Local recurrence of gastric adenocarcinomas after gastrectomy

Journal of Surgical Oncology 18:47-53 (1981)

Local Recurrence of Gastric Adenocarcinomas After Castrectomy ........................................................................................................... ...........................................................................................................

DIMITRI N. PAPACHRISTOU, MD, FRCS ( c ) , and JOSEPH G. FORTNER, MD, FACS

A study of 257 gastric cancer patients treated with gastrectomy and followed to their death showed that (a) the incidence of local recurrence in the field of gastrectomy was 25%; (b) from the three types of gastrectomy used, extended total gastrectomy resulted in the lowest incidence of recurrence, followed by subtotal and total; (c) recurrences were more common in patients with TNM stages I, 11, and I11 tumors where extended total gastrectomy was proven superior (p < 0.05); (d) early stage tumors tended to recur in the gastric remnant and the esophagus; (e) narrow surgical margins and margins involved by disease predisposed to recurrence; (f) not every patient with histologically invaded margins developed recurrence; (8) the risk of recurrence did not decrease with time; (h) the longer the disease-free interval the better the prognosis; (i) the more advanced the original lesion the longer the disease-free interval; (i) of all patients with recurrence only 19% had resectable lesions on reexploration; (k) the longer the disease-free interval the higher the resectability rate; (1) the median interval from recurrence to death was 2 months; (m) the same interval of those undergoing resection was 18 months. Patients with early-stage tumors treated with gastrectomy should be followed closely for local recurrence and should recurrence develop they should be reexplored if there is no evidence of metastasis.

........................................................................................................... ........................................................................................................... Key words: recurrence, gastric cancer

INTRODUCTION

Curative surgical treatment of gastric adenocarcinomas has two goals: (a) To re- move the primary lesion with an adequate tumor-free margin around it. (b) To resect those of the regional lymph nodes of the stomach which are likely to contain meta- static deposits. Subsequent development of distant metastasis proves that the neoplasm had spread beyond the confines of a properly performed gastrectomy at the time the procedure was carried out. Recurrence of the tumor, however, at the site of gastrectomy implies that the procedure did not fulfill the two objectives listed above.

According to autopsy studies, 80% of the patients who undergo gastrectomy for carcinoma of the stomach eventually develop local recurrence with or without distant

From the Gastric and Mixed Tumor Service, Department of Surgery, Memorial Sloan Kettering Cancer Cen- ter, New York.

Address reprint requests to D. N. Papachristou, MD, Archilleos 6, Agia Paraskevi, Athens, Greece.

0022-4790/81/1801-07$02.50 @ 1981 Alan R. Liss, Inc.

48 Papachristou and Fortner

metastasis [ 11. Also, second-look procedures performed in gastric cancer patients by Wangensteen et a1 showed that 60% of the gastrectomies had failed to control the disease locally if the regional nodes were involved [2]. If local recurrence at the site of gastrectomy is so frequent, how does it affect the course of the disease? Are there any contributory factors? Is it possible to prevent it? Which is the best way to manage it? How effective is treatment? These questions have not been answered in the literature. The present study examines the problem of local recurrence in a group of gastric cancer patients who were treated with gastrectomy and were followed regularly until their demise.

PATIENTS AND METHODS

The study involves 257 patients who underwent gastrectomy for adenocarcinoma of the stomach, survived their treatment and were followed regularly until their demise. None of the patients received adjuvant chemo- or radiation therapy and none of them had a second primary neoplasm. Death was caused either by the disease (n = 201) or by other causes (n = 56), such as myocardial infarction, cerebrovascular acci- dents, mesenteric thrombosis, intestinal obstruction, etc, of the 257 patients, 65 developed recurrence of the disease at the site of gastrectomy, either with (n = 52) or without (n = 13) distant metastasis. Recurrences were detected either by exploratory laparotomy (n = 46) or by endoscopy (n = 19) and they were always confirmed histologically. One more patient was found to have local recurrence with generalized metastasis at the time of autopsy. This was the only autopsy out of 19 that revealed local recurrence.

Resected neoplasms were classified according to the TNM system [3] on the basis of data obtained from pathology and operative reports. Local recurrence was defined as histologic evidence of tumor in the gastrectomy field, ie, the gastric remnant, the esophageal anastomosis, the pancreas, the duodenal stump, the hepatic pedicle, the splenic hilum, the celiac axis area, and the abdominal wound. The surgical margin was defined as the distance from the macroscopic edge of the primary to the line of gastric, duodenal or esophageal transection, depending on the type of gastrectomy.

RESULTS

Incidence

The overall incidence of local recurrence was 25% and it was not influenced significantly by the type of gastrectomy or the stage of the disease at the time of gastrectomy (Table I). Extended total gastrectomy [4] had the lowest incidence followed by distal subtotal and total gastrectomy, although the difference between the three procedures was not statistically significant. TNM classification showed that recurrences were more frequent among patients with early rather than advanced lesions (Table I), the difference, however, between the various stages was not statistically significant. It also showed that extended total gastrectomy gave superior results particularly in patients with early-stage neoplasms but again, the difference between the three types of gastrectomy in each TNM stage was not statistically significant (Table I). Extended total gastrectomy is a more thorough operation involving resection of the entire stomach, the spleen, the distal pancreas, both omenta and celiac nodes [4]. In the absence of distant spread (stages I, 11, III), extended total gastrectomy resulted in a significantly lower incidence of local recurrence than total gastrectomy. Thus, when the disease was

Local Recurrence of Gastric Cancer 49

TABLE I. Incidence of Local Recurrence Per Stage and Procedure

TNM stage

Gastrectomy All I I1 111 IV

Extended total 19%

Total 32%

Distal subtotal 28%

All 25%

(21/108)

(1 7/52)

(28197)

(66/257)

18% (2/11) 27% (3111) 35% (8/23) 28%

(13/45)

26% (4115)

100% (212) 30% (6/20) 32%

( 12/37)

16% (11167)

36% (9125) 33%

(12136) 25%

(3211 28)

26% (4/ 15) 21% (3114) 11% (2/18) 19% (9/45)

TABLE 11. Location of Recurrent Lesions ~~

Distal. Ext. total Total Subtotal

Gastric remnant 21 21 Esoph. anastomosis 11 (1) 8 3 AM. wound 4 (1) 2 2 Pancreas 4 (1) 1 2 1 Hepatic pedicle 3 (0) 1 2 Duodenum 8 (0) 2 4 2 Celiac axis 15 (0) 9 6

All 66 (13) 21 17 28

"In parentheses, number of resected lesions.

confined in the region of the stomach (stages I, 11, and III), the incidence of recurrence was 18% (17/93) with extended total, 37% (14/38) with total (p < 0.05, x2 test) and 33% (26/79) with distal gastrectomy.

Location

Recurrent lesions, in the majority, were located in the perigastric tissues and or- gans (Table 11). Recurrences appearing after total and extended total gastrectomy were mainly extragastric, while those appearing after distal gastrectomy were mainly gastric stump recurrences. Also, stages I and I1 lesions had the tendency to recur in the gastric stump and the esophageal anastomosis, while stages 111 and IV tumors had the tendency to recur in the perigastric tissues. Thus, 7 out of 1 1 esophageal and 13 out of 21 gastric stump recurrences appeared in patients with stages I and I1 neoplasms.

arising in the gastric stump were easier to treat than lesions arising in extragastric sites. In fact, almost half of stump recurrences were resectable (Table 11), while, only

Location had an impact on management of recurrent lesions. Recurrent tumors


2 out of 34 patients who developed recurrence in extragastric sites had resectable lesions. Esophagojejunostomy recurrences were difficult to manage because, in the majority of cases, the tumor had caused mediastinal involvement at the time of reexploration. Thus, out of 11 patients with esophageal recurrence, only one had resectable disease. The seriousness of extragastric recurrence was indicated by one more fact. Although recurrences developing in the abdominal wound were easy to detect by virtue of their superficial location, only 1 of the 4 patients who had such lesions could be managed with resection, the other 3 having advanced disease intrabdominaly.

Surgical Margins

Thus, the median distance from the primary to the proximal end of the gastric specimen in patients who developed recurrence in the stump after distal gastrectomy was only 3.5 cm (0-5.3), while the same distance was 6.5 cm (0-20) in patients who did not develop recurrence. The difference was statistically significant (p < 0.05, Rank sum test). Also, the median length of proximal margins in patients developing esophageal recurrence after total or extended total gastrectomy was only 2.5 cm (0-6), as compared with 6.5 cm (3-10) in patients not developing recurrence (p < 0.05). Histo- logically positive proximal margins were found in 4 patients treated with distal gastrectomy and in 22 treated with total or extended total gastrectomy. One in the first group and 9 in the second developed local recurrence.

4.5 cm (0-10) in patients developing recurrence as compared to 6.5 cm (2-12) in those who did not develop local recurrence. The difference was not statistically significant. In 14 patients the duodenal margin of the specimen was invaded by neoplastic tissue. Five of the 14 developed local recurrence in the duodenal stump. In 8 patients the duodenum was transected in an area which was macroscopically free but micro- scopically invaded by the neoplastic process. The longest of those false negative surgical margins was 5 cm. False-negative margins were also detected in the proximal end of the specimen in 14 patients. The longest proximal false negative margin in distal gastrectomy was 5 cm and in total and extended total was 7 cm.

Disease-Free Interval

Resections performed with narrow margins predisposed to local recurrence.

The median distance from the primary to the duodenal end of the specimen was

The median time interval from gastrectomy to detection of local recurrence was 15 months, recurrences, however, appeared as early as 2 months and as late as 7 years postoperatively (Table 111). The earlier the stage of the disease at the time of gastrectomy was, the longer the disease-free interval was (Table 111). Thus, stage I tumors recurred within 84 months, stage I1 tumors recurred within 44 months, and stage 111 tumors recurred within 38 months. The difference between stages I and 111 was statistically significant (p < 0.05, Rank sum test) and so was the difference between stages I and IV (p < 0.01). Although 48 of the 66 recurrences (72.7%) appeared the first 2 years after gastrectomy, the risk did not decrease as the years passed (Table IV), because recurrences affected l l % of the patients the first year and 12% of those who were still alive and at risk 4 or more years later.

Management

conservatively because they had clinically detectable distant metastasis. Resection of Of the 66 patients with recurrence, 46 were reoperated and 19 were managed


TABLE 111. Median Time Interval From Gastrectomy to Recurrence and From Recurrence to Death (Months)

TNM stage

All I I1 111 IV

Gastrectomy-recurrence 15 36 13.5 9 8

Recurrence-death 2 8 2 2 1 (2-84) (7-84) ( 2 4 4 ) (2-38) (2-24)

(0-89) (1-53) (0-89) (0-14) ( 1 4 )

TABLE IV. Incidence of Recurrence and Resectable Recurrence in the Postoperative Period

Months after gastrectomy

0-12 13-24 25-36 37-48 > 48

Local recurrence 1 1 % 14% 10% 8% 12%

Resectable recurrence 10% 22% 25% 40% 40% (30/257) (1 81 13 1 ) (8178) (5157) (5/43)

(3139) (4118) (218) (2/5) (215)

the recurrent process was possible in only 13 of the patients who were reexplored, for a 19% resectability rate. Ten of the 13 patients had gastric stump recurrences. The resectability rate was influenced by the stage of the disease and by the disease-free interval. The longer the disease-free interval was, the higher the resectability rate (Ta- ble JV). Resectability rate was 54% in patients with recurrent stage I disease as compared to 25, 12 and 0% in patients with stages 11, 111, and IV tumors. Gastric stump recurrences were managed with total gastrectomy. The patient with recurrence at the esophageal anastomosis was managed with resection and reanastomosis. The one with recurrence in the area of the pancreatic head was treated with pancreatoduodenectomy . Finally, the patient with recurrence in the abdominal scar was treated with wide local excision. Of the 13 patients with resectable recurrinces, 8 survived for another 5 or more years. Six long-term survivors had stage I tumors and 2 had stage I1 tumors.

Prognosis

with gastric adenocarcinoma. Although the median time interval from recurrence to death was only 2 months (Table 111), the 5-year survival of those wbo developed recurrence was 16% and of those who did not develop local recurrence was 18%. The 4- year survival was 21% for both groups and the 3-year survival was 30 and 25%, respectively. Even in stages I and I1 where recurrences were more manageable, the 5- year survival rates were 40 and 47%, respectively. It should be pointed out, however, that resection of the recurrent process prolonged significantly the survival of those pa-

Recurrence did not aggravate significantly the already dismal fate of patients


tients. Thus, the median time interval from exploration to death in patients with unre- sectable recurrence was only a month and a half (range 0 to 14 months), while the same interval in patients with resectable recurrence was 18 months (6-89). The difference between the two groups was statistically significant (p < 0.01, Rank sum test).

DISCUSSION

Local recurrence of gastric adenocarcinoma following gastrectomy is common, although its exact incidence is difficult to determine. McNeer et a1 reviewed autopsy protocols of gastric cancer patients who were treated with subtotal gastrectomy and died from their disease [I]. They found that the procedure had failed to control the disease locally in no less than 80% of the cases. On the contrary, the incidence of local recurrence in a series of patients treated with “radical surgery” was, according to Iwanaga et al, only 13% (5). The first study, however, included only patients dying from the disease, while the second was concluded although some of the patients at risk were still alive. Since not all operated patients die from the disease and not all long- term survivors escape local recurrence (Table IV), it is likely that neither of the two quoted figures represents the true incidence of local recurrence. By the same token, the 25% incidence reported here must be lower than real because it is possible that some of the patients died with subclinical recurrence which was never discovered. It appears that the only way to reach the truth would be to follow a group of gastrectomized patients to their demise and conduct postmortems in every single one of them.

tions or with some form of adjuvant treatment. As far as the first alternative is con- cerned, extended total gastrectomy resulted in the lowest incidence of recurrence in the present series and it was proven superior to total gastrectomy in this regard. The procedure, which is a “more thorough operation” [6], was particularly effective in the presence of early-stage lesions confined in the stomach and the regional lymph nodes. These results indicate that extended total might be preferable to total gastrectomy, although a selection between the two should be influenced by factors other than local recurrence, such as operative mortality and ultimate survival.

the time of gastrectomy. Although narrow margins were responsible only in part for the development of recurrence, the study showed that patients with gastric stump and esophageal recurrences had been treated with rather narrow margins. Considering the ability of gastric adenocarcinomas to spread intramurally both toward the esophagus [7] and the duodenum [8], it was not surprising that recurrences occurred even with margins as wide as 10 cm. On the other hand, patients treated with extremely narrow or even zero margins did not always develop recurrence. Hence the question arises, which patients should be treated with wide margins? The study showed that early-stage tumors were more prone to local recurrence than advanced lesions and that they had the tendency to recur in the gastric remnant and the esophagus. Accordingly, wide margins should be obtained in patients with early-stage neoplasms.

Technically, one can obtain a wide margin more easily in the upper than the lower part of the specimen. This is because the duodenum is a much more difficult structure to resect than the upper stomach or even the distal esophagus. Thus, in order to resect the primary with an adequate margin, one may have to include tbe second part of the duodenum which can only be accomplished with pancreatoduodenectomy. To proceed, however, with such an extensive operation one should be certain that the

Theoretically, recurrences could be prevented either with more extensive resec-

Another way of preventing local recurrence would be to obtain wider margins at


benefit is worth the risk, and this is not the case here. Although, the so-called pyloric banier to the spread of gastric cancer is a myth [8], pancreatoduodenectomy for eradi- cation of the disease from the duodenal area will result in a prohibitively high mortality [9]. Thus, in dealing with tumors approaching the duodenum, the surgeon may have to resort to some kind of adjuvant treatment such as intraoperative radiation [lo] or chemotherapy, rather than extensive surgery.

was associated with distant metastasis. Thus, from all those who developed recurrence only 19% had resectable lesions at the time of reexploration. Would that figure be higher if the patients had been explored earlier? The study did not answer this question; it led, however, to certain guidelines which might be used to detect the recurrence at an earlier stage. It showed that the risk of developing recurrence does not di- minish with time and that it remains high 5 or more years following gastrectomy. It also showed that the longer the disease-free interval the higher the possibility for the recurrent lesion to be resectable. Accordingly, a persistent and uninterrupted follow up may lead to early diagnosis and treatment. Follow up with regularly performed endos- copic examinations might be particularly important in patients with early stage tumors because they are not only more prone to recurrence than others but they also tend to develop recurrence in the gastric remnant and the esophageal anastomosis rather than in less accessible extragastric locations. Recurrences in such individuals appear long after gastrectomy and are resectable in quite a few.

Although the 5-year survival rates of patients with and without local recurrence were almost the same, the appearance of recurrence was a very grave prognostic ele- ment. Thus, the median time interval from the onset of recurrence to death was only 2 months. More important, however, is the fact that 8 of the 13 patients whose recurrent lesions were resected successfully survived for another 5 or more years and that the median time interval from resection of the recurrent tumor to death was 18 months. This means that in the absence of distant metastasis patients with local recurrence should be managed with reexploration and possibly with resection. It also could mean that early treatment of the recurrence might salvage some of these patients.

Local recurrence was difficult to treat effectively because, in most patients, it

REFERENCES 1. McNeer G, Vandenberg H Jr, Donn FY, Bowden L: A critical evaluation of subtotal gastrectomy for the

2. Wangensteen OH, Griffen WO, Lewis FJ, Arhelger SW: “The Second Look Operation in the Manage- cure of cancer of the stomach. Ann Surg 134:2, 1951.

ment of Abdominal Cancer,” p.62. “Treatment of Cancer and Allied Diseases,” Vol 5 , New York: P. B. Hoeber, 1962.

3. American Joint Committee for Cancer Staging and End Results Reporting: Manual for Staging of Cancer, p 71, 1977.

4. McNeer G, Pack GT: Neoplasms of the stomach, p 312, Philadelphia:Lippincott, 1967. 5 . Iwanaga T, Koyama H, Furukawa H, Taniguchi H, Wada A, Tateishi R: Mechanisms of late recurrence

after radical surgery for gastric carcinoma. Amer J Surg 135:639, 1978. 6. McNeer G, Sunderland DA, McInnes G, Vandenberg H Jr, Lawrence W Jr: A more thorough operation

for gastric cancer: Anatomical basis and description of technique. Cancer 4:957, 1951. 7. Papachristou DN, Karas M, Fortner JG: Anastomotic recurrence in the esophagus complicating gastrec-

tomy for adenocarcinoma of the stomach. Br J Surg 66609, 1979. 8. Paulino R, Roselli A: Carcinoma of the stomach with special reference to total gastrectomy. C u r Probl

Surg December 1973, p 13. 9. Berry REL, Rottschafer W: The lymphatic spread of cancer of the stomach observed in operative speci-

mens removed by radical surgery including pancreato duodenectomy. Surg Gyn Obst 104:269, 1957. 10. Mitsuyuki A, Takahashi M, Yabumoto E, Adachi H, Yoshii M, Mori K: Clinical experiences with

intraoperative radiotherapy of locally advanced cancers. Cancer 45:40, 1980.

Documents

Local recurrence of gastric adenocarcinomas after gastrectomy