The impact of sentinel node biopsy and axillary staging strategy on hospital costs

R. Rönkä1,*, K. v. Smitten1, H. Sintonen4, T. Kotomäki4, L. Krogerus2, E. Leppänen3 and M. Leidenius1

1 Breast Surgery Unit, 2 Department of Pathology, 3 Department of Clinical Physiology and Nuclear Medicine, Maria Hospital, Helsinki University Hospital, Helsinki; 4 Department of Public Health, Helsinki University, Helsinki, Finland

Received 5 May 2003; revised 21 July 2003; accepted 3 September 2003


    ABSTRACT
 Top
 ABSTRACT
 Introduction
 Patients and methods
 Results
 Discussion
 REFERENCES
 
Background:

The aim of the study was to evaluate total hospital costs of three different sentinel node biopsy (SNB) protocols compared to those of diagnostic axillary lymph node dissection (ALND).

Patients and methods:

The study included 237 consecutive breast cancer patients who underwent SNB with frozen section diagnosis. The sequence of the treatment procedures for each patient was recorded. The sequences of treatment procedures for the same patients were evaluated using three hypothetical scenarios: diagnostic ALND, SNB without frozen section diagnosis and SNB as day case surgery prior to the breast operation. The total hospital costs were calculated in all protocols.

Results:

The hospital costs per patient were 3750{euro}. The hospital costs per patient would have been 3020{euro} when using the ALND model, 4087{euro} had the frozen section not been applied and 4573{euro} using ‘SNB as day case surgery’ model. The costs with or without frozen section diagnosis would have been equal with a threshold false negative rate of 35%.

Conclusions:

SNB seems to be associated with higher hospital costs than diagnostic ALND. Frozen section diagnosis seems to be worthwhile as long as the false negative rate is <35%.

Key words: axillary lymph node dissection, breast cancer, hospital costs, sentinel node biopsy


    Introduction
 Top
 ABSTRACT
 Introduction
 Patients and methods
 Results
 Discussion
 REFERENCES
 
Sentinel node biopsy (SNB) represents a major change in the surgical treatment and staging of breast cancer. The procedure is currently introduced in most breast surgery centres as an alternative to axillary lymph node dissection (ALND) for axillary staging in breast cancer. The method is less invasive, providing a benefit for postoperative recovery and also probably for long-time morbidity [16].

It is obvious that SNB leads to increasing costs due to lymphoscintigraphy, surgical equipment such as gamma detectors and frozen section analysis. All this may be compensated by the shorter operation time and hospital stay [7]

When performing SNB, the indication for axillary clearance is determined according to the results of intraoperative frozen section diagnosis. If the sentinel nodes are uninvolved, the patient avoids unnecessary axillary clearance. However, a false negative frozen section diagnosis leads to axillary clearance as a second operation. The economic impact of false negative findings in the intraoperative diagnosis of sentinel nodes may be significant and jeopardise the whole concept from an economic point of view. In many centres frozen section diagnosis is not applied because of disappointing false negative rates as well as a considerable workload and high costs in the pathology laboratory.

The aim of the study was to compare SNB to diagnostic ALND in terms of hospital costs in Helsinki University Hospital. Another purpose was to evaluate the costs of three different SNB protocols.


    Patients and methods
 Top
 ABSTRACT
 Introduction
 Patients and methods
 Results
 Discussion
 REFERENCES
 
From 4 September 2000 to 21 August 2001, 237 patients with clinical T1-2, node-negative breast cancer underwent lymphatic mapping and sentinel node biopsy at the Breast Surgery Unit of Helsinki University Hospital. All patients were included in a prospective observational study. The Ethical Committee of Helsinki University Hospital approved the project plan and written informed consent was obtained from each patient. The characteristics of the study subjects are presented in Table 1.


View this table:
[in this window]
[in a new window]
 
Table 1. Characteristics of 237 breast cancer patients who underwent sentinel node biopsy with intraoperative frozen section diagnosis of sentinel node metastases
 
Lymphoscintigraphy
Lymphoscintigraphy was performed the day before surgery, 4 h after a single intratumoral injection of 80–100 MBq 99mTc-labelled human albumin microcolloid (ALBU-RES or Nanocoll, Nycomed Amersham Sorin s.r.l. Saluggia, Italy) in a volume of 0.2 ml. Anterior and lateral planar views were obtained using a gamma camera. The estimated locations of sentinel nodes as well as the injection site of the tracer were marked with permanent pen on the skin.

Surgical methods
At least 5 min prior to incision 1 ml Patent Blue dye was injected intratumourally (Bleu Patenté V; Laboratoire Geuerbet, Aulnay-sous-Bois, France). The sentinel nodes were localised using a {gamma} probe (Neoprobe TM 2000; Johnson & Johnson Medical, Hamburg, Germany or NavigatorTM GPS, Auto Suture European Services Centre, SA, France) and by searching blue stained lymphatic vessels and nodes.

Axillary clearance comprising levels I and II was performed when the sentinel node(s) in the axilla were involved, were not identified, were blue only without any radioactivity or if the tumour proved to be multifocal. Axillary clearance was extended to level III only when there were palpable nodes suspicious for metastatic involvement. When metastases were found in the frozen section, axillary clearance was performed during the same operation. Patients with false negative results in frozen section diagnosis of sentinel lymph nodes underwent axillary clearance as a second operation.

Histological methods
The sentinel lymph nodes were sent to the pathology laboratory as separate samples labelled with the site of origin. The fresh specimens were cleaned from all extra-capsular fat tissue, measured, sliced into 1–1.5 mm thick sections and mounted on iced OCT® (Sakura, Finetek, Europe, B.V., The Netherlands). Touch preparations and frozen sections from two levels were made from these slices; these were stained with toluidine blue and viewed. Remaining tissue was fixed in phosphate-buffered 10% formalin, embedded in paraffin and sections were stained with haematoxylin–eosin (H&E) and with Cam 5.2 immunostain. When frozen section diagnosis was not required, the nodes were fixed directly into phosphate-buffered 10% formalin. After fixation the nodes were cleaned from fat, cleaved and embedded wholly. H&E sections were made from two levels of each lymph node and cytokeratin immunostaining was done from one level. The finding was considered a metastasis when Cam 5.2 positive, morphologically identifiable cancer cells were found in the lymph node.

In patients with axillary clearance all non-sentinel lymph nodes were identified and embedded wholly in paraffin. H&E sections were prepared from two levels, approximately 200 µm apart.

Hypothetical alternatives for axillary staging
The axillary staging in 237 patients was performed using SNB with intraoperative frozen section diagnosis of sentinel node metastases. The sequence of the staging and treatment procedures of each of the 237 patients was registered accordingly (Tables 2 and 3). In addition, three hypothetical alternative treatment protocols were developed for each of the 237 patients: diagnostic ALND, SNB as day case surgery prior to the breast operation and SNB without frozen section diagnosis. Thereafter, the hypothetical sequence of the staging and treatment procedures of each patient was estimated, firstly applying diagnostic ALND for axillary staging, secondly if SNB as day case surgery prior to the breast operation had been applied and thirdly applying SNB without frozen section diagnosis. The hospital costs according to the real or hypothetical sequences of surgical staging and treatment procedures using each axillary staging alternative were calculated (Figure 1, Table 4).


View this table:
[in this window]
[in a new window]
 
Table 2. The surgical staging and treatment procedures in 237 breast cancer patients with sentinel node biopsy and intraoperative frozen section diagnosis of sentinel node metastases
 

View this table:
[in this window]
[in a new window]
 
Table 3. Intraoperative and postoperative histological diagnosis of sentinel node metastases in 204 patients with successful identification of sentinel nodes
 


View larger version (38K):
[in this window]
[in a new window]
 
Figure 1. Decision tree showing paths of eventual surgical staging and treatment procedures in four different axillary staging alternatives: diagnostic axillary lymph node dissection; sentinel node biopsy (SNB) with frozen section diagnosis; SNB as day case surgery prior to the breast operation and SNB without frozen section diagnosis. The prefix ‘p’ indicates ‘probability’ of each procedure on a path indicating the proportion of patients who undergo the procedure. The symbol ‘#’ stands for residual probability (1 – the complementary probability). For abbreviations of the procedures, see Table 4. At the end of each ‘branch’ of the tree (terminal node) are shown the costs accumulated on the way to reach the terminal node. The unit costs and the abbreviations of the costs are listed in Table 5.

 

View this table:
[in this window]
[in a new window]
 
Table 4. Baseline probabilities (in alphabetical order) used in decision tree analysis (Figure 1)
 
Hospital costs
For calculation of costs we used unit costs based on the detailed internal accounting system of Helsinki University Hospital. They are estimated according to the production costs of the various services and procedures in 2001. The costing included hospital inpatient care, outpatient visits, lymphoscintigraphy, surgery and pathological analysis. The cost of hospital inpatient care comprises room and board, medication, blood products, laboratory costs, as well as pre- and postoperative nursing care. The cost of outpatient visits comprises pre- and postoperative check-ups. The cost of pathology analysis includes intraoperative frozen section diagnosis of sentinel nodes; postoperative histological examinations of sentinel nodes, breast specimens as well as the axillary node dissection specimens in patients with axillary clearance. The surgery costs comprise the costs arising from the operation including anaesthesia and recovery room (Table 4).

The patients were admitted to the hospital on the day of surgery. On average, the patients were discharged on the first day after breast surgery and SNB and on the second day after axillary clearance. These patients left with drainage, which was removed as an outpatient procedure on the fifth day after surgery by a breast nurse.

Statistical methods
The calculations were carried out using a computer-based decision tree model built on DATA 3.5 (TreeAge Software, Inc., Williamstown, MA, USA) (Figure 1, Tables 4 and 5).


View this table:
[in this window]
[in a new window]
 
Table 5. Unit costs used in decision tree analysis
 

    Results
 Top
 ABSTRACT
 Introduction
 Patients and methods
 Results
 Discussion
 REFERENCES
 
The hospital costs per patient with intraoperative frozen section diagnosis were 3750{euro}. The hospital costs per patient would have been 3020{euro} when using the ALND model, 4087{euro} had the frozen section not been applied and 4573{euro} using ‘SNB as day case surgery’ model. Thus the difference in costs between the most and the least costly axillary staging method was 1553{euro} per patient (Table 6).


View this table:
[in this window]
[in a new window]
 
Table 6. Total hospital costs of surgical staging and treatment of breast cancer using four different axillary staging alternatives
 
The false negative rate in the intraoperative frozen section diagnosis was 13% (9/68) of patients with tumour positive sentinel nodes. According to the sensitivity analysis, the economical advantage of intraoperative diagnosis would have been lost with a false negative rate of ≥35%. In other words, the costs with or without frozen section diagnosis would be equal with a threshold false negative rate of 35%.

Two operations were necessary for the treatment and axillary staging of breast cancer in 19 (8%) of these 237 patients. More than one operation would have been necessary for each patient when applying the ‘SNB as day case surgery’ model (Table 7).


View this table:
[in this window]
[in a new window]
 
Table 7. The number of operations in 237 breast cancer patients when applying four different axillary staging alternatives
 

    Discussion
 Top
 ABSTRACT
 Introduction
 Patients and methods
 Results
 Discussion
 REFERENCES
 
It is not surprising that ALND turned out as the least costly alternative for axillary staging in the present study. New methods such as SNB tend to be associated with elevated costs due to the specialised equipment and examinations which are necessary for a successful performance of the method.

On the other hand, Gemignani et al. found SNB is less costly compared to ALND because of the shorter hospital stay after SNB. However, only patients with a low probability of axillary metastases, those with T1 tumours, were included in their study [7]. Their lower, false-negative rate (4–9%) in frozen section diagnosis gave lower costs of SNB in relation with diagnostic ALND than in our study. The vast majority (73%) of their SNB patients with tumour negative findings in frozen section were discharged the same day and in general the post-operative hospital stay was shorter in their study [7]. We still have to discharge our metastasis-free SNB patients the next day because of postoperative nursing care, counselling and physiotherapy.

The most tempting alternative for SNB, at least from a theoretical point of view, would be SNB as an initial outpatient procedure under local anaesthesia. The histopathological report of sentinel nodes would be final and the patient would come to the subsequent breast operation with axillary staging already performed. However, according to our calculations, it would be the most expensive alternative. Moreover, according to our clinical experience, axillary clearance as a second operation seems to be more demanding and laborious due to scar tissue caused by SNB and sparing the intercostobrachial nerves may be difficult.

From the economical point of view, the frozen section analysis sentinel node metastases are probably the most crucial part of the whole procedure. Frozen section diagnostics is expensive due to substantial workload in the pathology laboratory but it enables axillary clearance in most axillary node-positive patients during the primary operation. A false negative result leads to re-operation and elevated costs of the surgical treatment. Recovery, including sick leave will be longer and risk for post-operative morbidity may be higher. The sensitivity analysis revealed that frozen section diagnosis seems to be economically worthwhile as long as the false negative rate is not >35%, which is far higher than the false negative rate observed in our clinic.

Because the lower costs of a shorter hospital stay may not fully compensate the costs of lymphoscintigraphy and the histological evaluation of sentinel nodes, the possible economic justification of the method has to be sought in the patient history after the operation. SNB is associated with less early post-operative morbidity and thus with a faster recovery compared to axillary clearance [2, 4]. Therefore we tried to evaluate the differences in sick leave, but the length of sick leave appeared to be more dependent on psychosocial factors than simply on physical recovery from surgery. Post-operative sick leave often continued until the end of adjuvant chemo- and/or radiotherapy because of factors such as adverse effects of the treatment, the patient’s psychological status and the distance between residence and the Department of Oncology.

For these reasons, the most significant advantage of SNB will probably be a decrease in long-term post-operative morbidity. As regard to chronic post-mastectomy pain and lymphoedema, the preliminary results are encouraging [13, 5, 6]. However, larger studies with longer follow-up periods are needed until the real benefit of SNB can be concluded. Patients have to be followed up for several years with registration of detailed data regarding need for physiotherapy due to arm morbidity, periods of sick leave, and change in work place and/or in work responsibilities. However, these work-related events are irrelevant for a substantial part of breast cancer patients because they are already retired due to old age or disability caused by other conditions than breast cancer and its treatment. Anyway, physiotherapy and especially the treatment of severe chronic lymphoedema are expensive and concern all breast cancer patients irrespective of age and working status. A recent report has shown a five-fold lower risk for lymphoedema and a 2.9-fold lower risk for impaired use of the arm after SNB without further axillary treatment compared with those after axillary clearance [8]. In the future this has to have a cost-saving effect in favour of SNB.

Comparisons of the costs of different procedures is demanding and the results may not be generalised; for instance, due to differences in treatment practices and unit costs of resources and services [7, 9]. Therefore, the applicability of our results for other units depends on factors such as the costs of lymphoscintigraphy and inpatient care, as well as the salaries of the pathologists, surgeons and nurses. In addition, the length of the hospital stay after SNB or ALND, the success rate in identification of sentinel nodes and the false negative rate in frozen section diagnosis vary, not only between different units, but may also vary over time in the same hospital. For example, in our clinic, the sentinel node identification rate has improved to >95% from 86% reported in the present study, while the sensitivity of frozen section diagnosis has dropped only a little and is currently 93%. It should also to be noted that our purpose was just to compare the hospital costs over the short term, and not to carry out a cost-effectiveness comparison with explicit measurement of health outcome and associated costs over the long term.

The reliability of our results may be disputed because our patients were not randomised to the different axillary staging alternatives (ALND, SNB as day case surgery, SNB with and without intraoperative diagnosis). The costs of ANB with frozen section reflect actual costs, but the costs in the hypothetical strategies were estimated and may therefore differ, at least in some respects, from the actual ones. In addition, we assumed that all patients are eligible for SNB as day case surgery under local anaesthesia. Because the costs of lymphoscintigraphy as well as those of the operations and histopathological assessment of breast and axillary lymph node specimens are averages based on a large number of cases, only the length of the hospital stay and the possible complications after different axillary staging modalities leave room for speculation. The shortest possible postoperative hospital stay was applied when estimating the costs of the hypothetical strategies, therefore the actual costs might be somewhat higher if several patients need a longer stay in hospital after axillary clearance. Anyway, it is unlikely that the hospital costs of diagnostic ALND exceed those of SNB, at least when all patients eligible for SNB are considered. Nevertheless, SNB may be less expensive than ALND in patients with a very low probability of axillary metastases [7].

In conclusion, with regard to hospital costs, SNB is not the least costly method for axillary staging in breast cancer. The benefits of SNB are supposed to be found in decreased long-term arm morbidity and have to be further evaluated. The intraoperative diagnosis of sentinel node metastases seems worthwhile as long as the false-negative rate is relatively low.


    FOOTNOTES
 
* Correspondence to: Dr Riitta Rönkä, Martinmetsä 14, 02300 Espoo, Finland. Tel: +358-503-087-655; E-mail: riitta.ronka{at}pp.fimnet.fi Back


    REFERENCES
 Top
 ABSTRACT
 Introduction
 Patients and methods
 Results
 Discussion
 REFERENCES
 
1. Schrenk P, Rieger R, Shamiyeh A, Wayand W. Morbidity following sentinel lymph node biopsy versus axillary lymph node dissection for patients with breast carcinoma. Cancer 2000; 88: 608–614.[CrossRef][ISI][Medline]

2. Burak WE, Hollenbeck ST, Zervos EE et al. Sentinel lymph node biopsy results in less postoperative morbidity compared with axillary lymph node dissection for breast cancer. Am J Surg 2002; 183: 23–27.[CrossRef][ISI][Medline]

3. Sener SF, Winchester DJ, Martz CH et al. Lymphedema after sentinel lymphadenectomy for breast carcinoma. Cancer 2001; 92: 748–752.[CrossRef][ISI][Medline]

4. Leidenius M, Leppänen E, Krogerus L, vSmitten K. Motion restriction and axillary web syndrome after sentinel node biopsy and axillary clearance in breast cancer. Am J Surg 2003; 185: 127–130.[CrossRef][ISI][Medline]

5. Haid A, Köberle-Wührer R, Knauer M et al. Morbidity of breast cancer patients following complete axillary dissection or sentinel node biopsy only: a comparative evaluation. Breast Cancer Res Treat 2002; 73: 31–36.[CrossRef][ISI][Medline]

6. Miguel R, Kuhn AM, Shons AR et al. The effect of sentinel node selective axillary lymphadenectomy on the incidence of postmastectomy pain syndrome. Cancer Control 2001; 8: 427–430.[Medline]

7. Gemignani ML, Cody HS, Fey JV et al. Impact of sentinel lymph node mapping on relative charges in patients with early-stage breast cancer. Ann Surg Oncol 2000; 7: 575–580.[Abstract/Free Full Text]

8. Schijven MP, Vingerhoets AJJM, Rutten HJT et al. Comparison of morbidity between axillary lymph node dissection and sentinel node biopsy. Eur J Surg Oncol 2003; 29: 341–350.[CrossRef][ISI][Medline]

9. Palit TK, Miltenburg DM, Brunicardi FC. Cost analysis of breast conservation surgery compared with modified radical mastectomy with and without reconstruction. Am J Surg 2000; 179: 441–445.[CrossRef][ISI][Medline]





This Article
Abstract
Full Text (PDF)
E-letters: Submit a response
Alert me when this article is cited
Alert me when E-letters are posted
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (5)
Disclaimer
Request Permissions
Google Scholar
Articles by Rönkä, R.
Articles by Leidenius, M.
PubMed
PubMed Citation
Articles by Rönkä, R.
Articles by Leidenius, M.