Patient-related quality assurance with different combinations of treatment planning systems, techniques, and machines : A multi-institutional survey

Risk analysis for radiotherapy at the Universitätsklinikum Erlangen

PURPOSE

Risk analysis is required by various laws and regulations in Germany and has an impact on each department of a large clinic. We provide an overview of the relevant laws and regulations in Germany and present the technical and organizational experience of introducing risk analysis in the Department of Radiation Oncology at the Universitätsklinikum Erlangen.

METHODS

Risk analysis was performed with an in-house developed extension of our intranet platform and ticketing system. Risks were classified according to occurrence and severity, each on a 5-level scale resulting into a risk matrix. An interdisciplinary team of six experienced members formed the core meeting weekly.

RESULTS

A total of 38 risks and 50 measures have been identified in 41 1h-meetings corresponding to approx. 260 working hours. Risk was distributed 8/20/13 to the categories critical (n=8), monitoring (n=20), and conditionally acceptable (n=13). Risk analysis has been evaluated before and after introducing measures.

CONCLUSION

The risk analysis method introduced has been successfully used in routine operations for over a year. Risk analysis takes time and effort. However, because experts from different disciplines meet each other every week, the overall workflow of the radiation oncology department can be improved efficiently and continuously.

Sensitivity of radio-photoluminescence glass dosimeters to accumulated doses

Background

This study investigated the effect of accumulated doses on radio-photoluminescence glass dosimeters (RPLGDs) from measurements involving mega-voltage photons.

Methods

Forty-five commercially available RPLGDs were irradiated to estimate their dose responses. Photon beams of 6, 10, and 15 MV were irradiated onto the RPLGDs inside a phantom, which were divided into five groups with different doses and energies. Groups 1 and 2 were irradiated at 1, 5, 10, 50, and 100 Gy in a sequential manner; Group 3 was irradiated 10 times with a dose of 10 Gy; and Groups 4 and 5 followed the same method as that of Group 3, but with doses of 50 Gy and 100 Gy, respectively. Each device was subjected to a measurement reading procedure each time irradiation.

Results

For the annealed Group 1, RPLGD exhibited a linearity response with variance within 5%. For the non-annealed Group 2, readings demonstrated hyperlinearity at 6 MV and 10 MV, and linearity at 15 MV. Following the 100 Gy irradiation, the readings for Group 2 were 118.7 ± 1.9%, 112.2 ± 2.7%, and 101.5 ± 2.3% at 6, 10, and 15 MV, respectively. For Groups 3, 4, and 5, the responsiveness of the RPLGDs gradually decreased as the number of repeated irradiations increased. The percentage readings for the 10th beam irradiation with respect to the readings for the primary beam irradiation were 84.6 ± 1.9%, 87.5 ± 2.4%, and 93.0 ± 3.0% at 6 MV, 10 MV, and 15 MV, respectively.

Conclusions

The non-annealed RPLGD response to dose was hyperlinear for the 6 MV and 10 MV photon beams but not for the 15 MV photon beam. Additionally, the annealed RPLGD exhibited a fading phenomenon when the measurement was repeated several times and demonstrated a relatively large fading effect at low energies than at high energies.

Implementing a new scale for failure mode and effects analysis (FMEA) for risk analysis in a radiation oncology department

PURPOSE

Patients and staffs are endangered by different failure modes during clinical routine in radiation oncology and risks are difficult to stratify. We implemented the method of failure mode and effects analysis (FMEA) via questionnaires in our institution and introduced an adapted scale applicable for radiation oncology.

METHODS

Failure modes in physical treatment planning and daily routine were detected and stratified by ranking occurrence, severity, and detectability in a questionnaire. Multiplication of these values offers the risk priority number (RPN). We implemented an ordinal rating scale (ORS) as a combination of earlier published scales from the literature. This scale was optimized for German radiation oncology. We compared RPN using this ORS versus use of a rather subjective visual analogue rating scale (VRS).

RESULTS

Mean RPN using ORS was 62.3 vs. 67.5 using VRS (p = 0.7). Use of ORS led to improved completeness of questionnaires (91 vs. 79%) and stronger agreement among the experts, especially concerning failure modes during radiation routine. The majority of interviewed experts found the analysis by using the ORS easier and expected a saving of time as well as higher intra- and interobserver reliability.

CONCLUSION

The introduced rating scale together with a questionnaire survey provides merit for conducting FMEA in radiation oncology as results are comparable to the use of VRS and the process is facilitated.

A multi-institutional initiative on patient-related quality assurance: Independent computational dose verification of fluence-modulated treatment techniques

PURPOSE

This multi-institutional study investigates whether computational verification of fluence-modulated treatment plans using independent software with its own Strahlerkopfmodel is an appropriate method for patient-related quality assurance (PRQA) in the context of various combinations of linear accelerators (linacs), treatment techniques and treatment planning systems (TPS).

MATERIALS AND METHODS

The PRQA-software's (Mobius3D) recalculations of 9 institutions' treatment plans were analyzed for a horseshoe-shaped planning target volume (PTV) inside a phantom. The recomputed dose distributions were compared to a) the dose distributions as calculated by all TPS's and b) the measured dose distributions, which were acquired using the same independent measuring system for all institutions. Furthermore, dose volume histograms were examined. The penumbra deviations and mean gamma values were quantified using Verisoft (PTW). Additionally, workflow requirements for computational verification were discussed.

RESULTS

Mobius3D is compatible with all examined TPSs, treatment techniques and linacs. The mean PTV dose differences (Mobius3D-TPS, <3.0%) and 3D gamma passing rates (>95.0%) led to a positive plan acceptance result in all cases. These results are similar to the outcome of the dosimetric measurements with one exception. The mean gamma values (<0.5) show a good agreement between Mobius3D and the TPS dose distributions.

CONCLUSION

Using Mobius3D was proven to be an appropriate computational PRQA method for the tested combinations of linacs, treatment techniques and TPS's. The clinical use of Mobius3D has to be complemented with regular dosimetric measurements and thorough linac and TPS QA. Mobius3D's computational verification reduced measurement effort and personnel needs in comparison to dosimetric verifications.