Technique and feasibility of radical para-aortic and pelvic lymphadenectomy for gynecologic malignancies: a prospective study

Int J Gynecol Cancer 1991,1,133-140

Technique and feasibility of radical para-aortic and pelvic lymphadenectomy for gynecologic malignancies: a prospective study P. BENEDETTI. PANICI, G . SCAMBIA, G . BAIOCCHI, S. GREGGI & S . MANCUSO Department of Gynecology and obstetrics, Catholic University, Rome

Abstract. Benedetti Panici P, Scambia G, Baiocchi G, Greggi S, Mancuso S. Technique and feasibility of radical para-aortic and pelvic lymphadenectomy for gynecologicmalignancies: a prospective study. Int J Gynecol Cancer 1991; 1: 133-140. Of 284 patients evaluated for entry into the study between January 1986 and June 1990, systematic para-aortic and pelvic lymphadenectomy was performed in 208 cases (108 cervical cancer, 43 and 57 ovarian and endometrial cancer, respectively). The median number of nodes removed was 58, 49 and 54 for cervical, ovarian and endometrial cancer, respectively. The operating data are divided into 2 groups according to the consecutive number of the cases. The median operating time and the median estimated blood loss of lymphadenectomy was 230 minutes (range 120270) and 390 ml (range 200-3300) in the first 95 cases. These operating data decreased to 150 minutes (range 100-240) and 250 ml (range 100- 2800) in the second 113 cases. No surgery-related deaths occurred. Severe hemorrages (blood loss exceeding 1000 ml) occurred in 6 patients. The obturator nerve was dissected in 1 patient and in 1 case the left ureter was cut. Formation of lymphoceles occurred in 20.4% of patients. Eighteen patients (8.8%)developed deep venous thrombosis. Nine of these patients experienced pulmonary microembolism. In 3 patients a retroperitoneal abscess was diagnosed. One patient developed a fistula of the most proximal part of the right ureter during the third postoperative week. The resection or coagulation of branches of the genito-femoral and obturator nerves determined mild paresthesis localized at the supero-anterior and internal side of thigh in 11 cases (5.4%). No statistically significant differences were found between the clinical (age, weight and previous chemotherapy) and pathological (type of cancer and lymph node status) parameters considered on one hand and postoperative complications on the other. pelvic lymphadenectomy, para-aortic lymphadenectomy, gynecologic malignancies, gynecologic surgery, surgical techniques.

KEYWORDS:

It is well recognized that the lymphatic system is a major pathway of spread in gynecologic cancer. Therefore the evaluation of the lymph node status plays an important role in the diagnosis and prognosis Address for correspondence: Salvatore Mancuso, MD, Catholic University, Department of Gynecology and Obstetrics, Largo A. Gemelli, 8 00168 Rome, Italy.

of patients with gynecological malignancies. Although the diagnosis of metastatic disease using non-invasive techniques (lymphangiography, immunoscintigraphy, ultrasound, computerized axial tomography, magnetic resonance imaging) is attractive, each Of these procedures has significant negative and false positive rates('-3). Therefore, lymphadenectomy remains the most reliable means

134 P.B. Punici et al. for ascertaining the presence of lymph node meta~tases(”~). Moreover, taking into account that complete eradication of all cancer tissues must be the goal of radical surgery for a malignant condition, the node dissection should also be considered as an integral part of the surgical treatment, in particular for tumors with locoregional spread. On the other hand, more aggressive surgery in staging and treating patients with gynecologic malignancies is made feasible by improvements in anesthesia and perioperative care. Chevassu first described the extraperitoneal node dis~ection(~). Since then many different approaches have been designed and performed for the management of testicular By contrast, few studies have reported the detailed technique of systematic pelvic and para-aortic lymphadenectomy for gynecologic turn or^(^^-'^). In particular, only sporadic papers have been published about its feasibility, morbidity and The lack of data is mainly due to the fact that the indications are debatable and the procedure is considered difficult and hazardous. Since 1986 a prospective study of the feasibility of radical para-aortic and pelvic lymphadenectomy, by transperitoneal approach, in addition to primary surgery was began at our Institution. The purpose of this paper is to describe the technique used by us and to review the impact of this surgical procedure on the perioperative complications in 208 consecutive patients with gynecologic malignancies.

Materiak and methods Patients All patients with untreated cervical, endometrial and ovarian cancer were candidates for entry into this study. The pretreatment evaluation consisted of a complete history, physical and gynecologic examinations, laboratory work-up, EKG, chest X-ray, intravenous pyelogram, hepatic and pelvic ultrasonography . More specific investigations were performed for different types of cancer according to FIGO staging. Eligibility requirements included a maximum age of 70 years, an ECOG performance status of 0-1, adequate renal, cardiac and hepatic function, and informed consent. Specific criteria used in selecting patients with cervical cancer were as follows. 1 IB-IIA Stage of disease (FIGO classification) if one or more of the following risk factors were present: adenocarcinoma, poorly differentiated tumor; positive common iliac lymph node detected by frozen sections. 2 Locally advanced disease (Ib-I1 tumor volume

greater than 4 cm, 111) which became operable after neoadjuvant Ovarian cancer patients were selected as follows: early stages of disease; patients with no residual tumor after cytoreductive surgery; during second-look procedures if macroscopic or microscopic disease was absent. The inclusion criteria for endometrial cancer were: Stage I patients with poorly differentiated tumors and/ or greater than one third miometrial invasion intraoperatively detected; Stage I1 patients. Exclusion criteria included severe obesity, leg varicosities with edema or venous stasis changes, distant metastases detected intraoperatively, and presence of second tumor except for basal cell carcinoma of the skin.

Radical para-aortic lymphadenectomy The abdominal cavity is opened by a cutaneous xiphopubic incision. The first step in exposing the retroperitoneal space includes 3 peritoneal incisions. The first is made along the right paracolic gutter up to the hepatocolic ligament. The second is a diagonal cut on the peritoneum which goes along the route of the mesentery from the caecum to Treitz’s ligament. The third incision is made in the left side along the homolateral paracolic gutter almost u p to the splenocolic ligament. This last incision is useful in order to increase the exposition of the left surgical field. Accurate hemostasis of the small vessels of the retroperitoneal areolar tissue is achieved with coagulating forceps. By mobilizing and elevating the caecum and the ascending colon it is possible to identify the anterior leaf of the renal fascia which represents the surgical plane to follow. This manouver is safe and allows a wide exposition of the retroperitoneal field. The entirely mobilized ascending colon and the small intestine are thus taken out of the peritoneal cavity, placed on the chest and wrapped in warm wet dressings or in appropriate plastic bags. A specifically provided retractor (height 18 cm, width 13 cm) held in the left hand of the third assistant is used to keep the bowel off the surgical field. At this point, the main structures of the retroperitoneum including the right kidney, the renal vessels, the ureter, the homolateral ovarian pedicle, the vena cava, aorta and the inferior mesenteric artery and vein are identified. The right lateral margin of the para-aortic lymphadenectomy is delineated by the medial margin of the ureter which has to be identified and carefully isolated from the iliac vessels and ovarian pedicle up to its most proximal portion. The upper margin of the dissection

Lymphadenectomy for gynecologic malignancies for ovarian and endometrial cancer is 1-2 cm above the renal vessels, while for cervical cancer the surgical limit is represented by the origin of the inferior mesenteric artery. The posterior surgical plane is delineated laterally by the psoas muscles, medially by the prevertebral fascia. While the identification of the right lateral margin is reasonably easy, on the left side this manouver is more difficult. In fact, the left ureter runs in contact with the posterior aspect of the descending colon and it is very close to the left ovarian pedicle especially the portion below the inferior mesenteric artery. The isolation of the left ureter is important in order to avoid damaging it and in order to mobilize and lateralize the descending colon so far that the exposition of the left surgical field is increased. After resecting the right ovarian pedicle, lymphadenectomy begins with the isolation and dissection of the right renal pedicle and vena cava, and then continues with the removal of the caval, paracaval, retrocaval and intercavo-aortic, lymph nodes. In order to remove these groups of nodes more easily it is useful to identify the lumbar vessels and to raise aorta, vena cava and from time to time and, at the same time, to separate the great vessels with small swabs. Before removing the retro-aortic and para-aortic nodes on the left margin of the aorta, the left renal vein must be identified, the most proximal portion of the homolateral ureter is laterally retracted with a valve and the left ovarian pedicle is resected. When the dissection proceeds downwards, the inferior mesenteric artery should be protected; however, its resection or ligation rarely causes ischemia of the colon owing to the presence of a collateral circulation. Particular attention is required in patients who have macroscopic metastatic nodes closely adherent to the great vessels. In these cases, the normal anatomic cleavage between the nodal tissue and the vessels may be obliterated. In one patient, clamping, resection and suturing of the vena cava was necessary to remove the whole tumor.

Radical pelvic lymphadenectomy The dissection is carried down towards the bifurcation of the aorta and continues at the level of the common iliac vessels. Removal of the common iliac nodes is not complicated because of the optimal surgical field. Presacral vessels are then identified and the presacral space is prepared by mobilizing anteriorly the sigmarectum. At this stage, it is often possible to identify the presacral artery which sometimes originates from the posterior aspect of the aorta and its origin is therefore difficult to see. The common iliac arteries and veins are now completely freed.

I 35

As far as the typical pelvic lymphadenectomy is concerned, our technique consists of the 'en block' removal of all lymph node tissue surrounding the external and internal iliac vessels and that located above the obturator nerve. After the preparation of the pararectal and para-vescical recesses, the dissection begins at the origin of the external iliac vessels and continues caudally around them along the medial border of the psoas muscle, by preserving the aponeneurotic fascia covering it. At this point, in order to decrease the risk of neurologic postoperative complications, it is important to spare the branches of the genito-femoral nerve. The external iliac vessels are completely separated from each other and are drawn medially by a small tape. The lower limit of external iliac lymphadenectomy is represented by the deep inferior epigastric vessels. The lateral boundaries of the dissection are superficially delineated by the fascia covering the psoas muscle and deeply by the fascia covering the internal obturator and levator ani muscles. The lateralization of the external iliac vessels and of the obturator nerve allows the identification of the medial margin of the lymphadenectomy, represented by an imaginary plane which is parallel to the umbilical artery and to the lateral aspect of the bladder and the rectum. The clearing of the obturator fossa begins with the mobilization of the superficial obturator nodes which are completely dissected after the identification of the upper face of the obturator nerve. These nodes are removed en block with the lymphatic fatty tissue which has been previously separated from the iliac internal vessels until the origin of the internal pudendal vessels. For patients with locally advanced cervical cancer, the pelvic lymphadenectomy continues with the dissection of the deep obturator nodes and gluteal nodes. The isolated and ligated obturator artery and vein are removed together with the lymphatic tissue of the inferior part of the obturator fossa. By moving the external and common iliac vessels medially and the psoas muscle laterally with a retractor, it is possible to remove node groups of this area and see the superior and inferior gluteal veins and the trunk of the sciatic nerve. This step ends pelvic lymphadenectomy. In order to reduce the lymphatic loss, each lymphatic channel dissected was always ligated with hemoclips. Three retroperitoneal suction drains were placed during the operation: one at the insertion of the inferior mesenteric artery, up to the caudal tract of the left external iliac artery; another at the level of the paracaval area, from the origin of the right ovarian vein up to the caudal tract of the right external iliac artery; and the third (T-tube) at the level of the obturator fossa coming out of the vagina.

136 P.B. Panici et al.

Perioperative care

Results

Adequate perioperative care of patients undergoing systematic para-aortic and pelvic lymphadenectomy was imperative. Appropriate bowel preparation was routine. The standard 72-h bowel prep consisted of semiliquid diet, (liquid 24 h before surgery), saline enema and metronidazole (250 mg per os, 3 times per day). In all patients, a nasogastric tube was inserted and maintained until bowel function was restored. Central venous lines were helpful in monitoring hemodynamic status. During the first 2 years of our experience, all patients received mini-dose heparin prophylaxis (5000 units every 12 h), beginning 1 h before the operation and continuing postoperatively until complete ambulation was restored. Since 1988 we have used a different antithrombotic agent, called defibrotide which acts with a distinct mechanism. The purpose of this change was to reduce the risk of bleeding and lymphatic loss during the postoperative period. In fact, heparin prophylaxis has been reported by most authors as a promoting factor of lymphocyst formation and of postoperative h e m a t ~ m a ( ' ~All , ~ ~patients ). received shortterm antibiotic prophylaxis. Patients who became pyrexial were revaluated and antibiotics were regulated according to the diagnosis. By using weekly ultrasound examinations, the development of postoperative fluid collections was serially investigated since April 1987.

Of 284 patients evaluated for entry into the study between January 1986 and June 1990, systematic paraaortic and pelvic lymphadenectomy was performed in 208 cases (73.2%). Severe obesity, medical disorders (cardiac and respiratory diseases) and leg varicosities represented the exclusion criteria in 8 (2.8%), 32 (11.3%) and 10 (3.5%) patients, respectively. The presence of distant metastases or more extensive disease than expected were detected intraoperatively in 16 patients (5.6%) with cervical cancer and in 9 cases (3.2%) with endometrial cancer, therefore systematic lymphadenectomy was not carried out. Moreover, radical node dissection was not feasible in one case with severe retroperitoneal fibrosis. The characteristics of the 208 operated patients are shown in Table 1. For the 108 cervical cancer patients, lymphadenectomy was carried out in addition to type I11 and IV radical hysterectomy (73 and 35 cases, respectively). Type I1 radical hysterectomy was performed in the 57 cases with endometrial cancer. Systematic lymphadenectomy was an integral part of the intensive surgical staging and of second look procedures for 33 and 10 ovarian cancer patients, respectively. As reported in Table 2, the median number of paraaortic and pelvic nodes removed was different in the three types of tumors. The explanation of these results is found in a more radical pelvic dissection performed

Table 1. Characteristics of the patients undergoing systematic para-aortic and pelvic lymphadenectomy Primary tumor

Median age

(Range)

Stage I I1 111 Total No. (Yo) No. (YO) No. (YO) cases

49 60 46

(23-70) (45-69) (22-66)

36(33) 43(75) 27(63)

~

Cervix Endom. Ovary

43(40) 14(25) -

29(27) 16(27)

108 57 43

Table 2. Median numbers of lymph nodes removed according to type of tumor Data

Cervix No. (range)

Ovary No. (range)

Endom. No. (range)

Median No. lymph nodes removed Para-aortic Pelvic

58 (38-128) 20 (14-41) 38 (20-88)

49 (37-85) 27 (19-48) 22 (17-40)

54 (38-105) 24 (15-47) 30 (22-56)

Lymphadenectomy for gynecologic malignancies

137

at the level of the gluteal nodes as well as at the level of the obturator nodes in the case of cervical cancer. Moreover for this tumor the upper surgical limit of the para-aortic dissection was lower (origin of the inferior mesentery artery). Table 3 shows the operating data divided into two groups according to the consecutive number of the cases. As the surgical team acquired experience, the operating time and the blood loss gradually diminished. In both series of patients, regardless of the type of tumor, longer operating times and greater blood losses were usually observed in patients with vessel anomalies or vessel varicosities. Other complicating factors were represented by macroscopically metastatic nodes closely adherent to the great vessels or pelvic inflammatory diseases.

tomy and was determined by a retractor incorrectly placed. Severe hemorrage caused by arterial lesions was rare, however, in our series, it occurs in only 2 patients. Blood loss was quantified a s 1100 and 1300 ml, respectively. In both cases, only after raising u p and rotating the aorta it was possible to identify an injured middle sacral artery which originated from the posterior aspect of aorta. An intraoperative nerve lesion occurred in one patient, the obturatory nerve was dissected and repaired by suturing with monofilament 5-0. The outcome was good. There was one urinary complication: a left ureter not recognized, was cut during the para-aortic node dissection. The lesion was immediately identified a,nd repaired without sequelae.

Table 3. Operating data according to the consecutive number of the cases

Post-operative complications

Operating time (min) Estimated blood loss (ml) Transfusion (U)

Group A* Median (range)

Group B Median (range)

230

(120-270)

150 (100-240)

390 1

(200-3300) (0-7)

250 (100-2800) 0 (0-5)

*Group A, first 95 cases; Group B, second 113 cases.

Intra-operative complications

As far as vessel injuries are concerned, small lesions of the vena cava occurred often. However, the hemorrhages were minimal and could easily be controlled by suturing with monofilament 5-0 prolene. These lesions were usually caused by lacerations of small vessels draining from fatty lymph node tissue adherent to the vena cava. When the isolation and ligation of these small vessels with hemoclips was introduced as a routine, the lymph node dissection became progressively bloodless and more rapid. On the other hand, severe hemorrages (estimated blood loss exceeding 1000 ml) occurred in 4 patients. In 2 of these cases (blood loss was 1200 ml and 1400 ml, respectively) a lumbar vein was lacerated at the level of the most proximal part of the left ovarian pedicle. In fact, in many cases, a lumbar vein goes into ovarian veins just a few millimeters before the confluence with the renal vein. In the third patient, severe blood loss (1800 ml) was caused by a lesion of a lumbar vein tributary to the right common iliac vein. In the last patient, (blood loss was 2500 ml) an extensive laceration of the right internal iliac vein occurred at the end of lymphadenec-

The duration of postoperative hospitalization averaged 18 days (range 15-50 day). The median duration of drainage was 15 days (range 8-22 day), with a median total drained fluid volume of 1210 ml (range 200-4700 ml). Table 4 shows that the most frequent postoperative complication was the formation of lymphoceles. In fact, 29 out of 142 patients (20.4%) had lymphocyst ecographically diagnozed between the 11-26th postoperative day (median 16th day). Eleven of these fluid collections were asymptomatic and smaller than 50 ml in volume; they were monitored sonographically and resorbed spontaneously in an average time of 47 days. Twelve symptomatic and 6 progressive asymptomatic collections were drained under ultrasonic guidance. Among the symptomatic cases, 4 developed a deep venous thrombosis, 5 complained of pelvic pain, 2 had mild fever and in 1 case a unilateral hydronephrosis was detected at ultrasound examination. Catheter drainage allowed complete clinical and sonographic remission in all cases; it was well tolerated and caused no side effect. Only two asymptomatic recurrences occurred at the 3- and 4-month Table 4. Postoperative complications Complication Lymphocyst Deep venous thrombosis Pulmonary microembolism Mild paresthesia Retroperitoneal abscess Ureteral fistula

20/142* 18 9 11 3 1

(20.4) (8.8)

(4.4) (5.4) (1.4) (0.5)

*Lymphocyst rate is calculated on the last 142 cases, since only these cases had serial ultrasound examination.

138 P.B. Punici et al. follow-up respectively, and were present unchanged 6 months later. Despite the anticoagulant prophylaxis and early mobilization, 18 patients (8.8%) developed deep venous thrombosis, and in 4 of these cases, as mentioned above, a pelvic lymphocyst causing mechanical obstruction of the major vessels was diagnosed. Nine of these patients experienced pulmonary microembolism, in 2 cases the complication was considered to be serious and required intensive care. However the outcome was good in both cases. The resection or coagulation of branches of the genito-femoral and obturator nerves determined mild paresthesies localized at the supero-anterior and internal sides of the thigh in 11 patients (5.4%). This neurologic complication was treated with physical therapy and an improvement was evident in all cases. As the surgical team paid more attention in using the coagulating forceps, the rate of this side effect gradually diminished. Three patients developed a retroperitoneal abscess on the loth, 14th and 15th postoperative day, respectively. These abscesses required operative treatment and appropriate medical therapy. Although it is hard to say whether complications are a result of the lymphadenectomy or of the surgery for the primary tumor, the ureteral lesions of the upper portion were considered to be caused by the lymph node dissection. In this regard, fistula of the most proximal part of the right ureter appeared in 1 patient during the third postoperative week. However, the complication was dealt with successfully. In order to define prognostic factors for the development of postoperative complications, clinical and pathologic variables were correlated to the incidence of side effects. No statistically significant differences were found between the clinical parameters considered (age, weight and previous chemotherapy) and the postoperative complications. The type of cancer and the lymph node status had no statistical influence on the development of complications either, even though a higher percentage of lymphoceles was observed in patients with cervical cancer and with more than 6 positive nodes. A trend was also found between lymphocyst rate and number of nodes removed. Even though it was not possible to reach any statistically significant result, lymphocele development was more frequently observed in patients who had more than 60 lymph nodes totally removed andlor more than 10 obturator nodes dissected (data not shown).

Discussion The present report shows that systematic pelvic and

para-aortic lymphadenectomy could be carried out with an acceptable morbidity and no mortality. This surgical procedure seems to be feasible and its related intra- and postoperative complications are manageable. The high median number of nodes removed in our series, never reported in the literature so far, represents the objective proof of the thoroughness and radicality of this surgical procedure. Its use could be introduced by a well-trained surgical team even at our Institution which did not have any experience before 1986. Once surgery was begun, it could be terminated in all cases, except one because of intense retroperitoneal fibrosis. No operative complication was so difficult to deal with that the lymphadenectomy had to be stopped and the node dissection was even possible in cases with grossly metastatic nodes with extracapsular spread to surrounding structures. The feasibility of our technique mainly depends on the type of surgical approach used. In fact, the transperitoneal access, including careful isolation of ureters, veins, arteries and nerves, offers a wide surgical field. This results in the following major advantages, the surgical team works without tension, the potential intraoperative complications are reduced and a high number of lymph nodes can be removed. Vascular, ureteral and nerve lesions could be repaired successfully in each circumstance, given the wide space available. Moreover, as the surgical team acquired experience, the operating time and the blood loss gradually diminished. Although vascular injuries did occur, there were no permanent sequelae. Obviously, the surgical team must have previous experience in managing vessel, intestinal and genitourinary complications. The incidence rate of major vascular complications reported in this study (6/208, 2.9%) is comparable to that which occurred in Donohue and Rowland's series(*l). These authors noted 9 vessel lacerations among 235 men (3.8%) undergoing retroperitoneal lymphadenectomy for testicular cancer. There are several ways to determine the status of the para-aortic nodes. Some authors choose the rightsided approach, on the basis of anatomical considerations("). First, it is believed that a larger number of lymph nodes is located on the right side of the great vessels. Secondly, on the left side it is more difficult to isolate the ureter and a good exposure can be obtained by dividing the inferior mesenteric vessels. In our experience, we found that an appreciable number of node^(^-^) is attached to the posterior aspect of the aorta as well as to its left lateral face. Moreover, as reported also by Belinson et aZ.(lO) and confirmed by our experience, it is possible to find positive nodes on the left side with no involvement of those on the right

Lymphadenectomy for gynecologic malignancies

I 39

side. This evidence suggests that in order to assess the Systematic pelvic and para-aortic lymphadenecreal extension of the disease, an adequate number of tomy represents the only reliable method for the nodes should be collected from each side of the vena assessment of lymphatic spread in gynecologic cava and aorta. Using our approach, the inferior oncology, given the low accuracy of the imaging techmesenteric vessels do not need to be divided, since a niques. The question of whether lymph node dissecwide exposure of the left surgical field is achieved by tion can be considered also a therapeutic tool remains mobilizing and lateralizing the descending colon. unanswered. However, as the policy in the operative It has to be stressed that some measures and premanagement of cancer is to try to remove all the primcautions adopted intra- and perioperatively can reary tumor and its loco-regionalspread, retroperitoneal duce morbidity-related surgery without sacrificingthe lymphadenectomy could have a therapeutic role. In completeness of the lymphadenectomy. Hemoclips to fact, this surgical procedure will not only lessen the ligate small vessels and lymphatic channels are useful tumor burden by removing the cancer-laden nodes, it to shorten the operating time. might hopefully also block the further upward spread A relevant role must also be given to the perioperative. of the diesease. Moreover, it seems reasonable to care. Improvements in anesthesia and in monitoring hypothesize that a good chance of cure can be offered of the hemodynamic status made the surgery possible to patients with only 1-2 micrometastaticnodes which in most of the patients; in fact only 32 cases (11.3%) are removed by surgery. In the literature, reports on were excluded because of severe medical disorders. the evaluation of lymphadenectomy in the treatment Despite the antithrombotic prophylaxis, 9% of the of gynecologic malignancies are few and there is a lack patients developed deep venous thrombosis. These of randomized controls. The availability of a feasible failures suggest that new anticoagulant agents could surgical technique could extend its use and the value be useful for these patients with high risk of thromof lymph node dissection on prognosis of patients boembolic disease. More satisfying results were obwith gynecologic malignancies could therefore be tained by using antibiotic prophylaxis and suction better defined. drainages. With these preventative measures only 3 retroperitoneal infections strictly attributable to the lymph node dissection were documented. Finally, References adequate bowel preparation and use of the nasogastric 1 Hann LE, Crivello MS. Imaging techniques in the staging of tube appear to be important precautions in order to gynecologic malignancy. Clin Obstet Gynecoll986; 29: 715-27. facilitate the surgical manouvers and to restore the 2 Piver MS, Barlow JJ. Para-aortic lymphadenectomy, aortic bowel functions quickly. None of the patients comnode biopsy and aortic lymphangiography in staging paplained of paralytic ileus lasting more than 3 days. tients with advanced cervical cancer. Cancer 1973; 3 2 367-70. 3 King LA, Talledo OE, Gallup DG, El Gamma1 TAM. Among the postoperative complications, the deComputed tomography in evaluation of gynecologic maligvelopment of lymphocysts was the most frequent and nancies: A retrospective analysis. Am J Obstet Gynecol 1986; troublesome problem. The trend observed between 155: 960-4. the lymphocyst rate and the total number of nodes re4 Averette HE, Donato DM, Lovecchio JL, Sevin BU. Surgical moved as well as the number of those dissected at the staging of gynecologic malignancies. Cancer 1987; 60: obturator fossa level suggest that the thoroughness of 2010-20. 5 Chevassu M. Deux cas depitheliome du testicule traites par lymphadenectomy represents the main risk factor for la castration et l’ablation des ganglions lomboaortique. Bull lymphocele development. Among the different reMem Sac Chir Paris 1910; 3 6 236-9. gions exposed, the obturator fossa seems to be the 6 Lewis EL, Johnston RE, Rowe RB et al. Retroperitoneal most critical area for postoperative fluid collections. lymph node resection: the intercostoinguinal approach. These observations could explain the higher incidence Uroll952; 67: 338-42. 7 Nagamatsu G. A new extraperitoneal approach for bilateral of lymphocyst which occurred in cervical cancer paretroperitoneal lymph node dissection in testis tumor. J Urol tients. In fact, for this tumor the clearing of obturator 1963; 90: 588-92. fossa is more radical and the total number of nodes 8 Donohue JP. Retroperitoneal lymphadenectomy. Urol Clin removed is greater. Similarly to other a~thors(~’,’~), North Amer 1977; 4: 517-23. patients with positive nodes tended to have a higher 9 Donohue JP, Perez JM, Einhorn LH. Improved management of nonseminomatous testis. J Uroll979; 121: 425-8. lymphocyst rate than did those with negative nodes. 10 Belinson JL, Goldberg MI, Averette HE. Para-aortic lymphLymphocyst formation remains an open problem and adenectomy in gynecologic cancer. Gynecol Oncol 1979; 7: new strategies should be investigated to prevent this 188-98. complication, however it has to be noted that our 11 Lubicz S. Approach to the abdominal retroperitoneum in management scheme proved to be safe and highly patients with gynecologic malignancies. Gynecol Oncoll985; 2 2 32-9. effective.

140 P.B. Panici et al. 12 Winter R, Petru E, Haas J. Pelvic and para-aortic lymphadenectomy in cervical cancer. In: Burghardt E, Monaghan JM, eds. Operative treatment of cervical cancer. London: Bailliere, Tindall 1988; 4: 857-866. 13 Di Re F, Fontanelli R, Raspagliesi F, Di Re E. Pelvic and paraaortic lymphadenectomy in cancer of the ovary. In: Burghardt E, Monaghan JM, eds. Operative treatment of ovarian cancer. London: Bailliere, Tindall 1989; 5: 131-42. 14 Powell JL, Burrel MO, Franklin 111EW. Radical hysterectomy and pelvic lymphadenectomy. Gynecol Oncoll981; 12: 23-32. 15 Weiser EB, Bundy BN, Hoskins WJ, Heller PB, Whittington RR, Di Saia PJ, Curry SL, Schlaerth J, Thigpen JT. Extraperitoneal versus transperitoneal selective para-aortic lymphadenectomy in the pretreatment surgical staging of advanced cervical carcinoma (a gynecologic oncology group study). Gynecol Oncoll989; 33: 283-9. 16 Moore DH, Fowler WC, Walton LA, Droegemueller W. Morbidity of lymph node sampling in cancers of the uterine corpus and cervix. Obstet Gynecol 1989; 74: 180-4. 17 Benedetti Panici P, Scambia G, Greggi S, Di Roberto P, Baiocchi G, Mancuso S. Neoadjuvant chemotherapy and radical surgery in locally advanced cervical carcinoma: A pilot

study. Obstet Gynecol1988; 71: 344-8. 18 Benedetti Panici P, Scambia G, Baiocchi et al. Neo-adjuvant chemotherapy and radical surgery in locally advanced cervical cancer: prognostic factors for response and survival. Cancer (in press). 19 Catalona WJ, Kadmon D, Crane DB. Effect of mini-dose heparin on lymphocele formation following extraperitoneal pelvic lymphadenectomy. J Uroll980; 123: 890-2. 20 Kragt H, Bouma J, Aalders JG. Anticoagulants and the formation of lymphocysts after pelvic lymphadenectomy in gynecologic and oncologic operations. Surg Gynecol Obstet 1986; 162: 361-4. 21 Donohue JP, Rowland RG. Complications of retroperitoneal lymph node dissection. J Uroll981; 125: 338-40. 22 Rutledge F, Dodd GD, Kasilag FB. Lymphocysts: a complication of radical pelvic surgery. Am Obstet Gynecol 1959; 7 7 1165-75. 23 Petru E, Tamussino K, Lahousen M, Winter R, Pickel H, Haas J. Pelvic and paraaortic lymphocysts after radical surgery because of cervical and ovarian cancer. Am 1 Obstet Gynecol 1989; 161: 937-41. Accepted for publication 28 February 1991