Laparoscopic pelvic lymphadenectomy in an anatomical model: results of an experimental comparative trial

European Journal of Obstelrics & Gynecology and Reproductive Biology 72 (19971 51 55

ELSEVIER

oBsTna GYNECOLgGY

Laparoscopic pelvic lymphadenectomy in an anatomical model: results of an experimental comparative trial F. Ldcurw ~*, F. Robin ~, K. Nej?, C. Darles '', P. De Bievres ~', F. Vild6 ~', R. ~Scrt'ice dc G.i,n~¥oh)gie-Obst(,lrique,

T a u r e l l e ~'

H@ila/ Bottcicau/. 78 rue d~' h~ ('ol~lentiom 75015 Paris. Fram'c

bSercic# d'mll(llOItli(' Pal/lo/oL, iqtte. H@ila/ Boucicaut. 78 ru~' d{' /a (olll'ellliOtl, 75 015 Paris. t:)'al/{~' Received 5 August 1996: revised II October 1996: accepted 24 October 1996

Abstract Objectives: The aim of this paper was to compare the accuracy of laparoscopic versus open pelvic lymphadenectomy in an experimental trial. Study design: We performed unilateral laparoscopic pelvic lymphadenectomies (LPL) in 33 non-embalmed cadavers between the external lilac vein, the obliterated umbilical artery and the obturator nerve. Then a Iaparotomy was performed to inspect the LPL limits, look for laparoscopic complications and finally realize a controlateral lymphadenectomy. The LPL side was randomly decided. A pathologist counted the number of lymph nodes collected with both techniques. We compared the number of retrieved lymph nodes, the completeness of the dissection and the complication rate with those two procedures. Student's t-test, ze-test and non-parametric tests were used when appropriate. Results: No dissection had to be aborted. One hundred and twelve nodes ~ere removed laparoscopically (mean. 3.73: S.E., 2.9) and 84 at laparotomy {mean. 2.77: S.E., 2.06). There was no significant difference in the number of nodes retrieved with both procedures. Effectiveness of laparoscopy was not significantly different in the first ten procedures, in the second ten or in the last ten LPL. Residual tissue was observed after LPL in 13.3% of the procedures whereas all open lymphadenectomies were complete. LPL sensitivity reached at least 86% in this paper. Failures were more frequent at the beginning of the study {50% among the first ten dissections), in obese subjects or in subjects with prior history of laparotomy (but the difference was not significant). Two venous injuries occurred during LPL (6.7%). Complication rates for the two techniques were not significantly different. However, the LPL complication rate was higher at the beginning of the study and increased significantly in subjects with prior history of laparotomy (P < 0.05). Conchlsions: This randomized study shows that LPL and laparotomy have similar effectiveness. Incomplete dissections and complications are more frequent in obese subjects or in case of prior history of laparotomy. Fifteen procedures seems necessary to learn the technique and provide constant and safe results in routine practice. {? 1997 Elsevier Science Ireland Ltd. K e y w o r d s : Accuracy: Experimental trial: Laparoscopy; Pelvic lymphadenectomy

1. Introduction Laparoscopic pelvic l y m p h a d e n e c t o m y (LPL) has been proposed in the staging as well as in the treatmenl of pelvic cancer such as cervical or endometrial carc i n o m a s [1,2]. Descriptive clinical studies have been published with e n c o u r a g i n g results, but c o m p a r a t i v e trials are rare a n d the effectiveness of this procedure is *Corresponding author. Tel: --33 1 53788199: fax: +33 I 53788209. 0301-2115 97,,'SI7.00

still controversial [3]. Tile aim of this paper was to c o m p a r e the accuracy of laparoscopic versus open pelvic l y m p h a d e n e c t o m y , in a r a n d o m i z e d experimental trial.

2. Material and methods We performed t r a n s p e r i t o n e a l l y m p h a d e n e c t o m y by means of laparoscopy and l a p a r o t o m y in n o n - e m balmed cadavers.

1997 Elsevier Science Ireland Ltd. All rights reserxed

PII SO ?O1-21 1 5 1 9 6 ) 0 2 6 5 2 - 8

52

F. L&'uru et al./ European Journal o! Obsletrics & Gynecolog:v and Repro&~ctive BiologT 72 (19971 51 55

Subjects were placed in the Trendelenburg position. The technique consisted of a three ports laparoscopic access with disposable tools (Ethicon Ethnor Endosurgery). One was placed trans-umbilically (12 mm diameter) and two in a suprapubic position, one medial (5 or 12 mm diameter) and one lateral to the epigastric vessels (5 mm diameter) (this allowed sufficient access for unilateral lymphadenectomy). A 15 mmHg pneumoperitoneum was maintained using a high flow insufflator (Olympus). A panoramic vision was obtained by means of a 10 mm laparoscope (Olympus) attached to a video camera (Olympus) (Video monitor, Sony). First we inspected the peritoneal cavity for the presence of adhesions or other pathological conditions. Lymphadenectomy was then performed using grasping forceps and scissors between the external iliac artery laterally and the obliterated umbilical artery medially. Dissection was conducted from the femoral canal to the bifurcation of the common iliac artery. The obturator nerve at the bottom of the obturator fossa constituted the inferior limit of the dissection. The whole lymphatic tissue removed between these boundaries was left in the pouch of Douglas. The side of the laparoscopic lymphadenectomy was randomly decided. A large transversal laparotomy was then perfbrmed. Margins of the laparoscopic lymphadenectomy were carefully inspected to check the completeness of the endoscopic dissection. The abdomen and the surgical site were closely examined to look for any complication secondary to the laparoscopy. The lymphatic tissue was placed in a 10% formalin solution. An open lymphadenectomy was conducted on the opposite side, using the same margins. The tissue was also placed into a 10% formalin solution. Finally, a careful inspection controlled the quality of the limits and noted any complications for the open procedure. Lymphatic tissues were given to a pathologist in anonymous pots. After fixation, serial sections were stained with hematoxylin and eosin (H and E) and studied on light microscopy. The pathologist counted the number of lymph nodes in each lymphadenectomy specimen without any knowledge of the procedure used for each sample. Statistical analysis used z-Ltest Student's t-test or nonparametric tests. A P value of 0.05 or less was considered significant.

3. Results

Thirty-three non-embalmed cadavers were used. Three procedures were excluded from further analysis. In one case laparoscopy was suspended at the beginning of the procedure because of an abundant ascitis in an obese woman. In two cases the lymphatic tissue was poorly preserved and could not lead to a correct evaluation by the pathologist. No laparoscopic

Table I Comparison of the efficiency of laparoscopy and laparotomy according to experience (number of retrieved lymph nodes) Experience ~

Laparoscopy

Laparotomy

1 10 11 20 21 30

4 . 8 + 3 . 6 ( 0 II) 3.1+1.5(1 5) 3.3_+3.09(0 11)

1.9±2.18(0 6) 1 . 9 ! 1 . 6 ( 0 5) 4.5+_1.08(2 6)

P

> 0.46

< 0.006

" 'H' test. Kruskal Wallis.

lymphadenectomy had to be aborted during the dissection. Subjects were principally post-menopausal women. Mean age was 77.6 years (S.D., 9.8; range, 49 96). Twenty-three per cent were obese (weight over 90 kg), and 23.3% had a history of laparotomy. Fifteen laparoscopic lymphadenectomies were performed on the right side and 15 on the left side. One hundred and twelve lymph nodes were removed laparoscopically (mean number, 3.73; S.E., 2.9). The average number was 3.3 + 2 . 9 (range, 0 111 on the right and 4.1 -+2.9 (range, 0-111 on the left side (NS). Eightyfour lymph nodes were obtained with the open technique (mean number, 2.77: S.E., 2.06). The average number was 1.73 (range, 0-6) on the right and 3.8 (range, 0 6) on the left side. There was no significant difference in the number of nodes removed laparoscopically or with open surgery (P = 0.1). We compared the mean number of lymph nodes retrieved by the two techniques according to our experience (Table 1). Effectiveness of laparoscopy was not significantly different in the initial ten procedures than in the second ten or the last ten lymphadenectomies (Wilcoxon t-test and linear regression). On the other hand, the number of lymph nodes obtained at laparotomy increased significantly from the beginning to the end of our study (P < 0.006). Laparoscopy retrieved more lymph nodes than laparotomy in the first 20 subjects of the trial. Efficiency of laparoscopy and open surgery decreased in subjects with prior history of laparotomy or pelvic surgery in a non-significant manner (4.04 + 3.05 and 2.8 -+ 1.96 vs. 2.7 + 2.2 and 2.5, respectively) (Table 2). The number of harvested lymph nodes increased with the corpulence of the cadavers for the two techniques (NS, Table 3). Table 2 Efficiency of laparoscopy and laparotomy according to history of laparotomy (number of retrieved lymph nodes)

No history of laparotomy History of laparotomy 'H" test, Kruskal Wallis.

Laparoscopy

Laparotomy

4.04 _+ 3.05 2.7 +_2.2

2.8 + 1.96 2.5 i 2.2

F. L&'uru et al./ European Journal of Obstetrics & Qvnecolow and Reproductive Biology 72 (1997) 51 55

Table 3 Efficiency of laparoscopy and laparotomy according to the corpulence of the cadavers (number of retrieved lymph nodes) Size (kg)

Laparoscopy

Laparotomy

Thin t < 6 0 ) Normal (60 90) Obese ( > 9 0 )

2.6 (:k 1.8) 4 ( :k 3.4) 4.2( + 2.7)

2( + 1.7) 2.9( 2 2) 3.2( ~ 2.4)

53

100 9O 8O

60 50 40 30 20 10

P = 0.53, using the 'H' test of Kruskal Wallis.

Residual tissue was found at laparotomy in 13.3% of the procedures. In three cases, (10%) fatty tissue remained in the area of the common iliac artery bifurcation, and in one case (3.3%) at the other extremity, near the femoral canal. Half of these failures were observed among the first ten cases and 75% among the first 15 cases. Incomplete laparoscopic lymphadenectomies were more frequent in obese compared to 'normal' subjects (40 vs. 7%) and also in subjects with a prior history of pelvic surgery, but differences were not significant (Fisher test). Open lymphadenectomies were judged anatomically complete in 100% of cases. Two complications occurred during laparoscopic procedures (6.7%) and none during open lymphadenectomies (NS). One consisted of an external iliac vein injury, and the other of the section of a venous anastomosis between the obturator and external iliac vein. These two events occurred during the first ten cases. The complication rate was significantly higher in subjects with a history of pelvic surgery (40 vs. 0%, P < 0.05 Fisher test). Corpulence had no significant influence on morbidity. The duration of each laparoscopic lymphadenectomy is detailed in Fig. 1. The duration of each intervention was measured from the peritoneum incision to the end of the dissection. The mean operating time was 29.8 min (S.E., 9.9) and it significantly decreased with experience (P < 0.0001).

50 45

20

:i 0

Fig. 1. Duration of laparoscopic lymphadenectomy according to the experience (length in minutes).

0

-,

Fig. 2. Percentage of nodes removed at laparoscopy.

4. Discussion

Lymph node involvement is one of the major prognostic factors for cervical or endometrial carcinomas [4 6]. Radiological methods are not specific or particularly sensitive enough in the diagnosis of pelvic lymph node metastases and surgical pelvic lymphadenectomy actually remains the most accurate procedure to assess the spread of the disease to pelvic lymph nodes [7,8]. Surgical staging by means of laparotomy is a major procedure with a high cost/benefit ratio [9]. Laparoscopic staging has been proposed because it combines the accuracy of surgery with the minimal invasiveness of endoscopy. Clinical studies have reported encouraging results for LPL but only few data concerning randomized trials have been published. This trial was conducted on non-embalmed cadavers. Although this model differs from living patients, it allows the comparison of the two techniques. Moreover, this experimental model had some advantages over using animals. Anatomy was the same as in our clinical practice so that anatomical landmark recognition was easy. Finally, this model asked no ethical considerations, even in a randomized trial. We had previously learned the technique of laparoscopic lymphadenectomy and we used a transperitoneal technique, as in our routine clinical practice. Therefore, we can assume that the two procedures were compared in an equal manner. Moreover, our paper is the largest randomized trial published in recent literature. This leads us to think that our results are strong. Our work showed no statistical difference in the total number of lymph nodes removed by laparoscopic or open procedures. Moreover, laparoscopy harvested more nodes than laparotomy in most of the subjects (node yield, 57.4%; Fig. 2). Laparoscopic lymphadenectomy efficiency was constant all along this work and did not differ significantly, no matter what side the procedure was performed upon. Clinical studies have reported learning curves, showing an improving efficiency with time and number of procedures performed [1,10,11]. It is obvious that some authors, even in large series, reported their initial expe-

54

F. L&'uru ctal.

European Journal o/Ob,sletrics & (~vnecolo~,:v and Reproductire Biolog:v 72 (1997) 51 55

Table 4 Efficiency of laparoscopic lymphadenectomyin the literature Author

No. of cases

20 Fowler [15] 12 Guazzoni [10] 30 Parra [13] 24 Querleu [1] 110 Rukstalis [11] 103

Age 62 23 56 54 28

Boike [12]

Rightside 14.5

Left side 14.5

60 23.5 73 8.7 77 5.2 81 5.5 8.2_+4.9

23.5 8.8 5.5 5.1 8.2+4.9

rience. No learning curve was observed for laparoscopy in this trial. This might be explained by the fact that this work was performed by a skilled laparoscopic surgeon, who as a preliminary, learned the laparoscopic lymphadenectomy technique. On the other hand effectiveness of laparotomy significantly increased during the trial, probably because the dissection differs when applied to non-embalmed cadavers rather than to living subjects. In comparison with laparotomy, laparoscopy easily handled this problem. Effectiveness of laparoscopy and laparotomy was reduced in a non-significant manner, in subjects with a prior history of pelvic surgery. The two techniques were still equivalent in this situation. This is probably due to adhesions and induced anatomic variations. In comparison with other papers, our number of collected lymph nodes was lower with laparoscopy as well as with the open technique (Table 4). Our surgical margins resembled those carried out by the majority of gynecologic or urologic surgeons. Conversely, the number of collected lymph nodes in our clinical practice is equivalent to those reported in the literature (data not published). We can assume that this difference is probably due to our experimental model since the pathologist and the techniques used for the node count are the same in this trial and in our routine activity. Exhaustivity of laparoscopic lymphadenectomy has been poorly described. No real randomized trial has been published and comparative reports are generally numerically limited [1,10 12,14 16]. This point is nevertheless of major importance. It governs the procedure's sensitivity. In this trial, laparotomy showed

residual tissue in four cases (13.3%). If we consider that each incomplete laparoscopic lymphadenectomy should leave one residual lymph node in place, sensitivity reaches 96.5'7,,. These results are similar to those observed in published comparative studies (Table 5). In this trial we postulate that open lymphadenectomies were always complete. Some studies have demonstrated the contrary. Lymphangiography performed after radical surgery and side wall recurrences in non-irradiated N + women show evidence that laparotomy gives residual lymph nodes in at least 5% of cases [17,18]. This result is not so different from that achieved by laparoscopy. Failures principally occurred during the first half of the trial (75%) and were mainly due to an incomplete dissection of the iliac bifurcation area (75%). In the second half of our paper we attached a special importance on dissection and inspection of the iliac bifurcation, using the medial suprapubic port for the laparoscope at the end of the procedure. This artifice prevented this kind of failure occurring again. Our complication rate was limited to 6.7%. Accidents occurred principally in the beginning of the trial and in subjects with prior history of pelvic surgery. All those accidents were venous injuries, we did not observe either ureteral, bladder or bowel injury. No procedure was aborted because of a surgical complication. This work could only evaluate intra-operative complications and our complication rate is similar to those previously reported with LPLs [19] and quite similar to the rate observed with open techniques [20]. It has to be pointed out that incomplete laparoscopic lymphadenectomies and complications occurred principally during the first half of the paper and among subjects with prior history of pelvic surgery probably because of induced anatomic variations and adhesions. A minimal number of laparoscopic lymphadenectomies seems necessary to maintain a constant efficiency in routine practice. This point was previously described by Kerbl who observed an increasing number of lymph nodes retrieved by laparoscopy and a decreasing number with open procedure alter the fifteenth case [21]. In our trial, 15 lymphadenectomies have been necessary to get a constant effectiveness in routine practice. More-

Table 5 Residual lymph nodes rate after laparoscopic pelvic lymphadenectomy

Doublet [14] f:ovder [I 5] Guazzoni [ 10] Lang [161 Parra [I 2] Querleu [I] Rukslalis [I 1] " Number ullknown.

No. of procedures

No. of residual lymph nodes

No. of residual metastatic lymph nodes (% of cases)

29 12 30 8 24 58 20

~' 7.9% of cases 3i0{~; of lymph nodes " 0 ~' ~

3 0 10 12 0 0 5

F. L&'uru et al. European Jourmd of Ot~slet;h.s & Gvm'colo,,4v and Rel,odmvitv B;o/