Chapter 14 Image quality

Ken Holmes, Marc Griffiths

CHAPTER CONTENTS

Introduction 161

Signal-to-noise ratio 162

Page 162

Magnification 162

Sharpness 163

Geometry of imaging 164

Radiographic density 164

Contrast 165

Latitude 167

Viewing a radiograph 168

Acceptable image guidelines 169

Common faults with remedy 170

KEY POINTS

Image quality affects the ease of extracting information from an image.

Good image quality will ensure the maximum amount of diagnostic information is gained from the image.

Sharpness is essential as blurring of an image will lower the image quality.

Contrast describes the difference in density between two adjacent structures.

Density is the degree of blackening on the film.

Fault finding is essential to ensure that image quality is optimal.

Understanding the cause of faults will ensure rectification of the cause and improved image quality.

INTRODUCTION

One of the fundamental roles of practitioners in radiography is to produce optimum radiographs for reporting. Image quality is difficult to define because it is subjective in its nature, but an optimum quality image enables the observer to extract information from the image and make an accurate diagnosis. Poor quality images have a poor signal-to-noise ratio and detract from the process of extracting information. However, there are characteristics of a radiograph, which may be evaluated, and enable the practitioner to distinguish the diagnostic quality of the image. These factors include:

• magnification and distortion

• sharpness

• radiographic density

• contrast

• latitude of the film or the resolution of the monitor (VDU).

These factors can be affected during the production and processing of the image (both analogue and digital processing) and once the image has been produced or displayed. It is essential that if a film is produced it is viewed under optimum viewing conditions and must not be held up to the nearest light bulb! Digital images must also be viewed on an appropriate monitor with diagnostic resolution.

SIGNAL-TO-NOISE RATIO

One concept used to define image quality is the signal-to-noise ratio. The signal is the useful information from the patient being imaged and the noise is anything that detracts from accessing the information. Useful information is derived from photoelectric interactions within the patient (absorption by body structures) and noise is derived from scattered photons, because the image receptor does not have the ability to determine their origin. In radiographic images if the signal level is high compared to the noise, structures within the body will be clearly seen, but if the signal level is similar to or less than the noise level then the structure will become obliterated.

Images produced in radiography are often a compromise between obtaining a perfect signal and reducing the noise. The production of images is constrained by a number of factors including the radiation dose level and each component of the imaging chain. For example, a faster film-screen system is utilised when imaging the abdomen than for the extremities and this enables the practitioner not only to produce a diagnostic image with minimal noise but to minimise the radiation dose to sensitive organs.

MAGNIFICATION

When producing a radiographic image all objects in the image are larger than the object being X-rayed. This magnification is due to the geometry of imaging. The ideal situation is to have the object being X-rayed as close as possible to the image receptor, parallel to the X-ray tube and the image receptor, with the radiation beam at right angles to the object. This minimises the magnification of the image but also, more importantly, the magnification of the unsharpness in the image (Figs 14.1 and 14.2).

Figure 14.1 Image with minimal magnification.

Figure 14.2 Image demonstrating magnification – if the body part is not in contact with the cassette, the result is magnification.

If the object being imaged is not parallel to the image receptor it will be magnified; however, each end will be magnified differently and this will produce distortion. This may be elongation or foreshortening of the image (Fig. 14.3).

Figure 14.3 Image demonstrating distortion – if the body part is not parallel with the cassette, the result is distortion. Distortion also occurs when the X-ray beam is angled across the object being X-rayed.

The focus to image receptor (SID; see p. 91) should be as long as possible and the distance between the patient and the image receptor (object–film distance, OFD) as short as possible. For practical reasons the SID is usually 110 cm for techniques on the X-ray table and 180 cm for chest and cervical spine erect film work. The OFD should be as close as possible and ideally the object is in contact with the image receptor. Using a Bucky assembly also causes magnification of the image but the grid is necessary to reduce scatter and improve the contrast of the image. Practically, the mechanism which moves the grid and houses the Bucky is kept as small as possible.

Page 163

SHARPNESS

In order to determine the image quality the image must be sharp. Blurring will reduce the image quality and also reduce the diagnostic quality of the image. The factors affecting sharpness are:

• movement

• the resolution of the imaging system (film-screen combination or VDU)

• the geometry of the imaging system (relationship between the focus, patient and imaging system).

MOVEMENT UNSHARPNESS

Movement unsharpness should be kept to a minimum by careful radiographic technique. Making the patients comfortable, giving them precise instructions on breath holding and planning the procedures all help reduce this. Using short exposure times also helps reduce blur in the image. Exposure times as low as 0.003 seconds for a chest radiograph enable the practitioner to ensure minimum movement, even in paediatric patients.

RESOLUTION OF THE IMAGING SYSTEM

This is covered in more detail in other chapters of the book (see pp. 141, 154). However, it should be noted that the fastest film-screen combination or exposure index compatible with diagnostic quality should be used. When determining the quality required in an image the practitioner must be aware of what structures need to be defined. An image to determine the position of bones in a plaster cast following an orthopaedic reduction needs less resolution than the original image to diagnose the fracture. Ideally the smaller of the foci of the X-ray tube should be used (fine focus) but if this does not enable a short exposure time to be used on a patient likely to move then the practitioner may need to use a broad focus. This is another example of where the practitioner needs to make a decision that is a compromise between the ideal conditions and getting a diagnostic image.

Page 164

GEOMETRY OF IMAGING

The positioning of the patient (geometry) to produce an image has a direct relationship to the quality of that image. The section on magnification (p. 162) outlined the ideal conditions needed to produce radiographic images and dealt with the requirements to reduce magnification (p. 163). It also emphasised that any increase in the size of the object also increases the unsharpness of the image and this will be explored in this section.

Figure 14.4 is a diagrammatic representation of image production and shows that a penumbra (unsharpness) is formed with any image that is produced from a finite source (focal spot). The diagram uses a large distance between the object and image receptor to illustrate the principle of the penumbra. In practice the amount of geometric unsharpness (U_g) in an image is small and may be much less than 1 mm.

Figure 14.4 Diagram to demonstrate the geometry of imaging.

Measurement of the penumbra (U_g) is a straightforward calculation using similar triangles. Figure 14.5 demonstrates the diagrammatic representation of similar triangles and the calculation to determine unsharpness in an image ofa finger due to geometric unsharpness. The values used are typical in radiography and the unsharpness is only 0.005 mm.

Figure 14.5 Diagram to demonstrate geometry unsharpness when taking an X-ray of a finger.

MEASUREMENT OF UNSHARPNESS IN AN IMAGE

The sharpness of the image can be measured using a test tool (resolution) or by viewing the image and determining whether fine structures can be visualised (definition). The resolution can be determined by measuring the ability of the imaging system to resolve the smallest object of the test tool and is expressed in line pairs per millimetre (lp mm⁻¹). This is perhaps an abstract concept to illustrate sharpness and it would be more sensible to determine the definition within the image; for example, can I see the bony trabeculae?

Again, however, definition within the image is subjective and depends on a number of factors including the viewing conditions and the experience of the practitioner viewing the image.

RADIOGRAPHIC DENSITY

Radiographic density is generally applied to hard copy images and can be defined as the amount of overall blackening on the film.

It is an easy concept to understand at each extreme of density. All practitioners will be familiar with an image which has so much blackening that it is impossible to see any detail within the image, making it undiagnostic. Similarly, a film which has insufficient blackening, where the structures cannot be defined at all, has no diagnostic value. The secret to good practice is to produce an image with the ‘optimum density’, enabling the practitioner viewing the film to make an accurate evaluation of the patient’s condition and therefore the correct diagnosis. Unfortunately in conventional film-screen radiography, once the image is processed there is little that can be done to rectify a poor quality image except use a bright light to look at images with too much blackening.

Page 165

The equivalent term when viewing digital images is ‘brightness’ and this may be defined as the intensity of light that represents the individual pixels in the image. Brightness is controlled by the processing software and can be adjusted following image processing.

Numerous factors can affect the radiographic density, as shown in Box 14.1.

Box 14.1 Factors affecting the radiographic density

Milliampere-second (mA s)	This is the product of the mA and duration of the exposure (exposure time). The mA s directly affects the density of the image. Increasing the mA s increases the intensity (number of photons) in the X-ray beam and hence the number incident on the image receptor.
Kilovoltage (kV_p)	Increasing the kV_p increases the intensity of the X-ray beam. Unlike mA s there is no direct relationship between density and kVp. Changes in the kV_p also affect the quality of the beam (the range of photons) and the radiographic contrast.
Source–image distance (SID)	The intensity of the X-ray beam changes due to the inverse square law. As the distance to the source of the radiation (focus) increases the intensity decreases as a product of the square of the distance.
Use of grids	These absorb both the primary beam but, more predominantly, the scattered radiation. Grids reduce the overall density of the image and improve the radiographic contrast.
Film-screen speed	When using a film-screen combination the relative speed of the system will affect the density. Higher film-screen speeds are more efficient and increase the density of the image.
Collimation	Increasing the area of the patient that is irradiated will produce more scattered radiation and hence reduce the density of the image.
	Good collimation increases the contrast of the image.
Tube filtration	This is fixed in modern tubes but altering the filtration may affect the range of photons produced and hence the density.
Processing	Both the development of film in a film processor and the operating characteristics of the digital processor can affect the density.

When viewing a radiograph that is suboptimal, due to the radiographic density, it can be difficult to define the reason for the problem. This will be discussed at the end of the chapter.

CONTRAST

Contrast is a more difficult concept to understand in an image and can be influenced by a number of factors (Box 14.2). There are a number of ways of defining contrast in an image. Generally when we refer to contrast we mean the radiographic contrast, but practitioners should be precise in their terminology.

Box 14.2 Factors affecting radiographic contrast

Kilovoltage (kV_p)	Increasing the kV_p increases the intensity of the X-ray beam. The predominant process above 60 kV_p is Compton scatter. Generally speaking, as the kV_p increases the radiographic contrast decreases and this is due to the increased scattered radiation produced.
Subject contrast	The composition of the patient varies depending on the anatomical area. Images of the chest and abdomen have very different inherent contrast.
Collimation	It is essential to collimate (restrict the size of the beam) to the area being X-rayed. Increasing the area irradiated increases the scatter produced and therefore reduces the contrast.
Film-screen combination	The type and speed of the film-screen combination determines the inherent contrast of the image.
Grids	Grids eliminate scattered radiation and increase the contrast of the image.
Air gap	Air gaps have a similar function to grids. If there is a large distance between the object and the image receptor (OFD) a number of scattered photons, as they travel forward, will miss the image receptor.
Processing	Both the development of film and the manipulation of the digital image can affect the contrast.
VDU characteristics	The contrast can be adjusted by windowing the image post acquisition. Together with brightness adjustment, windowing allows the practitioners to view a wide range of structures.
Contrast agents	These are used to enhance the contrast of the image. They generally occupy an internal structure of the body, e.g. bowel. The subject contrast is therefore altered, affecting the contrast of the image.

Radiographic contrast

Differences in optical density on a radiograph:

• A high contrast image has few differences in density and is black and white (Fig. 14.6).

• Low contrast images have a large range of densities and more shades of grey (Fig. 14.7).

Figure 14.6 High contrast images with few shades of grey and large differences in densities.

Figure 14.7 Low contrast image with many shades of grey and small differences in densities.

Page 166

The radiographic contrast of any image will affect the ability to make an accurate diagnosis: if it is not optimum important structures may be poorly defined and an incorrect diagnosis made.

Page 167

1 Subject contrast (differences in densities within the patient):

• depends on the proton number of the structures and the electron density (see p. 112)

• varies depending on the area of the body

– the thorax has high inherent contrast and adjacent areas have large differences in proton number (air in lungs vs. dense heart muscle)

– the abdomen has low inherent contrast with adjacent areas having a similar proton number (muscle, fat and organs).

2 Subjective contrast (how we perceive the contrast in an image):

• depends on a number of factors including

– radiographic contrast

– viewing conditions

– experience of the practitioner viewing the image.

3 Objective contrast:

• is where the optical density is measured using a densitometer and two areas compared to determine the contrast. The results are then generally illustrated by a characteristic curve.

Practitioners may think that resolution of an image is dependant on the contrast qualities of the image. Unfortunately, due to the subjective nature of viewing images, the concept that these two factors are dependent on each other still confuses inexperienced practitioners. High contrast images appear sharper (higher resolution) to the inexperienced practitioner. When viewing images the ability to see small structures is the true measure of definition. The resolution should always be measured in an image, which has an optimum radiographic density and contrast.

LATITUDE

The latitude of the film-screen combination is determined during manufacture. Latitude can be defined as the region of useful exposure which will produce an image in the diagnostic range of densities.

In film-screen imaging an optimum exposure will produce an image with the details required in the useful density range (Fig. 14.8). Unfortunately it is not possible to manufacture a film with both high contrast and wide latitude. Generally speaking, the faster the film (the less exposure it needs to reach the maximum density) the higher the radiographic contrast and the narrower the latitude. A film-screen system with a speed index of 100 has the highest resolution and a wide exposure latitude. A film-screen system with a speed index of 400 requires less exposure to produce an image and has less resolution and narrow latitude. Figure 14.9 demonstrates the characteristic curves for two films – film A is a 400-speed system and film B a 100-speed.

Figure 14.8 Characteristic curve for a film-screen imaging system.

Figure 14.9 Characteristic curves for a high and low latitude film-screen imaging system.

Page 168

Overexposure will produce an image with too high a radiographic density whilst underexposure will produce an image with one that is too low. In a faster speed film-screen combination, which has narrower latitude, it is easier to over- or underexpose the image as the useful density region is narrow. The contrast of the image in both of these circumstances is reduced, affecting the practitioner’s ability to see detail within the image.

VIEWING A RADIOGRAPH

Having taken all the knowledge and skills of the practitioner to produce an optimum image it is essential that the image produced is viewed under optimum conditions. The equipment used and the environment used to view the image will affect the practitioner’s ability to extract information from the image. It is important not only to have correct viewing conditions but also to concentrate on viewing the image without distraction from colleagues.

The viewing box is the main equipment for viewing hard copy images. It is essential that all viewing boxes are part of a Quality Assurance programme and that the colour and intensity of light output are comparable. The viewing box must have even illumination, with a northern or tropical daylight output and a colour temperature of 6500 K.

There are six specific factors associated with the viewing of images that need to be considered when viewing images. The following information is taken from EUR 16260EN:

‘The proper assessment of image quality and accurate reporting on the diagnostic information in the radiographs can best be achieved when the viewing conditions meet the following requirement:

• The light intensity incident on the viewer’s eye should be about 100 cd/m². To achieve this, the brightness of the film illuminator should be between 2000 and 4000 cd/m² for films in the density range 0.5 to 2.2.

– This can be measured with a photometer.

• The colour of the illumination should be white or blue and should be matched throughout a complete set of film illuminators.

• Means should be available to restrict the illuminated area to the area of the radiograph to avoid dazzling.

– Only one viewing box should be illuminated and small images masked using black templates.

• Means for magnifying details in the displayed radiographic image should be available. These means should magnify by a factor of 2 to 4 and contain provisions to identify small image details of sizes down to 0.1 mm.

– This is achieved using a magnifying glass.

• For viewing exceptionally dark areas in the radiographic image an additional spotlight with iris diaphragm providing a brightness of at least 10 000 cd/m² should be available.

• A low level of ambient light in the viewing room is essential.’

Page 169

ACCEPTABLE IMAGE GUIDELINES

Image quality is normally considered in terms of the practitioner’s ability to perceive normal anatomy or detect any potential pathology. The image must have optimum contrast and density, maximum image sharpness and minimal noise. The image should demonstrate the anatomy or pathology of interest as conspicuously as possible; for example in extremity radiography this is bony trabeculae and for a chest radiograph a sharp reproduction of the vascular pattern in the whole lung, particularly the peripheral vessels, is required.

EUR 16260 EN gives details of guidelines for quality criteria for a number of radiographic examinations (chest, skull, lumbar spine and pelvis)¹ and EUR 16261EN outlines quality criteria for paediatrics.² Images should only be repeated if they are not of diagnostic quality (Box 14.3). If the clinical question is answered by the image but it is of poor aesthetic quality the image should not be repeated.

Box 14.3 Factors affecting acceptable image quality

Image annotation	The patient identification, the date of examination, positional markers and the name of the facility must be present and legible on the film. These annotations should not obscure the diagnostically relevant areas of the radiograph.
Patient positioning	Correct patient positioning plays a major role in determining the success of any radiographic examination. The area of interest must be included and the patient must be correctly positioned to demonstrate the relevant anatomy.
Image definition	The relevant anatomy is visualised and the area of interest sufficiently penetrated, e.g. bone trabeculae, skin surface. The structures are shown without misleading distortion.
Collimation	Image quality is improved and the radiation dose to the patient is reduced by limiting the X-ray beam to the smallest field giving the required diagnostic information. Radiosensitive organs need to be excluded whenever possible from the primary beam. On no occasion should the X-ray beam fall outside the image receptor area.
Protective shielding	for radiation protection purposes, standard protection devices should be available to shield radiosensitive tissues or organs whenever possible.
Film blackening and contrast	This has a major influence on image quality. The range of the mean optical density (D) of a clinical radiograph should normally lie between D = 1.0 and D = 1.4 and the optical densities of fog and film base should not exceed D = 0.25. For the diagnostically relevant parts of the film, the overall range of optical densities should lie between 0.5 and 2.2. These densities are difficult to assess subjectively; therefore the definition of one or a few critical points of the particular radiographic projections would be desirable where the optical density of a specific anatomical feature and its contrast relative to the surrounding image could be measured; e.g. bony trabeculae and skin surface.
Film processing	Optimal processing of the radiographic film has important implications both for the diagnostic quality of the image and for the radiation dose to the patient. Film processors should be maintained at their optimum operating conditions as determined by regular and frequent (i.e. daily) quality control procedures.
Radiographic exposure conditions	Knowledge and correct use of appropriate radiographic exposure factors, for example kV_p, focal spot value, tube filtration, and source–image distance (SID) is necessary because they have a considerable impact on image quality.
Artefacts	The image should be free from artefacts, especially clothing which may obscure relevant details.

Page 170

COMMON FAULTS WITH REMEDY

This is a difficult concept to articulate and I would strongly advise the practitioner to perform the task of image evaluation with an experienced radiographer or clinical tutor. Often it is difficult to determine the exact cause of the fault and a systematic approach is desirable to determine what may have caused your image to be of poor quality (Table 14.1).

Table 14.1 Common faults with remedy

Fault	Possible causes and remedies
Poor contrast	Poor collimation – collimate to the area of interest Obese patients – consider compression binder or using a grid
Artefacts	Clothes, ECG leads
Unsharp image	Imaging geometry, patient movement, poor screen/film contact (only affects limited areas of the image) Increase source–object distance • Use fine focus • Use short exposure times • Give clear instructions • Make patient comfortable • Consider immobilisation with pads/sandbags • Good QA and identify cassette responsible and remove from practice
Increased film blackening (too dark an image)	Overexposure (this is the most obvious reason for an image with increased film blackening and the easiest one to remedy) Simply check the exposure factors on the control panel Incorrect film/screen combination Again, check the cassette for the film/screen speed. You may have used a regular (400 speed) cassette for an extremity (100 speed) Automatic Exposure Device (AED) This may occur for a number of reasons: • Collimation within the chamber selected • Incorrect chamber selected (the structure selected has a higher proton number or is thicker than the structure of interest) e.g. an outside chamber over the chest when undertaking a spine image Film fogging Fogging of films may occur along the edge of the film due to light leakage into the cassette or globally due to light leak or faulty safelights in the darkroom. Neither of these processes are likely in a modern department due to QA programmes Film processing There are a number of reasons for dark images due to faulty processing Overactive developer may be due to over-replenishment, too high a developer temperature or too long in the developer. These faults are unlikely to occur due to processor QA but will affect several images, not just the one the practitioner is concerned about.
Reduced film blackening (too light an image)	Underexposure This is the most obvious reason for an image with decreased film blackening and the easiest one to remedy; simply check the exposure factors on the control panel Incorrect film / screen combination Again check the cassette for the film/screen speed. You may have used an extremity (100 speed) cassette for a shoulder (400 speed) Automatic Exposure Device This may occur for a number of reasons: • Too short an exposure time to allow the density to be achieved • The incorrect chamber selected (area selected has a lower atomic number or is thinner than the required area), e.g. an inside chamber over the chest when undertaking a spine image • Area of body does not cover the relevant chamber Film processing There are a number of reasons for light images due to faulty processing: • Underactive developer due to over-replenishment • Too low a developer temperature • Contamination of the chemistry These faults are unlikely to occur due to processor QA but will affect several images, not just the one the practitioner is concerned about.

Page 171

References

1 European guidelines on quality criteria for diagnostic radiographic images. June 1996 EUR 16260

2 European guidelines on quality criteria for diagnostic radiographic images in paediatrics. July 1996 EUR 16261 EN