A Novel Sphere Packing Solution Via Principal Component Analysis for Lattice SFRT Treatment of Large, Bulky, and Unresectable Tumors 📝

Author: James Robert Castle, Joshua Misa, Damodar Pokhrel, William St. Clair 👨‍🔬

Affiliation: University of Kentucky, Department of Radiation Medicine 🌍

Abstract:

Purpose: We propose a novel methodology for lattice deployment in spatially fractionated radiotherapy (SFRT), using principal component analysis (PCA) to improve sphere packing within the tumor. Improving sphere packing to increase the volume of peak dose may amplify tumor debulking and sensitization for further combination therapy.
Methods: The split-PCA (s-PCA) method uses PCA to rotate the reference frame so one of the Cartesian coordinate axes is aligned between the tumor’s first and second principal axes. This study included 20 previously treated SFRT (289.0-1266.8cc) prescribed to 15Gy. These patients were replanned using no PCA optimization (n-PCA), s-PCA, and 1-PCA (alignment along the tumor’s first principal axis), all with 4-full arcs, collimator angles of ±15°, 6MV-FFF energy, and a lattice diameter of 1.5cm and spacing of 3cm. Dose metrics extracted were: D50%, Dmean, and V50% to analyze direct cell-kill; D10%, D90%, and peak-to-valley-dose-ratio (PVDR=D10%/D90%) for indirect cell death; and Dmax to bladder, rectum, and bowel.
Results: The s-PCA method demonstrated a significant increase in spheres placed within the tumor volume, 13.7±7.8, compared to n-PCA, 12.3±7.8 (p<0.001), and 1-PCA, 12.8±8.5 (p=0.044). Additionally, s-PCA resulted in enhanced direct cell-kill metrics of D50% (Δ=1.3Gy, p<0.001; Δ=0.6Gy, p=0.060), Dmean (Δ=0.9Gy, p<0.001; Δ=0.3Gy, p=0.032), and V50% (Δ=7.1%, p<0.001; Δ=2.2%, p=0.121) compared to 1-PCA and n-PCA respectively. All three methodologies showed no statistical differences in PVDR, but s-PCA provided higher D10% (Δ=0.6Gy, p=0.003; Δ=0.3Gy, p=0.042) and D90% (Δ=1.2Gy, p<0.001; Δ=0.7Gy, p=0.012) compared to 1-PCA and n-PCA respectively. Lastly, all three methods showed insignificant differences in Dmax to the bladder, rectum, and bowel.
Conclusion: This s-PCA method enhances sphere packing, increasing direct cell-kill to bulky tumors while maintaining indirect cell death and maximum dose to adjacent critical organs. This methodology may improve the clinical outcomes of patients whose tumors are highly irregular and where sphere packing poses an obstacle.