- Email: [email protected]
14 Lymph node metastasis and retroperitoneal lymphadenectomy in ovarian cancer
14 Lymph node metastasis and retroperitoneal lymphadenectomy in ovarian cancer PAO-CHEN WU JING-HE LANG RONG-LI HUANG JING-YI QU HE WANG MIN-YI TANG RONG-GUO ZHAO LI-JUAN LIAN
Cancer of the ovary has become increasingly important in the last few decades and ranks as the leading cause of death among all gynaecological malignancies. While clinicians are devoting themselves to the study of the behaviour of ovarian cancer, as well as to the search for more effective therapeutic modalities, little attention has been paid to an important route of metastasis in this group of diseases: retroperitoneal spread through the lymphatic pathway. Two possible explanations can be tentatively made for this negligence. In the first place, the silent metastasis of ovarian cancer along the lymphatic pathway is overshadowed by the outstanding extensive intraperitoneal implantation and local invasion to adjacent tissues and/or organs which divert the attention of most gynaecologists and oncologists. Secondly, the traditional anatomical conception about lymphatic drainage of the ovary leads clinicians to believe that once retroperitoneal lymphatic metastasis has occurred the first group of lymph nodes involved will always be the para-aortic nodes--nodes which are situated high above the pelvis and far beyond the reach of conventional surgical and radiation therapy. In 1974, Knapp and Friedman first reported their experience with aortic lymph node metastasis in 26 patients with Stage I ovarian cancer subjected to para-aortic lymphadenectomy. In a prospective study by selective nodal biopsy in 61 unselected cases of ovarian cancer, Chen and Lee (1983) found that the incidence of metastasis was 37.7% for para-aortic nodes and 14.8% for pelvic nodes. Burghardt et al (1984) first emphasized the significance of retroperitoneal lymph node dissection in ovarian cancer from both the diagnostic and therapeutic points of view. In 48 patients undergoing systemic pelvic lymphadenectomy, the incidence of positive pelvic nodes was 56.3%, and the incidence of positive para-aortic nodes in 16 evaluated cases Bailli~re's Clinical Obstetrics and Gynaecology--Vol. 3, No. 1, March 1989
143
144
P.-c. w u ET AL
was 31.2%. Beginning in 1982, retroperitoneal lymphadenectomy has been performed at the Peking Union Medical College Hospital as a routine procedure in addition to the cytoreductive surgery for most patients with ovarian malignancies. The purpose of this paper is to present our five-year experience of a prospective study on lymph node metastasis in order to assess the role of lymphadenectomy in the management of ovarian cancer. Some of these cases were reported previously (Wu et al, 1984; 1986). MATERIAL AND METHODS During a period of 5 years from June 1982 through May 1987, retroperitoneal lymph node dissection was performed in 105 cases of ovarian cancer. The general clinical data are shown in Table 1. Ovarian cancer of epithelial Table 1. Subjects under study (n = 105). Number of cases (%) Histological type Epithelial cancer Germ-cell tumour
77 (73.3) 28 (26.7)
Clinical staging (FIGO) I II III IV
16 9 72 8
Primary or secondary surgery Primary Secondary
66 (62.9) 39 (37.1)
(15.2) (8.6) (68.6) (7.6)
origin constituted the majority of the cases, of which 53 (68.8%) were of the serous type. Clinical staging was based on the FIGO criteria assessed during laparotomy without considering the pathological findings of the lymph nodes. It can be seen that patients with Stage III and IV disease accounted for 76.2% of the total series, and only 16 patients had Stage I disease. Sixty-six patients (62.9%) were operated on for the first time (primary operation), and secondary operations were performed in the remaining 39 cases for recurrent or incompletely resected tumours. Lymphography In order to evaluate the role of roentgenological examination in predicting lymph node metastasis in ovarian cancer, lymphography was performed preoperatively in 30 cases. Six millilitres of Myodil were injected into the lymphatics between the second and third toes on the dorsum of each foot within 2 hours by means of an injection pump. X-ray films were taken of the anterio-posterior, left anterior oblique and right anterior oblique views immediately after injection of the radio-opaque media, and were repeated 24 hours later.
LYMPH NODE METASTASIS AND RETROPERITONEAL LYMPHADENECTOMY
145
Methods of surgery
The surgical methods used are indicated in Table 2. Table 2, Methods used in surgery. Number of cases Cytoreductive surgery Retroperitoneal lymphadenectomy Systemic Selective sampling
105 69 36*
*Aortic nodes were not removed in 25 cases.
Cytoreductive surgery With the exception of a few young patients with unilateral Stage I germ-cell tumour, total hysterectomy and bilateral salpingo-oophorectomy were performed as a routine. Omentectomy and appendectomy were done for each case. All the tumour-bearing tissue, including pelvic and abdominal peritoneum, serosa of the small bowel, colon and rectum, was resected to the best of the surgeons' skill to achieve a complete or optimal resection. Partial resection of the intestine, with anastomosis or colostomy, was performed when the intestinal wall had been deeply invaded (see Chapter 10). All the excised tissues and organs were labelled according to the sites and put into separate vessels for pathological examination.
Retroperitoneal lymphadenectomy Because the removal of the bulky pelvic mass always provides an easier access to the retroperitoneal structures (see Chapter 5), lymph node dissection was usually carried out at the end of the cytoreductive operation for the intraperitoneal tumours but before resection of the intestine if the latter procedure was indicated. Systemic lymphadenectomy denotes complete lymph node dissection covering all pelvic groups of nodes together with a segment of para-aortic lymph tissue about 4-5 cm in length from the bifurcation of the aorta up to the level of the inferior mesenteric artery. This was accomplished in 69 cases (65.7%). In the remaining 36 cases, only selective sampling of lymph nodes was performed, and in 25 of these cases the para-aortic nodes were not removed. Therefore, a total of 80 samples of para-aortic lymphatic tissue were available for study in the present series. The resected lymphatic tissue was grouped and labelled as para-aortic, presacral, left and right common iliac, external iliac, internal iliac, and obturator, and each tissue was subjected to histological examination individually. RESULTS
Histologically confirmed lymph node metastasis was found in 57 of the 105 cases, providing an overall incidence of 54.3%. The incidence of positive
146
P.-c. w u ET AL
pelvic nodes was 46.7% (49/105), and that of positive para-aortic nodes was 37.5% (30/80). The metastatic rates of the different groups of lymph nodes are shown in Figure 1; these data indicate that ovarian cancer may spread to any group of pelvic nodes and that there was no marked difference of incidence of positive nodes between left and right. The results of the study are analysed as follows.
I• '"'"' I I I I ' "
37.St(30/80)
16.3%
(151921
22.7% (22197)
,,o,..,
,,,..,
20.5% (17183)
21.7% (18/83)
Figure 1. Incidence of metastasis in different groups of lymph nodes. The groups of lymph nodes shown from above downward, and from lateral to medial, are: para-aortic, common iliac, external iliac, obturator and internal iliac. The incidence of presacral nodes is indicated trader the bifurcation of the aorta.
Exploratory evaluation and pathological findings of lymph nodes The gross appearance of the lymph nodes documented at operation was correlated with the pathological findings to assess the accuracy of evaluation during surgical exploration (Table 3). All the evaluations were carried out by two senior surgeons. The hard, enlarged, and/or fixed nodes were tentatively diagnosed as suspicious. In 55 cases with suspicious nodes, 41 (74.5%) proved to be positive histologically. Of the 50 cases in which the nodes were grossly normal, 16 (32%) were found to be positive. Table 3. Exploratory evaluation and pathological findings of lymph nodes. Pathological findings Positive Gross appearance Suspicious Normal Total
Number of cases
Negative
Number
(%)
Number
(%)
55 50
41 16
74.5 32.0
14 34
25.5 68.0
105
57
48
LYMPH NODE METASTASIS AND RETROPERITONEAL LYMPHADENECTOMY
147
Correlation of roentgenological and pathological findings of lymph nodes A comprehensive consideration of the lymphographic appearance of both lymphatics and lymph nodes was used for the radiological diagnosis of metastasis. As shown in Table 4, of the 18 cases with positive lymphographic findings, 15 were proven microscopically to have lymph node metastasis. Of the 12 cases with normal lymphograms, lymph node metastasis was found in two. The positive and negative correspondence rates were both 83.3%. There was no false negativity in the lymphographic diagnosis of the pelvic node metastasis. The only two false negatives in this series occurred on the roentgenograms of the para-aortic nodes. In one case of serous cancer, a Table 4. Correlation between lymphographic findings and pathological
diagnosis of lymph nodes. Pathological findings Positive Lymphographic findings
Negative
Number
(%)
Number
(%)
Total
15 2
83.3 16.7
3 10
16.7 83.3
18 12
Positive Negative
large lymph node, measuring 5 cm in diameter and adherent to the aorta, was resected during operation. Complete destruction of the normal lymphatic structure by the metastatic cancer, as revealed under the microscope, may provide an excuse for the failure of lymphography. The other patient was a case of Stage III endodermal sinus tumour with a 4 x 3 cm para-aortic lymph node metastasis. The lymphography was successful only on the left side. The failure of the radio-opaque media to ascend to a desired level along the para-aortic chain--resulting from the injection of an insufficient amount of media into the lymphatic channels--may explain the negative lymphographic findings.
Correlation of pelvic node and para-aortic node metastasis Systemic lymphadenectomy, in which a complete set of pelvic nodes-together with para-aortic nodes from the same patients--were available for Table 5. Correlation between pelvic node
and para-aortic node metastasis in 69 patients who underwent systemic lymphadenectomy. Lymph node Pelvic
Aortic
Number of cases
(%)
+ve +ve -ve -ve
+ve -ve +ve -ve
19 13 7 30
27.5 18.8 10.2 43.5
69
100.0
Total
148
P.-C. w u
ET AL
comparison, was performed in 69 cases in this series, and lymph node metastasis was found in 39 cases (56.5%). The results are shown in Table 5. Both pelvic and para-aortic nodes were histologically positive in 19 of the 39 cases (48.7%); aortic nodes were positive and pelvic nodes negative in 7 cases (18.0%); and pelvic nodes were positive and aortic nodes negative in 13 cases (33.3%).
Clinical staging and lymph node metastasis Among 16 patients with Stage I disease, 3 (18.8%) were found to have positive nodes, indicating that lymph node metastasis may occur even in very early cases (Table 6). For epithelial cancer, there seems to be a Table 6. Clinical staging and lymph node metastasis.
Epithelial cancer Stage
Germ-cell tumour
Number Number Number Number of cases +ve +ve (%) of cases +ve +ve (%)
I II III IV
7 8 59 3
1 3 38 3
14.3 37.5 64.4 100.0
9 1 13 5
2 1 7 2
22.2 100.0 53.9 40.0
Total
77
45
58.4
28
12
42.9
tendency towards an increased incidence of positive nodes in relation to more advanced clinical stages. However, the difference is not significant on account of the small number of patients with disease of Stages II and IV.
Histological type and lymph node metastasis The incidence of lymph node metastasis in different histological types of tumour is presented in Table 7. The incidence of positive nodes was 58.4% Table 7. Histological type and lymph node metastasis.
Number of cases
Number positive
Positive (%)
Epithelial cancer Serous Clear-ceU Endometrioid Mixed type Brenner Mixed mesodermal Total
53 13 5 3 2 1 77
35 5 2 1 2 0 45
66.0 38.5 40.0 33.3 100.0 0 58.4
Germ-cell turnout Endodermal sinus Immature teratoma Dysgerminoma Mixed type Total
13 9 3 3 28
4 6 2 0 12
30.0 66.7 66.7 0 42.9
105
57
54.3
Histological type
Total
LYMPH NODE METASTASIS AND RETROPERITONEAL LYMPHADENECTOMY
149
for epithelial cancer and 42.9% for germ-cell tumour. The lower incidence of lymphatic metastasis of the latter may be explained by the inclusion of more Stage I patients [9 in 28 (32.1%)] as compared with that of epithelial cancer [7 in 77 (9.1%)] (see Table 6). Histological grading and lymph node metastasis Among 77 cases of epithelial cancer the primary tumour was available for histological grading in 74. The relationship between lymph node metastasis and tumour cell differentiation is shown in Table 8. No positive nodes were Table 8. Histological grading and lymph node metastasis in epithelial cancer.
Number of cases
Number positive
Positive (%)
Borderline Highly differentiated Moderately differentiated Poorly differentiated
4 12 42 16
0 7 28 9
0 58.3 66.7 56.3
Total
74
44
59.5
Histological grading
found in four patients with borderline tumours. In the remaining 70 cases, the degree of differentiation seemed to have no apparent effect on the incidence of lymph node metastasis. Site of the primary cancer and pelvic node metastasis As shown in Table 9, the cancer was bilateral in 40% (42/105) of the cases. The incidence of positive pelvic nodes was the highest (71.4%) in this group of patients. In 38 cases in which the primary cancer originated in left ovary, 17 (44.7%) were found to have positive pelvic nodes, and the chances of spread to lymph nodes of the left and right pelvic wall were equal. However, of 25 cases in which primary cancer arose in the right ovary, only 2 (8%) had metastasis of the ipsilateral pelvic nodes. The difference is statistically highly significant (P<0.01), although the mechanism remains to be clarified. There was no marked difference in the incidence of para-aortic node involvement between cancer originating in left and right ovaries. Table 9. Site of ovarian cancer and pelvic node metastasis.
Positive node Total Site of primary Left ovary Right ovary Bilateral Total
Number of cases
Left
Right
Bilateral
Number
(%)
38 25 42
5 0 3
5 2 5
7 0 22
17 2 30
44.7 8.0 71.4
105
8
12
29
49
46.7
150
P.-C. wu ET AL
Primary or secondary operation and lymph node metastasis In 66 cases the lymphadenectomy was carried out during the primary operation, 54.4% were found to have positive nodes (Table 10); in the remaining
Table 10. Primaryor secondaryoperationin relation to lymph node metastasis. Number Number Operation of cases positive Positive(%) Primary 66 36 54.5 Secondary 39 21 53,9 Total 105 57 54:3 39 cases the operation was performed as a secondary surgical procedure, and the incidence of positive nodes was 53.9%. The result indicates that a previous cytoreductive surgery exerts little influence on the occurrence of lymph node metastasis.
The response of lymph node metastasis to chemotherapeutic agents Twenty-two patients with positive nodes, of which 16 underwent secondary surgical procedures for recurrent tumours and 6 underwent primary operations, received sufficient chemotherapy (at least one full course within 3 months) prior to lymphadenectomy. The drugs used were mainly alkylating agents, with or without combined Cis-platinum for epithelial cancer, and VAC, PVB or 5-fluorouracil (5FU) plus actinomycin D for germ-cell tumours. The lymph nodes obtained from these patients were carefully examined under the microscope for drug effects. AS shown in Table 11,
Table 11. Response of lymph node metastasis to chemotherapeuticagents. Response to chemotherapy Histologicaltype Epithelial cancer Immature teratoma Endodermal sinus tumour
Number Cellular of cases No response degeneration Necrosis 15 3
14 3
1 0
0 0
4
0
4
4
some cellular degeneration of the lymph nodes metastasis was observed in only 1 of the 15 patients with epithelial cancer, and no response at all was noted in the metastatic glial tumour of the lymph nodes in the three cases of teratoma. On the other hand, the endodermal sinus tumour showed a definite response to the chemotherapeutic agents. Cellular degeneration accompanied by extensive necrosis was demonstrated in the metastatic tumours of the lymph nodes in all four cases.
LYMPH NODE METASTASIS AND RETROPERITONEAL LYMPHADENECTOMY
151
Lymph node metastasis and the fate of the patients
The fate of the first 72 patients in this series correlates with lymph node metastasis, All patients were followed-up for at least 21/2 years prior to March, 1988. As shown in Table 12, 28 of the 72 patients (38.9%) survived Table 12. Lymph node metastasis and the fate of the patients.
Lymph node metastasis Yes No
Remission Number of cases Number (%) 42 30
14 14
33.3 46.7
Death Number
(%)
28 16
66.7 53.3
without evidence of recurrence. The rate of complete remission in patients without lymph node metastasis was 46.7% (14/30), but only 33.3% (14/42) for those with lymph node metastasis. However, the difference was not statistically significant ( P > 0.05) on account of the relatively small number of cases.
DISCUSSION Lymphatic drainage of the ovary and routes of lymphatic metastasis in ovarian cancer
According to the traditional conception of anatomy, the lymphatics of an organ always run parallel to the blood vessels. Therefore, it has been believed for many years that the lymphatics of the ovary should take a course identical to that of the ovarian veins, passing between leaves of the infundibulopelvic ligament to the lumbar chains along the abdominal aorta and inferior vena cava. On the basis of this knowledge, a misguided impression has been formed that the destination of the lymphatic drainage of the ovary--and consequently the spreading of ovarian cancer along the lymphatic pathway--should be the para-aortic nodes rather than the pelvic nodes. Although a high incidence of pelvic node involvement in patients with ovarian cancer was reported by Bergman in 1966 according to the autopsy findings, it had been generally regarded as a phenomenon occurring only in the terminal cases resulting from a retrograde flow of tumour cells to the pelvic lymphatic channels due to the blockage of the lumbar chain by tumour emboli and bearing little clinical significance. On the basis of pathological findings in lymph node biopsies, Chen and Lee (1983) found positive lymph nodes in 9 of the 61 cases of epithelial ovarian cancer without concomitant para-aortic node metastasis. Burghardt et al (1984) first reported that the incidence of pelvic lymph node metastasis in ovarian cancer was higher than that of para-aortic node metastasis, and suggested that para-aortic nodes were positive only in cases where pelvic lymph nodes were involved. The high incidence of positive lymph nodes obtained in the
152
P.-C. wu ET hE
present series demonstrates the importance of lymphatic metastasis in ovarian cancer. The higher incidence of positive pelvic nodes (46.7 %) compared with that of positive aortic nodes (37.5%) agrees well with the findings of Burghardt et al (1984). Theoretically, a classical para-aortic lymphadenectomy should include the lymph node dissection of the whole lumbar chain from the bifurcation of the aorta up to the level of crus of diaphragm similar to that carried out by urologists in the retroperitoneal lymphadenectomy for non-seminomatous testicular cancer. However, such an extensive aortocaval lymphadenectomy seems impracticable for ovarian cancer at present in view of the trauma resulting from the intraperitoneal cytoreductive surgery performed at the same time. Therefore, in our cases only the lower segment--instead of the total lumbar lymphatic chain--was removed. Thus, the rate of positive nodes obtained might have been somewhat lower than the real incidence of para-aortic lymph node metastasis. However, it is still reasonable to presume that malignant turnouts of the ovary may have an equal chance of metastasizing along the lymphatic pathways to the aortic and pelvic nodes. Particularly interesting is the result obtained from the 69 patients who underwent systemic lymphadenectomy, a procedure by means of which the pathological findings of the pelvic nodes can be used to correlate with those of the aortic nodes in the same patient. In 39 cases with positive nodes, both pelvic and para-aortic nodes were positive in 19 cases (48.7%). Seven cases (18.0%) had only positive aortic nodes, and 13 (33.3%) had only positive pelvic nodes. In a monograph, Plentl and Friedman (1971) pointed out that, in addition to the classical lymphatic pathway along the ovarian vessels, there are lymph trunks from the ovary coursing within the folds of the broad ligaments and terminating directly in the iliac nodes. The findings of our study further confirm the statement that the lymphatic drainage from the ovary may carry the tumour cells directly to nodes on the pelvic wall without traversing the para-aortic nodes. Clinical significance of lymphadenectomy in ovarian cancer
For staging evaluation Since the revision of the FIGO staging classification for ovarian cancer in 1975 (Day and Smith, 1975), positive retroperitoneal lymph nodes have been included in the criteria of Stage III disease. Whenever the retroperitoneal nodes are involved, the patient should be categorized as Stage III instead of Stage I--even if the primary cancer appears to be confined to the ovary. This revision of the staging criteria requires a re-evaluation of the traditional methods in assessing the stage of patients with ovarian cancer. A thorough exploration during operation, although very important for the accurate assessment of the extent of the intraperitoneal tumours, may be misleading in judging the status of the retroperitoneal lymph nodes. In this series, only three-quarters of the cases with positive nodes and two-thirds of the cases with negative nodes could be correctly diagnosed during exploration, indicating that evaluation of the lymph node involvement by gross
LYMPH NODE METASTASIS AND RETROPERITONEAL LYlVlPHADENECTOMY
153
appearance is not adequately reliable. Lymphography may be a very useful measure for detecting lymph node metastasis, but difficulties in interpreting the lymphographic manifestation on account of lack of studies correlating radiological abnormalities with histological findings limit the clinical application of this technique. In 30 of the cases in this series, both lymphographic diagnosis and pathological examination of the resected lymphatic tissue were available from the same patient for correlation. Of the 18 cases with positive lymphographic findings, three were proven microscopically to be false-positive cases (16.7%). Of the 12 cases with normal lymphograms, two were proven to be false-negative cases (16.7%). Therefore, the most reliable method of assessing lymph node involvement in ovarian cancer at present is to subject the resected lymphatic tissue to pathological examination. In this series, 7 of the 25 cases (25%) categorized as Stage I or II during surgical exploration were found to have positive nodes and should be reclassified as Stage III.
For therapeutic purposes The current policy in the operative management of ovarian cancer is for surgeons to do their best to remove all the primary and metastatic turnouts in the peritoneal cavity. According to the findings of the present series, and others (Chert and Lee, 1983; Burghardt et al, 1984), the incidence of retroperitoneal nodal involvement is about 50%. Therefore, all attempts at improving the fate of patients with ovarian cancer will be futile if the presence of this important route of metastasis--which may transport cancer cells to any part of the body--is overlooked. On the basis of the findings of a limited number of cases in this series, the metastatic tumours of the lymph nodes were found to be unresponsive to chemotherapeutic agents in both epithelial cancer and immature teratoma. Therefore, from a therapeutic point of view, retroperitoneal lymphadenectomy will not only lessen the tumour burden by removing the cancer-laden nodes, it might possibly block the further upward spread of the disease along the lymphatic channels. In the literature, reports on the evaluation of lymphadenectomy in the treatment of ovarian cancer are few. Burghardt et al (1986), in his pioneering work, compared the 5-year actuarial survival rate for Stage III ovarian cancer after operative treatment. In 70 cases in which lymphadenectomy was carried out the survival rate was 53.0%, while in 40 cases in which lymphadenectomy was not carried out the survival rate was only 13.0%. Unfortunately, our data cannot be used to assess the role of lymphadenectomy in the current series on account of the lack of randomized controls. However, the results achieved by Burghardt et al enlighten us as to the possibility of improving the prognosis of ovarian cancer through retroperitoneal lymphadenectomy. SUMMARY
While clinicians are devoting themselves to the study of the behaviour of
154
p.-c. w u ET AL
ovarian cancer as well as to the search for more effective therapeutic modalities, little attention has been paid to an important route of metastasis in this group of diseases: retroperitoneal spread through the lymphatic pathway. The purpose of this report is to present a 5-year experience of a prospective study on lymph node metastasis in patients with ovarian cancer through retroperitoneal lymphadenectomy at the Peking Union Medical College Hospital. From June 1982 through May 1987, retroperitoneal lymph node dissection was performed in 105 cases of ovarian cancer. Seventyseven (73.3%) were histologically diagnosed as cancer of epithelial origin, and 28 (26.7%) as germ-cell tumours. The overall incidence of retroperitoneal positive nodes was 54.3% (57/105). The incidence of positive pelvic nodes was 46.7% (49/105), and that of positive para-aortic nodes was 37.5% (30/80). In 69 patients who underwent systemic lymphadenectomy, 39 were found to have glandular involvement; in these 39 patients both aortic and pelvic nodes were positive in 19 cases (48.7%), aortic nodes were positive and pelvic nodes negative in 7 cases (18.0%), and pelvic nodes were positive and aortic nodes negative in 13 cases (33.3%). Preoperative lymphography was performed in 30 cases. The positive and negative correspondence rates with the pathological findings were both 83.3%. In 38 cases in which the primary cancer originated in the left ovary, 17 (44.7%) were found to have positive pelvic nodes, whereas in 25 cases with primary cancer arising in the right ovary only 2 (8%) had metastasis of the ipsilateral pelvic nodes. The lymph nodes obtained from 22 patients with positive nodes and sufficient preoperative chemotherapy were carefully examined under the microscope for the effects of the drugs. Some cellular degeneration of the lymph node metastasis was observed in only one of the 15 cases of epithelial cancer, and no response at all was noted in three cases of immature teratoma. Cellular degeneration accompanied by extensive necrosis was demonstrated in the metastatic tumours of the lymph nodes in all four cases of endodermal sinus tumour. Seventy-two patients were followed-up for at least 21/2 years. The rate of complete remission was 46.7% (14/30) in patients with negative nodes, but only 33.3% (14/42) in those with positive nodes. However, the difference was not statistically significant.
REFERENCES Bergrnan F (1966) Carcinoma of the ovary: a clinicopathological study of 86 autopsied cases with special reference to mode of spread. Acta Obstetricaet GynecologicaScandinavica 45: 211-232. Burghardt E, Pickel H & Stettner H (1984) Management of advanced ovarian cancer. European Journal of Gynecologic Oncology 3: 155-159. Burghardt E, Pickel H, Lahousen M & Stettner H (1986) Pelvic lymphadenectomy in operative treatment of ovarian cancer. American Journal of Obstetrics and Gynecology 155: 315319. Chen SS & Lee L (1983) Incidence of para-aortic and peMc lymph node metastasis in epithelial carcinoma of the ovary. Gynecologic OncoIogy 16: 95-100. Day TG & Smith JP (1975) Diagnosis and staging of ovarian cancer. Seminar Oncology 2: 217-222,
LYMPH NODE METASTASISAND RETROPERITONEALLYMPHADENECTOMY
155
Knapp RC & Friedman EA (1974) Aortic lymph node metastasis in early ovarian cancer. American Journal of Obstetrics and Gynecology 119: 1013-1017. Plentl AA & Friedman EA (1971) Lymphatic System of the Female Genitalia. Philadelphia: WB Saunders. Wu PC, Qu JY, Lang JH, Huang RL, Tang NY & Lian LJ (1984) Lymph node metastasis in ovarian malignancy. Chinese Journal of Obstetrics and Gynecology 19(supplement): 76-80. Wu PC, Qu JY, Lang JH, Huang RL, Tang MY & Lian LJ (1986) Lymph node metastasis of ovarian cancer: a preliminary survey of 74 cases of lymphadenectomy. American Journal of Obstetrics and Gynecology 155: 1103-1108.
Cancer of the ovary has become increasingly important in the last few decades and ranks as the leading cause of death among all gynaecological malignancies. While clinicians are devoting themselves to the study of the behaviour of ovarian cancer, as well as to the search for more effective therapeutic modalities, little attention has been paid to an important route of metastasis in this group of diseases: retroperitoneal spread through the lymphatic pathway. Two possible explanations can be tentatively made for this negligence. In the first place, the silent metastasis of ovarian cancer along the lymphatic pathway is overshadowed by the outstanding extensive intraperitoneal implantation and local invasion to adjacent tissues and/or organs which divert the attention of most gynaecologists and oncologists. Secondly, the traditional anatomical conception about lymphatic drainage of the ovary leads clinicians to believe that once retroperitoneal lymphatic metastasis has occurred the first group of lymph nodes involved will always be the para-aortic nodes--nodes which are situated high above the pelvis and far beyond the reach of conventional surgical and radiation therapy. In 1974, Knapp and Friedman first reported their experience with aortic lymph node metastasis in 26 patients with Stage I ovarian cancer subjected to para-aortic lymphadenectomy. In a prospective study by selective nodal biopsy in 61 unselected cases of ovarian cancer, Chen and Lee (1983) found that the incidence of metastasis was 37.7% for para-aortic nodes and 14.8% for pelvic nodes. Burghardt et al (1984) first emphasized the significance of retroperitoneal lymph node dissection in ovarian cancer from both the diagnostic and therapeutic points of view. In 48 patients undergoing systemic pelvic lymphadenectomy, the incidence of positive pelvic nodes was 56.3%, and the incidence of positive para-aortic nodes in 16 evaluated cases Bailli~re's Clinical Obstetrics and Gynaecology--Vol. 3, No. 1, March 1989
143
144
P.-c. w u ET AL
was 31.2%. Beginning in 1982, retroperitoneal lymphadenectomy has been performed at the Peking Union Medical College Hospital as a routine procedure in addition to the cytoreductive surgery for most patients with ovarian malignancies. The purpose of this paper is to present our five-year experience of a prospective study on lymph node metastasis in order to assess the role of lymphadenectomy in the management of ovarian cancer. Some of these cases were reported previously (Wu et al, 1984; 1986). MATERIAL AND METHODS During a period of 5 years from June 1982 through May 1987, retroperitoneal lymph node dissection was performed in 105 cases of ovarian cancer. The general clinical data are shown in Table 1. Ovarian cancer of epithelial Table 1. Subjects under study (n = 105). Number of cases (%) Histological type Epithelial cancer Germ-cell tumour
77 (73.3) 28 (26.7)
Clinical staging (FIGO) I II III IV
16 9 72 8
Primary or secondary surgery Primary Secondary
66 (62.9) 39 (37.1)
(15.2) (8.6) (68.6) (7.6)
origin constituted the majority of the cases, of which 53 (68.8%) were of the serous type. Clinical staging was based on the FIGO criteria assessed during laparotomy without considering the pathological findings of the lymph nodes. It can be seen that patients with Stage III and IV disease accounted for 76.2% of the total series, and only 16 patients had Stage I disease. Sixty-six patients (62.9%) were operated on for the first time (primary operation), and secondary operations were performed in the remaining 39 cases for recurrent or incompletely resected tumours. Lymphography In order to evaluate the role of roentgenological examination in predicting lymph node metastasis in ovarian cancer, lymphography was performed preoperatively in 30 cases. Six millilitres of Myodil were injected into the lymphatics between the second and third toes on the dorsum of each foot within 2 hours by means of an injection pump. X-ray films were taken of the anterio-posterior, left anterior oblique and right anterior oblique views immediately after injection of the radio-opaque media, and were repeated 24 hours later.
LYMPH NODE METASTASIS AND RETROPERITONEAL LYMPHADENECTOMY
145
Methods of surgery
The surgical methods used are indicated in Table 2. Table 2, Methods used in surgery. Number of cases Cytoreductive surgery Retroperitoneal lymphadenectomy Systemic Selective sampling
105 69 36*
*Aortic nodes were not removed in 25 cases.
Cytoreductive surgery With the exception of a few young patients with unilateral Stage I germ-cell tumour, total hysterectomy and bilateral salpingo-oophorectomy were performed as a routine. Omentectomy and appendectomy were done for each case. All the tumour-bearing tissue, including pelvic and abdominal peritoneum, serosa of the small bowel, colon and rectum, was resected to the best of the surgeons' skill to achieve a complete or optimal resection. Partial resection of the intestine, with anastomosis or colostomy, was performed when the intestinal wall had been deeply invaded (see Chapter 10). All the excised tissues and organs were labelled according to the sites and put into separate vessels for pathological examination.
Retroperitoneal lymphadenectomy Because the removal of the bulky pelvic mass always provides an easier access to the retroperitoneal structures (see Chapter 5), lymph node dissection was usually carried out at the end of the cytoreductive operation for the intraperitoneal tumours but before resection of the intestine if the latter procedure was indicated. Systemic lymphadenectomy denotes complete lymph node dissection covering all pelvic groups of nodes together with a segment of para-aortic lymph tissue about 4-5 cm in length from the bifurcation of the aorta up to the level of the inferior mesenteric artery. This was accomplished in 69 cases (65.7%). In the remaining 36 cases, only selective sampling of lymph nodes was performed, and in 25 of these cases the para-aortic nodes were not removed. Therefore, a total of 80 samples of para-aortic lymphatic tissue were available for study in the present series. The resected lymphatic tissue was grouped and labelled as para-aortic, presacral, left and right common iliac, external iliac, internal iliac, and obturator, and each tissue was subjected to histological examination individually. RESULTS
Histologically confirmed lymph node metastasis was found in 57 of the 105 cases, providing an overall incidence of 54.3%. The incidence of positive
146
P.-c. w u ET AL
pelvic nodes was 46.7% (49/105), and that of positive para-aortic nodes was 37.5% (30/80). The metastatic rates of the different groups of lymph nodes are shown in Figure 1; these data indicate that ovarian cancer may spread to any group of pelvic nodes and that there was no marked difference of incidence of positive nodes between left and right. The results of the study are analysed as follows.
I• '"'"' I I I I ' "
37.St(30/80)
16.3%
(151921
22.7% (22197)
,,o,..,
,,,..,
20.5% (17183)
21.7% (18/83)
Figure 1. Incidence of metastasis in different groups of lymph nodes. The groups of lymph nodes shown from above downward, and from lateral to medial, are: para-aortic, common iliac, external iliac, obturator and internal iliac. The incidence of presacral nodes is indicated trader the bifurcation of the aorta.
Exploratory evaluation and pathological findings of lymph nodes The gross appearance of the lymph nodes documented at operation was correlated with the pathological findings to assess the accuracy of evaluation during surgical exploration (Table 3). All the evaluations were carried out by two senior surgeons. The hard, enlarged, and/or fixed nodes were tentatively diagnosed as suspicious. In 55 cases with suspicious nodes, 41 (74.5%) proved to be positive histologically. Of the 50 cases in which the nodes were grossly normal, 16 (32%) were found to be positive. Table 3. Exploratory evaluation and pathological findings of lymph nodes. Pathological findings Positive Gross appearance Suspicious Normal Total
Number of cases
Negative
Number
(%)
Number
(%)
55 50
41 16
74.5 32.0
14 34
25.5 68.0
105
57
48
LYMPH NODE METASTASIS AND RETROPERITONEAL LYMPHADENECTOMY
147
Correlation of roentgenological and pathological findings of lymph nodes A comprehensive consideration of the lymphographic appearance of both lymphatics and lymph nodes was used for the radiological diagnosis of metastasis. As shown in Table 4, of the 18 cases with positive lymphographic findings, 15 were proven microscopically to have lymph node metastasis. Of the 12 cases with normal lymphograms, lymph node metastasis was found in two. The positive and negative correspondence rates were both 83.3%. There was no false negativity in the lymphographic diagnosis of the pelvic node metastasis. The only two false negatives in this series occurred on the roentgenograms of the para-aortic nodes. In one case of serous cancer, a Table 4. Correlation between lymphographic findings and pathological
diagnosis of lymph nodes. Pathological findings Positive Lymphographic findings
Negative
Number
(%)
Number
(%)
Total
15 2
83.3 16.7
3 10
16.7 83.3
18 12
Positive Negative
large lymph node, measuring 5 cm in diameter and adherent to the aorta, was resected during operation. Complete destruction of the normal lymphatic structure by the metastatic cancer, as revealed under the microscope, may provide an excuse for the failure of lymphography. The other patient was a case of Stage III endodermal sinus tumour with a 4 x 3 cm para-aortic lymph node metastasis. The lymphography was successful only on the left side. The failure of the radio-opaque media to ascend to a desired level along the para-aortic chain--resulting from the injection of an insufficient amount of media into the lymphatic channels--may explain the negative lymphographic findings.
Correlation of pelvic node and para-aortic node metastasis Systemic lymphadenectomy, in which a complete set of pelvic nodes-together with para-aortic nodes from the same patients--were available for Table 5. Correlation between pelvic node
and para-aortic node metastasis in 69 patients who underwent systemic lymphadenectomy. Lymph node Pelvic
Aortic
Number of cases
(%)
+ve +ve -ve -ve
+ve -ve +ve -ve
19 13 7 30
27.5 18.8 10.2 43.5
69
100.0
Total
148
P.-C. w u
ET AL
comparison, was performed in 69 cases in this series, and lymph node metastasis was found in 39 cases (56.5%). The results are shown in Table 5. Both pelvic and para-aortic nodes were histologically positive in 19 of the 39 cases (48.7%); aortic nodes were positive and pelvic nodes negative in 7 cases (18.0%); and pelvic nodes were positive and aortic nodes negative in 13 cases (33.3%).
Clinical staging and lymph node metastasis Among 16 patients with Stage I disease, 3 (18.8%) were found to have positive nodes, indicating that lymph node metastasis may occur even in very early cases (Table 6). For epithelial cancer, there seems to be a Table 6. Clinical staging and lymph node metastasis.
Epithelial cancer Stage
Germ-cell tumour
Number Number Number Number of cases +ve +ve (%) of cases +ve +ve (%)
I II III IV
7 8 59 3
1 3 38 3
14.3 37.5 64.4 100.0
9 1 13 5
2 1 7 2
22.2 100.0 53.9 40.0
Total
77
45
58.4
28
12
42.9
tendency towards an increased incidence of positive nodes in relation to more advanced clinical stages. However, the difference is not significant on account of the small number of patients with disease of Stages II and IV.
Histological type and lymph node metastasis The incidence of lymph node metastasis in different histological types of tumour is presented in Table 7. The incidence of positive nodes was 58.4% Table 7. Histological type and lymph node metastasis.
Number of cases
Number positive
Positive (%)
Epithelial cancer Serous Clear-ceU Endometrioid Mixed type Brenner Mixed mesodermal Total
53 13 5 3 2 1 77
35 5 2 1 2 0 45
66.0 38.5 40.0 33.3 100.0 0 58.4
Germ-cell turnout Endodermal sinus Immature teratoma Dysgerminoma Mixed type Total
13 9 3 3 28
4 6 2 0 12
30.0 66.7 66.7 0 42.9
105
57
54.3
Histological type
Total
LYMPH NODE METASTASIS AND RETROPERITONEAL LYMPHADENECTOMY
149
for epithelial cancer and 42.9% for germ-cell tumour. The lower incidence of lymphatic metastasis of the latter may be explained by the inclusion of more Stage I patients [9 in 28 (32.1%)] as compared with that of epithelial cancer [7 in 77 (9.1%)] (see Table 6). Histological grading and lymph node metastasis Among 77 cases of epithelial cancer the primary tumour was available for histological grading in 74. The relationship between lymph node metastasis and tumour cell differentiation is shown in Table 8. No positive nodes were Table 8. Histological grading and lymph node metastasis in epithelial cancer.
Number of cases
Number positive
Positive (%)
Borderline Highly differentiated Moderately differentiated Poorly differentiated
4 12 42 16
0 7 28 9
0 58.3 66.7 56.3
Total
74
44
59.5
Histological grading
found in four patients with borderline tumours. In the remaining 70 cases, the degree of differentiation seemed to have no apparent effect on the incidence of lymph node metastasis. Site of the primary cancer and pelvic node metastasis As shown in Table 9, the cancer was bilateral in 40% (42/105) of the cases. The incidence of positive pelvic nodes was the highest (71.4%) in this group of patients. In 38 cases in which the primary cancer originated in left ovary, 17 (44.7%) were found to have positive pelvic nodes, and the chances of spread to lymph nodes of the left and right pelvic wall were equal. However, of 25 cases in which primary cancer arose in the right ovary, only 2 (8%) had metastasis of the ipsilateral pelvic nodes. The difference is statistically highly significant (P<0.01), although the mechanism remains to be clarified. There was no marked difference in the incidence of para-aortic node involvement between cancer originating in left and right ovaries. Table 9. Site of ovarian cancer and pelvic node metastasis.
Positive node Total Site of primary Left ovary Right ovary Bilateral Total
Number of cases
Left
Right
Bilateral
Number
(%)
38 25 42
5 0 3
5 2 5
7 0 22
17 2 30
44.7 8.0 71.4
105
8
12
29
49
46.7
150
P.-C. wu ET AL
Primary or secondary operation and lymph node metastasis In 66 cases the lymphadenectomy was carried out during the primary operation, 54.4% were found to have positive nodes (Table 10); in the remaining
Table 10. Primaryor secondaryoperationin relation to lymph node metastasis. Number Number Operation of cases positive Positive(%) Primary 66 36 54.5 Secondary 39 21 53,9 Total 105 57 54:3 39 cases the operation was performed as a secondary surgical procedure, and the incidence of positive nodes was 53.9%. The result indicates that a previous cytoreductive surgery exerts little influence on the occurrence of lymph node metastasis.
The response of lymph node metastasis to chemotherapeutic agents Twenty-two patients with positive nodes, of which 16 underwent secondary surgical procedures for recurrent tumours and 6 underwent primary operations, received sufficient chemotherapy (at least one full course within 3 months) prior to lymphadenectomy. The drugs used were mainly alkylating agents, with or without combined Cis-platinum for epithelial cancer, and VAC, PVB or 5-fluorouracil (5FU) plus actinomycin D for germ-cell tumours. The lymph nodes obtained from these patients were carefully examined under the microscope for drug effects. AS shown in Table 11,
Table 11. Response of lymph node metastasis to chemotherapeuticagents. Response to chemotherapy Histologicaltype Epithelial cancer Immature teratoma Endodermal sinus tumour
Number Cellular of cases No response degeneration Necrosis 15 3
14 3
1 0
0 0
4
0
4
4
some cellular degeneration of the lymph nodes metastasis was observed in only 1 of the 15 patients with epithelial cancer, and no response at all was noted in the metastatic glial tumour of the lymph nodes in the three cases of teratoma. On the other hand, the endodermal sinus tumour showed a definite response to the chemotherapeutic agents. Cellular degeneration accompanied by extensive necrosis was demonstrated in the metastatic tumours of the lymph nodes in all four cases.
LYMPH NODE METASTASIS AND RETROPERITONEAL LYMPHADENECTOMY
151
Lymph node metastasis and the fate of the patients
The fate of the first 72 patients in this series correlates with lymph node metastasis, All patients were followed-up for at least 21/2 years prior to March, 1988. As shown in Table 12, 28 of the 72 patients (38.9%) survived Table 12. Lymph node metastasis and the fate of the patients.
Lymph node metastasis Yes No
Remission Number of cases Number (%) 42 30
14 14
33.3 46.7
Death Number
(%)
28 16
66.7 53.3
without evidence of recurrence. The rate of complete remission in patients without lymph node metastasis was 46.7% (14/30), but only 33.3% (14/42) for those with lymph node metastasis. However, the difference was not statistically significant ( P > 0.05) on account of the relatively small number of cases.
DISCUSSION Lymphatic drainage of the ovary and routes of lymphatic metastasis in ovarian cancer
According to the traditional conception of anatomy, the lymphatics of an organ always run parallel to the blood vessels. Therefore, it has been believed for many years that the lymphatics of the ovary should take a course identical to that of the ovarian veins, passing between leaves of the infundibulopelvic ligament to the lumbar chains along the abdominal aorta and inferior vena cava. On the basis of this knowledge, a misguided impression has been formed that the destination of the lymphatic drainage of the ovary--and consequently the spreading of ovarian cancer along the lymphatic pathway--should be the para-aortic nodes rather than the pelvic nodes. Although a high incidence of pelvic node involvement in patients with ovarian cancer was reported by Bergman in 1966 according to the autopsy findings, it had been generally regarded as a phenomenon occurring only in the terminal cases resulting from a retrograde flow of tumour cells to the pelvic lymphatic channels due to the blockage of the lumbar chain by tumour emboli and bearing little clinical significance. On the basis of pathological findings in lymph node biopsies, Chen and Lee (1983) found positive lymph nodes in 9 of the 61 cases of epithelial ovarian cancer without concomitant para-aortic node metastasis. Burghardt et al (1984) first reported that the incidence of pelvic lymph node metastasis in ovarian cancer was higher than that of para-aortic node metastasis, and suggested that para-aortic nodes were positive only in cases where pelvic lymph nodes were involved. The high incidence of positive lymph nodes obtained in the
152
P.-C. wu ET hE
present series demonstrates the importance of lymphatic metastasis in ovarian cancer. The higher incidence of positive pelvic nodes (46.7 %) compared with that of positive aortic nodes (37.5%) agrees well with the findings of Burghardt et al (1984). Theoretically, a classical para-aortic lymphadenectomy should include the lymph node dissection of the whole lumbar chain from the bifurcation of the aorta up to the level of crus of diaphragm similar to that carried out by urologists in the retroperitoneal lymphadenectomy for non-seminomatous testicular cancer. However, such an extensive aortocaval lymphadenectomy seems impracticable for ovarian cancer at present in view of the trauma resulting from the intraperitoneal cytoreductive surgery performed at the same time. Therefore, in our cases only the lower segment--instead of the total lumbar lymphatic chain--was removed. Thus, the rate of positive nodes obtained might have been somewhat lower than the real incidence of para-aortic lymph node metastasis. However, it is still reasonable to presume that malignant turnouts of the ovary may have an equal chance of metastasizing along the lymphatic pathways to the aortic and pelvic nodes. Particularly interesting is the result obtained from the 69 patients who underwent systemic lymphadenectomy, a procedure by means of which the pathological findings of the pelvic nodes can be used to correlate with those of the aortic nodes in the same patient. In 39 cases with positive nodes, both pelvic and para-aortic nodes were positive in 19 cases (48.7%). Seven cases (18.0%) had only positive aortic nodes, and 13 (33.3%) had only positive pelvic nodes. In a monograph, Plentl and Friedman (1971) pointed out that, in addition to the classical lymphatic pathway along the ovarian vessels, there are lymph trunks from the ovary coursing within the folds of the broad ligaments and terminating directly in the iliac nodes. The findings of our study further confirm the statement that the lymphatic drainage from the ovary may carry the tumour cells directly to nodes on the pelvic wall without traversing the para-aortic nodes. Clinical significance of lymphadenectomy in ovarian cancer
For staging evaluation Since the revision of the FIGO staging classification for ovarian cancer in 1975 (Day and Smith, 1975), positive retroperitoneal lymph nodes have been included in the criteria of Stage III disease. Whenever the retroperitoneal nodes are involved, the patient should be categorized as Stage III instead of Stage I--even if the primary cancer appears to be confined to the ovary. This revision of the staging criteria requires a re-evaluation of the traditional methods in assessing the stage of patients with ovarian cancer. A thorough exploration during operation, although very important for the accurate assessment of the extent of the intraperitoneal tumours, may be misleading in judging the status of the retroperitoneal lymph nodes. In this series, only three-quarters of the cases with positive nodes and two-thirds of the cases with negative nodes could be correctly diagnosed during exploration, indicating that evaluation of the lymph node involvement by gross
LYMPH NODE METASTASIS AND RETROPERITONEAL LYlVlPHADENECTOMY
153
appearance is not adequately reliable. Lymphography may be a very useful measure for detecting lymph node metastasis, but difficulties in interpreting the lymphographic manifestation on account of lack of studies correlating radiological abnormalities with histological findings limit the clinical application of this technique. In 30 of the cases in this series, both lymphographic diagnosis and pathological examination of the resected lymphatic tissue were available from the same patient for correlation. Of the 18 cases with positive lymphographic findings, three were proven microscopically to be false-positive cases (16.7%). Of the 12 cases with normal lymphograms, two were proven to be false-negative cases (16.7%). Therefore, the most reliable method of assessing lymph node involvement in ovarian cancer at present is to subject the resected lymphatic tissue to pathological examination. In this series, 7 of the 25 cases (25%) categorized as Stage I or II during surgical exploration were found to have positive nodes and should be reclassified as Stage III.
For therapeutic purposes The current policy in the operative management of ovarian cancer is for surgeons to do their best to remove all the primary and metastatic turnouts in the peritoneal cavity. According to the findings of the present series, and others (Chert and Lee, 1983; Burghardt et al, 1984), the incidence of retroperitoneal nodal involvement is about 50%. Therefore, all attempts at improving the fate of patients with ovarian cancer will be futile if the presence of this important route of metastasis--which may transport cancer cells to any part of the body--is overlooked. On the basis of the findings of a limited number of cases in this series, the metastatic tumours of the lymph nodes were found to be unresponsive to chemotherapeutic agents in both epithelial cancer and immature teratoma. Therefore, from a therapeutic point of view, retroperitoneal lymphadenectomy will not only lessen the tumour burden by removing the cancer-laden nodes, it might possibly block the further upward spread of the disease along the lymphatic channels. In the literature, reports on the evaluation of lymphadenectomy in the treatment of ovarian cancer are few. Burghardt et al (1986), in his pioneering work, compared the 5-year actuarial survival rate for Stage III ovarian cancer after operative treatment. In 70 cases in which lymphadenectomy was carried out the survival rate was 53.0%, while in 40 cases in which lymphadenectomy was not carried out the survival rate was only 13.0%. Unfortunately, our data cannot be used to assess the role of lymphadenectomy in the current series on account of the lack of randomized controls. However, the results achieved by Burghardt et al enlighten us as to the possibility of improving the prognosis of ovarian cancer through retroperitoneal lymphadenectomy. SUMMARY
While clinicians are devoting themselves to the study of the behaviour of
154
p.-c. w u ET AL
ovarian cancer as well as to the search for more effective therapeutic modalities, little attention has been paid to an important route of metastasis in this group of diseases: retroperitoneal spread through the lymphatic pathway. The purpose of this report is to present a 5-year experience of a prospective study on lymph node metastasis in patients with ovarian cancer through retroperitoneal lymphadenectomy at the Peking Union Medical College Hospital. From June 1982 through May 1987, retroperitoneal lymph node dissection was performed in 105 cases of ovarian cancer. Seventyseven (73.3%) were histologically diagnosed as cancer of epithelial origin, and 28 (26.7%) as germ-cell tumours. The overall incidence of retroperitoneal positive nodes was 54.3% (57/105). The incidence of positive pelvic nodes was 46.7% (49/105), and that of positive para-aortic nodes was 37.5% (30/80). In 69 patients who underwent systemic lymphadenectomy, 39 were found to have glandular involvement; in these 39 patients both aortic and pelvic nodes were positive in 19 cases (48.7%), aortic nodes were positive and pelvic nodes negative in 7 cases (18.0%), and pelvic nodes were positive and aortic nodes negative in 13 cases (33.3%). Preoperative lymphography was performed in 30 cases. The positive and negative correspondence rates with the pathological findings were both 83.3%. In 38 cases in which the primary cancer originated in the left ovary, 17 (44.7%) were found to have positive pelvic nodes, whereas in 25 cases with primary cancer arising in the right ovary only 2 (8%) had metastasis of the ipsilateral pelvic nodes. The lymph nodes obtained from 22 patients with positive nodes and sufficient preoperative chemotherapy were carefully examined under the microscope for the effects of the drugs. Some cellular degeneration of the lymph node metastasis was observed in only one of the 15 cases of epithelial cancer, and no response at all was noted in three cases of immature teratoma. Cellular degeneration accompanied by extensive necrosis was demonstrated in the metastatic tumours of the lymph nodes in all four cases of endodermal sinus tumour. Seventy-two patients were followed-up for at least 21/2 years. The rate of complete remission was 46.7% (14/30) in patients with negative nodes, but only 33.3% (14/42) in those with positive nodes. However, the difference was not statistically significant.
REFERENCES Bergrnan F (1966) Carcinoma of the ovary: a clinicopathological study of 86 autopsied cases with special reference to mode of spread. Acta Obstetricaet GynecologicaScandinavica 45: 211-232. Burghardt E, Pickel H & Stettner H (1984) Management of advanced ovarian cancer. European Journal of Gynecologic Oncology 3: 155-159. Burghardt E, Pickel H, Lahousen M & Stettner H (1986) Pelvic lymphadenectomy in operative treatment of ovarian cancer. American Journal of Obstetrics and Gynecology 155: 315319. Chen SS & Lee L (1983) Incidence of para-aortic and peMc lymph node metastasis in epithelial carcinoma of the ovary. Gynecologic OncoIogy 16: 95-100. Day TG & Smith JP (1975) Diagnosis and staging of ovarian cancer. Seminar Oncology 2: 217-222,
LYMPH NODE METASTASISAND RETROPERITONEALLYMPHADENECTOMY
155
Knapp RC & Friedman EA (1974) Aortic lymph node metastasis in early ovarian cancer. American Journal of Obstetrics and Gynecology 119: 1013-1017. Plentl AA & Friedman EA (1971) Lymphatic System of the Female Genitalia. Philadelphia: WB Saunders. Wu PC, Qu JY, Lang JH, Huang RL, Tang NY & Lian LJ (1984) Lymph node metastasis in ovarian malignancy. Chinese Journal of Obstetrics and Gynecology 19(supplement): 76-80. Wu PC, Qu JY, Lang JH, Huang RL, Tang MY & Lian LJ (1986) Lymph node metastasis of ovarian cancer: a preliminary survey of 74 cases of lymphadenectomy. American Journal of Obstetrics and Gynecology 155: 1103-1108.