT for head and neck [14][15][16] and other treatment sites [17][18][19][20], to our knowledge, only a single study thus far has directly compared these methods on MRI for any treatment site [21]. As MR-guided adaptive RT becomes more accessible, evaluating these autosegmentation methods on MRI is crucial.…”