Appropriate radiographic images with good technical quality are essential for proper radiographic interpretation. Improper radiographic technique or positioning can hinder interpretation of such images by veterinarians; this could cause lesions to be undetected, which may lead to misdiagnosis and incorrect treatments. Effects of the use of suboptimal radiographic images and the importance of radiographic image quality have been described by other authors.1 Many radiology textbooks2–20 are available in which proper radiographic techniques are described. Results of a recent study21 indicate 49.6% of repository radiograph sets for Australian Thoroughbreds in yearling sales contained at least 1 nondiagnostic radiograph (ie, radiograph with insufficient technical quality for accurate interpretation). In that study,21 errors in positioning and overexposure were the most common reasons radiographs were considered of nondiagnostic or poor quality. To the authors' knowledge, similar studies of the frequency of errors in radiograph sets for small animals have not been performed. The objective of the study reported here was to identify common technical errors in radiographic images of small animal patients obtained by referring veterinarians of a university hospital, determine the effect of such errors on the perceived quality of the radiographic images, and discuss the reasons such errors should be corrected by veterinarians to improve the diagnostic quality of radiographs and reduce the risk of misinterpretation.
Materials and Methods
Sample—One hundred thirty-five radiographic image sets submitted by referring veterinarians to the Kansas State University Veterinary Health Center were collected during July through October 2012. Analog (ie, conventional) radiographic film sets were mailed or brought by clients to the hospital. Digital radiographic image sets were mailed or brought by clients to the hospital on compact discs.
Evaluation of radiographs—The following observational information was recorded for each set of radiographs: region of interest (abdomen, thorax, musculoskeletal, or miscellaneous [including skull, neck, spine, or whole body]) and the radiographic format (digital or film). Each image and radiograph set was assessed by 1 author (EKN) for grading of 8 variables: exposure, collimation, positioning, centering of the region of interest on an image, inclusion of all appropriate views for the region of interest in a set, adherence to proper radiation safety procedures, presence of radiographic artifacts, and appropriate label information. Radiographs were graded for each of those categories as follows: acceptable performance of the category for all images in a set (yes) or detection of an error in at least 1 image in a set (no; Appendix). For each radiograph set, the total number of categories for which at least 1 error was detected was determined. Radiographic artifacts were identified by 1 author (EKN), radiographic artifacts were grouped into categories, the frequency of detection of artifacts in each category was recorded, and the frequency of detection of artifacts was compared between sets of film and digital radiographs. In addition, the overall diagnostic quality of each radiograph set was subjectively determined by 2 investigators: a board-certified radiologist (LJA) and a second-year radiology resident (EKN). Each of those authors identified radiograph sets as diagnostic or nondiagnostic by means of subjective assessments regarding whether the study could be determined to be radiographically normal or have radiographically detectable lesions on the basis of evaluation of the available images. After assessment by the 2 investigators, a radiograph set was determined to be of diagnostic quality only if both investigators identified it as diagnostic and a radiograph set was determined to be nondiagnostic if at least 1 investigator identified it as nondiagnostic.
Statistical analysis—For each evaluated radiographic variable, the proportions of radiograph sets for which the variable was or was not correctly performed were compared between sets determined to be diagnostic and those determined to be nondiagnostic by means of χ2 frequency analysis.a In addition, the effect of the type of radiograph (digital or film) on the assessment of overall diagnostic quality and for each of the variables was compared by means of χ2 frequency analysis. Values of P ≤ 0.05 were considered significant.
Results
A total of 135 sets of radiographs of animals were evaluated. Nine of 135 (7%) sets had no errors (ie, acceptable technical characteristics for all evaluated categories of variables); 126 of 135 (93%) radiograph sets had ≥ 1 error in 1 or more of the 8 categories evaluated. Fifty-one of 135 (38%) studies had errors in exposure, 66 (49%) had errors in collimation, 34 (25%) had errors in positioning, 56 (41%) had errors in centering, 23 (17%) had errors regarding inclusion of all appropriate views, 11 (8%) had errors regarding use of appropriate radiation safety, and 64 (47%) had errors in labeling. Fifty-five of 135 (41%) sets of radiographs had ≥ 1 artifact detected in at least 1 image; 67 artifacts were detected. The 2 most common categories of artifacts detected in radiograph sets were processing artifacts (n = 20 [film radiograph sets only]) and motion artifacts (20; 18 film and 2 digital radiograph sets; Table 1).
Frequency of detection of various types of radiographic artifacts in referral film and digital images of small animal patients evaluated at a university hospital.
| Artifact | Film (n = 93) | Digital (n = 42) |
|---|---|---|
| Processing artifacts (roller lines, fingerprints, smudges, discoloration, drip lines, overlap of films, exposure error attributable to light being turned on before film was completely inside the processor) | 20 | — |
| Motion | 18 | 2 |
| Bending of film or fingernail marks | 8 | — |
| Scratched film | 8 | — |
| Grid lines | 4 | 1 |
| Debris in cassette | 3 | 0* |
| White line artifact | — | 2* |
| Black specks over image (attributable to metal or grit deposits in film emulsion) | 1 | — |
Data are the number of radiograph sets for which the artifact was detected in at least 1 image.
Only applicable for digital images acquired by means of a computed radiography technique.
— = Not applicable.
Forty-two (31%) sets of radiographs were acquired digitally, and 93 (69%) were acquired on film. No significant (P = 0.533) difference was detected between types of radiographic images (film vs digital) regarding classification as diagnostic or nondiagnostic. No significant effect of the type of radiograph (film vs digital) was detected on the variables of exposure, collimation, positioning, centering, inclusion of all appropriate views, or radiation safety. Digital radiographs had significantly lower frequency of artifacts (P < 0.001) and higher frequency of proper labeling (P < 0.001) versus film radiographs.
One of the evaluators (EKN) considered 90 of 135 (67%) radiograph sets to be of diagnostic quality and 45 (33%) to be of nondiagnostic quality. The variables correct exposure, correct positioning, inclusion of all appropriate views, and a lack of artifacts were significantly associated with determination of a radiograph set as diagnostic for this investigator. The other evaluator (LJA) considered 83 of 135 (61%) radiograph sets to be of diagnostic quality and 52 (39%) to be of nondiagnostic quality. Correct exposure, inclusion of all appropriate views, and a lack of artifacts were significantly associated with determination of a radiograph set as diagnostic for this investigator.
Radiograph sets that were considered diagnostic by both reviewers and those that were considered nondiagnostic by at least 1 of the reviewers were identified, and the proportion of sets with correct and incorrect performance of each evaluated radiographic variable was compared for those 2 groups (Table 2). Seventy-five of 135 (56%) radiograph sets were considered to be of diagnostic quality by both evaluators, and 60 (44%) were considered to be of nondiagnostic quality by at least 1 evaluator. A lack of errors in the categories of exposure (P < 0.001), positioning (P = 0.002), inclusion of all appropriate views (P = 0.008), and artifacts (P = 0.021) was significantly associated with classification of radiograph sets as diagnostic. The frequency of detection of errors in the categories of labeling, radiation safety, centering, and collimation was not significantly different between radiograph sets determined to be of diagnostic quality and those determined to be of nondiagnostic quality. The determination of a radiograph set as being of diagnostic or nondiagnostic quality was significantly associated with a lower (P < 0.001) or higher (P = 0.018) number of categories of errors, respectively. There was no significant (P = 0.755) effect of the region of interest of radiograph sets (ie, thorax, abdomen, or extremity) on whether both evaluators considered a radiograph set to be of diagnostic quality.
The number of radiograph sets determined to be of di agnostic quality by 2 investigators (n = 75 [50 sets of film and 25 sets of digital radiographic images]) and those determined to be of nondiagnostic quality by at least 1 of 2 investigators (n = 60 [43 sets of film and 17 sets of digital radiographic images]) for which various radiographic variables were (yes) or were not (no) properly performed.
| Diagnostic | Nondiagnostic | |||
|---|---|---|---|---|
| Radiographic variable | Yes | No | Yes | No |
| Proper exposure* | 66 | 9 | 18 | 42 |
| Proper collimation | 36 | 39 | 33 | 27 |
| Proper patient positioning* | 64 | 11 | 37 | 23 |
| Proper centering of anatomic structures | 47 | 28 | 32 | 28 |
| All appropriate views acquired* | 68 | 7 | 44 | 16 |
| Proper use of radiation safety | 70 | 5 | 54 | 6 |
| No artifacts detected* | 51 | 24 | 29 | 31 |
| Proper labeling | 40 | 35 | 31 | 29 |
Indicates significant (P ≤ 0.05) difference between diagnostic and nondiagnostic radiograph sets regarding expected frequencies (determined with a χ2 test).
Discussion
Results of this study indicated the radiographic variables exposure, positioning, absence of artifacts, and inclusion of all appropriate views were associated with consideration of a radiograph set as being of diagnostic quality. Therefore, these variables were considered important for image quality.
Results of this study indicated incorrect exposure was an important technique error that had a negative effect on perceived radiograph quality. An underexposed radiograph does not have as much overall contrast as a properly exposed radiograph.8 For example, underexposure may prevent discrimination of small intestinal margins in the abdomen and cause errors in interpretation of serosal detail. Underexposure in spinal or extremity radiographs also causes low image contrast, which may hinder evaluation of fine detail in bones; for example, this may cause failure of detection of subtle lytic lesions. Overexposure is similarly detrimental to radiographic quality by causing irreversible loss of information in dark portions of an image; such information cannot be retrieved by viewing films with a bright light (film radiographs) or using window and level adjustments (digital radiographs; Figure 1). For example, in an overexposed thoracic radiographic image, lungs may appear black, parenchymal detail is lost, and small metastatic pulmonary nodules may not be detected.
Representative lateral digital radiographic images of the thorax of a dog indicating overexposure (A) and proper exposure (B). In the overexposed image (A), notice that the pulmonary parenchyma cranial to the cardiac silhouette (arrow) cannot be detected; pathological lesions in this area would not be detected in such an image. In the image with proper exposure (B), the lung parenchyma (which is radiographically normal in this image) can be evaluated. A ventrodorsal view of the thorax would be required for complete evaluation. These images were obtained from medical records of Kansas State University and were not evaluated in this study.
Citation: Journal of the American Veterinary Medical Association 244, 8; 10.2460/javma.244.8.961
Digital and analog film radiographs have some differences regarding effects of exposure errors. Radiographic film is sensitive to small changes in exposure to x-rays. Film requires a narrow range of exposure settings (peak kilovoltage and milliampere combinations) to produce a high-quality image; small errors in settings that cause overexposure or underexposure may yield film radiographic images of unacceptable quality. Digital radiographic images, however, have a wider range of exposure for which acceptable images may be produced (ie, high exposure latitude). The high exposure latitude is an advantage of digital radiography versus film radiography. The ability to change window or level settings to increase or decrease contrast in digital radiographic workstations allows high-contrast tissues (ie, bone) and soft tissues to be appropriately displayed simultaneously in a single image.22 For digital radiographs with apparent errors in exposure, evaluators should first adjust the window or level settings in an attempt to correct the exposure. Additional radiographic images should be acquired with different settings only when such adjustments do not yield an image of acceptable exposure. This avoids unnecessary x-ray exposure of patients and personnel. Despite the high exposure latitude of digital radiographs and the ability to adjust window or level settings, such images may have exposure errors that cannot be corrected with postprocessing adjustments. In the present study, 51 radiograph sets had errors in exposure; 15 of those were sets of digital radiographs. Underexposed film radiographs have uniform whitening because of an inadequate quantity of x-rays, and inadequate x-ray penetration of tissues. However, underexposed digital images have quantum mottle, which yields an overall grainy appearance of images (Figure 2).22–24 The effects of overexposure in digital radiographic images are similar to the effects of such exposure on film radiographic images; images appear dark and contrast is low, which cannot be compensated with window or level adjustments. Use of digital radiography may yield improvements in image quality, compared with film radiography; however, good radiographic technique is essential with either method.21
Representative lateral digital radiographic images of a stifle joint of a dog indicating underexposure (A) and proper exposure (B). In the underexposed image (A), notice the quantum mottle that results in a grainy appearance, which impairs evaluation of fine trabeculation in osseous structures, periarticular osteophytes, and soft tissues. In the properly exposed image of the same stifle joint (B), such structures can be more easily identified. These images were obtained from medical records of Kansas State University and were not evaluated in this study.
Citation: Journal of the American Veterinary Medical Association 244, 8; 10.2460/javma.244.8.961
Patients must be positioned correctly for identification of asymmetry and detection of abnormalities; incorrect positioning can affect proper interpretation (Figure 3). In images with obliquity, patient anatomy can be distorted, and interpretation of images may be incorrect.25,26 For example, obliquity can change the shape of the cardiac silhouette, which may cause false identification of cardiac chamber enlargement. Other references3,9,10,13,15–20 are available in which positioning techniques for radiography of small animal patients are described.
Representative lateral digital radiographic images of the abdomen of a dog with improper (A) and proper (B) positioning. In the image with improper positioning (A), notice that the anatomic structures are distorted and evaluation of the kidneys and retroperitoneal space is not possible. Notice that proper positioning of the same animal (B) allows detection of effusion in the retroperitoneal space that could not be identified in the radiograph with improper positioning. In a ventrodorsal view (not shown) of the abdomen of this animal, the kidney margins were not visible. A round mass associated with the spleen is also evident in the cranioventral aspect of the abdomen. These images were obtained from medical records of Kansas State University and were not evaluated in this study.
Citation: Journal of the American Veterinary Medical Association 244, 8; 10.2460/javma.244.8.961
Radiographic artifacts are variable in size, location, and extent of interference with image interpretation.12,16 For example, a small fingernail scratch mark in the corner of a radiographic film would likely not interfere with proper radiograph interpretation; however, substantial development errors that cause discoloration of an entire film can make interpretation of anatomic features difficult. In the present study, radiographic artifacts were detected in radiograph sets determined to be diagnostic and in those determined to be nondiagnostic. Some artifacts are unique to film processing, and some are unique to digital radiography techniques; veterinarians should be familiar with the artifacts for the type of radiographic technique they are using. When artifacts interfere with the ability to properly interpret a radiograph, veterinarians should identify the artifact, determine how to correct it, and obtain a new radiograph of sufficient quality for accurate interpretation.
The 2 most common artifacts detected in radiographs in the present study were those involving film processing or darkroom techniques (n = 20) and motion artifacts (20). Radiographic film processing or darkroom technique artifacts included the following: roller lines, fingerprints or smudges caused by handling before film was completely dry, discoloration attributable to improper fixation or developing, liquid drip lines attributable to inadequate drying, line artifacts caused by overlap of films in an automatic processor, and exposure to a light that was turned on before a film was completely inside an automatic processor. Such artifacts can typically be avoided with training of personnel regarding proper development techniques and the proper use of automatic processors.2,12,16,27 In addition, personnel should attend to detail when performing such tasks.
Motion is a substantial problem during radiography of veterinary patients and may be attributable to lack of patient compliance and respiratory motion. Motion degrades image quality by decreasing image sharpness. Motion attributable to respiration should be minimized, especially when obtaining thoracic radiographs (Figure 4).
Representative lateral digital radiographic images of the thorax of a dog with (A) and without (B) motion artifact. In the image with motion artifact (A), notice that osseous and pulmonary detail cannot be identified. In the radiographic image of the same patient without motion artifact (B), notice the lesion of the third rib (arrow) that cannot be identified in the image with motion artifact. These images were obtained from medical records of Kansas State University and were not evaluated in this study.
Citation: Journal of the American Veterinary Medical Association 244, 8; 10.2460/javma.244.8.961
At least 2 orthogonal views are typically necessary for thorough evaluation and accurate localization of abnormalities. In this study, we frequently reviewed radiograph sets that included only a single radiograph; we believed this limited the diagnostic value of the available image. Radiograph sets including only a single image are considered incomplete and may lead to misdiagnoses. Anatomic areas that have a radiographically normal appearance in a single radiographic view may have pathological lesions in an orthogonal view. For example, fractures may be displaced in only 1 plane; radiographs of such fractured bones may have minimal changes in 1 view and obvious displacement in the orthogonal view. At least 2 orthogonal views are necessary for complete evaluation of such fractures (Figure 5). In anatomic areas with complex features (eg, carpal, tarsal, or phalangeal regions), 2 additional oblique views obtained at 45° to the other views may be routinely acquired. Such oblique views provide additional images of anatomic structures to increase a clinician's ability to identify and localize lesions. Oblique views also provide tangential views of additional surfaces (eg, dorsolateral or palmaromedial surfaces) of anatomic structures.
Representative craniocaudal (A) and lateral (B) digital radiographic images of the tibia of a dog. Notice that the type II Salter-Harris fracture of the distal aspect of the tibia can be identified only in the lateral view because it is displaced in 1 plane. Two radiographic views are necessary for accurate interpretation. A fracture of the distal portion of the fibula is also evident. These images were obtained from medical records of Kansas State University and were not evaluated in this study.
Citation: Journal of the American Veterinary Medical Association 244, 8; 10.2460/javma.244.8.961
Repeated radiography was not performed for all animals in this study. For those animals that had undergone such imaging, repeated radiographic images were not assessed in the study. Because of this, no gold standard images were available for identification of possible lesions in orthogonal views for animals that had radiograph sets with only a single radiographic view.
In the present study, digital radiography did not result in images of higher quality than film radiography for the purpose of determination of a diagnosis. The type of image did not affect the variables exposure, collimation, positioning, centering, acquisition of all appropriate views, or radiation safety. This was likely because these variables were dependent on operators (ie, exposure settings, positioning of animals, and collimation were manually performed). Thus, it was not surprising that such operator-dependent errors were identified with similar frequency for digital and film radiographs.
Digital radiographic images had fewer artifacts and higher frequency of proper labeling versus film radiographs. The low number of digital radiographic images with artifacts was attributed to the fact that a high percentage of observed artifacts were caused by film handling, processing, and darkroom technique errors, all of which are eliminated during digital radiography. Compared with film images, digital radiographic images have fewer variables that interfere with diagnostic quality after acquisition of images because more processing steps are required for film images after exposure.
Radiographs are a part of a patient's medical record, and it is important to have labeling information included permanently in each image. Legally acceptable labeling methods for film radiographs require information to be present in the film emulsion in a permanent manner; such labels must include the client and patient name or identification number; the name of the hospital, clinic, or veterinarian; and the date on which the radiograph was obtained.28 Inclusion of additional information may be appropriate. To ensure radiographs can be identified as authentic, such information must be permanently included in a film or digital image in a manner that will not allow fading or other changes over time.28 Not all states have specific regulations regarding radiograph labeling; however, failure to include such information on radiographs is considered to be below the standard of care for veterinary patients.28 Many radiographic films evaluated in the present study had no label information, or such information was written in a nonpermanent format (ie, written with a marker on an edge of the film or on paper attached to the film); such radiographs could not be used as evidence in a court of law. Permanent labeling of each radiograph is also necessary for organization of radiographic image files. Many digital radiography systems require input of patient information before such information can be included in an image, which is analogous to preparing a label for analog film radiography. Nevertheless, we attributed the low rate of detection of proper labeling errors in digital images to the ability to associate patient, veterinary hospital, time, and date information with metadata of images in Digital Imaging and Communications in Medicine (DICOM) format. For film images, label information must be added to the film by an operator prior to development; because this is a distinct step, it is often omitted.
A limitation of the present study was that radiographs were determined to be of diagnostic or nondiagnostic quality on the basis of subjective evaluations of 2 investigators. A gold standard method was not available with which to compare results of radiographic image evaluations. The study design would have been improved if we had included only radiograph sets of patients with available error-free repeated radiographic images, to which referral images could have been compared. Furthermore, radiograph sets of high quality without artifacts may not indicate pathological lesions of animals that may be detected with other methods such as surgery or necropsy. Despite this limitation, we believe results of this study provided important information regarding common radiographic errors that decrease perceived image quality and that the variables identified should be considered by veterinarians during radiography. Another limitation of the study was that investigators did not agree regarding identification of radiographs as diagnostic or nondiagnostic for all animals. We attempted to reduce bias by reporting results for radiograph sets that both reviewers considered to be of diagnostic quality and those that at least 1 reviewer considered to be of nondiagnostic quality. Statistical analyses were performed for each of these groups. Statistical power may have been increased by including a third evaluator so that a majority decision could be determined regarding perceived radiograph quality.
Accuracy of radiograph interpretation is affected by image quality. Results of the present study indicated the variables radiographic exposure, correct positioning, absence of artifacts, and acquisition of all appropriate views were significantly associated with perceived diagnostic quality of radiograph sets. Proper image labeling, adherence to radiation safety, proper centering of anatomic structures in images, and proper collimation were not significantly associated with perceived diagnostic quality. The number of categories of radiographic errors detected was significantly associated with perceived diagnostic quality. Digital radiographic images had significantly fewer artifacts and higher frequency of proper labeling versus film images. Knowledge of common radiography errors may aid veterinarians in acquisition of radiographic image sets of high quality.
Winks SDA 6, version 6.0.93, Texasoft, Cedar Hill, Tex.
References
1. Tuncer S, Aksu N, Dilek H, et al. Fractures of the fingers missed or misdiagnosed on poorly positioned or poorly taken radiographs: a retrospective study. J Trauma 2011; 71:649–655.
2. Ticer JW. Darkroom theory and technique. In: Radiographic technique in veterinary practice. 2nd ed. Philadelphia: WB Saunders, 1984;35–54.
3. Ticer JW. Radiographic positioning and technique: small animals. In: Radiographic technique in veterinary practice. 2nd ed. Philadelphia: WB Saunders, 1984;103–395.
4. Ticer JW. Exposure factors. In: Radiographic technique in veterinary practice. 2nd ed. Philadelphia: WB Saunders, 1984;55–69.
5. O'Brien TR. Basic radiography of the abdomen. In: Radiographic diagnosis of abdominal disorders in the dog and cat. Davis, Calif: Covell Park Vet Co, 1978;3–7.
6. Wolfgang L. The role of the darkroom in radiograph quality control. N Y State Dent J 1998; 64:26–29.
7. Horner K. Quality assurance: 2. The image receptor, the darkroom and processing. Dent Update 1992; 19:120, 122–124, 126.
8. Thrall DE, Widmer WR. Radiation protection and physics of diagnostic radiology. In: Thrall DE, ed. Textbook of veterinary diagnostic radiology. 6th ed. Philadelphia: WB Saunders, 2013;2–21.
9. Thrall DE. Principles of radiographic interpretation of the thorax. In: Thrall DE, ed. Textbook of veterinary diagnostic radiology. 6th ed. Philadelphia: WB Saunders, 2013;474–488.
10. Kleine LJ, Warren RG. Restraint and positioning. In: Mosby's fundamentals of animal health technology small animal radiography. St Louis: CV Mosby, 1983;62–119.
11. Kleine LJ, Warren RG. Radiographic quality and exposure factors. In: Mosby's fundamentals of animal health technology small animal radiography. St Louis: CV Mosby, 1983;120–155.
12. Morgan JP & Silverman S. Radiographic quality. In: Techniques of veterinary radiography. 4th ed. Ames, Iowa: Iowa State University Press, 1987;94–103.
13. Morgan JP & Silverman S. Small animal radiography. In: Techniques of veterinary radiography. 4th ed. Ames, Iowa: Iowa State University Press, 1987;170–209.
14. Lavin LM. Radiographic quality. In: Radiography in veterinary technology. 2nd ed. Philadelphia: WB Saunders, 1994;41–46.
15. Lavin LM. Radiographic positioning. In: Radiography in veterinary technology. 2nd ed. Philadelphia: WB Saunders, 1994;143–230.
16. Han CM, Hurd CD. Achieving radiographic quality and radiographic positioning. In: Practical diagnostic imaging for the veterinary technician. 3rd ed. St Louis: Elsevier Mosby, 2005;10–123.
17. Mauragis D, Berry CR. Small animal thoracic radiography. Vet Med 2011;(Sept/Oct):45–50.
18. Mauragis D, Berry CR. Small animal abdominal radiography. Vet Med 2011;(Nov/Dec):53–60.
19. Mauragis D, Berry CR. Small animal pelvic radiography. Vet Med 2012;(Jan/Feb):48–53.
20. Mauragis D, Berry CR. Small animal radiography stifle joint and crus. Vet Med 2012;(March/April):53–58.
21. Jackson MA, Vizard AL, Anderson GA, et al. Identification and prevalence of errors affecting the quality of radiographs submitted to Australian Thoroughbred yearling sale repositories. Vet Radiol Ultrasound 2011; 52:262–269.
22. Mattoon JS. Digital radiography. Vet Comp Orthop Traumatol 2006; 19:123–132.
23. Drost WT, Reese DJ, Hornof WJ. Digital radiography artifacts. Vet Radiol Ultrasound 2008; 49:S48–S56.
24. Jiménez DA, Armbrust LJ, O'Brien RT, et al. Artifacts in digital radiography. Vet Radiol Ultrasound 2008; 49:321–332.
25. Whatmough C, Moores AP, Magalhaes RS, et al. Factors affecting width of the canine femorotibial joint space in nonweightbearing radiographs. Vet Radiol Ultrasound 2008; 49:129–134.
26. Rhodes W. Positioning of the canine hips for radiography. Vet Radiol 1961; 2:39–40.
27. Watters JW. The radiographic darkroom and film processing. Compend Contin Educ 1982; 4:311.
28. Wilson JF. Law and ethics of the veterinary profession. Yardley, Pa: Priority Press, Ltd, 2008;348.
Appendix
Definitions and criteria used to determine grades (yes vs no) for variables used to evaluate sets of radiographs provided by referring veterinarians for small animals evaluated at a university hospital.
| Variable | Definition and grading criteria |
|---|---|
| Proper exposure | Radiographic exposure appropriate for the area of interest in all images (eg, lungs not overexposed, adequate contrast for abdominal structures, appropriate detail for bones in extremities; for digital images, a grade of yes indicated acceptable exposure could be obtained by manual adjustment of window and level controls and a grade of no indicated such adjustment did not result in an image of acceptable exposure). |
| Proper collimation | Radiographic exposure collimated appropriately to the area of interest in all images (eg, entire length of bone included but not the entire limb, or entire abdomen with diaphragm included but not entire thorax). |
| Proper positioning | Animal properly positioned without obliquity in all images so that anatomic structures could be properly identified (did not include intentional obliquity for images of the carpal, tarsal, or phalangeal regions). |
| Centering of region of interest on image | All views centered on the region of interest (eg, image centered at the caudal border of the scapula [for thoracic images], caudal extent of ribs [for abdominal images], or the mid-diaphyseal region of the femur [for femoral images]). |
| Inclusion of all appropriate views for body region of interest | At least 2 orthogonal views of the region of interest included. |
| Radiation safety adherence | Adherence to proper general radiation safety protocols (eg, no visible gloved or ungloved hands in the primary x-ray beam and no ungloved hands visible outside the primary x-ray beam that were visible in the radiograph because of scattered radiation). |
| Absence of artifacts | All images free of radiographic artifacts (exposure errors were not included in this category). |
| Legal label information included | All images had the proper permanently printed or embedded legally required labeling information in images (including patient name or identification number, date, name of hospital or veterinarian, and laterality marker for images of extremities or the spinal column). |
