- Email: [email protected]
Liposuction-assisted nerve-sparing extended radical hysterectomy: Oncologic rationale, surgical anatomy, and feasibility study
Liposuction-assisted nerve-sparing extended radical hysterectomy: Oncologic rationale, surgical anatomy, and feasibility study Michael Höckel, MD, PhD,a Moritz A. Konerding, MD,b and Claus Peter Heußel, MDc Mainz, Germany OBJECTIVE: Our purpose was to improve the therapeutic index of radical hysterectomy by extending the resection of parametrial tissue without further impairing pelvic autonomic nerve functions. STUDY DESIGN: We studied the topographic anatomy of the parametrial tissue with high-resolution magnetic resonance imaging and by dissection of fresh human cadavers. We then performed a clinical feasibility study of the liposuction-assisted nerve-sparing extended radical hysterectomy. RESULTS: Magnetic resonance imaging demonstrated that the perispinous adipose tissue is retained after type III radical hysterectomy and is a frequent site of tumor recurrence. The perispinous adipose tissue contains the pelvic plexus, the pelvic splanchnic nerves, small blood vessels, and lymphatic tissue. We developed the liposuction-assisted nerve-sparing extended radical hysterectomy and applied it to seven consecutive patients with cervical or vaginal cancer. No intraoperative or postoperative complications occurred. Postoperative magnetic resonance imaging assured us that perispinous adipose tissue was cleared in all cases. A metastatic lymph node was found in the perispinous adipose tissue removed by liposuction from one patient. Suprapubic cystostomies could be removed after a median period of 12 days. CONCLUSION: The nerve-sparing removal of perispinous adipose tissue by liposuction is a feasible addition to wide en bloc parametrectomy in anatomically defined planes. (Am J Obstet Gynecol 1998;178:971-6.)
Key words: Radical hysterectomy, pelvic autonomic nerves, cervical cancer, liposuction
Radical hysterectomy is a time-honored surgical treatment of International Federation of Gynecology and Obstetrics stages IA2, IB, and IIA cervical cancer and stage IIB endometrial cancer.1-3 Some European authorities in gynecologic oncology include International Federation of Gynecology and Obstetrics stage IIB cervical cancer in the indications for radical hysterectomy, whereas this is disputed by others.4, 5 After decades of conformity with the abdominal approach known as Wertheim-Meigs operation and its variants, the introduction of advanced endoscopic equipment has recently allowed minimally invasive radical hysterectomy techniques.6, 7 Likewise, fertility-preserving surgical techniques have been developed and are being offered in experimental settings to young patients with early cervical cancer.8 Irrespective of the indication field and the operative
From the Departments of Obstetrics and Gynecology,a Anatomy,b and Radiology,c University of Mainz. Supported by a grant from Else Kröner-Fresenius-Stiftung to M.H. Received for publication July 28 1997; revised October 15, 1997; accepted October 30, 1997. Reprint requests: Michael Höckel, MD, PhD, University of Mainz Medical School, Langenbeckstraße 1, 55101 Mainz, Germany Copyright © 1998 by Mosby, Inc. 0002-9378/98 $5.00 + 0 6/1/87330
techniques there are three major surgicoanatomic points of radical hysterectomy that have not been adequately addressed so far: 1. The amount of “parametrial” tissue to be removed by radical hysterectomy has not been clarified according to exact anatomic landmarks in all planes. 2. Parametrectomy and pelvic and periaortic lymphadenectomy techniques as part of radical hysterectomy are not defined with respect to the pelvic autonomic nerves. 3. Parametrectomy techniques do not relate to the mesorectum, the embryologically determined plane with clearly documented relevance for rectal cancer surgery.9, 10 As a consequence, parametrectomy and periaortic and pelvic lymph node dissection may be suboptimal with respect to oncologic radicalness on one hand and to functional postoperative sequelae on the other hand. Anatomically more precise surgery is likely to improve the therapeutic index of this important operative treatment in gynecologic oncology. We used two new means to investigate this subject, magnetic resonance imaging (MRI) and liposuction. In a combined cadaver-clinical study we demonstrated the following: Residual parametrial tissue not resected with conventional hysterectomy techniques is a source for re971
972 Höckel, Konerding, and Heußel
May 1998 Am J Obstet Gynecol
Fig. 1. T2-weighted axial MRI scans of the female pelvis demonstrating (A) normal anatomy, (B) right infrailiac acetabular side wall recurrence of cervical cancer, and the postoperative situs (C) after conventional type III radical hysterectomy and (D) after nerve-sparing extended radical hysterectomy. C, Normal uterine cervix; T, tumor. Asterisks indicate perispinous adipose tissue; arrowheads point to anterolateral mesorectal plane. Circles highlight liquid signals demonstrating clearance of perispinous adipose tissue after nerve-sparing extended radical hysterectomy. In all panels the lettering locates the right ventral position.
currences; parametrectomy can be surgically defined with respect to the course of the pelvic plexus and to the anterolateral mesorectal planes; parametrectomy and systematic pelvic and periaortic lymph node dissection can be performed to spare most of the pelvic autonomic nerves, except the periureteral branches of the pelvic plexus. Material and methods MRI. Scans were performed on a Siemens Expert 1.0 T Magnetom (Siemens, Erlangen, Germany). A T2weighted turbo spin echo sequence was used in transverse and sagittal orientation. A T1-weighted spin echo sequence was used in transverse orientation and repeated after administration of contrast agent. Additionally a turbo spin echo sequence with fat suppression was used in coronal orientation. This sequence was optionally repeated in a sagittal orientation. The total time of data acquisition was about 25 minutes. Parametrial liposuction. Liposuction cannulas for parametrectomy were developed by Normed Medizintechnik Ltd. (Tuttlingen, Germany) according to specifications that emerged from the cadaver study. The cannulas are manufactured of high-grade stainless steel with one blunt hole at the inside of the slightly bent tip. A set of cannulas with lumen diameters of 3 and 4 mm and lengths of 17 and 20 cm provides the necessary versatility. The cannulas are connected to a high-performance
Fig. 2. The superior hypogastric plexus (SHP) and left and right hypogastric nerves (HN) isolated (A) in a female cadaver and (B) in a patient undergoing nerve-sparing extended radical hysterectomy, lateromedial view. LSN, Lumbar splanchnic nerves; U, left ureter; R, rectum; MR, left anterolateral mesorectal fascia.
vacuum pump producing 20 mm Hg residual pressure. Short strokes with the cannula in a piston-like manner are performed in anterior-posterior direction, with the hole directed to the mid pelvis beginning at the pelvic wall and moving medially. Cadaver study. Fresh female cadavers without evidence of pelvic operations or pathology were laparotomized. The retroperitoneum was entered by splitting the parietal peritoneum along the pelvic parts of the psoas muscles, paracolic gutters, and radix mesenterii. The visceral peritoneum was opened at the pouch of Douglas and in the uterovesical fold. The pararectal and paravesical spaces were fully opened. The rectum was dissected in the anterior and posterior mesorectal planes. The lymphatic tissue along the common, external, and internal iliac vessels and in the obturator fossa was stripped off. After complete removal of fat from the pararectal, parametrial, and paravesical tissues by liposuction, the topographic anatomy of the pelvic autonomic nerves was studied. Liposuction-assisted nerve-sparing extended radical hysterectomy. The possibility of nerve-sparing extended radical hysterectomy was offered to patients to be operated for cervical cancer (by M.H.). Patients who gave in-
Volume 178, Number 5 Am J Obstet Gynecol
Höckel, Konerding, and Heußel 973
formed consent underwent operation with these new surgical techniques. All surgical specimens, including the tissues obtained by liposuction, were fixed in formalin and routinely processed for histologic investigation. All patients underwent a pelvic MRI series in the second postoperative week. Results MRI studies. The sectional anatomy of the parametrial tissues was studied with pelvic MRI series of patients without gynecologic pathology. Sections 5 mm thick around the ischial spine clearly demonstrated a bilateral triangular zone of adipose tissue between the mesorectum (lower border), uterovaginal venous plexus (upper border), and obturator internus, coccygeus, iliococcygeus muscle (lateral border), designated perispinous adipose tissue (Fig. 1, A). Postoperative MRI series of patients treated with conventional (type III) radical abdominal hysterectomy by five experienced gynecologic surgeons revealed that the perispinous adipose tissue remained completely in situ (Fig. 1, C). MRI examinations were performed on 46 patients operated on in this institution from 1990 to 1996 with the laterally extended endopelvic resection–combined operative and radiotherapeutic treatment techniques for cervical cancer recurring at the pelvic side wall after radical hysterectomy and radiation. These scans demonstrated a topographic relationship of the recurrence to the nonresected perispinous adipose tissue in 21 cases (37%; Fig. 1, B), suggesting that residual parametrial tissue, in particular the perispinous adipose tissue, is a potential source of tumor recurrence after conventional radical hysterectomy in high-risk tumors. Liposuction-assisted cadaver dissection. The surgical anatomy of the female pelvic autonomic nerve system with respect to radical hysterectomy was investigated by liposuction-assisted pelvic dissection in five fresh human cadavers. The superior hypogastric plexus, a nerve band approximately 5 mm in width and between 4 and 5 cm in length, is formed at the level of the bifurcation of the aorta by two main lumbar splanchnic nerves arising from L2 and L3 positions laterally and medially to the aorta. The plexus descends at the surface of the left common iliac artery and vein to the level of the promontorium, where it divides into the two hypogastric nerves approximately 3 mm wide that run for about 10 cm at the left and right lateral mesorectum (Fig. 2). One to two centimeters below the level of the pouch of Douglas, each hypogastric nerve and three or four pelvic splanchnic nerves approaching laterally from the perispinous pelvic wall (side of attachment of coccygeus muscle to obturator internus muscle) form the pelvic plexus, a complex 3 to 4 × 2 to 3 cm flat network of nervous tissue. The pelvic plexuses on both sides are completely embedded in the perispinous adipose tissue (Fig. 3, A). After removal of the fat from
Fig. 3. Demonstration of left parametrium and perispinous adipose tissue in a cadaver (A) before and (B) after removal of fat with liposuction, craniocaudal view. U, Left ureter; UA, uterine artery; UV, uterine vein; HN, hypogastric nerve; PP, pelvic plexus; RUL, rectouterine ligament. Resection line for wide en bloc parametrectomy in nerve-sparing extended radical hysterectomy technique is indicated. Arrowhead points to pelvic splanchnic nerves.
the parametrial and paramesorectal tissue with liposuction, the pelvic plexus can be visualized (Fig. 3, B). A few medial branches of the pelvic plexus enter the mesorectum; most branches run forward laterally to the vagina and bladder. Most of these nerves are located below the uterine artery and vein at the level of the inferior vesical artery and vein. After giving off further medial branches to the cervix and vagina, the pelvic plexus radiates into the lateral bladder wall and trigone. Most of these branches run dorsally to the ureter and can be spared when the ureter is dissected out of the parametrial tissue (Fig. 4, A). On the basis of the topographic anatomic results of the cadaver study, the best compromise between maximum parametrectomy and sparing of the pelvic autonomic nerves appears to be the following: dissection of the rectouterine and rectovaginal ligaments exactly in the anterolateral mesorectal planes; separation of the uterine artery and veins immediately after branching from the parent vessels (hypogastric, vesicalis superior, or obturator vessels); transection of the pubocervical fascia (vesicouterine, vesicovaginal ligaments) at the projection
974 Höckel, Konerding, and Heußel
Fig. 4. A, Specimen of nerve-sparing extended radical hysterectomy. B, Postresection situs demonstrating right pelvis viewing mediolaterally. Most branches of the pelvic plexus to the lateral bladder wall and trigone (arrowheads) run dorsally of the ureter and are spared when dissecting the ureter out of the parametrial tissue and performing liposuction of the perispinous tissue. U, Ureter; V, vagina (open); MR, mesorectum; HN, hypogastric nerve; PP, pelvic plexus; PSN, pelvic splanchnic nerves. In addition to the en bloc specimen shown in A, between 5 and 10 ml of lymphatic adipose tissue is obtained by liposuction of the perispinous adipose tissue on each side.
of the parametrial course of the ureters; liposuction of the perispinous adipose tissue to expose the pelvic plexus; and lateral dissection of the cardinal ligament and the paracolpium with ventrolateral displacement of residual pelvic plexus after separation of the supraureteric branches during the dissection of the pubocervical ligament. Liposuction-assisted nerve-sparing extended radical hysterectomy. The operation starts with a hypogastric and epigastric midline laparotomy. The parietal peritoneum is incised at both paracolic gutters and along the radix mesenterii, the duodenum is mobilized from the aorta and vena cava, and the small and large bowel are packed into an intestinal bag. The pelvic parietal peritoneum is opened above the psoas muscles and the round ligaments are dissected. Depending on the age of the patient or on adnexal abnormality, the infundibulopelvic ligaments or ligamenta ovarii propria and proximal fallopian tubes are dissected to include or exclude the adnexa into the hysterectomy specimen.
May 1998 Am J Obstet Gynecol
The broad ligament of the uterus is incised anteriorly and posteriorly, and the paravesical and pararectal spaces are developed. Both uterine arteries are ligated and cut immediately after their branching from the hypogastric or obturator arteries. The pouch of Douglas is incised and the posterior vaginal wall is bluntly separated from the anterior rectum in the midline for between 4 and 5 cm. The visceral peritoneum in the vesicouterine fold is incised, and the bladder is completely detached from the midsagittal cervix and the proximal third of the anterior vaginal wall. At the aortic bifurcation the superior hypogastric plexus is identified and mobilized from the left common iliac artery and vein and from the promontorium. The presacral space is opened with electrocautery in the mesorectal plane to follow both hypogastric nerves without detaching them from the mesorectum. The rectouterine and rectovaginal ligaments are separated from the rectum by electrodissection exactly in the mesorectal plane. The posterior parametrial structures are clamped and ligated only toward the uterus to prevent lymphatic tumoral spread. No clamps are used toward the rectum, to allow the exact separation in the anterolateral mesorectal planes. Complete mobilization of the ureters out of the cardinal and pubocervical ligaments is facilitated by use of a scoop. Then the proper parametrial liposuction cannula (Normed Medizintechnik) is inserted into the parametrial fatty tissue at the parietal endopelvic fascia. Liposuction creates an approximately 1 cm fat-free space, skeletizing the distal uterine veins, which are then dissected between hemoclips immediately before branching from the hypogastric vein. Liposuction continues caudally at the level of the ischial spine along the iliococcygeus and coccygeus muscles to expose the middle hemorrhoidal vessels and the pelvic splanchnic nerves. These structures are spared. Multiple small (≤5 mm) lymph nodes are usually encountered during the maneuver. They can be easily harvested for histologic examination. The exposed pelvic splanchnic nerves and the hypogastric nerves delineate the pelvic plexus on both sides. The deep dissection of the cardinal ligament, pubocervical fascia, and paracolpium is performed with clamps above the level of the pelvic plexus, or with lateral displacement of the pelvic plexus if a more extensive vaginal resection is necessary. The vagina is incised at the oncologically adequate location between strong clamps placed both distally and proximally. Complete pelvic and periaortic lymphadenectomy follows, with the superior hypogastric plexus and the lumbar splanchnic nerves spared. After thorough washings, a peritoneal bridge between bladder and rectum is formed; this is used to support the middle pelvic parts of both ureters. The abdomen is closed with a running Smead-Jones suture. Clinical feasibility study. Nerve-sparing extended radi-
Höckel, Konerding, and Heußel 975
Volume 178, Number 5 Am J Obstet Gynecol
Table I. Clinical and histopathologic features of the first seven patients treated with liposuction-assisted nerve-sparing extended radical hysterectomy and parametrectomy
Patient
Age (yr)
Tumor type
FIGO stage
1 2* 3 4 5† 6 7
30 39 41 58 49 39 37
SC, cervix SC, cervix SC, vagina SC, cervix AC, cervix ASC, cervix SC, cervix
IIB IB IIB IIB IB IIB IB
pTpN
Lymph nodes removed
T2bN0 T1bN0 T1bN0 T2bN1 T1bN0 T2bN1 T1bN1
48 33 35 66 39 76 46
Total parametrial area resected (cm2) 25 18 85 48 35 50 46
Histopathologic Intraoperative tumor blood loss size (mm) (ml) 25 15 40 57 19 30 15
800 500 1000 1500 1000 600 800
Postoperative cystostomy (days) 8 14 21 12 7 30 11
FIGO, International Federation of Gynecology and Obstetrics; SC, squamous cell carcinoma; AC, adenocarcinoma; ASC, adenosquamous carcinoma. *After simple vaginal hysterectomy. †After conization.
cal hysterectomy was carried out in seven consecutive patients with cervical and vaginal cancer (Table I). No intraoperative complications occurred and total blood loss (including lymph node dissection) was relatively low, median 800 ml (range between 500 and 1500 ml). The removal of the perispinous adipose tissue was demonstrated by postoperative MRI in patients treated with the new technique (Fig. 1, D). The histopathologic workup of the surgical specimens showed a median parametrial en bloc resection area of 46 cm2 (range between 18 and 85 cm2; Fig. 4, B). The microscopic examination of the tissue obtained by liposuction after processing and staining found predominantly adipose tissue, blood cells, and lymphatic tissue. In patient 6, who had an adenosquamous cancer stage T2bN1, one small lymph node obtained by liposuction of the perispinous adipose tissue contained a metastasis. The postoperative courses of all patients were uneventful, without complications or deaths. The patients recovered within 3 days. Suprapubic cystostomies could be withdrawn after a median 12 days (range 8 to 30 days), when residual urine was constantly <50 ml. Comment According to the concepts of surgical oncology, radical treatment involves the wide en bloc resection of the tumor-bearing organ along with its primary lymphatic drainage compartment. The relevance of this principle has been recently reappraised with total mesorectal excision techniques for the surgical treatment of rectal cancer.9, 10 For the operative therapy of cervical cancer, radical hysterectomy adheres to the same oncologic principle, resection of the uterus en bloc with the parametria. However, liberation of the ureters and the pelvic plexus to sustain their function interferes with the continuity of the parametrial resection. Thus the extent of parametrectomy is a crucial issue in radical hysterectomy. Despite its importance, the surgical anatomy of parame-
trectomy has only been defined with respect to the uterine vessels and the ureter in European and American schools of gynecologic oncology.11 Only vague descriptions concerning the extend of parametrial resection with respect to the pelvic plexus are given in textbooks and surgical atlases. Moreover, parametrectomy has not to our knowledge been addressed with reference to anterolateral planes of the mesorectum. MRI clearly displays a distinct region of perispinous adipose tissue bordered mediocaudally by the anterolateral mesorectum, bordered laterally by the coccygeus, iliococcygeus, and obturator internus muscles, and bordered medioventrally by the vesicouterovaginal venous plexus. The MRI series from patients referred to us for the laterally extended endopelvic resection–combined operative and radiotherapeutic treatment12, 13 of pelvic sidewall recurrences of cervical cancer showed an involvement of the perispinous adipose tissue in 37% of cases. In the first series of seven patients treated with the nerve-sparing extended radical hysterectomy technique, we removed a metastatic lymph node 3 mm in diameter from the perispinous fatty tissue. Both observations point to the importance of the perispinous adipose tissue in tumoral spread in cervical cancer. However, perispinous adipose tissue is not resected in conventional type III radical hysterectomy procedures, as demonstrated by postoperative MRI scans. Perispinous adipose tissue was identified as the site of pelvic plexus in five female cadavers and corresponds to the pars nervosa of the cardinal ligament and the so called “lateral ligament of the rectum.” Recent anatomic studies revealed that the lateral ligament of the rectum is not a real anatomic ligament but rather a surgical construction from former rectal resection techniques as a consequence of the fusion of the pelvic plexus to the lateral mesorectum.10, 14 Likewise, we found no anatomic correlation to an “uterosacral ligament” between the sacrum and the uterus. The functional elongation of the
976 Höckel, Konerding, and Heußel
rectouterine condensation of the broad ligament from the rectum to the sacrum appears to be due to the pelvic plexus and pelvic splanchnic nerves adhering to the mesorectum. On the basis of the anatomic results of the study of the fat-free parametrial tissues in cadavers, we developed a new surgical technique for radical hysterectomy intended to compromise between maximum parametrial radicalness and sustained pelvic autonomic nerve function. The topographic anatomy of the pelvic plexus does not allow its complete preservation in radical hysterectomy because the detachment of the ureter out of the parametrial vesicouterovaginal venous plexus necessitates the disruption of its branches running to the bladder ventral to the ureter. However, (1) en bloc resection of the parametrial tissue down to the hypogastric nerves as landmarks for the dorsolateral resection line, (2) dissection of the rectouterine and rectovaginal ligaments in the anterolateral mesorectal planes, and (3) liposuction of the residual perispinous adipose tissue remove all parametrial lymphatic tissue and retain the pelvic splanchnic nerves and the pelvic plexus with its genitovesical branches dorsal to the ureteric level. Inherent in the new radical hysterectomy technique are the identification and sparing of the superior hypogastric plexus during upper pelvic and periaortic lymphadenectomy. Nerve-sparing radical hysterectomy techniques have also been developed in China and Japan. Their superiority to conventional techniques has been claimed; however, only a few publications are available in English.15-18 A detailed study of the surgical anatomy of the parametrial tissue and an anatomically defined nerve-sparing resection technique have been published by Yabuki et al.19, 20 These authors used an ultrasonic surgical aspirator for the removal of the parametrial alveolar tissue. We and others have reservations about the use of ultrasonically powered tissue removal techniques in oncologic situations, in which intraperitoneal tumor cell dissemination could be harmful.21, 22 Our preliminary results show that the new surgical techniques of liposuction-assisted extended radical hysterectomy are feasible. The histopathologic workup and the postoperative MRI prove there is a larger extent of removed parametrial tissue than is obtained with conventional type III radical hysterectomy techniques. However, postoperative bladder function, as preliminarily evaluated by measuring residual urine, is not more impaired. This suggests a gain in therapeutic index. We have initiated a prospective controlled study investigating bladder, vaginal, and rectal function after liposuction-assisted nerve-sparing extended radical hysterectomy by means of longitudinal questionnaires and urodynamic measurements. The long-term functional results and the oncologic outcome of patients with cervical
May 1998 Am J Obstet Gynecol
cancer, in particular, those with features carrying a high risk of local recurrence after surgical treatment, are awaited.
REFERENCES
1. Wertheim E. The extended abdominal operation for carcinoma uteri (based on 500 operative cases). Am J Obstet Dis Women Child 1912;66:169-232. 2. Okabayashi H. Radical abdominal hysterectomy for cancer of the cervix uteri. Surg Gynecol Obstet 1921;33:335-41. 3. Meigs JV. Carcinoma of the cervix: the Wertheim operation. Surg Gynecol Obstet 1944;78:195-8. 4. Burghardt E, Baltzer J, Tulusan AH, Haas J. Results of surgical treatment of 1028 cervical cancers studied with volumetry. Cancer 1992;70:648-55. 5. Friedberg V. Operative therapy for stage IIb cervical cancer. Baillieres Clin Obstet Gynaecol 1988;2:973-80. 6. Nezhat CR, Burrell MO, Nezhat FR, Benigno BB, Welander CE. Laparoscopic radical hysterectomy with paraaortic and pelvic node dissection. Am J Obstet Gynecol 1992;166:864-5. 7. Dargent D, Mathevet P. Schauta’s vaginal hysterectomy combined with laparoscopic lymphadenectomy. Baillieres Clin Obstet Gynaecol 1995;9:691-705. 8. Dargent D, Brun JL, Roy M, Rémy I. Pregnancies following radical trachelectomy for invasive cervical cancer [abstract]. Gynecol Oncol 1994;52:105. 9. Heald RJ, Husband EM, Ryall RDH. The mesorectum in rectal cancer surgery: the clue to pelvic recurrence? Br J Surg 1982;69:613-6. 10. Enker WE, Thaler HT, Cranor ML, Polyak T. Total mesorectal excision in the operative treatment of rectal cancer. J Am Coll Surg 1995;181:335-46. 11. Piver MS, Rutledge F, Smith JP. Five classes of extended hysterectomy for women with cervical cancer. Obstet Gynecol 1974;44:265-72. 12. Höckel M, Baußmann E, Mitze M, Knapstein PG. Are pelvic side-wall recurrences of cervical cancer biologically different from central relapses? Cancer 1994;74:648-55. 13. Höckel M, Schlenger K, Hamm H, Knapstein PG, Hohenfellner R, Rösler HP. Five-year experience with combined operative and radiotherapeutic treatment of recurrent gynecologic tumors infiltrating the pelvic wall. Cancer 1996;77:1918-33. 14. Sato K, Sato T. The vascular and neuronal composition of the lateral ligament of the rectum and the rectosacral fascia. Surg Radiol Anat 1991;13:17-22. 15. Sasaki H, Yoshida T, Noda K, Yachiku S, Minami K, Kaneko S. Urethral pressure profiles following radical hysterectomy. Obstet Gynecol 1982;59:101-4. 16. Wu YX. Preservation of pelvic nerve plexus during radical hysterectomy. Chung Hua Fu Chan Ko Tsa Chih 1991;26:368-71. 17. Tong XK, Huo RJ. The anatomical basis and prevention of neurogenic voiding dysfunction following radical hysterectomy. Surg Radiol Anat 1991;13:145-8. 18. Wu YX, Qui S, Meng J. Careful dissection of the pelvic plexus helps alleviate urinary retention following radical hysterectomy. Chung Hua Chung Liu Tsa Chih 1994;16:465-8. 19. Yabuki Y, Asamoto A, Hoshiba T, Nishimoto H, Kitamura S. Dissection of the cardinal ligament in radical hysterectomy for cervical cancer with emphasis on the lateral ligament. Am J Obstet Gynecol 1991;164:7-14. 20. Yabuki Y, Asamoto A, Hoshiba T, Nishimoto H, Satou N. A new proposal for radical hysterectomy. Gynecol Oncol 1996;62:3708. 21. van Dam PA, Coppens M, van Oosterom AT, Van Marck E, Buytaert P. Is there an increased risk for tumor dissemination using ultrasonic surgical aspiration in patients with vulvar carcinoma? Eur J Obstet Gynecol Reprod Biol 1994;55:145-7. 22. Nahhas WA. Potential hazard of use of surgical ultrasonic aspirator in tumor reductive surgery. Gynecol Oncol 1991;40:81-3.
Key words: Radical hysterectomy, pelvic autonomic nerves, cervical cancer, liposuction
Radical hysterectomy is a time-honored surgical treatment of International Federation of Gynecology and Obstetrics stages IA2, IB, and IIA cervical cancer and stage IIB endometrial cancer.1-3 Some European authorities in gynecologic oncology include International Federation of Gynecology and Obstetrics stage IIB cervical cancer in the indications for radical hysterectomy, whereas this is disputed by others.4, 5 After decades of conformity with the abdominal approach known as Wertheim-Meigs operation and its variants, the introduction of advanced endoscopic equipment has recently allowed minimally invasive radical hysterectomy techniques.6, 7 Likewise, fertility-preserving surgical techniques have been developed and are being offered in experimental settings to young patients with early cervical cancer.8 Irrespective of the indication field and the operative
From the Departments of Obstetrics and Gynecology,a Anatomy,b and Radiology,c University of Mainz. Supported by a grant from Else Kröner-Fresenius-Stiftung to M.H. Received for publication July 28 1997; revised October 15, 1997; accepted October 30, 1997. Reprint requests: Michael Höckel, MD, PhD, University of Mainz Medical School, Langenbeckstraße 1, 55101 Mainz, Germany Copyright © 1998 by Mosby, Inc. 0002-9378/98 $5.00 + 0 6/1/87330
techniques there are three major surgicoanatomic points of radical hysterectomy that have not been adequately addressed so far: 1. The amount of “parametrial” tissue to be removed by radical hysterectomy has not been clarified according to exact anatomic landmarks in all planes. 2. Parametrectomy and pelvic and periaortic lymphadenectomy techniques as part of radical hysterectomy are not defined with respect to the pelvic autonomic nerves. 3. Parametrectomy techniques do not relate to the mesorectum, the embryologically determined plane with clearly documented relevance for rectal cancer surgery.9, 10 As a consequence, parametrectomy and periaortic and pelvic lymph node dissection may be suboptimal with respect to oncologic radicalness on one hand and to functional postoperative sequelae on the other hand. Anatomically more precise surgery is likely to improve the therapeutic index of this important operative treatment in gynecologic oncology. We used two new means to investigate this subject, magnetic resonance imaging (MRI) and liposuction. In a combined cadaver-clinical study we demonstrated the following: Residual parametrial tissue not resected with conventional hysterectomy techniques is a source for re971
972 Höckel, Konerding, and Heußel
May 1998 Am J Obstet Gynecol
Fig. 1. T2-weighted axial MRI scans of the female pelvis demonstrating (A) normal anatomy, (B) right infrailiac acetabular side wall recurrence of cervical cancer, and the postoperative situs (C) after conventional type III radical hysterectomy and (D) after nerve-sparing extended radical hysterectomy. C, Normal uterine cervix; T, tumor. Asterisks indicate perispinous adipose tissue; arrowheads point to anterolateral mesorectal plane. Circles highlight liquid signals demonstrating clearance of perispinous adipose tissue after nerve-sparing extended radical hysterectomy. In all panels the lettering locates the right ventral position.
currences; parametrectomy can be surgically defined with respect to the course of the pelvic plexus and to the anterolateral mesorectal planes; parametrectomy and systematic pelvic and periaortic lymph node dissection can be performed to spare most of the pelvic autonomic nerves, except the periureteral branches of the pelvic plexus. Material and methods MRI. Scans were performed on a Siemens Expert 1.0 T Magnetom (Siemens, Erlangen, Germany). A T2weighted turbo spin echo sequence was used in transverse and sagittal orientation. A T1-weighted spin echo sequence was used in transverse orientation and repeated after administration of contrast agent. Additionally a turbo spin echo sequence with fat suppression was used in coronal orientation. This sequence was optionally repeated in a sagittal orientation. The total time of data acquisition was about 25 minutes. Parametrial liposuction. Liposuction cannulas for parametrectomy were developed by Normed Medizintechnik Ltd. (Tuttlingen, Germany) according to specifications that emerged from the cadaver study. The cannulas are manufactured of high-grade stainless steel with one blunt hole at the inside of the slightly bent tip. A set of cannulas with lumen diameters of 3 and 4 mm and lengths of 17 and 20 cm provides the necessary versatility. The cannulas are connected to a high-performance
Fig. 2. The superior hypogastric plexus (SHP) and left and right hypogastric nerves (HN) isolated (A) in a female cadaver and (B) in a patient undergoing nerve-sparing extended radical hysterectomy, lateromedial view. LSN, Lumbar splanchnic nerves; U, left ureter; R, rectum; MR, left anterolateral mesorectal fascia.
vacuum pump producing 20 mm Hg residual pressure. Short strokes with the cannula in a piston-like manner are performed in anterior-posterior direction, with the hole directed to the mid pelvis beginning at the pelvic wall and moving medially. Cadaver study. Fresh female cadavers without evidence of pelvic operations or pathology were laparotomized. The retroperitoneum was entered by splitting the parietal peritoneum along the pelvic parts of the psoas muscles, paracolic gutters, and radix mesenterii. The visceral peritoneum was opened at the pouch of Douglas and in the uterovesical fold. The pararectal and paravesical spaces were fully opened. The rectum was dissected in the anterior and posterior mesorectal planes. The lymphatic tissue along the common, external, and internal iliac vessels and in the obturator fossa was stripped off. After complete removal of fat from the pararectal, parametrial, and paravesical tissues by liposuction, the topographic anatomy of the pelvic autonomic nerves was studied. Liposuction-assisted nerve-sparing extended radical hysterectomy. The possibility of nerve-sparing extended radical hysterectomy was offered to patients to be operated for cervical cancer (by M.H.). Patients who gave in-
Volume 178, Number 5 Am J Obstet Gynecol
Höckel, Konerding, and Heußel 973
formed consent underwent operation with these new surgical techniques. All surgical specimens, including the tissues obtained by liposuction, were fixed in formalin and routinely processed for histologic investigation. All patients underwent a pelvic MRI series in the second postoperative week. Results MRI studies. The sectional anatomy of the parametrial tissues was studied with pelvic MRI series of patients without gynecologic pathology. Sections 5 mm thick around the ischial spine clearly demonstrated a bilateral triangular zone of adipose tissue between the mesorectum (lower border), uterovaginal venous plexus (upper border), and obturator internus, coccygeus, iliococcygeus muscle (lateral border), designated perispinous adipose tissue (Fig. 1, A). Postoperative MRI series of patients treated with conventional (type III) radical abdominal hysterectomy by five experienced gynecologic surgeons revealed that the perispinous adipose tissue remained completely in situ (Fig. 1, C). MRI examinations were performed on 46 patients operated on in this institution from 1990 to 1996 with the laterally extended endopelvic resection–combined operative and radiotherapeutic treatment techniques for cervical cancer recurring at the pelvic side wall after radical hysterectomy and radiation. These scans demonstrated a topographic relationship of the recurrence to the nonresected perispinous adipose tissue in 21 cases (37%; Fig. 1, B), suggesting that residual parametrial tissue, in particular the perispinous adipose tissue, is a potential source of tumor recurrence after conventional radical hysterectomy in high-risk tumors. Liposuction-assisted cadaver dissection. The surgical anatomy of the female pelvic autonomic nerve system with respect to radical hysterectomy was investigated by liposuction-assisted pelvic dissection in five fresh human cadavers. The superior hypogastric plexus, a nerve band approximately 5 mm in width and between 4 and 5 cm in length, is formed at the level of the bifurcation of the aorta by two main lumbar splanchnic nerves arising from L2 and L3 positions laterally and medially to the aorta. The plexus descends at the surface of the left common iliac artery and vein to the level of the promontorium, where it divides into the two hypogastric nerves approximately 3 mm wide that run for about 10 cm at the left and right lateral mesorectum (Fig. 2). One to two centimeters below the level of the pouch of Douglas, each hypogastric nerve and three or four pelvic splanchnic nerves approaching laterally from the perispinous pelvic wall (side of attachment of coccygeus muscle to obturator internus muscle) form the pelvic plexus, a complex 3 to 4 × 2 to 3 cm flat network of nervous tissue. The pelvic plexuses on both sides are completely embedded in the perispinous adipose tissue (Fig. 3, A). After removal of the fat from
Fig. 3. Demonstration of left parametrium and perispinous adipose tissue in a cadaver (A) before and (B) after removal of fat with liposuction, craniocaudal view. U, Left ureter; UA, uterine artery; UV, uterine vein; HN, hypogastric nerve; PP, pelvic plexus; RUL, rectouterine ligament. Resection line for wide en bloc parametrectomy in nerve-sparing extended radical hysterectomy technique is indicated. Arrowhead points to pelvic splanchnic nerves.
the parametrial and paramesorectal tissue with liposuction, the pelvic plexus can be visualized (Fig. 3, B). A few medial branches of the pelvic plexus enter the mesorectum; most branches run forward laterally to the vagina and bladder. Most of these nerves are located below the uterine artery and vein at the level of the inferior vesical artery and vein. After giving off further medial branches to the cervix and vagina, the pelvic plexus radiates into the lateral bladder wall and trigone. Most of these branches run dorsally to the ureter and can be spared when the ureter is dissected out of the parametrial tissue (Fig. 4, A). On the basis of the topographic anatomic results of the cadaver study, the best compromise between maximum parametrectomy and sparing of the pelvic autonomic nerves appears to be the following: dissection of the rectouterine and rectovaginal ligaments exactly in the anterolateral mesorectal planes; separation of the uterine artery and veins immediately after branching from the parent vessels (hypogastric, vesicalis superior, or obturator vessels); transection of the pubocervical fascia (vesicouterine, vesicovaginal ligaments) at the projection
974 Höckel, Konerding, and Heußel
Fig. 4. A, Specimen of nerve-sparing extended radical hysterectomy. B, Postresection situs demonstrating right pelvis viewing mediolaterally. Most branches of the pelvic plexus to the lateral bladder wall and trigone (arrowheads) run dorsally of the ureter and are spared when dissecting the ureter out of the parametrial tissue and performing liposuction of the perispinous tissue. U, Ureter; V, vagina (open); MR, mesorectum; HN, hypogastric nerve; PP, pelvic plexus; PSN, pelvic splanchnic nerves. In addition to the en bloc specimen shown in A, between 5 and 10 ml of lymphatic adipose tissue is obtained by liposuction of the perispinous adipose tissue on each side.
of the parametrial course of the ureters; liposuction of the perispinous adipose tissue to expose the pelvic plexus; and lateral dissection of the cardinal ligament and the paracolpium with ventrolateral displacement of residual pelvic plexus after separation of the supraureteric branches during the dissection of the pubocervical ligament. Liposuction-assisted nerve-sparing extended radical hysterectomy. The operation starts with a hypogastric and epigastric midline laparotomy. The parietal peritoneum is incised at both paracolic gutters and along the radix mesenterii, the duodenum is mobilized from the aorta and vena cava, and the small and large bowel are packed into an intestinal bag. The pelvic parietal peritoneum is opened above the psoas muscles and the round ligaments are dissected. Depending on the age of the patient or on adnexal abnormality, the infundibulopelvic ligaments or ligamenta ovarii propria and proximal fallopian tubes are dissected to include or exclude the adnexa into the hysterectomy specimen.
May 1998 Am J Obstet Gynecol
The broad ligament of the uterus is incised anteriorly and posteriorly, and the paravesical and pararectal spaces are developed. Both uterine arteries are ligated and cut immediately after their branching from the hypogastric or obturator arteries. The pouch of Douglas is incised and the posterior vaginal wall is bluntly separated from the anterior rectum in the midline for between 4 and 5 cm. The visceral peritoneum in the vesicouterine fold is incised, and the bladder is completely detached from the midsagittal cervix and the proximal third of the anterior vaginal wall. At the aortic bifurcation the superior hypogastric plexus is identified and mobilized from the left common iliac artery and vein and from the promontorium. The presacral space is opened with electrocautery in the mesorectal plane to follow both hypogastric nerves without detaching them from the mesorectum. The rectouterine and rectovaginal ligaments are separated from the rectum by electrodissection exactly in the mesorectal plane. The posterior parametrial structures are clamped and ligated only toward the uterus to prevent lymphatic tumoral spread. No clamps are used toward the rectum, to allow the exact separation in the anterolateral mesorectal planes. Complete mobilization of the ureters out of the cardinal and pubocervical ligaments is facilitated by use of a scoop. Then the proper parametrial liposuction cannula (Normed Medizintechnik) is inserted into the parametrial fatty tissue at the parietal endopelvic fascia. Liposuction creates an approximately 1 cm fat-free space, skeletizing the distal uterine veins, which are then dissected between hemoclips immediately before branching from the hypogastric vein. Liposuction continues caudally at the level of the ischial spine along the iliococcygeus and coccygeus muscles to expose the middle hemorrhoidal vessels and the pelvic splanchnic nerves. These structures are spared. Multiple small (≤5 mm) lymph nodes are usually encountered during the maneuver. They can be easily harvested for histologic examination. The exposed pelvic splanchnic nerves and the hypogastric nerves delineate the pelvic plexus on both sides. The deep dissection of the cardinal ligament, pubocervical fascia, and paracolpium is performed with clamps above the level of the pelvic plexus, or with lateral displacement of the pelvic plexus if a more extensive vaginal resection is necessary. The vagina is incised at the oncologically adequate location between strong clamps placed both distally and proximally. Complete pelvic and periaortic lymphadenectomy follows, with the superior hypogastric plexus and the lumbar splanchnic nerves spared. After thorough washings, a peritoneal bridge between bladder and rectum is formed; this is used to support the middle pelvic parts of both ureters. The abdomen is closed with a running Smead-Jones suture. Clinical feasibility study. Nerve-sparing extended radi-
Höckel, Konerding, and Heußel 975
Volume 178, Number 5 Am J Obstet Gynecol
Table I. Clinical and histopathologic features of the first seven patients treated with liposuction-assisted nerve-sparing extended radical hysterectomy and parametrectomy
Patient
Age (yr)
Tumor type
FIGO stage
1 2* 3 4 5† 6 7
30 39 41 58 49 39 37
SC, cervix SC, cervix SC, vagina SC, cervix AC, cervix ASC, cervix SC, cervix
IIB IB IIB IIB IB IIB IB
pTpN
Lymph nodes removed
T2bN0 T1bN0 T1bN0 T2bN1 T1bN0 T2bN1 T1bN1
48 33 35 66 39 76 46
Total parametrial area resected (cm2) 25 18 85 48 35 50 46
Histopathologic Intraoperative tumor blood loss size (mm) (ml) 25 15 40 57 19 30 15
800 500 1000 1500 1000 600 800
Postoperative cystostomy (days) 8 14 21 12 7 30 11
FIGO, International Federation of Gynecology and Obstetrics; SC, squamous cell carcinoma; AC, adenocarcinoma; ASC, adenosquamous carcinoma. *After simple vaginal hysterectomy. †After conization.
cal hysterectomy was carried out in seven consecutive patients with cervical and vaginal cancer (Table I). No intraoperative complications occurred and total blood loss (including lymph node dissection) was relatively low, median 800 ml (range between 500 and 1500 ml). The removal of the perispinous adipose tissue was demonstrated by postoperative MRI in patients treated with the new technique (Fig. 1, D). The histopathologic workup of the surgical specimens showed a median parametrial en bloc resection area of 46 cm2 (range between 18 and 85 cm2; Fig. 4, B). The microscopic examination of the tissue obtained by liposuction after processing and staining found predominantly adipose tissue, blood cells, and lymphatic tissue. In patient 6, who had an adenosquamous cancer stage T2bN1, one small lymph node obtained by liposuction of the perispinous adipose tissue contained a metastasis. The postoperative courses of all patients were uneventful, without complications or deaths. The patients recovered within 3 days. Suprapubic cystostomies could be withdrawn after a median 12 days (range 8 to 30 days), when residual urine was constantly <50 ml. Comment According to the concepts of surgical oncology, radical treatment involves the wide en bloc resection of the tumor-bearing organ along with its primary lymphatic drainage compartment. The relevance of this principle has been recently reappraised with total mesorectal excision techniques for the surgical treatment of rectal cancer.9, 10 For the operative therapy of cervical cancer, radical hysterectomy adheres to the same oncologic principle, resection of the uterus en bloc with the parametria. However, liberation of the ureters and the pelvic plexus to sustain their function interferes with the continuity of the parametrial resection. Thus the extent of parametrectomy is a crucial issue in radical hysterectomy. Despite its importance, the surgical anatomy of parame-
trectomy has only been defined with respect to the uterine vessels and the ureter in European and American schools of gynecologic oncology.11 Only vague descriptions concerning the extend of parametrial resection with respect to the pelvic plexus are given in textbooks and surgical atlases. Moreover, parametrectomy has not to our knowledge been addressed with reference to anterolateral planes of the mesorectum. MRI clearly displays a distinct region of perispinous adipose tissue bordered mediocaudally by the anterolateral mesorectum, bordered laterally by the coccygeus, iliococcygeus, and obturator internus muscles, and bordered medioventrally by the vesicouterovaginal venous plexus. The MRI series from patients referred to us for the laterally extended endopelvic resection–combined operative and radiotherapeutic treatment12, 13 of pelvic sidewall recurrences of cervical cancer showed an involvement of the perispinous adipose tissue in 37% of cases. In the first series of seven patients treated with the nerve-sparing extended radical hysterectomy technique, we removed a metastatic lymph node 3 mm in diameter from the perispinous fatty tissue. Both observations point to the importance of the perispinous adipose tissue in tumoral spread in cervical cancer. However, perispinous adipose tissue is not resected in conventional type III radical hysterectomy procedures, as demonstrated by postoperative MRI scans. Perispinous adipose tissue was identified as the site of pelvic plexus in five female cadavers and corresponds to the pars nervosa of the cardinal ligament and the so called “lateral ligament of the rectum.” Recent anatomic studies revealed that the lateral ligament of the rectum is not a real anatomic ligament but rather a surgical construction from former rectal resection techniques as a consequence of the fusion of the pelvic plexus to the lateral mesorectum.10, 14 Likewise, we found no anatomic correlation to an “uterosacral ligament” between the sacrum and the uterus. The functional elongation of the
976 Höckel, Konerding, and Heußel
rectouterine condensation of the broad ligament from the rectum to the sacrum appears to be due to the pelvic plexus and pelvic splanchnic nerves adhering to the mesorectum. On the basis of the anatomic results of the study of the fat-free parametrial tissues in cadavers, we developed a new surgical technique for radical hysterectomy intended to compromise between maximum parametrial radicalness and sustained pelvic autonomic nerve function. The topographic anatomy of the pelvic plexus does not allow its complete preservation in radical hysterectomy because the detachment of the ureter out of the parametrial vesicouterovaginal venous plexus necessitates the disruption of its branches running to the bladder ventral to the ureter. However, (1) en bloc resection of the parametrial tissue down to the hypogastric nerves as landmarks for the dorsolateral resection line, (2) dissection of the rectouterine and rectovaginal ligaments in the anterolateral mesorectal planes, and (3) liposuction of the residual perispinous adipose tissue remove all parametrial lymphatic tissue and retain the pelvic splanchnic nerves and the pelvic plexus with its genitovesical branches dorsal to the ureteric level. Inherent in the new radical hysterectomy technique are the identification and sparing of the superior hypogastric plexus during upper pelvic and periaortic lymphadenectomy. Nerve-sparing radical hysterectomy techniques have also been developed in China and Japan. Their superiority to conventional techniques has been claimed; however, only a few publications are available in English.15-18 A detailed study of the surgical anatomy of the parametrial tissue and an anatomically defined nerve-sparing resection technique have been published by Yabuki et al.19, 20 These authors used an ultrasonic surgical aspirator for the removal of the parametrial alveolar tissue. We and others have reservations about the use of ultrasonically powered tissue removal techniques in oncologic situations, in which intraperitoneal tumor cell dissemination could be harmful.21, 22 Our preliminary results show that the new surgical techniques of liposuction-assisted extended radical hysterectomy are feasible. The histopathologic workup and the postoperative MRI prove there is a larger extent of removed parametrial tissue than is obtained with conventional type III radical hysterectomy techniques. However, postoperative bladder function, as preliminarily evaluated by measuring residual urine, is not more impaired. This suggests a gain in therapeutic index. We have initiated a prospective controlled study investigating bladder, vaginal, and rectal function after liposuction-assisted nerve-sparing extended radical hysterectomy by means of longitudinal questionnaires and urodynamic measurements. The long-term functional results and the oncologic outcome of patients with cervical
May 1998 Am J Obstet Gynecol
cancer, in particular, those with features carrying a high risk of local recurrence after surgical treatment, are awaited.
REFERENCES
1. Wertheim E. The extended abdominal operation for carcinoma uteri (based on 500 operative cases). Am J Obstet Dis Women Child 1912;66:169-232. 2. Okabayashi H. Radical abdominal hysterectomy for cancer of the cervix uteri. Surg Gynecol Obstet 1921;33:335-41. 3. Meigs JV. Carcinoma of the cervix: the Wertheim operation. Surg Gynecol Obstet 1944;78:195-8. 4. Burghardt E, Baltzer J, Tulusan AH, Haas J. Results of surgical treatment of 1028 cervical cancers studied with volumetry. Cancer 1992;70:648-55. 5. Friedberg V. Operative therapy for stage IIb cervical cancer. Baillieres Clin Obstet Gynaecol 1988;2:973-80. 6. Nezhat CR, Burrell MO, Nezhat FR, Benigno BB, Welander CE. Laparoscopic radical hysterectomy with paraaortic and pelvic node dissection. Am J Obstet Gynecol 1992;166:864-5. 7. Dargent D, Mathevet P. Schauta’s vaginal hysterectomy combined with laparoscopic lymphadenectomy. Baillieres Clin Obstet Gynaecol 1995;9:691-705. 8. Dargent D, Brun JL, Roy M, Rémy I. Pregnancies following radical trachelectomy for invasive cervical cancer [abstract]. Gynecol Oncol 1994;52:105. 9. Heald RJ, Husband EM, Ryall RDH. The mesorectum in rectal cancer surgery: the clue to pelvic recurrence? Br J Surg 1982;69:613-6. 10. Enker WE, Thaler HT, Cranor ML, Polyak T. Total mesorectal excision in the operative treatment of rectal cancer. J Am Coll Surg 1995;181:335-46. 11. Piver MS, Rutledge F, Smith JP. Five classes of extended hysterectomy for women with cervical cancer. Obstet Gynecol 1974;44:265-72. 12. Höckel M, Baußmann E, Mitze M, Knapstein PG. Are pelvic side-wall recurrences of cervical cancer biologically different from central relapses? Cancer 1994;74:648-55. 13. Höckel M, Schlenger K, Hamm H, Knapstein PG, Hohenfellner R, Rösler HP. Five-year experience with combined operative and radiotherapeutic treatment of recurrent gynecologic tumors infiltrating the pelvic wall. Cancer 1996;77:1918-33. 14. Sato K, Sato T. The vascular and neuronal composition of the lateral ligament of the rectum and the rectosacral fascia. Surg Radiol Anat 1991;13:17-22. 15. Sasaki H, Yoshida T, Noda K, Yachiku S, Minami K, Kaneko S. Urethral pressure profiles following radical hysterectomy. Obstet Gynecol 1982;59:101-4. 16. Wu YX. Preservation of pelvic nerve plexus during radical hysterectomy. Chung Hua Fu Chan Ko Tsa Chih 1991;26:368-71. 17. Tong XK, Huo RJ. The anatomical basis and prevention of neurogenic voiding dysfunction following radical hysterectomy. Surg Radiol Anat 1991;13:145-8. 18. Wu YX, Qui S, Meng J. Careful dissection of the pelvic plexus helps alleviate urinary retention following radical hysterectomy. Chung Hua Chung Liu Tsa Chih 1994;16:465-8. 19. Yabuki Y, Asamoto A, Hoshiba T, Nishimoto H, Kitamura S. Dissection of the cardinal ligament in radical hysterectomy for cervical cancer with emphasis on the lateral ligament. Am J Obstet Gynecol 1991;164:7-14. 20. Yabuki Y, Asamoto A, Hoshiba T, Nishimoto H, Satou N. A new proposal for radical hysterectomy. Gynecol Oncol 1996;62:3708. 21. van Dam PA, Coppens M, van Oosterom AT, Van Marck E, Buytaert P. Is there an increased risk for tumor dissemination using ultrasonic surgical aspiration in patients with vulvar carcinoma? Eur J Obstet Gynecol Reprod Biol 1994;55:145-7. 22. Nahhas WA. Potential hazard of use of surgical ultrasonic aspirator in tumor reductive surgery. Gynecol Oncol 1991;40:81-3.