“EARLY” BREAST CANCER
AXILLARY CONTROVERSIES
Edward J. White
FOR DCIS
For patients with large “high risk” DCIS, or with DCIS and microinvasion, the probability of metastasis to axillary lymph nodes may be as high as 10-12%. The significance of these reports is not defined because of the known excellent results obtained historically in treating DCIS without axillary node dissection. In a study of 76 patients with high risk DCIS and 31 with DCIS and microinvasion underwent SLN biopsy. 12 patients overall had positive nodes. There were 9 in the pure high-risk group and 3 in the microinvasion group. Nine of the overall positive group were found to have micrometastases only. Six of the DCIS and 3 of the DCISM patients had completion ALND and one of the DCIS patients had an additional positive node. (Klauber-DeMore, Tan et al. 2000) A study at Moffitt of 87 patients with pure DCIS was conducted with SLN and IHC techniques. Five of the 87 patients (6%) had positive nodes. Three of these patients were only IHC positive and two were H&E positive. All SLNs that had only CK-positive cells were subsequently confirmed malignant by a more detailed histological examination of the nodes. Four of the patients with positive nodes had comedocarcinoma and one had a 9.5 cm tumor with mixed micropapillary and low grade cribriform. (Pendas, Dauway et al. 2000).
For small tumors (T1) with clinically a negative axilla, how useful is the sentinel node biopsy, what is the therapeutic significance of axillary dissection, what do we really learn from standard pathology, and what are the consequences of occult/micrometastatic disease?
Standard treatment for small T1 tumors has a good outcome in general. In a series of 962 patients (T1 to Small T2 N0-N1 M0) using standard pathology, those patients with T1a and T1b lesion had a combined 15-year survival of 90%, whereas those with T1c lesion dropped to 62%. Comparing node negative patients with node positive patient the results were 84% vs. 31%. Local relapse occurred in only 3.4% (33 of 980 treated cases). All patients were treated by BCT, including wide local excision, axillary dissection, postoperative radiation, and in the majority, adjuvant systemic therapy.(Vitucci, Tirelli et al. 2000) When Danish data for tumors less than or equal to 10mm was analyzed for 4771 patients, there was noted to be significantly better survival for those patients with 10 or more axillary nodes removed. They also noted that 8% of the patients had 4 or more nodes positive for metastases. These patients were retrospectively reviewed and had either mastectomy and ALND or BCT with ALND. (Axelsson, Rank et al. 2000). .
A small Swiss study of 44 patients found a SLN 93 % of the time, with 21 patients showing a positive SLN and 20 a negative node. One of the 20 had a positive node in the ALND specimen (5%); negative predictive value 95%. Of 17 patients with positive axillary nodes, only one had a negative SLN (5.9%) for a false negative rate of 5.9%. Two of the 41 patients had IHC identification of micrometastases (9.5%). They also noted that lymphoscintigraphy showed drainage to the internal mammary nodes in 2 of 28 patients (7%). (Langer, Zuber et al. 2000). For patients with invasive breast cancer, peritumoral lymphovascular invasion was most accurate in predicting for false negative sentinel nodes. (Noguchi, Kurosumi et al. 2000). Additional sections and special staining both increase the percentages of metastases identified. In a multi-center study of 214 node negative patients occult metastases were found in 15.9% of SLN and 4.2% of non-sentinel nodes.(Weaver, Krag et al. 2000) Another study of ALND in 423 patients with T1 lesions, showed a positive node in 1 of 31 T1a lesions (3%), 19 of 146 T1b (13%), and 61 of 246 T1c (25%). The average age was 61 and no focused pathological analysis was apparently used.(Lagares-Garcia, Garguilo et al. 2000). An Italian study of 102 patients with T1-T2 tumors showed a negative predictive value of 96.2% and a staging accuracy of 97.7%. They used Isotope technique, injecting the day before surgery, and found a sentinel node in 86.3% of the cases. Of 37 patients with positive axillary nodes 35 had positive SN (94.6% sensitivity), again a false negative rate of 5%. They also noted that in over 50% of their cases the sentinel node was the only positive node. (Casalegno, Sandrucci et al. 2000)
The question of recurrence after findings of negative lymph nodes is often discussed. The failure to identify nodal metastases is used in some patients to eliminate or downgrade subsequent adjuvant therapy, especially in small tumors. But around thirty percent of node negative patients are reported to develop recurrence within 10 years. Various series have reported that retrospective serial sectioning of the lymph nodes from these patients shows previously undetected metastases in 9-30% of the cases. And the accuracy of standard pathology studies of the axillary nodes has been studied many times. A recent study of 50 previously node negative patients showed a 10% micrometastasis by serial sectioning.(Karalak and Homcha-Em 1999). Another study the SLN specimens of 52 patients with invasive breast cancer compared H & E studies with serial sectioning and IHC staining. They found metastases in 12% of the former and 58% of the later. The 24 patients with IHC detection were divided into 12 patients with isolated cells and 12 with colonies.(Dowlatshahi, Fan et al. 1999). In a study of 736 patients with nodes negative by standard histology,occult nodal metastases were detected by serial sectioning and hematoxylin and eosin staining in 52 (7%) and by immunohistochemistry in 148 (20%). Only two (3%) of 64 invasive lobular or mixed invasive lobular and ductal cancers had node micrometastases detected by hematoxylin and eosin, compared with 25 (39%) by immunohistochemistry. Occult metastases, detected by either method, were associated with a significant decrease in disease free and overall survival in postmenopausal but not in premenopausal patients. Immunohistochemically detected occult lymph-node metastases remained an independent and highly significant predictor of recurrence even after control for tumor grade, tumor size, estrogen-receptor status, vascular invasion, and treatment assignment. (Cote, Peterson et al. 1999). Another recent report looked at the predictive factors for node metastases in 919 patients with T1a and T1b tumors. All patients underwent ALND. There were 199 patients with T1a tumors and 720 with T1b tumors. The overall incidence of metastases was 18%. Overall T1a metastasis was 16%(32 of 199). Overall T1b metastasis was 18.5%(133 of 720). These data were done on standard H & E studies and level I and II ALND. They noted significant influence on risk by patient age. For age > 50 the risk was 40 % less with small tumors. For patients < 40 the risk was 22.6% for T1a + T1B, but for > 70 years it was 10%. For tumors grade III the risk was 2.5X greater than grade I. Lymphovascular invasion also correlated strongly with increased risk of metastases. But even when they looked at patients with four favorable factors (T1a, well differentiated, no lymphovascular invasion, and age greater than 50), they found a 13% metastatic rate. For patients over 60 it dropped to 8.7%. They stated that because there is no “subgroup of patients that had an acceptable low risk of ALNM, the complete omission of assessing the axilla for metastatic disease in patients with small breast cancers cannot be advocated”.(Rivadeneira, Simmons et al. 2000). In the commentary on this paper a group of 2185 patients treated at Brown with tumors 1cm or less found a metastatic rate of 16% (11% for T1A and 17% for T1b). Age was again a major factor with a 31% rate in <40 and 15% for older patients.(Mustafa, Cole et al. 1997) In associating recurrence Rosen at Memorial Sloan Kettering in 1990 reviewed their data for patients treated 1964-1970 for tumors this size (1cm or less), treated by mastectomy and ALND, and showed an 18-year recurrence rate of 12% with mortality of 10% in node negative patients. This included 171 patients and the nodes were studied by standard pathology. If the nodes were positive the recurrence rate was 39% and the mortality 35%.(Rosen and Groshen 1990)
Haigh has also shown that method of tumor biopsy (by excision, core or needle), the volume of tissue removed and the timing of SLN biopsy do not affect the accuracy of the SLN technique. In 284 patients, treated with SLN biopsy and completion ALND, with a mean time of 17 days from tumor biopsy or excision to SLN biopsy they still found the SLN over 80% of the time and the negative predictive value remained over 95% (the false negative rate was 3.2%). They also noted that over 50% of the time the sentinel node was the only positive axillary node.(Haigh, Hansen et al. 2000)
It has also been shown that sentinel node biopsy is accurate and important following neoadjuvant chemotherapy. In nearly 50% of the patients it is the only positive node and it requires serial sectioning and IHC to identify 20% of the true metastases.(Breslin, Cohen et al. 2000; Cohen, Breslin et al. 2000)
Some sentinel nodes are outside the axilla. In a study of 113 patients 19% were found to have a sentinel node elsewhere. In only three cases was the nonaxillary node the only sentinel node. Twenty-two of the 30 identified nonaxillay sentinel nodes were harvested. Treatment changed in only three patients.(Jansen, Doting et al. 2000)
The risk of lymphedema is probably 20% with ALND and radiation therapy. It may be one half this if ALND is eliminated. Substituting SLN biopsy for ALND markedly reduces the early follow-up findings of lymphedema.(Schrenk, Rieger et al. 2000; Tengrup, Tennvall-Nittby et al. 2000)
What are the implications of missed axillary metastases, occult and micrometastatic? Multiple studies of node negative patients, re-reviewed for occult metastases by serial sectioning and/or IHC, as well as prospective studies of micrometastases, subsequently reviewed for outcome, have shown increased recurrence and worsened survival in patients with occult and/or micrometastases. These studies show a missed metastatic rate of 10 to 20 % and a reduction of disease free survival of 10 to 20%.(Friedman, Bertin et al. 1988; 1990; de Mascarel, Bonichon et al. 1992; Hainsworth, Tjandra et al. 1993) The data from the de Mascarel and Friedman studies showed that even single micrometastases implied a significant increase in recurrence.
What can be said about the potential benefit of axillary dissection with or without the use of radiation therapy and/or chemotherapy? Axelsson suggested that their data for T1a+T1 tumors implied a therapeutic advantage to axillary node dissection, with a decrease in axillary recurrence and an increase in survival if 10 or more nodes were removed. How radiation and chemotherapy were employed is not clear. (Axelsson, Rank et al. 2000). An upper bound for local axillary recurrence may be obtained from the ten-year follow-up of the NSABP-B04 study. In spite of the probable 40% positive nodes in the undissected patients, there was only a 17.8 % axillary failure rate in the non-treated axilla (i.e. no radiation and no dissection). These patients underwent delayed axillary dissection and ultimately only 4 of 365 patients failed in the axilla (1.1%).(Fisher, Redmond et al. 1985). Giuliano’s series of sentinel node negative patients, treated with no further axillary dissection, has shown no axillary recurrences in a 39-month mean follow-up.(Giuliano, Haigh et al. 2000) It should also be noted that the majority of those who did fail in the axilla in the NSABP study did so in the first 24 months. Osborne studied 211 patients with T1 and T2 lesions with clinically negative nodes who did not undergo axillary dissection, but had 6000 rads of external and interstitial radiation. They found only three recurrences (1.4%).(Osborne, Ormiston et al. 1984). Another study used to support the therapeutic value of axillary node dissection looked at 658 patients with T1 and T2 tumors. They were randomized to lumpectomy and radiation, with or without axillary node dissection. There was an improvement in survival in those who had axillary dissection, but only those who had pathologically proven disease on ALND received chemotherapy or tamoxifen.(Cabanes, Salmon et al. 1992) Vitucci establishes a lower bound for 15-year local/regional recurrence for standard BCT with ALND/Rad RX and adjuvant RX at 3.4%.(Vitucci, Tirelli et al. 2000). Another study that sought to clarify any survival benefit from axillary dissection combined results from six randomized studies for a meta-analysis. They reported a 5.4% improvement in survival with axillary dissection, however essentially no patients received adjuvant therapy and there were unusually few T1a tumors in the series.(Orr 1999)
In a criticism of the recent Canadian and Danish studies that appear to support the use of radiation to reduce locoregional recurrence in premenopausal high-risk patients, Silberman found that the median number of nodes removed in those studies was 7, compared to his 215 consecutive patients where the median number was 25. His contention is that the benefit from radiation was really a compensation for inadequate node removal. (Silberman, Sarna et al. 2000). Another study that looked at radiation therapy after either no AND (292 patients) or limited (5 nodes or less) AN sampling (126 patients) with an 8 year follow-up, found 1.4% developed RNF (6 patients). Four had simultaneous distant recurrence and 2 had isolated local recurrence. This study used axillary and supraclavicular ports. (Galper, Recht et al. 2000). Wong et al, in a study of 76 patients T1 tumors treated by excision and two field tangential standard radiation and no third nodal field, found no isolated nodal failure in a median 50 month follow-up.(Wong, Recht et al. 1997) Following up on this in a later paper studying 722 patients, Wong et al found that lymphovascular invasion predicted a significant increased risk of greater than four positive axillary nodes. They proposed that this might allow distinction between those who require addition of a third field for axillary/supraclavicular radiation.(Wong, O'Neill et al. 2000)
The success or failure of radiation therapy may depend considerably on the intact apoptotic pathway. This pathway is largely dependant on p53 and its related signaling proteins. These elements are damaged in a high percentage of breast cancers. It appears that identification of these abnormalities may predict radioresistence.(Dahm-Daphi 2000)
Kuerer et al showed the effect of chemotherapy on known positive axillary nodes in a study. Of 191 patients with locally advanced disease and cytologically proven axillary nodal metastases who were treated with neoadjuvant therapy, 23% were converted to a histologically negative axilla on subsequent AND. The lymph nodes from these 43 patients were studied with serial sectioning and IHC and only 4 (10%) were found to have occult metastases. Five-year survival was 87% in patients with pre-operative eradication of nodal disease and 51% on those with residual disease (it was 75% in those with occult residual disease).(Kuerer, Sahin et al. 1999) So it is probable that chemotherapy and radiation therapy can have a marked effect on the non-dissected axilla.
The question of the internal mammary nodes is also an issue. A recent prospective study of 80 consecutive patients showed that 12% had an identified IMSLN in addition to an axillary SLN.(Johnson, Soot et al. 2000) All patients underwent ALND and sentinel node biopsy. The quadrant did not predict well the presence of an IMSLN; 60% of the cases with a sentinel IM node were from lateral lesions. Metastatic disease was found in 3 of the 10 IMSLN cases, but all of these had positive axillary nodes. It is unclear whether they used IHC, but they did comment on historical data that showed in T1 and T2 tumors around 7% of nodal metastases may be isolated to the IM nodes. However, this data predated the more focused techniques of SLN evaluation by intense pathology and IHC now available. Numerous studies and reviews have failed to show any survival benefit in assessment or treatment of the internal mammary nodes. (Donegan 1977; Fisher, Redmond et al. 1985; Fowble, Hanlon et al. 2000; Freedman, Fowble et al. 2000; Sugg, Ferguson et al. 2000) An excellent review article on adjuvant radiation therapy for breast cancer supports the findings that there is a significant reduction in locoregional failure but that overall survival is not improved, largely because of the late cardiotoxicity. Without adjuvant radiation therapy, in node positive patients, the risk of locoregional failure at 10 years was 25% with 1-3 positive axillary nodes, and rose to 55% with 10 or more positive nodes. With radiation those recurrence rates dropped to 8% and 20% respectively.(Arriagada and Le 2000)
1000 ptsnegative
Negative SLN / 95%
Correct
/ 5%False Neg / Missed
50 Pts
1000 pts
eval by
SLN vs ALND / Find
5-10%
(Missed at
standard Path) / Gained
50 Pts
1000 pts
T1a+T1b
Standard Path
/ 20% + Nodes / 200 pts+ nodes / 200 pts
+ nodes
detected
By SLN / Get
95%
of the 200 / 190 pts
But with SLN and
SS and IHC / Get add’l
10% of
remaining
800 / 80 pts / 190 +80=
270 pts
Potential
Detected
+ nodes
Allow for only
Finding 90% SLN / 90% of 270 + / 243 pts / 243 patients
+ nodes / 243-200=
43 Extra pts
Detected
Node+ with
SLN vs Reg Path
For 1000
untreated
axillae / Minimum 25%
250 patients
positive (40%
Max)* / 180 *?*
axillary
recurrences
18% of 1000
(NSABP) / Less
Than 10%
Fail delayed
Rx / Less
Than
18 pts fail
sequence
No ALND
+ Salvage
Assume prior
Chemo+Rad
Recruits salvage in
Similar 1000 pts / Assume 50%
benefit / Less
Than
9 pts Fail after
no ALND+Rx
In 50% of cases
+SLN is only
positive axillary
node / 1000 pts
Assume 20%+
200+
100+ single / 900 pts
ALND adds
Nothing
Dx or Rx / Potential for
Benefit to
Only 10 % Max
AXILLARY AXIOMS
1)There is a subset of patients with small tumors and negative axillary nodes who have dissemination and are non-salvageable by maximum current therapy. (2-5%)
2)There is a persistent local failure rate that is not preventable by the maximum current therapy of mastectomy+ALND+chemo+radiation.
3)No randomized study shows a survival benefit to axillary dissection.
4)Delayed treatment of axillary nodes provides equivalent local control and survival benefit.
5)Treatment of the axilla reduces the risk of local recurrence, but does not improve overall survival.
6)Axillary dissection with removal of >10 lymph nodes may improve the recurrence free survival but not overall survival.
7)The axillary failure rate in clinically node negative patients, untreated by dissection, chemotherapy, or radiation is one-half the known metastatic rate (20% vs. 40%).
8)Standard pathology of the axillary nodes misses at least 10% of nodal metastases.
9)If knowledge of node status changes treatment recommendations, then you cannot rely on traditional pathology, especially with lobular carcinoma. (i.e. if you change the chemo rx based on positive or negative nodes).
10)New “Standard” requires SLN identification even if pre-decided on level I and II ALND, if micrometastasis would imply alteration of chemorx.
11)Sentinel node biopsy, though unsuccessful in 10% of patients, and false negative in 10% of patients, still provides an approximate 20% gain in the identification of axillary metastases over level l and ll ALND and standard pathology.