Artificial Intelligence (AI)-based auto-delineation technologies rapidly delineate multiple structures of interest like organs-at-risk and tumors in 3D medical images, reducing personnel load and facilitating time-critical therapies. Despite its accuracy, the AI may produce flawed delineations, requiring clinician attention. Quality assessment (QA) of these delineations is laborious and demanding. Delineation error detection systems (DEDS) aim to aid QA, yet questions linger about potential challenges to their adoption and time-saving potential. To address these queries, we first conducted a user study with two clinicians from Holland Proton Therapy Center, a Dutch cancer treatment center. Based on the study’s findings about the clinicians’ error detection workflows with and without DEDS assistance, we developed a simulation model of the QA process, which we used to assess different error detection workflows on a retrospective cohort of 42 head and neck cancer patients. Results suggest possible time savings, provided the per-slice analysis time stays close to the current baseline and trading-off delineation quality is acceptable. Our findings encourage the development of user-centric delineation error detection systems and provide a new way to model and evaluate these systems’ potential clinical value.@en

Objective.The integration of proton beamlines with x-ray imaging/irradiation platforms has opened up possibilities for image-guided Bragg peak irradiations in small animals. Such irradiations allow selective targeting of normal tissue substructures and tumours. However, their small size and location pose challenges in designing experiments. This work presents a simulation framework useful for optimizing beamlines, imaging protocols, and design of animal experiments. The usage of the framework is demonstrated, mainly focusing on the imaging part.Approach.The fastCAT toolkit was modified with Monte Carlo (MC)-calculated primary and scatter data of a small animal imager for the simulation of micro-CT scans. The simulated CT of a mini-calibration phantom from fastCAT was validated against a full MC TOPAS CT simulation. A realistic beam model of a preclinical proton facility was obtained from beam transport simulations to create irradiation plans in matRad. Simulated CT images of a digital mouse phantom were generated using single-energy CT (SECT) and dual-energy CT (DECT) protocols and their accuracy in proton stopping power ratio (SPR) estimation and their impact on calculated proton dose distributions in a mouse were evaluated.Main results.The CT numbers from fastCAT agree within 11 HU with TOPAS except for materials at the centre of the phantom. Discrepancies for central inserts are caused by beam hardening issues. The root mean square deviation in the SPR for the best SECT (90 kV/Cu) and DECT (50 kV/Al-90 kV/Al) protocols are 3.7% and 1.0%, respectively. Dose distributions calculated for SECT and DECT datasets revealed range shifts <0.1 mm, gamma pass rates (3%/0.1 mm) greater than 99%, and no substantial dosimetric differences for all structures. The outcomes suggest that SECT is sufficient for proton treatment planning in animals.Significance.The framework is a useful tool for the development of an optimized experimental configuration without using animals and beam time.

@en

Deep learning models for organ contouring in radiotherapy are poised for clinical usage, but currently, there exist few tools for automated quality assessment (QA) of the predicted contours. Bayesian models and their associated uncertainty, can potentially automate the process of detecting inaccurate predictions. We investigate two Bayesian models for auto-contouring, DropOut and FlipOut, using a quantitative measure – expected calibration error (ECE) and a qualitative measure – region-based accuracy-vs-uncertainty (R-AvU) graphs. It is well understood that a model should have low ECE to be considered trustworthy. However, in a QA context, a model should also have high uncertainty in inaccurate regions and low uncertainty in accurate regions. Such behaviour could direct visual attention of expert users to potentially inaccurate regions, leading to a speed-up in the QA process. Using R-AvU graphs, we qualitatively compare the behaviour of different models in accurate and inaccurate regions. Experiments are conducted on the MICCAI2015 Head and Neck Segmentation Challenge and on the DeepMindTCIA CT dataset using three models: DropOut-DICE, Dropout-CE (Cross Entropy) and FlipOut-CE. Quantitative results show that DropOut-DICE has the highest ECE, while Dropout-CE and FlipOut-CE have the lowest ECE. To better understand the difference between DropOut-CE and FlipOut-CE, we use the R-AvU graph which shows that FlipOut-CE has better uncertainty coverage in inaccurate regions than DropOut-CE. Such a combination of quantitative and qualitative metrics explores a new approach that helps to select which model can be deployed as a QA tool in clinical settings.
@en