Logo of the Physikalisch-Technische Bundesanstalt

Alanine dosimetry - from metrology to therapy?


A compact ESR spectrometer intended for use in industrial alanine dosimetry (irradiations of medical products, etc.), the so-called e-scan, is available from Bruker Biospin GmbH. PTB purchased such a device in autumn 2006 in order to investigate its possible use in radiation therapy dosimetry.

The magnetic field of the e-scan is produced mainly by permanent magnets, different from the EMX spectrometer which is used for the secondary standard measurement system. Consequently, the measurement range is restricted to g ≈ 2, but with the advantage that the required space is small and no water-cooling is necessary. The e-scan takes advantage of a built-in reference sample; the ESR amplitude of the substance under investigation is measured relative to that of the reference. The method had been adapted for the construction of PTB’s secondary standard, with the difference that the ESR amplitude is determined via a fitting method elaborated at PTB, which again is based on developments by the NPL, instead of the common peak-to-peak amplitude determination. The quantity of the reference substance contained in the available probe holders for the e-scan is adapted to requirements of industrial dosimetry, i.e. the measurement range is 20 Gy to some 104 Gy. At PTB, a modified holder with a smaller quantity of the reference substance is employed. With a thus accessible measurement range of 10 Gy to 50 Gy, the so-called therapeutic range of 1-10 Gy is just reached.

Measurement parameters were determined experimentally, which represent a compromise between an optimum attainable uncertainty and the (temporal) requirements of clinical work. The latter were determined in cooperation with the University Hospital Göttingen and the Ziekenhuis Middelheim in Antwerp, Belgium. The uncertainty attainable with the instrument and the given parameters were determined using the methods outlined in [1] and adapted to the e-scan. Alanine pellets of two suppliers (GammaService, Germany, and Harwell, UK) were compared.

The uncertainty which is attained using the methods developed at PTB is 40% smaller than the uncertainty obtained using the factory built-in peak-to-peak amplitude determination. The uncertainty is mainly due to the scatter of the ESR amplitude values, all other uncertainty contributions are, with the e-scan, almost negligible, which is different for the EMX spectrometer (see [1]). The nearly dose-independent reproducibility of the dose-normalised amplitude AD (see [1]) in the stated measurement range is 120 mGy for GammaService and 150 mGy for Harwell pellets. For the EMX, this value is much lower (20 mGy). According to the manufacturer, the ratio of 5-7 for the two spectrometers reflects the ratio of the attainable signal-to-noise ratios.

The relative uncertainty ur(D) of the determined dose when using 4 probes per dose and daily calibration lines is 1% for 10 Gy and 0.5% for D > 20 Gy for GammaService pellets; the corresponding values for Harwell are somewhat higher. This is valid in case the calibration line is set up using 10, 20, 30, 40 and 50 Gy, with four pellets per dose each. If the calibration curve is not determined on the same day as the signals for the test probes, the uncertainty is increased by up to 50% at the lower end of the measurement range, whereas at the upper end almost no change is observed. The uncertainties do not include the uncertainty of the primary standard.

Even without modifications of the hardware, the e-scan appears to be adequate for dose-measurement tasks in radiation therapy. Although in-vivo measurements for single irradiation fractions are certainly out of the question (D ≈ 2 Gy), the spectrometer may well be used in quality assurance. It is strongly recommended that the measurement methods developed at PTB be used for this purpose. The measurement time required for setting up the (daily) calibration curve is 1 h, for each dose value to be measured another 8 min are required. If only one probe per dose instead of 4 is used, the uncertainty is doubled.

Some constraints concerning the measurement procedure apply: to obtain the lowest possible uncertainties the user has to adhere to a certain sequence for the measurement of the test- and calibration probes. It is especially important to measure test- and calibration probes almost at the same time. Investigations of the temporal behaviour of the amplitudes within one measurement day showed that a drift may increase the relative uncertainties by 0.5% to 1% if, for example, the test probes are measured 3 h later than the calibration probes. The observed drift of the amplitude can be related to changes in the so-called base-line-drift. In spite of all efforts, the influence of this drift could not be eliminated by data analysis alone. Progress is expected from an active temperature stabilisation of the resonator. A further increase of the signal-to-noise ratio is also highly desirable, but this can only be achieved through development work by the manufacturer.

In spite of some reservations, use of alanine/ESR dosimetry in radiation therapy may become an alternative to thermoluminescence dosimetry (TLD). Considerable advantages are the non-destructive read-out, the simple handling (no pre- and postprocessing of the probes is required) and the very small dependence of the response on the radiation quality. The slightly longer times which are required for the irradiation of the alanine probes is no real counter-argument. With typical dose rates of 2-3 Gy/min for present-day clinical accelerators, the irradiation time itself is small compared to the preparation time (set-up of phantoms and probe holders, etc.) which is the same for both dose measurement systems. The financial investment necessary for implementing ESR/alanine with the e-scan should also be comparable to that of a TLD measurement system.


  1. Anton, M.:
    Uncertainties in alanine/ESR dosimetry at PTB,
    Phys. Med. Biol. 51 pp 5419-5440 (2006)