- Email: [email protected]
Increased mortality from breast cancer and inadequate axillary treatment
The Breast (2003) 12, 36–41 0960-9776/03/$-see front matter r 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0960-9776(02)00213-8
ORIGINAL ARTICLE
Increased mortality from breast cancer and inadequate axillary treatment D. B. Kingsmore, A. Ssemwogerere, D. J. Hole, C. R. Gillis and W. D. George The West of Scotland Cancer Surveillance Unit, The Department of Public Health, The University of Glasgow, Lilybank Gardens, Glasgow, G12 8RZ, Scotland, UK S U M M A R Y . Background. The assessment of axillary nodal status remains divisive: inaccurate staging may result in untreated axillary disease, and appropriate adjuvant therapy not being delivered. The impact of inadequate axillary treatment on survival remains controversial. We analyse the impact of failure to adequately assess the axillary nodal status on survival. Methods. All women with confirmed breast cancer in a 15-year period were identified, and the original pathology reports examined, and details of radiotherapy obtained. The survival of women by axillary sample size was compared to a reference group of women and corrected for nodal status, tumour size, age, deprivation category and speciality of treating surgeon. Findings. Sampling less than four nodes is associated with a significantly increased risk of death. This cannot be due to understaging the extent of axillary disease nor is fully explainable by differential prescription of adjuvant therapies. We conclude that the survival of the women studied may have been adversely effected by inadequate axillary treatment. r 2003 Elsevier Science Ltd. All rights reserved.
recurrent axillary disease on survival is however uncertain.7,8 There is still discussion on the most effective, reliable and minimally invasive means of achieving accurate axillary staging. Many consider axillary sampling to be as effective as axillary clearance in establishing axillary nodal status whilst lessening the associated morbidity.8–15 It is not clear how many axillary lymph nodes must be sampled to accurately reflect not only the presence, but also the extent of disease in the axilla. Some specialist units maintain that sampling three to five nodes is sufficient to identify node-positive women.16–19 Others maintain that ten nodes are required to minimise the risk of understaging the extent of axillary disease.20–22 This report clarifies the number of axillary lymph nodes that are required to be examined to accurately predict the survival of node-negative and node-positive women.
INTRODUCTION Axillary nodal status is the most important prognostic indicator in breast cancer. Prognosis can be independently predicted by both the number,1,2 and also the sequential level of axillary involvement.3–5 The nodal status influences both systemic and local treatment. Bonadonna6 first suggested systemic adjuvant chemotherapy targeted to node-positive women in 1975. Treatment of involved axillary lymph nodes is necessary to minimise axillary recurrence. Other prognostic variables do not identify women who require definitive axillary treatment and thus axillary staging is recommended in all women. The impact of untreated or
Address correspondence to: Prof. David J. Hole, M.Sc., The West of Scotland Cancer Surveillance Unit, The Department of Public Health, The University of Glasgow, Lilybank Gardens, Glasgow G12 8RZ, Scotland, UK. Tel.: +44 141 339 8855 ext. 0922; E-mail: [email protected] Received: 5 May 2002 Revised: 1 October 2002 Accepted: 9 October 2002
PATIENTS The regional cancer registry was used to identify all histologically verified cases of invasive breast cancer, 36
Increased mortality from breast cancer diagnosed between 1/1/80 and 31/12/94 in women aged under 75 who lived in a geographically defined area in the West of Scotland. Original pathology reports were obtained from local laboratories and details of pathological size, number and status of lymph nodes obtained. Information on grade was incompletely recorded and not included as a variable. This produced a population of 7144 cases of primary breast cancer of whom 4627 (65%) had axillary surgery performed and number of nodes stated, 2153 (30%) had no axillary surgery and 364 (5%) had axillary surgery but no specific number of axillary nodes stated. The primary cause of death was determined from a printout of the death certificates, obtained from the Registrar General for Scotland’s death list. Breast-cancer-specific mortality was defined as the excess deaths among the group of women diagnosed with breast cancer when compared to the mortality of an equivalent age-matched cohort obtained from the Registrar Generals’ list. Deaths from causes other than breast cancer were treated as intercurrent deaths in the analysis. Follow-up was complete to 12/98 (mean 9.5 years, range 3–18). Details of referral to radiotherapy were derived from the regional centre for oncology. Socio-economic status was derived from post-code sectors using the Carstairs and Morris Index.23
METHODS The number of nodes removed and reported was examined by tumour size to ensure that there was no bias in the extent of axillary surgery. A reference group of women was defined whose survival accurately represented true disease stage and in whom there was no understaging. We selected this reference group on the basis of previous reports, which indicated that removing less than ten lymph nodes underestimated the true incidence of nodal positivity.19–22 Therefore, women in whom ten or more lymph nodes had been removed were taken to represent a reference group of accurately staged axillae. We verified the prognostic accuracy of this reference group by examining nodal status and tumour size; firstly, confirming that they were directly related and secondly, that these factors influenced survival. The survival of women with fewer nodes removed could therefore be compared against this reference group. The minimal nodal sample size for prognostic accuracy was defined for node-negative women by comparing the outcome of the reference group to women with sample sizes less than ten lymph nodes.
37
The minimal nodal sample size for prognostic accuracy for node-positive women was similarly examined. Thus, with this calculated minimum axillary sample size for both node-positive and node-negative women, a comparison was made of the survival of inadequate vs adequate axillary sampling. Cox’s proportional hazards model24 and life table analysis was used to determine survival after adjustment for tumour size, age, speciality of surgeon, and deprivation category.
RESULTS There was no correlation between the number of nodes sampled and tumour size. Overall, 26–31% of all tumour sizes had 10 or more nodes sampled and 19– 22% had less than four nodes sampled. The percentage of women with any nodal sample size was similar in all categories of tumour size. In the reference group of women with ten or more nodes sampled, tumour size directly correlated to nodal involvement, ranging from 25% for tumours less than 10 mm to 76% for tumours greater than 40 mm (Fig. 1). In addition, the relative risk of death increased linearly with increasing tumour size and nodal status (Table 1). Women reported as node negative had a significantly poorer survival if less than four nodes were sampled when compared to the reference group (Table 2). The 5- and 10-year survival of women with less than four nodes sampled was 6% poorer than that of women with four or more nodes sampled. Sampling more than four nodes conferred no additional survival benefit. Women with one to three positive nodes had a significantly poorer survival if less than four nodes were sampled when compared to the reference group (Table 3). The 5- and 10-year survival of women with less than four nodes sampled was 19% and 21% poorer than those women with four or more nodes sampled. There was no significant difference in outcome if four or more lymph nodes were sampled when compared to the reference group. The outcome of inadequate axillary sampling (less than four nodes) was compared to the outcome of women with adequate axillary sampling (four or more nodes sampled) (Table 4, Fig. 2). Women with zero, one, two, three positive nodes had 5-year survival rates that were 6%, 15%, 27% and 29% poorer with inadequate sampling when compared to women with adequate axillary sampling. The 10-year survival rates were 6%, 27%, 24% and 8% poorer in the same groups. When all
38
The Breast 90
80
70
Percentage
60 Node Negative 50
Node Positive
40
30
20
10 0-0.9
1-1.9
2-2.9
3-3.9
4-4.9
5+
Tumour Size (cm) Fig. 1 The distribution of tumour size and nodal status in reference group (based on 1292 women with ten or more lymph nodes sampled and tumour size stated).
Table 1
The survival of the reference group by number of positive nodes and tumour size (based on women with 10+ nodes sampled)
Number of positive nodes
Node 1 2 3 4,5 6–10 11+
ve
Number of patients
609 168 102 65 72 160 137
% Surviving 5 years
10 years
85 79 59 72 49 46 33
73 64 34 32 19 30 15
n
w
(95% c.i.)
1 1.42 2.33 2.04 3.81 3.89 5.10
(Baseline) (0.94–2.15) (1.53–3.56) (1.22–3.40) (2.47–5.89) (2.82–5.38) (3.64–7.13)
RHRw
Tumour size (mm) o20 20–39 40+ No information
RHRn
534 528 230 21
83 67 49 50
73 47 30 29
1 2.64 4.62
(Baseline) (1.99–3.50) (3.40–6.26)
Relative hazard ratio (RHR) adjusted for age, deprivation category and tumour size. Relative hazard ratio (RHR) adjusted for age, deprivation category, and nodal status.
node-positive women were considered together, inadequate axillary sampling when compared to adequate sampling was associated with a survival difference of 6% and 7% at 5 and 10 years. The percentage of node-positive women with less than four nodes removed that were referred for radiotherapy ranged from 14% in 1980 to 100% in 1994. Overall, 42% (174/417) were referred.
DISCUSSION The major finding of this study is that sampling less than four axillary nodes is associated with a significantly poorer survival irrespective of the number of involved axillary nodes. Whilst some of the survival differences reported may reflect understaging, the analysis of outcome data has been corrected for tumour size that
Increased mortality from breast cancer
39
Table 2 The survival of node negative patients by the number of nodes sampled compared to the reference group (based on 2239 node-negative patients) Number of node sampled
Number of patients
% Surviving
RHR* (95% c.i.)
5 years
10 years
1 2 3 4–5 6–7 8–9 10+
211 189 95 408 430 362 609
83 82 79 87 83 89 85
72 66 70 76 71 78 73
1.31 (1.09–1.57)** 1.25 (0.86–1.80) 1.36 (0.95–1.96) 0.96 (0.70–1.32) 1.15 (0.84–1.56) 0.80 (0.56–1.16) 1 (Baseline)
1–3 4+
563 1367
81 87
70 76
1.31** (1.07–1.60) 1 (Baseline)
n
RHR adjusted for age, deprivation category and tumour size. Po0.01.
nn
Table 3 The survival of node-positive patients by number of nodes sampled compared to the reference group (based on 1473 patients with one to three nodes positive) Number of nodes sampled
Number of patients
RHRn (95% c.i.)
% Surviving 5 years
10 years
1 2 3 4–5 6–7 8–9 10+
144 122 152 259 246 215 335
57 53 49 72 67 76 72
36 35 34 56 53 64 50
1.66n n (1.21–2.27) 1.75n n (1.27–2.42) 1.86n n (1.38–2.56) 0.91 (0.67–1.22) 1.12 (0.83–1.52) 0.79 (0.56–1.10) 1 (Baseline)
1–3 4+
418 1055
53 72
35 56
1.85n n (1.54–2.21) 1 (Baseline)
n
RHR adjusted for age, deprivation category and tumour size. Po0.001.
nn
correlates strongly with true nodal status. Thus, understaging cannot adequately explain the survival differences seen. This poorer survival may be explained in two ways: either indirectly (understaging the extent of axillary nodal status may have influenced decisions regarding adjuvant therapy) or directly (untreated axillary disease itself may have led to poorer survival). The direct relationship of tumour size to nodal status was confirmed in the reference group. Of women with tumours smaller than 1 cm, 25% had involved axillary nodes. Thus, tumour size alone was not a sufficiently discriminating factor on which to base axillary surgery and systemic treatment in the population studied. All women therefore should have axillary staging performed. Several methods of pathological assessment of axillary nodal status have been described including axillary clearance, axillary sampling25 and sentinal node biopsy.26 In this cohort, axillary sampling has been performed poorly with 26% of women in whom axillary staging was attempted having less than four nodes pathologically identified. However, it has been pre-
viously emphasised that the responsibility of successful identification of four lymph nodes resides with the treating surgeon. Axillary sampling has been criticised due to the perceived weakness of definitive treatment after an inadequate or node-positive sample. However, although the method of axillary staging is debated, where axillary staging is performed there is neither argument nor logic to accept inadequate axillary staging. It was not possible to collect information on adjuvant therapies. However the large survival differences found in this study cannot be explained solely by differential use of adjuvant therapies that far exceed the reported absolute mortality benefits found by the Oxford overview from tamoxifen or chemotherapy at 5 years.27 Therefore, inadequate axillary treatment with surgery or radiotherapy may explain, at least in part, the outcome differences observed and appropriate axillary operations are still mandatory. These results are in keeping with trials of post-mastectomy radiotherapy in high risk women which showed a poorer survival subsequent to increased chest wall and axillary recur-
40
The Breast
Table 4 The survival of women by the number of positive nodes and the number of nodes sampled when compared to node-negative women with 4+ nodes sampled (based on 4637 patients) Number of positive nodes
4+ nodes sampled % surviving
1–3 nodes sampled % surviving
5 years
10 years
RHR* (95% c.i.)
5 years
10 years
RHR* (95% c.i.)
0
87
76
1 (Baseline)
81
70
1.40 (1.14–1.72)
1
77
66
1.43 (1.15–1.79)
62
39
2.90 (2.33–3.61)
2
71
54
1.82 (1.43–2.32)
44
30
4.25 (3.26–5.54)
3
65
38
2.55 (1.99–3.27)
36
30
5.12 (3.55–7.39)
4–11
44
25
4.34 (3.67–5.14)
12+
33
16
5.73 (4.28–7.68)
Node-negative women only
87
76
1 (Baseline)
81
70
1.34*** (1.09–1.65)
Node-positive women only
59
42
1 (Baseline)
53
35
1.20** (1.02–1.41)
n
RHR adjusted for age, deprivation category, tumour size. Po0.05; n n nPo0.01.
nn
7 6
R.H.R
5 4 3 4+ Nodes Sampled 1 - 3 Nodes Sampled
2 1 0 0
1
2
3
4 to 9
10 +
Number of Nodes Positive Fig. 2 The risk of death (RHR) by the number of positive nodes and number of nodes sampled.
rences.28,29 In these women, half of all local recurrences were in the axilla. We conclude that sampling less than four lymph nodes is associated with a greater risk of dying from breast cancer, which is not fully explained by under-
staging. This is important not only when analysing the outcome of variables and interventions on a population basis where equivalence of stage is required, but particularly in the provision and audit of breast careservices.
Increased mortality from breast cancer References 1. Carter C L, Allen C, Henson D. Relation of tumour size, lymph node status and survival in 24,740 breast cancer cases. Cancer 1989; 63: 181–187. 2. Fisher B, Bauer M, Wickerham D L, Redmund C K, Fisher E R. Relation of number of positive lymph nodes to the prognosis of patients with primary breast cancer: an NSABP update. Cancer 1983; 52:1551–1557. 3. Veronesi U, Rilke F, Luini A, Sacchini V et al. Distribution of axillary node metastases by level of invasion. An analysis of 539 cases. Cancer 1987; 59(4): 682–687. 4. Veronesi U, Luini A, Galimberti V, Marchini S, Sacchini V, Rilke F. Extent of metastatic axillary involvement in 1446 cases of breast cancer. Eur J Surg Oncol 1990; 16(2): 127–133. 5. Zurrida S, Morabito A, Galimberti V, Luini A, Greco M et al. Importance of the level of axillary involvement in relation to traditional variables in the prognosis of breast cancer. Int J Oncol 1999; 15(3): 475–480. 6. Bonadonna G, Brusamolino E, Valagussa P, Veronessi U. Adjuvant study with combination chemotherapy in operable breast cancer. Proc Am Assoc Cancer Res Am Soc Clin Oncol 1975; 16: 1135a. 7. Recht A, Houlihan M J. Axillary lymph nodes and breast cancer: a review. Cancer 1995; 76: 1491–1512. 8. Hayworth J, Caleffi M. The significance of local control in the primary treatment of breast cancer. Arch Surg 1987; 122: 1244–1247. 9. O’Dwyer P J. Axillary dissection in primary breast cancer. BMJ 1991; 302: 360–361. 10. Fentiman I S, Mansel R E. The axilla: not a no-go zone. Lancet 1991; 337: 221–223. 11. Sacks N D M, Barr L C, Allan S M, Baum M. The role of axillary dissection in operable breast cancer. Breast 1992; 1: 41–49. 12. Sacks N P M, Baum M. Primary management of carcinoma of the breast. Lancet 1993; 342: 1402–1408. 13. Chan H Y, Behen S, Greenall M J. Adequacy of axillary node sampling for evaluation of the axilla and control of axillary disease in breast cancer. Eur J Surg Oncol 1993; 19: 213–214. 14. Davidson T. Why I favour axillary node clearance in the management of breast cancer. Eur J Surg Oncol 1995; 21: 5–7. 15. Greenall M J. Why I favour axillary node sampling in the management of breast cancer. Europ J Surg Oncol 1995; 21: 2–5. 16. Kjaergaard J, Blichert-Toft M, Amdersen J A, Rank F et al. Probability of false negative nodal staging in conjunction with
17.
18.
19.
20. 21. 22.
23. 24. 25. 26. 27.
28. 29.
41
partial axillary dissection in breast cancer. Br J Surg 1985; 72: 365–367. Steele R J C, Forrest A P M, Gibson T, Stewart H J, Chetty U. The efficacy of lower axillary sampling in obtaining lymph node status in breast cancer. A controlled randomised trial. Br J Surg 1985; 72: 368–369. Blichert-Toft M, Rose C, Andersen J A et al. Danish randomized trial comparing breast conservation therapy with mastectomy: six years of life-table analysis. Danish Breast Cancer Cooperative Group. Monogr Natl Cancer Inst 1992; 11: 19–25. Graverson H P, Blichert-Toft M, Anderson J A, Zedeler K. Breast cancer: risk of axillary recurrence in node-negative patients following partial dissection of the axilla. Eur J Surg Oncol 1988; 14:407–412. Mathieson O, Carl J, Bonderup O, Panduro J. Axillary sampling and the risk of eroneous staging of breast cancer: an analysis of 960 consecutive patients. Acta Oncol 1990; 29: 721–725. Axelsson C K, Mouridsen H T, Zedeler K. Axillary dissection of level I aand II nodes is important in breast cancer classification. Eur J Cancer 1992; 28A: 1415–1418. Kiricuta C I, Tausch J. A mathematical model of axillary lymph node involvement based on 1446 complete axillary dissections in patients with breast carcinoma. Cancer 1992; 69: 2496–2501. Carstairs V, Morris R. Deprivation and Health in Scotland. Aberdeen: Aberdeen University Press, 1988. Cox DR. Regression models and life-tables. J R Stat Soc 1972; 34: 187–220. Steele R, Forrest A, Gibson M et al. The efficacy of lower axillary sampling in obtaining lymph node status in breast cancer: a controlled randomized trial. Br J Surg 1985; 72: 368–369. Sakorafas G, Tsitou A. Sentinel lymph node biopsy in breast cancer. A review. Am Surg 2000; 66(7): 667–674. Early Breast Cancer Trials Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic or immune therapy; 133 randomised trials involving 31 000 recurrences and 24 000 deaths among 75 000 women. Lancet 1992; 339: 1–15, 71–85. Ragaz J, Jackson S M, Nhu Le et al. Adjuvant radiotherapy and chemotherapy in node-positive premenopausal women with breast cancer. N Engl J Med 1997; 337: 956–962. Overgaard M, Hansen P S, Overgaard J et al. Postoperative radiotherapy in high-risk premenopausal women with breast cancer who receive adjuvant chemotherapy. N Engl J Med 1997; 337: 949–955.
ORIGINAL ARTICLE
Increased mortality from breast cancer and inadequate axillary treatment D. B. Kingsmore, A. Ssemwogerere, D. J. Hole, C. R. Gillis and W. D. George The West of Scotland Cancer Surveillance Unit, The Department of Public Health, The University of Glasgow, Lilybank Gardens, Glasgow, G12 8RZ, Scotland, UK S U M M A R Y . Background. The assessment of axillary nodal status remains divisive: inaccurate staging may result in untreated axillary disease, and appropriate adjuvant therapy not being delivered. The impact of inadequate axillary treatment on survival remains controversial. We analyse the impact of failure to adequately assess the axillary nodal status on survival. Methods. All women with confirmed breast cancer in a 15-year period were identified, and the original pathology reports examined, and details of radiotherapy obtained. The survival of women by axillary sample size was compared to a reference group of women and corrected for nodal status, tumour size, age, deprivation category and speciality of treating surgeon. Findings. Sampling less than four nodes is associated with a significantly increased risk of death. This cannot be due to understaging the extent of axillary disease nor is fully explainable by differential prescription of adjuvant therapies. We conclude that the survival of the women studied may have been adversely effected by inadequate axillary treatment. r 2003 Elsevier Science Ltd. All rights reserved.
recurrent axillary disease on survival is however uncertain.7,8 There is still discussion on the most effective, reliable and minimally invasive means of achieving accurate axillary staging. Many consider axillary sampling to be as effective as axillary clearance in establishing axillary nodal status whilst lessening the associated morbidity.8–15 It is not clear how many axillary lymph nodes must be sampled to accurately reflect not only the presence, but also the extent of disease in the axilla. Some specialist units maintain that sampling three to five nodes is sufficient to identify node-positive women.16–19 Others maintain that ten nodes are required to minimise the risk of understaging the extent of axillary disease.20–22 This report clarifies the number of axillary lymph nodes that are required to be examined to accurately predict the survival of node-negative and node-positive women.
INTRODUCTION Axillary nodal status is the most important prognostic indicator in breast cancer. Prognosis can be independently predicted by both the number,1,2 and also the sequential level of axillary involvement.3–5 The nodal status influences both systemic and local treatment. Bonadonna6 first suggested systemic adjuvant chemotherapy targeted to node-positive women in 1975. Treatment of involved axillary lymph nodes is necessary to minimise axillary recurrence. Other prognostic variables do not identify women who require definitive axillary treatment and thus axillary staging is recommended in all women. The impact of untreated or
Address correspondence to: Prof. David J. Hole, M.Sc., The West of Scotland Cancer Surveillance Unit, The Department of Public Health, The University of Glasgow, Lilybank Gardens, Glasgow G12 8RZ, Scotland, UK. Tel.: +44 141 339 8855 ext. 0922; E-mail: [email protected] Received: 5 May 2002 Revised: 1 October 2002 Accepted: 9 October 2002
PATIENTS The regional cancer registry was used to identify all histologically verified cases of invasive breast cancer, 36
Increased mortality from breast cancer diagnosed between 1/1/80 and 31/12/94 in women aged under 75 who lived in a geographically defined area in the West of Scotland. Original pathology reports were obtained from local laboratories and details of pathological size, number and status of lymph nodes obtained. Information on grade was incompletely recorded and not included as a variable. This produced a population of 7144 cases of primary breast cancer of whom 4627 (65%) had axillary surgery performed and number of nodes stated, 2153 (30%) had no axillary surgery and 364 (5%) had axillary surgery but no specific number of axillary nodes stated. The primary cause of death was determined from a printout of the death certificates, obtained from the Registrar General for Scotland’s death list. Breast-cancer-specific mortality was defined as the excess deaths among the group of women diagnosed with breast cancer when compared to the mortality of an equivalent age-matched cohort obtained from the Registrar Generals’ list. Deaths from causes other than breast cancer were treated as intercurrent deaths in the analysis. Follow-up was complete to 12/98 (mean 9.5 years, range 3–18). Details of referral to radiotherapy were derived from the regional centre for oncology. Socio-economic status was derived from post-code sectors using the Carstairs and Morris Index.23
METHODS The number of nodes removed and reported was examined by tumour size to ensure that there was no bias in the extent of axillary surgery. A reference group of women was defined whose survival accurately represented true disease stage and in whom there was no understaging. We selected this reference group on the basis of previous reports, which indicated that removing less than ten lymph nodes underestimated the true incidence of nodal positivity.19–22 Therefore, women in whom ten or more lymph nodes had been removed were taken to represent a reference group of accurately staged axillae. We verified the prognostic accuracy of this reference group by examining nodal status and tumour size; firstly, confirming that they were directly related and secondly, that these factors influenced survival. The survival of women with fewer nodes removed could therefore be compared against this reference group. The minimal nodal sample size for prognostic accuracy was defined for node-negative women by comparing the outcome of the reference group to women with sample sizes less than ten lymph nodes.
37
The minimal nodal sample size for prognostic accuracy for node-positive women was similarly examined. Thus, with this calculated minimum axillary sample size for both node-positive and node-negative women, a comparison was made of the survival of inadequate vs adequate axillary sampling. Cox’s proportional hazards model24 and life table analysis was used to determine survival after adjustment for tumour size, age, speciality of surgeon, and deprivation category.
RESULTS There was no correlation between the number of nodes sampled and tumour size. Overall, 26–31% of all tumour sizes had 10 or more nodes sampled and 19– 22% had less than four nodes sampled. The percentage of women with any nodal sample size was similar in all categories of tumour size. In the reference group of women with ten or more nodes sampled, tumour size directly correlated to nodal involvement, ranging from 25% for tumours less than 10 mm to 76% for tumours greater than 40 mm (Fig. 1). In addition, the relative risk of death increased linearly with increasing tumour size and nodal status (Table 1). Women reported as node negative had a significantly poorer survival if less than four nodes were sampled when compared to the reference group (Table 2). The 5- and 10-year survival of women with less than four nodes sampled was 6% poorer than that of women with four or more nodes sampled. Sampling more than four nodes conferred no additional survival benefit. Women with one to three positive nodes had a significantly poorer survival if less than four nodes were sampled when compared to the reference group (Table 3). The 5- and 10-year survival of women with less than four nodes sampled was 19% and 21% poorer than those women with four or more nodes sampled. There was no significant difference in outcome if four or more lymph nodes were sampled when compared to the reference group. The outcome of inadequate axillary sampling (less than four nodes) was compared to the outcome of women with adequate axillary sampling (four or more nodes sampled) (Table 4, Fig. 2). Women with zero, one, two, three positive nodes had 5-year survival rates that were 6%, 15%, 27% and 29% poorer with inadequate sampling when compared to women with adequate axillary sampling. The 10-year survival rates were 6%, 27%, 24% and 8% poorer in the same groups. When all
38
The Breast 90
80
70
Percentage
60 Node Negative 50
Node Positive
40
30
20
10 0-0.9
1-1.9
2-2.9
3-3.9
4-4.9
5+
Tumour Size (cm) Fig. 1 The distribution of tumour size and nodal status in reference group (based on 1292 women with ten or more lymph nodes sampled and tumour size stated).
Table 1
The survival of the reference group by number of positive nodes and tumour size (based on women with 10+ nodes sampled)
Number of positive nodes
Node 1 2 3 4,5 6–10 11+
ve
Number of patients
609 168 102 65 72 160 137
% Surviving 5 years
10 years
85 79 59 72 49 46 33
73 64 34 32 19 30 15
n
w
(95% c.i.)
1 1.42 2.33 2.04 3.81 3.89 5.10
(Baseline) (0.94–2.15) (1.53–3.56) (1.22–3.40) (2.47–5.89) (2.82–5.38) (3.64–7.13)
RHRw
Tumour size (mm) o20 20–39 40+ No information
RHRn
534 528 230 21
83 67 49 50
73 47 30 29
1 2.64 4.62
(Baseline) (1.99–3.50) (3.40–6.26)
Relative hazard ratio (RHR) adjusted for age, deprivation category and tumour size. Relative hazard ratio (RHR) adjusted for age, deprivation category, and nodal status.
node-positive women were considered together, inadequate axillary sampling when compared to adequate sampling was associated with a survival difference of 6% and 7% at 5 and 10 years. The percentage of node-positive women with less than four nodes removed that were referred for radiotherapy ranged from 14% in 1980 to 100% in 1994. Overall, 42% (174/417) were referred.
DISCUSSION The major finding of this study is that sampling less than four axillary nodes is associated with a significantly poorer survival irrespective of the number of involved axillary nodes. Whilst some of the survival differences reported may reflect understaging, the analysis of outcome data has been corrected for tumour size that
Increased mortality from breast cancer
39
Table 2 The survival of node negative patients by the number of nodes sampled compared to the reference group (based on 2239 node-negative patients) Number of node sampled
Number of patients
% Surviving
RHR* (95% c.i.)
5 years
10 years
1 2 3 4–5 6–7 8–9 10+
211 189 95 408 430 362 609
83 82 79 87 83 89 85
72 66 70 76 71 78 73
1.31 (1.09–1.57)** 1.25 (0.86–1.80) 1.36 (0.95–1.96) 0.96 (0.70–1.32) 1.15 (0.84–1.56) 0.80 (0.56–1.16) 1 (Baseline)
1–3 4+
563 1367
81 87
70 76
1.31** (1.07–1.60) 1 (Baseline)
n
RHR adjusted for age, deprivation category and tumour size. Po0.01.
nn
Table 3 The survival of node-positive patients by number of nodes sampled compared to the reference group (based on 1473 patients with one to three nodes positive) Number of nodes sampled
Number of patients
RHRn (95% c.i.)
% Surviving 5 years
10 years
1 2 3 4–5 6–7 8–9 10+
144 122 152 259 246 215 335
57 53 49 72 67 76 72
36 35 34 56 53 64 50
1.66n n (1.21–2.27) 1.75n n (1.27–2.42) 1.86n n (1.38–2.56) 0.91 (0.67–1.22) 1.12 (0.83–1.52) 0.79 (0.56–1.10) 1 (Baseline)
1–3 4+
418 1055
53 72
35 56
1.85n n (1.54–2.21) 1 (Baseline)
n
RHR adjusted for age, deprivation category and tumour size. Po0.001.
nn
correlates strongly with true nodal status. Thus, understaging cannot adequately explain the survival differences seen. This poorer survival may be explained in two ways: either indirectly (understaging the extent of axillary nodal status may have influenced decisions regarding adjuvant therapy) or directly (untreated axillary disease itself may have led to poorer survival). The direct relationship of tumour size to nodal status was confirmed in the reference group. Of women with tumours smaller than 1 cm, 25% had involved axillary nodes. Thus, tumour size alone was not a sufficiently discriminating factor on which to base axillary surgery and systemic treatment in the population studied. All women therefore should have axillary staging performed. Several methods of pathological assessment of axillary nodal status have been described including axillary clearance, axillary sampling25 and sentinal node biopsy.26 In this cohort, axillary sampling has been performed poorly with 26% of women in whom axillary staging was attempted having less than four nodes pathologically identified. However, it has been pre-
viously emphasised that the responsibility of successful identification of four lymph nodes resides with the treating surgeon. Axillary sampling has been criticised due to the perceived weakness of definitive treatment after an inadequate or node-positive sample. However, although the method of axillary staging is debated, where axillary staging is performed there is neither argument nor logic to accept inadequate axillary staging. It was not possible to collect information on adjuvant therapies. However the large survival differences found in this study cannot be explained solely by differential use of adjuvant therapies that far exceed the reported absolute mortality benefits found by the Oxford overview from tamoxifen or chemotherapy at 5 years.27 Therefore, inadequate axillary treatment with surgery or radiotherapy may explain, at least in part, the outcome differences observed and appropriate axillary operations are still mandatory. These results are in keeping with trials of post-mastectomy radiotherapy in high risk women which showed a poorer survival subsequent to increased chest wall and axillary recur-
40
The Breast
Table 4 The survival of women by the number of positive nodes and the number of nodes sampled when compared to node-negative women with 4+ nodes sampled (based on 4637 patients) Number of positive nodes
4+ nodes sampled % surviving
1–3 nodes sampled % surviving
5 years
10 years
RHR* (95% c.i.)
5 years
10 years
RHR* (95% c.i.)
0
87
76
1 (Baseline)
81
70
1.40 (1.14–1.72)
1
77
66
1.43 (1.15–1.79)
62
39
2.90 (2.33–3.61)
2
71
54
1.82 (1.43–2.32)
44
30
4.25 (3.26–5.54)
3
65
38
2.55 (1.99–3.27)
36
30
5.12 (3.55–7.39)
4–11
44
25
4.34 (3.67–5.14)
12+
33
16
5.73 (4.28–7.68)
Node-negative women only
87
76
1 (Baseline)
81
70
1.34*** (1.09–1.65)
Node-positive women only
59
42
1 (Baseline)
53
35
1.20** (1.02–1.41)
n
RHR adjusted for age, deprivation category, tumour size. Po0.05; n n nPo0.01.
nn
7 6
R.H.R
5 4 3 4+ Nodes Sampled 1 - 3 Nodes Sampled
2 1 0 0
1
2
3
4 to 9
10 +
Number of Nodes Positive Fig. 2 The risk of death (RHR) by the number of positive nodes and number of nodes sampled.
rences.28,29 In these women, half of all local recurrences were in the axilla. We conclude that sampling less than four lymph nodes is associated with a greater risk of dying from breast cancer, which is not fully explained by under-
staging. This is important not only when analysing the outcome of variables and interventions on a population basis where equivalence of stage is required, but particularly in the provision and audit of breast careservices.
Increased mortality from breast cancer References 1. Carter C L, Allen C, Henson D. Relation of tumour size, lymph node status and survival in 24,740 breast cancer cases. Cancer 1989; 63: 181–187. 2. Fisher B, Bauer M, Wickerham D L, Redmund C K, Fisher E R. Relation of number of positive lymph nodes to the prognosis of patients with primary breast cancer: an NSABP update. Cancer 1983; 52:1551–1557. 3. Veronesi U, Rilke F, Luini A, Sacchini V et al. Distribution of axillary node metastases by level of invasion. An analysis of 539 cases. Cancer 1987; 59(4): 682–687. 4. Veronesi U, Luini A, Galimberti V, Marchini S, Sacchini V, Rilke F. Extent of metastatic axillary involvement in 1446 cases of breast cancer. Eur J Surg Oncol 1990; 16(2): 127–133. 5. Zurrida S, Morabito A, Galimberti V, Luini A, Greco M et al. Importance of the level of axillary involvement in relation to traditional variables in the prognosis of breast cancer. Int J Oncol 1999; 15(3): 475–480. 6. Bonadonna G, Brusamolino E, Valagussa P, Veronessi U. Adjuvant study with combination chemotherapy in operable breast cancer. Proc Am Assoc Cancer Res Am Soc Clin Oncol 1975; 16: 1135a. 7. Recht A, Houlihan M J. Axillary lymph nodes and breast cancer: a review. Cancer 1995; 76: 1491–1512. 8. Hayworth J, Caleffi M. The significance of local control in the primary treatment of breast cancer. Arch Surg 1987; 122: 1244–1247. 9. O’Dwyer P J. Axillary dissection in primary breast cancer. BMJ 1991; 302: 360–361. 10. Fentiman I S, Mansel R E. The axilla: not a no-go zone. Lancet 1991; 337: 221–223. 11. Sacks N D M, Barr L C, Allan S M, Baum M. The role of axillary dissection in operable breast cancer. Breast 1992; 1: 41–49. 12. Sacks N P M, Baum M. Primary management of carcinoma of the breast. Lancet 1993; 342: 1402–1408. 13. Chan H Y, Behen S, Greenall M J. Adequacy of axillary node sampling for evaluation of the axilla and control of axillary disease in breast cancer. Eur J Surg Oncol 1993; 19: 213–214. 14. Davidson T. Why I favour axillary node clearance in the management of breast cancer. Eur J Surg Oncol 1995; 21: 5–7. 15. Greenall M J. Why I favour axillary node sampling in the management of breast cancer. Europ J Surg Oncol 1995; 21: 2–5. 16. Kjaergaard J, Blichert-Toft M, Amdersen J A, Rank F et al. Probability of false negative nodal staging in conjunction with
17.
18.
19.
20. 21. 22.
23. 24. 25. 26. 27.
28. 29.
41
partial axillary dissection in breast cancer. Br J Surg 1985; 72: 365–367. Steele R J C, Forrest A P M, Gibson T, Stewart H J, Chetty U. The efficacy of lower axillary sampling in obtaining lymph node status in breast cancer. A controlled randomised trial. Br J Surg 1985; 72: 368–369. Blichert-Toft M, Rose C, Andersen J A et al. Danish randomized trial comparing breast conservation therapy with mastectomy: six years of life-table analysis. Danish Breast Cancer Cooperative Group. Monogr Natl Cancer Inst 1992; 11: 19–25. Graverson H P, Blichert-Toft M, Anderson J A, Zedeler K. Breast cancer: risk of axillary recurrence in node-negative patients following partial dissection of the axilla. Eur J Surg Oncol 1988; 14:407–412. Mathieson O, Carl J, Bonderup O, Panduro J. Axillary sampling and the risk of eroneous staging of breast cancer: an analysis of 960 consecutive patients. Acta Oncol 1990; 29: 721–725. Axelsson C K, Mouridsen H T, Zedeler K. Axillary dissection of level I aand II nodes is important in breast cancer classification. Eur J Cancer 1992; 28A: 1415–1418. Kiricuta C I, Tausch J. A mathematical model of axillary lymph node involvement based on 1446 complete axillary dissections in patients with breast carcinoma. Cancer 1992; 69: 2496–2501. Carstairs V, Morris R. Deprivation and Health in Scotland. Aberdeen: Aberdeen University Press, 1988. Cox DR. Regression models and life-tables. J R Stat Soc 1972; 34: 187–220. Steele R, Forrest A, Gibson M et al. The efficacy of lower axillary sampling in obtaining lymph node status in breast cancer: a controlled randomized trial. Br J Surg 1985; 72: 368–369. Sakorafas G, Tsitou A. Sentinel lymph node biopsy in breast cancer. A review. Am Surg 2000; 66(7): 667–674. Early Breast Cancer Trials Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic or immune therapy; 133 randomised trials involving 31 000 recurrences and 24 000 deaths among 75 000 women. Lancet 1992; 339: 1–15, 71–85. Ragaz J, Jackson S M, Nhu Le et al. Adjuvant radiotherapy and chemotherapy in node-positive premenopausal women with breast cancer. N Engl J Med 1997; 337: 956–962. Overgaard M, Hansen P S, Overgaard J et al. Postoperative radiotherapy in high-risk premenopausal women with breast cancer who receive adjuvant chemotherapy. N Engl J Med 1997; 337: 949–955.