Pelvic exenteration for cervix cancer: Would additional intraoperative interstitial brachytherapy improve survival?

Int. J. Radiation

Biol. Phys., Vol. 38, No. 1, pp. 143-148, 1997 Copyright 0 1997 Elsevier Science Inc. Printed in the USA. All rights reserved 0360.3016/97 $17.00 + .OO

PI1 SO360-3016(96)00585-8

ELSEVIER

l

Oncology

Clinical Investigation PELVIC EXENTERATION FOR CERVIX CANCER: WOULD ADDITIONAL INTRAOPERATIVE INTERSTITIAL BRACHYTHERAPY IMPROVE SURVIVAL? JONATHAN J. BEITLER, M.D.,M.B.A.,* CAROLYN

PATRICK S.ANDERSON, M.D.,+Scorr WADLER, M.D.,* KATHERINE HAYES, M.D.,” ABBIE L. FIELDS, M.D.,i BFUJ SOOD, M.D.* AND GARY L. GOLDBERG,M.D.+

D. RUNO~ICZ,

M.D.,?

MARY

Departments of *Radiation Oncology and +Obstetricsand Gynecology, Division of Gynecologic Oncology; *Department of Medical Oncology, Albert Einstein Cancer Center and Montefiore Medical Center, Bronx, NY; and $Division of Radiation Oncology, New York Hospital, New York, NY Objective: Improved local control with the addition of brachytherapy to pelvic exenteration for recurrent cervical cancer has been reported to improve survival. We examined the sites of recurrence after pelvic exenteration to determine if these patients might have been salvaged by the improved local control promised by interstitial brachytherapy. We sought to identify risk factors available intraoperatively or perioperativeIy which might predict decreased local control. Methods: A retrospective review of 26 patients with recurrent cervical cancer who underwent total pelvic exen-since 1988 at our institution was performed. Results: Overall, the mean follow-up was 29.5 months (range 6.1-81.6). Of the 26 patients, 14 had no evidence ofse (NED), 1 was alive with disease (AWD), 9 were dead of disease (DOD), and 2 died of unrelated causes (DOC). Seven of 26 patients (27%) had margins r 5 mm, of whom 2 were NED, 4 DOD, and 1 AWD. Seven of 26 (27%) patients had lymphovascular involvement (LVI) or perineural invasion (PNI) with clear margins. Three of the seven with LVI or PNI and clear margins were NED, and four DOD. Of the 10 failures, 9 (90%) had close margins, PNI, or LVI. Conclusion: Our data reveal that 9 of 14 (64%) patients with close margins, LVI, or PNI were DOD or AWD, and 6 of 9 of those patients suffered local regional failure alone. Brachytherapy has the potential to cure 6 of 14 (43%) patients with these risk factors. Further study of brachytherapy at the time of pelvic extenteration is warranted. 0 1997 Elsevier ScienceInc. Cervix cancer, Brachytherapy, Pelvic exenterations, Margins, Lymphovascular involvement, Perineural invasion.

INTRODUCTION

currence and ultimately death in patients with this disease (8, 13, 17, 22). FIG0 Stage IB and IIA cervix cancer can be treated by either radical hysterectomy or external radiation and brachytherapy. FIG0 Stage IIB, IIIA, and IIIB cervix cancer is conventionally treated with a combination of external radiation therapy and brachytherapy. Among these patients there is a subset of patients who will recur locally with no evidence of nodal disease or distant metastases. Although there have been reports of patients with a small central recurrence who have been managed with radical hysterectomy after radiation therapy (6), these patients have generally been offered pelvic exenteration. Clinical predictors of survival after pelvic exenteration include interval from initial radiation treatment, size of the central mass, preoperative pelvic

Carcinoma of the cervix is the most frequent cause of disease and death from malignant neoplasms among women in developing countries, accounting for approximately 340,000 new cases and 160,000 deaths/year (16). The incidence is much lower in the developed world, where it ranks fifth in incidence and seventh in mortality among women. The American Cancer Society estimates that in 1996 there will be 15,700 new cases of cervical cancer and 4900 deaths attributed to cervix cancers in the United States (15). The Patterns of Care Outcome Study for uterine cervix cancer lists the 5year survival rates as 74%, 56%, and 33% for Stages I, II, and III, respectively (5). In-field recurrences are 12%, 27%, and 51%. Numerous reviews have identified risk factors predictive for re-

Presentedin part at the 27th Annnal Meeting of the Society of GynecologicOncologists,New Orleans,LA, February 11- 14, 1996. Reprint requests to: Jonathan Jay Beitler, M.D., M.B.A., Department of Radiation Oncology, Montefiore Medical Center,

111 East 210th St., Bronx, NY 10467-2490. authors acknowledge the expert secretarial assistanceof Ana Marie Sanchez. Accepted for publication 22 October 1996.

Acknowledgements-The

143

144

I. J. Radiation Oncology

l

Biology

l

Physics

sidewall fixation, and tumor within 1 cm of the lateral margin of resection (21). Until recently, pelvic exenteration was the last chance for cure for locally persistent disease. Now, despite prior radiation, expertly applied brachytherapy may improve local control (9- 11) in those patients with disease involving the pelvic sidewall. Patients with postradiation recurrences infiltrating the pelvic wall have been treated with laterally extended endopelvic resection and brachytherapy, with an actuarial 5-year survival probability of 40% (n = 51). Following staging laparotomy, exenteration of the infiltrated central pelvic organs, and maximum tumor debulking at the pelvic wall, afterloading catheters are inserted on the residual tumor or tumor bed. A pelvic wall plasty is performed to increase separation of the catheters from the small bowel, and routine surgical reconstruction of the bowel, bladder, and vulvoperineovaginal functions is performed. For those patients previously externally irradiated with 45 Gy to the pelvis, brachytherapy would begin during the second postoperative week. Patients receive 42-60 Gy in 6-Gy fractions delivered twice a week. Quality of life among those patients without evidence of disease was quite good, as assessed by a validated questionnaire (10). Using our patient population, we analyzed patterns of failure and then the pathologic predictors of failures after pelvic exenteration, to perioperatively select those patients who might benefit from improved local control with interstitial brachytherapy.

METHODS

AND MATERIALS

We performed a retrospective analysis of patients who underwent pelvic exenteration at the Albert Einstein College of Medicine and Montefiore Medical Center. Using our computerized database, 535 cervical cancer patients were identified as being evaluated at the Albert Einstein College of Medicine and Montefiore Medical Center from 1988 to 1995. Twenty-seven patients form the subgroup who underwent total pelvic exenteration and initial stage, tumor differentiation, prior therapy, pathological margins, histology, and clinical results were analyzed using Statistica for Macintosh, Release 4.1 (StatSoft, Tulsa OK). The pathology slides on all patients were reviewed and additional sections or blocks evaluated as necessary. Because of artifactual inaccuracies inherent in the pathological processing, any margin I 5 mm was considered close. Because of the difficulty in differentiating local from regional failure after pelvic exenteration, any pelvic failure after exenteration was considered to be a local-regional failure. Pelvic radiation was considered administered when any patient received a dose 2 39.6 Gy. One patient underwent pelvic exenteration because of multiple fistulas and a necrotic vagina, and never had recurrence of her cancer. Although this patient was sal-

Volume 38, Number 1, 1997

vaged, she will be excluded from analysis, and thus 26 patients form the basis of this analysis. RESULTS Demographic information is available in Table 1. Mean follow-up was 29.5 months (range 6.1-81.6) and the median follow-up was 20.5 months. Age, initial stage, clinicopathologic results, disease status, site of recurrence or distant failure, and follow-up are depicted in Table 2. Pre-exenteration treatment Twenty-three patients were originally intended to be treated by definitive radiation therapy. Three patients received “definitive” surgery prior to exenteration. Patient 25 was treated with a radical hysterectomy and never received curative radiation therapy. Patient 12 originally refused radiation therapy and was treated with radical hysterectomy, and then received external radiation after a vaginal cuff recurrence. Patient 19 underwent external radiation and after an unsatisfactory response underwent surgery. After persistent central disease, she had a total exenteration. Clinical results Sixteen of the 26 patients (62%) are currently without cancer recurrence. Two of these 16 patients died of other causes not related to the cancer or the therapy. Five-year actuarial disease-specific survival and localregional control are 63% and 66% (Figs. 1 and 2). The cumulative patterns of failure were largely local, with 80% of our failing patients maintaining distant control until their demise. Of the 10 failures (including Patient 12, who died of therapeutic complications), 9 (90%) had close margins,

Table 1. Demographic

characteristics n

Entered Resections for cancer Follow-up (months) mean (median)

range Squamous cell: adenocarcinoma Age (y=s) mean (median) range Initial stage Stage I Stage II Stage III Stage IVA Prior initial therapy External RT: no external RT Tandem and ovoid insertion No tandem and ovoid insertion

27 26 29.5 (20.5) 6.1-81.6 25:l 46.4 (50.0) 29-76 8129%) 9(33%) 7(26%) 3 (11%) 25:l

19 6

Pelvic exenterations and brachytherapy l Table 2. Clinical Patient no.

Age

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

29 71 62 57 67 53 76 49 67 39 40 49 41 46 47 55 55 47 32 69 38 59 53 40 51 52

IIB IVA IIB IIB IIIA IIIB IIB IB IIB IIIB IB IIIB IIA IB IB IVA IIB IIB IIB IIIB IB IIB IB IB IB IIIB

Disease status NED NED NED NED DOD DOC DOD DOD NED NED DOD DOD DOD DOD DOD DOD NED NED AWD NED NED NED NED NED DOC NED

+

+ + + + + + + + + + + + -

145

er al.

results

Close or positive margins

LVUI’NI

Stage

J. J. BEITLER

+ +

Follow-up months

Site of recurrence

65.6 55.1 8.4 28.6 8.2 14.1 7.6 11.0 55.7 35.6 7.6 6.1 7.9 6.7 21.1 13.6 64.6 20.0 17.3 55.9 16.5 22.5 48.7 28.8 57.2 81.6

Local/regional Local/regional Local/regional Distant None Local/regional Local/regional Local/regional Local/regional

& distant

Local/regional

NED, no evidence of disease; DOD, dead of disease; AWD, alive with disease; DOC, died of other causes.

lymphovascular invasion (LVI), or perineural involvement (PNI). The presence of either close margins, positive LVI, or positive PNI, or a combination, adversely affected both disease-specific survival (p = 0.006) and local control (p = 0.013) (Figs. 3 and 4). Seven of 26 patients (27%) had margins 5 5 mm; 2 had no evidence of disease (NED), 4 were dead of disease (DOD), and 1 was alive with disease (AWD). Close margins with negative LVI and negative PNI negatively impacted local control (p = 0.04) compared to the 12

Disease

Specific

patients with negative margins, and negative LVI and PNI (Fig. 5). Seven of 26 patients (27%) had LVI or PNI with clear margins. Three were NED and four DOD. Local control was adversely affected by either positive LVI and/or positive PNI (p = 0.018) (Fig. 6). Two patients had microscopic nodal disease diagnosed after exenteration, and both died of disease. In the eight patients who had local failure, seven had positive LVI,

Local

Survival

Regional n

Propoii~on Smving

n

q

q

Control

26

26

24

36

40

Months

Fig. 1. Disease-specific

60

0

12

24

36

46

Lcxal Regional Control in Monlhs

survival.

Fig. 2. Local-regional

control.

60

I. J. Radiation Oncology 0 Biology 0 Physics

146

Effect

of Close/Positive Margin and/or +LVII+PNI on Disease Specific Survival

Effect of Close or Positive Margins on Local Control 1

1 1 q .i.

,6

Volume 38, Number 1, 1997

/

Ii

.2 --

r-,

.-...

_...

.8

Clear Margins, No LVI or PNI; n =I2

i.

~-i.

-..j

~

Log Rank

24

36

40

had positive PNI, three had close margins, and two had microscopic nodal disease (Table 3). Twelve of the 26 patients had positive LVI and/or positive PNI, and 8 of 12 were DOD (including 1 death due to complications), 1 of 12 is AWD, and and 3 of 12 never developed recurrence. Seven of the 26 patients had positive or close (i.e., 1.5 mm) margins, and 3 of these patients suffered local-regional disease recurrence. One patient with close margins developed distant disease, and one patient died of therapeutic complications. The patients who experienced local failures are summarized in Table 3.

four

1

‘I 4 .0

--FI

.&ar .-

M$ps,

I

and/or

.

12

24

36

/ I 48

60

Fig. 5. Effect of close or positive margins on local control. LVI = lymphovascular invasion; PNI = perineural involvement; close or positive margins, 55-mm margin.

DISCUSSION Our data demonstrate that patients with negative nodes, clear margins, and no lymphovascular invasion or perineural involvement do quite well; 11 of 12 (92%) of these good-risk patients had no cancer recurrence. For these patients, no further therapy would seem to be necessary. Patients with the bad prognostic factors mentioned above do worse, and their primary pattern of failure is local. Both of our patients with microscopically positive nodes died of disease, and despite the more recent evidence that adjuvant para-aortic nodal treatment can reduce metastatic spread and improve survival (19), known nodal disease is presently a contraindication to pelvic exenteration (14, 20). Those few who become 5year survivors

Effect of +LVI and/or +PNI on Local Control 1

I

Ng.LVI qr PNi;I n=12

.0

~-

..-.-

.~.

Months

+LVII*PNI

.6

.,

Margins, No._ VI or PNI; n-11 .-..-.-.

I I

60

Fig. 3. Effect of close/positive margin and/or positive LVUpositive PNI on disease-specific survival. LVI = lymphovascular invasion; PNI = perineural involvement; close/positive margins, 55mm margin.

of Close/Positive Margins On Local Control I _ /

Cl+

1 Log-Rank Te$ p=O.O4 /

Months

Effect

..-.

I

.. 0

12

._~._

‘=0.006

0 0

_ .-_/___-

. . f ..... /

,

: ... -.

j.

I *c--------1

i

-‘k-,i

i

%

t

.6

- -- --

I 0

12

24

36

4%

60

Months

Fig. 4. Effect of close/positive margins and/or positive LVI/ positive PNI on local control. LVI = lymphovascular invasion; PNI = perineural involvement; close/positive margins, c5-mm margin.

0

12

24

36

40

60

Months

Fig. 6. Effect of positive LVI and/or positive PNI on local control. LVI = lymphovascular invasion; PNI = perineural involvement.

Pelvicexenterationsand brachytherapyl Table 3. Local failure Time of

Patient no.

LVI

PNI

Margins

Nodes

5

+-

-

-

7 8++ 13 14 15 16 19

+-

+ + +

+ +

++ t + +

+ +

-

local failure 7.7 5.5 4 3.2 3.3 15.7 3.53 18.4

Time of death (months) 8.2 7.6 11 7.9 6.7 21.1 13.6

AWD

LVI, lymphovascular invasion; PNI, perineural involvement; AWD, alive with disease. despite nodal disease appear to have only microscopic and unilateral disease (7). In a multivariate analysis, Shingleton et al. (21) found that tumor at the margin was the strongest adverse prognostic factor; our log rank analysis (Fig. 5) agrees that with negative LVI and PNI, clear margins do affect local control. We note that only one of our patients with close margins developed metastatic disease, which underlines the importance of local control. Vascular invasion predicts for a higher rate of lymph node involvement (3, 23). In initial surgical treatment of Stage IB cervical cancer, LVI predicts for decreased 5year survival despite negative nodes and negative margins (4). The importance of PNI is unclear from the literature, and because of concordance with LVI, it may be difficult to ascertain in future studies. We have shown that in the recurrent or persistent setting, LVI and PNI predicted failure; even in the presence of clear margins, four of seven of these patients died of their disease. Three of these four failures maintained distant disease-free survival, once again implying that improved local-regional control might be expected to improve survival. The rate of unexpected involved margins may be 1625 % ( I,2 1), and the rate of lymphovascular invasion and/ or perineural involvement in our experience was 46%. Nine of our 10 failures had either close margins, positive LVI, or positive PNI. All five patients with close mar-

J.

J. BEITLER et al.

147

gins and positive LVI and/or positive PNI are DOD or AWD. Local-regional failure in those patients with close margins, LVI, and PNI appears to be quite predictable. Considering that 9 of 14 (64%) patients with close margins, positive LVI, and positive PNI were either DOD or AWD, and considering that 67% of failing patients suffered local regional failure alone, we believe that further local therapy is warranted. Brachytherapy has the potential to cure 43% of patients with close margins, positive LVI, or positive PNI treated with pelvic exenteration alone. These prognostic factors would be available during the periexenteration time frame, and we believe it would be prudent to place afterloading catheters and reserve the decision about using these catheters until it is known that these factors are absent. Despite a pelvic wall plasty which includes the complete covering of the residual tumor and tumor bed using omentum majus, rectus abdominus, and gluteal thigh flaps (12), a German group has reported a censored severe complication rate of 33% (10). We believe that 6-Gy fractions unnecessarily increase the late complication rate, and would prefer treating patients twice a day using 2-2.5-Gy fractions. The soft-tissue sarcoma trial proved that brachytherapy significantly increased local control in high-grade lesions after gross surgical resection (2, 18). Sarcomas are relatively radioresistent, but 42-45 Gy delivered over 46 days using low-dose rate Iridium-192 produced a 5-year local control rate of 89%. We would suggest that 40 Gy of geographically appropriate low-dose-rate brachytherapy and a lesser total dose of high-dose-rate brachytherapy would be sufficient to affect local control. We note little risk in the insertion of afterloading catheters at the time of pelvic exenteration; brachytherapy could be used after final pathology has been reviewed and the correct location of the catheters vis-8-vis the target volume is confirmed. Improved local regional control would be expected to positively impact survival; insertion of afterloading catheters in those patients with possible poor prognostic factors would seem to be justifiable. We believe that the study of this technique in patients undergoing total pelvic exenteration for recurrent cervical cancer who have unfavorable prognostic factors warrants further efforts.

REFERENCES 1. Averette, H. E.; Lichtinger, M.; Sevin, B.-U.; Girtanner, R. E. Pelvic exenteration: A 15-year experience in a general metropolitan hospital. Am. J. Obstet. Gynecol. 150:179184; 1984. 2. Brennan, M. F.; Hilaris, B.; Shiu, M. H.; Lane, J.; Magill, G.; Friedrich, C.; Hajdu, S. 1. Local recurrence in adult softtissue sarcoma. Arch. Surg. 122:1289- 1293; 1987. 3. Burghardt, E.; Pickel, H.; Haas, J.; Lahousen, M. Prognostic factors and operative treatment of Stage IB to IIB cervical cancer. Am. J. Obstet. Gynecol. 156:988-996; 1987. 4. Burke, T. W.; Hoskins, W. J.; Heller, P. B.; Bibro, M. C.; Weiser, E. B.; Park, R. C. Prognostic factors associated with

radical hysterectomy failure. Gynecol. Oncol. 26: 153- 159; 1987. 5. Coia, L.; Won, M.; Lanciano, R.; Martial, V. A.; Martz, K.; Hanks, G. The patterns of care outcome study for cancer of the uterine cervix. Cancer 66:2451-2456; 1990. 6. Coleman, R. L.; Keeney, E. D.; Freedman, R. S.; Burke, T. W.; Eifel, P. J. Radical hysterectomy for recurrent carcinoma of the uterine cervix after radiotherapy. Gynecol. Oncol. 55:29-35; 1994. 7. Disaia, P. J.; Creasman, W. T. Clinical gynecologic oncology, 3rd ed., St. Louis: C. V. Mosby; 1989:122- 123. 8. Eifel, P. J.; Burke, T. W.; Delclos, L.; Wharton, J. T.; Oswald, M. J. Early stage I adenocarcinoma of the uterine cer-

148

9. 10.

11.

12.

13. 14. 15. 16.

I. J. Radiation Oncology l Biology l Physics vix: Treatment results in patients with tumors 4 cm in diameter. Gynecol. Oncol. 41:199-205; 1991. Hockel, M.; Baussmann, E.; Mitze, M.; Knapstein, P. G. Are pelvic side-wall recurrences of cervical cancer biologically different from central relapses? Cancer 74:648-655; 1994. Hockel, M.; Hamm, H.; Knapstein, P. G.; Schlenger, K. Five-year experience with the Combined Operative and Radiotherapeutic Treatment (CORT) of recurrent gynecologic tumors infiltrating the pelvic wall. Gynecol. Oncol. 60:98; 1996. Hockel, M.; Knapstein, P. G. The combined operative and radiotherapeutic treatment (CORT) of recurrent tumors infiltrating the pelvic wall: First experience with 18 patients. Gynecol. Oncol. 46:20-28; 1992. Hockel, M.; Knapstein, P. G.; Kutzner, J. A novel combined operative and radiotherapeutic treatment approach for recurrent gynecologic malignant lesions infiltrating the pelvic wall. Surg. Gynecol. Obstet. 173:297-302; 1991. Hoskins, W. J. Prognostic factors for risk of recurrence in Stages Ib and IIa cervical cancer. Baillieres Clin. Obstet. Gynaecol. 2:817-828; 1988. Morley, G. W.; Hopkins, M. P.; Lindenauer, S. M.; Roberts, J. A. Pelvix exenteration, University of Michigan: 100 patients at 5 years. Obstet. Gynecol. 74:934-943; 1989. Parker, S. L.; Tong, T.; Bolder, S.; Wingo, P. A. Cancer Statistics 1996. CA 46:5-28; 1996. Parkin, D. M.; Pisani, P.; Ferlay, J. Estimates of the worldwide mortality from 18 major cancers in 1985. Int. J. Cancer 54:594-606: 1993.

Volume 38, Number 1, 1997 17. Perez, C. A.; Grigsby, P. W.; None, S. M., et al. Effect of tumor size on the prognosis of carcinoma of the uterine cervix treated with irradiation alone. Cancer 69:2796-2806; 1992. 18. Pisters, P. W. T.; Harrison, L. B.; Leung, D. H. Y.; Woodruff, J. M.; Casper, E. S.; Brennan, M. F. Long-term results of a prospective randomized trial of adjuvant brachytherapy in soft tissue sarcoma. J. Clin. Oncol. 14:859-868; 1996. 19. Rotman, M.; Pajak, T. F.; Choi, K.; Clery, M.; Mar&l, V.; Grigsby, P. W.; Cooper, J.; John, M. Prophylactic extendedfield irradiation of para-aortic lymph nodes in stages IIB, and bulky IB and IIA cervical carcinomas: Ten year results of RTOG 79-20. JAMA 274:387-393; 1995. 20. Rutledge, F. N.; Smith, J. P.; Wharton, J. T.; O’Quinn, A. G. Pelvic exenteration; Analysis of 296 patients. Am. J. Obstet. Gynecol. 129:881-892; 1977. 21. Shingleton, H. M.; Soong, S.-J.; Gelder, M. S.; Hatch, K. D.; Baker, V. V.; Austin, J. M. Clinical and histopathologic factors predicting recurrence and survival after pelvic exenteration for cancer of the cervix. Obstet. Gynecol. 73: 1027- 1034; 1989. 22. Stehman, F. B.; Bundy, B. N.; DiSaia, P. J.; Keys, H. M.; Larson, J. E.; Fowler, W. C. Carcinoma of the cervix treated with radiation therapy. A multivariate analysis of prognostic variables in the Gynecologic Oncology Group. Cancer 67:2776-2785; 1991. 23. van Nagel, J. R.; Donaldson, E. S.; Wood, E. G.; Parker, J. C. The significance of vascular invasion and lymphocytic infiltration in invasive cervical cancer. Cancer 41:228-234; 1978.