Changes between Version 2 and Version 3 of Research/LhARA/RadiationBiology/Meetings/2026-02-24


Ignore:
Timestamp:
Mar 2, 2026, 2:24:37 PM (3 months ago)
Author:
ccd24
Comment:

--

Legend:

Unmodified
Added
Removed
Modified

  • Research/LhARA/RadiationBiology/Meetings/2026-02-24

    v2 v3  
    1111Phase 2 used post-plating, which allows for more statistics and provides better survival than pre-plating. Due to previous controls, a lack of survival, 1mm of media was added to the cell dish, which improved survival.
    1212Completed a dry run at SCAPA, where the cells were held vertically for 20minutes. This test saw good survival in both HeLa and FaDu cell lines. The results were studied 10 days post-seeding.
    13 However, in the actual irradiations, only the FaDu saw any survival. Two obvious factors for this are the uncertainty in the delivered dose and the increased time the cells were vertical for. The general operation of the irradiations meant that the cells were probably out for 1-1.5 hours. Most of this was because we were unaware that we needed to limit the time the cells were out for. This can be reduced further by reducing the time between shots, as well as not shooting RCF in the same carousel. Secondly, the actual irradiations were aiming to deliver 1, 2 and 4Gy using 4, 7 and 14 shots, but mroe likely delivered nearly twice that dose, with a large shot-to-shot variation in the dose delivered. This meant the seeding densities were off.
     13However, in the actual irradiations, only the FaDu saw any survival. Two obvious factors for this are the uncertainty in the delivered dose and the increased time the cells were vertical for. The general operation of the irradiations meant that the cells were probably out for 1-1.5 hours. Most of this was because we were unaware that we needed to limit the time the cells were out for. This can be reduced further by reducing the time between shots, as well as not shooting RCF in the same carousel. Secondly, the actual irradiations were aiming to deliver 1, 2 and 4Gy using 4, 7 and 14 shots, but more likely delivered nearly twice that dose, with a large shot-to-shot variation in the dose delivered. This meant the seeding densities were off.
    1414In the FaDu, a survival curve was seen, though with the dose uncertainty, it is hard to draw any conclusions. The plating efficiency was also at a lower level across the board, most likely due to the time out rather than the variation in dose. Having said this, the survival curve did look realistic.
    1515
    … …  
    3535* Proton Focus Imager (Lanex)
    3636* Cell Irradiation (Carousel)
    37 In phase 1 saw a maximum source energy of 12±2MeV. This increased to 14.4±1.2MeV, which could be due tp higher pulse intensity, but this does not fully explain the result. There was a comparison made between Kapton and Steel targets. Steel produces a much more circular proton beam but leads to EMP issues.
    38 Starting to investigate proton source instabilities as the potential cause of the shot-to-shot variation. There is a noted decrease in max source energy by 1MeV over 120 shots. This occurs due to debris buildup in the laser near-field spatial-intensity profile. Robbie also has a student that has started to reproduce dose variations byt varying the source energy spectrum. Would be interesting to further vary the spatial distribution and the angular distribution and see if these are enough to explain the change in dose. Needs to include the PMQs in the models as well.
     37In Phase 1, we saw a maximum source energy of 12±2MeV. This increased to 14.4±1.2MeV, which could be due to higher pulse intensity, but this does not fully explain the result. There was a comparison made between Kapton and Steel targets. Steel produces a much more circular proton beam, but it leads to EMP issues.
     38Starting to investigate proton source instabilities as the potential cause of the shot-to-shot variation. There is a noted decrease in max source energy by 1MeV over 120 shots. This occurs due to debris buildup in the laser near-field spatial-intensity profile. Robbie also has a student who has started to reproduce dose variations by varying the source energy spectrum. Would be interesting to further vary the spatial distribution and the angular distribution and see if these are enough to explain the change in dose. Needs to include the PMQs in the models as well.
    3939For the next beamtime, there are 3 key areas to work on:
    4040* Cell Studies
    4141* Developing/Adding Further Beamline Elements
    4242* Diagnostic Development, including Diagnosing the Source of the Variation
    43 The third of these is the most pressing
     43The third of these is the most pressing.
    4444Based on availability, the best time for new beamtime would be around July/August.
    4545
    … …  
    6161Firstly, Tony's project of CMOS. This is too thick to be used at PoPLaR, and also has an issue with the dose rate.
    6262
    63 A very thin phosphor sheet. This is work with Simon Joly. Can deposit on thin Kapton. Peter has also worked with phosphor experts at Brunel. This could in theory be added to the bottom of a newly designed cell dish. We are calling this the Instrumented Cell Dish Bottom technique. The main issue would be the optics and transport of the light. Could make the cell dish lid see through and carry the light out the other side?
     63A very thin phosphor sheet. This is work with Simon Jolly. Can deposit on thin Kapton. Peter has also worked with phosphor experts at Brunel. This could in theory be added to the bottom of a newly designed cell dish. We are calling this the Instrumented Cell Dish Bottom technique. The main issue would be the optics and transport of the light. Could make the cell dish lid see through and carry the light out the other side?
    6464
    6565Ultra-thin Secondary Electron Emission. Will not work for our energies at the current spec. It is comprised of nm Al on um Kapton. Tony is going to test some at Birmingham. There will be a problem with putting this in the vacuum chamber due to the EMP and the electronics involved in the detector. It is possible that there would also be interest in this from CNRS. If it becomes sensitive enough for our aims, then it may be possible to generate an x-y profile on a 10x10 grid.
    … …  
    6868
    6969
    70 
    71 
    7270==== RCF Use and Potential Diagnostic ====
    7371
    74 Diaza has written an RCF protocol explaining key issues, including film orientation and placement, scanner warm-up and the Callier effect. For the last one, we require a glass sheet to cover the film when being scanned. Further reading was undertaken on the RCF calibration and a technical document will be added to the wiki to describe the best practice. One key insight is that using the Polynomial Model but fixing c provides a much lower error when calculating the dose at little cost to the accuracy of the measurement. The error reduced from between 5 and 16% to 3-4.5% in a range of 0-18Gy. It was also suggested that a calibration could be undertaken using all three channels as this reduces the error further.
     72Diaza has written an RCF protocol explaining key issues, including film orientation and placement, scanner warm-up and the Callier effect. For the last one, we require a glass sheet to cover the film when being scanned. Further reading was undertaken on the RCF calibration, and a technical document will be added to the wiki to describe the best practice. One key insight is that using the Polynomial Model but fixing c provides a much lower error when calculating the dose at little cost to the accuracy of the measurement. The error reduced from between 5 and 16% to 3-4.5% in a range of 0-18Gy. It was also suggested that a calibration could be undertaken using all three channels, as this reduces the error further.
    7573
    76 The larger RCF films were then used to determine whether placing RCF around the edge of the cell dish would allow you to estimate accurately the mean dose in the cell dish. It was found that measurement error alone could not explain the residuals from the linear fit, so a constant error term to measure the film-to-film variation was added. Once this uncertainty was increased to 0.23Gy the reduced chi-squared dropped from around 6 to around 1. This means that this error will become substantial at low doses.
     74The larger RCF films were then used to determine whether placing RCF around the edge of the cell dish would allow you to estimate the mean dose in the cell dish. It was found that measurement error alone could not explain the residuals from the linear fit, so a constant error term to measure the film-to-film variation was added. Once this uncertainty was increased to 0.23Gy the reduced chi-squared dropped from around 6 to around 1. This means that this error will become substantial at low doses.
    7775
    78 Simulations were then undertaken to determine whether placing a piece of delaminated EBT3 would be a sensible approach. They suggested that the introduction of the RCF would increase the track-LET from 5.7 KeV/um to 6.5 KeV/um and the dose-LET from 6.8 KeV/um to 7.7 KeV/um, ie the RCF causes roughly a 1KeV/um increase in the LET. This was deemed acceptable, though other approaches are still preferred. There is also further analysis required on the ability to use this RCF to predict the delivered dose to the cell dish.
     76Simulations were then undertaken to determine whether placing a piece of delaminated EBT3 would be a sensible approach. They suggested that the introduction of the RCF would increase the track-LET from 5.7 KeV/um to 6.5 KeV/um and the dose-LET from 6.8 KeV/um to 7.7 KeV/um, i.e., the RCF causes roughly a 1KeV/um increase in the LET. This was deemed acceptable, though other approaches are still preferred. There is also further analysis required on the ability to use this RCF to predict the delivered dose to the cell dish.
    7977
    8078==== Summary ====
    8179
     80Suggested diagnostics:
     81* Sparse Sci-Fi Array
     82 * Prototype being developed
     83 * Main concern is transporting the light outside the chamber
     84* OAFM
     85 * A systematic study to find the correlation between dose and intensity detected is the next step for this
     86* CMOS
     87 * Too thick
     88* Thin phosphor sheet
     89 * Similar to Sci-Fi, the main concern is transporting the light
     90 * Better version of delaminated EBT3
     91* Ultra-thin Secondary Electron Emission
     92 * Will not work at the current spec
     93* Secondary Standard Calorimeter
     94 * Would need to be combined with a method to get the proton energy
     95 * Sensitive to 0.3Gy (possibly)
     96* RCF around the cell dish
     97 * Error is considerable below 1Gy
     98* RCF in front of the cell dish
     99 * Increase in LET is not an issue
     100 * Further study on the accuracy of estimation required
     101
     102
     103The RCF methods can be implemented almost immediately, with only minor research needed to verify the RCF in front of the cell dish. Combining this with several sparse scintillating-fibre arrays appears to be the best plan for the next run, and while the errors are still large, they are far smaller than the errors due to shot-to-shot variation.
     104It then seems reasonable to introduce a phosphor sheet to replace the RCF, as this would allow measurements to be analysed on shot, rather than waiting 24 hours.
     105This also gives time for further development of the secondary standard calorimeter and the OAFM, which could be used in future iterations of the beamline.
    82106
    83107=== Looking Forward ===
    84108
    85109==== Future Beamlines ====
     110Josie has completed several simulations varying the number of PMQs and the drifts between the beamline elements. Using Bayesian optimisation and with the restriction that the minimum drift to the first PMQ and the minimum drift between PMQs is 93mm, the best result was in a 2 PMQ setup. Currently running a Bayesian optimisation that releases all the constraints to find the best setup, whether physically feasible or not. We can then probe the restraints that need to be broken to see how flexible they are.
     111
     112==== Biology Next Steps =====
     113
     114Aim for good data is 3 independent experiments on 3 different days. The core data would be the survival curves, though this can be supplemented with IF and comet assays. There is also a clear benefit to increasing the number of cell lines being used.
     115
     116The simplest next step is to do the 3 independent experiments with the same cell lines and record the survival curves. This would be a repeat of Phase 2, but with hopefully a better understanding of the delivered dose and the cell handling.
     117It seems that adding comet assays to this would be feasible, and with the help of Marie Boyd, quite simple. This might be enough pilot data to make a plan going forward, and potentially a funding application, though MRC would require a lot more data.
     118
     119Additionally, results can be added by completing IF studies and increasing the number of cell lines we use.
     120
     121Then, in the future, we can look at improving the beamline to deliver the high-dose rates for FLASH.
     122
     123Emma will write up a plan for the biology going forward, to build a funding bid for October.
     124
    86125
    87126==== Discussion and Summary ====
    88 Robbie going to ELI soon
    89127
     128Bio results look better than expected based on the shot-to-shot variation, the cell handling and the variation across the dish. Can improve the cell handling by making a few changes, including reducing the time between shots and not irradiating RCF after the cells. Shot-to-shot variation means that an in-beam diagnostic is required. Spatial variation can be reduced by moving the scatterer closer to the laser target, but hopefully, a better approach using multiple PMQs can be achieved. This is what Josie is investigating with Bayesian Optimisations and studying the ELI beamline. Robbie is going to ELI soon, so that will hopefully help answer any questions.
    90129
    91 == Actions Required ==
    92 * Get a glass sheet
    93 * Investigate dose prediction from RCF infront of cells
     130Robbie also has a student who is trying to understand how the variations in the source lead to the variation in dose and trying to diagnose that problem more clearly. He also advocates for the next beamline time to be used to help complete this work and not cell irradiations. This would also allow information to be gathered so that a write-up of the beamline can be conducted.
    94131
     132The in-beam diagnostic quick fix is RCF, and hopefully, by the next beam time, we will also have a sparse sci-fi array. The phosphor sheet is the next obvious upgrade after this. Diagnostics improving on this require further development but should be kept an eye on.
     133
     134Substantial bio results are required, and the next beamtime would most likely be used to complete repeats of Phase 2, but with a better beamline and Marie Boyd's help. Marie Boyd can also conduct comet assays, which can be incorporated. A decision is required about whether we use Marie's lab.
    95135
    96136
    … …  
    98138
    99139Phase 2 Analysis
    100 - **EM**: Evaluate the cell results
    101 - **CD, AFr**: Complete analysis of RCF with errors
    102 - **EM/ED/LJ**: Send calibration films to SCAPA and scan them in both orientations
    103  - **CD, RW and AFr**: Explain scanning procedure
    104 - **EM, JMcG**: Write up summary of PoPLaR Phase 2, including technical summary of cell irradiation procedure
    105 - **MB**: Comet analysis
    106 - **CD**: Invite Liverpool, Chris Armstrong and LMU to join the in-person meeting
     140- **DONE**
    107141
    108142Beam Diagnostics
    109 - **Unassigned**: Find all the errors associated with RCF
    110143- **RW, CD**: Study correlation between laser diagnostics and mean dose
    111 - **CD, TP**: Evaluate LET in the cells with the RCF in front
     144- **CD, TP**: Refine evaluations of LET in the cells with the RCF in front
     145- **CD**: Predict the cell dish dose from the RCF in front of it
     146- **PH**: Investigate how to get the light out of the vacuum chamber
     147- Investigate how a phosphor sheet could be incorporated into the design
     148- Acquire a glass sheet
    112149
    113 Improve Bio results
    114 - **EM**: Inform Roshine when to defrost Glioma cell line
     150Bio Next Steps
     151- **EM**: Write up a biology plan
     152 - Decide whether to use Marie's lab and whether to stick with HeLa
    115153- **Unassigned**: Obtain an inverted microscope
    116154
    117 Long-term
     155Improve the spatial variation
    118156- **CD, JMcG**: Investigate how to achieve uniformity without a scatterer in place
    119 - **KL, CW**: Cost 4 or 6 quad, and chicane systems
    120157
    121158----