Retroperitoneal node biopsy in paratesticular rhabdomyosarcoma

Retroperitoneal

Node Biopsy in Paratesticular

Rhabdomyosarcoma

By Eugene S. Wiener, Walter Lawrence, Daniel Hays, Thorn E Lobe, Richard Andrassy, Sarah Donaldson, William Crist, William Newton, Jean Johnson, Edmund Gehan, and Chantal Rodary 0 Retroperitoneal lymph node dissection (RPLND) was used in 121 Intergroup Rhabdomyosarcoma Study (IRS) III patients with nonmetastatic paratesticular (PT) rhabdomyosarcoma (RMS) to assess retroperitoneal lymph node (RPLN) involvement so as to determine the need for x-ray therapy (XRT). Clinical node evaluation (CNE) was accomplished by a computed tomography (CT) scan in 105 and a sonogram in six. Pathological node evaluation (PNE) was performed in 113: lymphaderiectomy (9 bilateral, 85 unilateral) and biopsy in 19. Vincristine and actinomycin D were used for 1 year postoperatively in 89%; all patients who had positive PNE received RPLN XRT. This study compares CNE with PNE and evaluates predictors of relapse and survival. Findings: There were clinically negative nodes (cN0) in 81% of the 121 patients. Among cN0 patients, 14% had positive nodes (pN1). Of the clinically positive (cN1) patients, 94% had pN1. RPLN relapse occurred in only two of the 121 patients. Initially both had cN0 and one had PNE that was negative. For all 121 patients,

the B-year survival was 91%.

For cN0

patients, the 5-year survival was 98% compared with 69% for cN1 patients (P < .OOl). Among the children in whom treatment failed, nodes were cN1 in 5 of 11 (45% j compared with 15 of 107 (14%) in those whose treatment did not fail (P c .008). Conclusions: (1) Results of RPLN imaging studies were negative in 81% of patients with PT RMS (specificity 99%, sensitivity 57%). (2) RPLN recurrence is uncommon (even when RPLN are initially involved) if regional XRT and appropriate chemotherapy are used. (3) RPLN biopsy is not necessary to confirm positivity in patients with clinically positive nodes. (4) Node positivity is associated with decreased patient survival and helps identify patients who have an increased risk for relapse and death. Copyright o 1994 by W.B. Saunders Company INDEX WORDS: Rhabdomyosarcoma; paratesticular rhabdomyosarcoma; testicular tumors; retroperitoneal node dissection.

P

(PT) tumors account for ARATESTICULAR 7% of rhabdomyosarcomas (RMS) and 12% of childhood scrotal tumors. l When radical orchidectomy alone was the treatment, the &-viva1 was 50%.2 More recently, the survival was 90% when adjuvant chemotherapy was routinely administered and when radiation therapy was directed to known nodal or residual disease.2-4 The initial use of retroperitoneal lymph node dissection (RPLND) for PT RMS was based on the management of non-RMS testicular malignancies and the high incidence of retroperitoneal lymph node (RPLN) involvement.5s6 RPLND is performed to identify regional node involvement so as to direct radiation therapy and possibly to surgically remove nodes that are involved in either microscopic or gross disease. These benefits have been questioned, and some have advocated selective node JournalofPediatric Surgery, Vol 29, No 2 (February), 1994: pp 171-178

biopsy or reliance on clinical node evaluation alone.2.7-9 This is due to the reported accuracy of radiographic imaging studies in the identification of node involvement,2*9J0the successful eradication of micrometastasis by systemic therapy,2J1J2 and the complications attributed to RPLND.‘J3J4 The Intergroup Rhabdomyosarcoma Study (IRS) III (1984 to 1991) recommended preoperative computed tomography (CT) scan evaluation of RPLN and initial RPLND for all children with nonmetastatic PT RMS. This study analyzes the patients’ conditions to (1) evaluate the accuracy of modern radiographic imaging in the determination of node involvement, (2) determine when retroperitoneal lymphadenectomy or node biopsy is necessary, and (3) define the relationship of lymph node involvement to relapse and outcome, including the impact of withholding regional lymphatic irradiation that would result from possible inaccuracies of radiographic clinical node evaluation. MATERIALS AND METHODS In IRS III there were 1,062 eligible patients. Of these, 121 children had nonmetastatic PT RMS, and six had metastatic RMS. This study evaluates the nonmetastatic cases only. The characteristics of the study patients are shown in Table 1. Since RPLN involvement is not readily determined by physical examination, clinical node evaluation (CNE) necessitated a computed tomography (CT) scan and/or ultrasound examination of retroperitoneal and pelvic nodes in all patients at the time of study entry. A CT scan was performed in 95% of evaluable children. Lymphography was not performed in any patient. The IRS III surgical guidelines recommended (1) resection by radical inguinal orchidectomy to the level of the internal inguinal ring, (2) resection of scrotal skin when there was fixation to or involvement of scrotal skin by the tumor, or where there was a prior transscrotal approach for biopsy or resection, and (3) initial retroperitoneal lymph node dissection for all patients. Complete excision of the primary tumor was accomplished in 113 children (93%). Pathological retroperitoneal node evaluation (PNE) was accomplished in 113 (93%) patients. Unilateral lymphadenectomy

From the Intergroup Rhabdomyosarcoma Study Committee of the Children’s Cancer Group! Pediatric Oncolo~ Group, and United Kingdom Children’s Cancer Study Group. Supported by VSPHS Grants CA-24507, CA-30138, CA-30969, CA-29139, and CA-13539. Presented at the 24th Annual Meeting of the American Pediam’c Surgical Association, Hilton Head, South Carolina, May 15-18, 1993. Address reptint requests to Eugene S. Wiener, MD, Children’s Hospital of Pittsburgh, Department of Pediatric Surgery, 3705 Fifth Ave, Pittsburgh, PA 15213. Copyright o I994 by W.B. Saunders Company 0022.3468/94/2902-0008$03.OOJO 171

172

WIENER ET AL

Table 1. Characteristics

of Patients With Nonmetastatic

Paratesticular

Rhabdomyosarcoma Characteristic

n

%

White

87

75

Black

19

16

Other

10

9

Category Race

l

Age (yr), median

age,6 vr

IRS clinical group

o-1

2

2

1-5

44

37

5-10

33

27

IO-15

16

13

>15

26

21

It

86

71

II*

29

24

III5 Clinical node

CT scan

evaluation Histology

excision

5

105

95

Ultrasound

6

5

Embryonal

113

96

Alveolar Primary tumor

6

Complete incomplete

5

4

113

93

8

7

72

65

Tumor size

s5cm >5cm

39

35

Retroperitoneal

Dissection

94

83

19

17

Yes

32

27

No

88

73

31

VA**

81

67

32

VA + XRTtt

27

22

33

VA + Adr + XRTSS

1

1

34

VAC + XRT§§

1

1

35

VadrC-VAC + CPDD +

36

VadrC-VAC + CPDD +

38

VadrlVAC

node procedures Radiation therapy

biopsy

IRS Ill regimen

2

XRT*** VP16 + XRTttt

3

+ CPDD +

XRT***

6

*Calculations of percentages exclude the patients with missing data in the particular category.

were used for the paraaortic and iliac areas by the “hockey stick” technique using megavoltage photon radiation. Thirty-two of the 121 patients received radiation therapy to the retroperitoneal nodes or local sites. The following data were reviewed by IRS surgeons: registration data forms; clinical pathological grouping and pretreatment surgical staging data forms, which were completed at the patient’s institution; the surgeon’s operative notes; and the pathology reports. Pathology slides from all patients were reviewed centrally by IRS pathologists. Radiotherapy records and dose calculations were reviewed by the Quality Assurance Review Center (QARC) and confirmed by members of the IRS Radiotherapy Committee, who independently reviewed all simulation and portal films. In addition, an independent review of all of the surgical, pathological, and radiotherapy data was carried out by the IRS statistical office and by one of the authors (CR.). Survival (S) curves were calculated from the time of start of treatment in all clinical groups, and all deaths were considered “failures.” Progression-free survival (PFS) characterizes the percentage of patients who are alive and had not experienced relapse, progressive disease, or death by a particular time. PFS and S curves were calculated according to the method of Kaplan and Meier.16 The differences between curves were determined using logrankl’Jx tests and the Gehan-Breslow test.19 Patients alive and relapse-free were considered censored observations at the time of the last follow-up examination. The P values obtained from statistical tests were used as a measure of the strength of the evidence against the null hypothesis being tested; .Ol < P < .05 indicated a statistically significant result with moderate evidence against the null hypothesis, and P < .Olindicated a highly significant result with strong evidence against the null hypothesis. x2 tests were used for comparison of frequency distributions of patient characteristics. Sensitivity (the proportion of patients with PNE-positive nodes identified as positive by CNE), specificity (the proportion of patients with PNE-negative nodes identified as negative by CNE), predictive positive value (the proportion of patients with positive CNE who are PNE-positive), and predictive negative value (the proportion of patients with negative CNE who are PNE-negative) were calculated from a 2 x 2 table as illustrated in Table 3.20Calculations of percentages excluded the patients with missing data in each particular subset.

tLocalized disease, completely resected. *Total gross removal with microscopic residual and/or regional node involvement. §Gross residual.

l*Vincristine,

dactinomycin.

ttvincristine,

dactinomycin + XRT.

*SVincristine, dactinomycin, Adriamycin + XRT. §§Pulse vincristine, dactinomycin, cyclophosphamide

+ XRT.

***Pulse

cisplatin,

vincristine, Adriamycin, cyclophosphamide,

dactinomycin + XRT. tttPulse

vincristine, Adriamycin, cyclophosphamide,

cisplatin,

dactinomycin, VP-16 + XRT.

was performed in 85 patients, bilateral lymphadenectomy in nine, and retroperitoneal lymph node biopsy in 19. All patients were treated with IRS III regimensI as shown in Table 1. Radiation therapy was determined by the patient’s clinical group: group I had no radiation; groups II and III had radiation to the area of known involvement (local residual or retroperitoneal nodes) with a 5-cm margin. The radiation dose was 3,060 centigray (cGy) in ISO-cGy fractions, with five treatments per week. In select cases, a local area of residual disease was given a small field boost to a total dose of 4,500 cGy. Anterior-posterior opposed portals

RESULTS

The findings of CNE and PNE in all 121 patients are shown in Table 2. Nodes were clinically negative (cN0) in 81% and clinically positive (cN1) in 17% (three patients had missing data). Nodes were pathologically negative (pN0) in 73% of the evaluable patients and positive (pN1) in 27%. Comparison of CNE and PNE

There were 110 evaluable patients who had both radiological CNE and available pathological data from PNE (Table 3). The results of these two node evaluations were discordant in 14 patients. There were 13 false-negative CT scans (cN0 -+ pN1) and only one false-positive CT scan (cN1 -+ pN0). In this series, CNE or RPLN by imaging studies has a specificity of 99% and a sensitivity of 57%. A positive imaging study has a predictive value of 94%; a negative study has a predictive value of 86% (Table 3).

PARATESTICULAR RHABDOMYOSARCOMA

173

NODES

Table 2. Findings in 121 Children With Nonmetastatic Total No. of

Pathological Node

Patients

Evaluation

Primary Excision

Clinical Status cN0

Complete

cN1

Complete

89

12

4

3

1

98 (81%)

92 (94%)

13 (14%)

16

15

14

4

3

3

Incomplete Subtotal

20 (17%)

N unknown

Rhabdomyosarcoma

Positive Nodes*

94

Incomplete Subtotal

Paratesticular

Complete

18 (90%)

3

3

0

3 (2%)

3 (100%)

0

Total

121 (100%)

30 (27%)

Total Relapse

2

5

2

6 (6%)

1

3 1

17 (94%)

Subtotal

113 (93%)

Nodal Relapse

4 (20%) 0 0 2

10 (8%)

*% = positive/total node evaluations.

There were eight patients who did not have PNE and in whom node status and treatment were based on imaging findings alone. Each had a CT scan as the clinical evaluation of RPLN. Two had cN1, neither of whom had node recurrence. Six had cN0. One of these patients, described below, had RPLN recurrence. All eight have been alive for a median of 176 weeks. Completeness of Resection

Complete resection of the primary tumor was accomplished in 113 patients. Among these, 24% had pN1. Comparatively, in the eight patients with incomplete resection of the primary, 67% had pN1 (P = .02). One patient who had local recurrence (see below) and one who had relapse at a distant site (see below) were the only ones with adverse events who had incomplete resection. Adverse Events

Twelve of the 121 patients had adverse events (Table 4). Retroperitoneal lymph node relapse occurred in only two patients (1.6%). One patient, who did not have PNE and had cN0 by a CT scan, had relapse simultaneously in RPLN and the bone marrow. The other patient’s relapse was in RPLN, despite an initial extensive unilateral RPLND with 25 Table 3. Concordance of Radiological With Pathological Retroperitoneal

Node Evaluation (Sensitivity and Specificity) Pathological N’Jl

PNO

Total

Radiological cN1

17

1

18

cN0

13

79

92

30

80

110

Total

NOTE. Values = number of patients in each category. Sensitivity = 17/30 (56.6%). Specificity = 79/80 (98.7%). Predictive value of positive imaging study = 17/18 study = 79/92 (86%).

(94%). Predictive value of negative imaging

histologically negative nodes. Neither patient with node recurrence received initial radiotherapy. Nonnodal recurrence occurred in nine patients (7%). Eight had distant recurrences to thoracic sites (4) cortical bone (3) and bone marrow (1). The latter also had simultaneous nodal recurrence and is also included in the group with nodal relapse. There was one local recurrence in a cNO/pNO patient who had incomplete resection of a small primary tumor but received 4,140 cGy to the initially uninvolved pelvic nodes rather than to the scrotum as recommended in the protocol. There were eight deaths. Two were in pN0 patients: one had local recurrence and died of diffuse abdominal disease at 184 weeks, and the other had pulmonary relapse and died at 185 weeks. There were 6 deaths in pN1 patients. Four patients had recurrence in cortical bones or the chest wall and died 37, 48,78, and 100 weeks after entry into the study. One patient died a treatment-related death at 67 weeks; no residual tumor was found during the autopsy. The other death was in a 17 year old who had a 3-cm embryonal PT tumor and biopsy-proven extensive paraaortic lymph node involvement (pN1). Regional control was never achieved, despite nodal radiation (4,500 cGy) and chemotherapy (regimen 34); the patient died of disease progression at 57 weeks. This case is considered one of nonresponse rather than of relapse. Survival (S) and Progression-Free Survival (PFS)

In the group as a whole, Kaplan-Meier S is 95% (SE, .0218) at 3 years and 91% (SE, .0309) at 5 years. The PFS is 89% (SE, .0301) at both 3 and 5 years. In the subset of cN0 patients, the j-year S PFS are 96% and 93%, respectively; whereas in the cN1 patients, S and PFS are each 69% (P < .OOl for both) (Figs 1 and 2). Patient age of less than 10 years is associated with a 97% 5-year S, compared with an 84% 5-year S for patients 10 yrs of age or older (P = .03).

174

WIENER ET AL

Table 4. Characteristics of Patients Who Had Relapse and/or Patient No.

Resection

Histology

Age (~1

Size

cN

Procedure

pN

Died

Regimen

Irradiation

RdapSe

Outcome

Patients with relapse 1

Complete

18

EMB

<5

NO

NX

None

31

No

Node, marrow

Alive

2

Complete

19

EMB

>5

NO

NO

Uni/dissect

31

No

Nodes

Alive

3

Complete

15

EMB

>5

NO

NO

Uni/dissect

31

No

Bone

Alive

4

Complete

16

EMB

>5

NO

NO

UnVdissect

31

No

Lung

Dead

5

Incomplete

3

EMB

<5

NO

NO

Uni/dissect

32

Yes

Local*

Dead

6

Complete

7

EMB

>5

NO

Nl

Uni/dissect

32

Yes

Mediastinal

Alive

7

Complete

17

EMB

>5

Nl

Nl

UnVdissect

32

Yes

Bone

Dead

6

Incomplete

14

EMB

<5

Nl

Nl

Unbdissect

36

Yes

Chest wall

Dead

9

Complete

16

EMB

<5

Nl

Nl

Bil/dissect

32

Yes

Chest wall

Dead

10

Complete

14

EMB

>5

Nl

Nl

Uni/dissect

32

Yes

Bone

Dead

Complete

17

EMB

Patients without relapse who died 11

17

12 Abbreviations:

cN, clinical node status, pN, pathological

*Relapse occurred

in the opposite

scrotum;

<5

Nl

Nl

Biopsy

34

Yes

Progressiont

Dead

>5

Nl

Nl

Uni/dissect

32

Yes

None*

Dead

node status; EMB, embryonal:

radiation was incorectly

tThere were grossly positive nodes at diagnosis;

a complete

administered

Uni, unilateral;

Bil, bilateral.

to pelvic nodes rather than the site of residual.

response was never achieved, and the patient died of progressive

disease.

*Toxic death; no disease found at autopsy.

The 11 patients (nos. 1 to 11 in Table 4) whose therapy failed (relapse or disease progression) were compared with the 110 patients who had no characteristics predictive of failure. There was no significant difference in tumor size, completeness of resection,

or histological findings. The median age of the 11 patients was 16 years; the median age of the others was 6 years. The time to tumor relapse or progression in patients aged 10 or older is significantly shorter

i ...._._......................................... N1

0.30 -

0.40 -

0.30 -

0.20 -

0.10 -

Fig 1. Kaplan-Meier survival in 118 patients for whom clinical node status is available. Solid line: cN0 clinically negative nodes, 2 of 98 failed, O/E 0.299. Percentage surviving: 1 yr (SE), 100 (1.0484); 3 yr, 100 (1.0484); 5yr. 98 (0294). Dotted line: cN1 clinically positive nodes, 8 of 20 failed, O/E 4.575. Percentage surviving: 1 yr (SE), 90 (6871); 3 yr, 89 (.1053); 5 yr, 89 (.1053); P c ,001.

Fig 2. Kaplan-Meier proogression-free survival in 118 patients for whom clinical node status is available. Solid line: cN0 clinically negative nodes, 6 of 98 failed, O/E 0.595. Percentage surviving: 1 yr (SE), 97 (.0184); 3 yr, 93 1.0279); 5 yr, 93 (.0279). Dotted line: cN1 clinically positive nodes, 6 of 20 failed, O/E 3.136. Percentage surviving: 1 yr (SE), 80 (0894); 3 yr, 69 (1046); 5 yr, 69 (.1046); P c ,001.

PARATESTICULAR

RHABDOMYOSARCOMA

175

NODES

orbit, superficiai head and neck, and vulva/vagina the best survival of all RMS sites.2r-25This survival has been attributed to the adjuvant treatment, which included routine retroperitoneal node dissection.“-5J6 The reported incidence of positive nodes when RPLND was employed in PT RMS was 43% in a series of 46 patients and 40% in 20 IRS II patients with completely excised primary tumors (reported by Raney et aP). However, in later studies4Jj that included a larger group of these IRS II patients, there was a 26% incidence of positive retroperitoneal nodes. In IRS III the incidence of pathologically involved nodes is similar (27%) to the latter reports.

‘I 0.90

_

0.60

_

0 .---, -_ _ a-_, 10+ yrs_____

L___._.................

0.70

-

0.60 -

0.50

-

0.40

-

How Does Retroperitoneal Node Evaluation by Radiographic Imaging Studies Compare With Pathological Examination?

Fig 3. Time to relapse or progression by age group in 121 patients. Age 0 to 5 yr: 1 of 48 failed; O/E 0.233. Percentage relapse/ progression-free: 1 yr, 98; 3 yr, 98; 5 yr, 98. Age 5+ to 10 yr: 1 of 33 failed; O/E 0.328. Percentage relapse/progression-free: 1 yr, 100; 3 yr. 98; 5 yr, 96. Age lO+ yr: S of 42 failed; O/E 2.485. Percentage relapse/progression-free: 1 yr, 85; 3 yr, 77: 5 yr, 77. Gehan-Breslow test: slOyrv>lOyr,P< ,001.

than in patients less than 10 years of age (Z’ = .003) (Fig 3). In children whose treatment failed, the node status was cN1 in 5 of 11 (45%) compared with 15 of 107 (14%) evaluable patients whose treatment did not fail (P = .008). Using the same criteria, if the node status was pN1 the findings are 6 of 10 (60%) and 24 of 103 (23%), respectively (P = .012). The failures when node status and age are combined are shown in Table 5. DISCUSSION

RMS is a heterogeneous tumor with different outcomes at different sites. PT RMS shares with Table 5. Treatment

Failures (Relapse or Progression) by Age and

Node Status in 11 Patients Whose Treatment

Failed

Age W

n Failed

n Total

510

cN0

2

71

3

>lO

cN0

4

26

15

510

CNl

0

6

0

>lO

cN1

5

14

36

Gehan-Breslow

Retropetitoneal Node Evaluation by Node Dissection or Biopsy Indicated for These Patients?

Is

Node status

NOTE.

Our study indicates that clinical node evaluation, using CT scans, is not entirely reliable. There is a low sensitivity, eg, when nodes were pathologically proven to be positive, only 57% of imaging studies had demonstrated node positivity. Specificity was quite high, eg, when nodes were pathologically negative, 99% of the imaging study results had been negative. The predictive value of a positive imaging study is quite good, but the predictive value of a negative study is less reliable. This less-than-satisfactory correlation of CNE with PNE is similar to a finding of Kulakowski et al;14in 152 patients with nonseminomatous testicular tumors, they found a disparity of 35% between CNE and PNE. Studies in non-RMS testicular tumors suggest that CT and magnetic resonance imaging (MRI) scans are both equally sensitive and specific.14J7,28A study in adolescents and adults with testicular malignancies compared CT scan and ultrasound evaluations of RPLN to lymphography.‘* The diagnostic accuracy of CT and ultrasound were almost the same, with sensitivities of 80% and 76% (respectively) and specificities of 100% for both. Although pedal lymphography may assure a more accurate assessment of node involvement, the pain and technical difficulty associated with its use in children have diminished its utility.

tests:

~1O,cNOv>10,cNO;~=.027 slO,cNOv

>lO,cNl;P=

,660

>lO,cNOv

>lO,cNl;P=

,091

5 10, cN1 v > 10. cN1; P = ,107.

% Failed

Based on high relapse-free rates and 100% survival rates, several investigatorss*9 have recommended avoidance of RPLND when radical inguinal orchidectomy results in complete microscopic excision and when the lymphography result is normal. Olive et al9 studied 19 localized, completely resected PT RMS patients who had cN0 (16 of these by lymphography). Nodal recurrence occurred in two patients (10.5%),

176

one of whom did not receive chemotherapy. Both patients survived after salvage therapy. On the other hand, they recommended lymphadenectomy in children who had microscopic residual disease after orchidectomy because of the higher risk of recurrence. In our study, only two of 121 patients had regional node recurrences. However, the value of retroperitoneal node evaluation cannot be completely discounted. The cN0 patients who actuahy had pN1 a11 received nodal irradiation. Had they been treated according to clinical evaluation (cNO), radiation would not have been given, and the outcome may have been different. On the other hand, only one of six cN0 patients who did not have PNE (and did not receive irradiation) had recurrence in the retroperitoneal nodes. A negative RPLND does not guarantee relapse-free survival; one child who had an extensive negative RPLND had nodal recurrence. In our study, incomplete resection of the primary tumor is associated with a higher likelihood of pathologically positive nodes than is complete resection (P = .02). However, incomplete resection is not associated with increased disease recurrence or progression. Neither RPLND nor node biopsy is recommended in the current IRS IV trials in children with localized, completely resected tumors whose imaging study results are negative. These patients are being closely monitored for node relapse. The value of debulking RPLND when RPLN are clinically positive is unclear.“,9,26 Because in the IRS III study, the grossly involved nodes were all irradiated and the patients all received chemotherapy as well as RPLND, it is difficult to determine the specific value of RPLND. When nodes are only involved microscopically (cNO/pNl), the added benefit of radiation therapy following “therapeutic” RPLND has also been disputed 9~l.l2,14,27because effective chemotherapy treats micrometastases wherever they exist. In our study, none of 13 cNO/pNl patients had nodal relapse; however, all had received radiation to the nodes. Because of the design of the IRS III trial, we cannot determine whether node resection, radiation therapy, or both are required when appropriate chemotherapy is administered to children who have either gross or minimal nodal disease. Whatever value RPLND has in the high survival of patients with nonmetastatic PT RMS, the price of RPLND is significant. In a long-term follow-up study of IRS children with PT RMS, Heyn et al7 found significant complications related to RPLND: intestinal obstruction in lo%, loss of normal ejaculatory function in 8%, and leg lymphedema in 5%. Other

WIENER ET AL

studies of RPLND have shown significant surgeryrelated complications: reduced ejaculatory volume in most patients,29vascular injuries,13330absence of ejaculation in most patients, 31 dry ejaculations in 56%,12 and ejaculatory impotence in 75%.32 Technical modifications may decrease these complications,i2 but it is likely that RPLND will always have some associated morbidity. Refinement of therapy may decrease these long-term effects and should be included in the future strategies for management of these chiIdren whose overall outcome is so good. What Impact Did Withholding Regional Node Irradiation Have on Outcome? It appears that there was no adverse effect on survival or relapse from withholding nodal irradiation in the inaccurately radiographically staged children. There was one death and only one nodal recurrence in such patients. This finding supports the contention of Olive et al9 that withholding radiotherapy from minimally involved nodes (cN0, pN1) should not affect ultimate survival. Radiation therapy to the retroperitoneal node echelon cannot be evaluated separately from node dissection or systemic therapy. Neither patient with regional node relapse had received radiotherapy. In fact, in all but one patient who received irradiation, the disease was controlled in the irradiated areas. Tefft et aP3 asked whether radiation to regional nodes was necessary in a study of 16 children with PT RMS. All patients with known node involvement had irradiation. There were no node relapses. The question they posed cannot be answered from their study as it cannot be from ours. Is Retropetitoneal Lymph Node Involvement a Predictor of Relapse and Jar Death? The 5-year survival is significantly diminished by retroperitoneal lymph node involvement and for patients aged 10 or older. These are the only characteristics predictive of poor outcome in the children whose treatment failed. Despite reports that suggest otherwise,24,25134 histology had no impact on outcome; no child with alveolar histology had relapse or died. We cannot confirm the suggestion of Olive et al9 that the factors leading to incomplete resection of the primary tumor can identify a patient at risk for failure or one who would benefit from RPLND. The identification of the high-risk patient (ie, age r 10 years with positive retroperitoneal nodes) and the unpredictability of imaging studies in determining node positivity warrant reconsideration of RPLN sampling (not dissection) in older children whose nodes are negative according to imaging studies. This is especially true if

PARATESTICULAR

RHAEDOMYOSARCOMA

177

NODES

future adjuvant therapy is more effective for these high-risk children. ACKNOWLEDGMENT The authors acknowledge the support of the full Intergroup Rhabdomyosarcoma Study (IRS) Committee and the IRS Statisti-

cal Center, the Department of Health and Human Services, the USPHS grants, the National Cancer Institute, Clinical Investigations Branch, and the investigators and data managers in the participating institutions of the Children’s Cancer Group, the Pediatric Oncology Group, and the United Kingdom Children’s Cancer Study Group. Donna Carothers’ expert assistance in the preparation of the manuscript was invaluable.

REFERENCES 1. Bruce .I, Gough DCS: Long-term follow-up of children with testicular tumours: Surgical issues. Br J Urol67:429-433,199l 2. Hamilton CR, Pinkerton R, Horwich A: The management of paratesticular rhabdomyosarcoma. Clin Radio1 40:314-317,1989 3. LaQuaglia MP, Ghavimi F, Heller G, et al: Mortality in pediatric paratesticular rhabdomyosarcoma: A multivariate analysis. J Urol 142473-478, 1989 4. Raney RB Jr, Tefft M, Lawrence W Jr, et al: Paratesticular sarcoma in childhood and adolescence: A report from the Intergroup Rhabdomyosarcoma Studies I and II, 1973-1983. Cancer 60:2337-2343,1987 5. Raney RB Jr, Hays DM, Lawrence W Jr, et al: Paratesticular rhabdomyosarcoma in children. Cancer 42:729-736,1978 6. Lawrence Jr W, Hays DM, Heyn R, et al: Lymphatic metastasis with childhood rhabdomyosarcoma. Cancer 60:910-915, 1987 7. Heyn R, Raney B, Hays D, et al: Late effects of therapy in patients with paratesticular rhabdomyosarcoma. For the Intergroup Rhabdomyosarcoma Study Committee. J Clin Oncol10:614623,1992 8. Rodary C, Flamant F, Maurer H, et al: Initial lymphadenectomy is not necessary in localized and completely resected paratestitular rhabdomyosarcoma. Med Pediatr Oncol 20:430, 1992 (abstr) 9. Olive D, Flamant F, Zucker JM, et al: Paraaortic lymphadenectomy is not necessary in the treatment of localized paratesticular rhabdomyosarcoma. Cancer 54:1283-1287,1984 10. Hogeboom WR, Hoekstra HJ, Mooyaart EL, et al: The role of magnetic resonance imaging and computed tomography in the treatment evaluation of retroperitoneal lymph-node metastases of non-seminomatous testicular tumors. Eur J Radio1 13:31-36, 1991 11. Fossa SD, Qvist H, Stenwig AE, et al: Is postchemotherapy retroperitoneal surgery necessary in patients with nonseminomatous testicular cancer and minimal residual tumor masses? J Clin Oncol 10:569-573, 1992 12. Qvist H, Fossa SD, Ous S, et al: Retroperitoneal surgery in patients with nonseminomatous testicular cancer and minimal residual tumor. J Surg Oncol50:220-223, 1992 13. Morin JF, Provan JL, Jewett MAS, et al: Vascular injury and repair associated with retroperitoneal lymphadenectomy for nonseminomatous germinal cell tumours of the testis. Can J Surg 35:253-256, 1992 14. Kulakowski A, Dardzinski R: Retroperitoneal lymph node dissection in the treatment of patients with non-seminoma testicular tumours. Eur J Surg Oncol 17:581-584, 1991 15. Ragab A, Gehan E, Maurer H, et al: Intergroup Rhabdomyosarcoma Study (IRS) III: Preliminary report of the major results. Proceedings of the American Society of Clinical Oncology. 11:363, 1992 (abstr) 16. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 457-581,1958 17. Mantel N: Evaluation of survival data and two new rank order situations arising in its consideration. Cancer Chemother Rep 50:163-170,1966

18. Mantel N, Haenszel W: Statistical aspects of the analysis of data from retrospective studies of disease. J Nat1 Cancer Inst 22x719-748, 1959 19. Gehan EA: A generalized Wilcoxon test for comparing arbitrarily singly-censored samples. Biometrika 52:203-224, 1965 20. Fleiss JL: Statistical methods for rates and proportions (ed 1). New York, NY, Wiley, 14-20, 1973 21. Lawrence W Jr, Gehan EA, Hays DM, et al: Prognostic significance of staging factors of the UICC staging system in childhood rhabdomyosarcoma: A report from the Intergroup Rhabdomyosarcoma Study (IRS-II). J Clin Oncol5:4654,1987 22. Rodary C, Gehan E, Flamant F, et al: Prognostic factors in 951 nonmetastatic rhabdomyosarcoma in children: A report from the international rhabdomyosarcoma workshop. Med Pediatr Onco1 19:89-95, 1991 23. Rodary C, Rey A, Olive D, et al: Prognostic factors in 281 children with nonmetastatic rhabdomyosarcoma (RMS) at diagnosis. Med Pediatr Oncol 16:71-77, 1988 24. Donaldson SS: Rhabdomyosarcoma: Contemporary status and future directions. The Lucy Wortham James Clinical Research Award. Arch Surg 124:1015-1020,1989 25. Pedrick TJ, Donaldson SS, Cox RS: Rhabdomyosarcoma: The Stanford experience using a TNM staging system. J Clin Oncol 4:370-378, 1986 26. Malek RS, Kelalis PP: Paratesticular rhabdomyosarcoma in childhood. J Urol 118:450-453, 1977 27. Stomper PC, Kalish LA, Garnick MB, et al: CT and pathologic predictive features of residual mass histologic findings after chemotherapy for nonseminatous germ cell tumors: Can residual malignancy or teratoma be excluded? Radiology 180:711714,199l 28. Damgaard-Pedersen K, von der Maase H: Ultrasound and ultrasound guided biopsy, CT and lymphography in the diagnosis of retroperitoneal metastases in testicular cancer. Stand J Urol Nephrol 137:139-144.1991 (suppl) 29. Johnson DE, Bracken RB, Blight EM: Prognosis for pathologic stage I non-seminomatous germ cell tumors of the testis managed by retroperitoneal lymphadenectomy. J Urol 116:63-65, 1976 30. Skinner DG: Non-seminomatous testis tumors: A plan of management based in 96 patients to improve survival in all stages by combined therapeutic modalities. J Urol 115:65-69, 1976 31. Staubitz W, Magoss I, Grace J, et al: Surgical management of testis tumors. J Urol 101:350-355, 1969 32. Walsh PC, Kaufman JJ, Coulson WF, et al: Retroperitoneal lymphadenectomy for testicular tumors. J Am Med Assoc 217:309312,197l 33. Tefft M, Hays D, Raney RB, et al: Radiation to regional nodes for rhabdomyosarcoma of the genitourinary tract in children: Is it necessary? Cancer 45:3065-3068,198O 34. Leuschner I, Newton WA Jr, Schmidt D, et al: Spindle cell variants of embryonal rhabdomyosarcoma in the paratesticular region. A report of the Intergroup Rhabdomyosarcoma Study Committee. Am J Surg Pathol 17:221-230, 1993

178

WIENER ET AL

Discussion C.A. Sheldon (Cincinnati, OH): Given the facts that the incidence of false-negative imaging was 14% and that all patients with microscopic lymph node involvement received radiation therapy, we do not really know the risk of nodal relapse that we will encounter when we remove lymph node dissection from our armamentarium. The one question I would like to ask the authors is, what has been the experience with salvage therapy for regional nodal relapse? J.L. Grosfeld (Indianapolis, IN): Dr Wiener has identified a poor-risk group of patients in a very good-risk tumor site for rhabdomyosarcoma in which more than 90% of the patients survive. Two factors have prognostic significance: one is age of the patient, the other is advanced disease. I have a number of questions, however. (1) When the lymph nodes are involved, does the retroperitoneal tumor burden play a role in the small group of patients who eventually meet their demise?; (2) What was the complication rate in the large number of patients that had a retroperitoneal lymph node dissection or a node biopsy?; and (3) Did the operative reports indicate that pediatric surgeons are aware of or perform lymph node dissections using nervesparing techniques? (current procedure of choice). There is still a significant error rate in imaging the retroperitoneum. I doubt whether imaging for retroperitoneal lymph node involvement in this disease is any better than imaging for retroperitoneal lymph nodes in Hodgkin’s disease, which has a 25% to 33% error rate. To be more accurate and yet less invasive, perhaps laparoscopic exploration and biopsy may be useful for staging in these patients. Is the IRS considering this as part of a protocol? M.P. LaQuaglia (New York, NY): Did you map the anatomic distribution of the nodal involvement? The real crux of the problem here is microscopic disease in retroperitoneal lymph nodes, and whether or not that needs to be irradiated or can be cured by chemotherapy. That is the real reason for doing lymph node sampling. Is there a plan in the IRS to randomize patients with microscopic disease to chemotherapy, or chemotherapy and radiation, to see if chemotherapy alone can eliminate that disease? S.J. Shochat (Stanford, CA): Dr Wiener, could you clarify a couple of points? Most of the patients have node-negative disease, and yet this is the group that is difficult to determine by CT. Are you recommending that these patients have node biopsy? And, do you have evidence that the surgeon can identify the

positive nodes without doing a complete node dissection? E.S. Wiener (response): I thank all the discussants. Just to answer the last question first, Dr Shochat, no, we don’t have that evidence. As I mentioned in the presentation, this study was not designed as a randomized study. All the patients were supposed to have node dissections. In fact, 93% did have them. The few that didn’t and had biopsies fared just as well as those who had dissection, so there are not enough data to determine whether it makes a difference. Just as was shown earlier in the first presentation today by Dr Smith, I think the most important finding in our study is that node positivity is a marker of poor outcome, and this is contrary to what we’ve said previously about this site. To answer Dr Sheldon’s questions: Only two patients had recurrence in the nodes. One of them survived. The other did not survive, and that was the patient who had 25 negative nodes removed at the time of a node dissection and still had recurrence, which led to death. To answer Dr Grosfeld’s questions: We again cannot truly answer the question as to whether tumor burden is important. It would appear from a small subset of these patients that it makes no difference whether you actually remove the nodes or just perform biopsies alone, so leaving them behind may not make a difference; but as has been alluded to by Dr LaQuaglia, the question is whether radiation is necessary in patients who just had microscopically involved nodes, that is, those who were clinically negative but pathologically positive. Whether that group of patients needs anything more than radiation cannot be answered from these data. We do not plan to do a randomized study because the oncologists don’t feel it is justified. There are groups of patients whose conditions worsen when their nodes are involved, such as in the extremity sites, and therefore in those sites we do recommend node biopsy to clearly determine node involvement since scanning techniques are not as good. Laparoscopy may be of value depending on the expertise and skill of the laparoscopist. A biopsy of the pelvic nodes can be performed very easily during laparoscopy, but the intra-aorto caval nodes and the renal hilar nodes are not so easy for those who haven’t had much experience. I don’t advocate that approach for everyone.