Hide menu

EPR dosimetry


A measurable biologically important effect of ionizing radiation is the creation of free radicals. In many crystalline organic materials like amino acids and sugars, in formates and dithionates as well as in body materials as keratin in nails and hair and in bone and tooth enamel very long-lived, almost stable radicals are formed during radiation. By means of Electron Paramagnetic Resonance, EPR, spectroscopy it is possible to identify and quantify the number of induced free radicals. The EPR-signal that is a measure of the radical density is proportional to the mean absorbed dose in the sample in a large interval from the lowest detectable dose up to several kGy, and the phenomena is thus well suited for dosimetry. If organic materials are chosen most of them have an atomic composition quite similar to that of tissue or water and therefore have the same energy absorption properties. Another important advantage is that the signal is not disturbed after read-out, an EPR dosimeter can therefore be used for integrating dose determinations and be analyzed between every irradiation.

Dosimetry for radiation therapy

Dosimetry based on EPR spectrometry is since many years a well-established method. However it has only quite recently been applied to radiation therapy together with the introduction of improved techniques with more sensitive spectrometers and above all with the development of more sensitive dosimeter materials. During the last 15 years we have developed a number of promising dosimeter materials out of which lithium formate has shown to be the most suitable material. For clinical applications we have used lithium formate dosimeters for verification of calculated dose distributions both in external and brachy therapy and for testing the whole treatment procedure between different clinics in an audit. Recently we succeeded to decrease the size of the dosimeters keeping the high accuracy by using a modern spectrometer equipped with a super high Q resonator; a Bruker E500 X-band/L-band EPR and EPRI with a SuperX High Power Microwave Bridge installed at SIMARC , IFM.

In radiation therapy the aim is to give as high dose as possible to the tumor and to save the healthy tissue. Before radiation therapy a computer calculation is performed, a dose plan over how the radiation will be distributed over the target and surrounding tissue. This must be verified i e by measuring doses directly on the patient during treatment or as is the case at intensity modulated radiotherapy IMRT before treatment  in a phantom which i dose planned exactly like the patient. We have compared such measurements with ionization chamber in specified points with dose determinations with passive EPR dosimeters of lithium formate and we found that EPR dosimetry is sufficiently accurate and a rapid method, technically easier than ion chamber measurements.

At IMRT many differently shaped radiation fields are used sometimes very narrow where the radiation field consists of a relatively large penumbra region. In these cases the dose calculations are complicated and might be inaccurate. Therefore the linear response over a wide dose range for the lithium formate dosemeters is an advantage for the dosedeterminations over very narrow radiation fields including the penumbra region.

Steep dose gradients are also typical in brachy therapy and the EPR dosimeters have shown to be very useful with high precision and reproducibility when dose distributions are verified in a phantom around a 192Ir source stepped forward along an applicator. Brachyterapi is usually performed with sources with gamma energies in the region 30-400 keV. Photons from these sources are generating electrons with higher LET than those in the calibration fields we normally use, 4 MV accelerator photons. This means that we have to know the LET dependence of the dosimeters. The LET dependence is investigated for X-ray fields with well-known spectra. Monte Carlo simulations have given the conversion factors from air kerma to absorbed dose in the dosimeters for the different radiation fields.

Another important clinical application is the use of lithium formate EPR dosimetry for investigation of the whole treatment procedure and comparison between different radiation therapy clinics, an audit.

For these investigations the lithium formate tablets are prepared to the highest possible homogeneity and carefully calibrated. Temperature dependence and radical stability are investigated and recommendations given for a standardized read-out procedure, which however is changeable depending on measuring purpose. 

The use of heavy charged particles for radiation therapy is increasing and hence also an increasing interest for robust dosimetry systems for these radiation qualities. Normally the EPR dosimeters are calibrated in the same radiation quality as in which they are intended to be used since the dose response is dependent on the LET of the radiation.

For some materials the EPR shape of the EPR signal is depending on the LET and thus offers an opportunity to determine not only the dose but also the LET of the radiation with EPR analysis. We have investigated the LET dependence of potassium dithionate in C6+ and N7+ ion beams at the The Svedberg laboratory in Uppsala.  We irradiated a 5 mm dosimeter tablet with a 30 MeV/u beam so that the Braggpeak was found at a depth of less than 3 mm in potassium dithionate.

Project group: Eva Lund, Emelie Adolfsson Håkan Gustafsson, Sara Olsson, Åsa Carlsson Tedgren, Gudrun Alm Carlsson, Mattias Karlsson.

EPR imaging, EPRI

Visualization of the radical distribution in a sample is possible when the sample is placed in a strong  magnetic gradient field. The method is quite similar to the early MRI techniques but the development has been quite slow because of low radical concentrations in the samples compared to hydrogen nuclei for MRI. The short relaxation times for the electrons also demand extremely strong magnetic field gradients and only recently commercial spectrometers have been available. We have used the Bruker E540 EPR and EPR imaging spectrometer equipped with an E540 GC2X two axis X band gradient coil resonator at IFM.

The dosimeter material potassium dithionate (K2S2O6) is especially interesting since the EPR signal consists of two almost overlapping peaks with relative intensities depending on the ionization density, the LET, of the interacting radiation. In collaboration with professor S Schlick in Detroit we have mapped the dose distribution in tablets of potassium dithionate irradiated with N7+ or C6+ -ions and as far as we know for the first time with EPRI visualized the dose and the LET distributions with a high spatial resolution in the order of 5 μm.

figure 1




Fig 1 EPR imaging with the intensity as a function of the penetration depth for the N7+ ions in one direction and the absorption spectrum in the other one. To the right the spectrum variations (1:st derivative)  with depth are shown in detail. 

There are specific requirements for high spatial resolution when the dose gradients are steep as in the penumbra region of a radiation field or at the interface between tissue and a metal implant and often at brachy therapy. To be able to visualize dose distributions with high resolution the requirements for the dosimeter material is:  high yield for radiation induced radicals and a spectrum with a single narrow signal.  The width of the EPR signal determines the possible spatial resolution for dose distributions.

We have started to visualize a homogenously irradiated potassium dithionate tablet in different positions in the cavity of the EPR spectrometer equipped with gradients, 1.7 T/m  and achieved a resolution of 0.3 mm without further deconvolution. 

An in-house MATLAB code has been used for reconstruction of the EPR images taking the inhomogeneous sensitivity in the cavity into account and possibilities for base-line restoration.

figure 2.1figure 2.2
Fig 2 The EPR image of a 5 mm homogenously irradiated dosimeter placed 5 mm off center in the cavity after direct deconvolution of raw date (to the left) and after base line correction and correction for intensity variations in the cavity (to the right)


The aim of this investigation is to visualize the dose distribution of a 2x2 cm2 radiation field  including the penumbra region by using a stack of 5 mm dosimeters.  For larger radiation fields the penumbra region contributes to a very low percentage but for very small fields < 3x3 cm2 the contributions from the penumbra could stand for more than 10% of the total dose.

Research group: Eva Lund, Håkan Gustafsson and Maria Magnusson ISY.

Retrospective dosimetry based on EPR analysis.

Retrospective dose determinations in tooth enamel.

The tooth enamel is a very good retrospective dosimeter with quite high sensitivity for ionizing radiation and a high stability. The EPR signal after irradiation consists of the CO2-.

radical together with a background signal from the organic material. It is however possible to separate these two components using a deconvolution technique. For retrospective dose determinations normally a tooth must be extracted and carefully prepared before analysis. It is important to mechanically separate the dentine from the enamel. The background signal also includes mechanically induced signals from the preparation and the radiation induced signal might partly consist of signal from dental X-rays and from UV irradiation from the sun or the blue light used for hardening composite fillings.

Till recently a whole extracted tooth was needed for dose determinations but with more sensitive EPR spectrometers and also use of higher microwave frequencies, Q band, a smaller amount of enamel. 5-10 mg, is sufficient. To choose the best part of a tooth for analysis our research aims at detecting the influence of X-rays, UV-light and caries on the radiation induced signal in the enamel.

Retrospective dose determinations in sugars and sweeteners, touch glass in smart phones and finger nails.

We have successfully investigated the radiation induced signal from xylitol and sorbitol and applied this to chewing gum, a material often found in the pockets. We achieved a detection limit of 0.8 Gy and the dose response is linear. An unknown dose can be determined with an uncertainty of 0.17 Gy and the composition of radicals seems to be quite stable from day 4 to day 8 which means that quite accurate determinations can be made within this time period.

Figure 3



Fig 3 EPR spectra of xylitol and V6 irradiated to 50 Gy. The xylitol spectra were measured A = 1 day, B = 4 days, C = 8 days, D = 16 days, E = 36 days and F = 84 days after irradiation, while the dashed V6 spectra were measured Q = 4 days and R = 8 days after irradiation.

Within an inter-laboratory comparison of irradiated touch screens from smartphones we achieved a linear dose response and possibility to determine blind doses above 2 Gy with a detection limit of 1.3 Gy. A conclusion of this EURADOS test including 11 laboratories was that the used protocol was easily transferred to participants making a network of laboratories possible in case of a large emergency event.

figure 4







Fig 4 The signal intensity is determined between the arrows for the irradiation doses 0 – 10 Gy.

Finger and toe nails are easily collected and can be used for dose determinations shortly after exposure. The nails are mainly composed of a-keratine in which fragile disulfide bonds are broken by ionizing radiation and sulphur radicals are created. Unfortunately free radicals are also easily induced by cutting and irradiated cut samples of nails show EPR spectra with three partly overlapping signals; RIS, and mechanically induced MIS1 and MIS2. Our study of preparation and storage of irradiated nails shows that after water treatment the mechanically induced signal varies more, 10%, between individuals compared to variation between different samples from the same individual, 4%. Therefore if only one hand is exposed a background signal could be obtained from the nail samples from the other hand. The mechanically induced signal stays almost stable if the samples are stored in a freezer; -13o. The dose response is found to be nonlinear for water treated nail samples and it is possible to determine doses down to around 2 Gy.


Figure 5

Fig 5 The dose response curve produced from a mix of nails from eight different donors. Here  MIS2 has been subtracted from RIS, MIS1 is decayed

Project group: Eva Lund, Håkan Gustafsson, Axel Israelsson together with masters students.

Eva Lund
Name: Eva Lund
Position: Professor emerita in Medical Radiation Physics
Department: Department of Medical and Health Sciences
Divison: Division of Radiological Sciences


Telephone: +46 13 286851
E-mail: eva.lund@liu.se

Visiting address:
Gula Villan
Entrance 20, floor 11
Campus US

Postal address:
Linköping University
Eva Lund
Gula Villan
Department of Medical and Health Sciences
SE-581 85 Linköping

Page manager: sandra.malmstrom@liu.se
Last updated: Wed Sep 23 10:47:46 CEST 2015