Umeå universitets logga

Diagnostic reference levels (DRL) is a tool used by radiation safety authorities to monitor the use of radiation in health care. It can also be used as an optimization tool in radiology to reduce the radiation dose to the patient. In Sweden it is The Swedish Radiation Safety Authority (SSM) that decides the national DRL. At regular intervals, the hospitals submit their local DRL to SSM to verify that they are within the limits of the national DRL. In Region Jämtland Härjedalen, the collection of dose values is done manually which is time consuming, ineffectively and statistically limited. The aim with this master thesis is to automate the collection, analysis and presentation of the local DRL.

The automation of the collection of data for the local DRL was done by extracting patients height and weight data from Sectra Data Warehouse (SDWH) and adding it to REMbox (Dicom Port) via the REST-API that the system has. From REMbox, data for the local DRL reports could be extracted and the reports were created in Python. For the analysis the weight range of adult patients who have undergone a CT-scan in Region Jämtland Härjedalen in 2024 was compared to the weight range that SSM uses for the national DRL. The radiation dose to the patient, Computed Tomography Dose Index (CTDIvol), as a function of Body Mass Index (BMI) was examined to see how the radiation dose was affected by BMI. The number of detected photons is given by the Beer-Lambert relation which means that the relationship between radiation dose and BMI should be exponential.

The results of this master thesis show that 37% of the patients that have undergone a CT-scan were within the weight range that SSM uses for the national DRL. A method has also been developed for how to clean the data and get the exponential relationship between CTDIvol and BMI. Finally, an automation of the collection of data for the local DRL reports has been created.