Medical Physicist's Notes · Delta4 Phantom+
Modern radiotherapy is moving toward higher dose-per-pulse delivery — flattening-filter-free (FFF) beams, hypofractionated SBRT/SRS, and the emerging clinical interest in FLASH. The integrating detectors that work at conventional dose-per-pulse don't extrapolate cleanly to these regimes, for reasons that are well documented in the dosimetry literature. This page lays out the physics, the documented limits of conventional QA detectors, and what Delta4 Phantom+ does about it.
~1 mGy
typical dose per pulse
Conventional 6 MV linac, single dose pulse at iso. — the manufacturer figure.
50 nGy
Delta4 integration resolution
Per pulse — 1 part in 20,000 of a typical pulse. From the Manufacturer brochure.
≥40 Gy/s
FLASH instantaneous threshold
Pre-clinical sparing demonstrated above this rate — Vozenin 2019, Bourhis 2019.
A clinical linac delivers dose as a discrete train of bremsstrahlung pulses, each on the order of a few microseconds and depositing roughly 1 mGy at the isocentre under conventional 6 MV operation. Modern delivery techniques push that dose-per-pulse higher: FFF beams remove the flattening filter and raise the pulse intensity several-fold; SBRT and SRS use those higher rates to keep treatment times manageable at high prescription doses; and FLASH radiotherapy operates an order of magnitude higher again, at ≥40 Gy/s instantaneous, where pre-clinical normal-tissue sparing has been documented.
Sources: the manufacturer Delta4 Phantom+ brochure · Vozenin et al., Clin Cancer Res 25(1):35–42 (2019) · Bourhis et al., Radiother Oncol 139:18–22 (2019).
None of the points below is rhetorical. Each is paraphrased from a peer-reviewed source and cited inline. The implication is consistent across them: the dosimetry chain that worked at 0.1–0.5 mGy per pulse does not extend without explicit correction to FFF, SBRT or UHDR delivery.
General (Boag) recombination in vented ion chambers scales with the charge density produced by a single pulse. In flattening-filter-free (FFF) beams the dose-per-pulse is several-fold higher than conventional flattened beams, and the standard two-voltage Boag correction becomes inadequate — Wang & Rogers showed that the correction itself becomes dose-per-pulse dependent and recommends a polynomial recombination formalism specific to FFF.
Wang LLW & Rogers DWO, J Appl Clin Med Phys 13(5):3758 (2012)
Commercially available silicon diodes are not perfectly dose-per-pulse independent: characterised response can vary by a few percent across the dose-per-pulse range encountered when SSD or field size changes. The effect is small in flattened beams but becomes clinically relevant in FFF and short-SSD setups unless explicitly corrected.
Saini AS & Zhu TC, Med Phys 31(4):914–24 (2004), PMID 15259635
At ultra-high dose-rate (UHDR; ≥40 Gy/s instantaneous), Romano et al. document detector saturation and recombination effects across active dosimeters that are designed against conventional dose-per-pulse. The dosimetric chain that worked at 0.1–0.5 mGy per pulse does not extend without verification to the 1–10 mGy per pulse regime.
Romano F et al., Med Phys 49(7):4912–4932 (2022), PMC9544810
The pre-clinical advantage of FLASH (≥30–40 Gy/s) is documented in Vozenin 2019 in mini-pig and cat-cancer patients; Bourhis 2019 reports the first human FLASH case at 15 Gy in 90 ms. The biological effect is governed by the dose-rate during the pulse — not by the average treatment time — so any QA chain claiming FLASH coverage must resolve the pulse, not the integrated cumulative dose.
Vozenin MC et al., Clin Cancer Res 25(1):35–42 (2019) · Bourhis J et al., Radiother Oncol 139:18–22 (2019)
What does a patient-QA detector look like when its time base matches the time base of the linac — and the correction chain is shorter, not longer, at higher dose rates?
Each row pairs a documented dosimetric problem with the architectural decision in Delta4 that addresses it. Sources are inline so the mapping is auditable.
One pulse at a time — 50 nGy resolution
The manufacturer product brochure specifies: "Delta4 integrates one dose pulse (typically 1 mGy) at a time with a resolution of 50 nGy." The detector array is read out per pulse, not by integrating a fixed time window — so the time base of the measurement matches the time base of the linac.
Source: Manufacturer brochure — Delta4 Phantom+ wireless phantom.
0.04 mm³ p-type silicon diodes — no recombination correction needed
The 1,069 p-Si disc diodes (1 mm × 0.05 mm; active volume 0.04 mm³) are intrinsically dose-rate independent across the clinical photon range — silicon does not exhibit the Boag general-recombination loss that vented ion chambers do. The published detector stability is <0.1%/kGy, typically 0.04%/kGy at 6 MV.
Source: the manufacturer Delta4 Phantom+ technical specification.
No add-on hardware for high dose-rate measurement
High dose-rate verification on chamber-based or film-based chains typically requires dedicated UHDR-tolerant hardware (Romano 2022 reviews the available options — radiochromic film, alanine, plastic scintillator, calorimetry). The Delta4 array as shipped — without add-ons — operates on the same per-pulse read-out across conventional, FFF and UHDR delivery.
Source: Romano F et al., Med Phys 49(7):4912–4932 (2022); the manufacturer Delta4 Phantom+ specification.
Scope of this page
The 50 nGy per-pulse integration spec is taken from the manufacturer Delta4 Phantom+ brochure. The implications for FFF, SBRT and FLASH dosimetry above are grounded in peer-reviewed dosimetry literature on dose-per-pulse dependence and UHDR detector behaviour — not in head-to-head Delta4 vs competitor measurements at UHDR, which to our knowledge have not been published. Treat this page as an architectural argument that can be checked against the cited physics, not as a UHDR commissioning comparison.
Sources cited on this page
Delta4 Phantom+
Sibling posts in the Delta4 Phantom+ family.
For Medical Physicist
At what magnitude of MLC, gantry, collimator or MU error does the Delta4 gamma passing rate cross the 95 % action threshold? Per Udee 2024, J Med Phys.
Read this Medical PhysicistDepartments routinely run separate equipment for patient-specific QA and routine machine QA. What changes when the same 1,069-detector array runs both workflows under one software ecosystem?
Read this