Pretreatment Surgical Staging of Patients with
The Case for Lymph Node Debulking
Jonathan A. Cosin, M.D.
Jeffrey M. Fowler, M.D.
M. Dwight Chen, M.D.
Pamela J. Paley, M.D.
Linda F. Carson, M.D.
Leo B. Twiggs, M.D.
The Women’s Cancer Center, The University of
Minnesota Hospital and Clinics, Minneapolis, Min-
Presented in part at the 10th International Meeting
of Gynecological Oncology (ESGO), Coimbra, Por-
tugal, April 26–May 2, 1997; and the 1997 Annual
Meeting of the American Radium Society, New
York, New York, April 30–May 4, 1997.
Address for reprints: Jonathan A. Cosin, M.D., Box
395, UMHC, 420 Delaware Street SE, Minneapolis,
Received September 8, 1997; accepted October
BACKGROUND. The routine use of extraperitoneal surgical staging prior to radiation
therapy in patients with bulky or locally advanced cervical carcinoma remains
METHODS. A review was performed of 266 patients with cervical carcinoma who
underwent extraperitoneal pelvic and paraaortic lymphadenectomy prior to re-
ceiving radiotherapy. Patients were divided into groups based on their lymph node
status. Group A had negative lymph nodes; Group B had resected, microscopic
lymph node metastases; Group C had macroscopically positive lymph nodes that
were resectable at the time of surgery; and Group D had unresectable lymph nodes.
All patients received standard radiotherapy utilizing external beam and brachy-
therapy. Patients with lymph node metastases received extended field irradiation.
Survival probabilities were computed by the Kaplan-Meier product limits method
with statistical significance determined by the Mantel-Cox log rank test.
RESULTS. Lymph node metastases were detected in 50% of patients. Five- and
10-year disease free survival rates were similar for all patients in Groups B and C.
All patients in Group D recurred. There was a 10.5% incidence of severe radiation-
related morbidity and a 1.1% incidence of treatment-related deaths.
CONCLUSIONS. Pretreatment extraperitoneal staging of patients with bulky or lo-
cally advanced cervical carcinoma may afford a survival benefit via the debulking
of macroscopically positive lymph nodes without significantly increasing treat-
ment-related morbidity or mortality. Cancer 1998;82:2241–8.
© 1998 American Cancer Society.
KEYWORDS: cervical carcinoma, radiation therapy, surgical staging, pretreatment,
findings, clinical staging is accurate in only approximately 60% of
cases.2,3In many instances, errors in staging are related to undiag-
nosed lymph node metastases.2,3Unfortunately, patients with lymph
node metastases have lower overall survival, disease free survival, and
survival after recurrence.4–6Although not permitted to affect Inter-
national Federation of Gynecology and Obstetrics staging, many mo-
dalities have been employed to determine the extent of disease prior
to therapy. The routine use of extraperitoneal surgical staging prior to
radiation therapy in patients with bulky or locally advanced disease
remains controversial. At issue is whether the identification and ag-
gressive primary treatment of involved lymph nodes results in im-
proved overall survival that is not overshadowed by an increase in
morbidity and mortality.
ervical carcinoma remains the only major gynecologic malig-
nancy that is clinically staged.1When compared with surgical
© 1998 American Cancer Society
Proponents of surgical staging cite the ability to
detect and then treat metastatic disease beyond the
standard radiation treatment fields.7,8Recent evi-
dence suggests a survival advantage both in patients
with paraaortic lymph node metastases who received
radiation to the paraaortic area as part of their initial
treatment6and in patients who received prophylactic
paraaortic radiation, despite having negative or un-
evaluated paraaortic lymph nodes.9Beyond this is the
potential therapeutic benefit of the debulking of large
lymph node metastases that are beyond the ability of
standard radiotherapy doses to sterilize.10,11
Those opposed to the routine use of pretreatment
surgical staging in cervical carcinoma argue that only
a small number of patients will benefit from extended
field treatment. This is based on the belief that most
patients with advanced disease will die of local failure,
making distant control irrelevant.12,13However, with
the use of multimodal therapies and with improve-
ments in brachytherapy, there is a trend toward im-
proved local control, thus increasing the importance
of measures that preferentially might improve distant
control.14,15Preirradiation surgical staging also has
been associated with an increase in late, radiation-
related morbidity and mortality, although this princi-
pally is for transperitoneal surgical approaches and
when radiation doses in excess of 5000 centigray (cGy)
are given to the extended fields.10,16
In 1989, Downey et al. published a series of 156
patients from the University of Minnesota with cervi-
cal carcinoma who underwent extraperitoneal staging
laparotomy prior to definitive radiation therapy.17
That report demonstrated similar survival for patients
with completely resected lymph nodes, whether mi-
croscopically or macroscopically involved (57% and
51%, respectively). However, there were only 18 pa-
tients with microscopically involved lymph nodes and
9 patients whose lymph nodes were unresectable.
Since that time, an additional 110 patients have un-
dergone surgical staging at our institution. With the
additional number of patients, we are now able to
expand on the original report by Downey et al. to
include more information on the impact of paraaortic
lymph nodes as well as the results of a multivariate
analysis of the data.
MATERIALS AND METHODS
Between 1978 and 1990, 276 patients with cervical
carcinoma underwent complete extraperitoneal pre-
treatment surgical evaluation at the University of Min-
nesota Hospital and Clinics. Two hundred and sixty-
six patients were evaluable because 10 patients were
excluded from the analysis because of nonstandard or
incomplete radiation not due to toxicity or disease
progression (8 patients) or loss to follow-up in ?12
months from diagnosis (2 patients). Median follow-up
Patients were divided into four groups based on
the histopathologic status of their lymph nodes.
Group A (133 patients) had no lymph node metasta-
ses, Group B (39 patients) had resected microscopi-
cally positive pelvic and/or paraaortic lymph nodes,
Group C (74 patients) had macroscopically positive
pelvic and/or paraaortic lymph nodes that were re-
sected completely at the time of surgery, and Group D
(20 patients) had unresectable lymph nodes. For the
purposes of this study, lymph nodes were considered
macroscopically positive if they were noted in the
surgical findings to be macroscopically enlarged and
matted, macroscopically enlarged with obvious tumor
on cut section, or if the pathology report noted the
lymph nodes to be positive macroscopically. Enlarged
lymph nodes with only microscopic foci of metastases
were classified as microscopically positive. In all pa-
tients, the preoperative intent was complete excision
of all involved lymph nodes unless there was evidence
nodes generally were unresectable because of dense
adherence to, or invasion of, vascular or nervous
Pretreatment selection and evaluation as well as
surgical and radiation therapy methods have been
described previously.17,18Briefly, all patients under-
went extraperitoneal exploration with complete pelvic
lymphadenectomy unless the lymph nodes were
deemed unresectable. The majority of patients (97%)
also underwent bilateral paraaortic lymphadenectomy
to at least 3–4 cm above the aortic bifurcation. Pelvic
washings and exploration of the abdominal contents
were performed through a small midline peritoneal
incision made just large enough to admit the sur-
geon’s hand. Radiation treatment planning then was
based on surgical and pathologic results and utilized a
combination of external beam and brachytherapy.
Standard pelvic fields for lymph node negative pa-
tients extended up to the level of the fifth lumbar
vertebral body. Patients with pelvic lymph node me-
tastases received an additional 4500–5000 cGy of ex-
tended field radiation up to the level of the second
lumbar vertebrae. If the paraaortic lymph nodes were
involved, the treatment field was extended to the level
of the tenth thoracic vertebrae. Prior to 1987, all pa-
tients with positive peritoneal cytology also received a
dose of 2000 cGy to the whole abdomen. This practice
was discontinued when data from the original series
showed that the presence of malignant cells in the
months (range, 48–217
2242 CANCER June 1, 1998 / Volume 82 / Number 11
peritoneal fluid was not of independent prognostic
Patients were considered to have a complete re-
sponse if there was no radiologic, clinical, or patho-
logic evidence of disease at 6 months from the onset of
treatment. All other patients were considered to have
All treatment-related complications were re-
corded. Complications were considered minor if the
patient had symptoms requiring no significant inter-
vention. This corresponds to Grade 3 based on the
Radiation Therapy Oncology Group (RTOG) classifica-
tion for treatment-related morbidity. All complica-
tions requiring surgery or major medical intervention
were considered severe (RTOG Grade 4).
Survival probabilities were computed by the
Kaplan–Meier product limits method with statistical
significance determined by the Mantel–Cox log rank
test. The uncorrected chi-square test was used to de-
termine the significance of the nonparametric statis-
tics on outcome. Factors predictive of survival in the
univariate analysis were compared in a multivariate
logistic regression analysis.
Lymph node metastases were found in 133 of the 266
evaluable patients (50%). These were pelvic only in 87
patients, pelvic and paraaortic in 44 patients, and
paraaortic only in 2 patients. In the original series by
Downey et al. patients were placed into groups based
on the status of their pelvic lymph nodes only.17For
this update, patients instead were classified based on
their overall lymph node status. Reclassification of
patients by pelvic lymph node status did not alter the
probability of disease free survival for any group sig-
nificantly (P ? 0.10 for all groups).
The breakdown of the groups by stage, grade, and
cell type is shown in Table 1. The groups did not differ
significantly with respect to age, grade, or cell type.
There was a higher percentage of advanced stage pa-
tients in Groups C and D. The only histologic subtypes
that were found to be of independent prognostic sig-
nificance by multivariate analysis were undifferenti-
ated and small cell. Patients with adenosquamous cell
histologic type did not have a poorer prognosis than
patients with squamous cell or adenocarcinoma sub-
The surgical procedure and radiation methods
both remained consistent throughout the period of
the study. A larger number of patients treated since
1987 received chemotherapy either as radiation po-
tentiation, in the adjuvant setting, or both (36% vs.
22%; P ? 0.023). Agents utilized include cisplatin alone
(48 patients), hydroxyurea alone (3 patients), cisplatin,
vincristine, and bleomycin (17 patients), or hydroxy-
urea followed by cisplatin (1 patient). Patients in
Groups C and D were more likely to receive chemo-
therapy than patients in Groups A and B (57% and 89%
vs. 4% and 23%, respectively; P ? 0.0001). These pa-
tients were included in the analysis because a univar-
iate analysis showed no survival difference between
the patients in Group C who received chemotherapy
and those who did not (P ? 0.10). Overall, the use of
chemotherapy was shown to have a negative impact
on both disease free and overall survival based on the
multivariate analysis (RR [relative risk] ? 2.01 and
1.90, respectively; P ? 0.005). This is presumed to
reflect selection bias.
Radiation doses to Points A and B were similar in
all groups, with overall mean doses of 8229 cGy to
Point A and 5857 cGy to Point B. All patients with
positive lymph nodes received irradiation to the
paraaortic fields as described earlier, with a mean
dose of 4576 cGy. The median number of days from
surgery to the first day of radiation therapy was 8
(range, 4–146 days). This included four patients
treated on an experimental ‘‘ultrahigh risk’’ protocol
that involved a planned four courses of chemotherapy
prior to radiation therapy.
The Kaplan–Meier probability of disease free sur-
vival by pelvic and paraaortic lymph node status is
shown in Figure 1 and Table 2. The disease free sur-
vival at 5 and 10 years for Group A (75% and 68%,
respectively) was significantly better than that for
Groups B (43% and 35%, respectively), C (50% and
46%, respectively), and D (0% at 3 years) (P ? 0.05 for
all groups). Groups B and C showed similar survival
probabilities, although the median disease free sur-
vival for Group C (macroscopically positive, resected
lymph nodes) was twice that of Group B (microscop-
ically positive lymph nodes only; 65 vs. 34 months
Prognostic Factors of Lymph Node Groups
Group AGroup B Group C Group D
IB or IIA
III or IV
Number in parentheses indicates percentage of lymph node group.
Pretreatment Surgical Staging of Cervical Carcinoma/Cosin et al. 2243
respectively; P ? 0.10) (Table 2). When corrected for
stage, Group C demonstrated significantly better sur-
vival than Group B for both low (FIGO I and II) and
high (III and IV) stage patients (P ? 0.05). Both Groups
B and C demonstrated significantly better survival
than Group D (P ? 0.001). Most remarkable is the
difference in median survival between patients in
Groups C and D (65 months vs. 2 months, respective-
ly). This difference was maintained when the survival
data were corrected for stage.
Two hundred and fifty-eight patients underwent
paraaortic lymph node dissection. Of these, 46 pa-
tients (18%) had positive paraaortic lymph nodes. This
included 2 patients (1.5%) with negative pelvic lymph
nodes, both of whom had macroscopically positive
but resectable paraaortic lymph nodes. Three of the
patients with microscopically positive pelvic lymph
nodes (7.5%) had positive paraaortic lymph nodes.
This included two patients with microscopically and
one patient with macroscopically positive but resected
paraaortic lymph nodes. There was a 45% incidence
(41 of 91) of positive paraaortic lymph nodes among
patients with macroscopically involved pelvic nodes.
Overall, there were 13 patients with microscopically
positive paraaortic lymph nodes, 20 patients with
macroscopically positive paraaortic lymph nodes, and
13 patients with unresectable paraaortic lymph nodes.
When compared with stage, paraaortic lymph node
metastases were present in 17 Stage I patients (16%),
15 Stage II patients (12%), 13 Stage III patients (41%),
and 1 Stage IV patient (33%). The 5-year disease free
survival of the patients with resected, macroscopically
positive paraaortic lymph nodes was 43% (Fig. 2). The
median survival for these patients was 29 months.
The most significant adverse prognostic factor for
disease free survival by multivariate analysis was the
presence of unresectable lymph nodes (RR ? 4.89; P ?
0.0001). Other independent prognostic factors include
advanced stage (RR ? 2.32; P ? 0.001), and pelvic or
paraaortic lymph node metastases (RR ? 2.44; P ?
0.0002 and RR ? 2.48; P ? 0.01, respectively). Patients
with histologic subtypes other than squamous, adeno-
carcinoma, or adenosquamous also had poorer overall
FIGURE 1. Disease free survival by
lymph node group. —: negative; –I–:
microscopic; –?–: macroscopic, re-
sected; –Œ–: unresectable.
Recurrence Free and Median Survival By Lymph Node Group and Stage
Lymph node group (% lymph node group)Stage (%)
ABCDI IIIII and IV
Median survival (mos)
2244CANCER June 1, 1998 / Volume 82 / Number 11
and disease free survival (RR ? 3.67 and 3.83, respec-
tively; P ? 0.005).
The association between survival and the number
of positive lymph nodes was evaluated in the group of
patients with completely resected pelvic and paraaor-
tic lymph nodes. A survival difference was not ob-
served until patients with ?4 positive lymph nodes
were compared with patients with ?3 positive lymph
nodes (P ? 0.05). This significance only was observed
in the multivariate analysis when overall survival and
not disease free survival was used as an endpoint.
Disease free survivals of 77 and 143 months were
observed in patients with 10 and 12 macroscopically
positive but resected lymph nodes, respectively. The
latter patient underwent resection of both pelvic and
paraaortic lymph node metastases.
Comparison was made between lymph node sta-
tus and recurrence patterns. Complete responses were
observed in 125 patients in Group A (94%), 35 patients
in Group B (90%), 69 patients in Group C (93%), and 7
patients in Group D (35%). Only the patients in Group
D were more likely to have persistent disease after
primary treatment (P ? 0.0001 vs. all other groups).
Among complete responders, there were a total of 22
recurrences in Group A (17.6%), 15 recurrences in
Group B (43%), 29 recurrences in Group C (42%), and
7 recurrences in Group D (100%). Patients in Groups B
and C were equally as likely to recur, but were more
likely to recur than patients in Group A (P ? 0.002).
Not only were patients in Groups B and C equally
likely to recur, but these recurrences had some ele-
ment of distal disease in 73% and 66% of patients,
respectively (P ? 0.94). Recurrence patterns are sum-
marized in Table 3.
The entire group was examined for the incidence
There were 3 (1.1%) treatment-related deaths. Two
were caused by peritonitis secondary to bowel perfo-
ration in patients who received paraaortic radiation.
Both patients were free of disease at the time of death
24 and 44 months, respectively, after their original
surgery. Both had macroscopically positive lymph
nodes resected at the time of pretreatment surgery
and received radiation to the upper and lower
paraaortic fields. One patient died of a pulmonary
embolus during her second intracavitary cesium ap-
plication. Another patient developed a nonfatal pul-
monary embolus during radiation therapy for an over-
all incidence of pulmonary embolus of 0.75%.
There was a 19.9% incidence of lymphocysts di-
agnosed. The majority of these cases (35) were asymp-
tomatic and found on routine physical examination.
Eighteen patients (6.7%) developed lymphocysts that
required intervention (transcutaneous drainage in 8
patients and laparotomy in 10 patients) because they
were symptomatic or infected. Lymphedema devel-
oped in 18.4% of patients, with equal distribution
FIGURE 2. Disease free survival by
paraaortic lymph node status. —: neg-
ative; –I–: microscopic; –?–: macro-
scopic, resected; –Œ–: unresectable.
Recurrence Patterns By Lymph Node Groups
Lymph node groups
Number in parentheses indicates percentage of complete responders.
Pretreatment Surgical Staging of Cervical Carcinoma/Cosin et al. 2245