The installation of Sectra DoseTrack is fairly straightforward from the user end,
involving the configuration of sending RDSR and MPPS metadata from an X-ray unit to be connected to the service.
During the installation process some assistance from certified service engineers is required due to restrictions in X-ray equipment software.
The following types of modalities,
of varying models,
from General Electric,
Siemens and Philips,
are connected to the service in VLL:
- 8 Computed Tomography (CT) units,
including SPECT/CT and PET/CT
- 10 Fluoroscopy units
- 12 Conventional X-ray units
- 4 Mammography units
Furthermore,
user end configuration of the service is also required for correct sorting of DICOM metadata into various categories,
e.g.
hospitals,
clinics,
and examination types.
The installation,
configuration,
and validation process required approximately 2 months.
The web-based user interface of Sectra DoseTrack offers a comprehensible overview of different imaging modalities.
Each modality view offers a variety of displays visualising different parameters,
well suited for follow-up of statistics that may be used for evaluating the need of work with justification and optimization,
for example:
- Plots of statistical radiation exposure metrics,
e.g.
median Dose Length Product (DLP) for comparison of different CT units and benchmarking against a Diagnostic Reference Level (DRL) for a given examination type to identify need of optimization (Fig.
1).
- Scatter plots of radiation exposure metrics,
e.g.
for studying the spread of radiation exposure to individual patients for a given type of examination to do individual follow-up of outliers (Fig.
2).
- Plots of radiation exposure of patients stratified for age,
e.g.
for follow-up of justification for different examinations and indications coupled with average radiation exposure (Fig.
3).
In VLL,
the processes of optimization and justification are continuously discussed between radiologists,
medical physicists and radiographers in formal modality specific forums.
This collaborative work has been performed regularly for more than a decade,
and the introduction of exposure monitoring has furthered the possibilities to make a strong positive impact on the day-to-day operation in radiology.
As seen in Fig.
1,
Sectra DoseTrack offers a quick and easy access to statistics on exposure data for making follow-up and identifying the need of optimization of exposure levels from different X-ray units for the same type of examination for the typical patient.
When working with statistical metrics from radiation exposure monitoring,
comparison with DRLs and other healthcare providers also offer possibilities of improving the optimization process.
On a detailed level,
individual patient examinations may also be reviewed to identify outliers and analyse the effect of examination protocol settings for the non-typical patient,
as shown in Fig.
2.
It is interesting to note that the majority of high radiation exposure outliers are found for the university hospital in Umeå,
as compared to the smaller hospitals in Skellefteå and Lycksele.
The university hospital take care of most of the highly specialized patient care,
resulting in examinations with high demands on image quality,
multi-phase examinations,
as well as post mortem examinations,
which all by design yield a higher radiation exposure compared to a typical CT thorax examination.
Furthermore,
as shown in Fig.
3,
the typical patient can be identified for a given examination or modality in regard to age,
which is a powerful tool when dealing with questions of justification and risk to a certain patient group.
From Fig.
3,
which contains data from all CT examinations in VLL for a period of 14 months,
it can be seen that the typical patient is approximately 70 years old and that the dose exposure is considerably lower for patients < 10 years of age.
This is important information when discussing justification and risk associated with CT examinations on a general level together with statistics on radiation exposure.
A small number of multi-phase thorax-abdomen examinations for a certain diagnostic question,
performed at the university hospital in Umeå that by protocol design yield high radiation exposure are the source of the peak in median DLP for young teenagers (14 – 16 years of age) seen in Fig.
3.
Compared to adults,
very few examinations are performed on children in VLL,
and thus a small number of high radiation exposure examinations can cause large fluctuations in a median exposure metric.
The information can also be used to identify the frequency of examinations performed at each age,
e.g.
the peak for young adults around 20-years of age,
which has been identified as predominantly trauma cases.
On a more detailed level,
it is possible to do follow-up and evaluation of justification by employing examination accession numbers and manually review each individual case and indication in PACS/RIS.
The X-ray units for Coronary Angiography (CA) and Percutaneous Coronary Intervention (PCI) were upgraded from Philips Allura Xper (installed in 2012) to Philips Allura Clarity in March 2015.
One aim with the upgrade was to achieve a further reduction of patient as well as staff radiation exposure.
The Philips ClarityIQ software focuses on the reduction of noise and motion blur,
which allows for reduction in radiation exposure.
Sectra DoseTrack allowed easy follow-up of the results in radiation exposure change following this investment,
as shown in Fig.
4.
A substantial reduction in radiation exposure can be seen after installation,
clinical training and optimisation.
The catchment area of VLL is large and contains a wide variety of X-ray units for conventional radiology found in hospitals as well as in small healthcare units.
There are very few radiologists and radiographers who have the competence to use all the various systems,
in particular those located in small healthcare units,
creating a challenge in upholding an acceptable level of optimization.
When Sectra DoseTrack was installed,
the different examination types in conventional radiology were compared in a macroscopic manner,
as shown in Fig.
1.
For conventional X-ray examinations of the shoulder,
a discrepancy was found as relatively high levels of radiation exposure were reported from one specific X-ray unit,
as shown in Fig.
5.
At the time,
there was another X-ray unit of the same vendor and model in VLL,
which had satisfying image quality and comparatively lower radiation exposure levels.
When this difference was identified,
the examination protocols were quickly harmonized.
A similar case for conventional X-ray examinations of the hip was discovered for the same X-ray units,
as shown in Fig.
6.
Sectra DoseTrack condenses all RDSR metadata from interventional radiology into a form containing accumulated radiation exposure levels for quick and easy comparisons between patients,
X-ray units and interventional procedures.
However,
the RDSR metadata from modern X-ray fluoroscopy units contains more detailed information from interventional procedures,
which may be accessed for local administrators of the Sectra DoseTrack database.
In VLL,
the complete set of RDSR metadata for some interventional examinations have been extracted for detailed study of the radiation exposures,
as shown in Fig.
7.
Each irradiation event (fluoroscopy and stationary acquisitions/spot images) is specified with exposure parameters such as projection angle,
collimation,
air kerma in the interventional reference point (IRP),
KAP etc.
that can be applied to manufacturer information on (or measured) radiation scatter levels.
In August 2015 an anaesthesiologist without a radiation protection apron was preparing and monitoring a patient in an interventional radiology suite,
unaware that the radiologist had initiated fluoroscopy.
After 15 minutes,
the anaesthesiologist was made aware of the situation and left the interventional suite.
As follow-up of occupational exposure in this situation information on KAP for each irradiation event (e.g.
as shown in Fig.
7),
vendor provided scatter levels and suite-placement of the anaesthesiologist were employed.
It could be concluded that the radiation exposure from 15 minutes of unprotected work in the suite was similar to the typical daily radiation exposure for interventional radiologists using lead aprons when comparing to routine occupational dosimeter measurements.
One of the intended uses of large-scale radiation exposure monitoring is to fulfil legal requirements on reporting statistics to radiation safety authorities.
However,
there are at present several challenges associated with using Sectra DoseTrack and other similar IT-solutions to comply with legal requirements.
The lack of patient height and weight data in RIS/PACS,
which prevents using radiation exposure monitoring IT-solution for comparing the exposure in clinical radiology with a standard patient according to the formalism of DRLs,
is one of the major drawbacks.
The Swedish Radiation Safety Authority is (as of yet) reluctant to use big data statistics on the typical patient per examination type,
i.e.
the actual average patient in VLL,
for national follow-up of radiation exposure.
A further legal requirement in Sweden is the determination and detailed follow-up of X-ray radiation exposure from fluoroscopic procedures,
in the form of fluoroscopy time,
on an individual operator level.
It is relatively straightforward to access such information for ordinary X-ray fluoroscopy labs from radiation exposure monitoring.
However,
mobile C-arms employed in surgery and orthopaedics without LAN connection cannot use RIS for booking of patients,
and cannot send any meaningful metadata to Sectra DoseTrack for radiation exposure monitoring.
This is not only a problem associated with reporting statistics to radiation safety authorities.
The services employing mobile C-arms are commonly in need of expert assistance with optimization due to inexperience with X-ray equipment.
In VLL,
radiation exposure monitoring for mobile C-arms has been solved by implementation of a separate database.
This data is automatically evaluated and presented on a VLL internal web page.
Integration with aggregated metadata from Sectra DoseTrack has been prepared but has not yet been implemented in this system due to the lack of DICOM conformance from one manufacturer leading to incomplete data.
Since the installation of Sectra DoseTrack in VLL several gaps have been identified in available DICOM metadata for the installed base of various X-ray units.
The information gaps are of two distinct categories,
either they relate to pertinent information not (yet) implemented in the DICOM standard,
or that the manufacturers of X-ray units do not properly handle the metadata,
i.e.
a lack of DICOM conformance.
The amount and detail of information incorporated in DICOM metadata differs depending on manufacturer and modality.
In several cases,
manufacturers place important information in private DICOM fields instead of the fields specified in the DICOM standard.
A central example on information that should be incorporated in the DICOM standard is a patient size metric in the form of attenuation properties.
For CT applications,
this metric could be the water equivalent diameter (WED)3 that can be used for Size-Specific Dose Estimate (SSDE)4.
Some manufacturers already report such patient size metrics in private DICOM tags,
and this information should be available on all X-ray units employing automatic exposure control (AEC) and tube current modulation (TCM).
There are alternatives to commercial IT-solutions for radiation exposure monitoring.
One of these is OpenREM5,
which has been adopted in VLL for further development and refinement of radiation exposure monitoring.
Employing open source software puts a higher demand on the competence available at the user site.
However,
in some cases,
e.g.
for modern X-ray fluoroscopy and CT units,
more detailed information is available to the user without having to manually extract and parse the RDSR metadata a second time.
An IT-solution like OpenREM also makes more advanced,
customized,
and automatic analysis of combined DICOM image- and metadata possible.