A motion prediction confidence estimation framework for prediction‐based radiotherapy gating

Abstract

Purpose: Motion prediction can compensate for latency in image-guided radiotherapy and has been an active area of research. However, motion predictions are subject to error and variations. We have developed and evaluated a novel motion prediction confidence estimation framework to improve the efficacy and robustness of prediction-based radiotherapy gating decision-making. The specific scenario of adaptive gating in magnetic resonance imaging (MRI)-guided radiotherapy is studied as an example, but the method generalizes to other modalities and motion management setups. Methods: The proposed prediction confidence estimator is based on a generic training/testing paradigm and consists of a weighted combination of three components: the prediction model's goodness of fit, variation in the prediction using a leave-one-out process and the velocity of the tracked target. Roughly, these terms quantify respectively the consistency between prediction and the training data, the robustness of model inference, and the stability due to target speed. The weight parameters and the action level in triggering beam-off decision are optimized. The method is assessed and validated in 8 healthy volunteer and 13 patient studies using a 0.35T MRI-guided radiotherapy system predicting 0.25-0.33 s ahead. The effect of the action level on the predicted gating decision accuracy, beam-on positive predictive value (PPV) and median distance between the predicted and ground-truth target centroids were evaluated. Statistical significance was evaluated using a paired t-test. The tradeoff between these performance metrics and gating duty cycle was assessed. Results: Use of the confidence estimator threshold increased gating accuracy by up to 2.42%, increased PPV by up to 3.00%, and reduced the median centroid distance up to 0.28 mm. The confidence estimator threshold on average increased gating accuracy to 96.5% (P = 2.08 9 10 À4 ), increased PPV to 96.7% (P = 1.46 9 10 À5 ), reduced the median centroid distance to 0.54 mm (P = 1.71 9 10 À5 ) at the cost of reducing the gating duty cycle by 14.3% to 48.5%. Hyperparameter tuning revealed that contrary to intuition, the velocity term offered only minimal performance improvement in some cases but also introduced potential stability issues. The combination of goodness of fit and leave-one-out prediction variation provided the most effective confidence estimator, yielding universally better performance in gating decisions. Conclusion: Confidence estimation utilizing prediction model fitness criterion and validation principles can complement prediction methods to guide MRI-guided radiotherapy gating. Results from both volunteer and patient studies showed improved gating quality.

Technical Challenges of Real-Time Adaptive MR-Guided Radiotherapy

“…Note that this process is currently 2D, due to a lack of commercially available real-time 3D imaging sequences. Recent studies however showed promising approaches towards realtime 3D MR image acquisition (53)(54)(55)(56) and reconstruction (57).…”

Adaptive confidence regions for indirect tracking of moving tumors in radiotherapy

“…A closer examination of the ability of the CR resulting from the proposed adaptive covariance estimates to detect large errors is performed. A suitable way to quantify this performance is to analyze the receiver operating characteristic (ROC) curve 17,19,25 under various conditions. More specifically, the sensitivity and specificity are computed.…”

“…Few studies have investigated the prediction uncertainty, but only in the specific context of addressing temporal prediction, that is, predicting future positions knowing the past ones. Ginn et al 17 combined the prediction model's goodness of fit and motion prediction variation as a measure of prediction confidence. Based on a Kernel density estimator, 18 Ruan et al 19 showed that the predicted variances can be used as a feature to estimate the prediction errors, which allows identifying in advance breathing points with a higher probability of large prediction errors.…”

“…Few studies have investigated the prediction uncertainty, but only in the specific context of addressing temporal prediction, that is, predicting future positions knowing the past ones. Ginn et al 17 . combined the prediction model's goodness of fit and motion prediction variation as a measure of prediction confidence.…”

Uncertainty-driven determination of target measurement times for indirect tracking validation in adaptive radiotherapy

“…Performing validation and/or updating at preselected regular intervals may be not optimal. Few studies (Ruan 2010, Bukhari and Hong 2014, Ginn et al 2020 investigated rather the use of an estimation of the uncertainty associated with the predictions to automatically trigger motion management processes when low confidence is expected, but they focused on the latency compensation only, i.e. predicting future positions knowing the past ones.…”