KEY POINTS
- Investigators evaluated uTPS planning for the Halcyon 2.0 dual-layer MLC accelerator using 20 retrospective patients: five each with hypopharyngeal, esophageal, breast and cervical cancer. Every case was replanned with uTPS and Eclipse using the same geometry, yielding 40 VMAT plans.
- Prescriptions reflected different clinical scenarios: hypopharyngeal plans used simultaneous integrated boost doses up to 69.96 Gy in 33 fractions, esophageal and breast plans used 50 Gy in 25 fractions, and cervical plans used 50.4 Gy in 28 fractions.
- To isolate optimizer performance, uTPS-generated plans were imported into Eclipse and recalculated with Acuros XB on a 2.5-mm grid, so target and organ-at-risk comparisons used the same final dose engine rather than comparing uTPS Monte Carlo directly with Eclipse.
- For esophageal and cervical cases, target conformity, homogeneity and all tested organ-at-risk metrics were statistically comparable between systems. Hypopharyngeal target conformity and homogeneity were also similar after multiple-testing correction.
- The largest significant difference was seen in hypopharyngeal plans: uTPS reduced brainstem D0.1cc from 30.51 to 19.99 Gy, a 34.5% reduction, left-parotid mean dose from 23.98 to 21.15 Gy, and left-parotid V30Gy from 29% to 27% (p=0.043 for each after correction).
- In the five breast cases, uTPS reduced spinal-cord D0.1cc by 49.6% relative to Eclipse; the other tested breast parameters were not significantly different. No significant differences in total monitor units were found for any disease group.
- All uTPS plans were successfully delivered on Halcyon. ArcCheck patient-specific QA using 3%/2-mm gamma criteria produced passing rates above 97% in every plan, with mean rates approximately 98.3–99.1% across the four disease groups and no significant difference from Eclipse.
CLINICAL TAKEAWAY
uTPS appears capable of optimizing technically deliverable VMAT plans for a Halcyon dual-layer MLC system despite being developed by another vendor, with plan quality broadly comparable to Eclipse. The study establishes technical feasibility, not superiority: five patients per disease site are insufficient to claim consistent dosimetric advantages.