- Email: [email protected]
Surgical strategy for the management of sacral giant cell tumors: a 32-case series
The Spine Journal 12 (2012) 484–491
Clinical Study
Surgical strategy for the management of sacral giant cell tumors: a 32-case series Guodong Li, MD, Dong Fu, MD, Kai Chen, MD, Xiaojun Ma, MD, Mengxiong Sun, MD, Wei Sun, MD, Jian Li, MD, Zhendong Cai, MD* Department of Orthopaedics, Shanghai Tenth People’s Hospital, Tongji University School of Medicine, 301 Yanchang Rd, Shanghai 200072, China Received 14 December 2011; revised 28 April 2012; accepted 21 June 2012
Abstract
BACKGROUND CONTEXT: Surgical treatment of sacral giant cell tumors (GCTs) is associated with a high rate of complications, and there is controversy over which type of surgical treatment is optimal. PURPOSE: To develop an optimal treatment strategy for sacral GCTs. STUDY DESIGN/SETTING: Retrospective/academic medical center. PATIENT SAMPLE: A total of 32 patients (18 women and 14 men) with sacral GCT who underwent surgery between August 1996 and August 2008. OUTCOME MEASURES: Local recurrence rate, surgical margins, blood loss, sacral nerve root preservation, and complications. METHODS: The medical charts of 32 patients were reviewed. RESULTS: Patients underwent either wide resection (n52), marginal resection (n511), marginal resection plus curettage (n512), or curettage alone (n57). The curettage group and the wide resection group had the highest and lowest amounts of blood loss (4,500 vs. 1,300 mL, respectively). During follow-up (median, 42 months), 12 patients (37.5%) had local recurrence, including five of seven in the curettage group. The recurrence rate was significantly lower in the marginal excision group compared with that in the curettage group (18.2% vs. 71.4%, respectively; p5.049). Five patients had bladder dysfunction, and two patients had bowel dysfunction. Four patients who underwent marginal resection had lower limb dysfunction. Overall survival was 93.6%, and 2-year recurrence-free survival was 84.4%. CONCLUSIONS: Choosing an optimal surgical margin in the treatment of sacral GCTs is of great importance for local recurrence control and sacral nerve root preservation. Curettage alone should not be used to treat sacral GCT. Ó 2012 Elsevier Inc. All rights reserved.
Keywords:
Giant cell tumor; Sacrum; Surgical treatment
Introduction Giant cell tumors (GCTs) of bones are one of the most common neoplasms, accounting for 5% to 10% of all the primary bone tumors [1]. The sacrum is the third most common site for GCTs, following the peripheral area of the
FDA device/drug status: Not applicable. Author disclosures: GL: Nothing to disclose. DF: Nothing to disclose. KC: Nothing to disclose. XM: Nothing to disclose. MS: Nothing to disclose. WS: Nothing to disclose. JL: Nothing to disclose. ZC: Nothing to disclose. GL and DF are co-first authors. * Corresponding author. Department of Orthopaedics, Shanghai Tenth People’s Hospital, Tongji University School of Medicine, 301 Yanchang Rd, Shanghai 200072, China. Tel.: (86) 21-66307311; fax: (86) 21-56305592. E-mail address: [email protected] (Z. Cai) 1529-9430/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.spinee.2012.06.014
knee joint and the distal radius, and accounts for 3% to 4% of all GCTs [2]. Compared with GCTs at other sites, sacral GCTs have unique characteristics such as very complicated local anatomy, aggressive behavior, and nonspecific symptoms that can result in a delayed diagnosis. Although sacral GCTs are benign from the pathologic point of view, treatment is very difficult, optimal management is controversial, and there are no standard treatment guidelines [3]. Conventional treatments include radiotherapy, curettage combined with radiotherapy, marginal resection combined with curettage, marginal resection, extensive marginal resection, en bloc resection, and serial arterial embolization. The significant advantage of radiotherapy is the avoidance of surgical morbidity [3]. The main disadvantage is poor local control of the tumor and secondary
G. Li et al. / The Spine Journal 12 (2012) 484–491
sarcomatous changes [4,5]. Curettage combined with radiotherapy and marginal resection plus curettage can retain the sacral nerve root and maintain the completeness of the pelvic ring, but the rate of local recurrence is relatively high. Extensive marginal resection and en bloc resection may maximally decrease the recurrence rate but increase the incidence of surgical complications [6,7]. The purpose of this study was to retrospectively review the management of 32 cases of sacral GCTs who underwent surgical treatment at our medical center to develop a standard method for the treatment of sacral GCTs.
Materials and methods General data We retrospectively reviewed the medical records of 32 consecutive patients with sacral GCTs treated at our center between August 1996 and August 2008. There were 14 men and 18 women with a median age of 31 years (range, 19–74 years). All 32 patients had a diagnosis of primary GCT by pathologic examination of a tissue sample. Patients with recurrent GCTs, a postoperative pathologic diagnosis of malignant sacral GCT, and those lost to follow-up were excluded. This study was approved by the Institutional Review Board of our hospital, and because of the retrospective nature of the study, the requirement of informed consent was waived. Preoperative and surgical management All patients received anteroposterior radiographs, computed tomography (CT), magnetic resonance imaging (MRI), digital subtraction angiography, and emission computed tomography of the sacrum. Radiography, CT, and MRI were considered the main examination methods for three-dimensional evaluation of the tumors. Digital subtraction angiography was routinely performed for tumor angiography 24 to 72 hours before surgery. Embolization of the middle sacral artery, tumor feeding arteries, and selective bilateral internal iliac artery was carried out with Gelfoam and polyvinyl alcohol because of the extensive blood supply to the tumors. Beginning January 2008, the operation was performed 24 hours rather than 72 hours after digital subtraction angiography plus embolization. The surgical approach was based on the size and anatomic location of the tumor [8]. Two surgical approaches were used. The first was a single posterior approach using a posterior paramidline curved incision, posterior midline straight incision, or S3-plane transverse S-shaped incision. The posterior midline straight incision was suitable for high sacral tumors in the midline, the posterior paramidline curved incision was suitable for high eccentric tumors, and the S3-plane transverse S-shaped incision was suitable for low tumors. The sacrospinal muscles were dissected bilaterally, and the posterior wall of the sacral canal was opened. The tumor was removed while preserving at least
485
Context The best surgical strategy for the treatment of sacral giant cell tumors continues to be refined over time. The authors present their experience with this rare tumor. Contribution Based upon their chart review of 32 patients, the authors recommend avoiding curettage, as it incurs too high of a risk for local recurrence, and, conversely, overly aggressive margins (can put the sacral nerves at greater-thanneeded risk), which did not seem to impact outcomes. Implication This group’s recommendations are reasonable and, generally, in line with the opinions of the orthopedic oncology community. Denosumab and related therapies appear, at this time, as though they might be adjuvants that allow the authors’ suggested ‘‘middle road’’ to be more easily traveled. —The Editors unilateral S1–S3 nerve roots. In two patients with low sacral GCT, transverse S-shaped incisions were used. The origins of the bilateral gluteus maximus muscles were cut bilaterally to the superior margin of the greater sciatic foramen. The sacrotuberous, sacrospinous, and coccygeal ligaments were also cut. Osteotomy was performed at the upper part of the tumor, and the tumor was removed completely. The second approach was a combined anteroposterior approach. The anterior approach was performed through a midline incision in the lower abdomen with the patient in the lithotomy position for transabdominal exposure of the pelvic cavity. The integrity of the pseudomembrane of the tumor was preserved as much as possible to avoid contaminating the organs inside the pelvic cavity. The tumor was completely resected, and the portion of the sacrum invaded by the lesion was also removed. When the tumor was large and sacral destruction was significant, the tumor was partially resected along the anterior margin of the sacrum. A piece of large abdominal gauze was placed over the area of resection, and the retroperitoneum, peritoneum, and abdominal cavity were closed. Next, the patient was turned over and the posterior surgery was performed. In this series, marginal resection is defined as the complete removal of the tumor including soft tissues 1 to 3 cm outside the tumor capsule after sufficient exposure of the tumor and identification of the tumor capsule. All surgical margins were confirmed by intraoperative frozen section analysis. A negative margin was defined as no tumor cells found in the removed soft tissue around the tumor by pathologic examination. En bloc resection was carried out for tumors involving S3–S5 (Fig. 1, Left). Marginal resection was carried out for high eccentric tumors that did not cross
486
G. Li et al. / The Spine Journal 12 (2012) 484–491
Fig. 1. Posterior view of the sacrum. (Left) En bloc resection for sacral GCTs involving S3–S5. Green line between tumor and normal sacrum represents the surgical margin. (Middle) Marginal resection for high eccentric tumors that do not cross the midline. (Right) For high sacral GCTs, curettage or curettage together with marginal resection was carried out for patients with early-stage disease. Marginal resection was performed for S3 and its distal portion, and curettage was performed for the proximal portion of the tumor (S1 and S2). GCT, giant cell tumor.
the midline (Fig. 1, Middle) while totally preserving the contralateral sacral nerve roots. For high sacral tumors, curettage or curettage together with marginal resection was performed for patients with early-stage disease. Specifically, marginal resection was performed for S3 and its distal portion, and curettage was performed for the proximal portion of the tumor (S1 and S2) (Fig. 1, Right). After tumor resection, before 2008, the residual cavity and wound were treated with high-speed burr grinding combined with absolute ethanol plus distilled water for 28 patients. From 2008, four patients in total were treated using an argon beam coagulator plus I125 radioactive particles, which were placed evenly in the tumor cavity. The ISOLA pedicle screw system (DePuy, Raynham, MA, USA) was used for the reconstruction of the lumbosacral area when indicated in 28 of 32 patients, including unilateral reconstruction in seven patients (25.0%) and bilateral reconstruction in 21 patients (75.0%). A general treatment algorithm is shown in Fig. 2. Follow-up Follow-up evaluations were performed at 1, 2, 3, 6, and 12 months during the first year after surgery, every 3 months during the second year, and then every 6 months. During follow-up, radiographs and CT scans of the sacrococcygeal region as well as CT scans of the lungs were obtained. The patients were assessed for tumor recurrence (primary outcome), urination and defecation functions, lower limb sensation, myodynamia, and wound healing. Statistical analysis Categorical variables were compared using the Fisher exact test and blood loss using the Kruskal-Wallis test.
When a difference between groups was apparent, multiple comparisons of means were performed using the Bonferroni procedure with Type I error adjustment. Categorical data were represented by number (%), and continuous data were presented as median (interquartile range). The Kaplan-Meier method was used to measure overall survival (OS) and 2-year recurrence-free survival (RFS). All statistical assessments were two-sided and evaluated at the .05 level of significance. Statistical analyses were performed using SPSS 15.0 statistics software (SPSS, Inc., Chicago, IL, USA).
Results Patient demographic data and nerve root involvement are summarized in Table 1, and operative and follow-up data are presented in Table 2. Two patients underwent wide resection, and 11 underwent marginal resection with preservation of all the contralateral sacral nerve roots. Among the 19 patients with high sacral GCTs, seven underwent curettage and 12 underwent marginal resection plus curettage. The ISOLA pedicle screw system was used for the reconstruction of the lumbosacral area in 28 patients, including unilateral reconstruction in seven patients (25.0%) and bilateral reconstruction in 21 patients (75.0%). Five of 32 patients received radiation therapy at other hospitals before surgery. Local radiation therapy was performed after the operation for eight of the 12 patients who had recurrence. Perioperative adjuvant treatment with bisphosphonates (zoledronic acid, 4 mg/wk; two courses of treatment before surgery and four courses after surgery) was carried out for three patients. Local tumor recurrence occurred in 12 patients (37.5%) during follow-up. The local recurrence rates were 0%,
G. Li et al. / The Spine Journal 12 (2012) 484–491
487
Fig. 2. General treatment algorithm. GCT, giant cell tumor.
18.2%, 41.7%, and 71.4% for patients with wide resection (0/2), marginal resection (2/11), marginal resection plus curettage (5/12), and curettage (5/7), respectively. The marginal resection group had a significantly lower recurrence rate than the curettage group (18.2% vs. 71.4%; p5.049). There was a significant difference in blood loss among the four types of resection (p5.004). The greatest blood loss occurred in patients who underwent curettage (4,500 [3,900–4,700] mL) and the lowest occurred in patients who received wide resection (1,300 [550–2,050] mL). There was no significant difference in blood loss between the marginal resection and marginal plus curettage groups (3,000 vs. 2,950 mL; p5.557). The median duration of follow-up was 42 months (range, 18–115 months). Six patients had recurrence within 2 years after surgery, and altogether, 12 patients developed recurrence during the follow-up period. Among them, two patients died at the 18th and 23rd months after surgery, respectively, because of widespread pulmonary metastasis. Sarcomatous changes were observed on pathologic examinations after surgery in two cases. The OS and 2-year RFS curves are shown in Fig. 3. The OS rate was 93.6%, and 2-year RFS rate was
84.4%. In addition, benign pulmonary metastasis occurred during follow-up in one patient. During follow-up, five patients (5/32, 15.6%) developed urinary bladder dysfunction and two patients (2/32, 6.3%) developed bowel dysfunction requiring medication. Four patients with marginal resections had lower limb dysfunction (12.5%). Eleven patients (34.4%) had wound complications; five underwent local debridement, drainage, and secondary closure; and in four patients the wounds healed within 3 months with conservative management. In addition, two patients required skin flap transplantation after radiotherapy because of the large skin lesions. There were no significant differences in urinary bladder and bowel dysfunction and complications among the four types of resections (Table 2). Representative cases One A 31-year-old man complained of low back pain for 1 year, which was aggravated at night and not relieved by over-the-counter analgesics. He experienced sensory loss and numbness on the lateral side of the left foot, foot
488
G. Li et al. / The Spine Journal 12 (2012) 484–491
Table 1 Patient data (N532) Demographic data/nerve root involvement
Median (range)/n (%)
Age (y) Gender Male Female Nerve roots involved Left S1 S1–S2 S1–S3 S1–S4 S1–S5 S2–S3 S2–S4 S4–S5 S5 Right S1 S1–S2 S1–S3 S1–S4 S1–S5 S2–S4 S2–S5 S4–S5 S5
31 (19–74) 14 (43.8) 18 (56.3)
1 7 6 3 1 1 4 1 1
(3.1) (21.9) (18.8) (9.4) (3.1) (3.1) (12.5) (3.1) (3.1)
1 11 5 4 1 3 1 1 1
(3.1) (34.4) (15.6) (12.5) (3.1) (9.4) (3.1) (3.1) (3.1)
dorsal flexion was impossible, and a left-side straight leg test of 25 . Back view of the three-dimensional CT reconstruction showed left S1–S3 bone destruction, which involved the left sacroiliac (SI) joint and ilium. Computed tomography cross-section view shows that the tumor involved the left SI joint and left S1 sacral foramen but did not involve the midline. The patient was placed prone and posterior paramidline curved incision was made (Fig. 4). Marginal tumor resection was performed via the posterior approach,
and the contralateral S1–S5 nerve roots and ipsilateral S4–S5 nerve roots were retained. An argon beam coagulator was used for the treatment of the wound. Postoperative X-ray showed the status of lumbar-pelvic reconstruction; I125 radioactive particles were evenly put into the operation field. Skin necrosis occurred after surgery, which was managed with local wound care. At 6 months follow-up, bowel and bladder functions were nearly normal, and lower extremity function was normal. There was no recurrence during follow-up.
Two A 41-year-old man complained of low back pain for 8 months and sciatic pain for 10 days. The patient also had lower abdominal pain, a change in bowel habits, and constipation. Straight leg tests on both sides were 40 . Anteroposterior radiograph of the sacrococcygeal area showed a large area of osteolytic destruction at the S1–S4 levels and central growth of a tumor. Sagittal MRI showed that the tumor involved the S1–S4 levels and extended into the pelvic cavity with an anteroposterior diameter of 20 cm. Coronal MRI showed central growth of the tumor, which involved bilateral SI joint (Fig. 5). The combined anteroposterior approach was used because of the large anteroposterior diameter of the tumor. Distal resection together with proximal curettage was performed, and the bilateral S1–S3 nerve roots were retained. Because the SI joint was involved, bilateral lumbar-pelvic reconstruction was performed. A superficial wound infection developed postoperatively and was treated by debridement and wound drainage. At 6-months followup, intermittent catheterization was still required and medication was required for control of bowel function while lower extremity function was completely normal. At 37-months postoperatively, bowel and bladder functions had recovered and no additional drug treatment was needed. No local recurrence occurred during follow-up.
Table 2 Operative and follow-up data stratified by type of resection Index
Wide (n52)
Marginal (n511)
Marginalþcurettage (n512)
Curettage (n57)
Local recurrence* Blood loss (mL)z Urinary dysfunction* Bowel dysfunction* Lower limb dysfunction* Complications*
0 1,300 1 0 0
2 3,000 1 0 4
5 2,950 3 2 0
5 4,500 0 0 0
(0.0) (550–2,050) (50.0) (0.0) (0.0)
1 (50.0)
(18.2) (2,600–3,200)x (50.0) (0.0) (36.4)
3 (27.3)
(41.7) (2,850–3,425)x (25.0) (16.7) (0.0)y
6 (50.0)
(71.4)y (3,900–4,700)y,x,k (0.0) (0.0) (0.0)
1 (14.3)
Data are displayed as number (percentage) except for blood loss, which is presented as median (interquartile range). Pair-wise multiple comparisons between groups were determined using the Bonferroni test with Type I error adjustment. * p-Values based on the Fisher exact test. y Indicates a statistically significant difference between the indicated group and the marginal resection group. z p-Values based on the Kruskal-Wallis test. x Indicates a statistically significant difference between the indicated group and the wide resection group. k Indicates a statistically significant difference between the marginalþcurettage and curettage groups. { Significant difference among the four resection types, p!.05.
p .098 .004{ .267 .417 .032{ .417
G. Li et al. / The Spine Journal 12 (2012) 484–491
489
Fig. 3. (Left) Overall survival and (Right) 2-year recurrence-free survival rates.
Discussion There are no standard treatment guidelines for the surgical management of sacral GCTs. Unlike limb GCT, most cases of sacral GCT are Campanacci Grade III tumors with invasive growth [1,3]. Symptoms of sacral GCTs are nonspecific, and when diagnosed, the tumor is often very large and structures such as the lumbosacral nerve roots, iliac vessels, ureter, bladder, and rectum are frequently involved. This study of 32 patients is the largest case series of patients with sacral GCTs. The previous largest case series that included 24 patients was reported by Guo et al. [3], and the aim was to assess the outcomes of conservative surgery, either intralesional curettage or partial excision.
The median duration of follow-up was 50 months; local recurrence occurred in seven patients (29.2%), lower than the rate of 37.5% in our study, and the 5-year local RFS rate was 69.6%. In addition, 29.2% (seven patients) had urinary dysfunction, and 33.3% (eight patients) had bowel dysfunction compared with 15.6% and 6.3%, respectively, in our study. One of the core issues in the treatment of sacral tumors is the preservation of the sacral nerve roots. At present, most scholars consider that preservation of bilateral S1–S3 nerve roots and preservation of unilateral S1–S5 nerve roots can preserve bowel and bladder function in most patients, and protecting the sacral nerve roots should be a goal in the treatment of sacral tumors [9,10]. In the present study, bladder control disorders occurred in only
Fig. 4. Representative case. A 31-year-old man with a 1-year history of back pain. (A) Back view of the three-dimensional computed tomography reconstruction shows left S1–S3 bone destruction, which involves the left SI joint and ilium. (B) The patient was placed prone and posterior paramidline curved incision was made. (C) Computed tomography cross-section view showed that the tumor involved the left SI joint and left S1 sacral foramen but did not involve the midline. (D) Postoperative X-ray shows the status of lumbar-pelvic reconstruction; I125 radioactive particles were evenly put into the operation field. SI, sacroiliac.
490
G. Li et al. / The Spine Journal 12 (2012) 484–491
Fig. 5. Representative case. A 41-year-old man with an 8-month history of back pain. (A) Preoperative anteroposterior X-ray of sacrococcygeal area showed a large area of osteolytic destruction at S1–S4 levels and central growth of the tumor. (B) Sagittal MRI showed that the tumor involved the S1–S4 levels and grew into the pelvic cavity, and the anteroposterior diameter was about 20 cm. (C) Coronal MRI showed central growth of the tumor, which involved the bilateral SI joint; there are many uneven necrotic tissue signals inside the tumor. (D) Postoperative anteroposterior X-ray shows the status of lumbar-pelvic reconstruction. MRI, magnetic resonance imaging.
five patients of those with bilateral S1–S3 nerve root preservation (four cases) or unilateral S1–S5 nerve root preservation (one case); no bowel dysfunction occurred in any of these patients. Surgical treatment of sacral GCTs includes intralesional curettage, marginal resection plus curettage, marginal resection alone, extensive marginal resection, and en bloc resection. Generally speaking, the more extensive the resection, the better the local control rate; however, the incidence of complications such as sacral nerve injury also becomes higher. Therefore, it is very important to select a ‘‘reasonable’’ surgical margin according to the anatomic location of the tumor and the tumor size. Most sacral GCTs are eccentric high sacral tumors, and a considerable number do not invade the midline. The anatomic location of the tumor is very important for the selection of a surgical margin [8]. For low sacral GCTs, en bloc tumor resection can be carried out while preserving the bilateral S1–S3 nerve roots while achieving a satisfactory surgical margin. High sacral GCTs can be divided into two types according to midline involvement. If the tumor does not involve the midline, then marginal tumor resection can be performed while preserving the contralateral S1– S5 nerve roots. If the tumor involves the midline, then marginal resection together with curettage can be performed. Marginal resection should be carried out for S3 and its
distal portion, and curettage should be carried out for the proximal part of the tumor (S1 and S2) while preserving bilateral S1–S3 nerve roots. In the present study, the three surgical resection methods mentioned previously were used for 25 patients, and simple intralesional curettage was performed for the other seven patients with early-stage disease. During follow-up, only seven of the 25 patients (21.9%) had mild urinary or bowel dysfunction, and the total recurrence rate of the 25 patients was 28%. Among these three resection groups, the local recurrence rate was 0% in the en bloc resection group, 18.2% in the marginal resection group, and 41.7% in the marginal resection plus curettage group, and all of these rates were lower than the reported local recurrence rate of nearly 50% [4]. This may be related to the reasonable marginal resection performed and effective bleeding control. In the simple curettage group, the local recurrence rate was 71.4%, and the mean blood loss was significantly greater than that in any of the other three groups. These data suggest that simple curettage should be abandoned in the treatment of patients with sacral GCTs. Selective arterial embolization (SAE) has been used to treat sacral GCTs for 3 decades, and as a minimally invasive interventional treatment, SAE has advantages including low morbidity and complication rates [11]. Reports
G. Li et al. / The Spine Journal 12 (2012) 484–491
have indicated 10- and 20-year recurrence rates of 31% and 43%, respectively, which is better than that achieved with surgical treatment [11,12]. Although these data suggest that SAE can be the first choice treatment for some large sacral GCTs, SAE alone cannot provide effective local control for most sacral GCTs, and tumor progression often occurs [2]. From 2010, we began performing SAE two to four times before surgery for patients with sacral GCTs and our preliminary observations showed that intraoperative bleeding was reduced significantly and surgery was less complicated. Thus, we believe that SAE can play a role as an adjuvant therapy rather than a stand-alone treatment. At present, bisphosphonates and denosumab (receptor activator of nuclear factor kappa-B ligand monoclonal antibody) are the two most promising drugs in the treatment of GCTs. Both inhibit GCT cells by the inhibition of osteoclast differentiation. An in vitro experiment [13] showed that both drugs inhibited the growth of GCT cells and promoted tumor cell apoptosis. In addition, Tse et al. [14] reported that intravenous bisphosphonates could reduce the local recurrence rate of limb GCTs. Radiotherapy used to be an important adjuvant therapy in the treatment of sacral GCTs; however, it was recently reported that there were no statistical differences between intralesional curettage combined with radiotherapy and any single treatment with respect to tumor recurrence or mortality [15]. In addition, a large dose of radiation increases the risk of local sarcomatous changes [15]. Other adjuvant treatments such as freezing, argon beam coagulation, radioactive particles, and filling the defective bone cement filling may be helpful, but there are few cases reported on these methods and there has been no welldesigned controlled study [1,16,17]. Based on our results and reports in the literature, we believe that there are a number of points that should be emphasized in the treatment of sacral GCTs. Preoperative and postoperative bisphosphonate treatment should be administered. Digital subtraction angiography intravascular thrombosis should be performed 24 hours before surgery. Surgical approaches and resection should be based on the anatomic location and size of the tumor. Intraoperative auxiliary argon and a high-speed drilling should be used during the operation to treat the surgical wound. Longterm follow-up is crucial as recurrences after 5 years have been reported. Our study was limited by the median follow-up duration of 42 months. Longer follow-up is needed to better evaluate the outcome of the treatment. Also, follow-up evaluations only included data regarding urination and defecation, mobility, and recurrence; there was no assessment of quality of life. Lastly, because all patients did not receive the same adjuvant treatments, definite conclusions regarding their effects cannot be drawn.
491
Conclusions In summary, wide surgical margins can reduce the recurrence rate of sacral GCTs but result in a high rate of sacral nerve injury. Marginal resection and/or curettage reduce the rate of sacral nerve injury but result in higher rates of recurrence. Curettage alone is associated with an unacceptably high rate of recurrence. Our results indicate that surgical resection with clear margins combined with SAE and adjuvant therapy can reduce the recurrence rate while maximizing preservation of sacral nerve function.
Acknowledgments This work was supported by the Natural Science Foundation of Shanghai (grant no.11ZR1428400). References [1] Randall RL. Giant cell tumor of the sacrum. Neurosurg Focus 2003;15:E13. [2] Thangaraj R, Grimer RJ, Carter SR, et al. Giant cell tumour of the sacrum: a suggested algorithm for treatment. Eur Spine J 2010;19:1189–94. [3] Guo W, Ji T, Tang X, Yang Yl. Outcome of conservative surgery for giant cell tumor of the sacrum. Spine 2009;34:1025–31. [4] Gibbs IC, Chang SD. Radiosurgery and radiotherapy for sacral tumors. Neurosurg Focus 2003;15:E8–12. [5] Ozaki T, Liljenqvist U, Halm H, et al. Giant cell tumor of the spine. Clin Orthop Relat Res 2002;401:194–201. [6] Marcove RC, Sheth DS, Brien EW, et al. Conservative surgery for giant cell tumors of the sacrum. The role of cryosurgery as a supplement to curettage and partial excision. Cancer 1994;74:1253–60. [7] Wallace S, Granmayeh M, deSantos LA, et al. Arterial occlusion of pelvic bone tumors. Cancer 1979;43:322–8. [8] Cai ZD, Li GD, Fu Q, et al. Primary discussion on surgical classification of sacral tumors. Chin J Orthop 2008;28:101–5. [9] Stener B, Gunterberg B. Resection of the sacrum for tumors. Chir Organi Mov 1990;75(1 Supp1):108–10. [10] Todd LT, Yaszemski MJ, Currier BL, et al. Bowel and bladder function after major sacral resection. Clin Orthop 2002;397:36–9. [11] Hosalkar HS, Jones KJ, King JJ, Lackman RD. Serial arterial embolization for large sacral giant cell tumours: mid to long term results. Spine 2007;32:1107–15. [12] Lin PP, Guzel VB, Moura MF, et al. Long term follow-up of patients with giant cell tumor of the sacrum treated with selective arterial embolization. Cancer 2002;95:1317–25. [13] Cheng YY, Huang L, Lee KM, et al. Bisphosphonates induce apoptosis of stromal tumor cells in giant cell tumor of bone. Calcif Tissue Int 2004;75:71–7. [14] Tse LF, Wong KC, Kumta SM, et al. Bisphosphonates reduce local recurrence in extremity giant cell tumor of bone: a case-control study. Bone 2008;42:68–73. [15] Leggon RE, Zlotecki R, Reith J, Scarborough MT. Giant cell tumor of the pelvis and sacrum: 17 cases and analysis of the literature. Clin Orthop Relat Res 2004;423:196–207. [16] Kollender Y, Meller I, Bickels J, et al. Role of adjuvant cryosurgery in intralesional treatment of sacral tumors: results of a 3–11-year follow-up. Cancer 2003;97:2830–8. [17] Takeda N, Kobayashi T, Tandai S, et al. Treatment of giant cell tumors in the sacrum and spine with curettage and argon beam coagulator. J Orthop Sci 2009;14:210–4.
Clinical Study
Surgical strategy for the management of sacral giant cell tumors: a 32-case series Guodong Li, MD, Dong Fu, MD, Kai Chen, MD, Xiaojun Ma, MD, Mengxiong Sun, MD, Wei Sun, MD, Jian Li, MD, Zhendong Cai, MD* Department of Orthopaedics, Shanghai Tenth People’s Hospital, Tongji University School of Medicine, 301 Yanchang Rd, Shanghai 200072, China Received 14 December 2011; revised 28 April 2012; accepted 21 June 2012
Abstract
BACKGROUND CONTEXT: Surgical treatment of sacral giant cell tumors (GCTs) is associated with a high rate of complications, and there is controversy over which type of surgical treatment is optimal. PURPOSE: To develop an optimal treatment strategy for sacral GCTs. STUDY DESIGN/SETTING: Retrospective/academic medical center. PATIENT SAMPLE: A total of 32 patients (18 women and 14 men) with sacral GCT who underwent surgery between August 1996 and August 2008. OUTCOME MEASURES: Local recurrence rate, surgical margins, blood loss, sacral nerve root preservation, and complications. METHODS: The medical charts of 32 patients were reviewed. RESULTS: Patients underwent either wide resection (n52), marginal resection (n511), marginal resection plus curettage (n512), or curettage alone (n57). The curettage group and the wide resection group had the highest and lowest amounts of blood loss (4,500 vs. 1,300 mL, respectively). During follow-up (median, 42 months), 12 patients (37.5%) had local recurrence, including five of seven in the curettage group. The recurrence rate was significantly lower in the marginal excision group compared with that in the curettage group (18.2% vs. 71.4%, respectively; p5.049). Five patients had bladder dysfunction, and two patients had bowel dysfunction. Four patients who underwent marginal resection had lower limb dysfunction. Overall survival was 93.6%, and 2-year recurrence-free survival was 84.4%. CONCLUSIONS: Choosing an optimal surgical margin in the treatment of sacral GCTs is of great importance for local recurrence control and sacral nerve root preservation. Curettage alone should not be used to treat sacral GCT. Ó 2012 Elsevier Inc. All rights reserved.
Keywords:
Giant cell tumor; Sacrum; Surgical treatment
Introduction Giant cell tumors (GCTs) of bones are one of the most common neoplasms, accounting for 5% to 10% of all the primary bone tumors [1]. The sacrum is the third most common site for GCTs, following the peripheral area of the
FDA device/drug status: Not applicable. Author disclosures: GL: Nothing to disclose. DF: Nothing to disclose. KC: Nothing to disclose. XM: Nothing to disclose. MS: Nothing to disclose. WS: Nothing to disclose. JL: Nothing to disclose. ZC: Nothing to disclose. GL and DF are co-first authors. * Corresponding author. Department of Orthopaedics, Shanghai Tenth People’s Hospital, Tongji University School of Medicine, 301 Yanchang Rd, Shanghai 200072, China. Tel.: (86) 21-66307311; fax: (86) 21-56305592. E-mail address: [email protected] (Z. Cai) 1529-9430/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.spinee.2012.06.014
knee joint and the distal radius, and accounts for 3% to 4% of all GCTs [2]. Compared with GCTs at other sites, sacral GCTs have unique characteristics such as very complicated local anatomy, aggressive behavior, and nonspecific symptoms that can result in a delayed diagnosis. Although sacral GCTs are benign from the pathologic point of view, treatment is very difficult, optimal management is controversial, and there are no standard treatment guidelines [3]. Conventional treatments include radiotherapy, curettage combined with radiotherapy, marginal resection combined with curettage, marginal resection, extensive marginal resection, en bloc resection, and serial arterial embolization. The significant advantage of radiotherapy is the avoidance of surgical morbidity [3]. The main disadvantage is poor local control of the tumor and secondary
G. Li et al. / The Spine Journal 12 (2012) 484–491
sarcomatous changes [4,5]. Curettage combined with radiotherapy and marginal resection plus curettage can retain the sacral nerve root and maintain the completeness of the pelvic ring, but the rate of local recurrence is relatively high. Extensive marginal resection and en bloc resection may maximally decrease the recurrence rate but increase the incidence of surgical complications [6,7]. The purpose of this study was to retrospectively review the management of 32 cases of sacral GCTs who underwent surgical treatment at our medical center to develop a standard method for the treatment of sacral GCTs.
Materials and methods General data We retrospectively reviewed the medical records of 32 consecutive patients with sacral GCTs treated at our center between August 1996 and August 2008. There were 14 men and 18 women with a median age of 31 years (range, 19–74 years). All 32 patients had a diagnosis of primary GCT by pathologic examination of a tissue sample. Patients with recurrent GCTs, a postoperative pathologic diagnosis of malignant sacral GCT, and those lost to follow-up were excluded. This study was approved by the Institutional Review Board of our hospital, and because of the retrospective nature of the study, the requirement of informed consent was waived. Preoperative and surgical management All patients received anteroposterior radiographs, computed tomography (CT), magnetic resonance imaging (MRI), digital subtraction angiography, and emission computed tomography of the sacrum. Radiography, CT, and MRI were considered the main examination methods for three-dimensional evaluation of the tumors. Digital subtraction angiography was routinely performed for tumor angiography 24 to 72 hours before surgery. Embolization of the middle sacral artery, tumor feeding arteries, and selective bilateral internal iliac artery was carried out with Gelfoam and polyvinyl alcohol because of the extensive blood supply to the tumors. Beginning January 2008, the operation was performed 24 hours rather than 72 hours after digital subtraction angiography plus embolization. The surgical approach was based on the size and anatomic location of the tumor [8]. Two surgical approaches were used. The first was a single posterior approach using a posterior paramidline curved incision, posterior midline straight incision, or S3-plane transverse S-shaped incision. The posterior midline straight incision was suitable for high sacral tumors in the midline, the posterior paramidline curved incision was suitable for high eccentric tumors, and the S3-plane transverse S-shaped incision was suitable for low tumors. The sacrospinal muscles were dissected bilaterally, and the posterior wall of the sacral canal was opened. The tumor was removed while preserving at least
485
Context The best surgical strategy for the treatment of sacral giant cell tumors continues to be refined over time. The authors present their experience with this rare tumor. Contribution Based upon their chart review of 32 patients, the authors recommend avoiding curettage, as it incurs too high of a risk for local recurrence, and, conversely, overly aggressive margins (can put the sacral nerves at greater-thanneeded risk), which did not seem to impact outcomes. Implication This group’s recommendations are reasonable and, generally, in line with the opinions of the orthopedic oncology community. Denosumab and related therapies appear, at this time, as though they might be adjuvants that allow the authors’ suggested ‘‘middle road’’ to be more easily traveled. —The Editors unilateral S1–S3 nerve roots. In two patients with low sacral GCT, transverse S-shaped incisions were used. The origins of the bilateral gluteus maximus muscles were cut bilaterally to the superior margin of the greater sciatic foramen. The sacrotuberous, sacrospinous, and coccygeal ligaments were also cut. Osteotomy was performed at the upper part of the tumor, and the tumor was removed completely. The second approach was a combined anteroposterior approach. The anterior approach was performed through a midline incision in the lower abdomen with the patient in the lithotomy position for transabdominal exposure of the pelvic cavity. The integrity of the pseudomembrane of the tumor was preserved as much as possible to avoid contaminating the organs inside the pelvic cavity. The tumor was completely resected, and the portion of the sacrum invaded by the lesion was also removed. When the tumor was large and sacral destruction was significant, the tumor was partially resected along the anterior margin of the sacrum. A piece of large abdominal gauze was placed over the area of resection, and the retroperitoneum, peritoneum, and abdominal cavity were closed. Next, the patient was turned over and the posterior surgery was performed. In this series, marginal resection is defined as the complete removal of the tumor including soft tissues 1 to 3 cm outside the tumor capsule after sufficient exposure of the tumor and identification of the tumor capsule. All surgical margins were confirmed by intraoperative frozen section analysis. A negative margin was defined as no tumor cells found in the removed soft tissue around the tumor by pathologic examination. En bloc resection was carried out for tumors involving S3–S5 (Fig. 1, Left). Marginal resection was carried out for high eccentric tumors that did not cross
486
G. Li et al. / The Spine Journal 12 (2012) 484–491
Fig. 1. Posterior view of the sacrum. (Left) En bloc resection for sacral GCTs involving S3–S5. Green line between tumor and normal sacrum represents the surgical margin. (Middle) Marginal resection for high eccentric tumors that do not cross the midline. (Right) For high sacral GCTs, curettage or curettage together with marginal resection was carried out for patients with early-stage disease. Marginal resection was performed for S3 and its distal portion, and curettage was performed for the proximal portion of the tumor (S1 and S2). GCT, giant cell tumor.
the midline (Fig. 1, Middle) while totally preserving the contralateral sacral nerve roots. For high sacral tumors, curettage or curettage together with marginal resection was performed for patients with early-stage disease. Specifically, marginal resection was performed for S3 and its distal portion, and curettage was performed for the proximal portion of the tumor (S1 and S2) (Fig. 1, Right). After tumor resection, before 2008, the residual cavity and wound were treated with high-speed burr grinding combined with absolute ethanol plus distilled water for 28 patients. From 2008, four patients in total were treated using an argon beam coagulator plus I125 radioactive particles, which were placed evenly in the tumor cavity. The ISOLA pedicle screw system (DePuy, Raynham, MA, USA) was used for the reconstruction of the lumbosacral area when indicated in 28 of 32 patients, including unilateral reconstruction in seven patients (25.0%) and bilateral reconstruction in 21 patients (75.0%). A general treatment algorithm is shown in Fig. 2. Follow-up Follow-up evaluations were performed at 1, 2, 3, 6, and 12 months during the first year after surgery, every 3 months during the second year, and then every 6 months. During follow-up, radiographs and CT scans of the sacrococcygeal region as well as CT scans of the lungs were obtained. The patients were assessed for tumor recurrence (primary outcome), urination and defecation functions, lower limb sensation, myodynamia, and wound healing. Statistical analysis Categorical variables were compared using the Fisher exact test and blood loss using the Kruskal-Wallis test.
When a difference between groups was apparent, multiple comparisons of means were performed using the Bonferroni procedure with Type I error adjustment. Categorical data were represented by number (%), and continuous data were presented as median (interquartile range). The Kaplan-Meier method was used to measure overall survival (OS) and 2-year recurrence-free survival (RFS). All statistical assessments were two-sided and evaluated at the .05 level of significance. Statistical analyses were performed using SPSS 15.0 statistics software (SPSS, Inc., Chicago, IL, USA).
Results Patient demographic data and nerve root involvement are summarized in Table 1, and operative and follow-up data are presented in Table 2. Two patients underwent wide resection, and 11 underwent marginal resection with preservation of all the contralateral sacral nerve roots. Among the 19 patients with high sacral GCTs, seven underwent curettage and 12 underwent marginal resection plus curettage. The ISOLA pedicle screw system was used for the reconstruction of the lumbosacral area in 28 patients, including unilateral reconstruction in seven patients (25.0%) and bilateral reconstruction in 21 patients (75.0%). Five of 32 patients received radiation therapy at other hospitals before surgery. Local radiation therapy was performed after the operation for eight of the 12 patients who had recurrence. Perioperative adjuvant treatment with bisphosphonates (zoledronic acid, 4 mg/wk; two courses of treatment before surgery and four courses after surgery) was carried out for three patients. Local tumor recurrence occurred in 12 patients (37.5%) during follow-up. The local recurrence rates were 0%,
G. Li et al. / The Spine Journal 12 (2012) 484–491
487
Fig. 2. General treatment algorithm. GCT, giant cell tumor.
18.2%, 41.7%, and 71.4% for patients with wide resection (0/2), marginal resection (2/11), marginal resection plus curettage (5/12), and curettage (5/7), respectively. The marginal resection group had a significantly lower recurrence rate than the curettage group (18.2% vs. 71.4%; p5.049). There was a significant difference in blood loss among the four types of resection (p5.004). The greatest blood loss occurred in patients who underwent curettage (4,500 [3,900–4,700] mL) and the lowest occurred in patients who received wide resection (1,300 [550–2,050] mL). There was no significant difference in blood loss between the marginal resection and marginal plus curettage groups (3,000 vs. 2,950 mL; p5.557). The median duration of follow-up was 42 months (range, 18–115 months). Six patients had recurrence within 2 years after surgery, and altogether, 12 patients developed recurrence during the follow-up period. Among them, two patients died at the 18th and 23rd months after surgery, respectively, because of widespread pulmonary metastasis. Sarcomatous changes were observed on pathologic examinations after surgery in two cases. The OS and 2-year RFS curves are shown in Fig. 3. The OS rate was 93.6%, and 2-year RFS rate was
84.4%. In addition, benign pulmonary metastasis occurred during follow-up in one patient. During follow-up, five patients (5/32, 15.6%) developed urinary bladder dysfunction and two patients (2/32, 6.3%) developed bowel dysfunction requiring medication. Four patients with marginal resections had lower limb dysfunction (12.5%). Eleven patients (34.4%) had wound complications; five underwent local debridement, drainage, and secondary closure; and in four patients the wounds healed within 3 months with conservative management. In addition, two patients required skin flap transplantation after radiotherapy because of the large skin lesions. There were no significant differences in urinary bladder and bowel dysfunction and complications among the four types of resections (Table 2). Representative cases One A 31-year-old man complained of low back pain for 1 year, which was aggravated at night and not relieved by over-the-counter analgesics. He experienced sensory loss and numbness on the lateral side of the left foot, foot
488
G. Li et al. / The Spine Journal 12 (2012) 484–491
Table 1 Patient data (N532) Demographic data/nerve root involvement
Median (range)/n (%)
Age (y) Gender Male Female Nerve roots involved Left S1 S1–S2 S1–S3 S1–S4 S1–S5 S2–S3 S2–S4 S4–S5 S5 Right S1 S1–S2 S1–S3 S1–S4 S1–S5 S2–S4 S2–S5 S4–S5 S5
31 (19–74) 14 (43.8) 18 (56.3)
1 7 6 3 1 1 4 1 1
(3.1) (21.9) (18.8) (9.4) (3.1) (3.1) (12.5) (3.1) (3.1)
1 11 5 4 1 3 1 1 1
(3.1) (34.4) (15.6) (12.5) (3.1) (9.4) (3.1) (3.1) (3.1)
dorsal flexion was impossible, and a left-side straight leg test of 25 . Back view of the three-dimensional CT reconstruction showed left S1–S3 bone destruction, which involved the left sacroiliac (SI) joint and ilium. Computed tomography cross-section view shows that the tumor involved the left SI joint and left S1 sacral foramen but did not involve the midline. The patient was placed prone and posterior paramidline curved incision was made (Fig. 4). Marginal tumor resection was performed via the posterior approach,
and the contralateral S1–S5 nerve roots and ipsilateral S4–S5 nerve roots were retained. An argon beam coagulator was used for the treatment of the wound. Postoperative X-ray showed the status of lumbar-pelvic reconstruction; I125 radioactive particles were evenly put into the operation field. Skin necrosis occurred after surgery, which was managed with local wound care. At 6 months follow-up, bowel and bladder functions were nearly normal, and lower extremity function was normal. There was no recurrence during follow-up.
Two A 41-year-old man complained of low back pain for 8 months and sciatic pain for 10 days. The patient also had lower abdominal pain, a change in bowel habits, and constipation. Straight leg tests on both sides were 40 . Anteroposterior radiograph of the sacrococcygeal area showed a large area of osteolytic destruction at the S1–S4 levels and central growth of a tumor. Sagittal MRI showed that the tumor involved the S1–S4 levels and extended into the pelvic cavity with an anteroposterior diameter of 20 cm. Coronal MRI showed central growth of the tumor, which involved bilateral SI joint (Fig. 5). The combined anteroposterior approach was used because of the large anteroposterior diameter of the tumor. Distal resection together with proximal curettage was performed, and the bilateral S1–S3 nerve roots were retained. Because the SI joint was involved, bilateral lumbar-pelvic reconstruction was performed. A superficial wound infection developed postoperatively and was treated by debridement and wound drainage. At 6-months followup, intermittent catheterization was still required and medication was required for control of bowel function while lower extremity function was completely normal. At 37-months postoperatively, bowel and bladder functions had recovered and no additional drug treatment was needed. No local recurrence occurred during follow-up.
Table 2 Operative and follow-up data stratified by type of resection Index
Wide (n52)
Marginal (n511)
Marginalþcurettage (n512)
Curettage (n57)
Local recurrence* Blood loss (mL)z Urinary dysfunction* Bowel dysfunction* Lower limb dysfunction* Complications*
0 1,300 1 0 0
2 3,000 1 0 4
5 2,950 3 2 0
5 4,500 0 0 0
(0.0) (550–2,050) (50.0) (0.0) (0.0)
1 (50.0)
(18.2) (2,600–3,200)x (50.0) (0.0) (36.4)
3 (27.3)
(41.7) (2,850–3,425)x (25.0) (16.7) (0.0)y
6 (50.0)
(71.4)y (3,900–4,700)y,x,k (0.0) (0.0) (0.0)
1 (14.3)
Data are displayed as number (percentage) except for blood loss, which is presented as median (interquartile range). Pair-wise multiple comparisons between groups were determined using the Bonferroni test with Type I error adjustment. * p-Values based on the Fisher exact test. y Indicates a statistically significant difference between the indicated group and the marginal resection group. z p-Values based on the Kruskal-Wallis test. x Indicates a statistically significant difference between the indicated group and the wide resection group. k Indicates a statistically significant difference between the marginalþcurettage and curettage groups. { Significant difference among the four resection types, p!.05.
p .098 .004{ .267 .417 .032{ .417
G. Li et al. / The Spine Journal 12 (2012) 484–491
489
Fig. 3. (Left) Overall survival and (Right) 2-year recurrence-free survival rates.
Discussion There are no standard treatment guidelines for the surgical management of sacral GCTs. Unlike limb GCT, most cases of sacral GCT are Campanacci Grade III tumors with invasive growth [1,3]. Symptoms of sacral GCTs are nonspecific, and when diagnosed, the tumor is often very large and structures such as the lumbosacral nerve roots, iliac vessels, ureter, bladder, and rectum are frequently involved. This study of 32 patients is the largest case series of patients with sacral GCTs. The previous largest case series that included 24 patients was reported by Guo et al. [3], and the aim was to assess the outcomes of conservative surgery, either intralesional curettage or partial excision.
The median duration of follow-up was 50 months; local recurrence occurred in seven patients (29.2%), lower than the rate of 37.5% in our study, and the 5-year local RFS rate was 69.6%. In addition, 29.2% (seven patients) had urinary dysfunction, and 33.3% (eight patients) had bowel dysfunction compared with 15.6% and 6.3%, respectively, in our study. One of the core issues in the treatment of sacral tumors is the preservation of the sacral nerve roots. At present, most scholars consider that preservation of bilateral S1–S3 nerve roots and preservation of unilateral S1–S5 nerve roots can preserve bowel and bladder function in most patients, and protecting the sacral nerve roots should be a goal in the treatment of sacral tumors [9,10]. In the present study, bladder control disorders occurred in only
Fig. 4. Representative case. A 31-year-old man with a 1-year history of back pain. (A) Back view of the three-dimensional computed tomography reconstruction shows left S1–S3 bone destruction, which involves the left SI joint and ilium. (B) The patient was placed prone and posterior paramidline curved incision was made. (C) Computed tomography cross-section view showed that the tumor involved the left SI joint and left S1 sacral foramen but did not involve the midline. (D) Postoperative X-ray shows the status of lumbar-pelvic reconstruction; I125 radioactive particles were evenly put into the operation field. SI, sacroiliac.
490
G. Li et al. / The Spine Journal 12 (2012) 484–491
Fig. 5. Representative case. A 41-year-old man with an 8-month history of back pain. (A) Preoperative anteroposterior X-ray of sacrococcygeal area showed a large area of osteolytic destruction at S1–S4 levels and central growth of the tumor. (B) Sagittal MRI showed that the tumor involved the S1–S4 levels and grew into the pelvic cavity, and the anteroposterior diameter was about 20 cm. (C) Coronal MRI showed central growth of the tumor, which involved the bilateral SI joint; there are many uneven necrotic tissue signals inside the tumor. (D) Postoperative anteroposterior X-ray shows the status of lumbar-pelvic reconstruction. MRI, magnetic resonance imaging.
five patients of those with bilateral S1–S3 nerve root preservation (four cases) or unilateral S1–S5 nerve root preservation (one case); no bowel dysfunction occurred in any of these patients. Surgical treatment of sacral GCTs includes intralesional curettage, marginal resection plus curettage, marginal resection alone, extensive marginal resection, and en bloc resection. Generally speaking, the more extensive the resection, the better the local control rate; however, the incidence of complications such as sacral nerve injury also becomes higher. Therefore, it is very important to select a ‘‘reasonable’’ surgical margin according to the anatomic location of the tumor and the tumor size. Most sacral GCTs are eccentric high sacral tumors, and a considerable number do not invade the midline. The anatomic location of the tumor is very important for the selection of a surgical margin [8]. For low sacral GCTs, en bloc tumor resection can be carried out while preserving the bilateral S1–S3 nerve roots while achieving a satisfactory surgical margin. High sacral GCTs can be divided into two types according to midline involvement. If the tumor does not involve the midline, then marginal tumor resection can be performed while preserving the contralateral S1– S5 nerve roots. If the tumor involves the midline, then marginal resection together with curettage can be performed. Marginal resection should be carried out for S3 and its
distal portion, and curettage should be carried out for the proximal part of the tumor (S1 and S2) while preserving bilateral S1–S3 nerve roots. In the present study, the three surgical resection methods mentioned previously were used for 25 patients, and simple intralesional curettage was performed for the other seven patients with early-stage disease. During follow-up, only seven of the 25 patients (21.9%) had mild urinary or bowel dysfunction, and the total recurrence rate of the 25 patients was 28%. Among these three resection groups, the local recurrence rate was 0% in the en bloc resection group, 18.2% in the marginal resection group, and 41.7% in the marginal resection plus curettage group, and all of these rates were lower than the reported local recurrence rate of nearly 50% [4]. This may be related to the reasonable marginal resection performed and effective bleeding control. In the simple curettage group, the local recurrence rate was 71.4%, and the mean blood loss was significantly greater than that in any of the other three groups. These data suggest that simple curettage should be abandoned in the treatment of patients with sacral GCTs. Selective arterial embolization (SAE) has been used to treat sacral GCTs for 3 decades, and as a minimally invasive interventional treatment, SAE has advantages including low morbidity and complication rates [11]. Reports
G. Li et al. / The Spine Journal 12 (2012) 484–491
have indicated 10- and 20-year recurrence rates of 31% and 43%, respectively, which is better than that achieved with surgical treatment [11,12]. Although these data suggest that SAE can be the first choice treatment for some large sacral GCTs, SAE alone cannot provide effective local control for most sacral GCTs, and tumor progression often occurs [2]. From 2010, we began performing SAE two to four times before surgery for patients with sacral GCTs and our preliminary observations showed that intraoperative bleeding was reduced significantly and surgery was less complicated. Thus, we believe that SAE can play a role as an adjuvant therapy rather than a stand-alone treatment. At present, bisphosphonates and denosumab (receptor activator of nuclear factor kappa-B ligand monoclonal antibody) are the two most promising drugs in the treatment of GCTs. Both inhibit GCT cells by the inhibition of osteoclast differentiation. An in vitro experiment [13] showed that both drugs inhibited the growth of GCT cells and promoted tumor cell apoptosis. In addition, Tse et al. [14] reported that intravenous bisphosphonates could reduce the local recurrence rate of limb GCTs. Radiotherapy used to be an important adjuvant therapy in the treatment of sacral GCTs; however, it was recently reported that there were no statistical differences between intralesional curettage combined with radiotherapy and any single treatment with respect to tumor recurrence or mortality [15]. In addition, a large dose of radiation increases the risk of local sarcomatous changes [15]. Other adjuvant treatments such as freezing, argon beam coagulation, radioactive particles, and filling the defective bone cement filling may be helpful, but there are few cases reported on these methods and there has been no welldesigned controlled study [1,16,17]. Based on our results and reports in the literature, we believe that there are a number of points that should be emphasized in the treatment of sacral GCTs. Preoperative and postoperative bisphosphonate treatment should be administered. Digital subtraction angiography intravascular thrombosis should be performed 24 hours before surgery. Surgical approaches and resection should be based on the anatomic location and size of the tumor. Intraoperative auxiliary argon and a high-speed drilling should be used during the operation to treat the surgical wound. Longterm follow-up is crucial as recurrences after 5 years have been reported. Our study was limited by the median follow-up duration of 42 months. Longer follow-up is needed to better evaluate the outcome of the treatment. Also, follow-up evaluations only included data regarding urination and defecation, mobility, and recurrence; there was no assessment of quality of life. Lastly, because all patients did not receive the same adjuvant treatments, definite conclusions regarding their effects cannot be drawn.
491
Conclusions In summary, wide surgical margins can reduce the recurrence rate of sacral GCTs but result in a high rate of sacral nerve injury. Marginal resection and/or curettage reduce the rate of sacral nerve injury but result in higher rates of recurrence. Curettage alone is associated with an unacceptably high rate of recurrence. Our results indicate that surgical resection with clear margins combined with SAE and adjuvant therapy can reduce the recurrence rate while maximizing preservation of sacral nerve function.
Acknowledgments This work was supported by the Natural Science Foundation of Shanghai (grant no.11ZR1428400). References [1] Randall RL. Giant cell tumor of the sacrum. Neurosurg Focus 2003;15:E13. [2] Thangaraj R, Grimer RJ, Carter SR, et al. Giant cell tumour of the sacrum: a suggested algorithm for treatment. Eur Spine J 2010;19:1189–94. [3] Guo W, Ji T, Tang X, Yang Yl. Outcome of conservative surgery for giant cell tumor of the sacrum. Spine 2009;34:1025–31. [4] Gibbs IC, Chang SD. Radiosurgery and radiotherapy for sacral tumors. Neurosurg Focus 2003;15:E8–12. [5] Ozaki T, Liljenqvist U, Halm H, et al. Giant cell tumor of the spine. Clin Orthop Relat Res 2002;401:194–201. [6] Marcove RC, Sheth DS, Brien EW, et al. Conservative surgery for giant cell tumors of the sacrum. The role of cryosurgery as a supplement to curettage and partial excision. Cancer 1994;74:1253–60. [7] Wallace S, Granmayeh M, deSantos LA, et al. Arterial occlusion of pelvic bone tumors. Cancer 1979;43:322–8. [8] Cai ZD, Li GD, Fu Q, et al. Primary discussion on surgical classification of sacral tumors. Chin J Orthop 2008;28:101–5. [9] Stener B, Gunterberg B. Resection of the sacrum for tumors. Chir Organi Mov 1990;75(1 Supp1):108–10. [10] Todd LT, Yaszemski MJ, Currier BL, et al. Bowel and bladder function after major sacral resection. Clin Orthop 2002;397:36–9. [11] Hosalkar HS, Jones KJ, King JJ, Lackman RD. Serial arterial embolization for large sacral giant cell tumours: mid to long term results. Spine 2007;32:1107–15. [12] Lin PP, Guzel VB, Moura MF, et al. Long term follow-up of patients with giant cell tumor of the sacrum treated with selective arterial embolization. Cancer 2002;95:1317–25. [13] Cheng YY, Huang L, Lee KM, et al. Bisphosphonates induce apoptosis of stromal tumor cells in giant cell tumor of bone. Calcif Tissue Int 2004;75:71–7. [14] Tse LF, Wong KC, Kumta SM, et al. Bisphosphonates reduce local recurrence in extremity giant cell tumor of bone: a case-control study. Bone 2008;42:68–73. [15] Leggon RE, Zlotecki R, Reith J, Scarborough MT. Giant cell tumor of the pelvis and sacrum: 17 cases and analysis of the literature. Clin Orthop Relat Res 2004;423:196–207. [16] Kollender Y, Meller I, Bickels J, et al. Role of adjuvant cryosurgery in intralesional treatment of sacral tumors: results of a 3–11-year follow-up. Cancer 2003;97:2830–8. [17] Takeda N, Kobayashi T, Tandai S, et al. Treatment of giant cell tumors in the sacrum and spine with curettage and argon beam coagulator. J Orthop Sci 2009;14:210–4.