MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Motion Prediction Confidence Estimation for MRI-Guided Radiotherapy

J Ginn*, D Low , J Lamb , D Ruan , University of California, Los Angeles, Los Angeles, CA

Presentations

(Sunday, 7/14/2019) 5:00 PM - 6:00 PM

Room: Stars at Night Ballroom 2-3

Purpose: Motion prediction can overcome magnetic resonance image (MRI) guided radiotherapy gating latency. However, motion predictions are subject to error and variations. We have developed and evaluated a novel motion prediction confidence estimation method to improve the efficacy and robustness of prediction-based gating.

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 prediction model fitting process and the velocity of the tracked target. Roughly, these terms quantify respectively the prediction model’s representativeness of the training data, the robustness of model inference, and the potential variation due to target speed. The hyperparameters and the confidence estimator threshold for overriding gating predictions are optimized. The method is validated in 8 healthy volunteer and 13 patient studies using a 0.35T MRI-guided radiotherapy system predicting 0.25-0.33 seconds ahead. The effect of the confidence estimator threshold 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 on average increased gating accuracy 0.69% to 96.5% (p = 3.62x10-4), increased PPV 0.89% to 96.6% (p = 1.68x10-5), reduced the median centroid distance 0.09 mm to 0.54 mm (p = 4.01x10-5) at the cost of reducing the gating duty cycle 13.2% to 49.5%. Hyperparameter tuning revealed model goodness of fit and leave-one-out prediction variation provide the most effective confidence estimator.

Conclusion: We have developed a novel confidence estimation method to complement prediction methods to guide MRI-guided radiotherapy gating. Results from both volunteer and patient studies showed gating quality is improved.

Funding Support, Disclosures, and Conflict of Interest: James Lamb has previously received speaking and consulting fees from ViewRay.

Keywords

MRI, Radiation Therapy, Modeling

Taxonomy

IM/TH- MRI in Radiation Therapy: MRI/Linear accelerator combined- IGRT and tracking

Contact Email