- © 1999 by American Society of Clinical Oncology
Conservative Treatment for Girls With Nonmetastatic Rhabdomyosarcoma of the Genital Tract: A Report From the Study Committee of the International Society of Pediatric Oncology
- Hélène Martelli,
- Odile Oberlin,
- Annie Rey,
- Jan Godzinski,
- Richard D. Spicer,
- Nathalie Bouvet,
- Christine Haie-Meder,
- Marie-Josée Terrier-Lacombe,
- José Sanchez de Toledo,
- David Spooner,
- Danielle Sommelet,
- Françoise Flamant and
- Michaël C.G. Stevens
- From the Hôpital Necker–Enfants Malades, Paris; Institut Gustave Roussy, Villejuif; and CHU Vandœuvre, Nancy, France; Marciniak Hospital, Wroclaw, Poland; Royal Hospital for Sick Children, Bristol, and Children's Hospital, Birmingham, United Kingdom; and Hospital Materno-Infantil Vall Hebron, Barcelona, Spain.
- Address reprint requests to Hélène Martelli, MD, Service de Chirurgie Pédiatrique, Hôpital Necker–Enfants Malades, 149, rue de Sèvres, 75743 Paris Cedex 15, France; email helene.martelli{at}nck.ap-hop -paris.fr.
Abstract
PURPOSE: To report the results of a conservative multimodal approach in girls with nonmetastatic rhabdomyosarcoma (RMS) of the genital tract, treated in International Society of Pediatric Oncology (SIOP) Malignant Mesenchymal Tumors 84 and 89 protocols.
PATIENTS AND METHODS: From 1984 to 1994, 38 girls with RMS of the genital tract (vulva, vagina, uterus) were treated in SIOP protocols. With the exception of patients with rare small tumors, which were resected at the start of the studies, all patients received initial chemotherapy (CHT) (ifosfamide, vincristine, and actinomycin D). Local treatment including surgery, brachytherapy (BT), and external-beam radiotherapy (ERT) was given only to girls who did not achieve complete remission (CR) with CHT or who subsequently relapsed.
RESULTS: The primary tumor originated in the vulva or vagina in 27 girls and in the uterus in 11. The overall survival rate (± SE) was 91% ± 6% at 5 years, and the event-free survival rate was 78% ± 7%. At a median follow-up of 5 years, 30 girls were alive and in first CR and five were alive and in second CR. Four patients treated with complete resection of the tumor at diagnosis received less CHT. Thirteen patients were treated with CHT alone. In 17 patients, local treatment was necessary to achieve complete local control, for a residual mass after initial CHT (10 patients), for viable tumor on biopsy (three patients), or for local relapse (four patients). The local treatment used was radiotherapy (RT) (ERT in three patients, BT in seven), radical surgery with uterine ablation (three patients), RT and radical surgery (three patients), and conservative surgery with RT (one patient).
CONCLUSION: Girls with nonmetastatic RMS of the genital tract have an excellent prognosis. We found no difference in outcome between uterine and vulvovaginal RMS. Local treatment does not seem necessary in patients who have a complete response to CHT. When a local treatment is needed, BT may be an alternative to radical surgery or ERT.
RHABDOMYOSARCOMA (RMS) of the female genital tract is recognized as one of the most curable forms of RMS,1,2 and because of the efficacy of multiple-agent chemotherapy (CHT), the emphasis on radical surgery has been considerably reduced over the past decade. Approaches to local tumor control, including conservative surgery or radiation therapy (RT) (brachytherapy [BT] or localized external-beam radiotherapy [ERT]), have been developed in an attempt to limit long-term sequelae without jeopardizing survival.3-5 In this article, we review the experience of the International Society of Pediatric Oncology (SIOP) Malignant Mesenchymal Tumor (MMT) Committee in the management of RMS of the vulva, vagina, and uterus in girls.
PATIENTS AND METHODS
All patients whose data were examined were girls with nonmetastatic primary tumors of the genital tract who were included in the MMT 84 (1984 to 1988) and MMT 89 (1989 to 1994) studies. Histopathology of all patients was centrally reviewed by an international panel of pathologists and classified according to the new international classification as botryoid, embryonal, or alveolar.6 Distinctions between primary vulvovaginal and uterine sites were made by (1) initial examination (clinical examination, examination with the patient under anesthesia, and radiologic investigations); (2) posttherapy observations, including vaginoscopic findings and observations after surgical procedures; and (3) histologic assessment of surgical specimens obtained at any time.
Disease was staged according to the SIOP–International Union Against Cancer (UICC) tumor-node-metastasis (TNM) staging classification7 (Table 1). Treatment schedules used in the MMT 843 and MMT 89 studies are shown in Figs 1, 2, 3, and 4 . The standard first-line CHT in both studies was combination therapy with ifosfamide, vincristine, and actinomycin D (IVA), but the dosage of ifosfamide used per course differed (MMT 84 study, 6 g/m2; MMT 89 study, 9 g/m2). The goal of both protocols was to limit the amount of local therapy administered, basing such limitation on response to initial CHT. In the MMT 89 study, an attempt was made to decrease the amount of therapy given to patients in whom tumors had been completely resected at primary surgery by using the combination of vincristine and actinomycin D instead of IVA and by administering a more intensive multiagent CHT combination (incorporating carboplatin, epirubicin, vincristine, ifosfamide, actinomycin D, and etoposide) in patients with stage III nodal involvement who had been shown to have a particularly poor prognosis in the MMT 84 study.3
SIOP Classification for Clinical TNM Staging in Childhood RMS
MMT 84 study treatment schedules. Group 84.1, patients with completely excised stage I tumors; group 84.2, patients with completely excised stage II tumors and patients with microscopic residual disease after initial surgery for any tumors; group 84.3, patients with macroscopic residual disease after initial surgery or node-positive stage III tumors. IVA: ifosfamide 3 g/m2 on days 1 and 2, vincristine 1.5 mg/m2 on day 1 with maximum dose of 2 mg, and actinomycin D 1.5 mg/m2 on day 1 with maximum dose of 2 mg; DP: doxorubicin 60 mg/m2 on day 1 and cisplatin 100 mg/m2 on day 1; Res: tumor response.
MMT 89 study treatment schedules. Group 89.1, patients with completely excised stage I tumors; group 89.2, patients with residual disease after initial surgery for stage I tumors in genitourinary tract or head and neck, orbital, or nonparameningeal sites. VA: vincristine 1.5 mg/m2 on day 1 with maximum dose of 2 mg and actinomycin D 1.5 mg/m2 on day 1 with maximum dose of 2 mg; IVA: ifosfamide 3 g/m2 on days 1-3, vincristine 1.5 mg/m2 on day 1, and actinomycin D 1.5 mg/m2 on day 1; VCE: vincristine 1.5 mg/m2 on day 1, carboplatin 600 mg/m2 on day 3, and VM26 150 mg/m2 on day 4; IVE: ifosfamide 3 g/m2 on days 1-3, vincristine 1.5 mg/m2 on day 1, and etoposide 200 mg/m2 on days 1-3; S: surgery.
MMT 89 study treatment schedules. Group 89.3, patients with residual disease after initial surgery for stage I tumors in limbs and “other sites” and patients with any stage II tumors except parameningeal tumors. See Fig 2 for definitions of abbreviations.
In the MMT 84 study, in patients who achieved complete clinical and radiologic remission after initial CHT, a “second-look” operation with biopsies of the tumor bed was advised and often performed. Local relapse rates for patients with negative biopsy specimens who were in complete remission (CR) were not different from those for patients in clinical and/or radiologic CR who did not undergo second-look operations. Moreover, in the few patients found to have demonstrable histologic disease at second-look operation, there was no survival advantage, even when intensified therapy was given.8 Consequently, in the MMT 89 protocol, second-look operations were not routinely recommended to confirm CR in patients without residual disease. Local therapy was administered to residual tumor in patients with clinically or radiologically demonstrable masses at completion of CHT. Whenever possible, an attempt was made to avoid radical surgery (hysterectomy, total vaginectomy), and intracavitary BT was encouraged, when feasible, so that the use of radical ERT was limited. BT was administered after ovarian transposition, according to a previously described method.4,9
RESULTS
Histology
Thirty-eight girls with primary tumors of the genital tract were registered onto the MMT 84 (13 patients) and MMT 89 (25 patients) studies. Tumors were classified as RMS in 36 patients (22 botryoid, 12 embryonal, and two alveolar) and embryonal sarcoma (desmin-negative; ie, muscle marker–negative) in two patients.
Localization
The primary tumor originated in the vagina in 23 patients (with extension to the perineum in one girl and to the pelvis in two), in the vulva in four patients (with extension to the vagina in one girl), and in the uterus in 11 girls. Uterine tumors were localized to the cervix in seven patients (with an extension to the vagina in two girls) and to the corpus in one patient, but in three girls the whole uterus was invaded, with extension to the vagina.
Age at Diagnosis
The median age at diagnosis of girls with vulvovaginal tumors was 21 months (range, 9 months to 15.6 years), whereas in patients with uterine tumors, the median age at diagnosis was 15 years (range, 10 months to 16.6 years).
Treatment Groups
Allocation to clinical groups for treatment depended on TNM staging and the outcome of the initial surgical procedure (Table 2). Four patients underwent initial microscopically complete tumor resection (pT1) with conservative surgery (partial resection of the cervix in three patients with cervical tumors and partial vaginectomy in one patient with a vaginal tumor). As adjuvant treatment after surgery, one girl received two courses of IVA therapy (MMT 84 protocol) and three girls received only two courses of combination therapy with vincristine and actinomycin D (MMT 89 protocol). None of the patients had RT.
SIOP Classification for Clinical and Postsurgical TNM Staging and Tumor Size and Status in Childhood RMS
MMT 89 study treatment schedules. Group 89.5, patients with node-positive stage III tumors. CEV: carboplatin 500 mg/m2 on day 1, epirubicin 150 mg/m2 on day 1, and etoposide 200 mg/m2 on days 1-3. See Fig 2 for definitions of all other abbreviations.
At the time of diagnosis, 34 patients were found to have had incomplete resection: 10 girls had microscopic residual disease (pT3a) (eight stage I, one stage II, and one stage III) and 24 girls had macroscopic residue or only an initial biopsy for diagnosis (pT3b) (15 stage I and nine stage II). Only one girl had nodal involvement (stage III). The outcome of these 34 patients was as follows:
One girl never achieved local control of the primary tumor and died despite radical surgery (pelvic exenteration).
Twenty-two girls (65%) achieved complete clinical and radiologic remission with CHT alone (IVA therapy). Remission was assessed by vaginoscopy, with the patient under general anesthesia, performed during and after the end of CHT. In 12 cases, biopsies of the tumor bed were performed. Five of these girls were found to have tumor and underwent local treatment (intracavitary BT, three patients) or second-line CHT (vincristine, carboplatin, and teniposide [VM26] therapy, one patient; vincristine, carboplatin, and VM26 therapy and one course of cisplatin and adriamycin therapy, one patient). Pathologic examination did not reveal any evidence of tumor in seven patients; nevertheless, four girls subsequently relapsed locally. However, among the 10 girls who achieved complete local remission and who did not undergo biopsy of the tumor bed, only two relapsed locally.
In 11 patients, local treatment was necessary for a residual mass after initial CHT. RT was administered in five cases (BT in four and ERT in one), radical surgery was performed in four cases, and both modalities were used in two cases. In the six patients who underwent radical surgery, the tumor was initially localized to the uterine body in three patients and to the vagina in three. In all cases, the surgical procedure performed was hysterectomy. In addition to hysterectomy, total vaginectomy was performed in three of these patients and partial vaginectomy was performed in the other three. In five of these six girls, pathology showed microscopic residual tumor. Only one girl underwent an immediate reconstructive surgical procedure with replacement of the vagina using a colon graft.
Outcome
The 5-year overall survival rate (± SE) was 91% ± 6%, and the 5-year event-free survival rate was 78% ± 7%. Three girls died. One never achieved complete local control of the tumor, despite BT and radical surgery (pelvic exenteration). The second patient had alveolar RMS of the vulva with inguinal nodal involvement. Disease recurred in iliac nodes after initial complete local control (negative biopsy specimens from the tumor bed) with CHT alone, and despite lymphadenectomy and ERT to the para-aortic nodes (45 Gy), the patient developed lung metastases and died 29 months after diagnosis. The third patient had embryonal sarcoma of the uterine corpus. Total hysterectomy was performed after initial CHT left a residual mass, and no tumor cells were found in the pathologic specimen. Locoregional relapse occurred 11 months after diagnosis. The patient never achieved second CR, despite second-line CHT (vincristine, carboplatin, and VM26 therapy) and ERT (45Gy), and she died 23 months after diagnosis.
Thirty-five patients remained in CR at a median follow-up of 62 months (range, 16 to 129 months). Thirty of these girls were in first CR and five were in second CR. Four patients in whom microscopically complete excision of the tumor had been achieved at the beginning of the study received short CHT and were in CR 60, 70, 85, and 104 months after diagnosis. Thirteen girls (36%) were treated with CHT alone after incomplete initial resection and remained in first CR at a median follow-up of 44 months (range, 22 to 122 months). One girl relapsed locally and achieved second CR with IVA therapy but was lost to follow-up 6 months after relapse.
Seventeen patients (49%) required local treatment to achieve CR, either for a residual mass after initial CHT (10 patients) or for microscopic disease at biopsy (three patients) or after relapse (four patients). Local treatment was RT (ERT in three patients, BT in seven), radical surgery (hysterectomy in three; one of these patients also had a total vaginectomy), radical surgery (hysterectomy in three; two of these patient also had total vaginectomies) and RT (ERT in three, BT in one), and conservative surgery and ERT (one patient).
Local Relapse
Local relapse at the primary tumor site occurred in seven patients, with additional regional nodal involvement in two. The median interval between diagnosis and relapse was 16 months (range, 6 to 26 months). Two patients (stage III alveolar RMS of the vulva and embryonal sarcoma of the uterine corpus) died after relapse, and all other relapsing patients achieved and remained in second CR. Each of these patients had initially stage I tumors. One girl achieved second CR with IVA therapy but was lost to follow-up 6 months after relapse. The four remaining girls received CHT (IVA therapy in one patient and vincristine, carboplatin, and VM26 therapy in three patients) and local treatment (ERT in two, total vaginectomy and ERT in one, and complete conservative excision of a pelvic residual mass and ERT in one). They were all in second stable CR 5 to 9 years after diagnosis.
Uterine Tumors
Three patients with uterine tumors presented with small polypoid tumors of the cervix, which could be completely resected at diagnosis. All were alive at 60, 70, and 104 months. Of eight other patients (four with cervical tumors, one with RMS of the uterine corpus, and three with RMS of the whole uterus) in whom only biopsy was performed at diagnosis, complete local control was achieved in three by CHT alone (IVA therapy in two and IVA therapy and second-line CHT in one) 25, 61, and 122 months after diagnosis. In the five other girls, residual masses after initial CHT required local treatment (ERT in two girls; one girl also had BT) or hysterectomy (three girls; two girls also had total vaginectomies). One patient had no evidence of tumor in the resected specimen but died after locoregional relapse. All of the other patients were alive 33, 99, 113, and 129 months after diagnosis.
DISCUSSION
Our findings confirm the excellent prognosis of RMS arising in the female genital tract, both in the uterus and in the vagina, and emphasize the efficacy of a conservative approach to local therapy in decreasing long-term functional sequelae.
The favorable outcome of treatment for RMS arising in the vagina or vulva has been previously reported by several authors, in Intergroup Rhabdomyosarcoma Studies,2,10-12 and in European studies.4,13-15 In 1985, Hays et al12 reported 22 cases of nonmetastatic vaginal RMS treated on Intergroup Rhabdomyosarcoma Study (IRS) I and II protocols between 1972 and 1984. The only patient who died was treated in the neonatal period. Death was due to myelosuppression and infection. Of the 21 survivors, only three were treated without organ ablation; two of the 21 underwent primary exenteration (before 1974), 11 had hysterectomies and vaginectomies, and five had total or partial vaginectomies. Aggressive local treatment with organ ablation was considered mandatory in these early protocols.
In IRS III,5 26 girls with vulvovaginal primary tumors were treated with preoperative CHT and less radical surgery. Twenty-three patients became disease-free, although 13 (56%) of these patients underwent radical surgery (cystectomy in one, hysterectomy in five, and partial or complete vaginectomy in seven). The authors specified that no viable tumor was identified in the resected specimen in six patients.
Results of similar treatment in France during the same period as IRS I and II were published in 1990 by Flamant.4 Sixteen girls with nonmetastatic vaginal RMS received CHT and BT without organ ablation, with the exception of one patient who underwent hysterectomy. Fifteen girls remained disease-free after more than 10 years. Four of these patients had serious sequelae from low-dose BT, but the remaining 11 had no or minor vaginal sequelae.
Other reports from European centers involve small series of patients from individual institutions. There were no deaths among four patients treated in Amsterdam,15 of whom two retained the uterus and vagina, and among two patients treated in London without organ ablation.16
In the present study, 25 of 27 girls with vulvovaginal tumors survived and the uterine salvage rate was 88% (22 of 25 patients), because only three girls underwent hysterectomy. Of the 22 girls who retained their uteri, seven received intracavitary BT, four received ERT, and all will need long-term follow-up to ensure adequate evaluation of possible functional sequelae. However, 11 girls (44%) who had no local treatment and received only CHT after initial surgery may be considered cured and without any risk of local sequelae. This compares favorably with the six (23%) of 26 girls in IRS III who were also cured with CHT alone.5
RMS arising in the uterus has generally been identified as distinct from vaginal RMS, which is characterized by older age at diagnosis, propensity to local recurrence, and poor prognosis.11,12,17 The first report from IRS I and II by Hays et al12 included discussion of eight cases of uterine RMS, with three deaths from tumor progression after relapse. Among the five surviving girls, three underwent hysterectomy.
In 1995, Corpron et al6 reported results from IRS III and IV pilot of nonmetastatic uterine RMS. Of 11 patients, only six survived; four died from CHT toxicity and one died from tumor progression. Two of the six surviving patients had hysterectomies and vaginectomies. The authors suggested that less vigorous operative resection may be possible in combination with primary CHT but emphasized consideration of the four patients who died from sepsis or other treatment complications.
In the present study, however, the outcome of patients with uterine tumors was excellent, in that 10 of 11 girls survived, with eight of them retaining their uteri. The three girls who underwent partial excision of the cervix and the three other girls treated with CHT alone can be considered to have little or no risk of local sequelae, although the two girls who received RT may have long-term functional sequelae.
The results of this study confirm the strong possibility of cure for girls with RMS of the genital tract when an approach designed to limit the use of local therapy is taken. However, these patients must be followed up very closely, especially in the 2 years after diagnosis, when local relapse is most likely. In our experience, such close surveillance may require vaginoscopy with the patient under general anesthesia, initially repeated every 2 months.
- Received October 26, 1998.
- Accepted February 22, 1999.
