The influence of dose reductions on diagnostic quality using a series of high-resolution ultralow-dose computed tomography (CT) scans for computer-assisted planning and surgery including the most recent iterative reconstruction algorithms was evaluated and compared with the fracture detectability of a standard cranial emergency protocol. A human cadaver head including the mandible was artificially prepared with midfacial and orbital fractures and scanned using a 64-multislice CT scanner. The CT dose index volume (CTDI vol ) and effective doses were calculated using application software. Noise was evaluated as the standard deviation in Hounsfield units within an identical region of interest in the posterior fossa. Diagnostic quality was assessed by consensus reading of a craniomaxillofacial surgeon and radiologist. Compared with the emergency protocol at CTDI vol 35.3 mGy and effective dose 3.6 mSv, low-dose protocols down to CTDI vol 1.0 mGy and 0.1 mSv (97% dose reduction) may be sufficient for the diagnosis of dislocated craniofacial fractures. Non-dislocated fractures may be detected at CTDI vol 2.6 mGy and 0.3 mSv (93% dose reduction). Adaptive statistical iterative reconstruction (ASIR) 50 and 100 reduced average noise by 30% and 56%, and model-based iterative reconstruction (MBIR) by 93%. However, the detection rate of fractures could not be improved due to smoothing effects.
Complex craniofacial midface and orbital fractures are frequently sustained during sports activities and predominantly occur in teenagers and young and middle-aged adults. Modern emergency imaging algorithms may include a computed tomography (CT) scan of the brain, CT angiography of the head and neck area, and a full body contrast-enhanced CT to evaluate life-threatening intracranial bleeding, vertebral fractures, blunt vascular trauma, or thoracic and abdominal injury. There is a high requirement for functional and aesthetic success in the surgical treatment of craniofacial fractures. The precise alignment of displaced bone fragments when the bony edges are difficult to identify, orbital floor reconstruction with correction of enophthalmos combined with double vision and reduced eye mobility, the treatment of bilateral midface fractures, and the removal of foreign bodies can be challenging. For such advanced surgery, CT-based computer-assisted planning, including mirror imaging, fabrication of rapid prototyping models, customized surgical plate modelling, and computer-guided surgery, can be very helpful. However, the associated exposure to ionizing radiation has become a major concern as a source of radiation-induced cataract and because of the increased risk of lifetime associated salivary, thyroid, and brain cancer.
It is not commonly known that today’s CT scanners enable high-resolution protocols at doses that are well within the levels of a cone beam CT (CBCT). Unfortunately, a reduction in radiation dose increases noise, which may influence the diagnosis of fractures. More recent iterative image reconstruction algorithms, such as adaptive statistical iterative reconstruction (ASIR) and model-based iterative reconstruction (MBIR), can significantly reduce image noise compared with the standard filtered backward projection (FBP), and have a promising potential to improve the quality of low-dose images.
The aim of the present study was to evaluate the influence of dose reductions on diagnostic image quality of fractures of the midfacial bone and orbit using a series of ultralow-dose CT scans including the most recent iterative reconstruction algorithms and to compare the fracture detectability with a standard cranial emergency protocol.
Materials and methods
One cadaver head with intact soft tissues was obtained from the Department of Anatomy. Following the legal and ethical framework, bodies are donated by people who have given their informed consent for their use for scientific and educational purposes prior to death. The cadaver was preserved using an arterial injection of an alcohol–glycerine solution and immersion in phenolic acid in water for 1–3 months. This head was artificially prepared with midface fractures in the anatomy theatre.
The entire cadaver head including the mandible was scanned using a 64-multislice CT scanner (Discovery CT750 HD; GE Healthcare, Vienna, Austria). The following protocols were used: (1) the standard emergency protocol of the head used in our department, (2) a high resolution reference protocol for navigated surgery, (3) a sinusitis protocol, and (4) a series of four ultralow-dose protocols (I–IV). The emergency protocol used 120 kV, 700 mAs reference upper level in dose modulation mode, and ASIR-50. The reference, sinusitis, and each series of ultralow-dose protocols used fixed kV and mA levels, without dose modulation. In addition to FBP, the following iterative reconstructions were provided: ASIR-50, ASIR-100, and MBIR. The ASIR percentage can be selected across a spectrum of 0–100%, where 0% means all FBPs and 100% means all ASIRs. Protocol details can be found in Table 1 .
|Protocol||kV||mA||Care dose||FOV (mm)||Scan length (mm)||Collimation||Pitch||Rotation time (s)||Slice thickness (mm)||Reconstruction increment (mm)||Pictures|
|Emergency||120||314–488||Yes||224||240||20 × 0.625||1||0.6||0.625||0.625||384|
|Reference||120||100||No||226||230||20 × 0.625||0.5||1||0.625||0.625||373|
|Sinusitis||100||35||No||226||230||20 × 0.625||0.5||1||0.625||0.625||373|
|Low dose I||100||35||No||226||230||20 × 0.625||0.5||0.5||0.625||0.625||373|
|Low dose II||80||40||No||226||230||20 × 0.625||0.5||0.5||0.625||0.625||373|
|Low dose III||80||15||No||226||230||20 × 0.625||0.5||0.5||0.625||0.625||373|
|Low dose IV||80||10||No||226||230||20 × 0.625||0.5||0.4||0.625||0.625||373|
The images were exported in Digital Imaging and Communications in Medicine (DICOM) format using the IMPAX EE (Agfa HealthCare, Bonn, Germany) Picture Archiving and Communication System (PACS).
The computed tomography dose index volume (CTDI vol ) and dose length product (DLP) were recorded from the DICOM tags. In addition, CT-Expo v.2.1 (Medical University Hannover, Germany), an MS Excel application for assessing the radiation doses delivered to patients undergoing CT examinations based on the scan parameters of the scanner type used, was used to calculate CTDI weighted (CTDI w ), CTDI vol , DLP, and effective doses (International Commission of Radiation Protection – ICRP 103). In addition, effective doses were calculated for children by virtually assuming a smaller scan length of 18–19 cm (the true scan length of the adult study head including the mandible was 22–23 cm).
Analysis of diagnostic image quality
The emergency protocol was read by two readers, a consultant craniomaxillofacial surgeon (DDT) and a consultant head and neck radiologist (GW). The multiplanar reconstruction features of the IMPAX EE application software (Agfa HealthCare) and high-resolution diagnostic colour LCD monitors (Totoku CCL254i, Totoku Europe; Rein Medical GmbH, Willich-Anrath, Germany) were used. W/L was set to 4000.00/500.00. The following craniofacial fractures were identified: lateral wall of the right orbit, lateral wall of the left orbit, floor of the right orbit, floor of the left orbit, anterior wall of the left maxillary sinus, lateral wall of the left maxillary sinus, posterior wall of the left maxillary sinus at the pterygopalatine fossa, left zygoma, left zygomatic arch.
All other CT examinations were blinded and the detectability of the fractures defined above was assessed for each combination of protocol and reconstruction technique. A positive result was only given when the fracture could be clearly diagnosed. The images were evaluated by the two readers in consensus, as would occur in the clinical routine in the emergency setting.
Analysis of image noise
Image noise was measured by one observer (GW) as the standard deviation (SD) in Hounsfield units (HU) within an identical region of interest (ROI) of 1500 mm 2 within a homogeneous region in the posterior fossa, using the same axial slice position for all images.
The CTDI vol and DLP calculations of the CT-Expo software correlated well with the data from the DICOM tags. All calculations can be seen in Table 2 .
CTDI vol (mGy)
|CT-Expo CTDI w
|CT-Expo CTDI vol
Effective dose, adults
Effective dose, children
|Low dose I||4.2||106.2||2.0||4.0||96.0||0.4||0.3|
|Low dose II||2.6||66.9||1.3||2.6||62.0||0.3||0.2|
|Low dose III||1.0||25.1||0.5||1.0||23.0||0.1||0.1|
|Low dose IV||0.5||13.4||0.3||0.6||15.0||0.1||Below calculable|
The emergency and navigation reference protocols using CT-Expo had the highest radiation doses, with CTDI vol of 35.3 and 34.7 mGy, respectively. The sinusitis and subsequent ultralow-dose protocols showed 8.0, 4.0, 2.6, 1.0, and 0.6 mGy.
The effective dose for adults of the emergency and navigation reference protocols showed 3.6 and 3.5 mSv, respectively. The sinusitis and ultralow-dose protocols had 0.8, 0.4, 0.3, 0.1, and 0.1 mSv.
Assuming a scan length of 18–19 cm and otherwise unchanged scan parameters, the calculated effective doses for children were 3.0 and 2.5 mSv for the emergency and reference protocols, respectively. The sinusitis and ultralow-dose protocols showed 0.9, 0.3, 0.2, and 0.1 mSv; the effective dose for children of the ultralow-dose protocol IV was below the calculable range.
Diagnostic image quality
Dislocated fractures were detected in all protocols (see Fig. 1 , left column). The tiny and only minimally dislocated fracture of the lateral wall of the right orbit was clearly detected in all protocols except ultralow-dose protocol IV (see Fig. 2 ). The very tiny fractures of the floor of the right orbit and posterior wall of the left maxillary sinus could not be clearly detected by ultralow-dose protocols III and IV (see Fig. 1 , right column). Compared with FBP, the use of ASIR-50, ASIR-100, and MBIR could not improve the detection rate of fractures (see Fig. 2 ).