Commissioning of a MOSFET in-vivo patient dose verification system

Full text: TLD dosimetry has long been used for in-vivo measurements in estimating absorbed dose to critical structures on patients. Preparing TLDs for measurement, and then obtaining the results is a time consuming process taking many hours. The Thomson-Neilson 'MOSFET 20' (Metal Oxide Semiconducting Field Effect Transistor) dose assessment system, allows for in-vivo measurements (preparation and results) within minutes. Before being used clinically for dose verification, the MOSFETs were tested against the manufacturer's technical specifications, and compared with results from TLDs measured under controlled experiments and patient measurements. Standard sensitivity MOSFETs (TN-502RD) were used with the bias supply set to High sensitivity range. MOSFETs were tested for linearity (5-100cGy) and their calibration factors obtained for all energies (6MV, 18MV, 6MeV, 12MeV, 16MeV, 20MeV) using the method described by Ramani. MOSFETs and TLDs were exposed to a 6MV beam for 50MU at various depths (RW3 solid water phantom) and field sizes and compared to results taken with an ion chamber. Measurements using both systems were also taken at beam edge and 5mm and 10mm out of the field. Eleven patients, who had lens dose assessment requests were measured with both TLDs and MOSFETs and a paired t-test was performed on the results. On two patients, multiple (nine and four) MOSFET measurements were taken and the range of results compared to the range obtained from the TLDs. MOSFET linearity obtained co-efficients of R2 ≥ 0.996 for all energies, this compared to R2 ≥ 0.996 recorded by both Ramani and Chaung. The y-intercept values varied from 0 to -2.0mV. Greatest variation between calibration factors, measured for each energy, was 7.5%, this is substantially greater than 3.8% quoted by the manufacturer. For the measurements taken at varying depths and field sizes both TLDs and MOSFETs agreed with the ion chamber results ±IcGy. Measurements taken at beam edge varied ±6cGy. 5mm and 10mm out of the field MOSFETs and TLDs recorded the same dose. The paired t-test performed on the data taken from 11 separate lens measurements did not show any significant difference between the means at the 95% level of significance (left lens p=0.13, right lens 0.67). The negative y-intercept suggests that they should be added to the raw reading before being converted to dose, this is important when measuring in low dose regions. The problem with a larger than expected variation between energies, can easily be overcome by using individual calibration factors for each energy. For central beam and scatter doses results between TLDs and MOSFETs varied by 1cGy. The greatest variation of 6cGy occurred at beam edge where the dose gradient is greatest. This can be attributed to TLD chips measuring over a 9mm2 area while a MOSFET measures over 0.04mm2. Multiple readings per patient are now being performed relying on slight positional changes to provide a larger area for the MOSFETs to record 'point' measurements. After a careful commissioning procedure, the MOSFET 20 dose verification system provides reliable, consistent and accurate measurements equal to that of a TLD system. However, the MOSFETs can be prepared for use within seconds, and give results of total dose within minutes, and can be re-used instantaneously making them a far more efficient method of patient dose verification. Copyright (2004) Australasian College of Physical Scientists and Engineers in Medicine