377
Adult Rhabdomyosarcoma Outcome following Multimodality Treatment
Darren J. Little, M.D.1 Matthew T. Ballo, M.D.1 Gunar K. Zagars, M.D.1 Peter W.T. Pisters, M.D.2 Shreyaskumar R. Patel, M.D.3 Adel K. El-Naggar, M.D.4 Adam S. Garden, M.D.1 Robert S. Benjamin, M.D.3 1
Department of Radiation Oncology, University of Texas M.D. Anderson Cancer Center, Houston, Texas.
2
Department of Surgical Oncology, University of Texas M.D. Anderson Cancer Center, Houston, Texas.
3
Department of Medical Oncology, University of Texas M.D. Anderson Cancer Center, Houston, Texas.
4
Department of Pathology, University of Texas M.D. Anderson Cancer Center, Houston, Texas.
BACKGROUND. Childhood rhabdomyosarcoma (RMS) has a relatively good prognosis. Outcome for adults with this disease is poorly documented due to its rarity. METHODS. The clinicopathologic features, treatment methods, and disease outcome were reviewed retrospectively for 82 adults with locoregional RMS treated between 1960 and 1998. Patients with distant metastasis at diagnosis were excluded. Actuarial univariate and multivariate statistical methods were used to evaluate outcome. RESULTS. Patient ages ranged from 17 to 84 years (median, 27 years). Histologic subtypes were embryonal (34%), pleomorphic (43%), and alveolar (23%). Anatomic sites of origin were head and neck (52%), trunk (26%), and extremity (7%). Tumor size was 5 cm or smaller in 51% of patients. Regional lymph node metastasis was present in 33% of patients at presentation. Treatment consisted of radiation alone in 11%, radiation and surgery in 18%, radiation and chemotherapy in 34%, and all three modalities in 37%. With a median follow-up of 10.5 years, the 10-year actuarial disease-free and overall survival rates were 41% and 40%, respectively. The 10-year actuarial local, lymph node, and metastatic control rates were 75%, 82%, and 53%, respectively. The major determinant of metastatic control and survival was primary tumor size (ⱕ 5 vs. ⬎ 5 cm). Local control was satisfactory (10-year rate of 87%) for sites other than parameningeal (50% at 10 years). Patients whose disease responded to chemotherapy had a significantly better metastasis free period (72% at 10 years) than those whose disease failed to respond (19% at 10 years). CONCLUSIONS. Adult RMS is a highly malignant tumor with a significant incidence of metastatic recurrence. Continuing investigation of new and potentially more effective chemotherapy is crucial. Local control is satisfactory for sites other than parameningeal where new radiation technologies such as intensity-modulated therapy may be necessary to safely deliver adequate doses. Cancer 2002;95: 377– 88. © 2002 American Cancer Society. DOI 10.1002/cncr.10669 KEYWORDS: adult rhabdomyosarcoma, radiation therapy, surgery, chemotherapy, prognostic factors.
Supported in part by Grant CA 06294 awarded by the National Cancer Institute, U.S. Department of Health and Human Services. Address for reprints: Matthew T. Ballo, M.D., Department of Radiation Oncology, Box 97 M.D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030; Fax: (713) 794-5573; E-mail:
[email protected] Received November 12, 2001; revision received January 30, 2002; accepted February 15, 2002. © 2002 American Cancer Society
R
habdomyosarcoma (RMS) is a highly malignant, but uncommon tumor that accounts for 15–20% of all soft tissue sarcomas.1 It occurs largely, but not exclusively, in children among whom about 250 new cases are diagnosed each year in the United States.2 The Intergroup Rhabdomyosarcoma Study Group (IRSG), formed in 1972, has recruited the majority of children (by their definition, those younger than 21 years of age) with newly diagnosed RMS in the United States into protocols designed to investigate the therapy and biology of this disease. Four successive protocols—IRS-I (1972–1978), IRS-II (1978 –1984), IRS-III (1984 –1991), and IRS-IV (1991–1997)— have accrued more than 4000 patients eligible for evaluation of out-
378
CANCER July 15, 2002 / Volume 95 / Number 2
come 3 and results have been reported in detail.4 –7 The currently active pediatric protocol for RMS is IRSV (1997–present).3 Similar cooperative groups have been established in a number of countries.8,9 Based on the results of these studies, generally accepted treatment guidelines for childhood RMS include gross total resection with preservation of function, systemic chemotherapy using combinations of actinomycin-D, vincristine, and cyclophosphamide, and radiation therapy for all but completely resected tumors of embryonal subtype. As a result of this multimodality approach, the prognosis for children with RMS has improved dramatically from a once near uniformly fatal disease to one with long-term survival rates of 70 – 80%.6,7 RMS in adults, however, is another issue. Although experiences from childhood RMS are extrapolated widely to adults with this disease, therapeutic success has been limited and long-term survival rates remain poor in the range of 35– 45%.10 –15 Even in the IRSG studies, an adverse effect of increasing age on outcome has been documented. In IRS-IV, age was prognostically significant and, in particular, adolescent boys 10 years of age or older fared significantly worse (3-year failure-free survival, 63%) than boys younger than 10 years (3-year failure-free-survival, 90%).7 Other investigators have also reported that increasing age is an adverse prognostic factor.16 The reasons for this age effect are unclear but may reflect differences in distribution of histopathologic subtypes, diminishing chemosensitivity as tumors arise in older patients, or lower adult patient tolerance for intensive therapy. RMS in adults is very rare and clinical experience is limited to case reports, small series, and nonrandomized therapeutic approaches. Moreover, the pathologic diagnosis and subclassification of RMS remain difficult and have only recently improved with the availability of specific immunohistochemistry markers.17–20 Detailed reports of multimodal treatment outcome, patterns of failure, and prognostic factors in adult patients with RMS are few.14,15 This study is an attempt to further clarify these issues in a cohort of adults with pathologically confirmed RMS.
MATERIALS AND METHODS Between 1960 and 1998, inclusive, 157 consecutive patients 17 years of age and older with the diagnosis of RMS were irradiated in the Department of Radiation Oncology at the University of Texas M.D. Anderson Cancer Center (MDACC) as part of their overall management. Seventeen years of age is the cutpoint used at our institution for pediatric patients. Seventy-five patients were excluded from this analysis because they were irradiated with palliative intent for massive
local disease (15 patients) or for distant metastases (49 patients), they received high-dose radiation with tourniquet limb hypoxia (5 patients), or because the histopathologic diagnosis was revised on current review (6 patients). The remaining 82 patients are the subjects of this review. A retrospective review of medical records was performed. Each patient had been evaluated with a full history, physical examination, routine blood tests, chest radiography, and other studies as appropriate and available during the years encompassed by this review. Histologic re-review of archived material was performed by a single pathologist (A.K.E.) and tumors were subclassified into embryonal, alveolar, or pleomorphic subtypes. Immunohistochemical staining for desmin, myoglobin, and actin was used to support the histologic diagnosis. Following treatment, patients were typically followed at 3 to 6-month intervals for the first few years and yearly thereafter. Follow-up information was obtained from the medical record or by correspondence with the referring physician. The duration of follow-up for the 31 patients alive at last contact was 20 –337 months (median, 126 months, 10.5 years). Disease recurrence was defined as any clinicoradiographic evidence of tumor recurrence. Local recurrence was any persistence or regrowth of tumor at the primary site. Recurrences were scored as regional lymph node recurrence only if they occurred within the expected draining lymphatics of the primary site; otherwise they were scored as distant. Response to treatment was assessed using a combination of radiologic, pathologic, and clinical evaluation. Actuarial plots were calculated using the Kaplan–Meier method21 and tests of significance were based on the log rank statistic. Univariate analyses of factors potentially correlated to all outcomes included sex, age, initial versus recurrent disease at presentation, histologic subtype, primary site, lymph node status, treatment, and pathologic resection margin status. Multivariate analysis was done with the proportional hazards model, using the log-linear relative function of Cox.21 The age-adjusted expected survival was calculated from U.S. vital statistics data.22 The significance of differences between proportions was tested with the chi-square or two-sided Fisher exact tests and the nonparametric Mann–Whitney test was used for differences between continuous variables.23
RESULTS Patient and Tumor Characteristics and Treatment Histologic re-review revealed embryonal subtype in 28 patients (34%), alveolar in 19 (23%), and pleomorphic in 35 (43%). There were 45 males and 37 females.
Adult Rhabdomyosarcoma/Little et al.
Patient’s ages ranged from 17 to 81 years (median, 27 years). Twenty-four patients were 21 years of age or younger. The primary tumor site was as follows: head and neck, 43 (52%); trunk, 21 (26%); lower extremity 12 (15%); and upper extremity, 6 (7%). Trunk tumors included seven arising from the body wall, two in the retroperitoneum, four in the uterus, three in the vagina or vulva, three in the perianal area, and two in the paratesticular area. Within the head and neck, 25 tumors were located in parameningeal sites24 –26 (nasopharynx, nasal cavity, paranasal sinuses, middle earmastoid, infratemporal fossa, or pterygoid-palatine fossa). Only two tumors arose in the orbit. Primary tumor size was documented in 72 cases (88%) and ranged from 1 to 15 cm (median, 5 cm); 37 were 5 cm or smaller and 35 were larger than 5 cm in largest dimension. Tumor size was significantly smaller for head and neck lesions (median, 4.5 cm) compared with lesions of the trunk (median, 6.7 cm), lower extremity (median, 9.2 cm), or upper extremity (median, 6.8 cm; Mann–Whitney, P ⬍ 0.001). Regional lymph node metastases were present in 27 patients (33%) at presentation (patients with hematogenous metastases were not included in this study). Although the incidence of lymph node disease appeared higher in the head and neck group (18 of 43) than in the non– head and neck group (9 of 39), this difference was not statistically significant (two-tailed Fisher exact test, P ⫽ 0.100). Within the head and neck tumors, the incidence of lymph node metastases was similar in parameningeal (10 of 25) and nonparameningeal (8 of 18) primary sites. Although it appeared that alveolar tumors were associated with a higher incidence of regional lymph node disease at presentation (9 of 19, 47%) than embryonal/pleomorphic disease (18 of 63, 29%), this difference was not statistically significant (two-tailed Fisher exact test, P ⫽ 0.165). There were, however, significant differences in histopathologic tumor subtype according to the patients’ ages. Among 24 patients older than 40 years of age, 17 (71%) had pleomorphic RMS. Among 22 patients 20 years of age or younger, only 5 (23%) had this histologic subtype (two-sided Fisher exact test, P ⫽ 0.001). There were no statistically significant differences in patient gender, primary tumor location, or tumor size according to histologic subtype. According to the American Joint Committee on Cancer (AJCC) staging system for soft tissue sarcomas,27 the disease was Stage IIB (T1, N0) in 26 patients, Stage III (T2b, N0) in 25 patients, and Stage IV (Any T, N1, M0) in 27 patients. A stage could not be assigned to four patients whose tumor size was not documented and who had no lymph node metastases
379
at presentation. According to the IRSG Presurgical Staging Classification,28 the tumors were grouped as follows: Stage 1, 25 patients; Stage II, 19 patients; and Stage III, 31 patients. An IRSG stage could not be assigned to eight cases. Seventy patients were seen for their first presentation of disease, whereas 12 had undergone one (8 patients), two (3 patients), or three (1 patient) previous resections and presented to MDACC for management of recurrent disease. On presentation to MDACC, 65 patients had evident gross tumor and 17 did not, having undergone resection immediately before referral. The therapeutic goal for all patients was complete surgical resection with sparing of both function and cosmesis where possible. Estimation of the adequacy of resection in the 17 patients with no gross disease was based on physical examination, review of pertinent radiologic, operative, and pathologic data, and communication with the operating surgeon. In 15 of these patients, resection could not be improved without significant functional deficits and radiation without additional surgery was the locoregional adjuvant therapy. In two cases, further resection was performed: in one, radiation preceded resection and in the other it followed resection. Of the 65 patients with gross disease, complete gross resection in 37 was judged to be impossible without significant morbidity or attempted resection failed to remove all gross tumor and these patients received radiation as the definitive locoregional treatment. Twenty-nine patients underwent complete gross resection at MDACC with postoperative (20 patients) or preoperative (9 patients) radiation. In summary, radiation was delivered preoperatively to 10 patients, postoperatively to 35 patients, and as definitive treatment to 37 patients. For the 45 patients undergoing definitive resection, the final resection margins were pathologically negative in 28 and positive or uncertain in 17. Combined therapy (surgery and radiation) was employed more frequently in patients with non– head and neck primary sites (28 of 39) compared with those with head and neck primary sites (17 of 43, two-sided Fisher exact test, P ⫽ 0.004). Combined therapy (surgery and radiation) was also employed more frequently in patients with extremity primary sites (14 of 18) compared with those with nonextremity primary sites (31 of 64, two-sided Fisher exact test, P ⫽ 0.03). Radiation therapy was delivered using techniques appropriate to each primary tumor site and regional lymph nodes where applicable. External beam (60Co or higher energy photons) radiation alone was used in 77 patients and was combined with low-dose rate brachytherapy in 5 patients. For preoperative radiation, the primary tumor with a margin of 5–7 cm was
380
CANCER July 15, 2002 / Volume 95 / Number 2
TABLE 1 Management of Regional Lymph Nodal Basins Modality
Elective treatment (N0)
Definitive treatment (N1)
Radiation alone Surgery alone Surgery and radiation None
18 3 4
23 1 3 30
irradiated to a median dose of 50 Gy (range, 45–91 Gy). Postoperative radiation was delivered to a significantly (Mann–Whitney, P ⫽ 0.005) higher median dose of 60 Gy (range, 40 – 80 Gy) using a shrinking field technique after 50 Gy. Patients treated with radiation alone for gross tumor received a median dose of 60 Gy (range, 41–78 Gy). The high doses in each dose range resulted from the additional interstitial radiation given to five patients; total doses in these patients ranged from 70 to 91 Gy. Management of regional lymph node basins is outlined in Table 1. All 27 patients with gross lymph node metastases had their disease either resected, irradiated, or both. The median lymph node radiation dose was 50 Gy (range, 40 –75 Gy), regardless of treatment sequencing and combination— elective doses did not differ significantly from definitive lymph node radiation doses. There was no correlation between the likelihood of elective lymph node treatment and primary tumor size, tumor site, or histopathologic subtype. There did not appear to be any consistent policy regarding elective lymph node treatment. Systemic chemotherapy was delivered to 58 of the 82 patients (71%). In 56 of these 58 patients (97%), the regimen contained either actinomycin-D or doxorubicin, generally combined with vincristine and cyclophosphamide. The timing of chemotherapy in patients with respect to local irradiation was as follows: preradiotherapy alone, 10; preradiotherapy and concurrent followed by postradiotherapy, 21; concurrent followed by postradiotherapy, 13; concurrent alone, 8; and postradiotherapy only, 6. Four additional patients received subtherapeutic doses of chemotherapy concurrently with radiation (methotrexate in one, vincristine in two, and hydroxyurea in one) without additional systemic therapy and were analyzed as if no chemotherapy had been given. There was no correlation between the use of chemotherapy and tumor size, tumor site, lymph node status, or histopathologic subtype. However, systemic therapy was given substantially more frequently to this group of patients than to sarcoma patients in general treated at MDACC during the same time period.29 –31 In summary, only 9 patients (11%) received single-
FIGURE 1. Kaplan–Meier overall survival and disease-free survival curves for all patients. Vertical ticks indicate censored observations. Vertical bar is the 95% confidence interval. Also shown is the age and gender-matched expected survival.
modality treatment with radiation; 43 (52%) received two modalities (radiation and chemotherapy, 28; radiation and surgery, 15); and 30 (37%) received all three modalities.
Disease Outcome and Patient Survival At the time of analysis, disease had recurred in 47 (57%) patients and 51 (62%) had died. The sites of recurrence were local sites in 16; regional lymph nodes in 12; and distant metastatic sites in 35. The sites of first recurrence were metastatic sites in 22, lymph nodes in 11, local in 11, local and metastatic in 2, and lymph nodes and metastatic sites in 1. The median time to first recurrence was 6.4 months (range, 1–52 months). The actuarial 5, 10, and 15-year overall survival (OS) rates were 44%, 41%, and 36%, respectively. The actuarial 5, 10, and 15-year disease-free survival (DFS) rates were 40% at those times (Fig. 1). Median survival time for all patients was 38 months. In univariate analysis, the OS and DFS rates were inferior for patients with large (⬎ 5 cm) primary tumors. Tumors 5 cm or smaller were associated with actuarial 10-year OS and DFS rates of 49% and 57%, respectively, compared with 35% (P ⫽ 0.06) and 26% (P ⫽ 0.003), respectively, for tumors larger than 5 cm. For DFS, there was a suggestion of an inferior outcome for patients with alveolar compared with embryonal or pleomorphic subtypes (27% vs. 44%, P ⫽ 0.07). On multivariate analysis, the only significant prognostic determinant of OS and DFS was tumor size. When the OS and DFS analysis was limited to the 56 patients receiving acti-
Adult Rhabdomyosarcoma/Little et al.
381
TABLE 2 Analysis of Factors Potentially Affecting 10-Year Outcome (Significant and Borderline Significant Differences are Highlighted) Factor Gender Male Female Age (yrs) ⱕ 28 ⬎ 28 Primary size (cm) ⱕ5 ⬎5 Nodal status Negative Positive Treatment delivered XRT alone Preoperative radiotherapy Postoperative radiotherapy Treatment delivered XRT alone Surgery/XRT Prior treatment No Yes Primary site Non-H&N H&N Histology Emb/Pleo Alveolar IRS pretreatment stage Stage 1 Stage 2 Stage 3 Unknown Systemic therapyb Yes No
n
OS (%)
P value
DFS (%)
P value
DMFS (%)
P value
LC (%)
P value
45 37
40 42
0.9
42 37
0.8
54 52
0.8
77 72
0.7
42 40
45 37
0.4
42 38
0.9
50 56
0.4
75 74
0.9
37 35
49 35
0.06
57 26
0.003a
74 32
0.0002a
80 72
0.4
55 27
42 38
0.5
38 45
0.8
53 52
0.5
73 78
0.7
37 10 35
43 41 40
0.7
39 34 42
0.9
52 34 59
0.3
68 100 75
0.1
37 45
43 40
0.9
39 40
0.7
52 54
0.5
68 80
0.08
70 12
38 57
0.6
39 47
0.6
50 67
0.6
77 64
0.4
39 43
41 35
0.8
42 38
0.9
46 58
0.2
89 64
0.02a
63 19
44 28
0.2
44 27
0.07
58 32
0.04
76 71
0.6
25 18 31 8
40 49 37
0.9
51 44 30
0.4
53 70 42
0.4
87 59 77
0.5
56 26
47 31
0.1
45 30
0.1
59 41
0.1
80 63
0.2
XRT: radiotherapy; H&N: head and neck; emb/pleo: embryonal and pleomorphic subtypes; IRS: Intergroup Rhabdomyosarcoma Studies; OS: overall survival; DFS: disease-free survival; DMFS: metastasis-free survival; LC: local control. a Remained significant on multivariate analysis. b Actinomycin-D or doxorubicin containing regimens only.
nomycin-D or doxorubicin-containing chemotherapy regimens, the prognostic significance of large tumors remained (OS: ⱕ 5 cm, 59% vs. ⬎ 5 cm, 40%, P ⫽ 0.05; DFS: ⱕ 5 cm, 63% vs. ⬎ 5 cm, 29%, P ⫽ 0.01). Patient gender, age, lymph node status, primary versus recurrent presentation, tumor location, AJCC stage, IRSG stage, treatment (surgery and radiation vs. radiation alone), use of chemotherapy, and resection margin status did not correlate significantly with either OS or DFS (Table 2). The actuarial 5, 10, and 15-year distant metastasis-free survival (DMFS) rates were 55%, 53%, and 53%, respectively (Fig. 2). Median time to metastatic recurrence was 6.5 months (range, 1–72 months). Sites
of metastatic recurrence among the 35 patients who developed metastases were as follows: lung, 24; bone, 11; brain, 1; meningeal, 1; and other (including distant lymph nodes, bone marrow, soft tissues), 6 (several patients had more than one site of metastasis). Only one patient developed meningeal recurrence— he had a primary parameningeal site. Univariate analysis of factors potentially correlated with DMFS (Table 2) revealed that tumors larger than 5 cm were associated with a 10-year DMFS rate of 32% compared with 74% for tumors 5 cm or smaller (P ⬍ 0.001). Tumors exceeding 10 cm were associated with a particularly poor outcome (Fig. 3). Alveolar subtype was also associated with significantly poorer DMFS rates than embryonal
382
CANCER July 15, 2002 / Volume 95 / Number 2
FIGURE 2. Actuarial lymph node, local, and metastatic control curves for all patients. Vertical bars indicate 95% confidence intervals.
FIGURE 3.
Freedom-from-metastasis curves according to primary tumor
size.
or pleomorphic subtypes, with 10-year DMFS rates of 32% vs. 58%, respectively (P ⫽ 0.041). In multivariate analysis, the only significant determinant of DMFS was tumor size. When the DMFS analysis was limited to the 56 patients receiving actinomycin-D or doxorubicin-containing chemotherapy regimens, the prognostic significance of large tumors remained (DMFS: ⱕ 5 cm, 81% vs. ⬎ 5 cm, 40%, P ⫽ 0.002). The actuarial 5, 10, and 15-year local control rates were 75%, 75%, and 75% (Fig. 2). The median time to local failure was 10.5 months (range, 1–51 months). The site of local failure, although difficult in some cases to ascertain with certainty, appeared to be within the radiation field in all but one patient, in whom recurrence was marginal. Analysis of factors potentially correlated with local control (Table 2) re-
FIGURE 4. Actuarial local control curves according to site of primary tumor. Numbers in parentheses represent the number of patients in each group.
vealed that tumors of the head and neck had a significantly poorer 10-year local control rate (64%) than tumors at other sites (89%; P ⫽ 0.02). As illustrated in Figure 4, the poor local control for head and neck tumors was due entirely to a very poor control rate of parameningeal tumors (50% at 10 years). Nonparameningeal head and neck tumors had a 10-year local control rate of 87%, similar to the 89% rate for non– head and neck primary tumors. The difference in local control between parameningeal and all other sites was highly significant, 50% versus 83% at 10 years (P ⫽ 0.003). The use of radiation alone for gross disease appeared to decrease the likelihood of local control with a 10-year local control rate of 68% for radiation alone compared with 80% for combined surgery and radiation (P ⫽ 0.08). No other factor correlated with local control. In multivariate analysis, only parameningeal tumor location correlated with poorer local control (P ⫽ 0.003). Analysis of the potential influence of radiation parameters on local control was hampered by the relatively small number of events. Forty-five patients were treated with combined gross total resection and radiation. For all nonparameningeal sites treated with combined gross total resection and radiation, the 10-year local control rate was 91% with only 2 of 38 patients developing local recurrence. In this group of 38 patients, 10 received preoperative radiation (median dose, 50 Gy) and none developed local failure whereas 2 of 28 receiving postoperative radiation developed local recurrence. The two who experienced local recurrence in this group had embryonal tumors and received more than 60 Gy. Of the 26 in this group who did not develop local recurrence, 15 received less than 60 Gy. It was not possible to develop any dose-
Adult Rhabdomyosarcoma/Little et al.
response relation for radiation when combined with gross total resection, except that 25 patients whose disease was controlled did receive doses less than 60 Gy. (range, 45–58 Gy). Only seven patients with parameningeal tumors were treated with gross total resection and radiation, four of whom developed local recurrence. These four received a median dose of 57 Gy, whereas the three whose disease was controlled received a median dose of 54 Gy. Thirty-seven patients were treated with radiation alone for gross disease. For nonparameningeal sites treated in this manner, the 10-year local control rate was 76%, with 4 of 19 developing local recurrence. There were no significant differences in dose between recurring and controlled tumors. However, all four patients whose disease recurred locally received 60 Gy or less, whereas 11 of 15 whose disease was controlled received doses of 60 –70 Gy. For parameningeal sites treated with definitive radiation, the 10-year local control rate was 60%, with 6 of 18 patients developing local recurrence. Again, there were no statistically significant differences in dose between recurring and controlled tumors. However, only four patients in this group received doses of 60 – 63 Gy. The actuarial 5, 10, and 15-year lymph node control rates were 82% at each of the time intervals. The median time to lymph node failure was 5.2 months (range, 1–22 months). Univariate analysis did not reveal any significant correlation between lymph node recurrence and initial lymph node status, tumor histology, tumor site, tumor size, elective treatment of regional lymph node basins, or modality of regional lymph node treatment. Lymph node-negative patients receiving no adjuvant nodal treatment had a 10-year lymph node control rate of 75%, compared with 91% in those receiving such treatment (P ⫽ 0.244). Lymph node-positive patients, all of whom received lymph node treatment, had a 10-year lymph node control rate of 80%. Lymph node control did not appear to correlate with lymph node radiation dose. Whether radiation was delivered electively or for gross lymph node disease, there were no significant differences in lymph node control for doses of 50 Gy or less versus more than 50 Gy. However, the dose range employed was narrow with few cases irradiated to less than 45 or more than 55 Gy to their lymph node basins. Overall, 35 of the 82 patients (43%) encountered lymph node disease either at presentation or as recurrence or as both. There was a suggestion that tumors of alveolar type had a higher propensity for lymph node metastasis than either pleomorphic or embryonal types. Overall lymph node metastasis was evident in 12 of 19 (63%) alveolar tumors and in 23 of 63 (37%) combined
383
FIGURE 5. Survival after first relapse. Median survival time after relapse was 7 months.
embryonal-pleomorphic tumors (two-tailed Fisher exact test, P ⫽ 0.063). In 27 patients, preradiotherapy chemotherapy was delivered and response data were documented. Within this subgroup, there were 6 (22%) complete responses (CR), 14 (52%) partial responses (PR), and 5 nonresponses (NR). In addition, two patients had progression of disease (PD). The overall response rate was 74%. There were significantly better 5-year DMFS and local control rates among responders compared with nonresponders (DMFS:72% vs. 19%, P ⫽ 0.004; local control: 77% vs. 27%, P ⫽ 0.03). Twenty-five patients received concurrent chemotherapy and radiation for gross tumor and were evaluable for response. In this group, there were 17 CRs (68%), 5 PRs (20%), and 3 NRs (12%), for an overall response rate of 88%. In addition, 13 of 16 patients received radiotherapy alone for gross disease. Their CR rate was 81%. With respect to the number of treatment modalities used, there were no significant differences in any outcome measure between those receiving two or three modalities (data not shown). However, the nine patients treated with radiation alone (no chemotherapy or surgery) appeared to experience a particularly poor outcome as six developed recurring disease (four cases of metastasis, three of local recurrence, and one of lymph node recurrence). Their 10-year survival rate was 22% compared with 43% for those treated with at least two modalities (P ⫽ 0.007). Survival after recurrence was poor (Fig. 5) with a median survival time of 7 months and an actuarial 3-year survival rate of 10%. Death after recurrence appeared to be rapid in all patients and tumor categories (Table 2) and no significant prognostic factors
384
CANCER July 15, 2002 / Volume 95 / Number 2
TABLE 3 Treatment-Related Complicationsa Complication
Grade 1
Grade 2
Grade 3
Grade 4
TABLE 4 Reported Five-Year Survival Rates for Nonmetastatic Adult Rhabdomyosarcoma and for Children with Nonmetastatic Disease in the IRSG Studies
Necrosis Fracture Edema Osteonecrosis Enteritis Infection
2
1
1 2 1 1
1b
No. of patients
1 1
1c
a
Grade 1: requiring no treatment; Grade 2: requiring medical treatment; Grade 3: requiring surgical treatment; Grade 4: fatal. b Brain necrosis and death related to radiation therapy. c Postoperative sepsis and death related to surgery.
for postrecurrence survival were found. Death after isolated local recurrence was as likely as after other sites of initial recurrence (3-year survival rate of 18% vs. 7%, P ⫽ 0.44). There were 12 treatment-related complications, including two deaths (Table 3). One patient died after surgery secondary to massive sepsis and one patient died of radiation-induced brain necrosis after treatment for an extensive orbital tumor. The 5, 10, and 15-year actuarial radiation-related complication rates were 19%, 19%, and 24%, respectively. The median time to development of a radiation complication was 13.6 months (range, 2–127 months). The only factor displaying any correlation with the likelihood of a radiation-related complication was radiation dose. For doses exceeding 60 Gy, the 15-year complication rate was 29% compared with 11% for doses of 60 Gy or less (P ⫽ 0.070). Other factors, including tumor size, tumor location, sequence of surgery and radiation, the use of adjuvant chemotherapy, and the use of concurrent radiation and chemotherapy, did not correlate with the incidence of radiation-related complications.
DISCUSSION The results of our study are comparable to those reported by others for adult patients with RMS (Table 4), i.e., disease outcome is considerably less favorable for adults than for children. Apart from its poorer response to treatment, RMS in adults also appears to differ from childhood RMS in several other features. Although head and neck sites are common in children where they account for some 35% of cases, the 50% incidence in our study appear higher. There are few other studies on adult disease with which to compare this incidence. In one recent report by Hawkins et al.,14 head and neck sites accounted for only 24% of sites in patients. These differences may reflect referral patterns. However, most reports on adult disease
Adults 113 28 10 12 47 26 82 Children 557 827 912
5-yr survival (%)
Report
26 ⬃18a ⬃40a 36 ⬍ 50b 44 44
Ariel and Briceno10 Lloyd et al.11 Prestidge and Donaldson12 Seidel et al.13 Hawkins et al.14 Esnaola et al.15 Present study
63 71 79
Maurer et al. (IRS-I)5 Maurer et al. (IRS-II)5 Crist et al. (IRS-III)6
IRSG: Intergroup Rhabdomyosarcoma Study Group; IRS-I, IRS-II, IRS-III: successive IRSG protocols. a Approximate estimate from publication. b Exact data not given, but survival less than 50%.
agree that orbital presentations are distinctly uncommon.11,13,14,18,32 There were only two (2%) such cases in our series. Orbital presentations are among the most favorable for childhood RMS, as demonstrated in all four IRSG studies4 –7 and the relative paucity of these among adult patients might partly explain their worse overall prognosis. In IRS-I, II, and III, orbital tumors were associated with 5-year survival rates exceeding 90%, whereas 5-year survival rates for other sites were between approximately 50% and 88%.4 – 6 This prognostic difference was also evident in IRS-IV where orbital disease was associated with a long-term survival rate of approximately 95%, but nonorbital sites gave survival rates of 60 – 85%.7 As an unfavorable compensation for the paucity of orbital tumors, adults appear to present with more parameningeal tumors. In our series, parameningeal disease was present in 30%, which is higher than the 14 –18% incidence reported by the IRSG.4,5 At least one other series of adult head and neck RMS also noted a high proportion of this unfavorable anatomic site.33 The concept of parameningeal RMS has undergone some vicissitudes since it was first introduced in 1978 following an observation during IRS-I that children with RMS arising in the paranasal sinuses, nasopharynx, and the infratemporal-pterygopalatine region had a 35% incidence of meningeal recurrence following treatment.24 The prognosis for such presentations was believed to be inferior to that associated with other head and neck nonorbital sites.25 There followed recommendations for cranial radiation and intrathecal therapy for children with parameningeal tumors if there was evidence of base of skull or cranial nerve
Adult Rhabdomyosarcoma/Little et al.
invasion.34,35 However, it subsequently became evident that the radiation portals utilized in IRS-I for head and neck presentations were not wide enough to cover disease extension. Providing that the primary tumor is covered with radiation portal margins of at least 2 cm around gross disease, meningeal recurrence was uncommon (⬍5%).36,37 Later IRSG reports indicated that meningeal recurrence is not a common event, but the prognosis for parameningeal RMS remains relatively poor with 5-year survival rates of approximately 70 –75%5–7 and the parameningeal site is classified as unfavorable in the current IRS-V protocol.3 The major recurrence pattern for children with RMS in this site is local,38 as it was in our adult series. Another significant difference between childhood and adult RMS is the proportion of histopathologic subtypes. In IRS-IV, the relative subtype proportions were embryonal, 70%; alveolar, 20%; and others, 10%.7 In our series, the proportion of embryonal subtype was only 34%, counterbalanced by pleomorphic tumors which made up 43%. The relative paucity of embryonal tumors and the relative increase in alveolar and pleomorphic types are characteristic of all adult RMS series.10,14 –16,18 Although the recognition and subclassification of RMS remains difficult with significant disagreement even among expert pathologists,39 – 41 the use of immunohistochemical techniques, as in this study, increases objectivity.20 Recognizing the advantageous outcome associated with the embryonal subtype, it is classified in IRSG protocols as favorable in contrast to alveolar and other types.3 The imbalance in distribution of favorable versus unfavorable histotype between children and adults may place adults at a disadvantage. However, the prognostic significance of RMS subtype in adults is controversial. In our series, patients with alveolar tumors experienced significantly poorer metastasis-free rates than those with embryonal or pleomorphic tumors. This difference did not translate into significant disease-free or survival differences. In addition, it appeared that the adverse effect of alveolar histology was mediated through a larger tumor size because there was no effect of histology on metastasis-free survival when size was included in the multivariate model. In one recent adult RMS series,15 there was a trend toward poorer survival for those with alveolar histology, but this was not significant. In another similar study,14 histologic type was not predictive of outcome. The prognostic significance of RMS subtype in adults remains uncertain, but is unlikely to be a major factor. To what extent these differences in tumor characteristics between children and adults with RMS account for the poorer prognosis of the latter remains a matter for speculation.
385
The most common pattern of disease recurrence in adults with RMS is distant metastasis10,14,15 (Fig. 2), most often to the lungs. In our series and in other adult series,14,15 the major determinant of metastatic propensity was tumor size as specified by the AJCC TNM system (ⱕ5 cm vs. ⬎5 cm). Moreover, tumors exceeding 10 cm had a particularly high metastatic rate. In our study and in one other,15 all patients with primary lesions in this size range were predicted to develop metastasis. The group and stage classifications employed by the IRSG6,7,28,41 did not appear to be useful for our patients. Chemotherapy as used in this series and in two other recent adult series14,15 did not satisfactorily abrogate metastases. Nevertheless, there is evidence from our data as well as from another study15 that adult RMS is chemosensitive with objective tumor response observed in approximately 75% of patients. This is similar to the response rates reported by IRSG for children.5,6 Disease that responded to chemotherapy was associated with a significantly better freedomfrom-metastasis than disease that did not, an observation also made by others.15 Although this phenomenon might simply reflect selection of favorable tumors, it does encourage the continued investigation of multiagent chemotherapy for adult RMS. The IRSG recently reported favorable results in children with metastatic disease, whose long-term outcome is also poor. They were treated with a combination of standard vincristine, actinomycin-D, and cyclophosphamide combined with ifosfamide and etoposide.42 The high metastatic rate and chemoresponsiveness of this sarcoma mandate the continuing investigation of multiagent chemotherapy in adults with this disease. Whereas local tumor control was satisfactory for lesions arising in most locations (87% at 10 years), the control rate for parameningeal tumors was poor (50% at 10 years). There is little published adult data with which to compare our results. One study on head and neck RMS, including a high proportion of parameningeal sites, reported that only 2 of 12 patients were free of tumor at variable follow-up times.33 Our patients did not experience a significant rate of meningeal dissemination; the problem was one of local recurrence, as also noted in recent IRSG studies.3,7,38 The retrospective nature of our analysis and small dataset precluded a clear definition of the reasons for, and solutions to, this problem. Tumors arising in these sites are rarely resectable with clear margins and local control depends on delivering adequate radiation doses to an appropriate tumor volume while respecting the radiation tolerance of numerous nearby critical normal structures. The local anatomic distribution
386
CANCER July 15, 2002 / Volume 95 / Number 2
of parameningeal tumors tends to be highly complex and cannot be evaluated adequately without sophisticated imaging techniques. Most of the patients in our study were treated before the availability of sophisticated tumor imaging and radiation delivery systems such as intensity-modulated therapy. This likely accounts for the relatively low doses delivered to our patients with gross tumors in this site (40 – 63 Gy; median, 60 Gy). Although we were unable to define a dose-response relationship from our data, general principles as well as the high recurrence rate suggest that these tumors were not irradiated adequately. Utilizing current imaging modalities, treatment planning facilities, and radiation delivery technologies, it may now be possible to deliver higher doses (66 –70 Gy) to gross disease and spare critical normal structures. For this presentation, it may also be advantageous to employ the IRSG strategy of utilizing radiation early in the treatment course with appropriate concomitant chemotherapy.3,26 For sites other than parameningeal sites, the overall local control was satisfactory. Our data did not reveal any significant prognostic factors for local control other than the parameningeal site. In particular, primary tumor size was not significant. This finding was consistent with that for other adult soft tissue sarcomas,29 –31 but was at variance with IRSG results where increasing tumor size correlated with an increased risk of local recurrence.43 Although our data permit no definitive conclusions on the relation between radiation dose and local control, we believe that the following statements are consistent with our findings. Adult RMS is somewhat more radiosensitive than most adult sarcomas where a control rate of 68% for gross disease would be unexpected. However, the disease in adults is not as radiosensitive as in children where doses in the range 40 –55 Gy result in local control in approximately 80% of patients treated for gross disease (IRS-III).43 We believe that optimal local control is achieved with a combination of gross total resection and local radiation. For completely resected disease with negative margins, a dose of 50 –56 Gy would seem adequate. For margin-positive disease, it is likely that a higher dose of 60 Gy is appropriate. If gross tumor cannot be resected completely, then radiation to a dose of 66 –70 Gy should be employed utilizing modern field-defining technologies. In children, IRS-IV failed to show any improvement in local control when hyperfractionated radiation (59.4 Gy in 1.1-Gy fractions twice daily) was compared with conventionally fractionated therapy (50.4 Gy in 1.8-Gy fractions daily).7 Whether radiation can be withheld for completely resected tumors as recommended by the IRSG for their Group I embryonal case remains uncertain in adults. The two
randomized trials on adult sarcoma in general44,45 found that the addition of radiation resulted in significant improvements in local control over surgery alone. Also, there are data from the International Society of Pediatric Oncology that children with incompletely resected or inoperable RMS achieve complete remission of disease with chemotherapy. These children, in whom local remission is confirmed by the results of more than one biopsy and who receive no further local treatment, experience a local recurrence rate of 50%.46 Some form of local treatment in addition to chemotherapy, regardless of apparent complete response, should be administered. Finally, we consider the question of regional lymph node disease. RMS is one the few soft tissue sarcomas with a known propensity for regional lymph node metastasis.47 In our series, 33% of patients had clinical evidence of lymph node disease at presentation. Another adult series reported a 46% incidence of lymph node involvement.15 These frequencies are substantially higher than those reported for children in whom typical figures are 15–20%.7,9,28,48,49 In children, lymph node metastasis is uncommon in head and neck sites. In our data there was no significant correlation between primary site and the likelihood of lymph node metastases. When lymph node disease was calculated as involvement at any time, either at presentation or at recurrence, the overall incidence of lymph node metastasis was 43%. The evidence suggests that alveolar tumors were associated with a higher risk of lymph node metastasis (overall incidence, 63%) than embryonal-pleomorphic types (overall incidence, 37%). In children, lymph node status is an important determinant of outcome.9,28,48,49 In adults, unexpectedly, it has not been shown to be a significant prognostic factor. In our series, the DFS, DMFS, and OS rates did not correlate with initial lymph node status. Similar findings were reported by the only other adult series that addresses this issue.15 The reasons for this discrepancy are not apparent, but may result from the overall poor disease outcome in adults. However, the high incidence of lymph node disease suggests that regional treatment should be a component of the overall treatment strategy, at least for patients with alveolar tumors. Patients with evident lymph node disease are likely best managed along the same lines that we recommend for the treatment of the primary tumor. Although our data did not reveal any benefit from elective lymph node treatment, this may reflect the fact that no consistent policy was followed. We currently recommend elective lymph node radiation to a dose of approximately 50 Gy for all patients with alveolar RMS. In conclusion, adult RMS is a highly malignant
Adult Rhabdomyosarcoma/Little et al.
tumor with a poor prognosis. Metastatic recurrence is a major finding and the development of effective multiagent chemotherapy is an urgent priority. The optimal treatment remains undefined, but multimodality approaches combining surgical resection, radiation, and systemic chemotherapy need further investigation. Local control of tumors arising in most sites is satisfactory following surgical resection and radiation. Tumors arising in head and neck parameningeal sites pose significant problems for local control. Early irradiation using advanced conformal radiation techniques with appropriate tumor imaging need to be employed to achieve acceptable local control. Lymph node metastasis is a significant risk, particularly for tumors of alveolar histology and elective lymph node treatment should be considered. Because salvage therapy after recurrence was unsuccessful in most cases, aggressive initial therapy is necessary if the patient is to have any realistic chance for cure.
REFERENCES 1.
2.
3.
4.
5. 6. 7.
8.
9.
10.
11. 12.
13.
Russell WO, Cohen HJ, Enzinger FM, et al. A clinical and pathological staging system for soft tissue sarcomas. Cancer. 1977;40:1562–1570. Young JL, Jr., Ries LG, Silverberg E, Horm JW, Miller RW. Cancer incidence, survival, and mortality for children younger than age 15 years. Cancer. 1986;58:598 – 602. Raney RB, Maurer HM, Anderson JR, et al. The Intergroup Rhabdomyosarcoma Study Group (IRSG): major lessons from the IRS-I through IRS-IV studies as background for the current IRS-V treatment protocols. Sarcoma. 2001;5:9 –15. Maurer HM, Beltangady M, Gehan EA, et al. The Intergroup Rhabdomyosarcoma Study-I. A final report. Cancer. 1988;61: 209 –220. Maurer HM, Gehan EA, Beltangady M, et al. The Intergroup Rhabdomyosarcoma Study-II. Cancer. 1993;71:1904 –1922. Crist W, Gehan EA, Ragab AH, et al. The Third Intergroup Rhabdomyosarcoma Study. J Clin Oncol. 1995;13:610 – 630. Crist WM, Anderson JR, Meza JL, et al. Intergroup Rhabdomyosarcoma Study-IV: results for patients with nonmetastatic disease. J Clin Oncol. 2001;19:3091–3102. Flamant F, Rodary C, Rey A, et al. Treatment of non-metastatic rhabdomyosarcomas in childhood and adolescence: results of the second study of the International Society of Paediatric Oncology—MMT84. Eur J Cancer. 1998;34:1050 – 1062. Rodary C, Rey A, Olive D, et al. Prognostic factors in 281 children with nonmetastatic rhabdomyosarcoma (RMS) at diagnosis. Med Pediatr Oncol. 1988;16:71–77. Ariel IM, Briceno M. Rhabdomyosarcoma of the extremities and trunk: analysis of 150 patients treated by surgical resection. J Surg Oncol. 1975;7:269 –287. Lloyd RV, Hajdu SI, Knapper WH. Embryonal rhabdomyosarcoma in adults. Cancer. 1983;51:557–565. Prestidge BR, Donaldson SS. Treatment results among adults with childhood tumors: a 20-year experience. Int J Radiat Oncol Biol Phys. 1989;17:507–514. Seidel T, Kinblom L-G, Angervall L. Rhabdomyosarcoma in middle-aged and elderly individuals. APMIS. 1989;97:236 – 248.
387
14. Hawkins WG, Hoos A, Antonescu CR, et al. Clinicopathologic analysis of patients with adult rhabdomyosarcoma. Cancer. 2001;91:794 – 803. 15. Esnaola NF, Rubin BP, Baldini EH, et al. Response to chemotherapy and predictors of survival in adult rhabdomyosarcoma. Ann Surg. 2001;234:215–223. 16. La Quaglia MP, Heller G, Ghavimi F, et al. The effect of age at diagnosis on outcome in rhabdomyosarcoma. Cancer. 1994;73:109 –117. 17. Miettinen M. Rhabdomyosarcoma in patients older than 40 years of age. Cancer. 1988;62:2060 –2065. 18. Schurch W, Begin LR, Seemayer TA, et al. Pleomorphic soft tissue myogenic sarcomas of adulthood. A reappraisal in the mid-1990s. Am J Surg Pathol. 1996;20:131–147. 19. Asmar L, Gehan EA, Newton WA, et al. Agreement among and within groups of pathologists in the classification of rhabdomyosarcoma and related childhood sarcomas. Report of an international study of four pathology classifications. Cancer. 1994;74:2579 –2588. 20. Weiss SW, Goldblum JR, Enzinger and Weiss’s soft tissue tumors (2nd edition). St. Louis: Mosby, 2001:785– 835. 21. Harris EK, Albert A. Survivorship analysis for clinical studies. New York: Marcel Dekker, 1991:5–125. 22. U.S. Department of Health and Human Services. Vital statistics of the United States. Hattsville, MD: U.S. Public Health Service, National Center for Health Statistics, 1988. 23. Altman DG. Practical statistics for medical research. London: Chapman and Hall, 1991. 24. Tefft M, Fernandez C, Donaldson M, Newton W, Moon TE. Incidence of meningeal involvement by rhabdomyosarcoma of the head and neck in children: a report of the Intergroup Rhabdomyosarcoma Study (IRS). Cancer. 1978;42:253–258. 25. Wharam MD, Foulkes MA, Lawrence W, et al. Soft tissue sarcoma of the head and neck in childhood: nonorbital and nonparameningeal sites. A report of the Intergroup Rhabdomyosarcoma Study (IRS)-1. Cancer. 1984;53:1016 –1019. 26. Benk V, Rodary C, Donaldson SS, et al. Parameningeal rhabdomyosarcoma: results of an international workshop. Int J Radiat Oncol Biol Phys. 1996;36:533–540. 27. Fleming ID, Cooper JS, Henson GE, et al. AJCC cancer staging manual. (5th edition) Philadelphia: Lippincott-Raven, 1997:149 –156. 28. Lawrence W, Jr., Anderson JR, Gehan EA, Maurer H. Pretreatment TNM staging of childhood rhabdomyosarcoma: a report of the Intergroup Rhabdomyosarcoma Study Group. Children’s Cancer Study Group. Pediatric Oncology Group. Cancer. 1997;80:1165–1170. 29. Zagars GK, Mullen JR, Pollack A. Malignant fibrous histiocytoma: outcome and prognostic factors following conservation surgery and radiotherapy. Int J Radiat Oncol Biol Phys. 1996;34:983–994. 30. Mullen JR, Zagars GK. Synovial sarcoma outcome following conservation surgery and radiotherapy. Radiother Oncol. 1994;33:23–30. 31. Fleming JB, Berman RS, Cheng S-U, Hunt K, Feig BW, et al. Long-term outcome of patients with American Joint Committee on Cancer stage IIB extremity soft tissue sarcomas. J Clin Oncol. 1999;17:2772–2780. 32. Gaffney EF, Dervan PA, Fletcher CD. Pleomorphic rhabdomyosarcoma in adulthood. Analysis of 11 cases with definition of diagnostic criteria. Am J Surg Pathol. 1993;17:601– 609.
388
CANCER July 15, 2002 / Volume 95 / Number 2
33. Nakhleh RE, Swanson PE, Dehner LP. Juvenile (embryonal and alveolar) rhabdomyosarcoma of the head and neck in adults. A clinical, pathologic, and immunohistochemical study of 12 cases. Cancer. 1991;67:1019 –1024. 34. Berry MP, Jenkin RD. Parameningeal rhabdomyosarcoma in the young. Cancer. 1981;48:281–288. 35. Gerson JM, Jaffe N, Donaldson MH, Tefft M. Meningeal seeding from rhabdomyosarcoma of the head and neck with base of the skull invasion: recognition of the clinical evolution and suggestions for management. Med Pediatr Oncol. 1987;5:137–144. 36. Chan RC, Sutow WW, Lindberg RD. Parameningeal rhabdomyosarcoma. Radiology. 1979;131:211–214. 37. Gasparini M, Lombardi F, Gianni MC, Massimino M, Gandola L, Fossati-Bellani F. Questionable role of CNS radioprophylaxis in the therapeutic management of childhood rhabdomyosarcoma with meningeal extension. J Clin Oncol. 1990;8:1854 –1857. 38. Benk V, Rodary C, Donaldson SS, et al. Parameningeal rhabdomyosarcoma: results of an international workshop. Int J Radiat Oncol Biol Phys. 1996;36:533–540. 39. Caillaud JM, Gerard-Marchant R, Marsden HB, et al. Histopathological classification of childhood rhabdomyosarcoma: a report from the International Society of Pediatric Oncology pathology panel. Med Pediatr Oncol. 1989;17:391– 400. 40. Tsokos M, Webber BL, Parham DM, et al. Rhabdomyosarcoma. A new classification scheme related to prognosis. Arch Pathol Lab Med. 1992;116:847– 855. 41. Newton WA, Jr., Gehan EA, Webber BL, et al. Classification of rhabdomyosarcomas and related sarcomas. Pathologic aspects and proposal for a new classification—an Intergroup Rhabdomyosarcoma Study. Cancer. 1995;76:1073– 1085. 42. Breitfeld PP, Lyden E, Raney BR, et al. Ifosfamide and eto-
43.
44.
45.
46.
47. 48.
49.
poside are superior to vincristine and melphalan for pediatric metastatic rhabdomyosarcoma when administered with irradiation and combination chemotherapy: a report from the Intergroup Rhabdomyosarcoma Study Group. Am J Pediatr Hematol Oncol. 2001;23:225–233. Wharam MD, Hanfelt JJ, Tefft MC, et al. Radiation therapy for rhabdomyosarcoma: local failure risk for clinical group III patients on Intergroup Rhabdomyosarcoma Study II. Int J Radiat Oncol Biol Phys. 1997;38:797– 804. Pisters PWT, Harrison LB, Leung DHY, Casper ES, Brennan MF. Long-term results of a prospective randomized trial of adjuvant brachytherapy in soft tissue sarcoma. J Clin Oncol. 1996;14:859 – 868. Yang JC, Chang AE, Baker AR, et al. Randomized prospective study of the benefit of adjuvant radiation therapy in the treatment of soft tissue sarcomas of the extremity. J Clin Oncol. 1998;16:197–203. Godzinski J, Flamant F, Rey A, Praquin M-T, Martelli H. Value of postchemotherapy bioptical verification of complete remission in previously incompletely resected (Stage I and II pT3) malignant mesenchymal tumors in children: International Society of Pediatric Oncology 1984 Malignant Mesenchymal Tumor Study. Med Pediatr Oncol. 1994;22:22– 26. Mazeron J-J, Suit HD. Lymph node metastases from sarcomas of soft tissue. Cancer. 1987;60:1800 –1808. Lawrence W, Hays DM, Heyn R, et al. Lymphatic metastases with childhood rhabdomyosarcoma. A report from the Intergroup Rhabdomyosarcoma Study. Cancer. 1987;60:910 – 915. Mandell L, Ghavimi F, LaQuaglia M, Exelby P. Prognostic significance of regional lymph node involvement in childhood extremity rhabdomyosarcoma. Med Pediatr Oncol. 1990;18:466 – 471.
