Sacral chordoma is a rare, low to intermediate-grade tumor that is often diagnosed late because of the insidious onset of symptoms and a location deep within the pelvic cavity1-5. For these reasons, many sacral tumors are quite large at the time of diagnosis and may involve adjacent neurovascular structures and vital organs. The large size, the proximity to neurovascular structures, and the complex anatomy of the true pelvis make complete or wide excision difficult. Sacrectomy has been reported as a form of surgical treatment for malignant and benign primary tumors and metastatic tumors from distant sites1,4-18. Surgical approaches for sacrectomy include a combined sequential anterior and posterior approach, a combined simultaneous anterior and posterior approach, a posterior approach alone, and combined extended ilioinguinal and posterior approaches9-12,15,17. Few reports in the literature have analyzed the functional and oncologic results in patients following sacrectomy for a chordoma. A recent study that evaluated the operative management of sacral chordoma concluded that negative surgical margins are the most important predictor of survival and local recurrence19. In the present study, we evaluated the oncologic and functional outcomes and complications of sacrectomy in a series of patients treated for sacral chordoma at a single institution, by a single surgeon, using combined sequential anterior and posterior approaches.
Materials and Methods
We reviewed our institutional database for all patients who underwent sacrectomy, performed by the senior author (F.J.E.), for sacral chordoma from May 1985 to December 2001. We identified sixteen patients and retrospectively analyzed the clinical records and pathology slides for each patient. Patients undergoing sacrectomy for other types of tumors were excluded from the present study. Variables analyzed included the symptom onset, tumor size, treatment, presence or absence of local recurrence and metastases, survival, length of hospital stay, use of adjuvant radiation therapy, use of instrumentation, allograft use, creation of a colostomy compared with no colostomy, and complications associated with surgery. The minimum duration of follow-up was two years or until the time of death. The tumor size was recorded as the maximum tumor dimension measured on the gross specimen or on the preoperative magnetic resonance images. The tumors were classified according to the Musculoskeletal Tumor Society staging system20. Table I summarizes the demographic and pertinent clinical information for each patient. The functional and oncologic outcomes were assessed clinically with use of clinical examinations and imaging studies at six-month intervals. Every patient had a computed tomography scan or magnetic resonance imaging evaluation, or both, at each six-month examination. Data were recorded in a Microsoft Excel 2002 spreadsheet (Redmond, Washington) and were analyzed with use of SPSS statistical software (SPSS, Chicago, Illinois). Nonparametric and parametric variables were analyzed with use of the Kruskal-Wallis and analysis-of-variance tests, respectively. Significance was assigned to p values of <0.05.
Preoperatively, all patients were examined clinically and evaluated by the senior treating surgeon. A magnetic resonance imaging scan of the pelvis from the L4 level to the anus was acquired, with and without gadolinium contrast, for all patients. If more detail was needed, then computed tomography scans and/or magnetic resonance imaging scans specifically of the sacrum were made. All patients were also evaluated with a total-body bone scan and a computed tomography scan of the chest to ensure that no distant metastatic disease was present. In our series, all patients had involvement of levels cephalad to or including at least a portion of S2; therefore, a combined sequential anterior and posterior approach was used for every patient. When it was possible, ligation of the internal iliac vessels was avoided because it always reduces circulation to the gluteal muscles through the superior gluteal artery and vein. Loss of this circulation often compromises planned rotational gluteal flaps, which aid wound closure21. In patients who would certainly lose bowel control, a colostomy was performed by the general surgery team. No tumor excision was ever performed through the anterior approach. Prior to closure, a laparotomy sponge was placed over the anterior aspect of the sacrum and the freed end of the omental flap was sutured to the laparotomy sponge so that, when the sponge was encountered posteriorly, the flap could be easily located. The anterior abdominal incision was then closed.
The patient was then placed in the prone position. The incision was made from the spinous process of L5 to the tip of the coccyx. With reference to the preoperative imaging studies, a dissection was started through the subcutaneous tissue and the fascia, staying at least 2 cm away from the tumor margins posteriorly and laterally. The neural elements were exposed at least 2 cm cephalad to the intraspinal portion of the tumor. In most patients, this involved a laminectomy of L5 and, in some patients, a partial laminectomy of L4. Any nerve that traverses the tumor was excised. If a nerve was definitely not involved with the tumor, then it was preserved. Decisions were made clinically for the nerves that were adjacent to the tumor but were not definitively encased in the tumor. At least one completely free nerve root above the tumor was well exposed for electrophysiologic testing purposes. The dural sac was isolated just caudad to the origin of the most caudad nerve root to be spared. A number-0 silk suture was placed around the dural sac, tied securely, and was then oversewn with a 3-0 nylon running suture. A Valsalva maneuver was performed to verify that there was no spinal fluid leak. Osteotomies were then made through the sacrum with use of osteotomes. The cuts were planned at least 2 cm from the tumor in the cephalad and lateral directions. The lateral dissection was carried out to expose the sciatic notch. If the sacroiliac joint was involved, then the ilium on that side was cut in an oblique fashion toward the sciatic notch. The sciatic nerve was identified, and care was taken to preserve it when possible. The sacrotuberous and sacrospinous ligaments were identified and cut caudally. Anterior dissection was then begun over the coccyx. The levator muscles were then cut to expose the retroperitoneal area. Dissection was carried out bluntly on the anterior aspect of the sacrum up to the anterior aspect of the tumor. Once the tumor had been completely freed anteriorly, osteotomes were used to make cuts under fluoroscopic control approximately 2 cm cephalad to the tumor in the sacrum. These cuts were extended laterally, providing a 2-cm margin from the tumor. At the point that the osteotomies were made, the surgical team confirmed that adequate blood products were available in the operating room for rapid administration because a large proportion of the blood loss usually occurred during the final ten to fifteen minutes of the tumor resection.
After the tumor had been resected, all osseous and soft-tissue margins of both the tumor specimen and the patient were inspected. Any questionable areas were biopsied again, and wider margins were obtained if necessary. The plastic surgery team then placed the omentum from the omental flap at the base of the wound to cover the bowel, and a surgical mesh was placed over the floor of the perineum to prevent herniation of the abdominal contents. Gluteal flaps were then rotated medially to fill the large, deep soft-tissue defect, which was always present. Adequate drains were placed to protect the flaps from hematoma development. The skin edges were closed with nylon sutures, which were left in place for at least three weeks. The wound was covered with a dry sterile dressing. The duration of surgery ranged from eight to twenty-four hours. When it was necessary, reconstruction included fixation to at least two or three levels of the remaining lumbar spine and instrumentation into the pelvis.
Postoperatively, the patients were managed on an antidecubitus air mattress to protect the skin. They were initially mobilized within the first five postoperative days and were allowed to walk freely but were discouraged from sitting while the incision was healing. The drains were kept in place until the daily output was <60 mL. Many patients were placed on hyperalimentation, if they could not take food by mouth during the first two postoperative days. These patients often required additional treatment in a rehabilitation hospital in order to become familiar with bowel and bladder care and to resume walking in a safe fashion. In the first month after surgery, baseline pelvic computed tomography and magnetic resonance imaging scans at the previously described levels were acquired, with and without contrast, to serve as a comparison with later surveillance scans acquired at subsequent six-month intervals.
Eight male patients and eight female patients underwent sacrectomy for sacral chordoma. The mean age at the time of diagnosis was 61.2 years (range, sixteen to eighty years). Presenting symptoms included pain in all patients, urinary dysfunction in six, bowel dysfunction in three, a prominent mass in four, paresthesias in four, and sexual dysfunction in one. The mean tumor size was 15.2 cm (range, 9.5 to 30 cm) in the greatest diameter. At the time of presentation, fifteen lesions were stage I-B and one was stage III-B, according to the system of the Musculoskeletal Tumor Society. The patient with a stage III-B lesion was sixteen years old and underwent surgery because of severe skin compromise locally following a poorly planned intralesional procedure performed elsewhere and because of severe local pain. The mean duration of follow-up was sixty-six months (range, fifteen to 174 months). Seven patients underwent marginal resections, and three patients were found to have a violation of the tumor margins on intraoperative histologic examination. The three patients had repeat excision of the margins, and negative margins were ultimately obtained during the same surgical procedure. Nine patients underwent wide local excision. The mean operative blood loss was 5 L (range, 1500 to 8000 mL).
The mean length of hospital stay was thirty days (range, seventeen to fifty-one days). Patients with more cephalad levels of resection (cephalad to S1) had longer hospital stays (p = 0.022). One patient had received chemotherapy in the distant past for an unrelated diagnosis of lymphoma. Two patients received postoperative chemotherapy. One of them was the sixteen-year-old girl with a stage-III-B lesion, and she was placed on a soft-tissue sarcoma chemotherapy protocol by our medical oncology team because of the aggressive nature of the metastatic chordoma. The other patient to receive postoperative chemotherapy was started on thalidomide by the oncologists after several resections of metastatic lesions and local recurrences. Seven patients were treated with adjuvant radiation therapy postoperatively. The average dose of radiation was 5900 cGy. One patient did not initially receive adjuvant radiation therapy but was treated with radiation therapy for a local recurrence.
The tumor recurred locally in twelve of the sixteen patients. The mean time from the index surgical procedure to the first recurrence was twenty-nine months (range, twelve to sixty-six months). The mean time to the second recurrence was sixty-two months (range, thirty to 125 months) in five patients, and the mean time to the third recurrence was 71.5 months (range, thirty to 140 months) in four patients. At the conclusion of our study, six patients had clinical and/or radiographic evidence of metastases. The mean time to metastasis from the index procedure was fifty months (range, sixteen to 112 months). Four patients were alive with no evidence of disease at a mean follow-up period of 94.5 months. Six patients were alive with disease at a mean follow-up period of fifty-five months. Five patients had died as a result of progressive local or metastatic disease at an average of 31.4 months (range, fifteen to forty months). One patient died in the perioperative period because of a myocardial infarction two weeks following resection of a recurrence fourteen and one-half years after the index procedure (Figs. 1-A, 1-B, 1-C, 1-D, 1-E). With the numbers available, negative surgical margins did not have a significant effect on survival. Negative margins were associated with decreased rates of local recurrence or metastases (Table II). All patients in whom contaminated margins were obtained during surgery had local recurrence or metastases. Three of the four patients in whom an inadequate margin (a marginal excision) had been obtained had a local recurrence, and two of the four patients had metastases. Six of the nine patients in whom a wide local excision had been achieved had local recurrence and two of them had metastases. With the numbers studied, the use of adjuvant radiation therapy was not associated with a significant increase in survival nor was it associated with decreased local recurrence. Patients with a larger tumor had a significant increase in the number of perioperative complications (p = 0.034).
Three patients were able to walk without assistive devices, four patients used a single cane, eight patients used a walker, and one patient relied on a standard wheelchair. With the numbers studied, neither the tumor size nor the level of resection was found to have a significant impact on the ability to walk.
One patient was continent of urine postoperatively, and fifteen were dependent on an indwelling urinary catheter, intermittent straight catheterizations, or the Crede maneuver in order to void (see Appendix). More cephalad levels of resection (cephalad to S2) were associated with poorer bladder control (p = 0.01). One patient retained normal bowel control. Seven patients used a colostomy. Eight patients used a bowel program or digital stimulation to defecate (see Appendix). More cephalad levels of resection (cephalad to S2) were associated with poorer bowel control (p = 0.01).
Eight patients had wound complications, ranging from superficial infections that did not require operative intervention to deep wound infections, seromas, and major wound dehiscences that required one or more additional operative procedures. There were four documented deep wound infections, five wound dehiscences, and three seromas. A deep-vein thrombosis was documented in one patient. No patient had a clinically important pulmonary embolism. One patient died secondary to a perioperative acute myocardial infarction after undergoing anterior resection of a recurrent sacral chordoma fourteen and one-half years after the index procedure. With the numbers studied, the only factor that was found to be significantly associated with increased postoperative complications was a larger tumor size (p = 0.034).
Sacrectomy has been described for the treatment of sacral tumors, including chordoma, osteosarcoma, chondrosarcoma, teratoma, recurrent anorectal carcinoma, giant-cell tumor, and giant sacral schwannoma1,5,7,9,11-13,18,22. These tumors present a therapeutic challenge because most malignant and benign aggressive tumors require wide local excision for favorable results; however, ensuring adequate margins may come at the expense of normal adjacent neurovascular structures and associated function. Gunterberg et al. contributed a series of articles evaluating urologic, anorectal, and sexual function following nerve root sacrifice following sacrectomy23-25. Todd et al., in a study on bowel and bladder function in fifty-three patients who had a major sacral resection, showed that unilateral resection of the sacral roots or preservation of at least one S3 root preserves bowel and bladder function in most patients3. In our series, one patient retained normal bowel and bladder control with intact S2 nerve roots bilaterally. One additional patient retained a single S3 nerve root but did not exhibit continence as may have been predicted on the basis of other published reports23,24. There appears to be some variability in clinically observed neurophysiologic control; therefore, we caution other surgeons to advise patients that retention of a single S3 nerve root improves the probability, but does not guarantee, that urinary and rectal continence will be preserved.
Wuisman et al. presented oncologic and functional outcome data on nine patients treated with sacrectomy over a ten-year period5 and noted that three patients died of the disease and six were alive and disease-free following sacrectomy. These results were similar to those in the present study in terms of mortality (six of sixteen patients had died at an average follow-up period of five and one-half years), although there was a higher rate of local recurrence in the present study. This is likely due to the fact that all tumors except chordomas were excluded from the present study. In a more recent study, Fuchs et al.19 reported a rate of local recurrence of 44% in fifty-two patients who had been followed for an average of 7.8 years after operative treatment of a sacral chordoma. Those authors reported that inadequate margins had been obtained at the time of surgery in 60% of the patients, and the mortality rate was 26%, 48%, and 53% at five, ten, and fifteen years, respectively. Local recurrence in twelve of our sixteen patients with a chordoma is similar to that observed by Berven et al., who reported local recurrence in eighteen (78%) of twenty-three patients with a chordoma26. We identified several important differences between the patients described by Fuchs et al.19 and our own series of patients. These differences include patient age, tumor size, and the cephalad extent of the tumors. The mean age of the patients in our series was 61.2 years compared with fifty-six years in the series of Fuchs et al.19. The mean tumor size in their series was 9 cm, with two tumors that were >15 cm in diameter. This contrasts sharply with the very large tumors in our series, which had a mean size of 15.2 cm, with a diameter of 30 cm in the largest tumor and 9.5 cm in the smallest tumor, which was larger than the mean size in their study. Last and perhaps most importantly, the cephalad extent of the tumors in the fifty-two patients described in their study reached the S2 level in less than two-thirds of the patients, whereas all patients in the present series had tumor involvement cephalad to or involving the S2 vertebral level. These key differences may explain why the rate of local recurrence in the present series was higher, despite the fact that the proportion of the patients who had so-called adequate margins (nine of sixteen patients) was greater than that in the study by Fuchs et al. (40%). Twelve of our sixteen patients had a local recurrence despite surgical attempts to achieve adequate margins. Many tumors recurred more than once. Six of seven patients who had two or more recurrences eventually had metastatic disease develop. One patient had tumor recurrences at sites cephalad and caudad to the resection and at noncontiguous intrathecal sites. Metastases occurred relatively late in most patients, with a mean time to diagnosis of metastasis of 49.8 months. Our results underscore the difficulty of obtaining adequate margins, especially in very large chordomas and those that are more cephalad; the high rate of local recurrence of chordoma despite negative margins; and the propensity of sacral chordoma to metastasize late in the course of the disease.
The functional outcome measures in our study were limited to those of bowel and bladder control and walking status. One patient with preserved bilateral S2 nerve roots retained bowel and bladder control. Although three patients were able to walk without the use of assistive devices, the remaining thirteen had to use a cane or a walker or relied on a wheelchair. The ability to walk was also affected in some patients because of local recurrences, such as those affecting the sciatic nerve, rather than the nerve root sacrifice at the time of the index procedure.
The complications in our patients were similar, with respect to type and prevalence, to those reported by other investigators. In the study by Wuisman et al., the prevalence of deep-vein thrombosis was 11% and the prevalence of wound complications was 45.5%5. Fuchs et al. reported that the most common complication in their series was local wound break-down and added that the majority of such complications occurred before the transverse rectus abdominis musculocutaneous flap became a routine portion of their surgical protocol. The wound complications in our patients were treated with aggressive débridement and antibiotic therapy. Because of the small number of patients studied, the routine use of omental flaps, a more recent part of the surgical protocol at our institution, did not prove to have a significant effect on the number of complications in the present series. We believe, however, that the use of an omental flap is worthwhile since it limits the occurrence of complications such as radiation enteritis. The use of rotational flaps for coverage is a standard part of our sacrectomy protocol and will continue to be used in an effort to avoid wound complications in the future.
The treatment of sacral chordoma is an arduous clinical undertaking that requires a multidisciplinary approach and attention to detail from the outset. Despite aggressive, well-planned surgical management and adherence to strict surveillance protocols, frequent recurrence and late onset of metastatic disease are to be expected in a substantial proportion of patients treated. Adequate surgical treatment results in substantial functional impairment and numerous complications; however, it does offer the possibility of long-term disease-free survival. We advocate an attempt at complete resection, when there is still a possibility of cure, and aggressive treatment of local recurrences.
Tables showing urologic and bowel functional outcomes of all patients are available with the electronic versions of this article, on our web site at jbjs.org (go to the article citation and click on “Supplementary Material”) and on our quarterly CD-ROM (call our subscription department, at 781-449-9780, to order the CD-ROM). ▪
In support of their research for or preparation of this manuscript, one or more of the authors received grants or outside funding from Miami Center for Orthopaedic Research and Education (Miami CORE). None of the authors received payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity. No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, educational institution, or other charitable or nonprofit organization with which the authors are affiliated or associated.
Investigation performed at the Department of Orthopaedics and Rehabilitation, University of Miami School of Medicine, Miami, Florida
1. . Resection of the sacrum for benign giant cell tumor; a case report. Ann Surg. 1953;138: 115-20.
2. . Chordoma. J Bone Joint Surg Am. 1981;63: 501-5.
3. , Yaszemski MJ, Currier BL, Fuchs B, Kim CW, Sim FH. Bowel and bladder function after major sacral resection. Clin Orthop Relat Res. 2002;397: 36-9.
4. , Tsuchiya H. Total sacrectomy and reconstruction for huge sacral tumors. Spine. 1990;15: 1223-7.
5. , Lieshout O, Sugihara S, van Dijk M. Total sacrectomy and reconstruction: oncologic and functional outcome. Clin Orthop Relat Res. 2000;381: 192-203.
6. , Byrne PO, Robertson ID, Gullan RW, Montgomery AC. Radical excision of sacrococcygeal tumours. Brit J Surg. 1994;81: 460-1.
7. . Giant cell tumor of the sacrum: a case report. Ann Surg. 1948;128: 1164-72.
8. , Batsakis JG, Owano LR. Unusual manifestations of chordoma. A report of two cases. J Bone Joint Surg Am. 1968;50: 1618-28.
9. , Azzarelli A, Quagliuolo V. A posterior approach for the excision of sacral chordoma. J Bone Joint Surg Am. 1987;69: 565-8.
10. , Dawson EG, Eilber FR. Abdominosacral resection for malignant tumors of the sacrum. Am J Surg. 1984;148: 157-61.
11. , Francis KC, Rossano PG. Abdominosacral resection of sacrococcygeal chordoma. Ann Surg. 1967;166: 394-402.
12. , Springfield DS, Suit HD, Mankin HJ. Operative treatment of sacrococcygeal chordoma. A review of twenty-one cases. J Bone Joint Surg Am. 1993;75: 1476-84.
13. , Mitsunaga MM, Lockett JL. Total sacrectomy for a giant sacral schwannoma. A case report. Clin Orthop Relat Res. 1993;294: 285-9.
14. , Yamamuro T, Kotoura Y, Mikawa Y, Iida H, Maetani S. Total sacrectomy and reconstruction for primary tumors. Report of two cases. J Bone Joint Surg Am. 1988;70: 122-5.
15. , Porter A, Davis A, Griffin A, McLeod RS, Bell RS. Cephalad sacral resection with a combined extended ilioinguinal and posterior approach. J Bone Joint Surg Am. 1995;77: 405-11.
16. , Richardson WJ, Scully SP, Harrelson JM. Long-term survival following total sacrectomy with reconstruction for the treatment of primary osteosarcoma of the sacrum. A case report. J Bone Joint Surg Am. 1999;81: 848-55.
17. , Gunterberg B. High amputation of the sacrum for extirpation of tumors. Principles and technique. Spine. 1978;3: 351-66. Erratum in: Rev Chir Orthop. 1987;73:following 217.
18. , Nelson H, Gunderson LL, Sim FH. Sacropelvic resection for recurrent anorectal cancer. A multidisciplinary approach. Clin Orthop Relat Res. 2000;372: 231-40.
19. , Dickey ID, Yaszemski MJ, Inwards CY, Sim FH. Operative management of sacral chordoma. J Bone Joint Surg Am. 2005;87: 2211-6.
20. . A system of staging musculoskeletal neoplasms. Clin Orthop Relat Res. 1986;204: 9-24.
21. , McDonald WS, Armstrong M, Eismont F, Hellinger M, Thaller S. Reconstruction after extirpation of sacral malignancies. Ann Plast Surg. 2003;51: 126-9.
22. , Sim FH, Unni KK. Giant cell tumor of the sacrum. Clin Orthop Relat Res. 1993;291: 215-21.
23. , Norlen L, Stener B, Sundin T. Neurourologic evaluation after resection of the sacrum. Invest Urol. 1975;13: 183-8.
24. , Kewenter J, Petersen I, Stener B. Anorectal function after major resections of the sacrum with bilateral or unilateral sacrifice of sacral nerves. Br J Surg. 1976;63: 546-54.
25. , Petersen I. Sexual function after major resections of the sacrum with bilateral or unilateral sacrifice of sacral nerves. Fertil Steril. 1976;27: 1146-53.
26. , Zurakowski D, Mankin HJ, Gebhardt MC, Springfield DS, Hornicek FJ. Clinical outcome in chordoma: utility of flow cytometry in DNA determination. Spine. 2002;27: 374-9.