Pyloric Carcinoma #

As pointed out by Castleman (l936), there was a time when it was widely believed that gastric carcinoma did not invade the duodenum. In most of the earlier text-books of surgery and surgical pathology it was stated that, while carcinoma of the stomach spread extensively in many directions by means of lymphatic and haematogenous metastases and through direct extension, it stopped abruptly at the pyloric ring. However, on reviewing the literature, Castleman (l936) collected 38 case reports of gastric carcinoma in which duodenal spread of the tumor had occurred. In the majority the extent of duodenal invasion was not more than 1.0 to 2.0 cm, but in isolated cases it could be as much as 5.0 cm. In view of these findings it was necessary to remove most of the first part of the duodenum in resections for gastric carcinoma.

In his own investigation Castleman (l936) examined microscopic preparations of 134 surgical and 65 autopsy specimens of pyloric carcinoma which had accumulated in his department during the previous 34 years. There was microscopic invasion of the duodenum in 6 (9 percent) of the autopsy, and in 15 (11 percent) of the surgical specimens. The extent of duodenal spread in the 21 cases varied from 4.0 mm to 2.3 cm. (In most of the surgical specimens carcinoma cells were found at the distal cut edge, i.e. the tumor had been transected). Spread usually occurred along the submucosa, often in the lymphatics, and rarely involved the mucosa. It was stated that spread had taken place deep to Brunner's glands (i.e. between Brunner's glands and the muscularis externa) and in the accompanying illustrations it appeared as if these glands were not involved. Usually duodenal infiltration was so slight that the surgeon had difficulty in palpating the thickened duodenal wall and it could only be detected microscopically. It was recommended that at least 3.0 cm of the duodenum be removed during gastric resection for carcinoma at the pylorus.

Dixon and Stevens (l936) drew attention to 6 cases of linitis plastica of the stomach in which histologically similar but separate, discrete lesions were found elsewhere in the gastrointestinal tract, namely in the oesophagus, jejunum, colon and rectum. No mention was made of direct extension into the duodenum in these cases. The authors had encountered 37 similar case reports in the literature.

Coller et al. (l941) studied the routes and extent of spread in 53 cases of gastric carcinoma. Duodenal involvement was seen in 14 cases (26.4 percent). In some it had occurred through direct extension of the tumor, and in others via submucosal and intermuscular lymphatic channels. The exact extent of duodenal spread was not mentioned.

Zinninger and Collins (l949) studied microscopic sections of gastric resection specimens in 36 cases of carcinoma of the stomach without any gross evidence of invasion of the duodenum. Microscopically 9 (25 percent) showed duodenal invasion by carcinoma. If only those cases were considered in which the lesion was situated within 5.0 cm of the pylorus, there were 9 cases out of 30 (30 percent) in which duodenal invasion had occurred. The extent of duodenal spread was as follows: in 2 cases it was 3.0 mm, in 4 cases 7.0 to 15.0 mm, in 2 cases 20.0 mm and in one case 60.0 mm. (The latter was an unusual case. It appeared to be a benign gastric ulcer, but was diagnosed microscopically as a mucoid scirrhous carcinoma with widespread extension. The cell types in the other 8 cases were not mentioned). The invasion of the duodenum seemed to be mostly by direct infiltration of the muscle or extension through the subserosal lymphatics. In 3 of the 9 cases spread had taken place in the submucosa. In some cases an isolated lymph node adjacent to the duodenum beyond the pylorus contained metastases without any carcinoma being found in the duodenal wall. No mention was made of carcinomatous infiltration in relation to Brunner's glands.

Eker (l951) studied the lymphatic spread of gastric carcinoma in 70 total and 100 partial gastrectomy specimens. Generally speaking extension into the oesophagus was much more frequent than spread into the duodenum. In the 70 specimens of total resection, 32.8 percent showed involvement of the oesophagus, while in the entire material of 170 cases only 2 tumors were found to have extended into the duodenum. It appeared as if the pyloric "sphincter" formed a barrier which prevented distal spread of carcinoma; similar circumstances did not pertain at the gastro-oesophageal junction.

Subsequently Eker and Efskind (l952) analyzed the frequency and extent of spread of gastric carcinoma within the different layers of the wall. For this purpose 42 total and 38 partial gastrectomy specimens were examined. Duodenal extension of the tumor was apparent in 14 of the 80 specimens. In all 14 cases the distance of duodenal infiltration beyond Brunner's glands was said to be short. Microscopically the cases were divided into adenocarcinomas, mucinous scirrhus carcinomas and colloid carcinomas.

It was found that gastric adenocarcinoma spread mainly in the mucosa and submucosa of the stomach, with spread decreasing gradually in the muscle layers toward the subserous layer. Spread in the subserous layer was slight, a fact which was especially obvious in the highly differentiated adenocarcinomas. In the 6 cases of gastric adenocarcinoma with duodenal extension, the spread in the mucous membrane stopped at the point where Brunner's glands commenced, i.e. at the line of demarcation between the stomach and duodenum. In other words, the duodenal mucosa was not involved by spread of gastric adenocarcinoma across the pylorus, while the other duodenal layers were.

In gastric mucinous scirrhus carcinomas conditions were found to be different. In most of these cases there was little spread in the gastric mucosa, most of the spread occurring in the submucosa and deeper layers of the gastric wall. In 6 of these cases duodenal extension of the tumor had occurred; in 4 of the 6 cases mucosal involvement stopped at the gastroduodenal junction, the duodenal mucosa being spared. In the other 2 cases the duodenal mucosa was involved for a distance of 2.0 mm in the first and 7.0 mm in the second case. All the other duodenal layers were involved in the spread. In gastric colloid carcinoma, the characteristics of spread in the gastric wall resembled those of adenocarcinoma in some cases, and mucinous scirrhus carcinoma in others. On the whole the greatest spread occurred in the submucous layer. In 2 cases infiltration of the duodenum had occurred, but further particulars of these were not given. The results seemed to justify the conclusion that the less differentiated carcinomas spread more readily to the duodenum than the highly differentiated tumors. In addition, the tendency to spread in the various layers of the wall was stronger in the stomach than in the duodenum.

In a wide-ranging study of gastrectomy specimens obtained from cases of gastric carcinoma, Eker and Efskind (1960) found the following: In 865 partially resected stomachs infiltration of the lower border was seen in 2.4 percent. In 256 totally resected stomachs infiltration of the lower border was present in 0.8 percent. Tumors involving the sphincteric cylinder (both adenocarcinoma and scirrhus carcinoma) infiltrated the lower border in 3.6 percent. The rarity with which distal extension of tumor occurred was ascribed to a barrier effect of the pyloric "sphincter", or to different peristaltic effects in the duodenum.

Majima et al. (l964) examined 833 subtotal resection specimens of cases of gastric carcinoma. Duodenal invasion was encountered in 151 cases (18.1 percent); the frequency of duodenal invasion depended on the location of the tumor in the stomach. If only those tumors within 1.0cm of the pyloric ring were considered, the incidence of duodenal invasion was 37.0 percent. In cases where the gastric tumor was located 3.0 to 4.0 cm proximal to the ring, the incidence was 4.1 percent. (Almost all these cases were associated with spread of carcinoma to other structures, e.g. the peritoneum and lymph nodes).

Of the cases with duodenal invasion, the serosa of the duodenum was involved in 115 (76.2 percent); in none of these cases could a continuous extension into the mucosal coat of the duodenum be demonstrated. However, in a few cases carcinoma of the prepyloric region extended directly into the duodenal submucosa at an early stage, in the absence of extragastric spread. In 92 percent of cases the extent of spread into the duodenum was 1.0 cm or less. In 6.6 percent it was between 1.0 and 2.0 cm, and in 0.7 percent it varied from 2.0 to 2.4 cm. In some of these cases the line of division of the duodenum was involved by carcinomatous infiltration, indicating that the duodenal invasion extended further than the distance given. The pattern of duodenal invasion was continuous in 54 percent of cases, discrete in 34 percent, and both continuous and discrete in 10 percent. In the 5 cases of duodenal mucosal invasion, spread had occurred via the lymphatic channels and not through continuous direct extension.

Fernet et al. (l965) described 7 cases of scirrhous carcinoma of the stomach (linitis plastica) which had spread within the wall of long segments of the alimentary tract, producing an appearance of leathered induration similar to that seen in the stomach. Occasionally the process extended downward as far as the sigmoid colon. This tubal spread was due to a neoplastic infiltrate in the intramural tissue spaces, producing a diffuse induration and separation of the mucosa, submucosa and muscularis. Spread occurred in the longitudinal lymph channels, mainly in the submucosa and subserosa, leaving the muscular layer relatively intact; the lymph spaces were clogged with small round or oval undifferentiated cells with a positive mucin reaction. Metastatic spread also occurred to the mesenteric lymph nodes and the intestinal serosa. It appeared as if the longitudinal spread in linitis plastica was quite different from that of glandular carcinoma, as the appearance of leathered induration was not observed in well- differentiated adenocarcinoma. Tubal spread seemed to occur only in linitus plastica or similar undifferentiated forms of carcinoma. No detailed description of conditions at the pyloro-duodenal junction in these cases was given.

Paramanandhan (l967) reviewed the literature on duodenal spread of gastric carcinoma for the period 1865 to 1965, and found great variations in the reported incidences. He studied the frequency and extent of duodenal invasion in 29 necropsy specimens of gastric adenocarcinoma. Invasion of the serosa and lymph nodes was apparent in all cases. The subpyloric group of lymph nodes was involved in 24 cases, 20 of these showing invasion of the duodenum. Thus duodenal spread had occurred in 20 of the 29 cases, giving an incidence of 68.9 percent. There was a discrepancy between the macroscopic and microscopic appearances, as only 3 of the 20 cases showed macroscopic evidence of duodenal invasion. Microscopically the spread often involved the duodenal mucosa and submucosa. The submucosal layer was the most frequently affected, being followed by the serosa and the other layers of the wall. In 17 cases duodenal invasion occurred chiefly via the lymphatics, but it was difficult to determine whether this was primarily via the submucous or via the subserosal lymph vessels. In 3 cases extension appeared to be by direct infiltration of sheets of anaplastic cells. The duodenal surface epithelium appeared to remain intact. Although they were often compressed by dilated lymphatics containing tumor cells, Brunner's glands appeared to be particularly resistant. In all cases there was some degree of lymphocytic infiltration of the duodenum, due to lymphatic stasis. As all cases with duodenal invasion showed involvement of the subpyloric group of glands, spread could be due to a downward or retrograde lymph flow resulting from a block of the nodes. The extent of duodenal spread varied from 0.5 cm to 22.6 cm. In more than half the cases no tumor tissue occurred beyond the first 3.0 cm of the duodenum, while in one case extension was seen for a distance of 22.6 cm.

In discussing the pathology and prognosis of carcinoma of the stomach, Hawley et al. (l970) analyzed the findings in 205 patients who underwent total or partial gastrectomy. Cases were divided into the following types: (1) well-differentiated adenocarcinoma; (2) poorly differentiated or anaplastic carcinoma; (3) linitis plastica type. One of the difficulties in the grading was the great variability in the degree of cellular differentiation in different parts of the same tumor. The amount of lymphocytic and plasma cell infiltration had a significant effect on the prognosis; five out of 7 patients with a heavy infiltration of such cells survived over 5 years. The feature with the worst influence on prognosis was lymph gland involvement. Linitis plastica was a type of carcinoma which had to be considered separately; here submucosal and muscular infiltration involved part or all of the stomach, sometimes without mucosal ulceration. Three cases of mucosal carcinoma, in which areas of malignant change were confined to the mucous membrane, were encountered; these were discovered incidentally during histological examinations of gastrectomy specimens removed for peptic ulceration. According to Öhman et al. (l972) gastric carcinomas could be graded histologically into the following types: (1) highly differentiated adenocarcinoma; (2) moderately differentiated adenocarcinoma; (3) poorly differentiated adenocarcinoma; (4) undifferentiated or expansive carcinoma; (5) linitis plastica; (6) gelatinous type. The highly differentiated adenocarcinomas had the best prognosis while there was no prognostic difference between the moderately and poorly differentiated and undifferentiated types. The prognostic significance of the degree of differentiation was probably linked to dissimilarities in longitudinal growth along the wall and the occurrence of lymph node metastases. Linitis plastica, which was also undifferentiated, spread in a different, infiltrative manner and had the worst prognosis of all. While all gastric carcinomas contained some amount of mucus, abundant mucinous masses were present in gelatinous carcinoma.

Ming (l973) reiterated that gastric carcinomas are highly infiltrative tumors, extending rapidly to the serosa. From here tumor cells may implant on other parts of the peritoneum; lymphatic and vascular spread is common. On gross examination of his cases the lesion often appeared to end sharply at the gastroduodenal junction, but it had been documented repeatedly that tumor extension into the duodenum was not uncommon. In many of those cases the duodenal invasion was mostly subserosal and the mucosa was not involved. The reason for the latter phenomenon was not known. Intramural spread of a highly infiltrative tumor to other parts of the small bowel and colon as reported by Fernet et al. (l965) appeared to be very rare.

Koehler et al. (l977) reviewed 111 consecutive cases with gastric adenocarcinoma. Spread of tumor to the duodenum was evident on gross inspection of the resected specimens in 7 cases (6 percent). In 6 patients (5 percent) radiographic abnormalities of the duodenum due to the transpyloric extension of gastric carcinoma were evident. (In the other case gastric outlet obstruction prevented radiological evaluation of the duodenum). In 2 patients without duodenal invasion, radiographic deformities of the duodenal bulb were due to inflammatory changes adjacent to the gastric tumor, and in one case the duodenal bulb was deformed due to metastatically enlarged adjacent lymph nodes. Microscopically the duodenum was involved in 20 cases (18 percent). Tumor cells were noted predominantly in the muscular and submucous layers of the duodenum, but mucosal invasion was noted occasionally. From the literature they concluded that in as many as half the cases with duodenal involvement, the tumor extended no more than 1.0 cm beyond the pylorus.

Joffe et al. (l977) described 4 cases with intraluminal filling defects of the duodenal bulb secondary to transpyloric prolapse of polypoid gastric carcinomas. It was pointed out that primary adenocarcinoma of the duodenum practically never arose in the bulb, and that intraluminal filling defects of the duodenum due to malignant lymphoma or metastatic malignancy were rare. Radiologically the cases had to be differentiated from prolapse of gastric mucosa (Chap. 38).

Present Investigations #

Patients and Methods #

In an effort to further elucidate duodenal spread and some other features of pyloric carcinoma, we did a retrospective study of 50 consecutive cases encountered in the radiological department over a period of 3 years. These patients had been referred for examination from the out-patient department during the ordinary course of events. At the time a barium radiographic study was the primary investigation for suspected gastric carcinoma. Because of the heavy work load and other reasons endoscopy was generally reserved for conditions such as haematemesis and radiologically negative dyspepsia, and for obtaining biopsies in cases of gastric ulceration. Moreover endoscopy was considered to have certain shortcomings in the diagnosis of gastric malignancy, especially in the pyloric region. Whereas the oral border of the lesion is usually clearly visualized endoscopically, narrowing and deformity may preclude passage of the instrument and the full extent of the lesion, as well as possible duodenal involvement, may not be evident. The barium suspension, on the other hand, usually traverses most narrowings to enter the duodenum, and the aboral border of the lesion may be seen.

Pyloric carcinoma was defined, somewhat arbitrarily, as a gastric carcinoma which, at the radiographic investigation, was seen to extend to the pyloric ring or to within 2.0 cm of the ring. Cases in which the aboral border of the lesion was more than 2.0 cm away from the ring were excluded. On average the duration of symptoms in these patients was from 4 to 12 months, i.e. they presented at a rather late stage of the disease; radiologically the lesions left little room for doubt. Cases of early gastric carcinoma, in which the lesion is limited to the mucosa, were not encountered, presumably because of the late presentations.

All cases were subsequently confirmed by endoscopic biopsy and/or operation, and the histology was obtained in 40 (Table 33.1). In 10 of the unresectable cases the surgeons deemed histological confirmation unnecessary. The 44 operations consisted of 24 Billroth II and 2 Billroth I partial gastrectomies, 16 palliative gastro-enterostomies and 2 exploratory laparotomies (in which widespread metastases precluded surgical palliation).

The following are representative cases:

Case Reports

Case 33.1 A.S., 63 year old male, presented with loss of weight, epigastric pain and haematemesis of one year's duration. Radiology showed a constant, lobulated filling defect 4.0 to 5.0 cm in length in the region of the pyloric sphincteric cylinder; aborally it extended to within one or two millimeters of the pyloric ring (Fig. 33.1). The pyloric aperture remained patent throughout. Mucosal folds were absent in the affected region. There was total absence of cyclical contraction and relaxation of the cylinder, the appearance remaining unchanged. The base of the duodenal bulb appeared normal. Emptying of barium suspension was not delayed to any appreciable extent.

At Billroth II partial gastrectomy an ulcerating malignant tumor measuring 6.0 x 2.5 cm was removed. Microscopy showed a poorly differentiated adenocarcinoma infiltrating into the serosa. The adjacent gastric mucosa showed chronic inflammatory cell infiltration and intestinal metaplasia. Both excision lines as well as the draining lymph glands were free of tumor cells.

Fig. 33.1.Case A.S. Lobulated filling defect in pyloric sphincteric cylinder. Mucosal folds and cyclical activity absent. Pyloric aperture patent. Base of duodenal bulb normal.

Case 33.2 J.B., 65 year old female, presented with a large epigastric mass and left supraclavicular Virchow Trosier lymphadenopathy. Radiology showed a constant, irregular narrowing of the distal 8.0 cm of the stomach, with an intraluminal filling defect extending to the pyloric ring. There was narrowing of the aperture and delayed emptying of barium suspension, indicating partial obstruction of the gastric outlet (Fig. 33.2). Food residues were present in the proximal stomach. Mucosal folds were absent in the affected region. There was total absence of cyclical activity of the pyloric sphincteric cylinder. The lesser curvature side of the base of the duodenal bulb showed a smooth, concave identation, apparently due to external pressure; in other respects the bulb appeared normal.

At laparotomy a large carcinoma of the distal part of the stomach, causing partial obstruction, was found. Aborally the tumor boundary was at the pyloric ring; no duodenal infiltration was noted. There was widespread infiltration in the region of the pancreas, in the gastrohepatic and gastrocolic omenta, as well as lymph gland extension. The tumor proved to be unresectable and a palliative gastro-enterostomy was done. Histology of involved lymph nodes showed a papillary adenocarcinoma.

Fig. 33.2.Case J.B. Irregular narrowing distal stomach with intraluminal filling defect. Mucosal folds and cyclical activity of pyloric sphincteric cylinder absent. Pyloric aperture narrowed. Food residues proximal stomach. Concave indentation base of bulb.

Case 33.3 C.J., 57 year old male, presented with loss of appetite and weight of 3 months' duration. Radiology showed a constant, irregular, constricting filling defect of the distal 5.0 cm of the stomach, with absent mucosal folds. Aborally, it extended as far as the pyloric ring (Fig. 33.3). There was a total absence of cyclical contraction and relaxation of the sphincteric cylinder; emptying of fluid barium was not delayed significantly. The base of the duodenal bulb showed shallow, concave indentations on either side of the pyloric aperture, the appearance being suggestive of external impressions. The duodenal "tail" appeared normal (Chap. 13).

Gastroscopy showed a friable, haemorrhagic, polypoid tumor in the corpus and "antrum" of the stomach; the pylorus and duodenum appeared normal. Endoscopic biopsy revealed a moderate to poorly differentiated adenocarcinoma. At laparotomy the aboral border of the tumor was at the pylorus; macroscopically the duodenum was not involved. On the oral side it extended to just below the gastro-oesophageal junction. The tumor was considered to be unresectable and the abdomen was closed.

Fig. 33.3.Case C.J. Irregular, constricting filling defect distal stomach. Mucosal folds and cyclical activity of sphincteric cylinder absent. Duondenal tail (arrow) normal. Concave indentations base of bulb on either side of pyloric aperture.

Case 33.4 A.S., 45 year old male, had nausea, early satiety and loss of weight for the previous 3 months. A hard epigastric mass was present. Radiology showed a constant, irregular, sharply defined filling defect 8.0 cm in length in the distal stomach, with absence of mucosal folds. It extended to the pyloric aperture, which was narrowed and not clearly visible most of the time (Fig. 33.4). Cyclical activity of the pyloric sphincteric cylinder was absent; emptying of fluid barium was delayed. The base of the duodenal bulb showed a deep, concave indentation with constant irregularity, suggestive of duodenal spread of the lesion.

Gastroscopy revealed a diffusely infiltrating, fungating mass in the pyloric region, extending to the pylorus and causing partial obstruction. As a result possible duodenal involvement could not be determined. Endoscopic biopsy showed a poorly differentiated, tubular adenocarcinoma. At laparotomy the tumor was confirmed. Metastases were present in the liver, it was considered to be unresectable, and a palliative gastro-enterostomy was done. The condition of the first part of the duodenum could not be determined precisely.

Fig. 33.4.Case A.S. Irregular filling defect distal stomach. Mucosal folds and cyclical activity of cylinder absent. Pyloric aperture unrecognizable. Concave indentation and irregularity of duodenal bulb.

Case 33.5 R.B., 60 year old female, presented with malaena of 5 months' duration, iron deficiency anaemia and an epigastric mass. Radiology showed a constant, lobulated and constricting filling defect in the distal 8.0 cm of the stomach, extending as far as the pyloric aperture, which remained patent. Emptying of fluid barium was not significantly delayed. Mucosal folds in the affected region and cyclical activity of the pyloric sphincteric cylinder were absent. The base of the duodenal bulb showed a smooth, regular, concave indentation, suggestive of external pressure rather than infiltration of the bulb itself (Fig. 33.5).

Gastroscopy revealed a large ulcerating carcinoma extending to the gastro-oesophageal junction on the lesser curvature. Endoscopic biopsy showed a poorly differentiated papillary adenocarcinoma. At laparotomy a large gastric carcinoma was found; there was infiltration of the transverse colon and gall bladder, with metastases in the liver. It was difficult to evaluate the duodenum precisely. The tumor was considered to be unresectable and palliative gastro-enterostomy was done.

Fig. 33.5.Case R.B. Lobulated and constricting filling defect distal stomach. Mucosal folds and cyclical activity of cylinder absent. Pyloric aperture patent. Smooth, concave indentation base of duodenal bulb.

Case 33.6 S.M., 49 year old male, presented with dyspepsia, loss of weight and postprandial vomiting of 7 months' duration. Radiology showed a constant, irregular narrowing with nodular filling defects in the distal 5.0 to 6.0 cm of the stomach; aborally it extended to the pyloric ring. Mucosal folds were not recognizable in the affected region. There was total absence of cyclical activity of the sphincteric cylinder. Emptying of fluid barium was not significantly delayed. A concave indentation of the base of the duodenal bulb, presumably due to external pressure, was seen (Fig. 33.6). In other respects the bulb appeared normal; the duodenal "tail" was normal.

Gastroscopy showed an infiltrating carcinoma of the "antrum", involving the lesser and greater curvatures and extending to within 8.0 cm of the gastro-oesophageal junction; the pyloric aperture could not be visualized. At Billroth II partial gastrectomy the tumor was removed; it extended to the pylorus but not into the duodenum. The transverse mesocolon was attached to it; there were no lymphatic or liver metastases. Microscopic examination revealed a poorly differentiated adenocarcinoma infiltrating deeply into the muscularis externa. No duodenal extension was seen. Draining lymph glands were normal. The surrounding gastric mucosa showed subacute gastritis.

Fig. 33.6. Case S.M. Irregular narrowing distal stomach. Mucosal folds and cyclical activity of cylinder absent. Pyloric aperture deformed. Concave indentation base of duodenal bulb. Duodenal "tail" (arrow) apparently unaffected.

Case 33.7 M.B., 68 year old female, presented with a large, hard epigastric mass. Radiology showed a constant narrowing of the distal 3.0 to 4.0 cm of the stomach, somewhat resembling partial contraction or spasm of the pyloric sphincteric cylinder (Fig. 33.7A). However, it contained a filling defect and mucosal folds were absent. While some degree of movement was seen, this was atypical, with total absence of cylical contraction and relaxation of the cylinder (Fig. 33.7B - E). The base of the duodenal bulb appeared normal. Emptying of barium suspension was not significantly delayed.

At Billroth II partial gastrectomy a large, ulcerated pyloric carcinoma, extending aborally as far as the pyloric ring, was removed. The duodenum appeared normal. Microscopy revealed a signet ring cell carcinoma extending as far as the lamina propria and not involving the muscularis. The duodenum was confirmed to be normal.

Fig. 33.7. A Case M.B. Narrowing distal 3.0 to 4.0 cm of stomach resembling partial contraction or spasm of sphincteric cylinder.

Fig. 33.7 B-E. Case M.B. Filling defect in narrowed region. Mucosal folds absent. Some movement evident but cyclical contraction and relaxation of sphincteric cylinder absent. Base of duodenal bulb appears normal.

Results #

Pyloric region #

Radiographic Anatomy of Sphincteric Cylinder In all 50 verified cases of pyloric carcinoma, alteration, deformity or destruction of the anatomical constituents of the pyloric sphincteric cylinder occurred to greater or lesser extent. In most cases the cylinder was totally unrecognizable; in 3 an appearance vaguely simulating the normal cylinder was seen (Fig. 33.7A). In some cases the process was of a mainly proliferative type, causing intraluminal filling defects (Fig. 33.1). In others it was mainly infiltrative, causing rigidity of the walls and "fixing" of the pyloric aperture in the patent or open position, with apparently a normal rate of emptying of fluid barium (Fig. 33.5). The process was of a mainly stenosing nature in other cases, with irregularity of the walls and narrowing of the lumen and pyloric aperture, causing partial or total obstruction (Fig. 33.2, 33.6). In many cases the tumor mass was ulcerated; not infrequently a combination of the above appearances was seen. In the few cases where the process somewhat resembled partial contraction of the normal cylinder, associated filling defects, mucosal destruction and atypical, restricted movements confirmed the diagnosis.
Of importance is the fact that the narrowing did not tally with any phase of the normal, cyclical contraction of the sphincteric cylinder. There was destruction of mucosal folds within the confines of the lesion in all cases.

Motility of Sphincteric Cylinder In all cases movements of the pyloric sphincteric cylinder were abolished or markedly altered. Only in 3 cases some movement occurred, but it was atypical (Fig. 33.7B); normal cyclical contraction and relaxation of the cylinder was absent in all.

Emptying of Liquid Barium In most cases there was no appreciable delay in the emptying of barium suspension; in a few cases incomplete or almost complete obstruction to the flow of liquid barium occurred.

Results #

Duodenum #

Radiographic Appearance of Duodenal Bulb Radiographically the base of the duodenal bulb, as well as its other walls and lumen appeared normal in 40 of 50 cases of pyloric carcinoma, showing no evidence of tumor extension (Fig. 33.1). In 2 of the remaining 10 cases a concave, smooth indentation of the base of the bulb, without other signs of duodenal involvement, was seen (Fig. 33.5). In 2 cases smaller, concave indentations of the base of the bulb on either side of the pyloric aperture were noted (Fig. 33.3). Between these indentations a small, triangular projection of barium extended into the aboral side of the pyloric aperture; this we have called the duodenal "tail" (Chap. 13). There was a suggestion of infiltration of the base of the duodenal bulb in 2 of the 10 cases. In the remaining 4 cases the bulb was not visualized adequately owing to technical factors (e.g. overlying partially obstructed stomach) and possible duodenal extension could not be determined.

Incidence of Duodenal Spread as Determined at Operation An attempt was made to determine duodenal involvement or non-involvement by inspection and palpation in 44 cases at operation. In 9 it was not possible to evaluate the pyloro-duodenal junction adequately on account of the tumor mass itself, infiltration of surrounding structures, liver metastases and/or ascites (Table 33.2); these cases were exluded. In other words, 35 cases could be evaluated.

In 23 evaluated Billroth II partial gastrectomies there was no evidence of macroscopic duodenal spread in 20; duodenal involvement was thought to be present in 3. In 2 Billroth I partial gastrectomies and 2 exploratory laparotomies no duodenal extension was evident. In 8 evaluated palliative gastroenterostomies the duodenum appeared to be uninvolved in 6 cases and duodenal extension was noted in 2. In other words, macroscopic duodenal spread was thought to be present in 5 of 35 cases evaluated at operation. (Histology subsequently disproved duodenal spread in one of the 5 cases; in the 2 cases with gastroenterostomies spread was not confirmed owing to lack of histological examination. In 3 of the 30 cases in which the duodenum appeared normal at operation, histology did show duodenal spread). It is confirmed that at operation the duodenum may appear normal macroscopically, in the presence of microscopic evidence of transpyloric spread of pyloric adenocarcinoma.

Incidence and Extent of Duodenal Spread as seen Microscopically Histological examination was obtained in 23 resection specimens. Of these 16 showed no evidence of duodenal extension. In 7 cases duodenal spread of pyloric carcinoma was seen microscopically; these will be described briefly.

Case Reports #

Case 33.8 J.P., 39 year old male. Radiology: 5.0cm long constricting pyloric filling defect. Smooth, concave indentation base of duodenal bulb (Fig. 33.8). Operation: Entire lesser curvature of stomach, part of greater curvature and pyloric region infiltrated. Serosal spread, lymph node and probable liver metastases. Duodenum appears normal. Billroth II. Gastric histology: well differentiated adenocarcinoma with widespread infiltration of gastric submucosa and mucosa. Duodenal histology: spread into submucosa and to lesser extent into mucosa of proximal 2.0 cm of duodenum.

Fig. 33.8 A-D. Case J.P. Constricting pyloric filling defect. Smooth, concave indentation base of duodenal bulb.

Case 33.9 A.F., 66 year old male. Radiology: 2.5 cm mass lesion partially obstructing pyloric aperture. Base of gas-filled duodenal bulb normal (Fig. 33.9). Operation: small ulcerated pyloric mass causing partial obstruction. Serosal spread and lymph node metastases. Spread into proximal 2.0 cm of duodenum. Billroth II. Gastric histology: well differentiated adenocarcinoma, intestinal type. Duodenal histology: spread into submucosa for 2.0 mm and into mucosa for 1.0 mm.

Fig. 33.9. Case A.F. Small mass lesion at pyloric aperture (arrow). Base of gas-filled duodenal bulb normal.

Case 33.10 S.F., 49 year old male. Radiology: 6.0 cm nodular and constricting pyloric filling defect. Smooth, concave indentation base of duodenal bulb (Fig. 33.10). Operation: Pyloric mass with serosal spread, adherent to pancreas. Lymph node metastases. Billroth II. Gastric histology: well differentiated adenocarcinoma. Duodenal histology: spread into serosa and muscularis of duodenum. Duodenal mucosa and Brunner's glands free of tumor cells.

Fig. 33.10.Case S.F. Nodular and constricting pyloric filling defect. Smooth, concave indentation base of duodenal bulb.

Case 33.11 K.B., 69 year old female. Radiology: 6.0 cm long constricting pyloric filling defect. Base of duodenal bulb normal (Fig. 33.11). Operation: Pyloric mass with serosal spread and lymph node metastases. Billroth II. Gastric histology: poorly differentiated adenocarcinoma. Marked infiltration of pyloric ring region. Duodenal histology: some tumor cells in muscularis and small blood vessels. Brunner's glands not affected.

Fig. 33.11.Case K.B. Constricting pyloric filling defect. Base of duodenal bulb normal.

Case 33.12 E.L.J., 44 year old male. Radiology: 7.0 cm long constricting pyloric filling defect. Base of duodenal bulb normal (Fig. 33.12). Operation: large mass pyloric region with serosal spread. Duodenum appears normal. Widespread lymphatic metastases. Coeliac and para-aortic glands involved. Metastases in transverse mesocolon and spleen. Billroth II. Gastric histology: poorly differentiated adenocarcinoma (mucinous type with signet ring cells). Duodenal histology: spread into submucosa up to commencement of Brunner's glands. No infiltration of Brunner's glands or superficial mucosa.

Fig. 33.12 A-D Case E.L.J. Constricting pyloric filling defect. Base of duodenal bulb normal.

Case 33.13 G.T., 58 year old female. Radiology: 8.0 cm long nodular, constricting filling defect pyloric region. Base of duodenal bulb normal. Operation: large tumor mass with serosal extension and lymph node metastases. Billroth II. Gastric histology: mucinous adenocarcinoma extending through all layers of gastric wall. Duodenal histology: spread into first 2.0 cm of duodenal serosa, muscularis, submucosa and mucosa. No comment on Brunner's glands.

Case 33.14 A.A., 39 year old female. Radiology: ulcerated pyloric filling defect. Base of duodenal bulb normal. Operation: fungating pyloric mass with serosal extension and lymph node metastases. Duodenum appears normal. Billroth II. Gastric histology: poorly differentiated adenocarcinoma with signet ring cells. Duodenal histology: spread into duodenal mucosa for distance of 5.0 cm. Brunner's glands free of tumor cells.

As far as the histologic evaluation of duodenal spread of pyloric adenocarcinoma is concerned, it should be kept in mind that resections were not performed in cases with very extensive disease; in these cases specimens were not available for examination, and the extent of duodenal spread remains undetermined.

A striking feature in the 7 cases with microscopic evidence of duodenal spread is the fact that spread occurred for very short distances. In 3 cases it involved not more than the proximal 2.0 cm of the duodenum; in one it was found to extend to the commencement of Brunner's glands. In 2 cases the extent of spread was not mentioned, but judging by the descriptions it appeared to be very short; only in one case did spread extend into the duodenum for a distance of 5.0 cm.

Duodenal Spread in Relation to Gastric and Extragastric Extension

All 7 cases with microscopic evidence of duodenal spread had serosal involvement on the gastric side with extragastric extension to draining lymph nodes; one of the cases also had extension to the transverse mesocolon and spleen.
Most of the 16 cases without microscopic evidence of duodenal spread had widespread gastric and extragastric extension, e.g. to the serosa, omenta, draining lymph glands, liver and pancreas.

Discussion #

Role of Radiography #

Radiography is highly sensitive and specific in the diagnosis of pyloric adenocarcinoma. In the present series of 50 cases the diagnosis was confirmed by endoscopy and/or operation in all. In only one case was a differential diagnosis considered; in this exceptional case the lesion had to be differentiated from malignant lymphoma (Chap. 34), cicatrization and obstruction due to benign pyloric ulceration (Chap. 29), acid corrosive stricture (Chap. 39) and eosinophilic infiltration. Malignant processes with rather different radiographic appearances are Carman ulcer, linitus plastica and early gastric carcinoma. Radiography also plays a useful role in the assessment of individual cases, as the state of the pyloric aperture and other features not readily visible endoscopically, may be determined.

Emptying of Solids and Liquids #

It is surmized that destruction of the various anatomical constituents of the pyloric sphincteric cylinder, seen radiographically in all 50 cases, with associated lack of cyclical contraction and relaxation of the cylinder, will result in absent or severely impaired propulsion and trituration of solids (Chap. 18). Mucosal destruction and consequent lack of mucosal fold movements should aggravate these effects (Chap. 13).
Radiography shows that in the absence of actual obstruction, i.e. in the majority of cases, emptying of liquid barium is not usually delayed to an appreciable extent.

Duodenal Spread #

Radiography is less accurate as far as evaluation of the duodenum in cases of pyloric adenocarcinoma is concerned. However, the smooth and regular appearance of the base of the duodenal bulb, seen in 40 of the 50 cases, is of importance. This "normal" appearance, seen in association with a constant filling defect involving the pyloric sphincteric cylinder, is corroborative evidence of pyloric carcinoma. The same holds true for smooth, concave indentations of the base of the bulb; this appearance, seen in 4 cases, is presumably due to external indentation by the pyloric tumor. In conditions such as malignant lymphoma, eosinophilic granuloma, corrosive stricture and tuberculosis, the duodenal bulb is usually deformed. Even in the presence of histologically proven duodenal spread, the duodenum may appear normal radiologically. In 20 cases of transpyloric extension, Koehler et al. (l977) noted duodenal abnormalities in 6, 14 appearing normal. In our 7 cases with microscopic spread duodenal abnormalities were seen in 2 only.

Macroscopically, at operation, duodenal spread may not be evident (Castleman l936; Paramanandhan l967). In 3 of our 7 cases the duodenum was considered to be normal at operation. In one duodenal spread was noted; in 3 the condition of the duodenum was not commented on.

The only unequivocal evidence of duodenal spread of pyloric adenocarcinoma is obtained by microscopic examination (Castleman l936). In the following cases microscopic duodenal spread of pyloric carcinoma was present in all:

Extent of Duodenal Spread as Determined Microscopically #

In the majority of 38 cases culled from the literature, Castleman (l936) found that the extent of duodenal invasion was not more than 1.0 to 2.0 cm; in isolated cases it could be 5.0 cm. In 6 of his autopsy specimens the extent of duodenal spread varied from 4.0 mm to 2.3 cm.

In 8 of 9 cases Zinninger and Collins (l949) found the extent of duodenal invasion to vary from 3.0 mm to 2.0 cm; in an unusual case it was 6.0cm. In 14 cases Eker and Efskind (l952) found the distance of duodenal infiltration beyond Brunner's glands to be short; 6 of these were adenocarcinomas and in these cases the invasion stopped at Brunner's glands.

In 92 percent of 151 cases Majima et al. (l964) found duodenal spread to be 1.0 cm or less; in a further 6 percent it was 2.0 cm or less. Paramanandhan (l967) examined 29 cases. In more than half no tumor tissue was detected beyond the first 3.0 cm of the duodenum; in one it extended for 22.6 cm. Koehler et al. (l977) concluded than in as many as half the cases the tumor extended no more than 1.0 cm beyond the pylorus.

In 3 of our 7 cases duodenal extension was 2.0 cm or less; in 3 it appeared to extend as far as the commencement of Brunner's glands, and in one it was 5.0 cm.

Route of Spread #

According to Castleman (l936) pyloric adenocarcinoma usually spreads along the submucosa into the duodenum, and rarely along the mucosa; spread often occurs in lymphatics. In 14 cases Coller et al. (l941) found that direct spread along the walls occurred in some cases, and in others spread was via submucosal and intermuscular lymphatic channels.

Zinninger and Collins (l949) stated that duodenal spread occurred mostly through direct infiltration of the muscular layers or through subserosal lymphatics; in 3 of 9 cases it also occurred in the submucosa. Spread may also take place via lymphatics to isolated lymph nodes beyond the pylorus, without the duodenal walls being involved.

In 6 cases of gastric adenocarcinoma Eker and Efskind (l952) noted duodenal spread along all layers of the walls, but in the mucous membrane it stopped at the commencement of Brunner's glands. In 4 of 6 cases of mucinous scirrhus carcinoma the duodenal mucosa was spared; in the remaining 2 cases the duodenal mucosa was involved for distances of 2.0 mm and 7.0 mm (all the other layers were also involved). In 2 cases of gastric colloid carcinoma the greatest spread occurred in the submucosa.

In 151 cases Majima et al. (l964) found serosal spread to the duodenum in 115; in a few cases there was direct spread in the submucosa. In none could continuous extension in the mucosa be demonstrated; in 5 cases of duodenal mucosal invasion, spread had occurred via lymphatic channels. In 29 necropsy specimens Paramanandhan (l967) found invasion of the duodenal submucosa the most frequent route of spread, followed by the serosa and other layers of the walls; the duodenal surface epithelium appeared to remain intact. Spread also commonly occurred to the subpyloric and other lymph nodes.

Ming (l973) stated that in most cases duodenal invasion occurred subserosally and that the duodenal mucosa was not involved; the reason for this phenomenon was not known. In 20 cases Koehler et al. (l977) found tumor cells predominantly in the duodenal muscular and submucous layers; mucosal invasion was noted occasionally.

Three of our 7 cases were well differentiated adenocarcinomas. In 2 spread had taken place along the submucosa and to a lesser extent in the mucosa; in one spread occurred along the serosa and muscular layers. Three cases were poorly differentiated adenocarcinomas; in one spread was seen in the muscular layer and small blood vessels, in one in the submucosa but not in the superficial mucosa, and in one in the mucosa. One case was a mucinous adenocarcinoma; in this case spread occurred in all layers of the wall.

Linitis plastica was a type of carcinoma in which tubal spread occurred, and it had to be considered separately (Fernet et al. l965; Hawley et al l970; Öman et al. l972).

Does a Duodenal Barrier Exist?

From the foregoing it appears that in the majority of cases duodenal spread of pyloric adenocarcinoma is limited to the proximal 2.0 to 3.0 cm of its first part; this is seen even in the presence of widespread extension in the stomach itself, contiguous spread to other organs, e.g. the pancreas and widespread haematogenous and lymphatic metastases.

The question arises whether a barrier exists in the first part of the duodenum (as opposed to a barrier at the pylorus), preventing further downward spread. It is assumed that a possible barrier will have to be of an anatomical or a biochemical nature. Paramanandhan (l967), while commenting on the infrequency with which pyloric adenocarcinoma invades the duodenal mucosa (it has less respect for the deeper components of the duodenal wall), stated that the following factors had to be taken into account in explaining the phenomenon: (1) scarcity of lymphatic connections between the stomach and duodenum; (2) normal upward flow of lymph from duodenum to stomach; (3) alkalinity of duodenum; (4) spasmodic contraction of the pyloric "sphincter". Duodenal Spread and Brunner's Glands

Castleman (l936) did not describe his 21 cases with microscopic duodenal spread individually. He noted that spread had generally taken place deep to Brunner's glands, i.e. between the glands and the muscularis externa. On examining the accompanying illustrations, it appears if Brunner's glands were not involved.

In 14 cases with duodenal extension Eker and Efskind (l952) stated that the distance of duodenal spread beyond Brunner's glands was short in all. Six of the cases were gastric adenocarcinomas, and in these spread in the duodenal mucous membrane stopped at the point where Brunner's glands commenced. (In the remaining 8 cases of mucinous scirrhus and colloid carcinomas no mention was made of Brunner's glands).

Paramanandhan (l967) noted in 29 necropsy specimens that Brunner's glands were often compressed by dilated lymphatics containing tumor cells. He stated that the glands themselves appeared to be particularly resistant to gastric carcinoma spreading across the pylorus.

In our 7 cases Brunner's glands were stated to be free of tumor cells in 4 (3 cases of poorly and one of well differentiated adenocarcinoma). In 3 cases no comment was given on Brunner's glands; the extent of spread in these cases was 2.0 cm, 2.0 cm and 2.0 mm respectively. It is concluded that there appears to be a possibility that Brunner's glands are resistant to the spread of gastric adenocarcinoma across the pylorus.

Epidermal Growth Factor in Gastric Carcinoma

Epidermal growth factor (EGF) and homologous alpha-tumor growth factor are mitogenic polypeptides which act by binding to epidermal growth factor receptors (EGFR). Pfeiffer et al. (l990) investigated whether increased production of EGF or increased density of EGFR may occur in gastric carcinomas as compared with normal mucosal tissue. The EGF binding capacity was found to be significantly higher in carcinomas than in normal mucosa. In 15 cases of gastric carcinoma EGFR showed an increase in 9, a decrease in 2 and no change as compared with normals in 3. In a case of mucinous adenocarcinoma there was an extreme, 320-fold increase of EGFR. In 2 of 22 carcinomas EGF activity was increased. It was concluded that a relative overexpression of EGFR occurs in some cases of gastric carcinoma.

Lee et al. (l989, l991) using a retrospective immunohistochemical evaluation for EGF and EGFR in 167 cases of benign and malignant gastric disease, made similar observations. In approximately 20 percent of gastric carcinomas increased amounts of EGF and EGFR were detected in the tumor as well as the adjacent mucosa, the intratumor values being significantly higher than the mucosal. Overexpression of EGF and EGFR was found to identify a definite subgroup of gastric carcinomas; tumors in this subgroup appeared to be deeply invasive.

Brunner's Glands and Epidermal Growth Factor

Epidermal growth factor is described as a polypeptide present in human urine, probably playing some role in the regulation of cell growth; it was originally found to inhibit gastric acid secretion and was called urogastrone (Pfeiffer et al. l990). An EGF homologous protein called alpha tumor growth factor may be produced in high amounts by transformed cells. It binds to the EGF receptor and has activities similar to EGF (Pfeiffer et al. l990).

Lee et al. (l989, 1991) and Pfeiffer et al. (l990) seem to implicate EGF in the pathogenesis of a subgroup of gastric carcinomas. These authors did not mention possible involvement of Brunner's glands. It has been shown (Chap. 4) that in man Brunner's glands secrete EGF (Elder et al l978; Heitz et al. l978). EGF is also produced in the submandibular salivary glands (Hollenberg l979). Mouse EGF and human urogastrone are closely related structures and seem to be identical in biologic activity; data on human EGF suggest that it is urogastrone. Experimentally both substances are powerful mitogens (Elder et al. l978), increasing the synthesis and contents of DNA and RNA in the gastroduodenal mucosa (Dembinsky et al. l982).

Kirkegaard et al. (l98l, l983) demonstrated a dense network of vasoactive intestinal polypeptide (VIP) immunoreactive nerve fibres around the acini of Brunner's glands. According to Ferri et al. (l984) few of these fibres enter the glands to reach the acinar cells in cases of gastric carcinoma. Skov Olsen et al. (l985) pointed out that cholinergic, adrenergic and VIP-containing nerves innervate and thereby influence secretion from Brunner's glands; VIP was found to increase the secretion and total output of EGF.

Whether there is a relationship between Brunner's glands and some types of pyloric adenocarcinoma is speculative. Too few cases have been studied for firm conclusions; however, available evidence suggests that an association cannot be excluded. On the one hand Brunner's glands seem to escape direct infiltration in duodenal spread of some cases of pyloric adenocarcinoma. On the other hand overexpression of EGF, which is produced by Brunner's glands (and by the submandibular salivary glands) may occur in subgroups of gastric carcinoma.

The fate of APUD cells in the pyloric mucosal zone in cases of pyloric carcinoma, is not known.

Conclusion: Pyloric Carcinoma, Brunner’s Glands and Epidermal Growth Factor #

It appears that Brunner's glands of the duodenum may be resistant to pyloric adenocarcinoma spreading across the pylorus (Eker and Efskind 1952, Paramanandhan 1967, Keet 1993). However, further patho-anatomical studies would be necessary to confirm this finding.

Overexpression of EGF, which is produced by Brunner's glands (and by the submandibular salivary glands) may occur in subgroups of gastric carcinoma. The exact relationships between pyloric carcinoma, Brunner's glands and EGF remain to be elucidated.

References #

1. Castleman B. Extension of gastric carcinoma into the duodenum. Ann Surg l936, 103, 348-352.
2. Coller FA, Kay EB, McIntyre RS. Regional lymphatic metastases of carcinoma of the stomach. Arch Surg l941, 43, 748-761.
3. Dembinski A, Gregory H, Konturek SJ, et al. Trophic action of epidermal growth factor on the pancreas and gastroduodenal mucosa in rats. J Physiol l982, 325, 35-42.
4. Dixon CF, Stevens GA. Carcinoma of linitis plastica type involving the intestine. Ann Surg l936, 103, 263-272.
5. Eker R. Carcinomas of the stomach: investigation of the lymphatic spread from gastric carcinomas after total and partial gastrectomy. Acta Chir Scand l951, 101, 112-126.
6. Eker R. Efskind J. Investigations on the intramural spread of gastric carcinoma. Acta Path Microbiol Scand l952, 30, 371-383.
7. Eker R, Efskind J. The pathology and prognosis of gastric carcinoma. Acta Chir Scand l960, Suppl 264, 1-182.
8. Elder JB, Williams G, Lacey E, et al. Cellular localization of human urogastrone- epidermal growth factor. Nature l978, 271, 466-467.
9. Fernet P, Azar HA, Stout AP. Intramural (tubal) spread of linitis plastica along the alimentary tract. Gastroenterology l965, 48, 419-424.
10. Ferri GL, Botti P, Biliotti G, et al. VIP, substance P and metenkephalin immunoreactive innervation of the human gastroduodenal mucosa and Brunner's glands. Gut l984, 25, 948-952.
11. Hawley PR, Westerholm P, Morson BC. Pathology and prognosis of carcinoma of the stomach. Brit J Surg l970, 57, 877-883.
12. Heitz PU, Kasper M, Van Noorden S, et al. Immunohistochemical localization of urogastrone to human duodenal and submandibular glands. Gut l978, l9, 408-413.
13. Hollenberg MD. Epidermal growth factor-urogastrone, a polypeptide acquiring hormonal status. Vitam Horm l979, 37, 69-110.
14. Joffe N, Goldman H, Antonioli DA. Transpyloric prolapse of polypoid gastric carcinoma. Gastroenterology l977, 72, 1326-1330.
15. Kirkegaard P, Lundberg JM, Poulsen SS, et al. Vasoactive intestinal peptidergic nerves and Brunner's gland secretion in the rat. Gastroenterology l98l, 81, 872-878.
16. Kirkegaard P, Skov Olsen P, Poulsen SS, et al. Exocrine secretion of epidermal growth factor from Brunner's glands: stimulation by VIP and acetylcholine. Regulatory Peptides l983, 7, 367-372.
17. Koehler RE, Hanelin LG, Laing FC, et al. Invasion of the duodenum by carcinoma of the stomach. Amer J Roent Rad Ther Nucl Med l977, 128, 201-205.
18. Lee EY, Wang TC, Clouse RE, et al. Mucosal thickening adjacent to gastric malignancy: association with epidermal growth factor. Mod Pathol l989, 2, 397-402.
19. Lee EY, Wang TC, Clouse RE. Gastric carcinoma, epidermal growth factor and epidermal growth factor receptor (letter). Gastroenterology l99l, 100, 289.
20. Majima S, Yamagucki I, Yoshida K, et al. Duodenal extension of carcinoma of the stomach. Tohoku J Exp Med l964, 83, 159-167.
21. Ming SC. Tumors of the esophagus and stomach. Atlas of Tumor Pathology, Fasc 7, US Armed Forces Inst Path, Washington DC l973, pp 200-201, pp 231-240.
22. Öhman U, Wetterfors J, Moberg A. Histologic grading of gastric cancer. Acta Chir Scand l972, 138, 384-390.
23. Paramanandhan TL. The duodenal spread of gastric carcinoma. Brit J Surg l967, 54, 169-174.
24. Pfeiffer A, Rothbauer E, Wiebecke B, et al. Increased epidermal growth factor receptors in gastric carcinoma. Gastroenterology l990, 98, 961-967.
25. Skov Olsen P, Poulsen SS, Kirkegaard P. Adrenergic effect on secretion of epidermal growth factor from Brunner's glands. Gut l985, 26, 920- 927.
26. Zinninger MM, Collins WT. Extension of carcinoma of the stomach into the duodenum and esophagus. Ann Surg l949, 130, 557-566.