arXiv Science⌕ Search
Home/Archive/arXiv

arXiv · 2105.07024

A hyperparameter-tuning approach to automated inverse planning

Kelsey Maass·Aleksandr Aravkin·Minsun Kim
View original source ↗Read source document ↗Export metadata ↓

Abstract

Radiotherapy inverse planning often requires planners to modify parameters in the treatment planning system's objective function to produce clinically acceptable plans. Due to the manual steps in this process, plan quality can vary depending on the planning time available and the planner's skills. This study investigates two hyperparameter-tuning methods for automated inverse planning. Because this framework does not train a model on previously-optimized plans, it can be readily adapted to practice pattern changes, and plan quality is not limited by that of a training cohort. We selected 10 patients who received lung SBRT using manually-generated clinical plans. We used random sampling (RS) and Bayesian optimization (BO) to tune parameters using linear-quadratic utility functions based on 11 clinical goals. Normalizing all plans to have PTV D95 equal to 48 Gy, we compared plan quality for the automatically-generated and manually-generated plans. We also investigated the impact of iteration count on the automatically-generated plans, comparing planning time and plan utility for RS and BO plans with and without stopping criteria. Without stopping criteria, the median planning time was 1.9 and 2.3 hours for RS and BO plans. The OAR doses in the RS and BO plans had a median percent difference (MPD) of 48.7% and 60.4% below clinical dose limits and an MPD of 2.8% and 3.3% below clinical plan doses. With stopping criteria, the utility decreased by an MPD of 5.3% and 3.9% for RS and BO plans, but the median planning time was reduced to 0.5 and 0.7 hours, and the OAR doses still had an MPD of 42.9% and 49.7% below clinical dose limits and an MPD of 0.3% and 1.8% below clinical plan doses. This study demonstrates that hyperparameter-tuning approaches to automated inverse planning can reduce active planning time with plan quality that is similar to or better than manually-generated plans.

Explore related subjects

physics.med-phcs.LG

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Kelsey Maass, Aleksandr Aravkin, Minsun Kim. 2021-12-10. A hyperparameter-tuning approach to automated inverse planning. https://doi.org/10.1002/mp.15557

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Fail-closed conformal multiresolution error control for conservative three-dimensional dose remapping in synthetic phantoms

Conservative dose remapping can be formulated by transporting mass and deposited energy across nonmatching grids, but fixed high-sampling calculations use the same work for easy and difficult cases. We developed conformal multiresolution error control (CoMERC), a fail-closed multilevel quasi-Monte Carlo (QMC) controller for reference-relative target-cell allocation error in a fixed piecewise-constant source model with known deformation. Two randomized nested replicates were evaluated at a probe level of 4 and candidate levels of 8 and 16. Four endpoints were controlled jointly: reference-mass-weighted global, high-gradient, and density-gradient-threshold root-mean-square error, plus eligible-cell maximum absolute error. A split-conformal multiplier was calibrated from 240 synthetic cases, frozen, and evaluated on 400 independent cases from the same generator and 120 cases from a prespecified shifted generator. In the primary set, joint coverage was 394/400 (98.50%; one-sided 95% lower limit, 97.06%), 388/400 cases (97.00%; lower limit, 95.18%) received an output, and 0/388 released outputs exceeded any tolerance (one-sided 95% upper limit, 0.769%). Final actions were level 8 for 190 cases, level 16 for 198, and ABSTAIN for 12. The frozen policy implied mean relative production-sample work of 0.5844 compared with always running both randomized replicates through level 16; the prespecified 95th-percentile bootstrap upper limit was 0.6194. All seven primary gates passed. In the shifted-generator set, 113/120 cases were released and one released output exceeded a tolerance. CoMERC provided marginal finite-sample error control and lower modeled production-sample work in the locked synthetic population, but not patient-level, registration-level, conditional-on-release, or clinical safety validation.

physics.med-ph↗
arXiv2026

Dosimetric equivalence of deep learning prostate contours after LDR brachytherapy: pre-declared margins, patient-level acceptance thresholds and the incremental predictive value of DVH indices

Background and purpose: Dose Volume Histogram (DVH) indices remain the main dose-effect metrics for toxicity prediction, but they depend on how contours were made. Inter-observer variability (IOV) is unavoidable and clinically accepted, so the question for automatic segmentation addresses equivalence: do automatic contours produce DVH errors comparable to human IOV, and can these indices predict patient-reported toxicity? Materials and methods: In 429 patients treated with iodine-125 LDR-BT monotherapy, indices from expert manual delineation were compared with a deterministic and a Bayesian nnUNet on a fixed dose distribution, by two-one-sided tests against margins set from CT contouring IOV. Logistic regression gave patient-level thresholds at 90\,\% probability of equivalence. In 380 patients, eleven DHV indices were added to a clinical baseline predicting change in International Prostate Symptom Score (IPSS) at six horizons over 5 years, with nested cross-validation, bootstrap intervals and corrected $t$-tests. Results: All cohort-level comparisons of DVH indices were declared equivalent, with no interval consuming more than 44\,\% of its equivalence margin. Individual agreement was weaker with equivalence rates from 55.9\,\% to 89.7\,\%, depending on the DVH index and automatic segmentation model. Thresholds ranged from 0.864 to 0.958 Dice. No DVH block improved IPSS prediction at any horizon. The largest improvement declared by bootstrap intervals was 0.12 IPSS points, well below the minimal clinically important difference, across all learners. Conclusion: Automatic contours matched expert dosimetry within human IOV on average, but individual equivalence requires a volume dependent quality metric threshold definition. Our DVH indices panel provides no significant predictive power for IPSS, irrespective of segmentation source and horizon.

physics.med-ph↗
arXiv2026

The Electrodynamic Basis of Dichroism-Mediated Polarization Perception

Humans see the polarization of light through entoptic percepts arising from the macula's Henle fiber layer, where xanthophyll pigments absorb preferentially across the radiating fibers. From the layer's complex dielectric tensor alone, Maxwell's equations in Berreman $4\times4$ form yield its Mueller matrix, set by four scalars $\{A,B,C,D\}$. Its intensity channel depends only on the dichroic pair $A$ and $B$, which fix the percepts at a maximum contrast $|B|/A\approx0.05$. One relation generates the whole dichroism-mediated family: Haidinger's brushes under a uniform field, their dark arms perpendicular to the $\mathbf{E}$-vector; fractured brushes under spatially varying fields; and $N$-fold brushes under vector-vortex illumination.

physics.med-ph↗