The technologies

Radiation dose monitoring software automatically gathers, stores and analyses information on patients' radiation exposure from medical imaging involving ionising radiation. Between 2012 and 2013, over 22.6 million X‑rays and 4.7 million CT scans were done in the UK (NHS annual imaging and radiodiagnostic activity in England, 2012/13).

This briefing describes 8 software technologies that analyse patient-level radiation doses from different imaging modalities and examination types. Dose information can be collected:

  • directly from the imaging device

  • using the picture archiving and communication system (PACS), a technology that stores and handles medical images and related information or

  • from the radiology information system.

The software uses digital imaging and communications in medicine (DICOM)-standard data sources to gather information relating to the radiation dose. DICOM (National Electrical Manufacturers Association) is an international standard used in the NHS for storing and exchanging medical images and image-related information. The most common DICOM methods used by dose monitoring software for recording and analysing dose data are:

  • Radiation dose structured report (RDSR), a report included in the DICOM dataset which contains various dose-related parameters, for example, the dose length product (DLP) and the volumetric CT dose index (CTDIvol) for CT scans.

  • Modality-performed procedure steps (MPPS), a report generated by the scanner that contains information about an examination including data about the images, length of scan and dose delivered.

  • Dose reports using optical character recognition (OCR) of CT. Each CT machine generates a patient record, which consists of an image containing, among other information, the CTDIvol and DLP values. In newer scanners (produced in 2012 and onwards) or older scanners with newer software, the RDSR has replaced the need to retrieve data in this way.

  • Image file header, which is part of each DICOM dataset that contains general parameters for the specific imaging examination.

The main functions of the software are to:

  • Collect, store and monitor radiation dose data across different medical imaging modalities, regardless of the system manufacturer, hospital or hospital unit, including: CT scanners, interventional radiology systems, cardiovascular systems, mammography, radiography systems and surgical/mobile C-arms.

  • Analyse radiation dose data, including:

    • change in the number of examinations done per year

    • radiation doses for all examinations done

    • a comparison of patient-level data with population-based information on radiation dose

    • set protocol-specific diagnostic reference levels (DRLs), which are pre-specified dose levels based on national or local criteria. DRLs are currently based on people with BMI of 20 to 25.

  • Create reports and automated alerts, including:

    • run reports that review the range of doses given for each type of study and identify examinations with the highest dose

    • when an examination exceeds the DRL, the system can create alert notifications to display in a chart or to be sent by email.

  • Using the dose index data to help find the lowest reasonable radiation dose for acceptable image quality.

Table 1 outlines the main characteristics, differences and similarities of the 8 technologies included in this briefing (adapted from Boos et al. 2016).

Table 1 Summary of included technologies

Technology name (company)

Data acquisition

Installation

User access

Modalities

CE mark*

DOSE

(Qaelum)

RDSR, MPPS, OCR, header

Local

Web

CT, XA, DR, MG, RF, NM, PET

IIb

DoseM

(Infinitt)

RDSR, MPPS, OCR, header

Local

Web

CT, XA, DR, MG, RF, NM, DXA

I

DoseMonitor (PACS Health)

RDSR, MPPS, OCR,

header

Local

Web

CT, XA, DR, MG, RF, NM, DXA

I

DoseTrack

(Sectra)

RDSR, MPPS, OCR

Cloud

Web

CT, XA, DR, MG, RF, PET

I

DoseWatch

(GE Healthcare)

RDSR, MPPS,

OCR,

header

Local

Web

CT, XA, DR, MG, PET

I

DoseWise

(Philips)

RDSR, MPPS, OCR, header

Local

App

CT, XA, DR, MG

I

OpenREM

(OpenREM)

RDSR,

header

Local

Web

CT, XA, RF, DR, MG

Not applicable*

teamplay

(Siemens Healthcare)

RDSR,

OCR,

header

Local and cloud

Web

CT, XA, DR, MG, RF, NM, PET

Not applicable*

Abbreviations: DR, digital radiography; DXA, dual-energy X‑ray absorptiometry; header, DICOM‑Header; MG, mammography; MPPS, modality-performed procedure step; NM, nuclear medicine; OCR, optical character recognition; PET, positron emission tomography; RDSR, radiation dose structured report; RF, radiofluoroscopy; XA, angiography.

* NICE understands that the technology does not need CE marking as a medical device.

Innovations

The use of radiation dose monitoring software may improve the collection, analysis and reporting of radiation dose data compared with current manual or semi-automated methods. The detailed information provided by dose monitoring software allows for the best image quality possible while minimising radiation exposure to the patient.

Dose monitoring software can also be used to alert healthcare professionals to radiation exposure when DRLs are consistently exceeded. Some of the technologies may help facilitate management of protocols, contrast media and staff dose as well as image quality.

The systematic monitoring and analysis of radiation dose data can potentially reduce radiation exposure for people having multiple imaging procedures. It can also help hospitals meet legal and policy requirements. Based on the Medical Exposure Directive 97/43, in some European countries (currently including the UK), radiation protection legislation mandates the recording of individual patient doses (or parameters from which dose can be calculated). Current UK legislation includes Ionising Radiation (Medical Exposure) Regulations 2000 (IRMER) from the Department of Health, which details DRLs and what to do in cases of excessive radiation exposure. Systematic dose monitoring may also help to support quality assurance in terms of meeting directives such as the EU Council Directive 2013/59/EURATOM.

Current guidelines and arrangements

Public Health England currently gathers and collates radiation dose data for common examinations from a sample of UK hospitals through manually compiled databases. The Department of Health's response to COMARE's 16th report on the increased use of CT scans in the UK recommended that more frequent UK dose surveys need to be done. These surveys will provide data to support regular updating of national DRLs, including those specifically for children.

Manual and semi-automatic recording of radiation dose data requires data entry in the radiology information system, a spreadsheet or on paper. This is time consuming and may result in an error rate of up to 6% (Noumeir 2005, Boos et al. 2015).

The 2011 review of the Public Health England report CRCE-013: Doses from CT examinations in the UK specifies that for a national audit on radiation dose data, a healthcare professional (either a radiographer or a physicist) with access to PACS should acquire the data and a data manager (a radiographer or a physicist) should verify the data before transferring it to a spreadsheet or other record.

Population, setting and intended user

The technologies are likely to be used for retrospective analysis by healthcare professionals specialising in radiation protection and with appropriate training. These would most likely be medical physicists, radiographers and radiologists. The technologies would be used in secondary care in the NHS to record and analyse data in the trust.

Radiation dose data can be collected from anyone having medical imaging with ionising radiation.

Costs

Table 2 shows the costs associated with each technology.

Table 2 Costs of radiation dose monitoring software

Technology name (company)

Cost (excluding VAT)

Additional information

DOSE

(Qaelum)

Average £10,000 to £15,000 (per hospital, per year; based on study volume).

DoseM

(Infinitt)

Average £8,000 to £10,000 (per hospital, per year; based on multi-modality, multi-manufacturer data collection and archiving from ionising radiation sources through PACS and extending to modalities which do not support DICOM SR/MPPS).

DoseMonitor (PACS Health)

Average £10,000 to £15,000 (per hospital, per year; based on study volume).

DoseTrack

(Sectra)

Average £10,000 to £15,000 (per hospital, per year; based on 30 modalities per year).

DoseWatch

(GE Healthcare)

Average £10,000 to £20,000 (per hospital, per year).

Operational lease model for connecting all ionising radiation scanning equipment.

DoseWise

(Philips)

£1,300 per modality, plus a server fee of £10,400 per year.

OpenREM

(OpenREM)

Free to download and use.

Available under open-source licence (GPL V3).

teamplay

(Siemens Healthcare)

£7,400 to £20,000 (per trust, per year).

Basic version at no additional cost with Siemens scanners. Costs are per installation on a PACS system, which may be shared across a trust.

Non-Siemens scanners/mixed fleets cost £7,400 per year.

Dose and usage monitoring apps for Siemens scanners.

Abbreviations: PACS, picture archiving and communication system; DICOM SR, digital imaging and communications in medicine, structured reporting; MPPS, modality-performed procedure steps.

Costs of standard care

The main cost associated with manual dose data recording is the clinical time it takes. The cost of a hospital radiographer's time is £35 per hour (Agenda for Change band 5) and a medical physicist's time is £56 per hour (Agenda for Change band 7; Personal Social Services Research Unit [PSSRU] 2016). This includes all remuneration, qualifications, department overheads and capital costs.

The cost depends on whether data are recorded manually (by radiographers filling in forms) or downloaded from the radiology information system.

Specialist commentators estimate that it takes 20 minutes of a radiographer's time for data collection per examination. The data are then transferred to a medical physicist for verification of data entry, review, analysis and report production. This takes at least 1 hour per examination. Combining the radiographer's and medical physicist's time, the estimated average cost is £68 per examination.

If data are instead taken from the radiology information system, little or no radiographer time is needed. However, extra medical physicist input may be needed to understand the data, eliminate outliers, confirm the validity of results and remove zero values before the data can be analysed. One specialist commentator estimated that this is at least 1.5 hours of a medical physicist's time is needed per examination, with an estimated average cost of £84.

Resource consequences

If adopted, the technologies would likely to be used with the available ionising radiation imaging equipment. A mid-sized hospital trust could have an average volume of 100,000 images per year, whereas a large trust may do 250,000 images per year.

These technologies are software packages to be used with current hardware and so no additional facilities or technologies are likely to be needed. However, the hospitals will need to reallocate staff to manage information governance and software compatibility arrangements such that the technologies can be properly installed and used. This may be time consuming. The technologies will also need IT involvement to set up and staff to maintain them.

No published evidence on the resource consequences of adopting the technologies was found, including either economic evaluations or costing studies.