Beam Application and Monitoring System response to low-fluence and low- low-intensity irradiationslow-intensity irradiations

3.3. Low-fluence beam irradiation and monitoring

Figure 3.6: Typical response of the𝑡_𝑅𝑃 on low-dose (500 primaries/RPat 1 mm separation) irradiations with the first nominal intensity enforced (a) and when the reference intensity allowed by theBAMShas been modified (b).

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000

Sample number [RPs]

0 0.5 1 1.5

tRP [s]

#10^-3

t_RP typical low-dose irradiation FOV = 16 # 3 cm² 500 ions/RP, 5022 RPs expected min = 0.16 ms

max =1.43 ms mean = 0.28 ms std (<) = 0.09 ms

(a) Behavior of the 𝑡_𝑅𝑃 in a low-dose scenario (500 ions/RP at 1 mm separation). A FOV of 16.1 × 3.0 cm² was scanned with a 299.94 MeV/u ¹²C ion beam of 3.9 mmFWHM, using the lowest nominal intensity from theLIBC (2 ⋅ 10⁶ pps) acquired with the shortest integration time (𝑡_𝑖𝑛𝑡= 55 𝜇s) allowed by the read-out electronics (cf. Section3.4). ThePBPneeded less than one-full spill and about 1.3 s to be concluded, leading to minimum intervals betweenRPsof 165𝜇s in∼14%of the samples. ∼6%of the expected number of samples (5022RPscounts in this instance) were missed.

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000

Sample number [RPs]

0 1 2 3 4

tRP [s]

#10^-3

t_RP typical low-dose irradiation Intensity reduced FOV = 16 # 3 cm² 500 ions/RP, 5022 RPs min = 0.33 ms max = 4.40 ms mean 1.16 ms std < = 0.38 ms

Longer times elapsed in between two spills

(b) The low-dose-PBP timing showed on the left lies on the border of the electronics capabilities to acquire synchronous data to the active scanning. Therefore the BAMS reference intensity allowance was reduced from 3.65 nA/RPto 0.8 nA/RPin order to slow the irradiation time by about a factor of four, so that 𝑡_𝑠 is enlarged (𝑡_𝑠 = 165 𝜇s) and all the RPs are registered. Note that when the beam is requested through theMBP, the dynamic intensity control [Schoemers et al. 2015] is not activated, hence the rising slope of the spill is not as steep as the one in Figure3.3, which has an impact in the𝑡_𝑅𝑃 along the irradiation.

The solution to this issue is to slow down the irradiation rate. Based upon the principle described in Section 3.3.1 to determine the reference intensity value allowed by the BAMS via equation3.1, the low-dose fast-PBP, whose timing is represented in Figure 3.6a, yielded a reference current value of 3.65 nA/RP, corresponding to a reference charge measurement by the firstBAMS PPICof 0.92 pC/RP, since its integration time was also extended to 250 𝜇s to allow longer signal accumulation. The third measurement range was used, which guaranteed reliable current measurements up to 170 nA.

The current allowance in theMBPwas then downsized to 0.8 nA/RPin order to slow down the plan by increasing the irradiation time to 5.8 s to be completed with∼1.3 spills (cf. Figure 3.6b). This configuration assured the detection of the total expected number of RPs with an extended sampling time of𝑡_𝑠= 55 𝜇s ⋅3^𝑐/𝑠= 165 𝜇s.

Consequences of these time fluctuations affecting the DAQ performance,detection efficiency and noise mitigation methodswill be addressed in the subsequent sections.

3.3.4 Beam Application and Monitoring System response to fluence and

explored at HIT before, in consequence, the BAMS have never been calibrated for the low number of particles intended for transmission imaging purposes (cf. Section 3.2). In order to give a realistic estimation of the dose deposited in the imaged phantoms, the exact number of ions delivered per scanning step must be known accurately. Therefore, theBAMS response to low-fluence and low-intensity irradiations was investigated by means of aBragg-Peak Ionization Chamber (BPIC)^C as external reference device, in order to verify whether the number of irradiated particles registered by theBAMSand recorded in thePBRscoincide with the actual number of particles exiting the nozzle. The experimental description of these investigations are explained after introducing briefly theBAMS Ionization Chambers (ICs)calibration procedure in the clinical scenario at the HIT facility.

As it was described in theSection 2.3.3, theBAMSconsists of threeICscommitted to monitor and control the beam fluence before and during the delivery of the treatment. The BAMS ICs work in a threshold-mode, which is determined by a pedestal value and its deviation margin, given in terms ofAnalog-to-Digital-Converter (ADC)counts. Thus, theICswill provide inexact particle counts whenever the induced charge signals are close to the established threshold in the integrator.

TheTCSrelies mainly on the feedback from the firstIC(IC1), being its charge accumulation constraints the highest for safe clinical applications. In other words, the IC1 reading will show the number of particles which is closest to the requested ions in the PBP, keeping the dose safely bounded. IC2 widens the charge pedestal tolerance to a certain extent, thus showing a more realistic response to be compared with the BPIC measurements, even when the dose security-thresholds are not fulfilled.

The BAMS has its own independent dosimetric calibration method, owing to the active beam delivery system for which it is used. This calibration is based on the code of practice for dosimetry (absorbed dose to water) established by theIAEA, however, the reference conditions are adjusted according to the raster scanning system and facility-specific dynamic energy range.

This energy-dependent calibration is experimentally validated by absorbed dose measurements in homogeneous fields at a few representative energies using a Farmer-type ionization chamber^D. The dose is quantified at the plateau region of the BCs. The full energy range is then obtained by an interpolating fitting function. The agreement between measured and calculated doses is usually better than 1% for square mono-energetic fields. The procedure is very similar to the one described by Jäkel et al.[2004], although atHIT, the fitting curve is based on the Bethe-Bloch formulawith two additional free parameters. The collected charge to number of particles conversion for each beam energy and ion species relies on the geometry and parameters of the BAMS ICs (cf. Table 2.1) and stopping power tables obtained both, analytically and by MC (FLUktuierende KAskade (FLUKA)) simulations, including the full HIT beam-line [Parodi et al. 2012b].

TheBPIC is a plane parallelICwith disk shaped air-filled sensitive volume of 2.5 cm³with an electrode spacing of 2.1 mm. The electrode radius is 40.8 mm. The entrance and exit windows have a respective thickness of ∼3.5 mm and ∼6 mm. The housing of the BPIC is made of PMMA, hence these thicknesses correspond to water equivalent path lengths of ∼4

C© Bragg-Peak ionization chamber by PTW Freiburg GmbH (http://www.ptw.de/bragg_peak.html)

DFarmer ionization chamber by PTW Freiburg GmbH (http://www.ptw.de/farmer_chambers0.html)

3.3. Low-fluence beam irradiation and monitoring mm and∼7 mm, respectively.

In the same way as it is regularly done for the above described QA protocols, the measure-ments with theBPIC were performed at the entrance region of the beam, i.e., at the plateau region of theBC. TheBPIC was located at 13 cm downstream of the beam-line nozzle outlet, aligned with the beam direction. This choice is justified by three main reasons; first, to ob-tain a reasonable comparison between the external device (BPIC) and theBAMS ICsoutputs.

Second, in transmission imaging applications, the plateau region is the most important zone for dose assessment. And third, owing to the energy dependency of ionization chamber out-comes, measuring in the steady part of theBC allows to fairly extrapolate a similar behavior for other energies. Absolute charge measurements were obtained from theBPIC, adjusting the measurement range according to the charge level expected from each beam irradiation configu-ration. The performance comparison between theBAMS ICsand theBPICwas done in terms of particle number. The PBRs of the first two ICs of the BAMS provide these numbers for each irradiatedPBP, while the number of ions detected byBPIC is estimated as follows:

𝑁_{𝑖𝑜𝑛𝑠}

𝐵𝑃 𝐼𝐶 = 𝑄 ⋅ 𝑊 (^𝑑𝐸_𝑑𝑥)

𝑤

𝑥_𝑤

, (3.5)

where:

𝑥_𝑤≈(2.1±0.1)⋅10⁻³ mm WEchamber air gap.

𝑊_{𝑐𝑎𝑟𝑏𝑜𝑛}=34.5±0.5 J/C (in dry air).

𝑊_{𝑝𝑟𝑜𝑡𝑜𝑛}=34.23 ±0.14 J/C (in dry air);𝑊_{𝑝𝑟𝑜𝑡𝑜𝑛}=34.8 ±0.5 eV (in humid air).

𝑄is given in Coulombs as the absolute charge measured by the BPIC with an uncertainty of 1%given by the manufacturer documentation.

(^𝑑𝐸/𝑑𝑥)

𝑤

represents the FLUKA depth dose in water (from vacuum), evaluating the dose per unitary fluence (^𝑑𝐸/𝑑𝑥 in water for 𝜌= 1 ^𝑔/𝑐𝑚³) at the specified depth of 6 mm [3.11mm (BPIC) + 2.89 mm (Vacuum Window, Air and BAMS) WE]. This calculation is based on FLUKAv2011.2c.

Bragg Peak Ionization Chamber (BPIC) energy calibration

In order to rule out any energy dependence of theBPICon the charge collection and to compare its performance with respect to theBAMS ICs, theBPICwas exposed to proton and carbon-ion beams of six differentLIBCenergy settings: E255, E209, E157, E104, E58, E21, corresponding to 221.05 MeV, 190.48 MeV, 160.09 MeV, 130.52 MeV, 100.46 MeV and 70.03 MeV for proton beams and to 430.1 MeV/u, 366.78 MeV/u, 305.27 MeV/u, 246.57 MeV/u, 187.97 MeV/u and 129.79 MeV/u for carbon-ion beams, yielding approximately the same ion range in water (cf.

Figure 3.7). For both ion types, the TCS irradiation was configured to deliver a fixed high number of particles at the isocenter, thus discarding any BAMS ICs inaccuracy due to low-fluence acquisition. The required number of particles to be delivered at the center of theBPIC was 2 ⋅ 10⁹ protons and 5 ⋅ 10⁷ carbon-ions. The beam intensity was the one automatically chosen by the TCS PDG, i.e., I10 = 3.2 ⋅ 10⁹ pps for protons and 8 ⋅ 10⁷ pps for carbon-ion beams, in all the irradiation configurations.

For both ion types, the lowest energies show a slightly larger discrepancy from the originally required number of primaries, due to the proximity of the BP to the chamber. Without

Figure 3.7: Bragg Peak Ionization Chamber (BPIC) energy calibration. A fixed high number of (a) protons (2 ⋅ 10⁹) and (b) carbon-ions (5 ⋅ 10⁷) was requested to be delivered at the isocenter at six representative energies, covering the energy span available at HIT. The number of ions counted by theBPICwas determined with the Equation3.5 and compared to the readings registered in thePBRsof the first twoBAMS ICs, as well as the originally requested number of primaries.

1.85 1.9 1.95 2 2.05 2.1 2.15 2.2

////

221.05 ////

190.48 ////

160.09 ////

130.52 ////

100.46 ////

70.03

Particle/counts/////

////

x/10⁹

Energy/[MeV/u]////

////

Requested/primaries ////BAMS/IC1 ////BAMS/IC2 ////BPIC

(a) Proton counts comparison between the BPIC out-come and the number of protons registered in thePBR by theBAMS ICs. As reference, the target dose (2 ⋅ 10⁹ protons) is indicated in light gray. Protons were irradi-ated at the center of theBPIC, located in the plateau region of theBC(13 cm downstream of the nozzle exit-window), at six different beam energies: 221.05 MeV, 190.48 MeV, 160.09 MeV, 130.52 MeV, 100.46 MeV and 70.03 MeV.

]]]]

4.55 ]]]]

4.65 ]]]]

4.75 ]]]]

4.85 ]]]]

4.95 ]]]]

5.05 ]]]]

5.15 ]]]]

5.25

]]]]

430.1 ]]]]

366.78 ]]]]

305.27 ]]]]

246.57 ]]]]

187.97 ]]]]

129.79

Particle]count]]]]]

]]]]

x]10⁷

Energy][MeV/u]]]]]

]]]]

Requested]primaries ]]]]BPIC ]]]]BAMS]IC ]]]]BAMS]IC2

(b) Carbon-ion counts comparison between the BPIC outcome and the number of ions registered by theBAMS ICs. The expected number of carbon-ions (5 ⋅ 10⁷) is represented in gray. Analogously, carbon-ions were irra-diated at the center of theBPIC, located in the plateau region of theBC(13 cm downstream of the nozzle exit-window) at six different beam energies: 430.1 MeV/u, 366.78 MeV/u, 305.27 MeV/u, 246.57 MeV/u, 187.97 MeV/u and 129.79 MeV/u.

taking into account the lowest energy, the BPIC estimation of protons and carbon-ions is in agreement to the target fluence within a difference lower than1% and 1.5%, respectively, and not beyond their corresponding estimated error. Transmission imaging applications usually require energies higher than 200 MeV/u, for which the BPIC exhibits a stable performance with no strong energy-dependence.

BPIC response for typical fluences available at HIT

In order to investigate the response of theBPICon the application of different typical fluences used in a standard treatment scenario, the highest carbon-beam-energy available atHIT(430.1 MeV/u) was chosen and2 ⋅ 10⁶,5 ⋅ 10⁶,8 ⋅ 10⁶,1 ⋅ 10⁷,2 ⋅ 10⁷,5 ⋅ 10⁷ primaries were requested to be delivered at the nominal intensity chosen by the PDG (I10 = 8 ⋅ 10⁷ pps) (cf. Figure 3.8). The BPIC and the BAMS were found to accurately coincide with the target fluences requested. As expected, the IC1 shows the best agreement (−0.18% difference), followed by the IC2 (−0.54% difference). Nonetheless, on average, the BPIC does not exceed the 1.5%

uncertainty for this specific energy.

BPIC and BAMS ICs response to reduced beam intensity

Once theBPICperformance is characterized for the standard irradiation conditions, the accel-erator procedure to decrease the particle delivery rate (cf. Section 3.3.3) should be tested. To this aim, a mono-energetic 430.1 MeV/u ¹²C beam was used, the targeted number of particles was settled to be 2⋅10⁶, which is determined by the number of particles delivered by the lowest nominal intensity available at HIT (2⋅10⁶ delivered in one second), and the particle delivery

3.3. Low-fluence beam irradiation and monitoring

BBBB -2.00 BBBB -1.50 BBBB -1.00 BBBB -0.50 BBBB 0.00 BBBB 0.50 BBBB 1.00 BBBB 1.50 BBBB 2.00 BBBB 2.50 BBBB 3.00 BBBB 3.50

0 1 2 3 4 5

BBBB

0.00E+000 BBBB1.00E+007 BBBB2.00E+007 BBBB3.00E+007 BBBB4.00E+007 BBBB5.00E+007

SBdifferenceBBBBB

ParticleBcountsBBBBB xB10BBBBBB7

RequestedBparticlesBBBB

BBBB

BPIC BBBBBAMSBIC1 BBBBExpected BBBBBAMSBIC2 BBBBSBdifferenceBfromBexpected

Figure 3.8: BPIC calibration for typical fluences available at HIT. To perform a scan over the usual number of primaries requested by treatment plans, a 430.1 MeV/u¹²C beam was used to assess the accuracy of theBPICto recover the amount of particles requested. The target dose (number of particles) was reduced from5 ⋅ 10⁷to2 ⋅ 10⁶ and a linear behavior was confirmed among the three ICs compared. On average, the BPIC chamber ion-counts coincides with the expected primaries within1.5%, although differences up to3%can be found.

rate was reduced by decreasing the current allowance of the BAMS IC1, starting from the reference current value defined by the PDG (cf. Figure 3.9). Already on the forth reduction step (1 nA), theBPIC shows a difference of∼20% from the expected particle count, which is not detected by the BAMS IC1. Further reduction of the reference current value allowed by the BAMS (∼fA), might yield discrepancies up to 80% primaries count. This BAMS limita-tion should be taken into account when low-dose irradialimita-tions are performed to determine the optimal and reliableBAMS ICreference current.

BPICandBAMS ICs response to a fluence reduction under nominal particle rates The accelerator research working-mode also allows to fix the desired number of particles to be delivered while deciding which beam nominal-intensity should be used, so that the PDG is sidestepped. Under this configuration, a carbon-ion beam of a typical energy used for transmission imaging was chosen (E189 = 339.8 MeV/u) to impinge 2 ⋅ 10⁶, 1 ⋅ 10⁵, 1 ⋅ 10⁴, 1 ⋅ 10³,5 ⋅ 10² and1 ⋅ 10² ions on theBPIC (cf. Figure3.10). The higher numbers of primaries requested perRP(2 ⋅ 10⁶ and1 ⋅ 10⁵) were irradiated without inducing safety interlocks at the tenth carbon-beam LIBCintensity, I10 = 8 ⋅ 10⁷ pps. In order to deposit 1 ⋅ 10⁴ per RP, the beam rate was decreased by one order of magnitude to the I4 =8 ⋅ 10⁶ pps, while the rest of the particle requests demanded the lowest carbon-beam intensity available, I1 =2 ⋅ 10⁶pps. As it was mentioned in Section3.2, the dose schemes used in this work fall within the limit of the HITcapabilities to deliver low number of particles and register them accurately by theBAMS records, as it is evident from the percentage of discrepancy between theBPIC and the BAMS IC2 readings indicated on the right axis of Figure 3.10. By extrapolating the curve tendency (due to missing data point), thehigh-dose scheme might display up to27%disagreement from theBAMS IC2 readings, while the middle-and low-dosescenarios may differ from the PBRs up to90% and 142%, respectively.

0 [[[[

0.5 1 [[[[

1.5 2 [[[[

2.5 3 [[[[

3.5 4 [[[[

4.5

Particle[counts[[[[[

[[[[

x[10[⁶

BAMS[IC[Current[Reference[value[[A][[[[

[[[[

Requested[primaries [[[[BAMS[IC1 [[[[BPIC

Figure 3.9: BPICandBAMS IC1 response to reduced beam intensity. 2⋅10^{6 12}C-ions at 430.1 MeV/u were requested for delivery while the reference current accepted by theBAMSIC1 was reduced in 12 steps. When the current allowance by the IC1 is reduced to∼1 nA theBAMSfails to detect∼20%ions of the expected particle goal. Further reduction of theBAMSIC1 reference current leads to higher inaccuracies, which should be avoided. Optimally, currents of∼10 nA should guarantee the best trade-off between the correct detection ofRPsby the read-out electronics, long-enough integration times and with reliable dose delivery.

0 llll 50 llll 100 llll 150 llll 200 llll 250 llll 300 llll 350 llll 400 llll 450

llll 1.0E+00 llll 1.0E+01 llll 1.0E+02 llll 1.0E+03 llll 1.0E+04 llll 1.0E+05 llll 1.0E+06 llll 1.0E+07

llll

2.0E+06 llll1.0E+05 llll1.0E+04 1.0E+03llll llll5.0E+02 llll1.0E+02 plofldifferencelbetweenlBPIClandlBAMClIC2lllll

Pa rti cl el co u n tslllll

Requestedlprimariesllll

llll

Requestedlprimaries llllBPlIC llll

BPlIC llllBAMSlIC1

llll

BAMSlIC2 % underestimation of BAMS IC2

Figure 3.10: BPIC and BAMS ICs response to fluence reduction under nominal particle rates. A 339.8 MeV/u

12C-ion beam, which is in the energy-range used normally for transmission-imaging purposes, was chosen to irradiate 2 ⋅ 10⁶, 1 ⋅ 10⁵,1 ⋅ 10⁴, 1 ⋅ 10³, 5 ⋅ 10² and1 ⋅ 10² ions at the isocenter, traversing the BPIC in the plateau of theBC. For this investigation, thePDGwas evaded by enforcing a certain nominal beam-intensity before the particle delivery. To deposit2 ⋅ 10⁶and1 ⋅ 10⁵carbon-ions perRP, I10 =8 ⋅ 10⁷ppswas used. Requesting1 ⋅ 10⁴ ions per RPalready required a reduced beam-intensity by one order of magnitude, while the last three doses (1 ⋅ 10³,5 ⋅ 10² and1 ⋅ 10²carbon-ions) were achieved by selecting the I1 =2 ⋅ 10⁶ ppsbefore the irradiation started.

Dose implications

The dosimetric implications perRPof the particle counting mismatches presented in theformer section, at thethree different dose-levelsinvestigated in this work, are summarized in the Table

3.4. Synchronized data acquisition with the active delivery system and updated electronics 3.1. The dose is estimated by the Equation2.19, in the plateau region of a 339.8 MeV/u¹²C-ion beam. Exemplary scanning steps of 1 mm and the particle fluences obtained with the different ICsare considered for the calculation.

Table 3.1: Dose calculation comparison given by theBAMS ICsandBPICparticle outcome when irradiated at the three different dose-schemesconsidered in the experimental investigations of this work.

requested primaries BAMS IC 1 BAMS IC 2 BPIC

Low-dose 500 532 477 1208 ions

Dose 0.95 1.01 0.91 2.29 mGy

Middle-dose 1000 1069 1226 1898 ions

Dose 1.90 2.04 2.33 3.61 mGy

High-dose 5000 5166 5759 6348 ions

Dose 9.52 9.83 10.96 12.09 mGy

The results presented in this section show the limitations of theHITaccelerator and monitor-ing system under the non-clinical dose regime. They provide indicators that must be taken into consideration for accurate image dose estimation when acquiring low-dose transmitted-images.

3.4 Synchronized data acquisition with the active delivery system and updated

Im Dokument Low-dose ion-based transmission radiography and tomography for optimization of carbon ion-beam therapy (Seite 81-87)