(For a printable version without graphics, please click here)
Learning Objective - To better understand the complexity of image interpretation and to acquire some visual experience with radiographic images.
An important thing to understand about interpreting conventional radiographic images is that it is not possible to "teach" this skill. The only way to become a skillful (as skillful as it is possible to be) interpreter of conventional radiographic images is to carefully study a large number of these images. In other words acquiring skill is almost totally a function of visual experience. Within that limitation, some guidance can be given to improve the efficiency of the process by which you gain that visual experience. There is very helpful website on this subject which is linked below. Please use this website to supplement this discussion. There are many specific topics listed on the website so after you read this text, use the URL listed below to provide additional information on any aspect of the discussion about which you still have questions.
**For further instructions for the interpretation of chest x-rays, see the links below.**
Reading the Chest X-Ray
You have probably heard comments about the importance of a systematic or even rigid search pattern when interpreting radiographs. When tested, the specific search pattern does not seem to have much to do with interpretation accuracy. In the best known research article (Visual Search Patterns and Experience with Radiological Images. Diagnostic Radiology 103:523-528. June, 1972) Kundel makes the comment that instead of trying to develop a consistent search pattern it would be more useful to develop a well-defined sense of what is normal. If you have a clear expectation for what the normal image pattern should be, any abnormality will quickly attract your eye. But, it is not easy. It takes a lot of experience to have a well-developed expectation for what the normal image pattern should be for various imaging methods and multiple body parts. It is important to remember that your skill will improve with experience so you should take every opportunity to study radiographs critically, not just look at the film but actually analyze the image in a detailed way. As you become more critical and look at the subtle details of the film, you will see density variations which cannot be explained. As you resolve these uncertainties by reviewing teaching files or reading textbooks, you will gradually extend your knowledge of imaging and develop your own understanding of what is the normal image pattern.
Remember, all density variations on the film may not be related to the patient's anatomy or pathology. The imaging system is complex and there are many possibilities for creating artifactual density variations. Of course, the imaging system includes the technologist who controls the equipment. The image shown below is an interesting anatomical study but it is hard to believe. Take a close look at this image and try to explain why the inferior thoracic vertebrae shown on the image are so large. Your knowledge of human anatomy should make it difficult for you to believe this image is a realistic representation of the patient. Because of that, you might begin to reconstruct the situation in which this image was made. When I did that, the technologist quickly realized that she did not change the film from a previous localized examination of the lateral lumbar spine on another patient. Because of that, the lower cervical spine of this patient was superimposed (double exposed) on the lower lumbar spine film of the previous patient. That is the explanation for the unusual size of the vertebrae seen on this film. The important thing to understand from this example is that even though a film may superficially look anatomic, it is not always believable so be very critical in your analysis of each image.
Routine Radiographic Examination of the Chest
Chest radiography is the most frequent radiographic examination done in the United States. Actually, chest radiographs account for almost one half of the radiographic procedures each year in this country. The exact reason for this frequent use is not clearly understood. It may be a carry over from the period in which chest radiography was done to detect infectious pulmonary disease. There may also be some feeling that this part of the body has the most important organs.
The routine chest examination usually consists of two 14x17 inch films. The first is a frontal view of the chest recorded from the posterior to anterior projection (PA) the second film is a lateral view which can be recorded with the beam entering from either side . There is currently some controversy about when the lateral projection should be used. Some radiologists believe that every chest examination should include both the PA and lateral films. A more modern view is that the use of the lateral film should be a second level decision based on the findings obtained from the PA projection. The reason for this question about the utility of the lateral film is that the lateral radiograph requires considerably more radiation exposure than the PA projection while the lateral film usually contributes considerably less useful clinical information. Therefore, the lateral film has a much lower benefit/risk ratio than the PA film.
When interpreting the PA projection, it is important to specifically view all structures on the image. Every interpreter develops a unique process for doing this. A common method would be to begin with the heart and mediastinum, evaluating these structures for size and configuration then moving to a general examination of the lungs looking for areas of abnormal density. Of course, normal structures (such as the pulmonary arteries and veins) of water density are present in the lung fields and the interpreter must develop an understanding of the normal pattern by studying a large number of films (Tuddenhams article on chest Roentgenology may be helpful*). After carefully evaluating all areas of the lungs, specific questions about the pleural space could be addressed such as, is there any evidence of a pneumothorax or pleural effusion? Then, the portion of the thoracic skeleton included on the film should be looked at bone by bone. Destructive lesions of the ribs are particularly hard to recognize on a chest radiograph so you should be very careful to look at each rib in detail when interpreting the chest radiograph of a patient who might have rib metastases.
If abnormalities are found which are difficult to localize or additional information is needed, oblique views can be used. These are usually done in the posterior to anterior projection and are named for the side of the body adjacent to the film. For instance, a film recorded with the beam entering the right side of the patient's back and exiting through the left anterior aspect of the chest would be called a left anterior oblique. With a little practice, you will be able to tell which projection was used from the shape of the heart and mediastinum on the radiograph.
The radiation exposure from a chest radiograph is quite low considering the thickness of the body part. With good technique, a single film of the chest usually gives a mid plane chest exposure of approximately twenty (20) millirads (.00020 Gray). The radiation exposure from the same projection of the abdomen would require approximately ten (10) times as much radiation. Can you figure out why the chest radiograph gives so little radiation exposure?
Below is an outline (but not the only method outline) for interpreting a chest radiograph.
Reading the Chest X-Ray
a. In a dark room; on a view box (not using an overhead light) with unused viewboxes turned off.
b. "Hanging" the films:
1. Old films should always be viewed if available.
2. AP/PA placed facing you.
3. Lateral film with spine on the right.
2. Name, Age, Date & History
3. Film type
a. AP vs. PA (marker, vetebral body sign, clavicle and ant/post rib size, rib shape)
b. Supine vs. upright (marker, gastric bubble, less lung aeration due to diaphragm and more horizontal ribs)
4. Film Quality
5. Tubes, Wires &Lines
6. Soft Tissue & Bones
a. Neck, supraclavicular, axilla, chest wall, breast, abdomen & gastric bubble.
b. Humorous, shoulder joint, scapula, clavicle, vetebrae, ribs & sternum
(note bone density as well as structure)
a. Upper, middle & lower ant/middle/post;
a. Size: CT ratio (must evaluate AP/PA, rotation, diaphragm elevation-inspiration, mass, etc., thoracic cage, lungs- COPD, effusions)
b. Shape (chambers)
b. Costophrenic angle
a. Trachea & Bronchi
f. Behind Heart
12. Examine from close and far.
(From Jud Gash - Class of 1990)
The next radiographic examination of the chest will illustrate the difficulty of finding lesions in certain areas on the chest radiograph. It will also raise the question of the utility of the lateral examination. Although many radiographic examinations of the chest are done using both the PA and lateral as routine films, the trend in the United States is to do a single film examination (the PA only in individuals less than 40 years old) and then obtain a lateral view only if an abnormality is seen on the frontal projection. The reason for this trend is the benefit to risk ratio of the films. Much more information is available on the PA film (remember the lungs are superimposed on the lateral film) while the risk of the lateral film is much higher. Radiation exposure for the lateral radiograph is usually about three (3) times greater than the PA (most people are thicker side to side than they are front to back ............. although there are exceptions). Carefully study the PA film on this patient. Can you find any abnormality? Now look at the lateral view. Is there any abnormality on the lateral view which would match any problem you found on the PA view?
No matter how careful you are in analyzing a complex image such as a chest film, there will be errors. A review of chest radiograph interpretation by radiologists, (Herman et al) showed that 41% of chest radiograph reports contained potentially significant errors and 56% showed indeterminate disagreements. Among these significant errors and indeterminate disagreements, 78% were false negatives and 22% were false positives (Disagreements in Chest Roentgen Interpretation. Chest 68:3:278-282. September, 1975). This problem is not surprising, the interpretive process is highly complex and is dependent upon the psychophysical characteristics of the image and the perceptional and judgmental abilities of the interpreter. This kind of performance error is not limited to radiographic interpretation, other diagnostic tasks in medicine, such as taking histories and interpreting electrocardiograms have been reported to have a similar, high rate of disagreement.
Even when the densities are related to the patient, human observers frequently misinterpret real density variations by creating subjective images. Subjective images are formed by the interpreter combining multiple elements of a complex image pattern into a recognizable shape. Since most medical images involve a relatively thick body part, it is entirely possible that elements combined into a subjective image may not even be in the same object plane. You might be surprised how often this occurs in a complex image such as the chest radiograph in which the interpreter will connect elements of pulmonary arteries and rib densities into what appears to be a very believable pulmonary nodule or pulmonary cavity. An example of this phenomenon is shown in the simple image below. This image is actually composed of "v" shaped corners and notched circles. Most observers will use these individual picture elements to construct a white triangle which is superimposed upon the triangle outlined by the black line.
In the next image which is composed entirely of short line segments, most observers will create small circular structures at the intersection of the lines.
An even more striking example of image subjectivity is the variation in interpretation of the image printed below. This is a well-known image and I am sure you have probably seen it before. Most observers feel very strongly about what the image depicts but you will find that about one-half of the observer population sees an old lady facing your left with her head wrapped in a shawl while approximately the same number of observers will interpret the image as showing a younger woman viewed from behind her left ear. To get you oriented, the nose of the old woman becomes the left side of the young woman's jaw. Anyway the important part of this discussion is that humans tend to see whatever they want to see and it requires a great deal of self discipline to actually interpret the film correctly.
Another problem which is common in clinical radiographic interpretation is the phenomenon of subtle change. Comparison with previous radiographs greatly increases the sensitivity of detecting abnormal features which have occurred since the last examination. However, it is possible for changes to occur very gradually and this makes it difficult to detect a change if only the recent, previous films are compared. If a series of radiographs is done, it is sometimes surprising how little change occurs between each sequential image even though the change is very apparent if the first image is compared with the most recent image. In clinical radiography, you will frequently observe the film interpreter only reviewing the last previous image. This may result in a report that says "no change". If the interpreter included the first (oldest) available image in the comparison, the change might be considerably more apparent. When does the change become apparent in the image sequence below?
As an example of a common conventional radiographic study for interpretation practice, a four-view examination of the chest has been included below. Although you can view these on the computer screen, you will not be able to see the subtle detail that can be recognized if the full size image is examined on a light box.
As you can see, these images show a small pulmonary nodule projected in the right upper lung field. The nodule is not very dense and can be easily missed. In practice, nodules are frequently missed. In clinical practice, nodules such as this are frequently described in the radiographic report as being projected relative in position to a rib. Most of the time, the number of the posterior rib(s) is used to give the position. How would you describe the position of this nodule relative to ribs?
You might also find it helpful to carefully analyze the pleural fissures that can be identified on these radiographs. The horizontal fissure is visible. In some views, one or more oblique fissures are visible. From your observation of these pleural fissures, you should be able to predict whether the nodule lies in the right upper lobe or the right lower lobe. What do you think?
Remember, you are unlikely to recognize a subtle abnormality if you have never seen it (and have been shown what it is) before. If you have seen it at least once, the probability of recognizing it on a subsequent radiograph is greatly increased. Because of that, your ability to accurately interpret radiographs will increase very rapidly as you gain visual experience with these images.
In summary, the four rules below will help you make more accurate radiographic interpretations.
1) Do not trust what you think you see.
2) Be aware of and avoid common perception problems.
3) Confirm findings with other projections.
4) Use contrast material and/or other techniques to increase certainty.
1. Is there a clear understanding of the limitations of human observation?
2. Are film interpreters aware of common subjective images?
3. What is normal? Is calcification of the aortic arch in a 70 year old, normal or abnormal?
4. How can you detect a pulmonary nodule? How big must it be? How dense?
5. Do you clearly understand the anatomy of the pleural fissures?
1. Herman PG, et al. Disagreements in Chest Roentgen Interpretation. Chest 68:278-282.
2. Kundel HL, LaFollette PS. Visual Search Patterns and Experience with Radiological Images. Radiology 103:523-528.
Send us comments: Dr. David Adcock, DAVID@uscmed.sc.EDU.
Return to the School of
Return to the University of South Carolina homepage.
This page was last updated 26 August 2004.