How I Do It Marc S. Levine, MD Stephen E. Rubesin, MD Hans Herlinger, MD

Igor Laufer, MD

Double-Contrast Upper Gastrointestinal Examination: Technique and Interpretation1

T HE development of routine double- contrast techniques for examining

the upper gastrointestinal (GI) tract has dramatically improved our ability to di- agnose a variety of inflammatory and neoplastic diseases in the esophagus, stomach, and duodenum. Despite in- creasing acceptance of this technique, many radiologists still use convention- al single-contrast radiography as the primary modality for examining the upper GI tract. In a recent survey, more than 50% of the responding radiolo- gists at major academic institutions in- dicated that they did not perform dou- ble-contrast upper GI studies on a rou- tine basis (1). In many cases, double- contrast techniques are avoided because of lack of experience or train- ing in the technical aspects of perform- ing or interpreting these studies.

The issue of single- versus double- contrast technique has recently been overshadowed by another, more omi- nous development in the practice of GI radiology. Data indicate that there has been a gradual but steady decline in the total number of upper GI fluoro- scopic examinations p;rformed during the past decade (2,3). This trend can be attributed partly to the increased use of cross-sectional imaging modalities, such as computed tomography or ultra- sound, to evaluate GI problems. The greater availability of effective hista-

Index terms: Duodenum, abnormalities, 73.2 Duodenum, neoplasms, 73.3 Duodenum, ra-

diography, 73.123 Esophagus, abnormalities, 71.2 Esophagus, neoplasms, 71.3 Esophagus, radiography, 71.123 Radiology and radiolo- gists, How 1 Do I t Stomach, abnormalities, 72.2 Stomach, neoplasms, 72.3 Stomach, ra- diography, 72.123

Radiology 1988; 168:593-602

I From the Department of Radiology, Hospi- tal of the University of Pennsylvania, 3400 Spruce St, Philadelph~a, PA 19104. Received March 22, 1988; accepted and revision request- ed April 22; revision received May 13. Address reprint requests to M.S.L.

RSNA, 1988

mine-blocking agents has also changed the clinical approach to patients with dyspepsia or ulcer symptoms, since many of these patients are now treated empirically without further diagnostic evaluation. However, another major factor in the declining number of fluo- roscopic examinations is the increasing use of endoscopy as the initial screen- ing study in these patients. This trend has been spurred by a number of cor- relative radiologic-endoscopic studies from the gastroenterology literature in which endoscopy was found to be a more accurate diagnostic examination (4-6). However, these studies tended to be biased in favor of endoscopy, be- cause the primary authors were usually endoscopists. It also remains unclear whether the ability of endoscopy to de- pict radiographicaily missed leiions in the upper GI tract has any significant effect on the management or eventual outcome of the conditions in these pa- tients.

While fiberoptic endoscopy has been recognized as a highly accurate tech- nique for examining the upper GI tract, it is also an invasive technique with a small but measurable risk of gastric perforation or other complications. Furthermore, it is an expensive tech- nique, costing three to four times more in the United States than double-con- trast upper GI examinations. Because barium studies are safer and less expen- sive than endoscopy, radiologic evalua- tion of the upper GI tract remains a via- ble alternative as long as its radiologic accuracy approaches that of endoscopy for clinically significant disease. We be- lieve that a carefully performed dou- ble-contrast upper GI study provides the best opportunity for radiology to be competitive with endoscopy as a diag- nostic modality. A detailed description of the technical aspects of performing and interpreting these examinations is therefore presented.

TECHNIQUE

The routine double-contrast upper GI examination should be performed as

a biphasic study in which both double- contrast and single-contrast views of the esophagus, stomach, and duode- num are obtained. In the double-con- trast portion of the study, a series of maneuvers is required to achieve ade- quate gaseous distention of the lumen while a thin layer of high-density bari- um is spread on the mucosa. The dou- ble-contrast examination is facilitated by the routine use of hypotonic agents. Subsequently, the double-contrast study should be supplemented by prone or upright single-contrast views of the esophagus, stomach, and duode- num obtained with low-density barium and varying degrees of compression. The basic elements of the double-con- trast examination, including mucosal coating, gaseous di'stention, hypotonia, fluoroscopic maneuvers, and a step-by- step approach to the routine examina- tion, are discussed separately in the fol- lowing sections.

Mucosal Coating

Adequate mucosal coating is achieved in the upper GI tract by flow- ing a high-density barium suspension over the mucosa. Because mucosal coat- ing depends on the physical and chem- ical properties of the barium suspen- sion, the choice of barium directly af- fects the quality of the examination. In general, the best results are obtained with high-density (ie, 200% wt/vol), in- termediate-viscosity barium, such as E-Z-HD (E-Z-EM, Westbury, NY) or HD 85 (Lafayette Pharmacal, Lafayette, Ind). Adequate mucosal coating is usu- ally present when a thin, uniform white line is observed with fluoroscopy along the contour of the stomach. The quality of mucosal coating may also be judged on the basis of whether an areae gastricae pattern is visible in the stom- ach. With standard barium suspen- sions, however, areae gastricae can be detected in only about 70% of patients (7). Since visualization of these struc- tures depends on multiple factors, in- cluding the amount and viscosity of mucus in the stomach, failure to dem-

onstrate an areae gastricae pattern does not necessarily indicate that mucosal coating is inadequate.

Food or fluid in the stomach may also interfere with the quality of muco- sal coating. Patients are therefore in- structed to fast overnight before under- going a double-contrast examination, in order to minimize the amount of flu- id or secretions in the stomach. Patients should also be discouraged from smok- ing on the day of the examination, as cigarette smoke increases gastric secre- tions and impairs mucosal coating (8). Nevertheless, some patients with gas- tric outlet obstruction, gastroparesis, or hypersecretory states may have so much residual fluid or debris in the stomach that adequate mucosal coating cannot be achieved despite these pre- cautions. In such cases, a nasogastric tube may be used to aspirate fluid from the stomach before the examination.

Gaseous Distention

Gaseous distention is required to ef- face normal folds, in order to visualize the overlying mucosa. If adequate dis- tention is achieved, the surface features of the mucosa can be assessed both en face and in profile. If distention is in- adequate, however, prominent folds may obscure mucosal lesions, and bari- um trapped between the folds can mimic ulceration. By contrast, overdis- tention may impair mucosal coating or may obliterate abnormal folds or vari- ces. As a result, spot radiographs should be obtained at varying degrees of distention during the fluoroscopic examination.

The simplest method for introducing gas into the stomach and duodenum is to have the patient swallow efferves- cent tablets, granules, or powder that rapidly release 300-400 mL of carbon dioxide on contact with fluid in the stomach. Occasionally, however, bub- ble formation may occur in the esopha- gus or stomach as an undesirable arti- fact. Thus, most effervescent agents used for double-contrast examinations contain simethicone, an antifoaming agent, to minimize this problem.

While the stomach and duodenum are readily distended by effervescent agents, the patient must gulp a high- density barium suspension as quickly as possible to achieve adequate gaseous distention of the esophagus. With rapid swallowing, the peristaltic sequence is interrupted, and the esophagus be- comes hypotonic. As the patient gulps the high-density barium, swallowed air also distends the esophagus, contribut- ing further to the double-contrast ef- fect. However, some elderly or debili- tated patients may only be able to take small sips of barium, so that adequate esophageal distention cannot always be

achieved. If esophageal disease is be- lieved to be present, a soft rubber cath- eter may be inserted into the esopha- gus and a tube esophagogram obtained by insufflating air through the catheter as the patient swallows high-density barium (9). Similarly, gas may be intro- duced via a tube in the stomach or duo- denum when gaseous distention of these structures is inadequate on rou- tine double-contrast studies in patients who are believed to have disease.

Hypotonia

Glucagon is an effective hypotonic agent for double-contrast examinations of the upper GI tract. A standard dose of 0.1 mg administered intravenously produces adequate gastric hypotonia in most patients within 45 seconds of the injection (10). With the use of glucagon to decrease gastric peristalsis, it is pos- sible to obtain better double-contrast views of the antrum and body of the stomach in a relaxed, hypotonic state. Because glucagon also tends to delay gastric emptying, double-contrast views of the stomach can be obtained in most patients before it is obscured by overlapping loops of barium-filled duodenum or small bowel. At the same time, delayed filling of the duodenum may be an undesirable side effect of glucagon, prolonging the examination in some patients. Otherwise glucagon has virtually no side effects, except for the remote possibility of a hypersensi- tivity reaction. The only contraindica- tions to the use of glucagon are pheo- chromocytoma, insulinoma, and brittle, insulin-dependent diabetes.

In the evaluation of known gastric lesions (ie, healing of gastric ulcers), larger doses of glucagon (usually 1.0 mg) may be given intravenously to achieve greater levels of gastric hypo- tonia for a detailed double-contrast ex- amination. Similarly, a hyyotonic duo- denogram may be obtained in patients who are believed to have duodenal dis- ease, by administering 1.0 mg of gluca- gon to induce duodenal hypotonia af- ter barium has entered the duodenum.

While glucagon is a useful hypotonic agent for the stomach and duodenum, it has no effect on esophageal peristal- sis and will not produce better double- contrast views of the esophagus. How- ever, it has been shown that glucagon relaxes the lower esophageal sphincter ( 1 1) and therefore increases the fre- quency of spontaneous gastroesopha- geal reflux (12). Anticholinergic drugs, such as propantheline bromide (Pro- Banthine; Searle, San Juan, Puerto Rico), may be given to paralyze the esophagus, but these drugs are rarely used because of their marly side effects. Another anticholinergic drug, Busco- pan (nyoscine N-butyl bromide), pro-

duces effective but transient hypotonia of the esophagus, stomach, and duode- num. Because this agent relaxes the py- lorus, excellent double-contrast views of the duodenum can also be obtained. While Buscopan is a popular hypotonic agent in Europe, it is not available in the United States.

Maneuvers For performance of the double-con-

trast examination, a standard set of ma- neuvers is required to achieve adequate mucosal coating and gaseous disten- tion. The examination must be per- formed quickly, since overlapping loops of barium-filled small bowel may impair visualization of the stomach and duodenum. The major purpose of fluo- roscopy is to determine the volume of barium and gas, to assess mucosal coat- ing, and to insure accurate positioning and timing of spot radiographs. Be- cause the quality of mucosal coating tends to deteriorate rapidly during the fluoroscopic examination, repeated turning of the patient is required to manipulate the barium pool and obtain a fresh mucosal coating before each ex- posure.

While it is important to develop a standard routine for performance of the examination, additional maneuvers and/or spot radiographs may be re- quired if an abnormality is suspected at fluoroscopy. When the double-contrast portion of the study has been complet- ed, prone and upright single-contrast views should be obtained with com- pression to supplement the double- contrast examination. Because careful fluoroscopic positioning of the patient is required, the double-contrast study consists only of fluoroscopic spot radio- graphs without routine overhead radio- graphs.

The Routine Examination

1. A standard dose of 0.1 mg of gluca- gon diluted to 0.25 mL with sterile wa- ter is given intravenously.

2. 'The patient swallows one packet of effervescent granules or "fizzies" (ie, E-Z Gas; E-Z-EM) followed by 10 mL of water.

3. The patient gulps a cup of high- density barium (120 mL) as quickly as possible. The two best barium suspen- sions for this purpose are E-Z-HD (E-Z- EM) and HD 85 (Lafayette Pharmacal).

4. One three-on-one or two two-on- one upright spot radiographs are ob- tained in rapid sequence to demon- strate the entire length of the esopha- gus in a left posterior oblique (LPO) projection (Fig la). (All radiographic projections are indicated with respect to the tabletop.)

5. The table is brought to the hori-

594 Radiology September 1988

zontal position with the patient's back to the tabletop.

6. The patient turns to the right through a 360" circle back to the supine position. If mucosal coating is ade- quate, a frontal spot radiograph of the stomach is obtained for a double-con- trast view of the gastric antrum and body (Fig lb), since barium pools in the fundus in the supine position. If muco- sal coating is inadequate, the patient may be rotated one or more additional

turns before obtaining a radiograph. While some barium may spill into the duodenum during this maneuver, 360' rotation of the patient provides optimal coating of all mucosal surfaces of the stomach by the high-density barium. (If the patient cannot be rotated 360" because of age, debilitation, or other reasons, a gentle rocking maneuver can be performed to achieve adequate mu- cosal coating in most cases.) -

7. A spot radiograph of the antrum

and body of the stomach is obtained in the LPO position.

8. The patient turns into a right later- al position for a double-contrast view of the gastric cardia and fundus (Fig Ic). This view also permits visualiza- tion of the retrogastric area.

9. "Uphill" flow technique is per- formed by turning the patient onto the back and then slowly into an LPO posi- tion. This maneuver causes the barium pool to flow gradually from the antrum toward the body and fundus of the stomach. While the flow of barium is carefully monitored with fluoroscopy, four-on-one spot radiographs of the upper and lower body and antrum of the stomach are obtained. These radio- graphs are deliberately taken while the posterior wall of the body and antrum are covered by a thin layer of barium, to delineate shallow depressed or pro- truded lesions (13).

10. With the table in a semiupright position, the patient is turned into a right posterior oblique (RPO) position f& a double-contrast view of the upper body and lesser curvature of the stom- ach. "Downhill" flow technique may also be performed by observing the stomach with fluoroscopy as the pa- tient is slowly turned to the RPO posi- tion. This maneuver causes the barium pool to flow from the fundus along the lesser curvature and vosterior wall into the body and antrum: better delineat- ing lesions high on the lesser curvature

I of the stomach. Additional four-on-one

Figure 1. Normal double-contrast upper GI I examination. (a) Upright LPO view of esophagus. (b) Supine view of stomach. (c) Right lateral view of cardia and fundus. 1 (d) Semiupright LPO view of duodenum.

Volume 168 Number 3 Radiology 595

Figures 2,3. (2) Flow artifact in esophagus. (a) Residual layer of high-density barium obscures mucosal detail in esophagus. (b) Repeat dou- ble-contrast view obtained moments later shows reflux esophagitis in distal esophagus, with mucosal nodularity and inflammatory esopha- gogastric polyp (arrow) not visible on earlier image. (3) Value of prone esophagogram for distal esophagus. (a) Double-contrast radiograph shows no definite abnormality, although distal esophagus is not optimally distended. (b) Prone esophagogram from same examination shows hiatal hernia and adjacent peptic stricture (arrow) not visible on double-contrast image.

spot radiographs may be obtained dur- ing this maneuver.

11. By this time, barium usually has emptied into the duodenum. The pa- tient is turned into an LPO position for four-on-one double-contrast spot radio- graphs of the duodenal bulb and de- scending duodenum (Fig Id). If the bulb is inadequately distended with gas or obscured by barium, the table may need to be elevated to a fully up- right position. This maneuver causes barium to pool in the antrum or de- scending duodenum and air to rise into the duodenal bulb, which tends to as- sume a vertical configuration, so that adequate double-contrast views of the bulb usually can be obtained.

12. An upright LPO spot radiograph of the stomach is obtained for an addi- tional double-contrast view of the fun- dus.

13. The table is lowered to the hori- zontal position, and the patient drinks a low-density barium suspension in a prone RAO position. To assess esopha- geal peristalsis, the patient is instructed to take single swallows of barium, so that the entire peristaltic sequence can be evaluated with fluoroscopy. A single three-on-one or two two-on-one spot radiographs of the esophagus are then

exposed during continuous drinking, to permit optimal distention of the dis- tal esophagus and gastroesophageal junction.

14. With the patient in a prone or RAO position, four-on-one spot radio- graphs of the gastric antrum and duo- denal bulb are obtained with varying degrees of compression using an inflat- able balloon or other prone compres- sion device positioned beneath the pa- tient's abdomen. These views are im- portant for showing depressed or protruded lesions on the anterior wall of the stomach or duodenum.

15. The patient is turned onto the left side and then onto the back, so that barium pools in the fundus. The gastro- esophageal junction is then monitored with fluoroscopy as the patient turns slowly to the right, in order to demon- strate spontaneous gastroesophageal re- flux. A straight-leg raising maneuver or Valsalva maneuver can also be per- formed to elicit reflux.

16. The patient is returned to the right lateral position, and the table is fully elevated for four-on-one upright spot radiographs of the barium-filled stomach and duodenum with varying degrees of compression.

The filming sequence and purpose of

each radiograph are summarized in Ta- ble 1. With some experience, the fluo- roscopist should be able to complete the study in 2-4 minutes of fluoroscopy time and about 10 minutes of room time. When a lesion is suspected at flu- oroscopy, the double-contrast examina- tion should be tailored to demonstrate the lesion both en face and in profile. In many cases, flow technique and prone or upright compression views may also be needed to better delineate a suspected abnormality. Ultimately, each radiologist must develop his or her own routine for performing the ex- amination. While individual maneu- vers or filming sequences may vary, the end result should be a high-quality double-contrast examination.

Problems and Pitfalls

Esop/lngus.-Technical factors related to the amount of barium and gas in the esophagus may greatly affect the quali- ty of the double-contrast examination. Because double-contrast radiographs of the esophagus are obtained with the patient in an upright position, pooling of barium in the distal portion of the esophagus may obscure mucosal detail in this region. As barium enters the

596 Radiology September 1988

a. 6.

Figure 4. Value of combined double-contrast technique and prone compression for anteri- or wall lesions in stomach. (a) Supine double-contrast view shows ring shadow (arrow) in antrum. (b) Prone compression view demonstrates filling of anterior wall ulcer (arrow).

stomach, a residual layer of high-densi- ty barium in the esophagus may also obscure the mucosa (Fig 2a). When this "flow artifact" is observed at fluorosco- py, however, repeat views obtained moments later should better delineate mucosal lesions as the volume of resid- ual high-density barium on the mucosa decreases (Fig 2b) (14). At the same time, esophageal peristalsis causes the esophagus to collapse almost immedi- ately after passage of the barium bolus into the stomach. The fluoroscopist therefore must time the exposures to capture the esophagus during a rela- tively brief period of optimal disten- tion and coating. With some experi- ence, it is possible to obtain satisfactory double-contrast views of the esophagus in about 75445% of patients (15,16).

When esophageal disease is believed to be present, the cup of high-density barium may be split into two portions for a more detailed double-contrast ex- amination. By having the patient drink half of the barium during the usual LPO projection and half during an RPO projection, protruded or depressed le- sions in the esophagus may be evaluat- ed both en face and in profile. Alterna- tively, a tube esophagogram may be

obtained when the routine double-con- trast examination is inconclusive (9).

While mucosal disease is best evalu- ated with double-contrast technique, abnormalities of the longitudinal folds are better seen on mucosal relief views of the collapsed or partially collapsed esophagus. For example, esophageal varices may be effaced or even obliter- ated by esophageal distention, so that mucosal relief views should be ob- tained with the patient in the recum- bent position whenever varices are sus- pected on the basis of clinical findings. Various types of esophagitis may also be recognized due to the presence of thickened, irregular longitudinal folds in the collapsed esophagus.

When a ring or stricture is suspected in the distal portion of the esophagus, particular emphasis should be placed on prone single-contrast views of the esophagus during continuous drinking of a low-density barium suspension, to produce optimal distention of the distal esophagus and gastroesophageal junc- tion. Both Schatzki rings and peptic strictures that are not visible on dou- ble-contrast radiographs may be detect- ed when the esophagus is maximally distended by means of this technique

(Fig 3) (17,18). Stomach.-The proper volume of bar-

ium and gas are required for optimal double-contrast views of the stomach. A 120-mL cup of high-density barium is usually adequate for this purpose. Occasionally, however, barium may empty rapidly from the stomach de- s ~ i t e intravenous administration of glucagon. In such cases, additional bar- ium may be given to the patient in a re- cumbent LPO position to obtain better mucosal coating. Similarly, if the pa- tient belches during the examination, an additional packet of effervescent agent may be required to produce ade- quate gaseous distention. Conversely, the patient may be asked to belch if the stomach is overdistended with gas, since overdistention may efface abnor- mal folds associated with ulcers or ul- cer scars. Even when the proper amounts of barium and gas a;e present, retained fluid or debris in the stomach may impair mucosal coating, so that ad- ditional turning of the patient is re- quired to wash the mucosal surface and achieve adequate coating.

In patients who are relatively unre- sponsive to glucagon, rapid emptying of barium into the duodenum may cause portions of the stomach to be ob- scured by overlapping loops of barium- filled duodenum or small bowel. This overlap may particularly impair visual- ization of the distal antrum. If it is a frequent problem, the first part of the double-contrast study may be modified so that the table is returned to the hori- zontal position with the patient facing the table immediately after swallowing the high-density barium suspension. The patient then may be turned from the prone position onto the left side and back in order to obtain unobscured double-contrast views of the distal an- trum before barium spills into the duo- denum. The usual routine then may be followed for the rest of the examina- tion.

In other patients who are unusually sensitive to glucagon, gastric hypotonia may significantly delay spillage of bari- um into the duodenum, prolonging the examination. When this problem is en- countered, the patient may be asked to wait outside the examinihg room for 10 or 15 minutes until the glucagon effect has subsided. Since another study can be performed while the patient is wait- ing, the fluoroscopy schedule need not be delayed by a glucagon-sensitive per- son.

Most lesions on the anterior wall of the stomach will be shown by the com- bination of supine double-contrast and prone compression radiographs. For ex- ample, a shallow ulcer on the anterior wall may appear as a "ring shadow" on double-contrast views (Fig 4a), but prone compression views of the an-

Volume 168 Number 3 Radiology 597

trum should demonstrate filling of a discrete ulcer niche (Fig 4b). When such lesions are not clearly shown by these routine maneuvers, however, a double-contrast study of the anterior wall can be performed (19). In such cases, the patient is placed in a prone LAO Trendelenberg position. This ma- neuver causes the barium to pool in the fundus of the stomach with only a thin residual layer of barium on the anterior wall, so that double-contrast views of this region can be obtained.

In patients who have undergone par- tial gastrectomy (ie, Billroth I1 opera- tion), the altered anatomy of the stom- ach requires modification of the rou- tine double-contrast study. The patient initially should be given a larger intra- venous dose of glucagon (ie, 1.0 mg rather than 0.1 mg) to prevent rapid spillage of barium through the gas- troenterostomy into the proximal small bowel. At the same time, a smaller dose of effervescent agent (ie, half of a pack- et of "fizzies") is sufficient to achieve adequate gaseous distention of the gas- tric remnant. Also, the patient should not gulp a full cup of high-density bar- ium, since a large volume of barium in the gastric remnant may obscure muco- sal detail in this region. Instead, several swallows of barium in a semiupright position should permit adequate muco- sal coating of the anastomosis and gas- tric remnant (Fig 5). The reader is re- ferred to other articles for a more de- tailed discussion of double-contrast techniques for examining the postoper- ative stomach (20,21).

Duodenum.-Once the duodenum has filled with high-density barium, ade- quate double-contrast views of the duo- denal bulb and descending duodenum can usually be obtained by placing the patient in a recumbent, semiupright, or upright LPO position to distend the bulb with gas. When the duodenal bulb is located in a posterior position, the patient may be turned into a left lateral or even a prone LAO position to achieve adequate gaseous distention of the bulb. If, despite these maneuvers, satisfactory double-contrast views of the duodenum cannot be obtained, the fluoroscopist should proceed to the sin- gle-contrast portion of the study rather than prolong the double-contrast exam- ination, since it is almost always possi- ble to obtain adequate prone or upright compression views of the bulb.

INTERPRETATION

Normal Appearances Esopllagus.-The esophageal mucosa

normally has a smooth, featureless en face appearance on double-contrast ra- diographs. Occasionally, delicate trans- verse folds may be observed as a tran-

Figure 5. Normal double-contrast appear- I a&e of postoperative stomach (ie, after Bill- roth I1 operation).

sient phenomenon due to contraction of the longitudinally oriented muscu- laris mucosae (Fig 6) (22). However, the normal longitudinal folds are usually effaced or obliterated by gaseous dis- tention of the esophagus. While the - - presence of mucosal nodularity usually indicates an inflammatory or neoplastic process, the mucosa occasionally may have a nodular appearance in some middle-aged or elderly individuals with glycogenic acanthosis, a benign degenerative condition in which there is accumulation of glycogen within the squamous epithelial lining of the esophagus (Fig 7) (23,24).

Stomach.-While the appearance of the rugae in the stomach is important on conventional barium studies, these folds are effaced or obliterated by gas- eous distention on double-contras~ex- aminations. Instead, emphasis is placed on radiologic evaluation of the gastric mucosa. With adequate mucosal coat- ing, the normal areae gastricae are usu- ally most prominent in the antrum and body of the stomach. Enlargement or distortion of the areae gastricae may be related to the presence of superficial gastritis, hypersecretory states, intesti- nal metaplasia, or, rarely, early gastric cancer (25,26). However, further inves- tigation is required to understand bet- ter the significance of these variations in the surface pattern of the stomach.

Duodenum.-The duodenal bulb usu- ally has a smooth, featureless appear- ance on double-contrast radiographs. Occasionally, however, a lacy, reticular mucosal pattern may be observed in the duodenum as a normal variation (Fig 8a) (27). This appearance results from trapping of barium in intersecting grooves between tiny villous structures on the mucosa. In other patients, punc- tate collections of barium may be ob-

Figures 6,7. (6) Transverse folds in esoph- agus due to contraction of longitudinally oriented muscularis mucosae. (7) Glycogenic acanthosis with discrete plaques and nod- ules in midesophagus. Candida esophagitis could produce a similar appearance.

served in the duodenum both en face and in profile due to trapping of bari- um in normal mucosal pits (Fig 8b) (27). This finding is important because it can be mistaken radiographically for erosive duodenitis.

Depressed Lesions Esophagus.-Double-contrast radiog-

raphy is particularly valuable for the detection of shallow ulcers and ero- sions due to various types of esophagi- tis. In most cases, the underlying cause can be suggested on the basis of the ra- diographic findings and clinical histo- ry. In patients with reflux esophagitis, superficial ulceration in the distal por- tion of the esophagus may manifest on radiographs as multiple tiny collections of barium on the esophageal mucosa (Fig 9a) (28). Although some patients may have a more diffuse erosive esoph- agitis involving the distal one-third or two-thirds of the thoracic esophagus, there is almost always a continuous area of mucosal disease extending prox- imally from the gastroesophageal junc- tion. Thus, the presence of superficial ulceration in the midesophagus with distal esophageal sparing should indi- cate some other cause for the patient's

598 Radiology September 1988

esophagitis. Superficial ulcers predominantly in-

volving the midesophagus are usually caused by herpes or drug-induced esoph- agitis. In the early stages of herpes eso- phagitis, double-contrast radiographs may reveal discrete, superficial ulcers clustered together or widely separated by normal intervening mucosa (Fig 9b) (29,30). The ulcers are often surround- ed by a radiolucent halo of edematous mucosa. Similarly, drug-induced eso- phagitis may manifest as one or more discrete areas of superficial ulceration in the mid or, less commonly, distal esophagus, so that differentiation from herpes esophagitis may not be possible on the basis of the radiologic findings (Fig 9c) (31,32). However, a temporal relationship between ingestion of the offending medication (usually tetracy- cline or doxycycline) and the onset of esophagitis khould indicate the correct diagnosis.

Stomach.-Gastric erosions may be classified radiographically as complete or incomplete erosions. Most patients have complete or "varioliform" ero-

a. b. sions in which a punctate or slitlike Figure 8. Normal variations of duodenal mucosa. (a) Delicate reticular pattern in duodenal barium is surrounded by a bulb. (b) Tiny collections of barium in normal mucosal pits. Erosive duodenitis could pro- radiolucent halo of edematous mucosa duce a similar appearance. (Fig 10a) (33). Varioliform erosions are

most commonly found in the antrum of

Figure 9. Esophageal ulcers. (a) Reflux esophagitis with superficial ulceration in distal esophagus. (b) Herpes esophagitis with several shallow ulcers (arrows) separated by normal intervening mucosa. (Reprinted, with permission, from reference 29.) (c) Drug-induced esophagitis with discrete ulcers (arrows) in midesophagus. Although herpes 'sophagitis could produce identical findings (cf Fig 9h), this particular patient was receiving orally ad- ministered tetracycline.

the stomach a& are often aligned on rugae. While no etiologic significance is generally attributed to the shape or location of these gastric erosions, aspi- rin and other nonsteroidal antiinflam- matory drugs may produce linear or serpiginous erosions that tend to be clustered in the body of the stomach, on or near the greater curvature (Fig lob) (34). Because these lesions have a distinctive appearance on double-con- trast studies, a specific cause for the pa- tient's gastritis can be suggested on the basis of the radiographic findings.

Double-contrast radiography is a valuable technique for diagnosing gas- tric ulcers and for differentiating be- nign from malignant lesions (35,36). The majority of ulcers detected on dou- ble-contrast studies are less than 1 cm in size (36). While some benign ulcers are round and symmetric (Fig 1 la), oth- ers may have a rod-shaped or linear ap- pearaice. Enlarged areae gastricae are often observed surrounding the ulcer, due to edema and inflammation of the adjacent mucosa (Fig 12a) (36). Subse- quent ulcer healing may manifest not only as a decrease in the size of the ul- cer but as a change in its shape (Fig 12b). In most cases, ulcer healing pro- duces a radiographically visible ulcer scar with a central pit or depression, ra- diating folds, and/or retraction of the adjacent gastric wall (36,37). While a re- sidual depression could be mistaken for an active ulcer crater, the central por- tion of a reepithelialized scar some-

Volume 168 Number 3 Radiology 599

Figure 10. Erosive gastritis. (a) Typical varioli dominantly near greater curvature) due to rece

b. iform erosions in gastric antrum. (b) Linear and serpiginous erosions in body of stomach (pre- 'nt ingestion of indomethacin.

times can be differentiated from an ac- tive ulcer by the presence of normal areae gastricae within the scar (Fig 1 lb) (36).

Duodnlum.-Erosive duodenitis is di- agnosed less frequently than erosive gastritis on double-contrast studies be- cause of the difficulty in differentiating duodenal erosions from normal muco- sal pits (27). However, erosive duodeni- tis can be diagnosed when double-con- trast radiographs reveal one or more varioliform erosions in the proximal duodenum (Fig 13).

Unlike gastric ulcers, which occur primarily on the posterior wall of the stomach, as many as 50% of duodenal ulcers are located on the anterior wall (38). Since most double-contrast radio- graphs are obtained with the patient in a supine or supine oblique position, the anterior wall of the duodenum is not optimally coated with barium, so that anterior wall ulcers may be missed on the double-contrast portion of the study. For this reason, double-contrast views of the duodenum should be sup- plemented by prone compression views obtained with a low-density bari- um suspension to demonstrate ulcers on the anterior wall (38). In other pa- tients with anterior wall ulcers, a "ring shadow" may be detected on double- contrast radiographs due to a thin coat- ing of barium on the rim of a nonde- pendent ulcer. In such cases, the ulcer can almost always be filled with bari- um and diagnosed definitively on prone or upright compression views of the duodenum. Thus, biphasic tech- nique is particularly important in the evaluation of this area.

Protruded Lesions Esophng~ts.-CnirdidR esophagitis usu-

ally manifests on double-contrast ra- diographs as discrete plaquelike lesions

a. b. Figure 11. Benign gastric ulcer and ulcer scar in different patients. (a) Round, symmetric ulcer on posterior wall of stomach with smooth folds radiating to edge of ulcer crater. (b) Reepithelialized ulcer scar with radiating folds and normal areae gastricae (arrow) in central portion of scar.

with intervening segments of normal mucosa (Fig 14) (39). Because the le- sions have well-defined borders, they may be etched in white by a thin layer of barium trapped between the edge of the plaque and the adjacent mucosa. Glycogenic acanthosis may also pro- duce plaquelike lesions indistinguish- able from those of Candida esophagitis (Fig 7) (23,24). However, most patients with glycogenic acanthosis are asymp- tomatic, so that these conditions usual- ly can be differentiated on the basis of clinical findings.

Early esophageal cancers classically appear on radiographs as small pro- truded lesions less than 3.5 cm in diam- eter (40). They may be plaquelike le- sions (often with central ulceration) or flat, sessile polyps with a smooth or

slightly lobulated contour (Fig 15a) (40). Other early cancers may be super- ficial spreading lesions, causing local- ized nodularity or granularity of the mucosa (Fig 15b) (40). When a suspi- cious lesion is detected on double-con- trast radiographs, endoscopy and biop- sy should be performed for a definitive diagnosis.

Stomnch.-Early gastric cancers may be elevated, flat, or depressed lesions. The Japanese have developed an ele- gant system for classifying these can- cers based cn the predominant mor- phologic features of the lesion (41). Ele- vated lesions may be plaquelike or polypoid and occasionally may occupy a considerable burface area of the stom- ach without invading beyond the sub- mucosa. ~ n f o r t u n a t ~ l ~ , most patients

600 Radiology September 1988

a. b. Figure 13. Erosive duodenitis with varioli-

Figure 12. Benign gastric ulcer with healing. (Reprinted, with permission, from reference form erosions (arrows) in proximal duode- 36.) (a) Posterior wall ulcer (arrow) in upper body of stomach. Note enlarged, nodular areae num. gastricae adjacent to ulcer due to surrounding edema and inflammation. (b) Considerable ul- cer healing is seen 3 months later, with residual linear ulcer (arrow). Note return of normal areae gastricae in this region. (718 and 10.16 indicate dates of image acquisition.)

lated lymphoid hyperplasia of the duo- denum may be observed in patients who have no underlying disease.

14. 151. 15b. References

Figures 14,15. (14) C ~ l r d i d n esophagitis with discrete plaquelike lesions in the esophagus. 1. Ominsky SH, Margulis AR. Radiographic Note linear configuration of plaques. (15) Early esophageal cancer. (a) Small sessile polyp examination of the upper gastrointestinal (arrow) in midesophagus. (Courtesy of Seth N. Glick, MD, Philadelphia.) (b) Localized area tract: a survey of current techniques. Radi- of mucosal nodularity due to superficial spreading form of esophageal carcinoma. ology 1981; 139:ll-17.

2. Gelfand DW, Ott DJ. Declining volume of gastrointestinal fluoroscopies: a survey of 18 hospitals. Castrointest Radio1 1982;

in the United States with gastric carci- ized on double-contrast radiographs by 7:229-230. noma already have advanced lesions at angulated or polygonal 1-5-mm filling 3. Gelfand DW. Ott DJ. Chen YM. Decreas- the time of clinical presentation, so that defects that tend to be clustered near ing numbers of gastrointestinal studies:

early gastric cancer is unlikely to be de- the base of the bulb (Fig 16) (43). By report of data from 69 radiologic practices. AJR 1987; 149:1133-1136.

tected on double-contrast studies as hyperplastic Brunner glands 4. Martin TR, Vennes JA, Silvis SE, Ansel HJ. long as these examinations are per- and prominent lymphoid follicles in A comparison of upper gastrointestinal formed predominantly on symptomatic the duodenum may manifest as multi- endoscopy and radiography. J Clin Gas- patients (42). ple round, uniformly sized, 2-3-mm troenterol 1980; 2:21-25.

Duodet~utn.-Heterotopic gastric mu- nodules in the bulb and proximal duo- 5. Tedesco FJ, JW, Crisp WL* cosa in the duodenal bulb is observed denum (43). While diffuse lymphoid ny HF. "Skinny" upper gastrointestinal

endoscopy: the initial diagnostic tool-a as a relatively frequent finding on dou- hyperplasia of the duodenum and prospective comparison of upper gastroin- ble-contrast examinations of the upper small bowel is often associated with testinal endoscopy and radiology. J Clin GI tract. This abnormality is character- immunoglobulin deficiency states, iso- Gastroenterol 1980; 2:27-30.

Volume 168 Number 3 Radiology 601

Dooley CP, Larson AW, Stace NH, et al. Double-contrast barium meal and upper gastrointestinal endoscopy: a comparative study. Ann Intern Med 1984; 101:538-545. Rubesin SE, Herlinger H. The effect of barium suspension viscosity on the delin- eation of areae gastricae. AJR 1986; 146:35-38. Thoeni RF, Goldberg HI. The influence of smoking on coating of the gastric mu- cosa during double contrast examination of the stomach (abstr). Invest Radiol 1980; 15:388. Levine MS. Kressel HY, Laufer I, Her- linger H, Goren R. The tube esophagram: a technique for obtaining a detailed dou- ble-contrast examination of the esopha- gus. AJR 1984; 142:293-298. Miller RE, Chernish SM, Greenman GF, Maglinte DDT, Rosenak BD, Brunelle RL. Gastrointestinal response to minute doses of glucagon. Radiology 1982; 143:317-320. Hogan WJ, Dodds WJ, Hoke SE, Reid DP, Kalkhoff RK, Arndorfer RC. Effect of glucagon on esophageal motor function Gastroenterology 1975; 69:160-165. Haggar AM, Feczko PJ, Halpert RD, Simms SM. Spontaneous gastroesopha- geal reflux during double-contrast upper gastrointestinal radiography with gluca- gon. Gastrointest Radiol 1982; 7:319-321. Kikuchi Y, Levine MS, Laufer I, Herlinger H. Value of flow technique for double- contrast examination of the stomach. AJR 1986; 147:1183-1184. Maglinte DDT, Lappas JC, Chernish SM, Angler RT, Miller RE. Flow artifacts in double-contrast esophagography. Radiolo- gy 1985; 157:535-536. Balfe DM, Koehler RE, Weyman PJ, Baron RL, Reinus WR. Routine air-contrast esophagography during upper gastroin- testinal examinations. Radiology 1981; 139:739-741. Maglinte DDT, Schultheis TE, Krol KL, Caudill LD, Chernish SM, McCune WM. Survey of the esophagus during the upper gastrointestinal examination in 500 pa- tients. Radiology 1983; 147:65-70.

Chen YM. Ott DJ, Gelfand DW, Munitz HA. Multiphasic examination of the esoph- agogastric region for strictures, rings, and hiatal hernia: evaluation of the individual techniques. Gastrointest Radiol 1985; 10:311-316. Ott DJ, Chen YM, Wu WC, Gelfand DW, Munitz HA. Radiographic and endoscop- ic sensitivity in detecting lower esophage- al mucosal ring. AJR 1986; 147:261-265. Goldsmith MR, Paul RE, Poplack WE, Moore JP, Matsue H, Bloom S. Evaluation of routine double-contrast views of the anterior wall of the stomach. AJR 1976; 126:1159-1163. Gold RP, Seaman WB. The primary dou- ble-contrast examination of the postopera- tive stomach. Radiology 1977; 124:297- 305. Gohel VK, Laufer I. Double-contrast ex- amination of the postoperative stomach. Radiology 1978; 129:601-607. Gohel VK, Edell SL, Laufer I, Rhodes WH. Transverse folds in the human esophagus. Radiology 1978; 128:303-308. Click SN, Teplick SK, Goldstein J , Stead JA, Zitomer N. Glycogenic acanthosis of the esophagus. AJR 1982; 139:683-688. Chahremani GG, Rushovich AM. Glyco- genic acanthosis of the esophagus: radio- graphic and pathologic features. Gastroin- test Radiol 1984; 9:93-98. Koga M, Nakata H, Kiyonari H, lnakura M, Tanaka M, Nobe T. Minute mucosal patterns in gastric carcinoma. Radiology 1976; 120:199-201. Watanabe H, Magota S, Shiiba S, Ebata K, Yoshiya K. Coarse areae gastricae in the proximal body and fundus: a sign of gas- tric hypersecretion. Radiology 1983; 146:303-306. Bova JC, Kamath V, Tio FO, Peters JE, Coldstein HM. The normal mucosal sur- face pattern of the duodenal bulb: radio- logic-histologic correlation. AJR 1985; 145:735-738. Laufer I. Radiology of esophagitis. Radiol Clin North Am 1982; 20:687-699. Levine MS. Laufer I, Kressel HY, Fried- man HM. Herpes esophagitis. AJR 1981; 136:863-866. Levine MS, Loevner LA, Saul SH, Rubesin SE, Herlinger H, Laufer I. Herpes esoph- agitis: sensitivity of double-contrast esophagography. AJR 1988; 151:57-62.

Creteur V, Laufer I , Kressel HY, et al. Drug-induced esophagitis detected by double-contrast radiography. Radiology 1983; 147:365-368. Bova JG, Dutton NE, Goldstein HM, Ho- berman LJ. Medication-induced esopha- gitis: diagnosis by double contrast esoph- agogrdphy. AJR 1987; 148:731-732. Laufer I , Hamilton J, Mullens JE. Demon- stration of superficial gastric erosions by double contrast radiography. Gastroenter- ology 1975; 68:387-391. Levine MS. Verstandig A, Laufer I. Ser- piginous gastric erosions caused by aspi- rin and other nonsteroidal antiinflamma- tory drugs. AJR 1986; 146:31-34. Thompson G, Somers S, Stevenson GW. Benign gastric ulcer: a reliable radiologic diagnosis? AJR 1983; 141 :331-333. Levine MS, Creteur V, Kressel HY, Laufer I, Herlinger H. Benign gastric ulcers: di- agnosis and follow-up with double-con- trast radiography. Radiology 1987; 164:9- 13. Gelfand DW, Ott DJ. Gastric ulcer scars. Rad~ology 1981; 140:37-43. Laufer I. Double contrast gastrointestinal radiology. Philadelphia: Saunders, 1979; 345-353. Levine MS. Macones AJ, Laufer 1. Candi- da esophagitis: accuracy of radiographic diagnosis. Radiology 1985; 154:581-587. Levine MS. Dillon EC, Saul SH, Laufer I. Early esophageal cancer. AJR 1986; 146:507-512. Murakami T. Pathomorphological diag- nosis. In: Murakami T, ed. Early gastric cancer. Tokyo: University of Tokyo Press, 1971; 53-55. White RM, Levine MS. Enterline HT, Laufer I. Early gastric cancer: recent ex- perience. Radiology 1985; 155:25-27. Langkemper R, Hock AC, Dekker W, O p den Orth JO. Elevated lesions in the duo- denal bulb caused by heterotopic gastric mucosa. Radiology 1980; 137:621-624.

602 Radiology September 1988