Quantifying the spatial information loss of Dose-Volume Histograms in Gamma Knife radiosurgery via…

AI / informaticsRadiological physics and technology· Jul 2026

Quantifying the spatial information loss of Dose-Volume Histograms in Gamma Knife radiosurgery via unsupervised 3D dose geometry embedding.

Omar Hamzaoui, Yassine Oulhouq, Mohammed Rezzoug … Abdeslem Rrhioua

Stereotactic & FunctionalAI / informaticsQ3IF 1.6

Gamma Knife plans with matching dose-volume histograms can still have hotspots 2.81 mm apart.

Quantifying the spatial information loss of Dose-Volume Histograms in Gamma Knife radiosurgery via unsupervised 3D dose geometry embedding.

Omar Hamzaoui … Abdeslem Rrhioua

Radiological physics and technology · Jul 2026 · Q3 in General Medicine

Background

Dose-Volume Histograms (Dose-Volume Histograms) discard spatial information by summarizing 3D distributions into 1D curves.

Purpose

This study quantifies this loss in Gamma Knife radiosurgery and introduces TopoGK, a novel unsupervised deep learning framework capturing full 3D dose geometry.

Results

matched dose plans can still miss the same hotspot by several millimeters

TopoGK

0.451correl

PCA

0.287correl

More results

Dose grids were tumor-centered, resampled to [Formula: see text] voxels, and normalized to [Formula: see text].

A 3D Convolutional Variational Autoencoder compressed each dose-mask volume into a 64dimensional spatial embedding.

More results

Spatial information loss was quantified by comparing pairwise Dose-Volume Histograms and latent distances both globally and within volume-stratified subgroups.

Five-fold cross-validation ensured generalizability.

“

Conclusion

As a proof-of-concept in single-fraction vestibular schwannoma radiosurgery, these results motivate further investigation toward spatially-aware quality assurance in stereotactic radiosurgery.

Quantifying the spatial information loss of Dose-Volume Histograms in Gamma Knife radiosurgery via unsupervised 3D dose geometry embedding.

Quantifying the spatial information loss of Dose-Volume Histograms in Gamma Knife radiosurgery via unsupervised 3D dose geometry embedding.

Related papers