Symmetry of the submandibular glands in humans—a postmortem study assessing the linear morphometric parameters

Retrograde Migration of an Au-198 Grain to the Submandibular Gland Post Brachytherapy Treatment of Floor of Mouth Cancer

At our institution, radiation oncologists routinely treat early-stage oral cancer with low-dose-rate brachytherapy (LDR-BRT) using Au-198 grains. In this report, we show a unique case of a patient with a gold grain located within the submandibular gland, found incidentally during follow-up after LDR-BRT for floor of mouth cancer. One month after the implant, he showed sialadenitis-like symptoms, but the pain resolved two months later. All the grains were detected around the anterior sublingual area by computed tomography (CT) four months after the implant. Unexpectedly, 11 months after the implant, CT revealed that a grain was located in an intraglandular site of the submandibular gland. This finding clearly demonstrates that the grain entered Wharton's duct and retrogradely migrated to the submandibular gland through the duct. As a mechanism of the calculus formation within Wharton's duct, retrograde migration of foreign bodies to the inside of the duct has been proposed. Our incidental finding after LDR-BRT highlights the need for monitoring post-LDR-BRT using Au-198 grains for the treatment of floor of mouth cancer and sheds additional light on retrograde theory within Wharton's duct.

Investigating the potential of deep learning for patient-specific quality assurance of salivary gland contours using EORTC-1219-DAHANCA-29 clinical trial data

INTRODUCTION

Manual quality assurance (QA) of radiotherapy contours for clinical trials is time and labor intensive and subject to inter-observer variability. Therefore, we investigated whether deep-learning (DL) can provide an automated solution to salivary gland contour QA.

MATERIAL AND METHODS

DL-models were trained to generate contours for parotid (PG) and submandibular glands (SMG). Sørensen-Dice coefficient (SDC) and Hausdorff distance (HD) were used to assess agreement between DL and clinical contours and thresholds were defined to highlight cases as potentially sub-optimal. 3 types of deliberate errors (expansion, contraction and displacement) were gradually applied to a test set, to confirm that SDC and HD were suitable QA metrics. DL-based QA was performed on 62 patients from the EORTC-1219-DAHANCA-29 trial. All highlighted contours were visually inspected.

RESULTS

Increasing the magnitude of all 3 types of errors resulted in progressively severe deterioration/increase in average SDC/HD. 19/124 clinical PG contours were highlighted as potentially sub-optimal, of which 5 (26%) were actually deemed clinically sub-optimal. 2/19 non-highlighted contours were false negatives (11%). 15/69 clinical SMG contours were highlighted, with 7 (47%) deemed clinically sub-optimal and 2/15 non-highlighted contours were false negatives (13%). For most incorrectly highlighted contours causes for low agreement could be identified.

CONCLUSION

Automated DL-based contour QA is feasible but some visual inspection remains essential. The substantial number of false positives were caused by sub-optimal performance of the DL-model. Improvements to the model will increase the extent of automation and reliability, facilitating the adoption of DL-based contour QA in clinical trials and routine practice.

Using Spatial Probability Maps to Highlight Potential Inaccuracies in Deep Learning-Based Contours: Facilitating Online Adaptive Radiation Therapy

PURPOSE

Contouring organs at risk remains a largely manual task, which is time consuming and prone to variation. Deep learning-based delineation (DLD) shows promise both in terms of quality and speed, but it does not yet perform perfectly. Because of that, manual checking of DLD is still recommended. There are currently no commercial tools to focus attention on the areas of greatest uncertainty within a DLD contour. Therefore, we explore the use of spatial probability maps (SPMs) to help efficiency and reproducibility of DLD checking and correction, using the salivary glands as the paradigm.

METHODS AND MATERIALS

A 3-dimensional fully convolutional network was trained with 315/264 parotid/submandibular glands. Subsequently, SPMs were created using Monte Carlo dropout (MCD). The method was boosted by placing a Gaussian distribution (GD) over the model's parameters during sampling (MCD + GD). MCD and MCD + GD were quantitatively compared and the SPMs were visually inspected.

RESULTS

The addition of the GD appears to increase the method's ability to detect uncertainty. In general, this technique demonstrated uncertainty in areas that (1) have lower contrast, (2) are less consistently contoured by clinicians, and (3) deviate from the anatomic norm.

CONCLUSIONS

We believe the integration of uncertainty information into contours made using DLD is an important step in highlighting where a contour may be less reliable. We have shown how SPMs are one way to achieve this and how they may be integrated into the online adaptive radiation therapy workflow.

Strategies to improve deep learning-based salivary gland segmentation

BACKGROUND

Deep learning-based delineation of organs-at-risk for radiotherapy purposes has been investigated to reduce the time-intensiveness and inter-/intra-observer variability associated with manual delineation. We systematically evaluated ways to improve the performance and reliability of deep learning for organ-at-risk segmentation, with the salivary glands as the paradigm. Improving deep learning performance is clinically relevant with applications ranging from the initial contouring process, to on-line adaptive radiotherapy.

METHODS

Various experiments were designed: increasing the amount of training data (1) with original images, (2) with traditional data augmentation and (3) with domain-specific data augmentation; (4) the influence of data quality was tested by comparing training/testing on clinical versus curated contours, (5) the effect of using several custom cost functions was explored, and (6) patient-specific Hounsfield unit windowing was applied during inference; lastly, (7) the effect of model ensembles was analyzed. Model performance was measured with geometric parameters and model reliability with those parameters' variance.

RESULTS

A positive effect was observed from increasing the (1) training set size, (2/3) data augmentation, (6) patient-specific Hounsfield unit windowing and (7) model ensembles. The effects of the strategies on performance diminished when the base model performance was already 'high'. The effect of combining all beneficial strategies was an increase in average Sørensen-Dice coefficient of about 4% and 3% and a decrease in standard deviation of about 1% and 1% for the submandibular and parotid gland, respectively.

CONCLUSIONS

A subset of the strategies that were investigated provided a positive effect on model performance and reliability. The clinical impact of such strategies would be an expected reduction in post-segmentation editing, which facilitates the adoption of deep learning for autonomous automated salivary gland segmentation.

Sialolithiasis-Do Early Diagnosis and Removal Minimize Post-Operative Morbidity?

Background and objectives: Sialolithiasis is an inflammation of a salivary gland due to obstruction of salivary flow by a sialolith. We aim to assess potential factors that may predict lower morbidity following endoscopically assisted per-oral sialolith removal. Materials and Methods: Retrospective cohort study. Retrospective review of 100 records of patients with sialolithiasis, following surgical sialolith removal. A single medical center (Department of oral and maxillofacial surgery-Rabin Medical Center, Beilinson & Hasharon-Israel) survey. Data were gleaned from the patient files based on a structured questionnaire. Factors that may predict morbidity were evaluated using linear regression equation. Results: 59 of the subjects were men and 41 were women. The mean age of the patients in the study was 50 ± 17.5 years. Sialolith volume and past antibiotic treatment were positively associated while age was negatively associated with hospitalization duration. Conclusion: Early sialolith diagnosis and removal may lower postoperative morbidity.

Anatomical localization of submandibular gland for botulinum toxin injection

The aim of this study was to document the anatomical landmarks of the submandibular gland (SMG) for a botulinum toxin injection. Thirty-four SMGs from 20 cadavers were examined. The mean length of a reference line between the angle of the mandible and the gnathion was 94.8 ± 5.9 mm, the proximal and distal point of the SMG from the angle of the mandible was 10.6% (11.5 ± 3.5 mm) and 41.8% (40.9 ± 5.2 mm), respectively. The facial artery came out of the SMG at 11.6% (14.6 ± 3.4 mm) and the position of the intersection of the facial artery with the inferior border of the mandible was located at 24.4% (28.0 ± 5.5 mm) from the angle of the mandible. The shape of the SMG was generally triangular or irregular round on the anatomical position. The mean superior-inferior diameter, anterior-posterior diameter and medial-lateral diameter of the gland was 28.8 ± 4.1, 30.0 ± 6.1 and 15.1 ± 3.5 mm, respectively. The safety zone for the injection was 20-35% from the mandible angle on the inferior view and 1.5 cm below the inferior line of the mandible on the lateral view. In addition, the needle should be inserted to a depth of 2.0 cm from the skin surface on the inferior view. These results may assist in determining a accurate localization of injection sites for the SMG, particularly for injections without ultrasound guidance.