Estimation of the optimal brachytherapy utilization rate in the treatment of carcinoma of the uterine cervix

Estimation of the Optimal Brachytherapy Utilization Ratein the Treatment of Carcinoma of the Uterine CervixReview of Clinical Practice Guidelines and Primary Evidence

Stephen Thompson, MBBSGeoff Delaney, MBBS, MDGabriel Sam Gabriel, MBBS, MPHSusannah Jacob, MBBS, MD, MHAPrabir Das, MBBS, MPHMichael Barton, MBBS

Collaboration for Cancer Outcomes Research andEvaluation (CCORE), Liverpool Hospital, Sydney,Australia.

BACKGROUND. Brachytherapy (BT) is an integral part of cervical carcinoma treat-

ment. There have been no attempts to estimate the optimal proportion of new cervi-

cal carcinoma cases that should be treated with BT, that is, the optimal rate of

brachytherapy utilization (BTU).

METHODS. Evidence-based guidelines and primary evidence were used to construct a

BTU tree for carcinoma of the uterine cervix. Searches were performed of the epide-

miological literature to ascertain the proportion of patients who fulfilled criteria for

BT. The robustness of the model was tested by sensitivity analyses and by peer review.

A patterns of care study of BT in New South Wales for 2003 was conducted, and

actual BTU for cervical carcinoma determined. The differences between optimal and

actual rates of BTUwere assessed.

RESULTS. The optimal cervical carcinoma BTU was 49% (range, 42% to 50%). In New

South Wales in 2003, actual BTU was only 30% of 256 cervical carcinoma patients.

The major discrepancy was for FIGO stage IB-IIA disease, where there was an under-

utilization of BT, estimated to be 15% actual use compared with 47% optimal use. In

Surveillence, Epidemiology, and End Results (SEER) areas, there was underutilization

for stage IB-IIA (22% actual BTU versus 47% optimal BTU) and for stage IIB-IVA (54%

actual BTU versus 100% optimal BTU).

CONCLUSIONS. BT for cervical carcinoma is underutilized in New South Wales and in

SEER areas. The authors’ model of optimal BTU can be used as a quality assurance

tool to provide an evidence-based benchmark against which actual patterns of prac-

tice can be measured. The model can also be used to help determine adequacy of BT

resource allocation. Cancer 2006;107:2932–41.� 2006 American Cancer Society.

KEYWORDS: cervical carcinoma, carcinoma of the uterine cervix, brachytherapy,radiotherapy, utilization rate, patterns of care studies.

T he incidence of cervical carcinoma is decreasing in first world

countries such as Australia.1 Despite this, and the promise of

newly developed cervical carcinoma vaccines,2 cervical carcinoma is

still the third largest cancer killer of women world-wide, causing

274,000 deaths in 2002.3 Cervix cancer is a curable cancer, but achiev-

ing the best results depends on well-organized and appropriately

resourced cancer services. Previous studies have shown best results to

be critically dependent on service-related factors such as dose and

number of patients treated.4–8 Brachytherapy (BT) is an integral part

of the cervical carcinoma treatment armamentarium. It is a techni-

cally demanding and highly specialized method of radiotherapy (RT)

delivery. Depending on equipment usage, capital expenditures and

staff costs may be high. However, there have been no attempts to esti-

We thank the Court of Reviewers who critiqued theutilization tree: Prof David Allen, Prof JohnathanCarter, Prof Neville Hacker, Prof Roger Houghton,Prof Graeme Morgan, Prof Kailash Narayan, ProfMichael Quinn, Dr Colin Bull, Dr Catherine Clark,Dr Viet Do, and Dr Tony Proietto.

Address for reprints: Stephen Thompson, MBBS,Department of Radiation Oncology, Prince of WalesHospital, High St, Randwick, NSW 2031, Australia;Fax: (011) 61 2 9382 2550; E-mail: [email protected]

Received August 23, 2006; revision receivedSeptember 25, 2006; accepted September 26,2006.

ª 2006 American Cancer SocietyDOI 10.1002/cncr.22337Published online 15 November 2006 in Wiley InterScience (www.interscience.wiley.com).

2932

mate the optimal proportion of new cervical carci-

noma cases that should be treated with BT.

The current study had two aims: 1) The first was to

estimate the optimal brachytherapy utilization (BTU)

rate for treatment of carcinomas of the uterine cervix.

The optimal BTU was the ideal proportion, based on

the best available evidence, of those patients who

should have been treated with BT at least once during

the course of their illness. 2) The second aim was to

compare the optimal BTU with actual BTU. The actual

BTU was defined as the proportion of patients with cer-

vical carcinoma who received at least 1 course of BT

during the course of their illness.

MATERIALS AND METHODSOptimal Brachytherapy UtilizationMethods similar to those described by Tyldesley et al.9

and Delaney et al.10 in their estimates of optimal exter-

nal beam radiation therapy (EBRT) utilization rates in

the treatment of various cancers were used to develop a

model of the optimal BTU for cervical carcinoma.

Review of evidence for the efficacy of brachytherapySystematic Medline searches, manual review of biblio-

graphies, and expert advice were used to identify

English-language evidence-based clinical practice

guidelines for management of cervical carcinoma.

These were used to derive indications and contraindi-

cations for and against the use of BT. Table 1 shows the

guidelines that were identified and their order of prece-

dence.

When guidelines did not adequately cover indica-

tions for BT, the primary evidence itself was reviewed.

The evidence was prioritized according to the Austra-

lian National Health and Medical Research Council21

hierarchy of levels of evidence.

Construction of brachytherapy utilization treesThe indications and contraindications for BTwere used

to construct a utilization tree. TreeAge Data 3.5 software

was used to diagrammatically represent the utilization

tree and to calculate results.

Each of the patient, tumor, or treatment-related

attributes that affected the indication for or against BT

was represented on the utilization tree by a branch,

with each branch of the tree leading to further branches

that, in turn, represent other factors that affected the

management decision on the need for BT. Each termi-

nal branch of the tree has only 1 of 2 possible out-

comes; BT was indicated as a result of this particular

combination of circumstances (represented by a ‘‘1’’ in

the column to the immediate right of the terminal

branch), or BTwas not indicated in these circumstances

(represented by a ‘‘0’’ in the column to the immediate

right of the terminal branch).

Incidence of brachytherapy indicationsThe proportions of patients with particular attributes

that corresponded to each branch on the utilization

tree were determined from the best available epidemio-

logical evidence. These proportions are shown below

their corresponding branch on the tree. Systematic

Medline searches, manual review of bibliographies, and

TABLE 1Hierarchy of Precedence of Guidelines

Hierarchy Type Guideline

1 Multinational FIGO: Federation Internationale

de Gynecologie et d’Obstetrique

staging classifications and clinical

practice guidelines in the

management of gynaecologic

cancers11

2.1 Australian national None identified

2.2 Other national NIH: National Institutes of Health

Consensus Development Conference

Statement on Cervical Cancer12

PDQ: National Cancer Institute

PDQ Statement on theManagement

of Cervical Cancer13

NCCN: National Comprehensive Cancer

Network Clinical Practice Guidelines in

Oncology: Cervical Cancer14

3.1 Australian state NSW: New SouthWales Gynaecological

Oncology Study Group: Gynaecological

Oncology Clinical Practice Guidelines15

3.2 Other state/

provincial

BCCA: British Columbia Cancer Agency

Guidelines on Cervical Cancer16

YCN: Yorkshire Cancer Network

Guidelines for theManagement of

Gynaecological Cancers17

4.1 Australian

multi-institutional

SGOG: The Sydney Gynaecologic

Oncology Group, Royal Prince

Alfred and Liverpool Hospitals:

Clinical Practice and Management

Policies18

4.2 Other

multi-institutional

None identified

5.1 Australian

single-institutional

None identified

5.2 Other

single-institutional

None identified

6 Single-disciplinary ABS: American Brachytherapy

Society Recommendations for

Low-Dose-Rate Brachytherapy for

Carcinoma of the Cervix19

ACOG: American College of

Obstetricians and Gynaecologists

Clinical Management Guidelines:

Diagnosis and Treatment of

Cervical Carcinomas20

Use of Brachytherapy in Cervical Cancer/Thompson et al. 2933

expert advice were used to identify English-language lit-

erature that provided data used to estimate the size of

these proportions. Usually more than 1 epidemiological

source of data for each branch of these utilization trees

was identified. Competing epidemiological data were

ranked by quality as shown in Table 2 by using criteria

modified by Delaney et al.10 that were previously deter-

mined by Tyldesley et al.9 Precedence was given to

higher ranking epidemiological data, as these data were

less likely to suffer from referral bias and sampling

error. Where 2 or more sources of data of equivalent

quality (based on criteria in Table 2) were found, a

weighted average of the results was calculated. If large

differences in incidences existed between similar stu-

dies, then the effect of this uncertainty on the BTU esti-

mate was estimated by sensitivity analysis.

Calculation of the optimal brachytherapy utilization rateThe proportion of patients with each indication for BT

was calculated by multiplying the proportions of

patients with each individual attribute that constituted

the BT indication. This proportion is shown in the sec-

ond column to the right of the terminal branch. The

optimal BTU was calculated by adding the proportions

of patients for all terminal branches that end with an

indication for BT.

Expert reviewTo obtain multidisciplinary input and remove potential

for investigator bias, a panel of experts in cervical can-

cer radiation oncology and gynecological oncology

from across Australia formed a Court of Reviewers for

the model. This Court of Reviewers was drawn from the

group who were asked to review the study of optimal

rates of EBRT.10 Minor changes were recommended

and incorporated.

Sensitivity analysisUncertainties in the indications for BT or in the epide-

miological data were potential sources of error in the

model. The robustness of the utilization tree and the

magnitude of the potential sources of error were mod-

eled by sensitivity analyses. TreeAge Data 3.5 software

enables 1-way sensitivity analyses to be performed,

altering a single variable (such as an indication for BTor

the proportion of patients with a particular attribute

that affects a treatment decision) to assess the effect

that this change has on the optimal BTU.

Actual Brachytherapy UtilizationThe actual BTU for cervical carcinoma was determined

as part of a Patterns of Care Study for 2003 for New

SouthWales (NSW). NSW is the largest state of Australia,

with a population of 6.6 million in 2003, 63% of whom

live in Sydney.22 Site visits were made to all 9 radiation

oncology departments in NSW that deliver BT and de-

identified data on all patients who were treated with BT

in 2003 were extracted from patient records. Information

on all NSW cervical carcinoma patients for 2003 was

obtained from the Health Outcomes Information Statisti-

cal Toolkit (HOIST) database. HOIST contains data from

the NSW Central Cancer Registry. It does not contain

data on stage of disease for cervical carcinoma. The likely

stage distribution of cervical carcinoma in NSW was

assumed to be similar to that of the United States, and,

therefore, the incidence of each stage of cervical carci-

noma in NSW was extrapolated from the Surveillance,

Epidemiology, and End Results (SEER) Database.23 SEER

data for 1997–2001 were used to estimate the distribution

of stages and substages because no other source pro-

vided this level of information. Comparison of cervical

screening rates24,25 and overall cervical carcinoma stage

distribution in US and Australia23,26 showed similar

results, suggesting that the SEER substage distributions

are likely to be representative of those in Australia. The

actual BTU was calculated by dividing the number of

NSW cervical carcinoma patients treated with BT by the

estimated stage incidence of cervical carcinoma. Ethics

approval was obtained from South Western Sydney Area

Health Service (Western Zone) Human Research Ethics

Committee.

RESULTSOptimal Brachytherapy Utilization RateThe BT utilization tree for cervical carcinoma is shown

in Fig. 1.

Table 3 lists the patient, tumor, and treatment-

related attributes for each indication for BT, Austra-

lia’s National Health and Medical Research Council

(NHMRC) level of evidence for the indication, guide-

TABLE 2Hierarchy of Epidemiological Data9,10

Quality of

source Source type

a Australian National Epidemiological data

b Australian State Cancer Registry

g Epidemiological databases from other

large international groups (eg, SEER)

d Results from reports of a random sample from a population

e Comprehensive multi-institutional database

z Comprehensive single-institutional database

u Multi-institutional reports on selected

groups (eg, multi-institutional clinical trials)

l Single-institutional reports on selected groups of cases

m Expert opinion

2934 CANCER December 15, 2006 / Volume 107 / Number 12

FIGURE 1. This is the decision tree for optimal brachytherapy utilization to treat cervical carcinoma. LVI indicates lymphatic vascular space invasion; PLND, pelviclymph node dissection; LN, lymph nodes; GOG, Gynecology Oncology Group.

Use of Brachytherapy in Cervical Cancer/Thompson et al. 2935

lines that support it, and the proportion of patients with

uterine cervical carcinoma that have that indication for

BT. Of patients with cervical carcinoma, 49% have an in-

dication for BT (the optimal BTU). The optimal BTU for

FIGO stage IA disease is 14%, for IB-IIA disease 47%, for

IIB-IVA disease 100%, and for IVB disease 0%.

Table 4 summarizes epidemiological data related to

indications for or against BT: the proportion of patients

with cervical carcinoma with the BT indication or con-

traindication, the quality of epidemiological evidence,

and sources of data.

Two issues require further explanation. There were

no data available on the proportion of patients with

nonmetastatic cancer who were unfit for surgery

(branches G, H, K, and P, Fig. 1). We used NSW commu-

nity self-reported health status,30 adjusting the results

by the known age distribution of patients with cervical

cancer1: 19% of this population reported their health to

be in the 2 worst of 5 categories (‘‘fair’’ or ‘‘poor’’). It

was assumed that these patients would not be fit for

radical surgical resection, apart from therapeutic coni-

zation in the case of stage IA1 disease. All other medi-

cally unfit patients would be referred for RT including

BT. We chose this measure because it is NSW popula-

tion-based data and because self-reported health status

is a predictor of later morbidity and death.44,45 We per-

formed sensitivity analysis to estimate the effect of

considering only the 4% of patients with ‘‘poor’’ self-

reported health as nonoperable candidates. Note that

the 19% figure is similar to that obtained by using data

from Tyldesley et al.,31 who used the Canadian National

Population Health Survey to estimate the proportion of

Canada’s population to be of poor performance status

by age category (Eastern Cooperative Oncology Group

[ECOG] scores of 2 or 3 and 4). These data were as-

sumed to be representative of the performance status

of Australian patients with nonmetastatic cervical can-

cer, and data were adjusted for the age distribution of

Australian patients with cervical cancer1 to calculate

the average percentage of Australian patients with non-

metastatic cervical cancer who had poor performance

status. This average was 16%.

The other issue is whether patients with stage IB-

IIA disease should be managed with primary surgery or

RT (branches O, Fig. 1). There has been 1 randomized

trial of surgery versus RT including BT for operable cer-

vical carcinoma.33 Although disease-free survival rates

and overall survival rates were identical in both arms of

the study, the authors suggested that morbidity of RT

(particularly in postmenopausal women) was less.

TABLE 3Cervical Carcinoma: Indications for Brachytherapy—Levels and Sources of Evidence

Outcome no. Clinical scenarioTreatmentindicated

Level ofevidence References

Percentage of

all cervixcancer patients

3 Stage IA1 (Cone biopsy); LVI negative;

Margin positive; Medically inoperable

BT IV 12, 13, 14, 16, 19 1.0

5 Stage IA1 (Cone biopsy); LVI positive;

Medically operable (Hyst and PLND);

LN negative; Local recurrence

EBRTand BT IV 12, 13, 14, 15, 18 0.1

7 Stage IA1 (Cone biopsy); LVI positive;

Medically inoperable

EBRTand BT IV 12, 13, 14, 16, 19 0.2

9 Stage IA2 (Cone biopsy); Medically

operable (Hyst and PLND); LN negative;

Local recurrence

EBRTand BT IV 12, 13, 14, 15, 18 0.3

11 Stage IA2 (Cone biopsy); Medically inoperable EBRTand BT IV 12, 13, 14, 16, 19 2.2

13 Stage IB-IIA; <4 cm; Medically operable

(Hyst and PLND); Margin negative;

LN negative; GOG score<120; Local recurrence

(nil distant metastases); Central local recurrence

EBRTand BT IV 12, 13, 14, 15, 18 0.4

17 Stage IB-IIA; <4 cm; Medically operable

(Hyst and PLND); Margin positive;

Vaginal margin positive

EBRTand BT IV 12, 13, 14, 18, 19 0.9

19 Stage IB-IIA; <4 cm; Medically inoperable EBRTand BT IV 12, 13, 15, 16, 17, 18, 19 5.6

20 Stage IB-IIA; >4 cm EBRTand BT II, III 12, 13, 16, 18, 19 13.3

21 Stage IIB-IVA EBRTand BT III 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 25

Total 49

LVI indicates lymphatic vascular space invasion; Hyst, hysterectomy; PLND, pelvic lymph node dissection; LN, lymph node.

2936 CANCER December 15, 2006 / Volume 107 / Number 12

TABLE 4Cervical Carcinoma: Incidence of Attributes Used to Define Indications for Brachytherapy

Key*Population orsubpopulation of interest Attribute

Proportion of

population withthis attribute

Quality ofinformation References

A All registry cancers Gynecological cancer 0.045 a AIHW1

B All gynecological cancer Cervical carcinoma 0.194 a AIHW1

C All cervical carcinoma Stage IA 0.26 g SEER23

D Stage IA Stage IA1 0.55 g SEER23

E Stage IA1, (conization) LVI negative 0.92 e Ostor AG27

F Stage IA1, (conization);

LVI negative

Margin positive 0.60 z Jones et al.28

0.10 z Creasman et al.29

G Stage IA1, (conization);

LVI negative;

Margin positive

Medically operable (Hyst) 0.81 b NSWHealth Surveys30

0.84 g Tyldesley et al.31

H Stage IA1, (conization);

LVI positive

Medically operable

(Hyst and PLND)

0.81 b NSWHealth Surveys30

0.84 g Tyldesley et al.31

I Stage IA1, (conization);

LVI positive; Medically operable,

(Hyst and PLND

LN negative 0.97 e Ostor AG27

J Stage IA1, (conization);

LVI positive; Medically operable,

(Hyst and PLND); LN negative

Local recurrence 0.06 e Ostor AG27

K Stage IA2 Medically operable

(Hyst and PLND)

0.81 b NSWHealth Surveys30

0.84 g Tyldesley et al.31

L Stage IA2; Medically operable,

(Hyst and PLND)

LN negative 0.94 e Ostor AG27

M Stage IA2; Medically operable,

(Hyst and PLND); LN negative

Local recurrence 0.03 e Ostor AG27

N All cervical carcinoma Stage IB/IIA 0.43 g SEER23

O Stage IB/IIA ‘‘Non-bulky’’disease 0.69 z Fuller et al.32

0.67 u Landoni et al.33

P Stage IB/IIA;

‘‘Non-bulky’’disease

Medically operable

(Hyst and PLND)

0.81 b NSWHealth Surveys30

0.84 g Tyldesley et al.31

Q Stage IB/IIA; ‘‘Non-bulky’’disease;

Medically operable, (Hyst and PLND)

Margin negative 0.94 u Landoni et al.33

R Stage IB/IIA; ‘‘Non-bulky’’disease;

Medically operable; (Hyst and PLND);

Margin negative

LN negative 0.75 u Landoni et al.33

0.84 z Kim et al.34

0.86 z Kamura et al.35

S Stage IB/IIA; ‘‘Non-bulky’’disease;

Medically operable, (Hyst and PLND);

Margin negative; LN negative

‘‘Low risk’’ for recurrence

(GOG score< 120)

0.84 e Delgado et al.36

0.76 u Landoni et al.33

0.73 z Kridelka et al.37

T Stage IB/IIA; ‘‘Non-bulky’’disease;

Medically operable, (Hyst and PLND);

Margin negative; LN negative;

‘‘Low risk’’ for recurrence (GOG score< 120)

Local recurrence without

distant metastases

0.07 e Delgado et al.,36 Samlal et al.38

0.08 u Landoni et al.33

U Stage IB/IIA; ‘‘Non-bulky’’disease;

Medically operable, (Hyst and PLND);

Margin negative; LN negative;

‘‘Low risk’’ for recurrence

(GOG score< 120); Local recurrence

without distant metastases

Central local relapse 0.38 z Look et al.39

0.40 z Chung et al.40

0.42 z Wang et al.41

0.48 z Burke et al.42

V Stage IB/IIA; ‘‘Non-bulky’’disease;

Medically operable, (Hyst and PLND);

Margin positive

Vaginal margin positive 0.64 z Snidjers-Keilholz et al.43

W All cervical carcinoma Stage IIB-IVA 0.25 z SEER23

X All cervical carcinoma Stage IVB 0.06 g SEER23

LVI indicates lymphatic vascular space invasion; Hyst, hysterectomy; PLND, pelvic lymph node dissection; LN, lymph node; GOG, Gynecology Oncology Group.

* Key refers to the corresponding branch of the Utilization Tree in Figure 1.

Use of Brachytherapy in Cervical Cancer/Thompson et al. 2937

Others have contended that this trial actually showed

that morbidity of RT was greater than that of surgery

alone.46 Both FIGO11 and the available Australian

guidelines15,18 recommend surgery for patients with

nonbulky disease who are suitable surgical candidates.

In Australian practice, the vast majority of these pa-

tients are treated by surgery. The exceptions are usually

patients with poor performance status or significant

comorbidity or who refuse surgery. We have, therefore,

followed the Australian guidelines and Australian prac-

tice and calculated based on medically fit patients with

nonbulky IB-IIA disease being treated with primary

surgery and medically unfit patients receiving a defini-

tive RT including BT. Patients with bulky stage IB-IIA

disease have a high risk of locoregional recurrence after

surgery alone.33,47 For these patients, randomized trials

have shown EBRT and BT to be equal to surgery and

EBRT,33 surgery and EBRT to be superior to surgery

alone,47 and completion hysterectomy following RT to

have no benefit.48 Toxicity is increased by combined-

modality treatment.47 For this reason, and the lack of

evidence for a benefit for combined modality treatment

over RT (especially with chemotherapy) alone, guide-

lines12,13,16,18,19 recommend definitive treatment with

chemoradiation, rather than surgery and postoperative

EBRT, for patients with bulky stage IB-IIA disease.

Sensitivity analysisThere were 3 variables with considerable uncertainty in

the epidemiological data. By using reports in the litera-

ture of lowest and highest estimates for these variables,

we performed sensitivity analyses (see Fig. 2) on the

model to estimate the effect that these uncertainties

might have on the final BTU.

1. Proportion of patients with technically operable dis-

ease who were not fit for an operation (Branches G, H, K,

P, Fig. 1). As discussed above, this proportion was cal-

culated at 19%,1,30,31 with sensitivity analysis showing

that if this proportion was reduced to 4%, then the over-

all optimal BTU fell from 49% to 42%, with the rate for

stage IA disease falling from 14% to 4% and the rate for

stage IB-IIA disease falling from 47% to 37%.

2. Margin positivity rate after conization for stage IA1

(Branches F, Fig. 1). Two series have reported margin

positivity after conization for stage IA1 disease. In Jones

et al. 30 of 50 patients28 and in Creasman et al. 4 of 39

patients29 had an average margin positivity rate of 38%

(for these 89 patients). Sensitivity analysis was underta-

ken, varying margin positivity rate from 10% to 60%. As

a result, the BTU for stage IA disease varied from 12% to

16%, and the overall optimal BTU varied from 48% to

50%.

3. Proportion of patients with non-bulky stage IB-IIA

disease and negative surgical margins who also had neg-

ative lymph nodes (Branches R, Fig 1). Extrapolation

from data contained in Landoni et al.33 indicates that

between 73% and 80% of patients with negative mar-

gins also have negative lymph nodes. Varying these

values had no effect on the overall optimal BTU for cer-

vical carcinoma or the rate for stage IB-IIA disease.

Comparison with Actual PracticeA comparison of optimal BTU with actual BTU by stage

is shown in Table 5. Note that for the calculation of

BTU by stage, the denominator is the stage incidence of

cervical carcinoma in NSW, which is an extrapolation

from SEER data23 because local population-based data

FIGURE 2. A sensitivity analysis of optimal brachytherapy utilization for cervical carcinoma is presented here. LVI indicates lymphatic vascular space invasion; PLND,pelvic lymph node dissection; LN, lymph nodes; GOG, Gynecology Oncology Group.

2938 CANCER December 15, 2006 / Volume 107 / Number 12

were not available (see above). Therefore, the figures for

actual BTU by stage are estimates. In NSW in 2003, 77

(30%) of 256 cervical carcinoma patients were treated

with BT, less than two-thirds of the optimal percentage

of 49%. For all stages except stage IB-IIA, the optimal

and the actual BTU were similar. For FIGO stage IA, the

estimated actual BTU was 0%, compared with 14%

(range, 4%–16%) optimal; for stage IIB-IVA disease esti-

mated 92% actual BTU and 100% optimal BTU; and for

stage IVB disease estimated 7% actual and 0% optimal.

Themajor discrepancy was for stage IB-IIA, where there

was an underutilization of BT, estimated 15% actual

compared with 47% optimal BTU (range, 37%–47%).

Only 2 other sources of data on actual practice

were available. Prospective surveys of RT practice were

conducted in Sweden over 12 week periods in 199249

and in 2001.50 In 1992, 39% of patients with cervical car-

cinoma were treated with BT. In 2001, the actual rate

had fallen to 28%. According to the SEER database,23

the corresponding figures for 1992 and 2001 in the US

were 28% and 24%. The SEER figures are for ‘‘first

course of cancer directed treatment’’. BT for disease re-

currence would not be included but only accounts for

0.8% of all indications for BT. The discrepancy appeared

to be most marked for stage IB-IIA disease and for stage

IIB-IVA disease.

DISCUSSIONWe have constructed an evidence-based peer-reviewed

model that assesses the optimal brachytherapy utiliza-

tion rate in the treatment of cervical carcinoma. The

optimal BTU is 49%, (range, 48% to 50%).

Our model of optimal BTU can be used as a quality

assurance tool by comparing the optimal BTU with

actual practice. Compared with our model of optimal

practice, BTwas underutilized in the treatment of cervi-

cal carcinoma in NSW, Sweden, and the US. For NSW,

the underutilization was essentially restricted to FIGO

stages IB-IIA (approximately 15% actual versus 47%

optimal). This is surprising given that earlier research

has shown that the actual utilization of EBRT is close to

optimal, for all stages including early stage disease.51 In

that report, Delaney et al. hypothesized that the reason

the optimal and actual rates were similar was that per-

haps cervical carcinoma is most commonly treated by

specialist gynecological oncologists in a multidisciplin-

ary setting, as opposed to other tumor sites that may be

more commonly managed in a general setting. This

implies that patients with high-risk stage IB-IIA in

NSW are probably being inappropriately treated with

primary surgery and then require adjuvant EBRT,

rather than being treated in accordance with guide-

lines12,13,16,18,19 and evidence33,47,48 with definitive

radiotherapy including BT. The EBRT utilization will

remain optimal in either case, but, in the former case,

the BTU will be suboptimal. Disease control is likely to

be similar with either approach but with greater risk of

toxicity for the patients treated with both surgery and

EBRT.33,47

More concerning is the underuse of BT for patients

with stage IIB-IVA disease in the SEER registry data.

Earlier research showed that for this group of patients

utilization of EBRT was close to optimal.51 There are 3

possible explanations for these 2 observations. Possibly

these patients are being treated with surgery and EBRT

rather than EBRT and BT, against the recommendation

of all guidelines.11–20 Another possible explanation is

that these patients are being treated with definitive RT

using EBRTonly, against guideline recommendations11–20

and with a high likelihood of suboptimal dose, and

reduced local control and survival.4–8,52,53 The third

possibility is that SEER has not reliably recorded the

use of BT in these patients. There are no confirmatory

US BT utilization sources, but for EBRT, SEER’s figures

correlate well with those of the American College of

Surgeons.51 For Sweden, there is no explanation in the

literature of why BT is underutilized for cancer of the

uterine cervix.

This study of optimal and actual BTU for cervical

carcinoma has a number of limitations. As shown in Ta-

ble 3 above, there is a paucity of high-level evidence for

many of the BT indications and contraindications. It

may also be that the model is unable to fully account

TABLE 5Comparison Between Optimal and Actual BTU

FIGO stageOptimal BTUpercentage

Actual BTUpercentage

NSW 2003 Sweden 1992 Sweden 2001 SEER 1992 SEER 2001

IA 14 (4–16) 0 n/a n/a 2 2

IB-IIA 47 (37–47) 15 n/a n/a 28 22

IIB-IVA 100 92 n/a n/a 59 54

IVB 0 7 n/a n/a 11 18

All 49 (42–50) 30 39 28 28 24

Use of Brachytherapy in Cervical Cancer/Thompson et al. 2939

for individual patient circumstances. An example is

increasing patient age, which Delaney et al.10 found to

predict increased use of RT. Another issue, as discussed

above, is the use of the SEER database to give the stage

distribution of cervical cancer for the optimal model

and to provide an estimate of the incidence by stage of

cervical cancer in NSW to enable calculation of the

actual stage BTU in NSW in 2003.

Our model of optimal BT utilization can also be

used to assist in planning BT resource requirements.

In the report on Asia Pacific Region radiotherapy

resources, Tatsuzaki and Levin54 assumed that 50% of

cervical carcinoma patients in the first world would

require BT—similar to our evidence-based estimate of

49%. With the insertion of an appropriate stage distri-

bution into the model, it could be used to provide evi-

dence for or against another Tatsuzaki and Levin

assertion54 that 80% of cases in the developing world

would require BT.

ConclusionsThis study used an evidence-based peer-reviewed

method of calculating the optimal proportion of all

patients with carcinoma of the uterine cervix who

should be treated with BT at some time during the

course of their illness. This optimal brachytherapy utili-

zation rate was 49%. A patterns of care study of BT in

NSW in 2003 has shown that the actual rate of use of BT

is only 30%, with most of the undertreatment occurring

in patients with early stage, potentially operable dis-

ease. Comparison with the SEER database showed that

there is underuse of BT in the US, with this undertreat-

ment occurring in early and in more advanced stages of

disease.

We have used this model of optimal BTUas a quality

assurance tool, thus providing an evidence-based bench-

mark against which we have measured actual patterns of

practice. This model can also be used to determine the

adequacy of current BT resource allocation and to for-

mulate a plan for future BT resource allocation.

REFERENCES1. Australian Institute of Health and Welfare (AIHW). Cancer

incidence data: Cancer age-standardised database. (2000). At

URL: http://www.aihw.gov.au/cancer/datacubes/index.cfm

2. Lowndes CM, Gill ON. Cervical cancer, human papillomavi-

rus, and vaccination. BMJ. 2005;331:915–916.

3. Parkin DM, Bray F, Ferlay J, Pisani P. Global Cancer Statistics,

2002. CA Cancer J Clin. 2005;55:74–108.

4. Lanciano RM, Won M, Coia LR, Hanks GE. Pretreatment and

treatment factors associated with improved outcome in squa-

mous cell carcinoma of the uterine cervix: a final report of the

1973 and 1978 patterns of care studies. Int J Radiat Oncol Biol

Phys. 1991;20:667–676.

5. Hanks GE, Herring DF, Kramer S. Patterns of care outcome

studies. Results of the national practice in cancer of the cer-

vix. Cancer. 1983;51:959–967.

6. Coia L, Won M, Lanciano R, Marcial VA, Martz K, Hanks G.

The Patterns of Care Outcome Study for cancer of the uterine

cervix. Results of the Second National Practice Survey. Cancer.

1990;66:2451–2456.

7. Montana GS, Martz KL, Hanks GE. Patterns and sites of failure

in cervix cancer treated in the USA in 1978. Int J Radiat Oncol

Biol Phys. 1991;20:87–93.

8. Komaki R, Brickner TJ, Hanlon AL, Owen JB, Hanks GE. Long-

term results of treatment of cervical carcinoma in the United

States in 1973, 1978, and 1983: Patterns of Care Study (PCS).

Int J Radiat Oncol Biol Phys. 1995;31:973–982.

9. Tyldesley S, Boyd C, Shulze K, Walker H, Mackillop WJ. Esti-

mating the need for radiotherapy for lung cancer: an evi-

dence-based, epidemiologic approach. Int J Radiat Oncol Biol

Phys. 2001;49:973–985.

10. Delaney GP, Jacob S, Featherstone C & Barton MB. Radiother-

apy in cancer care: estimating optimal utilisation from a

review of evidence-based clinical guidelines. Collaboration

for Cancer Outcomes Research and Evaluation (CCORE),

Liverpool Hospital, Sydney, Australia. 2003. Available at URL:

http://www.ncci.org.au/supp/reports/genreports_2003-2005.

htm

11. Benedet JL, Bender H, Jones H 3rd, Ngan HY, Pecorelli S.

FIGO staging classifications and clinical practice guidelines

in the management of gynecologic cancers. FIGO Committee

on Gynecologic Oncology. Int J Gynecol Obstet. 2000;70:209–

262.

12. National Institutes of Health. Consensus Statement: National

Institutes of Health consensus development conference state-

ment on cervical cancer. Gynecol Oncol. 1997;66:351–361.

13. National Cancer Institute. US National Institutes of Health.

CancerNet PDQ Cancer Information Summaries: Treatment

of Cervical Cancer. 2003. Available at URL: http://cancernet.nci.

nih-gov/cancertopics/pdq/treatment/cervical/healthprofes-

sional

14. National Comprehensive Cancer Network. Clinical Practice

Guidelines in Oncology - v.1.2003 - Cervical Cancer. Available

at URL: http://www.nccn.org/professionals/physician_gls/

pdf/cervical.pdf. 2003. 2004.

15. NSW Gynaecological Oncology Study Group. Gynaecological

Oncology Clinical Practice Guidelines. 2004. Sydney: NSW

Gynaecological Oncology Study Group; 2004.

16. British Columbia Cancer Agency. Cancer Management Guide-

lines. Gynecology. Item 4, Uterine Cervix. 2001. Available at

URL: http://www.bccancer.bc.ca/HPI/CancerManagement-

Guidelines/Gynecology/UterineCervix1of2/default.htm [Ac-

cession date February 2004.]

17. Yorkshire Cancer Network. Guidelines for the Management of

Gynaecological Cancers. 2003. Available at URL: http://york-

shire-cancer-net.org.uk/html/publications/clinical_gynaecology.

htm [Accession date: June 15, 2004.]

18. Sydney Gynaecologic Oncology Group, Sydney Cancer Centre.

Clinical Practice and Management Policies. Johnathan Carter.

1–6-2004. Sydney: Johnathan Carter; 2004. Available at URL:

http://www.gynaecancer.org.au/

19. Nag S, Chao C, Erickson B, et al. The American Brachytherapy

Society recommendations for low-dose-rate brachytherapy

for carcinoma of the cervix. Int J Radiat Oncol Biol Phys. 2002;52:33–48.

20. Committee on Practice Bulletins-Gynecology. ACOG practice

bulletin. Diagnosis and treatment of cervical carcinomas,

number 35, May 2002.Obstet Gynecol. 2002;99(5 pt 1):855–867.

2940 CANCER December 15, 2006 / Volume 107 / Number 12

21. National Health and Medical Research Council. A guide to the

development, evaluation and implementation of clinical

practice guidelines. Appendix B, 56. Canberra: National

Health and Medical Research Council; 1999. Available at URL:

[email protected]

22. Australian Bureau of Statistics. New SouthWales. National Re-

gional Profile 1999 to 2003. Release date 24/11/2005. Available

at URL: www.abs.gov.au/

23. Surveillance, Epidemiology, and End Results (SEER) Program

(www.seer.cancer.gov) SEER*Stat 5.2.2 database. SEER Cancer

Incidence Public-Use Month 2004 Database (1973–2001),

National Cancer Institute, DCCPS, Surveillance Research Pro-

gram, Cancer Statistics Branch, released Month Year, based

on the Month Year submission.

24. NSW Department of Health. Report of The New South Wales

Chief Health Officer. 2006. Adelaide: Centre for Epidemiology

and Research, Population Health Division, NSW Department

of Health; 2006.

25. Sirovich BE, Welch HG. The frequency of Pap smear screening

in the United States. J Gen Intern Med. 2004;19:243–250.

26. South Australian Cancer Registry. 2000 Cancer Report—

Epidemiology of Cancer in South Australia 1977–1999. Ade-

laide: Department of Health, Government of South Australia;

2000.

27. Ostor AG. Pandora’s Box or Ariadne’s Thread? Definition and

prognostic significance of microinvasion in the uterine cervix.

Squamous lesions. Pathol Annu. 1995;(30 pt 2):103–136.

28. Jones WB, Mercer GO, Lewis JL Jr, Rubin SC, Hoskins WJ. Early

invasive carcinoma of the cervix. Gynecol Oncol. 1993;51:26–32.

29. Creasman WT, Fetter BF, Clarke-Pearson DL, Kaufmann L,

Parker RT. Management of stage IA carcinoma of the cervix.

Am J Obstet Gynecol. 1985;153:164–172.

30. NSW Department of Health. Electronic Report of the 1997

and 1998 NSW Health Surveys. Available at URL: www.

health.nsw.gov.au/public-health/nswhs

31. Tyldesley S, Zhang-Salomons J, Groome PA, et al. Association

between age and the utilization of radiotherapy in Ontario.

Int J Radiat Oncol Biol Phys. 2000;47:469–480.

32. Fuller AF, Elliott N, Kosloff C, Hoskins WJ, Lewis JL Jr. Deter-

minants of increased risk for recurrence in patients under-

going radical hysterectomy for stage IB and IIA carcinoma of

the cervix. Gynecol Oncol. 1989;33:34–39.

33. Landoni F, Maneo A, Colombo A, et al. Randomised study of

radical surgery versus radiotherapy for stage Ib-IIa cervical

cancer. Lancet. 1997;350:535–540.

34. Kim SM, Choi HS, Byun JS. Overall 5-year survival and

prognostic factors in patients with Stage IB and IIA cervi-

cal cancer treated by radical hysterectomy and pelvic

lymph node dissection. Int J Gynecol Cancer. 2000;10:305–

312.

35. Kamura T, Tsukamoto N, Tsuruchi N, et al. Multivariate analy-

sis of the histopathologic prognostic factors of cervical cancer

in patients undergoing radical hysterectomy. Cancer. 1992;69:

181–186.

36. Delgado G, Bundy B, Zaino R, Sevin BU, Creasman WT,

Major F. Prospective surgical-pathological study of disease-

free interval in patients with stage IB squamous cell carci-

noma of the cervix: a Gynecologic Oncology Group study.

Gynecol Oncol. 1990;38:352–357.

37. Kridelka FJ, Berg DO, Neuman M, et al. Adjuvant small field

pelvic radiation for patients with high risk, stage IB lymph

node negative cervix carcinoma after radical hysterectomy

and pelvic lymph node dissection. A pilot study. Cancer. 1999;

86:2059–2065.

38. Samlal RA, van der Velden J, Ten Kate FJ, Schilthuis MS,

Hart AA, Lammes FB. Surgical pathologic factors that pre-

dict recurrence in stage IB and IIA cervical carcinoma pa-

tients with negative pelvic lymph nodes. Cancer. 1997;80:

1234–1240.

39. Look KY, Rocereto TF. Relapse patterns in FIGO stage IB carci-

noma of the cervix. Gynecol Oncol. 1990;38:114–120.

40. Chung CK, Nahhas WA, Stryker JA, Curry SL, Abt AB, Mortel R.

Analysis of factors contributing to treatment failures in stages

IB and IIA carcinoma of the cervix. Am J Obstet Gynecol. 1980;

138:550–556.

41. Wang CJ, Lai CH, Huang HJ, et al. Recurrent cervical carci-

noma after primary radical surgery. Am J Obstet Gynecol. 1999;

181:518–524.

42. Burke TW, Hoskins WJ, Heller PB, Shen MC, Weiser EB, Park

RC. Clinical patterns of tumor recurrence after radical hyster-

ectomy in stage IB cervical carcinoma. Obstet Gynecol. 1987;

69(3 pt 1):382–385.

43. Snijders-Keilholz A, Hellebrekers BW, Zwinderman AH, van de

Vijver MJ, Trimbos JB. Adjuvant radiotherapy following radical

hysterectomy for patients with early-stage cervical cancer

(1984–1996). Radiother Oncol. 1999;51:161–7.

44. Idler EL, Benyamini Y. Self-rated health and mortality: a re-

view of twenty-seven community studies. J Health Soc Behav.

1997;38:21–37.

45. McCallum J, Shadbolt B, Wang D. Self-rated health and sur-

vival: a 7-year follow-up study of Australian elderly. Am J Pub-

lic Health. 1994;84:1100–1105.

46. Soutter P. Radical surgery or radiotherapy for stage IB-IIA cer-

vical cancer. Lancet. 1997;350:532.

47. Sedlis A, Bundy BN, Rotman MZ, Lentz SS, Muderspach LI,

Zaino RJ. A randomized trial of pelvic radiation therapy versus

no further therapy in selected patients with stage IB carci-

noma of the cervix after radical hysterectomy and pelvic

lymphadenectomy: a Gynecologic Oncology Group study.

Gynecol Oncol. 1999;73:177–183. Comment in: Gynecol Oncol.

1999;73:175–176.

48. Keys HM, Bundy BN, Stehman FB, et al. Radiation therapy

with and without extrafascial hysterectomy for bulky stage IB

cervical carcinoma: a randomized trial of the Gynecologic

Oncology Group. Gynecol Oncol. 2003;89:343–353.

49. [No authors listed]. A prospective survey of radiotherapy in

Sweden. Acta Oncol. 1996;35suppl 6:47–56.

50. Moller TR,BrorssonB,Ceberg J, et al. Aprospective surveyof radio-

therapypractice 2001 in Sweden.ActaOncol. 2003;42:387–410.

51. Delaney G, Jacob S, BartonM. Estimation of an optimal radio-

therapy utilization rate for gynecologic carcinoma: part I—

malignancies of the cervix, ovary, vagina and vulva. Cancer. 2004;

101:671–681.

52. Perez CA, Breaux S, Madoc-Jones H, et al. Radiation therapy

alone in the treatment of carcinoma of uterine cervix. I. Anal-

ysis of tumor recurrence. Cancer. 1983;51:1393–1402.

53. Lanciano RM, Martz K, Coia LR, Hanks GE. Tumor and treat-

ment factors improving outcome in stage III-B cervix cancer.

Int J Radiat Oncol Biol Phys. 1991;20:95–100.

54. Tatsuzaki H, Levin CV. Quantitative status of resources for

radiation therapy in Asia and Pacific region. Radiother Oncol.

2001;60:81–89.

Use of Brachytherapy in Cervical Cancer/Thompson et al. 2941