 
|
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Journal of the New Zealand Medical Association,
13-February-2009, Vol 122 No 1289
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Sentinel lymph node biopsy experience in Taranaki: a
prospective audit in a provincial New Zealand hospital
Emily Davenport, Michael W Fancourt, William T C Gilkison,
Steven M Kyle, Damien A Mosquera
Axillary lymph node status is considered the most important
prognostic factor for patients with early stage breast
cancer.1 While axillary clearance is accepted
as the gold standard in detection of metastatic nodal disease, sentinel lymph
node biopsy (SLNB) is increasingly adopted as an alternative approach. Early
research suggests that SLNB is a reliable method of predicting lymph node
status, and may spare women the morbidity associated with axillary
clearance.2,3 Nonrandomised studies of SLNB
followed by axillary clearance have demonstrated that one or more sentinel lymph
node can be identified in more than 90% of patients with invasive breast cancer,
with a false negative rate of less than 10%.4,5
In an era of increasing centralisation of surgical services,
little published data exists on adoption of SLNB in peripheral centres. In
Taranaki, a provincial New Zealand centre, SLNB techniques have been employed
since 2002. This study aims to provide a prospective comparison of SLNB in
Taranaki with established standards.
MethodsEligibility and study design—All
patients in the Taranaki region, with early stage operable breast cancer,
confirmed histologically or radiologically, and requiring axillary staging as
part of their surgical management are eligible for SLNB. Major exclusion
criteria include tumour size greater than 5.0 cm in diameter, palpable axillary
lymph nodes, failure to give consent, and previous axillary surgery. The initial
training set (49 patients) has provided information about outcome measures.
These 49 patients were included in the final data analysis (152 SLNB). Written
informed consent was received from all patients prior to surgery.
Surgery and patient
management—Sentinel node identification was performed by either
the blue dye or scintigraphic techniques, or a combination, depending on surgeon
preference and available equipment. All patients preparing for sentinel node
biopsy received a Technetium 99 antimony colloid subcutaneous injection either
the day of, or the day prior to surgery. The time between injection and surgery
was at least 3 hours. Nuclear medicine staff identified the sentinel node on
lymphoscintiscan and marked its location on the skin. In the operating room, 2
ml of patent blue dye in 0.5 ml aliquots were injected immediately after the
induction of anaesthesia.
Prior to June 2006 four peritumoural injection sites
were used both for technetium and blue dye injections (51 patients). Since June
2006, injection sites have been periareolar. SLNB was performed prior to tumour
resection or mastectomy, and nodes that were hot (at least 10 times background
activity), or blue were removed. Wide local excision or mastectomy was then
performed in the usual manner. In a single case, the sentinel lymph node was
sent as a frozen section for immediate cytological examination. This patient
underwent immediate axillary clearance for a positive SLNB. If no sentinel lymph
node was identified, an immediate axillary clearance was undertaken.
Patients in the training set underwent immediate
axillary clearance (level I and II nodes). Patients in the non-training set
underwent axillary clearance only if the SLNB was positive or not obtained. All
patients were otherwise treated identically.
The Taranaki Breast Cancer
Database—This computerised database was established in 2003 as a
collaborative effort of the specialist breast care nurse, general surgical
consultants, and information systems staff. It is a data storage system which
categorises data on patient characteristics, risk factors, diagnostic methods,
tumour characteristics, axillary status, surgical procedures, and adjuvant
therapy. All patients in Taranaki with histologically or cytologically confirmed
breast cancer are entered into the database on a prospective basis.
Confidentiality is maintained at all times.
Pathological aspects—Each
sentinel LN is submitted in its entirety in approximately <5 mm portions per
cassette. Each cassette specimen is then sectioned into five levels (3 micron
width) to be examined. One of these levels is stained with a pancytokeratin
AE1/AE3, the others with haematoxylin and eosin (H&E).
ResultsBetween October 2002 and August 2007, 152 SLNB were
undertaken in 151 patients. Of these, almost half were performed by one surgeon
(Table 1).
Table 1. Sentinel lymph node biopsy (SLNB) by
surgeon between October 2002 and August 2007
From initiation in 2002, the number of SLNB performed
per year has steadily increased. This trend seems to have reached a plateau in
2006, with around 50 SLNB per year (Figure 1).
Figure 1. Number of SLNBs per year performed in
Taranaki
 The overall identification rate for SLNB was 92.7%. When
analysed separately, the identification rate for non-training cases was 97.1%,
and for training cases was 83.7%, indicating a sharp increase in identification
rates pre and post training. An average of 2.0 nodes per SLNB were removed, with
a range of 0–15.
In 70 (46%) cases, surgical technique involved preoperative
lymphoscintigraphy and intraoperative use of the gamma probe alone. In 8 (5%)
cases blue dye alone was used. These cases occurred when a functional gamma
probe was not available, or when nuclear medicine staff were unavailable to
perform lymphoscintigraphy. In 74 (49%) cases both blue dye and
lymphoscintigraphy with intraoperative use of the gamma probe were employed.
Of the 152 SLNBs attempted, 141 were successfully identified
with 11 failures (7.2%). Of the 141 successful biopsies, 40 (26.3%) were
positive and 100 (65.8%) negative for nodal malignancy. The positive biopsies
can be further subdivided into 9 biopsies positive for micrometastases only, and
31 positive for macroscopic malignancy. One specimen was lost prior to
histological examination.
The test performance measures for SNB are available for the
training case data and are summarised in Table 2. They indicate a sensitivity of
75% (6 of 8), a false negative rate of 5% (2 of 40) and a negative predictive
value of 92% (23 of 25). Individual surgeon test performance is demonstrated in
Table 3. The lost SLN was picked up from theatre, but never arrived at the
labaratory. An internal investigation was conducted but failed to locate the
node.
Table 2. Test performance measures for sentinel
lymph node biopsy in Taranaki (training cases only, n=40)
Table 3. Individual Taranaki surgeon test
performance
† Cases where SLNB was combined with routine
axillary clearance; ‡ cases performed under mentor guidance from Surgeon
A; § surgeon A attended a formal training course in 2002; SLNB=sentinel
lymph node biopsy; SLN=sentinel lymph node.
Of the 152 cases of SLNB, axillary clearance was performed
in 77 (50.7%). This includes the 49 training cases who underwent routine
axillary clearance. The indications for axillary clearance are summarised in Figure 2. An average of 13 nodes were removed (range
1–25) with an average of 5 positive nodes (range 0–17).
Eighteen (23.4%) of 77 axillary clearances were positive: 1
for micrometastases only and 17 for invasive metastases. Results are summarised
in Figure 2. One patient had a positive SLNB (two of six nodes) but declined
further surgical treatment.
Average tumour size was 2.8 cm. 86 patients (59%) had
tumours ≤2.0 cm, 57 patients (39%) had tumours >2.0cm but ≤5.0cm,
and 4 patients (3%) had tumours >5.0 cm. Of the four patients with tumours
>5.0 cm, all were reported to have tumours estimated at <5.0 cm in
preoperative radiological reports. Increasing tumour size correlated with an
increased likelihood of positive nodal status (Figure 3).
Figure 3. Predicting nodal status based on
tumour size
 Histology results showed a predominance of invasive ductal
cell carcinoma, with or without an in situ component, (n=114 cases
[81%]). Invasive lobular carcinoma was diagnosed in 16 (12%), ductal cell
carcinoma in situ alone in 9 (6%). Single cases of adenomyoepithelioma,
medullary carcinoma, and invasive papillary carcinoma were also reported.
Using the modified Bloom-Richardson cytological grading
system, tumour grade was reported for 147 cases. Of these, 65 (44%) were
diagnosed as grade 1 tumours, 44 (30%) grade 2, and 38 (28%) grade 3.
Of the nine patients with SLNB positive for micrometastases
only, six had micrometastases >0.2 mm, two had micrometastases ≤0.2 mm
and one did not have size stated in the pathology report. Management of
micrometastases is summarised in Table 4.
Table 4. Management of micrometastases in
Taranaki compared to American Society of Clinical Oncology (ASCO) guideline
recommendations6
Note: For patients 2 and 3, management
decisions were made prior to the publication of ASCO guidelines. For patient 7,
the decision not to proceed to axillary clearance was multifactorial, and
included patient age, comorbidities, and reluctance for further surgery.
Ten patients who underwent SLNB had ductal carcinoma in
situ (DCIS) alone on final histology results. Of these 10 patients, 6
underwent mastectomy. Of the remaining 4, 2 underwent wide local excision (WLE)
for known DCIS with a possible invasive component on biopsy, and 2 underwent WLE
for DCIS only on biopsy.
DiscussionTaranaki is a provincial New Zealand centre staffed by four
general surgical consultants, who serve a population of just over 100,000. All
four surgeons regularly perform SLNB as part of management of patients with
early stage breast cancer.
151 patients underwent 152 SLNB over a period of almost 5
years. The numbers of SLNB increased steadily over the first 3 years after
introduction of the technique, and now seems to be levelling off at around 50
procedures per year. This equates to one SLNB per 2000 population.
Initial patient accrual was slow, with potential patients
being managed by any of the four general surgeons who adopted the technique at
differing dates throughout 2002 and 2003. The introduction of SLNB at Taranaki
followed the attendance of one general surgeon, Surgeon A, at a formal SLNB
training course in 2002. Adoption of the technique by the remaining three
surgeons was mentored by Surgeon A.
None of the remaining three surgeons attended a formal
course in the technique as recommended by the ASCO Guidelines for SLNB in early
stage breast cancer.6 However, prior research
has failed to identify a statistically significant difference in false negative
rates among surgeons who have undertaken a formal training course and those who
have not.7
The use of mentoring, proctored cases and formal training in
accredited continuing medical education courses is thought to reduce the
personal case experience necessary to achieve optimal results, but this effect
has yet to be quantified.
Accumulated data from many multi-centre trials continues to
support the need to perform 20 cases of SLNB in combination with axillary
dissection, or to perform 20 SLNB procedures with mentoring, as being necessary
to minimise the risk of false-negative
results.5,7
Among the four surgeons at Taranaki, the number of SLNB in
combination with routine axillary clearance (AC) performed as test cases was
highly variable (Table 3). Surgeons C and D, who performed fewer training cases,
and did not attend formal training, are shown to have higher identification
rates and lower false negative rates. However, this may represent a biased
figure due to the small numbers of cases and of positive axillae for these
surgeons.
There are two key parameters to successful SLNB: the
successful identification rate and the false negative rate. Identification rate
is defined as the proportion of patients in whom a SLN is identified and
removed. For SLNB to be a useful test, it is essential that a SLN is identified
in the majority (>90%) of patients. The overall successful identification
rate in Taranaki is 93%. In non-training cases, the successful identification
rate is 97%. This exceeds acceptable standards. It also identifies an expected
initial learning curve.
The false-negative rate is defined as the proportion of
patients with axillary nodal metastases who have a negative SLN biopsy. An
acceptable false negative rate has been previously defined as 10% or
less.4,5 However, there is a problem with
this calculation when a predetermined false
negative rate of 10% is set, as there is no way
of predicting the percentage of node
positivity. If 25% of 40 cases are node-positive (10 cases), then a
surgeon with only 1 false negative would have a
false negative rate of 10%. However, a surgeon
who has 50% node positivity in 40 cases (20 cases)
will have the same false negative rate (10%) with 2 false
negative cases.
To remove this bias, the false negative rate can be
calculated as a percentage of the total number
of patients rather than as a percentage of the
positive axillae. This issue was addressed in the statement
from the Consensus Conference Committee in
Philadelphia.8 This method of calculating false
negative rate as a percentage of total cases was used in the ALMANAC trial
validation phase, which assessed whether surgeons were competent to proceed to
the randomisation phase.9
The false negative rate in Taranaki is 5% when calculated as
a percentage of total cases. If calculated as a percentage of positive axillae,
the false negative rate is 25%. The latter is an unacceptably high false
negative rate, however (as explained above) it may represent significant bias.
A recent meta-analysis including more than 8000 patients
showed that the reported false negative rate ranged from 0.0% to 29.4% across
studies.10 The false negative rate was
significantly lower in the 23 studies that included >100 patients compared
with the 46 studies that included <100 patients (p=0.007). Twenty-one studies
(36.2%) reported a false negative rate >10%.
Janis, P et al11 show that
unobtainably large numbers of SLNB cases are required to make any reliable
conclusions regarding the quality of SLNB. They calculate that it will take 750
patients with 300 tumor-positive basins to establish with 95% certainty that a
surgeon who has a nonidentification rate of 5% and a false-negative rate of 5%
indeed has these capabilities within a range of 0% to 7%. Therefore
Taranaki’s high false negative rate when calculated as a percentage of
positive axillae may be secondary to bias due to small sample size and a small
number of positive axillae.
We have chosen to accept the false negative rate of 5% as a
percentage of total number of SLNB as a more accurate indicator of patient
safety. This false negative rate is within established acceptable standards.
The inclusion and exclusion criteria for consideration of
SLNB used in Taranaki may have affected test performance. In Taranaki, patients
with tumour diameters up to 5.0 cm, including those with multifocal or
multicentric disease, are included for SLNB. Many of the seminal reports on
SLNB, including the recent Australasian Sentinel Lymph Node versus Axillary
Clearance I (SNAC I) trial, only included patients with tumours <3.0 cm
diameter and excluded patients with multicentric or multifocal
disease.12
Currently accepted test performance standards, against which
Taranaki’s test performance measures have been compared, were set based on
results from these early reports. Increasing evidence suggests that SLNB for
tumours up to 5.0cm diameter (including multifocal and multicentric disease) is
feasible with similar test performance measures as smaller, unifocal
disease.13,14 Indeed, the SNAC II trial, which
follows on from the SNAC I trial has extended its inclusion criteria to include
tumours >3.0 cm and those with multifocal or multicentric disease.
Initial reports of SLNB in breast cancer were based on a
technique involving peritumoural injection of either radioactive colloid or blue
dye. Subsequent experience has shown that
subdermal,15
intradermal,16,17 and
subareolar18,19 routes of injection are
associated with greater success and a comparable false negative rate to that
associated with the peritumoural route. If indeed the same SLN is
“sentinel” to the entire breast, then this SLN can be identified in
cases of multicentric cancer by subareolar or intradermal injection.
Several small nonrandomised series in which such a approach
was evaluated have demonstrated that the test performance of SLNB in
multicentric or multifocal disease is similar to that for women with unifocal
disease, suggesting that the technique can be applied in this
setting.20–22 Taranaki has chosen to
include tumours up to 5.0 cm diameter and those with multifocal or multicentric
disease for consideration of SLNB based on current literature. Although this may
have affected the test performance measures, evidence suggests that including
larger or multicentric tumours should not effect these.
Sentinel lymph node biopsy technique in Taranaki may have
also affected both identification and false negative rates. There is now a
substantial amount of evidence suggesting that the use of blue dye and
radioactive colloid in combination as opposed to either method on its own,
increases identification rates while minimising false
negatives.6,23 In Taranaki, the majority of
SLNBs were performed using a single method only. Although this does not appear
to have adversely affected the identification rate, it may be a contributing
factor to the high false negative rate.
It remains unclear whether isolated tumour cells or
micrometastases represent an adverse prognostic indicator and whether AC should
be carried out on all cases. Likewise, there is insufficient data to determine
whether the presence of isolated tumour cells or micrometastases should be a
factor in treatment decisions. McCready et al24
suggest that metastasis is found in nonsentinel nodes in approximately 10% of
patients with isolated tumour cells in the SLN and in 20–35% of patients
with micrometastases in the SLN.
Some studies have demonstrated an adverse effect on survival
in patients positive for micrometastatic disease, others have
not.25–27 The definition of
micrometastases and detection methods throughout the literature have varied.
There are no clear New Zealand guidelines on the management of axillary
micrometastases in patients with early stage breast cancer. ASCO guidelines
recommend routine AC in patients with micrometastases >0.2 mm but
≤2.0mm until further studies are
completed.6
Definitive data from randomised trials is needed to decide
if axillary dissection is needed when the SLN is positive. Two large prospective
clinical studies, NSABP-32 and American College of Surgeons trial Z0011, are
hoped to definitively resolve questions regarding the optimal surgical
management of patients positive for micrometastases and isolated tumour
cells.28,29
In Taranaki, all patients are discussed at a
multidisciplinary meeting involving surgical, oncology, radiation therapy, and
radiology specialists and further management decisions are made based on current
evidence in association with individual patient factors.
Another area of ambiguity is the use of SLNB in patients
with a histological diagnosis of DCIS alone. A positive SLN has been reported in
6% to 13% of patients with DCIS.30 Although it
is well established that nodal status for invasive disease is prognostically
important, the clinical relevance of a positive SLN in patients with DCIS
remains undetermined.
New Zealand guidelines for the use of SLNB in DCIS are
inconclusive. ASCO guidelines recommend considering SLNB for patients with DCIS
when a mastectomy is indicated or when immediate reconstruction is planned, as
axillary staging by SNB is essentially impossible if an invasive tumour is
found.6 Although an invasive component will
subsequently be found in 10–20% of cases diagnosed by core biopsy as DCIS
alone, they do not recommend SLNB for patients with DCIS undergoing breast
conserving surgery. However, some experts argue that SLNB in patients with DCIS
undergoing breast conserving surgery will help identify those with unrecognised
invasive disease, and suggest that SLNB in those with high grade, or large areas
of DCIS is warranted so as to avoid a second operation on the axilla if invasive
cancer is found.27
Size >4 cm has been shown to be a predictor for invasive
breast cancer in patients with an initial diagnosis of DCIS
alone.31 In Taranaki, treatment was in
accordance with ASCO guideline recommendations in all but two cases of DCIS who
underwent SLNB. In those two cases, the area of DCIS histologically was 40 mm
and 50 mm respectively, indicating a high risk of unrecognised invasive disease.
Management of both cases was discussed at the multidisciplinary team
meeting.
The performance of SLNB in Taranaki is comparable to
international centres. This is despite a lack of statutory training
requirements. This article highlights the need to adhere to recommended
protocols when introducing a new technique in order to provide non-biased test
performance measures. In the case of SLNB, this involves at least 20 cases
either followed by routine axillary clearance, or with mentor supervision. For
cases of micrometastases and ductal carcinoma in situ, further evidence
is required to determine optimal management.
Adoption of SLNB in early stage breast cancer as routine in
Taranaki has been achieved with results equivalent to internationally
established standards. This may spare many Taranaki women the morbidity of
axillary clearance, without jeopardising safety.
Competing interests: None known.
Author information: Emily
Davenport, Surgical Registrar, Department of
Surgery, Dunedin Hospital, Dunedin; Michael W
Fancourt, General Surgical Consultant,
Department of Surgery, Taranaki Base Hospital, New
Plymouth; William T C Gilkison, General Surgical Consultant, Department of
Surgery, Taranaki Base Hospital, New Plymouth; Steven M Kyle, General Surgical
Consultant, Department of Surgery, Taranaki Base Hospital, New Plymouth; Damien
A Mosquera, Department of Surgery, Taranaki Base Hospital, New Plymouth
Correspondence: Emily Davenport, 7A Clark
St, Dunedin, New Zealand; Fax: +64 (0)3 4740999;
email: emily@med.co.nz
References:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| | Current
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads) Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals |
|