The pattern of lymph node involvement in carcinoma of the head of the pancreas

International Journal ofPancreatology, vol. 13, no. 1, 15-22, February 1993 �9 Copyright 1993 by Humana Press Inc. All rights of any nature whatsoever reserved. 0169-4197/93/13:15-22/$2.00

The Pattern of Lymph Node Involvement in Carcinoma of the Head of the Pancreas

A Histologic Study of the Surgical Findings in Patients Undergoing Extensive Nodal Dissections

Takukazu Nagakawa,* H. Kobayashi, Keiichi Ueno, Tetsuo Ohta, Masato Kayahara, Kazuhiro Mori, T. Nakano,

T. Takeda, I. Konishi, and ltsuo Miyazaki The Second Department of Surgery, School of Medicine,

Kanazawa University, 13-1 Takaramachi, Kanazawa, 920 Japan

Summary

To clarify the pattern of lymph node metastasis in carcinoma of the pancreas, lymph node involvement was examined in forty-two patients who underwent extensive nodal dissections, including the paraaortic lymph nodes. The correlation between the spread of the tumor and lymph node involvement was evalu- ated: The most common site of involved lymph nodes was the retropancreatic region. The prevalence of nodal metastases was 78.6%. Metastases to the paraaortic region were present in seven patients, among whom metastases in the paraaortic region were most common in the median region from the celiac artery to the inferior mesenteric artery and in the space between the aorta and the vena cava. The risk of lymph node metastases tended to increase with tumor size, except in the paraaortic region, where the correlation between the frequency of metastasis and tumor size was poor. The probability of lymph node metastases increased with the degree of lymphatic invasion (ly) and the growth pattern of the tumor (INF) and was high in patients with invasion into the retropancreatic tissue and in tumors with scirrhous histology. These results indicate that even in small cancers, lymph nodes of the paraaortic region frequently harbor metastases and should be dissected en block during radical resections of pancreatic cancer.

Key Words: Carcinoma of the head of the pancreas; lymph node involvement; paraaortic lymph nodes; tumor size; histological stage.

Introduction

It has been our practice since 1973 to perform an extended lymph node dissection as part of the surgical management of carcinoma of the pancreas. In

Received March 23, 1992; Revised June 8, 1992; Accepted June 22, 1992

*Author to whom reprint requests should be addressed: The Second Department of Surgery, School of Medicine, Kanazawa University, 13-1 Takaramachi, Kanazawa, 920 Japan

1977, a translateral retroperitoneal approach (1,2) was developed to make retroperitoneal dissection, including dissection of paraaortic lymph nodes, possible. This approach has resulted in a five year survival rate of 27.6% for carcinoma of the head of the pancreas. However, local recurrence caused by lymphatic metastases remains a serious problem (3-5).

Morphological studies of the retroperitoneal lymph nodes, particularly the paraaortic lymph

15

16 Nagakawa et al.

nodes, have generally been centered on anatomical location and drainage patterns (6-9). There have been few analyses that provide theoretical support for dissection of the paraaortic lymph nodes. There- fore, we examined the lymph node involvement in patients who underwent extensive nodal dissections, including dissection of the paraaortic lymph nodes, for carcinoma of the head of the pancreas. We then correlated the frequency of paraaortic lymph node involvement with the extent of the primary tumor in order to clarify the importance of paraaortic lymph node dissection in curative pancreatic resections.

Subjects and Methods

The pattern of lymphatic metastases was investigated in forty-two patients with cancer of the pancreatic head who were undergoing radical resection at the Second Department of Surgery, School of Medicine, Kanazawa University, during the period between November 1973 and December 1989. Sections cut at 5-mm intervals were prepared from excised specimens after fixation in 10% for- malin, and were stained with hematoxylin and eosin (H-E). The lymph nodes were excised from the specimens as soon as possible after surgical extirpation and fixed in the same manner. One representative histologic section, including the hilum of the lymph node, was analyzed in each patient. Small lymph nodes that were not seen grossly were identified in the serial sections microscopically. The histological findings were classified according to the General Rules for Cancer of the Pancreas published by the Japan Pancreas Society (10) with a modified classification, as shown in Tables 1 and 2. The nomen- clature of the major lymph nodes was defined as Figs. 1 and 2, according to the classification of the General Rules for Pancreatic Cancer and the proposal of the Japan Society for Cancer Therapy (11), or our own classification of paraaortic lymph nodes.

All of the neoplasms except one undifferentiated carcinoma were ductal adenocarcinoma. The relationship between tumor size, histologic type of the tumors, histologic invasion, and lymphatic meatastasis was studied using the chi square test.

Results

Distribution of Lymphatic Metastases Metastases to the lymph nodes were present in 33

of the 42 patients (78.6%). The sites of metastasis were in areas 6, 8, 9, 11-14, and 16-18. The prevalence of metastases to area 13 was the highest, followed by areas 17 and 14. There were seven patients with metastases to the paraaortic area 16 (Table 3).

There was single nodal metastasis in area 16 in four patients; multiple paraaortic nodes were involved in the other three patients, with 3, 6, and 9 nodes involved in one patient each. In patients harboring paraaortic nodal metastases, the average number of lymph nodes involved was 3.1. The average incidence of lymph nodes involved in area 16 was 3.2% of examined lymph nodes. The extent of lymphatic metastasis in area 16 increased with the number of involved lymph nodes. From an anterior view in area 16, the incidence of metastases to segment b2 was the highest, 85.7% (six of seven cases), followed by 42.9% (three of seven cases), for segment b l. In the transectional view, the incidence of "Inter" segment involvement was the highest, 71.4% (five of seven cases), followed by 57.1% (four of seven cases) for the Aor-pre segment (Fig. 3).

Relationship Between Tumor Size, Histologic Type of the Tumors, Histological Invasion, and Lymphatic Metastasis The extent of lymphatic metastases tended to

increase with the tumor size. However, two patients with t2a tumors had metastases to area 16, suggesting that the relationship between tumor size and risk of metastasis to the paraaotic region was poor (Table 4).

Regarding to histologic type of the tumors, the incidence and the extent of lymphatic metastasis in tubular adenocarcinomas tended to be higher and wider than those in paillary adenocarcinomas, but these differences were insignificant. Metastases to area 16 were seen in five patients with tubular adenocarcinoma and in two patients with papillotubular adenocarcinoma (Table 5).

International Journal of Pancreatology Volume 13, 1993

Lymph Node Involvement in Pancreatic Cancer 17

Table 1 Classification of Histological Findings of Pancreatic Cancer Tumors

Amount/type Size, t Growth patterns connective tissue

t. < 2 0 c m 1

t2a 2.1-3.0 cm t2b 3.1-4.0 cm t 3 4.1-6.0 cm t4 > 6.1 cm

INFot expansive with clear margins

INF~ intermediate between tz and "t

INFu infiltrative without clear margins

Medullary type Intermediate type Scirrhous type

Table 2 Classification of Histological Findings of Pancreatic Cancer Invasion

Retropancreatic Degree of invasion to Groups of invasion, rp lymphatic vessels, ly lymph nodes

rpo: tumors confined to pancreas

rpe: retropancreatic invasion

rpi: invasion to the other organs

lY0: no invasion ly.: minimal invasion lYe2: moderate invasion lY3: severe invasion

NO: no mestases NI: mestases to group 1

(areas 8, 12-14, and 17) N2: mestases to group 2

(areas 9, 11, 12, 15, 16, and 18) more than N2

Table 3 Distribution of Lymph Node Metastases in Carcinoma of the Head of the Pancreas

Nodal station a 3 4 5 6 7 8 9 10 11

No. of patients 0 0 0 1/42 0 6/42 2/42 0 2/42 with metastasis

Frequency of 2.4 14.3 4.8 8.3 metastasis, %

12 13 a b 14 a b c d 15 16 17 a b 18

9/42 29/42 16/42 0 7/42 17/42 5/20 22 19 8 11 3 6 13 9

21.4 69.0 38.1 16.7 40.5 25.0 52.4 45.2 19.0 26.2 7.1 14.3 31.0 21.4

a3-7, gastric lymph nodes; 8-18, see Fig. 1.

No lymphatic metastasis was observed in three o f four patients (75.0%) with tumors conf ined within the pancreas (rpo). In contrast, 23 of 37 patients (62.2%) had metastases to at least two lymph nodes if invasion into the retropancreatic tissue (rpe) was present. All patients with metastases to area 16 had rpe or rpi (Table 6). The extent of lymphatic metastases increased with invasion into

lymph vessels, and all patients with metastases to area 16 had this invasion (Table 7).

The risk of lymphatic metastases was correlated with the pattern of infiltration, with the extent of nodal metastases increasing from the o~ to the 13 and y patterns. In the y pattern particularly, metastases to at least two lymph nodes were observed in 73.7% (14 of 19 patients), and was significantly higher than

International Journal o f Pancreatology Volume 13, 1993

Fig. 1. Classification of the regional lymph nodes of the pancreas. Insert: Subdivision of area 14. (AMS) superior mesenteric artery: (AJ) jejunal artery; (APDI) inferior pancreaticoduodenal artery; medial colic artery; (8) lymph nodes around the common hepatic artery; (9) lymph nodes around the celiac trunk; (10) lymph nodes at the hilus of the spleen; (11) lymph nodes around the splenic artery; (12) lymph nodes of the hepatoduodenal ligament; (13) posterior pancreatoduodenal lymph nodes; (13a) above the papilla of Vater; (13b) below the papilla; (14) lymph nodes around the superior mesenteric artery; (14a) lymph nodes at the root of the superior mesenteric artery; (14b) lymph nodes at the root of the inferior pancreatoduodenal artery; (14c) lymph nodes at the root of the medial colic artery; (14d) lymph nodes at the roots of the first jejunal artery; (16) paraaortic lymph nodes; (17) anterior pancreatoduodenal lymph nodes; (17a) above the papilla of Vater; (17b) below the papilla; (18) subpancreatic lymph nodes.

A

:,,,, al

. . . . . . . . . . . . . . . . . . . . . . . . . . . "~ I a

i n t e r !

pre I pre ,

retro retro

Bif

l 16c

I

Fig. 2. Classification of the lymph nodes of area 16. (a) frontal classification; (b) transectionaI segment. (IVC) inferior vena cava; (Aor) abdominal aorta; (16a) upper region above the celiac artery; (16b) median region from the celiac artery to the inferior meseteric artery; (16c) lower region from the inferior meseteric artery. Cross-sectional segmentation was analyzed according to the proposal of the Japan Society for Cancer Therapy (11). The inferior vena cava (IVC) was divided into anterior (Ivc-pre), lateral (Ivc-latero), and posterior (Ivc-retro) segments, and a segment between the abdominal aorta and inferior vena cava (Ivc-inter), and the abdominal aorta (Aor), into anterior (Aor-pre), lateral (Aor-latero) and posterior (Aor-retro) segments.

International Journal of Pancreatology 18 Volume 13, 1993

case

mapping of

lymph nodes

positive LN examined LN

4

o~

3 19"

6 9 23

Fig. 3. Lymph node mapping in patients of paraaortic lymph node metastasis. (LN) lymph nodes. (O) anterior lymph node without metastasis; ( 0 ) anterior lymph node with metastasis; (O) posterior lymph node without metastasis; (11) posterior lymph node with metastasis.

Table 4 Relationship Between Tumor Size

and Lymph Node Involvement in Carcinoma of the Head of the Pancreas

Tumor No. of "t" stage cases N0 a N1 t' _>N2 c

t 1 2 2 0 0 2a 12 3 4 5 (2) d 2b 14 2 3 9 (3) 3 11 1 1 9 (3) 4 3 1 1 1

aN0, no nodal metastases. bN1, metastases to lymph nodes in group 1. CN2, metastases to lymph nodes group 2. din parentheses, the number of cases with

positive metastases to area 16.

Table 5 Relationship Between Histological Type of Tumors

and Lymphatic Metastases in Carcinoma of the Head of the Pancreas

Type a Cases NO N1 ~N2

Pap 3 1 2 0 Pap-tub 5 1 2 2 (2) b Tub 29 5 17 7 (5)

Tub1 10 3 5 2 (2) Tub2 15 2 9 4 (3) Por 4 0 3 1

Sq 1 0 1 0 Adsq 1 0 0 1 Und 1 1 0 0

apap, papillary adenocarcinoma; PaP-tub papillotubular adenocarcinoma; Tub, tubular adenocarcinoma; Tubl, well differentiated adenocarcinoma; Tub2, moderately differentiated adenocarcinoma; Por, poorly differentiated adenocarcinoma; Sq, squamous cell carcinoma; Adsq, adenosquamous carcinoma; Und, undifferentiated carcinoma.

bin parentheses, the number of cases with positive metastases to area 16.

Table 6 Relationship Between Invasion into the Retropancreatic Tissue and Lymph Node Involvement

in Carcinoma of the Head of the Pancreas

Extent of retroperitoneal No. of

spread a cases NO b

rpo 4 3 rpe 37 6 rpi I 0

N2 >_N2

1 0 8 23 (6) o 1 (1)

arpo, tumor remains within the pancreas ; rpe, invasion into retroperitoneal connective tissue; rpi, invasion into retroperitoneal viscera. NO, N1, N2, in parentheses, see Table 2.

Table 7 Relationship Between Invasion

into Lymph Vessels and Lymph Node Involvement in Carcinoma of the Head of the Pancreas

Extent of lymphatic No. of invasion a eases NO b N2 >_N2

ly0 5 3 1 1 1 10 1 5 4 ( 1 ) 2 z0 4 2 14 (3) 3 7 1 1 5 (3)

aly0, no invasion into lymphatic vessels; lyl, minimal invasion into lymphatic vessels; ly2, moderate invasion into lymphatic vessels; ly3, severe invasion into lymphatic vessels.

bN0, NI, N2, in parentheses, see Table 2.

International Journal of Pancreatology 19 Volume 13, 1993

20 Nagakawa et al.

Table 8 Relationship Between Growth Pattern

of the Primary Tumor and Lymph Node Involvement

in Carcinoma of the Head of the Pancreas Growth No. of pattern a cases NO b NI ~_.N2

INF ~ 3 2 l 0* 13 20 6 4 10 0)* ' / 19 1 4 14 (6)*

aINFct; expansive growth with clear margins; INF,, intermediate between ~ and'/; INF'/, infiltrative growth with clear margins.

t~N0, N1, N2, in parentheses, see Table 2. * p < 0.05.

that in patients showing the o~ pattern. Metastases to the paraaortic region were present in the patients with the 13 and Y patterns, and were particularly common in patients with the Y pattern (Table 8).

Regarding the amount of connective tissue in the tumor tissue, the incidence of lymphatic metastases was higher in patients with intermediate and scirrhous types than in patients with medullary type. The incidence of metastasis to area 16 showed a similar tendency, with none of the patients with medullary type having metastases to area 16 (Table 9).

Survival o f Patients with Metastasis to Area 16

Among seven patients with metastasis to area 16, four died within 1 yr and three survived more than 1 yr. The longest survival period was two years and three months, and this patient was alive until June 1992. Among 35 patients without metastasis to area 16, 11 have remained alive more than three years.

Discussion

The incidence of lymph node metastases in cancer of the head of the pancreas is high. Even in resectable patients, Cubilla et al. (12) observed nodal metastases in 19 of 22 patients undergoing pancreatoduodenectomy. Similar incidences of nodal metastases in Japan have been reported (13- 15). In our present study, the frequency of metastases (78.6%) was similar to earlier reports. Metastases to area 16 was 29, 21.1, and 16.7% in

Table 9 Relationship Between Amount of Corrective Tissue

in the Tumor Tissue and Lymph Metastases in Carcinoma of the Head of the Pancreas

No. of Type c a s e s NIP N1 >N2

Medullary type 6 2 2 2 Intermediate type 17 3 4 10 (3) Scirrhous type 19 4 3 12 (4)

*NO, N1, N2, in parentheses, see Table 2.

reports by Nakao et al. (14), our previous study (15), and this study, respectively.

According to previously published guidelines (10), the extent of dissection of area 16 is restricted to the paraaortic area ranging from the origin of the celiac artery (upper margin) to the origin of the inferior mesenteric artery (lower margin). However, this description of the area does not include the con- cept of "depth." To more adequately define the anatomic spread of the tumor, we have included the extent of disease in the cranial and caudal direc- tions, that is, the extent according to the general rules (10), to lie in a longitudinal plane viewed anter- iorly, whereas the extent in the ventral to dorsal di- rections is regarded was lying in a cross-sectional plane.

Pathways of lymphatic metastasis from the pancreatic head to area 16 have been studied by Evans (9), Nagal (16), and Deki et al. (8). Evans (9) injected pigment into the posterior part at the pancreatic neck between the pancreatic head and body and identified spread to the juxta-aortic node in 30 autopsy cases, but no details of the inflow pathway or of the anterior part within area 16 were provided. Nagai (16) studied 21 autopsy cases, injecting the pigment into areas 13a and 13b and then measured the flow into the lymphatic network on the posterior part of the pancreatic head and into area 16 via the lymph vessels on the right side or along the posterior part of the superior mesenteric artery. The pigment injected into area 17a flowed toward the contralateral nodes through two or more lymph vessels on the anterior part of the pancreatic head and arrived mainly in areas 14a, 14b, and 14d. Further- more, some of the pigment flowed directly into the major intestinal lymphatic trunk and arrived in area

International Journal o f Pancreatology Volume 13, 1993

Lymph Node Involvement in Pancreatic Cancer 21

16. In area 17b it flowed mainly into the lymphatic network on the posterior part of the pancreatic head, and arrived in area 16 via lymph vessels running along the pos ter ior surface o f the super ior mesenteric artery. Deki et al. (8) studied four autopsy cases and reported that the lymph vessels from the right side of the pancreas flowed into the upper and lower lymph nodes of the left renal veins between the aorta and the inferior vena cava prima- rily via the lymph nodes on the posterior surface of the pancreatic head. However, some reports have shown that the incidence of metastases to the lymph nodes in area 13 was 28.8-55% (13-15). We also found a high frequency (69%) of metastases to area 13. No patient had nodal metastases without area 13 also being involved. This suggests that if the lymphatic pathway from the pancreatic head to area 16 is orthodromic, then area 13 is the site of the primary interstitial lymph nodes. The second lymph node on the metastatic pathway would then be area 14, was emphasized by both our group (15) and by Hagiwara (13). Our finding that all patients with metastases to area 16 also had metastases to area 14 supports this hypothesis. Hagiwara (13) performed an experiment with dogs where the roots of the celiac and superior mesenteric arteries were occluded, and reported that area 14 was the final lymph vessel supplying area 16. Nagal (16) reported a pathway from areas 13 to 14, and then on to area 16, as described above. (From these results, there seems to be a close correlation between area 16 and area 14, and the main pathway of metastasis from the pancreatic head to area 16 seems to be from areas 13 ---> 14 ---> 16.)

Nagai (16) investigated the pattern of lymphatic metastases within area 14 in 3 patients with metastases to area 16. He found that three of these patients had metastases to area 14b, two to area 14d, and one to area 14a, respectively. As far as our patients were concerned, metastases to area 14b were present in three of seven cases, and there were some instances of a single metastasis to areas 14a, 14c, and 14d, respectively. These results suggest that metastases to any of the sites of 14a, 14b, 14c, and 14d are equally possible.

Hagiwara (13) found that the incidence of lymph node metastasis tended to increase with tumor size,

a trend that we observed. In area 16, however, only two of our patients with small t2a tumors had metastases, suggesting that the correlation between tumor size and metastasis to the paraaortic region is poor. We also studied the relationship between histological stage and the prevalence of nodal metastases, with more histologically aggressive patterns of lymphatic invasion (ly), and growth (INF, expressed was o~, 13, and y), the incidence of metastases increased. In addition, the incidence of metastases was also higher in rpe cases, rpi cases, and in intermediate and scirrhous types of tumors (defined according to the quanti ty o f stromal connective tissue in the cancer). Similar results were observed concerning metastases to area 16, but this relationship between metastases and histological type was indifferent from the relationship between tumor size and metastasis to this area, and as far as we know, there have been no previous reports showing this relationship. However, our results support the hypothesis that favorable prognostic features include no invasion into any lymph nodes (ly0), or veins (v0) and INFo~ (17).

In our patients, the frequency of cancer cells transferred to 16b2 was the highest of the anteri- orly-viewed segments, and that to segment IVC- Inter was the highest in the cross-sectional plane. when the number of aetastatic lymph nodes increased, the area of distribution broadened. Based on the autopsies of three patients with potentially operable cancer of the pancreatic head with nodal metastases but who died without attempt at resection because of poor general condition, Nagai (16) for- mulated a map of the pattern of spread to area 16. According to this report, there were 1-3 lymph nodes with metastases. Although there was no description of the site, cancer cells were located in area 16b in two cases, 16c in three, 16d in one, IV in two, V in two, and VI in two, according to Nagai's classification (16). According to our classification, cancer cells were located in area 16bl in two cases, 16b2 in three, 16c in one, segment IVC-Inter in two, segment Aor-pre in two, and segment Aor-latero in two. Thus, our results were similar to Natal 's (16). Nagai (16) also observed lymphatic flow toward the right of the aortic surface in the area extending from area 13 to areas 16b and 16c, toward the anterior

Internationod Journal of Pancreatology Volume 13, 1993

22 Nagakawa et al.

surface of the aorta in area 16b, and toward the right, anterior, and left surfaces of the aorta in the areas, respectively, corresponding to segment IVC-Inter of area 16b, segment Aor-pre of area 16bl and segments IVC-lnter, Aor-pre and latero of area 16b, according to our classification. Kutsuna (6) described direct transfer of cancer cells from the posterior surface of the pancreatic head to the sites corresponding to segment Aor-pre of area 16b according to our classification in a small number of the cases. The segment IVC-Inter of area 16b2 was therefore esti- mated to be the most important site in the early stages of metastases to area 16, with lymph distributing gradually and centering on segment IVC-Inter of b2 within area 16. This suggests that transfer toward the segment in the longitudinal direction tends to take precedence over that in the cross-sectional plane, and that it is necessary to dissect up to the depth of the right and left renal arteries. All reports related to the present study were restricted to the autopsy cases reported by Nagal (16). Clinical cases are ex- pected to be reported in the future by institutes where dissection of area 16 is being actively performed.

From these findings, we consider an extensive dissection including areas 14 and 16 necessary for radical resection of cancer of the pancreatic head. In the dissection of area 16, a radical resection, including the left side of segment aorta up to a level deeper than both of the renal arteries, is necessary to offer the patient the best chance for curative resection.

References

1 Nagakawa T, Kurachi M, Konishi K, Miyazaki I. Retroperitoneal dissection for operation of pancreatic cancer: Translateral retroperitoneal approach. Igakunoayumi 1979; 111: 339-341 (in Japanese).

2 Nagakawa T, Kurachi M, Konishi K, Miyazaki I. Trans- lateral retroperitoneal approach in radical surgery for pancreatic carcinoma. Jpn J Surg 1982; 12: 229-233.

3 Nagakawa T, Konishi I, Higashino Y, Ueno K, Ohta T, Kayahara M, Ueda N, Maeda K, Miyazaki I. The spread and prognosis of carcinoma in the region of the pancreatic head. Jpn J Surg 1989; 19: 510-518.

4 Nagakawa T, Konishi I, Ueno K, Ohta T, Akiyama T, Kayahara M, Miyazaki I. Surgical treatment of pancreatic cancer, lnt J Pancreatol 1991; 9: 135-143.

5 Kayahara M. Nagakawa T. Konishi I, Ueno K, Ohta T, Miyazaki I. Clinicopathological study of pancreatic carcinoma with particular reference to the invasion of the extrapancreatic neural plexus. Int J Pancreato11991; 10: 105-111.

6 KutsuaaS.AnatomyoftheLymphatic~ysteminthe Japanese. Kanehara Pub., Tokyo, Japan, 1967 (in Japanese).

7 Inoue Y. Lymphatic system of the stomach, duodenum, pancreas, and diaphragma. Jpn Anatomy 1936; 9:35-117 (in Japanese).

8 Delft H, Sate T. An anatomic study of the peripancreatic lymphatics. Surg Radial Anat 1988; 10: 121-135.

9 Evans BP Ochsner A. The gross anatomy of the lymphatics of the human pancreas. Surgery 1954; 36: 177-191.

10 Japan Pancreas Society. General Rules for Surgery and Pathological Studies on Cancer of the Pancreas. 3rd ed., Kanehara, Tokyo, Japan, 1986 (Abstract in English).

l l Japan Society for Cancer Therapy. General Rules for Reporting Clinical Oncology (JSCT General Rules). Kanehara, Tokyo, Japan, 1991 (in Japanese).

12 Cubilla AL, Fortner JG, Fitzgerald PJ. Lymph node involvement in carcinoma of the head of the pancreas area. Cancer 1978; 41: 880-887.

13 Hagihara K. Experimental and clinicopathological stidies on lymphatic flow of the pancreas. Igaku Kenkyu 1982; 52:61-85 (in Japanese).

14 Nakao A, Takeda N, Kasuga T. Significance of dissection of paraaortic lymph nodes in pancreatic cancer. Jpn J Surg Assoc 1990; 21:23-26 (in Japanese).

15 Nagakaa T, Konishi I, Maeda M, Ohta T, Aldyama T, Kanno M, Kayahara M, Miyazaki I. Lymph node involvement by carcinoma of the pancreatic head area: A grouping of regional lymph nodes involved by the tumors. Memoirs A1 Med. Kanazaa Univ. 1990; 14:1-11.

16 Nagai H. Involvement of paraaortic lymph nodes by pancreatic cancer and its lymphatic flow. Jpn JSurg 1987; 88:308-317 (Abstract in English).

17 Takagi K, Ohhashi I, Ohta H. Pancreatic cancer with bad prognosis. Stomach and Intestine 1984; 19:1193-1197 (Abstract in English).

International Journal of Pancreatology Volume 13, 1993

Documents

The pattern of lymph node involvement in carcinoma of the head of the pancreas