- Email: [email protected]
Randomized comparison of laparoscopic and open lymphadenectomy in pigs 1
Randomized Comparison of Laparoscopic and Open Lymphadenectomy in Pigs THOMAS J. HERZOG, MD, JUSTIN S. WU, DAVID M. HOVSEPIAN, MD, DONNA LUTTMAN, RN, AND ALAA ELBENDARY, MD Objective: To compare the efficacy of open and laparoscopic lymphadenectomy and validate an objective model of lymph node retrieval using lymphangiography in pigs. Methods: Twenty-five pigs weighing 54 –75 lbs were randomly assigned by side to open or laparoscopic pelvic and paraaortic lymphadenectomy. Lymph node yield, quantified by a masked pathologist, operative time, complications, blood loss, and other variables were recorded. Lymphangiography was done, and radiographs were taken before and after lymph node harvesting. Statistical analysis used McNemar test for nominal data and paired Student t test or Wilcoxon signed-rank test for continuous variables. Results: Lymph node yields were a mean of 11.5 with a standard deviation of 2.8 for open and 15.3 6 3.4 nodes for laparoscopic lymphadenectomy (P 5 .009). Mean operating time was 26.5 6 5.3 minutes for open versus 54.9 6 23.7 minutes for laparoscopy (P < .01). Mean blood loss was higher for laparoscopic cases, 35 mL for open versus 58 mL for laparoscopic lymphadenectomy (P 5 .048). The four major complications were evenly distributed between the two procedures. Lymphangiography was successful in 24 of 25 pigs. A total of 243 lymph nodes were identified with equal distribution by side. Lymphangiographic data correlated well with number of lymph nodes retrieved. Conclusion: In this randomized trial, laparoscopic lymphadenectomy was at least as effective as open lymphadenectomy, although operating time and blood loss were greater. Lymphangiography was a reliable, objective mode for documentation and assurance of lymph node recovery. (Obstet Gynecol 1999;93:603– 6. © 1999 by The American College of Obstetricians and Gynecologists.)
Laparoscopy has replaced many open surgical procedures because of shorter hospital stays, less postoperative pain, and faster return to employment. Although From the Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, and Departments of Surgery and Radiology, Washington University School of Medicine, St. Louis, Missouri. Supported in part by Washington University Minimally Invasive Surgery Institute. Funded in part by an educational grant from Ethicon-Endosurgery, Inc.
VOL. 93, NO. 4, APRIL 1999
widely used for many gynecologic procedures, it has been adopted more slowly by gynecologic oncologists, partly because of the complexity of many oncologic procedures and concern about laparoscopy providing comparably complete surgery. The area most important to widespread adoption of laparoscopic surgery by gynecologic oncologists is adequacy of lymph node dissection. Several studies showed feasibility, safety, and decrease in adhesion formation using laparoscopy in human or animal subjects.1– 6 There was still variance in node number and distribution between models, and no study prospectively compared lymph node yields by both modalities using an objective measure of thoroughness.7–9 To identify a difference in node yields between laparoscopic and open lymph node harvesting, each pig had laparoscopic dissection on one side followed by open lymphadenectomy on the other. To provide a standard for comparison, lymphangiography was used with independent review of before and after lymphadenectomy radiographs to find any nodes missed by lymphadenectomy.
Materials and Methods After institutional review board and animal studies committee approvals, 25 adult female pigs weighing 54 –75 lbs were block randomized by opening sequentially numbered sealed envelopes (containing side and modality), as determined by random number tables. Recorded data included weight, time of incision and complete closure, complications, and blood loss. Anesthesia used was a combination of intravenous tiletamine and zolazepam, ketamine, and xylazine. After induction, the pigs were intubated and anesthetized with 2% isoflurane. Laparoscopy was done with four ports in a diamond configuration with 10- to 12-mm umbilical and suprapubic ports, and two lateral 5-mm ports (Ethicon Endo-
0029-7844/99/$20.00 PII S0029-7844(98)00466-9
603
surgery, Cincinnati, OH). A three-chip camera and 30° laparoscope (Olympus, Lombard, IL), unipolar 5-mm endoscopic scissors, atraumatic graspers, endoClips (Ethicon Endosurgery), and suction aspirator were used standardly for laparoscopic nodal removal. Lymphadenectomies were done laparoscopically after achieving pneumoperitoneum to 14 mm Hg with the pigs in deep Trendelenburg (33 degrees) and video monitors appropriately positioned. Procedures were done by the first two authors (TJH and JSW), each of whom has done fewer than ten laparoscopic lymphadenectomies. The bladder was drained and the peritoneum over the external iliac vessels was incised. The paravesical and pararectal spaces were dissected bluntly. Lymph nodes were removed first from the obturator space, then from between the external and internal iliac vessels. The ureter was retracted medially with distal dissection and laterally with proximal dissection to ease node removal. Para-aortic dissection was done to the level of the renal vessels proximally after small bowel and lateral ureteral retractions. Hemostasis was achieved with monopolar cautery and endoclips, as required. Open lymphadenectomies were done through subxiphoid-to-pubis midline incision on the opposite side using standard laparotomy instruments and a similar surgical approach. Lymph nodes were immediately placed in formaldehyde and carefully labeled by date and pig identifier number before transfer to the pathology laboratory. The pathologist was blinded to side and mode of dissection. Lymph node yield was independently quantified by a single pathologist. Lymphangiography was done before lymphadenectomy by using a technique we developed, which is described in a previous publication.10 Briefly, after sterile preparation, 0.5 mL of isosulfan blue was injected into the hoof clefts and webbed spaces of each foot. Bilateral inguinal crease cut-down allowed identification of the blue-streaked lymphatics, which were then isolated and cannulated with 30-gauge needles. Four to 5 mL of ethiodized poppy seed oil was injected on each side. After lymphangiography but before lymph node harvesting, radiographs were taken of the abdomen and pelvis, then repeated after lymphadenectomy. Lymphangiograms were analyzed independently from the pathology data by a radiologist. Size and location of all visible nodes before and after lymphadenectomy were recorded. Nodes less than 1 cm were classified as small, nodes 1 to 2 cm as medium, and nodes greater than 2 cm as large. Lymph nodes designated as para-aortic required more than half the node to be above the L4-L5 disc space, otherwise they were classified as pelvic. Randomization codes were broken after the lymph nodes and lymphangiograms were analyzed. Those
604 Herzog et al
Laparoscopic Lymphadenectomy
Table 1. Comparison of Open and Laparoscopic Lymphadenectomy
Lymph node yield Operating time (min) Estimated blood loss (mL) Major complications Missed nodes on postoperative lymphangiography
Open
Laparoscopic
P
11.5 6 2.8 26.5 6 5.3 35 6 31 2 16
15.3 6 3.4 54.9 6 23.7 58 6 43 2 20
.009 ,.01 ,.048 NS NS
NS 5 not significant. Data are presented as mean 6 standard deviation or frequency.
variables and other data obtained were compared by modality of lymph node dissection using paired Student t test for continuous variables, which were expressed as mean 6 standard deviation (SD), or McNemar test for nominal data. Wilcoxon signed-rank test was used for comparison of estimated blood loss.
Results The results of this analysis are summarized in Table 1. Lymph node yield was significantly greater for laparoscopy (mean 15.3 6 3.4 per pig) versus open lymphadenectomy (mean 11.5 6 2.8 per pig) (P , .01). Operating time was significantly greater for laparoscopy. Mean estimated blood loss was statistically greater in the laparoscopic group than in the open lymphadenectomy group (Table 1). Ranges of estimated blood loss were 20 to 300 mL for laparoscopy and 10 to 100 mL for laparotomy. Despite inherent difficulties in measuring estimated blood loss, blood was meticulously collected and measured. The observed difference in estimated blood loss was negligible clinically, because all animals remained hemodynamically stable throughout the procedures. Major complications were noted in four cases, with an even distribution between modalities. Vena caval and external iliac lacerations occurred in two laparoscopic cases; one ureteral serosal injury and one laceration of a vena caval perforator happened during open surgery. There was significant blood loss with the vascular injuries in the laparoscopic procedures, with a combined loss of 600 mL of blood in those two cases. The laparoscopic caval laceration could not be completely controlled laparoscopically, and immediate laparotomy was done. The lymphadenectomy was nearly completed at the time of caval perforation. No anesthesia complications or deaths resulted. As shown in Table 2, the operating times and estimated blood losses for the first five laparoscopies were statistically greater than for the last five. Laparoscopy remained more time-consuming than open laparotomy throughout the study, with times ranging from 35 to 100
Obstetrics & Gynecology
Table 2. Comparison of Initial and Final Five Laparoscopies Variable
Mean 6 standard deviation
Time (min) First 5 Last 5 Blood loss (mL) First 5 Last 5 Complications First 5 Last 5
P
69.4 6 14 51 6 11.4
,.05
178 6 95 58 6 20
,.05
1 0
Not significant
minutes for laparoscopy, compared with 20 to 35 minutes for laparotomy. Despite increased operating time and blood loss for initial laparoscopies, no significant differences in node yields were found between the first and last five laparoscopies. Bilateral lymphangiograms (Figure 1) were feasible in 24 of 25 cases, of which 83.3% were completely interpretable. A total of 243 lymph nodes were identified by lymphangiography before lymph node harvesting (118 right, 125 left). After harvesting, 36 lymph nodes evenly distributed by side remained, as determined by radiography, a recovery rate of 85.2%. Inguinal nodes were well opacified but were not clinically relevant to this study, therefore not included in dissection or lymphangiographic analysis. Twenty nodes were missed by laparoscopy and 16 by lymphadenectomy (P 5 .62). Distribution between pelvic (19) and para-aortic (18) nodes missed was nearly equal. Only 28% of the nodes were distributed in the para-aortic region, but half of the missed nodes were from there. Most missed nodes were small (15) or medium sized (15), with no differences between modes of dissection.
Figure 1. Preoperative (left) and postoperative (right) lymphangiograms show removal of the pelvic nodes with vascular clips and opacified inguinal nodes on the postoperative film.
VOL. 93, NO. 4, APRIL 1999
Discussion The female pig is an excellent model for lymph node dissection because the scale and distribution of lymph nodes are similar to humans, except for a prominent node group lateral to the external iliac artery along the deep circumflex vessels and a relative paucity of nodes posterior to the external iliac vein. By radiography, nodal architecture appears similar to that of humans, although nodes are somewhat more coarse and clustered. Microscopically the pig’s nodal cortex was deep to the more peripheral medulla, as described.11 The present randomized trial showed that lymph node yield was greater for the laparoscopic group. The overall difference in node yields by arm was approximately four nodes per side, which was statistically significant (P , .01). The results showed that a difference of nearly eight additional nodes per pig would be harvested using laparoscopy. For both methods, the node yield exceeded the only previous porcine studies, which quantitated nodes harvested, indicating adequacy of technique.7,8 The finding that laparoscopy showed superior node yield might be due to magnification provided by the laparoscope, enhancing the view of the nodes, and facilitating dissection. Also, laparoscopic samples might be fragmented, artificially increasing the node count in favor of laparoscopy, but our pathologist did not believe that fragmentation was a major factor. Removal of nodes through 10- to 12-mm ports results in additional sectioning that might inadvertently result in a higher node count and might explain the increased laparoscopic node yield, despite no statistical difference in missed nodes by lymphangiography. Lymphangiography was used as an objective standard for comparison of node retrieval because redissecting the same surgical field, as done in some studies5,6 to confirm node retrieval, introduces potential operator bias. Lymphangiography results showed no significant differences in missed nodes between laparoscopy and laparotomy. The percentage of missed nodes from the para-aortic region, relative to total distribution, exceeded that missed from the pelvic region. The majority of missed nodes were less than 2 cm in largest dimension. Neither of those findings was unexpected, because surgical exposure for the para-aortic region is feasible yet more difficult, and the smaller lymph nodes were more likely to be missed. The learning curve effect (Table 2), with longer operating times and higher estimated blood loss, for the first five versus the final five laparoscopies was consistent with previous reports that retrospectively compared results of initial and later procedures.12,13 The complication rate, although not statistically significant, was
Herzog et al
Laparoscopic Lymphadenectomy
605
higher in the first five laparoscopies, during which a vena caval laceration occurred. The node recovery rate was not statistically different, but more time was required for harvest. The clinical importance of establishing at least equivalent node yields and recoveries between laparoscopy and laparotomy is important for laparoscopy to be adopted widely. The results of the present study provide strong evidence that laparoscopy is at least as efficacious as laparotomy for lymphadenectomy, not only in terms of diagnostic yield, but for satisfactory recovery rates. This result is important clinically because nodal dissection confers therapeutic value, such as with cervical carcinoma. Although no significant increased morbidity between modalities was noted, many additional subjects would be necessary to statistically confirm that finding. The results of our study support the potential for laparoscopic techniques to be applied to lymph node harvesting in humans.
7.
8.
9.
10. 11.
12.
13.
Carson LF, et al. Lymph node yield from laparoscopic lymphadenectomy in cervical cancer: A comparative study. Gynecol Oncol 1993;51:187–92. Lanvin D, Elhage A, Henry B, Leblanc E, Querleu D, DelobelleDeroide A. Accuracy and safety of laparoscopic lymphadenectomy: An experimental prospective randomized study. Gynecol Oncol 1997;67:83–7. Johnson N, Johnson V, McKie G, Knowles S. Laparoscopic pelvic lymphadenectomy versus traditional surgery in pigs. Br J Obstet Gynaecol 1994;101:901–2. Parra RO, Andrus C, Boullier J. Staging laparoscopic pelvic lymph node dissection: Comparison of results with open pelvic lymphadenectomy. J Urol 1992;147:875– 8. Hovsepian DM, Wu JS, Herzog TJ. Technique for performing direct lymphangiography in pigs. Acad Radiol 1997;4:361– 6. Spalding H, Heath T. Pathways of lymph flow through the superficial inguinal lymph nodes in the pig. Anat Rec 1987;217: 188 –95. Lecruru F, Taurelle R. Transperitoneal laparoscopic pelvic lymphadenectomy for gynecologic malignancies. Surg Endosc 1998;12: 1– 6. Kerbl K, Clayman RV, Petros JA, Chandhoke PS, Gill IS. Staging pelvic lymphadenectomy for prostate cancer: A comparison of laparoscopic and open techniques. J Urol 1993;150:396 – 8.
References 1. Fowler JM, Hartenbach EM, Reynolds HT, Borners J, Carter JR, Carlson JW, et al. Pelvic adhesion formation after pelvic lymphadenectomy: Comparison between transperitoneal laparoscopy and extraperitoneal laparotomy in a porcine model. Gynecol Oncol 1994;55:25– 8. 2. Kavoussi LR, Sosa E, Chandhole P, Chodak G, Clayman RV, Hadley HR, et al. Complications of laparoscopic pelvic lymph node dissection. J Urol 1993;149:322–5. 3. Querleu D, Leblanc E, Castelain B. Laparoscopic pelvic lymphadenectomy in staging of early carcinoma of the cervix. Am J Obstet Gynecol 1991;164:579 – 81. 4. Querleu D. Laparoscopic para-aortic node sampling in gynecologic oncology: A preliminary experience. Gynecol Oncol 1993;49:24 –9. 5. Childers JM, Hatch K, Surwit EA. The role of laparoscopic lymphadenectomy in the management of cervical carcinoma. Gynecol Oncol 1992;47:38 – 43. 6. Fowler JM, Carter JR, Carlson JW, Maslonkowski R, Byers LJ,
606 Herzog et al
Laparoscopic Lymphadenectomy
Address reprint requests to:
Thomas J. Herzog, MD Division of Gynecologic Oncology Washington University School of Medicine 4911 Barnes Hospital Plaza St. Louis, MO 63110 E-mail: [email protected]
Received March 26, 1998. Received in revised form September 29, 1998. Accepted October 8, 1998. Copyright © 1999 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
Obstetrics & Gynecology
Laparoscopy has replaced many open surgical procedures because of shorter hospital stays, less postoperative pain, and faster return to employment. Although From the Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, and Departments of Surgery and Radiology, Washington University School of Medicine, St. Louis, Missouri. Supported in part by Washington University Minimally Invasive Surgery Institute. Funded in part by an educational grant from Ethicon-Endosurgery, Inc.
VOL. 93, NO. 4, APRIL 1999
widely used for many gynecologic procedures, it has been adopted more slowly by gynecologic oncologists, partly because of the complexity of many oncologic procedures and concern about laparoscopy providing comparably complete surgery. The area most important to widespread adoption of laparoscopic surgery by gynecologic oncologists is adequacy of lymph node dissection. Several studies showed feasibility, safety, and decrease in adhesion formation using laparoscopy in human or animal subjects.1– 6 There was still variance in node number and distribution between models, and no study prospectively compared lymph node yields by both modalities using an objective measure of thoroughness.7–9 To identify a difference in node yields between laparoscopic and open lymph node harvesting, each pig had laparoscopic dissection on one side followed by open lymphadenectomy on the other. To provide a standard for comparison, lymphangiography was used with independent review of before and after lymphadenectomy radiographs to find any nodes missed by lymphadenectomy.
Materials and Methods After institutional review board and animal studies committee approvals, 25 adult female pigs weighing 54 –75 lbs were block randomized by opening sequentially numbered sealed envelopes (containing side and modality), as determined by random number tables. Recorded data included weight, time of incision and complete closure, complications, and blood loss. Anesthesia used was a combination of intravenous tiletamine and zolazepam, ketamine, and xylazine. After induction, the pigs were intubated and anesthetized with 2% isoflurane. Laparoscopy was done with four ports in a diamond configuration with 10- to 12-mm umbilical and suprapubic ports, and two lateral 5-mm ports (Ethicon Endo-
0029-7844/99/$20.00 PII S0029-7844(98)00466-9
603
surgery, Cincinnati, OH). A three-chip camera and 30° laparoscope (Olympus, Lombard, IL), unipolar 5-mm endoscopic scissors, atraumatic graspers, endoClips (Ethicon Endosurgery), and suction aspirator were used standardly for laparoscopic nodal removal. Lymphadenectomies were done laparoscopically after achieving pneumoperitoneum to 14 mm Hg with the pigs in deep Trendelenburg (33 degrees) and video monitors appropriately positioned. Procedures were done by the first two authors (TJH and JSW), each of whom has done fewer than ten laparoscopic lymphadenectomies. The bladder was drained and the peritoneum over the external iliac vessels was incised. The paravesical and pararectal spaces were dissected bluntly. Lymph nodes were removed first from the obturator space, then from between the external and internal iliac vessels. The ureter was retracted medially with distal dissection and laterally with proximal dissection to ease node removal. Para-aortic dissection was done to the level of the renal vessels proximally after small bowel and lateral ureteral retractions. Hemostasis was achieved with monopolar cautery and endoclips, as required. Open lymphadenectomies were done through subxiphoid-to-pubis midline incision on the opposite side using standard laparotomy instruments and a similar surgical approach. Lymph nodes were immediately placed in formaldehyde and carefully labeled by date and pig identifier number before transfer to the pathology laboratory. The pathologist was blinded to side and mode of dissection. Lymph node yield was independently quantified by a single pathologist. Lymphangiography was done before lymphadenectomy by using a technique we developed, which is described in a previous publication.10 Briefly, after sterile preparation, 0.5 mL of isosulfan blue was injected into the hoof clefts and webbed spaces of each foot. Bilateral inguinal crease cut-down allowed identification of the blue-streaked lymphatics, which were then isolated and cannulated with 30-gauge needles. Four to 5 mL of ethiodized poppy seed oil was injected on each side. After lymphangiography but before lymph node harvesting, radiographs were taken of the abdomen and pelvis, then repeated after lymphadenectomy. Lymphangiograms were analyzed independently from the pathology data by a radiologist. Size and location of all visible nodes before and after lymphadenectomy were recorded. Nodes less than 1 cm were classified as small, nodes 1 to 2 cm as medium, and nodes greater than 2 cm as large. Lymph nodes designated as para-aortic required more than half the node to be above the L4-L5 disc space, otherwise they were classified as pelvic. Randomization codes were broken after the lymph nodes and lymphangiograms were analyzed. Those
604 Herzog et al
Laparoscopic Lymphadenectomy
Table 1. Comparison of Open and Laparoscopic Lymphadenectomy
Lymph node yield Operating time (min) Estimated blood loss (mL) Major complications Missed nodes on postoperative lymphangiography
Open
Laparoscopic
P
11.5 6 2.8 26.5 6 5.3 35 6 31 2 16
15.3 6 3.4 54.9 6 23.7 58 6 43 2 20
.009 ,.01 ,.048 NS NS
NS 5 not significant. Data are presented as mean 6 standard deviation or frequency.
variables and other data obtained were compared by modality of lymph node dissection using paired Student t test for continuous variables, which were expressed as mean 6 standard deviation (SD), or McNemar test for nominal data. Wilcoxon signed-rank test was used for comparison of estimated blood loss.
Results The results of this analysis are summarized in Table 1. Lymph node yield was significantly greater for laparoscopy (mean 15.3 6 3.4 per pig) versus open lymphadenectomy (mean 11.5 6 2.8 per pig) (P , .01). Operating time was significantly greater for laparoscopy. Mean estimated blood loss was statistically greater in the laparoscopic group than in the open lymphadenectomy group (Table 1). Ranges of estimated blood loss were 20 to 300 mL for laparoscopy and 10 to 100 mL for laparotomy. Despite inherent difficulties in measuring estimated blood loss, blood was meticulously collected and measured. The observed difference in estimated blood loss was negligible clinically, because all animals remained hemodynamically stable throughout the procedures. Major complications were noted in four cases, with an even distribution between modalities. Vena caval and external iliac lacerations occurred in two laparoscopic cases; one ureteral serosal injury and one laceration of a vena caval perforator happened during open surgery. There was significant blood loss with the vascular injuries in the laparoscopic procedures, with a combined loss of 600 mL of blood in those two cases. The laparoscopic caval laceration could not be completely controlled laparoscopically, and immediate laparotomy was done. The lymphadenectomy was nearly completed at the time of caval perforation. No anesthesia complications or deaths resulted. As shown in Table 2, the operating times and estimated blood losses for the first five laparoscopies were statistically greater than for the last five. Laparoscopy remained more time-consuming than open laparotomy throughout the study, with times ranging from 35 to 100
Obstetrics & Gynecology
Table 2. Comparison of Initial and Final Five Laparoscopies Variable
Mean 6 standard deviation
Time (min) First 5 Last 5 Blood loss (mL) First 5 Last 5 Complications First 5 Last 5
P
69.4 6 14 51 6 11.4
,.05
178 6 95 58 6 20
,.05
1 0
Not significant
minutes for laparoscopy, compared with 20 to 35 minutes for laparotomy. Despite increased operating time and blood loss for initial laparoscopies, no significant differences in node yields were found between the first and last five laparoscopies. Bilateral lymphangiograms (Figure 1) were feasible in 24 of 25 cases, of which 83.3% were completely interpretable. A total of 243 lymph nodes were identified by lymphangiography before lymph node harvesting (118 right, 125 left). After harvesting, 36 lymph nodes evenly distributed by side remained, as determined by radiography, a recovery rate of 85.2%. Inguinal nodes were well opacified but were not clinically relevant to this study, therefore not included in dissection or lymphangiographic analysis. Twenty nodes were missed by laparoscopy and 16 by lymphadenectomy (P 5 .62). Distribution between pelvic (19) and para-aortic (18) nodes missed was nearly equal. Only 28% of the nodes were distributed in the para-aortic region, but half of the missed nodes were from there. Most missed nodes were small (15) or medium sized (15), with no differences between modes of dissection.
Figure 1. Preoperative (left) and postoperative (right) lymphangiograms show removal of the pelvic nodes with vascular clips and opacified inguinal nodes on the postoperative film.
VOL. 93, NO. 4, APRIL 1999
Discussion The female pig is an excellent model for lymph node dissection because the scale and distribution of lymph nodes are similar to humans, except for a prominent node group lateral to the external iliac artery along the deep circumflex vessels and a relative paucity of nodes posterior to the external iliac vein. By radiography, nodal architecture appears similar to that of humans, although nodes are somewhat more coarse and clustered. Microscopically the pig’s nodal cortex was deep to the more peripheral medulla, as described.11 The present randomized trial showed that lymph node yield was greater for the laparoscopic group. The overall difference in node yields by arm was approximately four nodes per side, which was statistically significant (P , .01). The results showed that a difference of nearly eight additional nodes per pig would be harvested using laparoscopy. For both methods, the node yield exceeded the only previous porcine studies, which quantitated nodes harvested, indicating adequacy of technique.7,8 The finding that laparoscopy showed superior node yield might be due to magnification provided by the laparoscope, enhancing the view of the nodes, and facilitating dissection. Also, laparoscopic samples might be fragmented, artificially increasing the node count in favor of laparoscopy, but our pathologist did not believe that fragmentation was a major factor. Removal of nodes through 10- to 12-mm ports results in additional sectioning that might inadvertently result in a higher node count and might explain the increased laparoscopic node yield, despite no statistical difference in missed nodes by lymphangiography. Lymphangiography was used as an objective standard for comparison of node retrieval because redissecting the same surgical field, as done in some studies5,6 to confirm node retrieval, introduces potential operator bias. Lymphangiography results showed no significant differences in missed nodes between laparoscopy and laparotomy. The percentage of missed nodes from the para-aortic region, relative to total distribution, exceeded that missed from the pelvic region. The majority of missed nodes were less than 2 cm in largest dimension. Neither of those findings was unexpected, because surgical exposure for the para-aortic region is feasible yet more difficult, and the smaller lymph nodes were more likely to be missed. The learning curve effect (Table 2), with longer operating times and higher estimated blood loss, for the first five versus the final five laparoscopies was consistent with previous reports that retrospectively compared results of initial and later procedures.12,13 The complication rate, although not statistically significant, was
Herzog et al
Laparoscopic Lymphadenectomy
605
higher in the first five laparoscopies, during which a vena caval laceration occurred. The node recovery rate was not statistically different, but more time was required for harvest. The clinical importance of establishing at least equivalent node yields and recoveries between laparoscopy and laparotomy is important for laparoscopy to be adopted widely. The results of the present study provide strong evidence that laparoscopy is at least as efficacious as laparotomy for lymphadenectomy, not only in terms of diagnostic yield, but for satisfactory recovery rates. This result is important clinically because nodal dissection confers therapeutic value, such as with cervical carcinoma. Although no significant increased morbidity between modalities was noted, many additional subjects would be necessary to statistically confirm that finding. The results of our study support the potential for laparoscopic techniques to be applied to lymph node harvesting in humans.
7.
8.
9.
10. 11.
12.
13.
Carson LF, et al. Lymph node yield from laparoscopic lymphadenectomy in cervical cancer: A comparative study. Gynecol Oncol 1993;51:187–92. Lanvin D, Elhage A, Henry B, Leblanc E, Querleu D, DelobelleDeroide A. Accuracy and safety of laparoscopic lymphadenectomy: An experimental prospective randomized study. Gynecol Oncol 1997;67:83–7. Johnson N, Johnson V, McKie G, Knowles S. Laparoscopic pelvic lymphadenectomy versus traditional surgery in pigs. Br J Obstet Gynaecol 1994;101:901–2. Parra RO, Andrus C, Boullier J. Staging laparoscopic pelvic lymph node dissection: Comparison of results with open pelvic lymphadenectomy. J Urol 1992;147:875– 8. Hovsepian DM, Wu JS, Herzog TJ. Technique for performing direct lymphangiography in pigs. Acad Radiol 1997;4:361– 6. Spalding H, Heath T. Pathways of lymph flow through the superficial inguinal lymph nodes in the pig. Anat Rec 1987;217: 188 –95. Lecruru F, Taurelle R. Transperitoneal laparoscopic pelvic lymphadenectomy for gynecologic malignancies. Surg Endosc 1998;12: 1– 6. Kerbl K, Clayman RV, Petros JA, Chandhoke PS, Gill IS. Staging pelvic lymphadenectomy for prostate cancer: A comparison of laparoscopic and open techniques. J Urol 1993;150:396 – 8.
References 1. Fowler JM, Hartenbach EM, Reynolds HT, Borners J, Carter JR, Carlson JW, et al. Pelvic adhesion formation after pelvic lymphadenectomy: Comparison between transperitoneal laparoscopy and extraperitoneal laparotomy in a porcine model. Gynecol Oncol 1994;55:25– 8. 2. Kavoussi LR, Sosa E, Chandhole P, Chodak G, Clayman RV, Hadley HR, et al. Complications of laparoscopic pelvic lymph node dissection. J Urol 1993;149:322–5. 3. Querleu D, Leblanc E, Castelain B. Laparoscopic pelvic lymphadenectomy in staging of early carcinoma of the cervix. Am J Obstet Gynecol 1991;164:579 – 81. 4. Querleu D. Laparoscopic para-aortic node sampling in gynecologic oncology: A preliminary experience. Gynecol Oncol 1993;49:24 –9. 5. Childers JM, Hatch K, Surwit EA. The role of laparoscopic lymphadenectomy in the management of cervical carcinoma. Gynecol Oncol 1992;47:38 – 43. 6. Fowler JM, Carter JR, Carlson JW, Maslonkowski R, Byers LJ,
606 Herzog et al
Laparoscopic Lymphadenectomy
Address reprint requests to:
Thomas J. Herzog, MD Division of Gynecologic Oncology Washington University School of Medicine 4911 Barnes Hospital Plaza St. Louis, MO 63110 E-mail: [email protected]
Received March 26, 1998. Received in revised form September 29, 1998. Accepted October 8, 1998. Copyright © 1999 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.
Obstetrics & Gynecology