Lymph node micrometastases from breast carcinoma

1188

Lymph Node Micrometastases from Breast Carcinoma Reviewing the Dilemma

Kambiz Dowlatshahi, M.D.1 Ming Fan, M.D.1 Howard C. Snider, M.D.2 Fahim A. Habib, M.D.1

BACKGROUND. The presence or absence of regional lymph node metastases has been one of the most important determining factors in recommending adjuvant chemotherapy for patients with breast carcinoma. However, because of the 15– 20% failure rate at 5 years for lymph node negative patients, other tumor-related prognostic factors have gained greater significance in this decision-making process.

1

Department of General Surgery, Rush-Presbyterian-St. Luke’s Medical Center, Chicago, Illinois. 2

Department of General Surgery, Baptist Medical Center, Montgomery, Alabama.

Many investigators have reported finding micrometastases that were not detected by routine sectioning of the lymph nodes but were identified by multiple sectioning and additional staining. This review attempts to evaluate the role of occult lymph node micrometastases and their relevance to disease recurrence. METHODS. A literature search of the entire MEDLINE data base was conducted. All relevant articles were reviewed for the criteria they used to define micrometastases. The frequency of detection of micrometastases by various methodologies and the prognostic significance of such deposits were examined. RESULTS. Tumor deposits involving the lymph nodes were found to be arbitrarily categorized as either micrometastases or macrometastases, with the cutoff point ranging from 0.2–2.0 mm. The detection rate of such deposits by conventional techniques was inadequate. Serial sectioning and immunohistochemistry appeared to increase the detection rate by 9–33%. A definite survival disadvantage was noted for patients with such occult metastases. CONCLUSIONS. Current routine histologic examination of regional lymph nodes underestimates breast carcinoma metastases. Serial sectioning and immunohistochemistry increase the yield but are too labor-intensive and expensive for routine use. However, the introduction of the sentinel lymph node biopsy in lieu of axillary lymph node dissection in cases of breast carcinoma holds promise for making these methods practical and cost-effective. Cancer 1997;80:1188–97. q 1997 American Cancer Society.

KEYWORDS: breast carcinoma, lymph nodes, micrometastases, occult metastases.

B Address for reprints: Kambiz Dowlatshahi, M.D., Department of Surgery, Rush-Presbyterian-St. Luke’s Medical Center, 1653 West Congress Parkway, Chicago, IL 60612. Received January 27, 1997; revision received May 14, 1997; accepted May 14, 1997.

reast carcinoma remains an important health care problem. It is the most common cancer in women (31%) and carries the second highest mortality rate, exceeded only by that of lung carcinoma.1 Traditional view holds that malignant cells disseminating from the primary tumor are first harbored in regional lymph nodes. They may remain quiescent there for a variable time but ultimately reach the systemic circulation. A more recent hypothesis is that most breast carcinomas are systemic from the outset, but the axillary lymph node status still has major prognostic implications.2 Therefore, it is understandable that in evaluation of the clinical stage of breast carcinoma and, consequently, therapy and outcome, emphasis continues to be placed on axillary lymph node status. Clinical examination of the axillary lymph nodes alone is inadequate, yielding many false-negative and even more false-positive findings. Thus, axillary lymph node

q 1997 American Cancer Society

/ 7b71$$1343

09-10-97 10:15:57

cana

W: Cancer

Lymph Node Micrometastases from Breast Carcinoma/Dowlatshahi et al.

dissection and subsequent histologic examination currently are the accepted ‘‘yardsticks.’’ In approximately 16% of patients in whom tumors are classified as lymph node negative by standard histologic examination, tumors recur within 5 years of initial treatment.3,4 Prospective randomized clinical trials have shown the beneficial effects of adjuvant therapy in these high risk patients.5 – 10 However, such therapy does carry attendant morbidity and mortality.11 Therefore, it would be of immense clinical value if this high risk group could be accurately identified. Technologic advances in diagnosis now allow the detection of smaller, frequently lymph node negative breast carcinomas. This earlier diagnosis augments the need to accurately identify the high risk subgroup so that they alone can receive adjuvant therapy and its injurious effects can be avoided in low risk patients. The discrepancy between lymph node status on conventional histologic examination and patient survival has been attributed to several factors.12 The carcinoma could be systemic at the time of diagnosis without involvement of regional lymph nodes, or the involved lymph nodes (e.g., internal mammary, Level 3) might not be examined. The involved lymph nodes might be removed and examined but metastases not detected because of inadequacies in diagnostic techniques. This review was undertaken to examine the various methods of detecting occult metastases in the regional lymph nodes, the distinction between microand macrometastases, and their prognostic significance. The ability of these occult metastases to define the high risk group requiring adjuvant therapy was examined, perhaps bridging the gap between negative lymph node status and outcome.

Review Shortcomings in the routine histologic evaluation of dissected axillary lymph nodes were first demonstrated by Saphir and Amromin in 1948.13 Employing the technique of serial sectioning and hematoxylin and eosin (H & E) staining of alternate sections, they detected additional metastatic disease in 10 of 30 patients (33%) with primary invasive carcinoma and negative axillary lymph nodes on routine examination. They concluded that the routine examination of axillary lymph nodes was inadequate to detect ‘‘obscure metastases.’’ In 1961 Pickren14 confirmed the increased detection of metastatic disease by serial sectioning negative lymph nodes at 12-mm intervals, finding occult metastases in 21 of 97 cases (22%). Follow-up at 5 years showed no recognizable difference in prognosis between the 2 groups. In 1971 Huvos et al.15 introduced an arbitrary distinction between axil-

/ 7b71$$1343

09-10-97 10:15:57

cana

1189

lary lymph node micrometastases (õ2 mm) and macrometastases (ú2 mm) and evaluated their significance. At 8 years of follow-up, survival rates were significantly better among patients with micrometastases than those with macrometastases and essentially the same as those without lymph node involvement. Attiyeh et al.16 reevaluated the group of Huvos et al. after 14 years of follow-up, still finding no adverse effect of micrometastases on survival. Because the 10- and 14year survival rates were the same, they believed that a 10-year follow-up was adequate for long term evaluation. In 1978 Fisher et al.17 serially sectioned negative lymph nodes at 5-mm intervals, staining every fourth section with H & E. The rationale for this sectioning interval was that because tumor cells were ú 20 mm in size, sections every 20 mm should reveal all metastatic foci. Conversion to positive lymph nodes was observed in 19 of 78 patients (24%), with tumor deposits occurring in 2 distinct patterns. In 10 of 19 patients (53%), free tumor emboli were observed in the peripheral sinus or subcapsular lymphatic spaces. Parenchymal involvement was present in 9 of the 19 patients (47%). The dimensions of the metastases ranged from 0.2 – 1.3 mm, averaging 0.35 mm. At a mean follow-up of 5 years, there was no statistically significant survival difference between patients with and those without occult micrometastases. In another study, Fisher et al.18 further subdivided micrometastases into those õ1.3 mm and those ú 1.3 mm but õ2 mm. They based this division on their previous report that 1.3 mm represented the largest metastasis detected by serial sectioning that was not found on routine histologic examination. Dividing metastases into those ú 2 mm or õ2 mm yielded similar disease free survival (DFS) and overall survival (OS) rates. The survival of patients with lymph node metastases õ 1.3 mm was similar to those patients without metastases but significantly better than those patients in whom the metastases were ú1.3 mm (P Å 0.01). Black et al.19 defined micrometastases as tumor occupying õ20% of the sectioned area. Even by this definition, no significant prognostic difference could be demonstrated between patients with and without micrometastases, but outcome was poorer among those with macrometastases (P õ 0.05). Lymph node histologic features were used to stratify risk. Only patients with operable disease by the 1974 TNM criteria were included. Lymphocyte predominance, an unstimulated state, and the presence of sinus histiocytosis were associated with a low recurrence rate (9%). Lymphocyte depletion, germinal center predominance, and the presence of micrometastases were associated with an intermediate recurrence rate (23%),

W: Cancer

1190

CANCER October 1, 1997 / Volume 80 / Number 7

whereas the presence of macrometastases was associated with a high recurrence rate (58%). Rosen et al.20 reviewed the original histology slides of 147 patients, using the 2 mm criterion to differentiate micro- and macrometastases. They also divided the study population into those with tumors õ 2 cm (T1N1M0) and tumors ú 2 cm (T2N1M0). In the first 3 years, only macrometastases in the T1N1M0 group were associated with a higher recurrence rate, and even that difference was not statistically significant. No difference could be observed between patients with and without micrometastases in the initial 6 years of study. A statistically significant difference evolved at 10 years of follow-up. The prognosis for women with a single micrometastasis and those with a single macrometastasis was nearly identical, but both groups had poorer survival than did lymph node negative patients (P Å 0.01). No statistically significant difference could be demonstrated for the T2N1M0 group. Rosen et al.21 further evaluated the association between intramammary lymphatic emboli and the presence of occult metastases with regard to treatment failures. The detection rate was 32% (9 of 28 cases) and all metastases were õ2 mm in greatest dimension. No significant effect on DFS or OS could be demonstrated. Wilkinson et al.12 retrospectively studied the increased detection rate with serial sectioning. A 17% conversion rate was found (89 of 525 cases). Of these 89, 20% were identified on review of the original sections. In 71 of the 89 (80%), the metastatic deposit could be detected only on resectioning. These data confirmed the findings of Huvos et al.15 and Attiyeh et al.16 that no significant DFS or OS disadvantage was observed in women with micrometastases compared with those without similar foci. They did find a significantly poorer outcome independent of lymph node status in patients with tumors õ 2 cm and in tumors with vascular or lymphatic invasion. The significance was greater when no occult metastases were present. Using immunohistochemical staining (IH) with 3 monoclonal antibodies (E29, HMFG2, and KL1), Wells et al.22 detected metastases in 7 of 45 cases (15.5%) previously believed to be lymph node negative. The increase in the detection rate was greater for lobular carcinoma than for infiltrating ductal carcinoma (IDC) (33% vs. 9%); most cases involved single or small clusters of tumor cells. In contrast, Sloane et al.,23 using a single monoclonal antibody to epithelial membrane antigen (EMA), found no increase in the detection rate. Bussolati et al.24 confirmed the value of IH, detecting metastases in 12 of 59 cases (20%) using the monoclonal antibodies EMA, HMFG2, and antikeratin compared with 5 of 50 cases (10%) using H & E alone. Metastases detected were mainly isolated noncohesive

/ 7b71$$1343

09-10-97 10:15:57

cana

cells either located in the sinuses or diffusely distributed in the lymphoid tissue, rendering the technique more useful in infiltrating lobular carcinoma (ILC), in which the neoplastic cells are smaller and may mimic sinus histiocytosis. The small number of patients and limited follow-up precluded evaluation of prognostic significance. Trojani et al.,25 using 5 monoclonal antibodies directed against epithelial antigens, found metastases in 21 of 150 cases (14%). They also noted a higher conversion rate for ILC. A mean follow-up of 10 years revealed a significant prognostic influence of micrometastases on both recurrence and survival (P Å 0.0025 and 0.02, respectively) for IDC. The sample size for ILC was too small to detect a similar significance. Bryne et al.26 evaluated IH by restaining the original slides with a monoclonal antibody to EMA, finding metastases in 4 of 40 cases (10%). Three of the four metastatic deposits were single cells. The increased yield was not believed to be of practical clinical value because of the additional time and expense involved. Using 2-mm serial sections and H & E stains, Friedman et al.27 found previously undetected tumor emboli in the sinus margins of lymph nodes in 43 of 456 cases (9.4%). They noted that these ‘‘clandestine metastases’’ increased the risk of distant recurrence (P Å 0.05; relative risk, 1.7) by the same proportion as parenchymal metastases and were important prognostic indicators. Berry et al.,28 using a panel of monoclonal antibodies (HMFG1, HMFG2, E29, and CAM 5.2) found metastases not detected by H & E staining in 17.3% of patients (13.1% for IDC and 37.5% for ILC). The limited number of patients and short follow-up precluded detailed statistical analysis. Sedmak et al.,29 using AE1/AE3 and anticytokeratin antibody, showed an 11% increase in the detection of metastases. No evaluation of prognosis was made. Apostolikas et al.30 obtained an increased detection rate with serial sectioning (7 of 50 patients [14%]) and noted that all occult metastases were õ1 mm in size. Raymond and Leong31 pointed out that the 2 mm limit for the differentiation of micrometastases was purely arbitrary. The greatest dimensions of metastases that they detected using CAM 5.2 and AE1/AE3 ranged from 0.2 – 0.6 mm. They found metastases in 7 of 30 patients (23%), noting an average of 50 carcinoma cells per metastatic deposit (range, 5 – 200 cells). The first major prospective study was that of the International (Ludwig) Breast Cancer Study Group.32 Serial sectioning and H & E staining in 921 patients with Stage I breast carcinoma revealed metastatic disease in 83 patients (9%). At 5 years the DFS rates for patients with micrometastases compared with those with negative lymph nodes were 58% versus 74% (P Å

W: Cancer

Lymph Node Micrometastases from Breast Carcinoma/Dowlatshahi et al.

0.003) and OS rates were 79% versus 88% (P Å 0.002). The addition of perioperative chemotherapy increased DFS in the former group. The authors concluded that ‘‘the pathological examination of a single H & E stained node section is probably no longer clinically tenable.’’ Neville et al.33 reported that at 6 years of follow-up, DFS (53% vs. 71%; P Å 0.0008) and OS (70% vs. 86%; P Å 0.0009) achieved even greater statistical significance. It was their opinion that the only remaining issue was that of achieving the increased detection of micrometastases without the tedium of serial sectioning. Springall et al.34 confirmed an increased detection rate with IH (8 of 40 [20%]) and reported that a larger study with longer follow-up was underway. Galea et al.,35 using IH staining, demonstrated a definite but smaller increase in the detection rate. In a study of 98 patients, 3-mm sections were cut from the previously sampled surface and stained with H & E and monoclonal antibodies to CAM 5.2 and NCRC-11. Occult metastases were detected in 9 of 98 sections (9%); all 9 were CAM 5.2 positive but 1 was missed by NCRC11. A 14-year follow-up revealed no survival disadvantage for patients with occult metastases. Chen et al.36 (1991) stained 4-mm sections in ‘‘negative’’ lymph nodes, using polyclonal and monoclonal antibodies to cytokeratin, a monoclonal antibody to milk fat globule membrane antigen, and polyclonal antibodies to carcinoembryonic antigen (CEA). They found previously unsuspected tumor cells or microfoci in 23 of 80 patients (29%). Follow-up in 61 patients revealed distant metastases in 18% of lymph node positive and 2% of lymph node negative patients (P õ 0.05). There was no difference in local recurrence. Bryne et al.,37 using monoclonal antibodies CA 153 and MCA, demonstrated an increased detection rate and showed that patients with micrometastases had a poorer outcome after a mean follow-up of 3.6 years (P õ 0.001). Using serial sectioning and a cutoff point of 0.5 mm to differentiate micro- from macrometastases, de Mascarel et al.38 found micrometastases in 7% of patients and macrometastases in 13%. There was no significant difference in prognosis between those groups, but there was a significant difference in recurrence (P Å 0.005) and survival (P Å 0.0369) between lymph node negative patients and those with even a single metastasis. Clayton and Hopkins39 found a statistically significant prognostic difference (P Å 0.05) between patients with micrometastases and those with no metastases only in a subset with primary tumors õ 1.8 cm. The authors hypothesized that there might be two types of micrometastases. One type was small simply because it was still in an early stage and eventually

/ 7b71$$1343

09-10-97 10:15:57

cana

1191

would become larger. The other might not be in an early stage but was small for other reasons such as an effective host response or insufficient angiogenesis. They postulated that the former would be associated with an effectively metastasizing tumor and would portend an unfavorable prognosis. The latter might be more common in larger tumors but might not necessarily be associated with a poor prognosis. Hainsworth et al.40 evaluated 343 cases with IH (BC2/BC3 and 3E1.2), finding occult metastases in 41 cases (12%), 10 of which also were found to be positive on blind review of H & E stains. Thirty-one of the occult metastases were õ2 mm in greatest dimension. At a median follow-up of 79 months, the presence of occult metastases predicted a higher recurrence rate (32% vs. 17%; P Å 0.05) but no difference in survival. When two or more lymph nodes were positive, there also was a decreased rate of survival (P Å 0.01). Elson et al.41 recut the original ‘‘negative’’ lymph nodes in 97 patients and stained the 5-mm sections with AE1/AE3 and DF3. Metastases were found in 20 of 97 patients (20.6%), 9 of whom (9.3%) had tumor cells that were overlooked on the original H & E sections. Nasser et al.42 took 5 sections from ‘‘negative’’ lymph nodes 150 mm apart and evaluated them with H & E and keratin immunostaining. Fifty of 159 patients (31%) had metastases, 28 of which (56%) were detected by both H & E and IH and an additional 22 (44%) only by IH. In 19 patients (38%), metastases were ú0.2 mm and in 31 (62%) they were °0.2 mm. Patients with metastases ú 0.2 mm had a worse recurrence rate (P Å 0.02), worse DFS (P Å 0.04), and worse OS (P Å 0.07) than did those with metastases õ 0.2 mm. Recurrence and survival in the latter group were comparable to those of lymph node negative patients. In a recent study of serial sectioning and IH staining, McGuckin et al.43 took 4 sections at 100-mm intervals from 208 patients. Occult metastases were detected in 53 (25%), 28 of which (53%) were located in the capsular or subcapsular region, 13 (25%) in the parenchyma, and 12 (23%) in both regions. Approximately 67% of the metastases were õ0.5 mm in greatest dimension. At a 5-year follow-up, DFS (P Å 0.007) and OS (P Å 0.021) were poorer in patients with occult metastases. Involvement of more than one lymph node, the size of metastases, and histologic and nuclear grade were significant prognostic indicators. Molland et al.44 demonstrated that the c-erb B-2 oncogene may be another useful way of detecting the presence of metastatic lymph node disease. In a study of 88 lymph node negative patients with primary operable breast carcinoma, 28 of 88 samples (32%) stained positive for the c-erb B-2 oncogene. Follow-up revealed a statistically significant DFS (P õ 0.001) and

W: Cancer

1192

CANCER October 1, 1997 / Volume 80 / Number 7

OS (P Å 0.003) difference that was evident at 2 years and persisted throughout the 5-year follow-up period. In recent years, several groups have used reversetranscriptase polymerase chain reaction (RT-PCR) to further improve the detection of lymph node metastases beyond that obtained with serial sectioning and IH. Using the marker cytokeratin 19, which is found in breast tissue but not in normal lymph nodes, Schoenfeld et al.45 examined 57 lymph nodes, 18 of which (32%) were known to contain metastatic disease and 39 of which (68%) were histologically negative. All 18 positive lymph nodes were positive with PCR as well. In addition, 4 of the histologically negative lymph nodes (10%) were positive with PCR and ethidium staining, and an additional 10 of 35 lymph nodes (28.6%) were positive after Southern blot hybridization. A two-stage amplification was too sensitive, detecting keratin 19 in patients with normal lymph nodes. Forty cycles of RT-PCR and Southern blot hybridization was the optimum method of distinguishing metastatic lymph nodes from normal lymph nodes. Mori et al.,46 studying a variety of cancers, evaluated the ability of RT-PCR to detect lymph node metastases by the expression of CEA mRNA. All 30 histologically positive lymph nodes also were positive with RT-PCR. RT-PCR detected evidence of tumor in 47 of 87 histologically negative lymph nodes, increasing the positive rate from 26% using histology to 66% using RT-PCR. Noguchi et al.47 studied 10 histologically positive and 53 histologically negative lymph nodes, using RT-PCR to detect MUC1 mRNA and keratin 19 mRNA. Both mRNAs were detected in all ten histologically positive lymph nodes. Keratin 19 was ten times as sensitive as MUC1 on dilution studies. Keratin 19 also detected evidence of tumor in 5 of 53 histologically negative lymph nodes (9%) in comparison with 3 (6%) with MUC1. In a phase of the study in which portions of all of the axillary lymph nodes were pooled before RTPCR, those in one patient were histologically positive but negative by RT-PCR. The authors speculated that this false-negative finding could have resulted from dilution by the normal lymph nodes in the axilla or possibly because the portion of the lymph node that was pooled contained no tumor.

DISCUSSION The advent of screening mammography has led to the increased detection of nonpalpable breast carcinomas, an increasing proportion of which are small and lymph node negative. This trend poses a dilemma for the clinician, who must decide which patients to treat with adjuvant systemic therapy. In ú15% of lymph node negative patients, tumors recur within 5 years of initial therapy. Several studies have shown that all

/ 7b71$$1343

09-10-97 10:15:57

cana

women with breast carcinoma, regardless of lymph node status, may derive some benefit from systemic therapy.5 – 10 Early results from the National Surgical Adjuvant Breast and Bowel Project (NSABP B-13 and B14) were considered significant enough for the National Cancer Institute to adopt a novel method of disseminating information even before its publication in peer-reviewed journals. They issued their first ever Clinical Alert9 concluding that ‘‘adjuvant hormonal or cytotoxic chemotherapy can have a meaningful impact on the natural history of node-negative breast cancer patients.’’ Fisher et al.7 reported the 8-year follow-up of NSABP B-13, one of the original studies that led to the Clinical Alert. They randomized 760 lymph node negative, estrogen receptor negative patients to receive either chemotherapy or placebo. DFS was significantly superior in the treated group compared with the placebo group (74% vs. 59%; P õ 0.001). This benefit was observed both in patients age ° 49 years (69% vs. 56%; P Å 0.006) and those age ¢ 50 years (89% vs. 80%; P Å 0.03). Similar advantages can be gained by the use of tamoxifen in lymph node negative, estrogen receptor positive tumors as shown by the NSABP B14 study. The early trialists’ large meta-analysis also revealed systemic therapy to be of benefit in both premenopausal and postmenopausal lymph node negative patients.5,6 On the basis of these data, many oncologists adopted a policy of treating all patients systemically, regardless of lymph node status, but this course often results in overtreatment of the majority of lymph node negative patients who would do well with locoregional therapy alone. These women are thus needlessly subjected to the risk, side effects, and expense of adjuvant therapy. The traditional method of evaluating axillary lymph nodes for metastases has been shown to be notoriously unreliable.12 – 43,45 – 48 The current challenge is to define a cost-effective method of determining which lymph nodes are truly negative and which harbor occult metastases. How this information then should be used in making therapeutic decisions needs clarification.

Definition of Micrometastases Lymph node metastatic involvement has been arbitrarily subdivided into micro- and macrometastases, usually according to the size of the tumor deposits, with the cutoff point ranging from 0.2 – 2.0 mm.15,42 One group of authors19 defined micrometastases as õ 20% involvement of the lymph node sectional area; this parameter is even more arbitrary because of the great variability in lymph node size and in enlargement due to factors other than metastases. The authors of the current study see no useful purpose in

W: Cancer

Lymph Node Micrometastases from Breast Carcinoma/Dowlatshahi et al.

continuing the arbitrary, inconsistent division of lymph node involvement into micro- and macrometastases. The presence of any tumor in a lymph node, regardless of size, must be considered in light of what it implies about tumor biology. A single malignant cell detected in a subcapsular sinus by IH has demonstrated the capacity to break through anatomic barriers (basement membrane, interstitium, etc.) but has not demonstrated its ability to survive in the hostile environment. Surrounded by lymphocytes and immune substrates, one of two fates awaits the intruder: it can be destroyed by the body’s immune system, or it can establish a new home, multiply, grow, and perhaps even spread from there to yet another location. Circulating tumor cells may be detected by RT-PCR as well, and their significance currently is unknown. In contrast, a group of cells in the parenchyma, regardless of how small, likely indicates the biology of the tumor allows survival and growth in sites other than the primary location. Whether some small clusters remain relatively dormant as postulated by Clayton and Hopkins39 whereas others grow aggressively, remains to be clarified.

Occult Metastasis Detection Rate Nearly half a century ago, it was clearly demonstrated that the procedure of obtaining a few sections through the middle of a lymph node and staining them with H & E was woefully inadequate for detecting metastases.13 Nevertheless, that technique remains the standard in most pathology laboratories. The increased yield of serial sectioning with H & E staining has been reported variously as ranging from 7–33% (Table 1).12–15,17–21,27,30,32 Analysis of these reports reveals great variability in terms of the sectioning interval. The Ludwig group32 sectioned each block at 6 levels and prepared 6 3-mm sections at each level. They found one of the lowest conversion rate (9%) of all studies to date and attributed their low yield to inadequate sectioning and the lack of IH staining. Use of a 48-mm sectioning interval21 increased the yield to 32%. Therefore, it is obvious that serial sectioning with H & E is superior to routine histopathology, but the sectioning interval has a significant impact on the detection rate. A number of studies have confirmed the increased ability of IH to detect metastatic cells in lymph node tissue. A variety of monoclonal antibodies have been used, with detection rates ranging from 10 – 23% (Table 1).22,23,25,26,34,37,41,48 The addition of IH has been shown consistently to increase the yield from 14% to 17% in comparison with H & E staining alone.24,28,29,42 The increased yield has been greater for infiltrating lobular carcinoma, in which the single small cells or clusters

/ 7b71$$1343

09-10-97 10:15:57

cana

1193

of cells may be mistaken for histiocytes on routine staining. The use of extensive serial sectioning and IH on all axillary lymph nodes is too expensive and laborintensive to be practical. However, it has been shown that the vast majority of lymph node metastases would be detected by taking 2 sections 0.3 mm apart and staining them with a single monoclonal antibody (BC2).43 It appears reasonable to implement such a policy for those patients in whom the detection of occult metastases would alter their adjuvant therapy. In many centers these patients would be those with tumors õ1 cm in size, who would not be treated unless lymph node metastases were demonstrated. The use of RT-PCR has been shown to detect occult lymph node metastases, but the usefulness of this technique awaits clarification. It even may be too sensitive, detecting single circulating cells that have not established the ability to survive in the lymph node and may have no clinical significance.

Occult Metastases and Their Relation to Prognosis The impact of occult metastases on prognosis has been an issue of great debate. Initial studies14 – 17 did not reveal any survival disadvantage in patients found to have ‘‘micrometastases.’’ Consequently, these small foci of tumor were not believed to be of clinical significance, and it was generally believed that a search for them probably was unwarranted. More recent studies have shown that opinion to be no longer tenable. Rosen et al.20 found that patients with T1 primaries had poorer DFS and OS at 10 years even if their lymph node metastases were õ2 mm in greatest dimension. No such survival disadvantage was found for T2 patients, supporting the later contention of Clayton and Hopkins39 that small lymph node metastases might be qualitatively different, depending on the size of the primary. In a large study, Wilkinson et al.12 failed to show a survival disadvantage for occult metastases, but every study since then of at least 100 patients has shown decreased DFS or OS for at least a subset of patients.25,27,32,38 – 40,42,43 This survival disadvantage has been demonstrated for small tumor deposits down to a size of only 0.2 mm.42 It is likely that with larger series and longer follow-up, even smaller metastases will be found to be significant. Rosen et al.49 showed that with 6-year follow-up, survival rates for patients with ‘‘micrometastases’’ were similar to those for lymph node negative patients, but at 10 years, survival rates were distinctly poorer for patients with ‘‘micrometastases’’ and similar to those of patients with ‘‘macrometastases.’’ Attiyeh et al.,16 although demonstrating no difference in outcome, did show that the rates at 10 and 14 years of

W: Cancer

1194

CANCER October 1, 1997 / Volume 80 / Number 7

TABLE 1 Studies of Micrometastases, their Detection Rate, and Prognostic Significance

Year

No. of cases

1948 1961 1971 1977 1978

30 51 208 105 78

Sloane et al.23 Black et al.19 Rosen et al.20 Rosen et al.21 Wilkinson et al.12

1980 1980 1981 1982 1982

39 172 147 28 525

Wells et al.22 Bussolati et al.24 Trojani et al.25 Byrne et al.26 Friedman et al.27 Berry et al.28 Sedmak et al.29 Apostolikas et al.30 Raymond and Leong31 Ludwig Breast Cancer Study Group32 Springall et al.34 Galea et al.35

1984 1986 1987 1987 1988 1988 1989 1988 1989 1990

45 50 150 40 456 50 5 50 30 921

1990 1991

40 98

Chen et al.36 Neville et al.33 Byrne et al.37 de Mascarel et al.38 Clayton and Hopkins39 Hainsworth et al.40 Elson et al.41 Nasser et al.42

1991 1991 1992 1992 1993 1993 1993 1993

80 921 39 336 399 343 97 159

McGuckin et al.43

1996

208

Author Saphir and Amromin Pickren14 Huvos et al.15 Attiyeh et al.16 Fisher et al.17

13

Sections 332 sections per block 12-mm sections — — Sec at 5 mm, stain every fourth Original section — Original section 48-mm sections 24-mm and 48-mm sections Serial sections 5 serial sections Original sections Original sections 1-2-mm serial sections 4-mm sections 4-mm sections 6 sec of 5 mm each Single section 6 levels per block — 4 3-mm sec from prv suf 4-mm sections 6 levels per block 2 new sections 1.5-mm slices — 4-mm sections 4-mm sections 5-mm sec at 150-mm intervals 4 levels 100 mm apart

Staining method

% of lymph nodes positive

Follow-up period (yrs)

DFS

OS

H&E H&E H&E H&E H&E

33 22 — — 24

3 5 8 14 5

— NS NS NS NS

— NS NS NS NS

IH H&E H&E H&E H&E

No increase 7.50 — 32 17

— 3 10 10 15

— NS — NS NS

— NS P Å 0.01b NS NS

H & E and IH H & E and IH IH IH H&E IH H & E and IH H&E IH H&E

20 24 14 10 9.40 17.40 30 14 23 9

— 3.5 10 5 10 2 — — — 5

— — P Å 0.0025 NS — — — — — P Å 0.003

— — P Å 0.02 NS P Å 0.05 — — — — P Å 0.002

IH H & E and IH

20 9

— 14

— NS

— NS

IH H&E IH IH H&E IH IH H & E and IH

29 9 12.80 7 — 12 20.60 31

3.2 6 3.6 7 16.7 6 5.7 11

P õ 0.05 P Å 0.0008 P õ 0.001 P Å 0.005 P õ 0.05c P ú 0.05 NS P Å 0.02c

— P Å 0.0009 — P Å 0.0369 — NS NS P Å 0.07

H & E and IH

25

5

P Å 0.021

P Å 0.007

DFS: disease free survival; OS: overall survival; sec: section; prv suf: previously sampled surface; H & E: hematoxylin and eosin; IH: immunohistochemistry; NS: not significant. a For metastases ú1.3 mm in greatest dimension. b For T1N1M0 tumors only. c For the subset of primary tumors õ1.8 cm. d For metastases ú0.2 mm.

follow-up were similar. This suggests that future studies should have at least 10 years of follow-up. At first glance, it is puzzling that many early studies failed to show survival differences in patients with occult metastases, whereas more recent studies have consistently shown such differences. Closer scrutiny reveals that, with the exception of the article by Wilkinson et al.,12 the differences can be accounted for by patient population size, length of follow-up, the use of IH staining, or a combination of the three. Many of the studies that have failed to show survival differences

/ 7b71$$1343

09-10-97 10:15:57

cana

had inadequate numbers of patients (range, 28 – 105 patients) to detect small differences in survival.14,16,17,21,26,35,41 Others had relatively short followup periods (range, 3 – 5.7 years), which are inadequate to detect survival differences in patients with breast carcinoma.14,17,19,26,42 All studies that showed survival differences had larger patient populations (range, 147 – 921 patients),20,25,27,32,38 – 40,42,43 follow-up of at least 6 years,20,25,27,33,39,42 or detection of occult lymph node metastases by IH.25,36 – 38,40,42,43 It is likely that larger numbers of patients, longer follow-up, or the conver-

W: Cancer

Lymph Node Micrometastases from Breast Carcinoma/Dowlatshahi et al.

sion of some of the occult metastases from the ‘‘negative’’ groups by IH would have resulted in the demonstration of survival differences in most of the series. The authors are unable to account for the lack of survival difference in the series by Wilkinson et al., which had a large patient population (525), long follow-up (15 years), and a reasonable conversion rate (17%).12 The relatively recent application of sentinel lymph node biopsy to patients with breast cancer holds promise for making the use of serial sectioning and IH staining practical and cost-effective for routine use. Such a technique already has been demonstrated to be reliable in detecting lymph node metastases in patients with melanoma.50 Presented with a single lymph node (or a small number of lymph nodes) most likely to harbor metastases, pathologists could devote more time and resources to an exhaustive evaluation of that lymph node. For sentinel lymph node biopsy either a radioactive (technetium Tc 99m sulfur colloid) or visual (isosulfan blue) tracer is injected into the region of the tumor, and the lymph node to which the tracer drains is determined intraoperatively. Theoretically, tumor cells would have coursed through the same lymphatics to the same lymph node. Using gamma probe localization, Krag et al.51 found the radiolabeled sentinel lymph node in 18 of 22 patients (82%). The sentinel lymph node was positive in all 7 patients with axillary lymph node metastases and was the only focus of metastases in 3 patients (42.8%). Using isosulfan blue, Giuliano et al.,52 found the sentinel lymph node in 114 of 174 patients (65.5%). The sentinel lymph node accurately predicted axillary lymph node status in 109 of the 114 patients (95.6%). With experience, Giuliano et al. were able to detect the sentinel lymph node in 78% of patients.52 The authors’ own experience with use of a gamma probe revealed a detection rate ú 90%, with a relatively short learning curve (unpublished data). When the sentinel lymph node was evaluated with serial sectioning and IH, Giuliano et al.53 found metastases in 68 of 162 patients (42%) in comparison with 39 of 134 patients (29%) when the full axillary lymph node dissection was processed in the usual manner (P õ 0.03). The metastases were smaller than 2 mm in 4 of 39 patients (10.3%) in the full axillary lymph node dissection group, compared with 26 of 68 patients (38.2%) in the sentinel lymph node group (P õ 0.0005). Although sentinel lymph node biopsy with serial sectioning and IH may offer a highly accurate, costeffective method of determining axillary lymph node status in the future, insufficient data currently exist to support that assumption. A large multicenter trial currently is in progress evaluating some aspects of sentinel lymph node biopsy. Its conclusion and publica-

/ 7b71$$1343

09-10-97 10:15:57

cana

1195

tion will no doubt shed some light on the accuracy of the procedure. However, serial sectioning and IH staining are not a part of that study so one would anticipate that a significant number of occult lymph node metastases will be missed. In a limited study currently in progress, the authors found occult metastases in 3 of 13 positive sentinel lymph nodes (23%) only by serial sectioning. The metastases would have been missed by routine sectioning. To the authors’ knowledge, no studies (including the current study) have examined both the sentinel lymph node and all other axillary lymph nodes by serial sectioning or IH. Until that is done, one cannot say that occult metastases are not being overlooked in nonsentinel lymph nodes when the sentinel lymph node is negative. The heterogeneous nature of the series in this review and the fact that most series did not analyze such prognostic factors as nuclear grade, DNA ploidy, Sphase fraction, hormone receptor status, and lymphatic/vascular invasion (LVI) precluded the authors’ ability to analyze their impact on adjuvant therapy or survival. Thus, the authors were unable to conclude that occult lymph node metastases are an independent predictor of survival separate from all other prognostic factors. It is possible that in the future other tumor parameters may render axillary lymph node status moot. Chadha et al.54 demonstrated that tumor size, nuclear grade, LVI, and age were all determinates for lymphatic metastases. Tumor size and LVI were independent predictors in multivariate logistic regression analyses. Seventeen of 25 patients (68%) with both features had lymph node metastases in contrast with 7 of 79 patients (9%) who had neither feature. Bland et al.55 evaluated Ha-ras, c-myc, c-fos, and p53 as prognostic discriminates for breast carcinoma, finding an increasing likelihood of recurrence with increasing oncogene coexpression. Expression of 1 oncogene predicated a 17.2% chance, 2 oncogenes a 56.3% chance, and 3 or 4 oncogenes a 100% chance of recurrence (P Å 0.001). Future studies should include a comparison between sentinel lymph nodes and the remainder of the axillary lymph nodes in which all lymph nodes are subjected to serial sectioning and IH. The number of patients evaluated should be large enough to substratify them according to a number of variables, including not only size but also oncogene expression and enzyme activity. Only then can a determination be made as to whether the findings on sentinel lymph node biopsy accurately predict prognosis and serve as independent variables.

CONCLUSIONS Advances in diagnostic techniques have led to the earlier detection of breast carcinomas, the vast majority of

W: Cancer

1196

CANCER October 1, 1997 / Volume 80 / Number 7

which are found to be lymph node negative by routine histology. Numerous studies have demonstrated that up to 33% of these patients harbor occult tumor deposits in the lymph nodes, which can be detected only with serial sectioning, IH, or RT-PCR. In the past decade, large studies with adequate follow-up have demonstrated consistently that these small metastases have important prognostic implications and therefore could be important in making therapeutic decisions. The routine evaluation of these occult metastases must be incorporated into clinical practice. It is hoped that the refinement and widespread use of sentinel lymph node biopsy will allow their accurate detection without requiring undue expenditure of time and resources.

12.

13. 14. 15.

16.

17.

REFERENCES 1.

Parker SL, Tong T, Bolden S, Wingo PA. Cancer statistics. 1996. CA Cancer J Clin 1996;46:5–27. 2. Fisher B, Redmond C, Fisher ER. The contribution of recent NSABP clinical trials of primary breast cancer therapy to an understanding of tumor biology—an overview of findings. Cancer 1980;46:1009–25. 3. Fisher B, Bauer M, Wickerham DL, Redmond CK, Fisher ER, Cruz AB, et al. Relation of number of positive axillary nodes to prognosis of patients with primary breast cancer; an NSABP update. Cancer 1983;52(9):1551–7. 4. Rosen PP, Saigo P, Braum DW Jr., Weathers E, Depalo A. Predictors of recurrence in Stage I (T1N1M0) breast carcinoma. Ann Surg 1981;193:15–25. 5. Early Breast Cancer Trialists’ Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic or immune therapy: 133 randomized trials involving 31,000 recurrences and 24,000 deaths among 75,000 women. Lancet 1992;339(8785):1–15. 6. Early Breast Cancer Trialists’ Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic or immune therapy: 133 randomized trials involving 31,000 recurrences and 24,000 deaths among 75,000 women. Lancet 1992;339(8785):71–85. 7. Fisher B, Dignam J, Mamounas EP, Constantino JP, Wickerham DL, Redmond C, et al. Sequential methotrexate and fluorouracil for the treatment of node-negative breast cancer patients with estrogen receptor-negative tumors: eightyear results from National Surgical Adjuvant Breast and Bowel Project (NSABP) B13 and first report of findings from NSABP B-19 comparing methotrexate and fluorouracil with conventional cyclophosphamide, methotrexate and fluorouracil. J Clin Oncol 1996;14(7):1982–92. 8. Goldhirsch A, Castiglione M, Gelber RD. The International Breast Cancer Study Group. A single perioperative chemotherapy course for node-negative breast cancer: five year results of trial V. Monogr Natl Cancer Inst 1992;11:89–96. 9. NCI. Alert on node-negative breast cancer. Breast Cancer Res Treat 1988;12:3–5. 10. Fisher B, Redmond C. The National Surgical Adjuvant Breast and Bowel Project. Systemic therapy in node-negative patients: updated findings from NSABP clinical trials. Monogr Natl Cancer Inst 1992;(11):105–16. 11. Ludwig Breast Cancer Study Group. Toxic effects of early

/ 7b71$$1343

09-10-97 10:15:57

cana

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

adjuvant chemotherapy for breast cancer. Lancet 1983; 2(8349):542–4. Wilkinson EJ, Hause L, Hoffman RG, Kuzma JF, Rothwell DJ, Donegan WL, et al. Occult axillary lymph node metastases in invasive breast carcinoma: characteristics of the primary tumor and significance of the metastases. Pathol Annu 1982;17(Pt 2):67–91. Saphir O, Amromin GD. Obscure axillary lymph node metastases in carcinoma of the breast. Cancer 1948;1:238–41. Pickren JW. Significance of occult metastases. A study of breast cancer. Cancer 1961;14:1266–71. Huvos AG, Hutter R, Berg JW. Significance of axillary macrometastases and micrometastases in mammary cancer. Ann Surg 1971;173(1):44–6. Attiyeh FF, Jensen M, Huvos AG, Fracchia A. Axillary micrometastases and macrometastases in carcinoma of the breast. Surg Gynecol Obset 1977;144(6):839–42. Fisher ER, Swamidoss S, Lee CH, Rockette H, Redmond C, Fisher B. Detection and significance of occult axillary node metastases in patients with invasive breast cancer. Cancer 1978;42:2025–31. Fisher ER, Palekar A, Rockette H, Redmond C, Fisher B. Pathologic findings from the National Surgical Adjuvant Breast and Bowel Project (Protocol #4) V. Significance of axillary nodal micro- and macrometastases. Cancer 1978; 42:2032–8. Black RB, Roberts M, Stewart HJ, Prescott R, Cant EL, Sumerling MD, et al. The search for occult metastases in breast cancer: does it add to established staging methods? Aust N Z J Surg 1980;50(6):574–9. Rosen PP, Saigo P, Weathers E, Fracchia AA, Kinne DW. Axillary micro- and macrometastases in breast cancer: prognostic significance of tumor size. Ann Surg 1981;194(5):585–91. Rosen PP, Saigo P, Braun DW Jr., Beattie EJ Jr., Kinne DW. Occult axillary lymph node metastases from breast cancers with intramammary lymphatic tumor emboli. Am J Surg Pathol 1982;6(7):639–41. Wells CA, Heryet A, Brochier J, Gatter KC, Mason DY. The immunohistochemical detection of axillary micrometastases in breast cancer. Br J Cancer 1984;50(2):193–7. Sloane JP, Ormerod M, Imrie SF, Coonbes RC. The use of antisera to epithelial membrane antigen in detecting micrometastases in histological sections. Br J Cancer 1980;42(3):392–8. Bussolati G, Gugliotta P, Morra I, Pietribiasi F, Berardengo E. The immunohistochemical detection of lymph node metastases from infiltrating lobular carcinoma of the breast. Br J Cancer 1986;54(4):631–6. Trojani M, de Mascarel I, Bonichon F, Coindre JM, Delsol G. Micrometastases to axillary lymph nodes from carcinoma of the breast. Br J Cancer 1987;55(3):303–6. Byrne J, Waldron R, McAvinchey D, Dervan P. The use of monoclonal antibodies for the histopathological detection of mammary axillary micrometastases. Eur J Surg Oncol 1987;13(5):409–11. Friedman S, Bertin F, Mouriesse H, Benchabat A, Genin J, Sarrazin D, et al. Importance of tumor cells in axillary node sinus margins (clandestine metastases) discovered by serial sectioning in operable breast cancer. Acta Oncol 1988; 27(5):483–7. Berry N, Jones D, Marshall R, Smallwood J, Taylor I. Comparison of the detection of breast carcinoma metastases by routine histological diagnosis and by immunohistochemical staining. Eur Surg Res 1988;20(4):225–32.

W: Cancer

Lymph Node Micrometastases from Breast Carcinoma/Dowlatshahi et al. 29. Sedmak DD, Meineke T, Knechtges DS. Detection of metastatic breast carcinoma with monoclonal antibodies to cytokeratins. Arch Pathol Lab Med 1989;113:786–9. 30. Apostolikas N, Petraki C, Agnantis AG. The reliability of histologically negative axillary lymph nodes in breast cancer. Pathol Res Pract 1988;184:35–8. 31. Raymond WA, Leong A-Y. Immunoperoxidase staining in the detection of lymph node metastases in Stage I breast cancer. Pathology 1989;21:11–5. 32. International (Ludwig) Breast Cancer Study Group. Prognostic importance of occult axillary lymph node micrometastases from breast cancer. Lancet 1990;335(8749):1565–8. 33. Neville AM, Price K, Gelber RD, Goldhirsch A. Axillary node micrometastases and breast cancer. Lancet 1991;337(8749): 1110. 34. Springall RJ, Rytina E, Millis RR. Incidence and significance of micrometastases in axillary lymph nodes detected by immunohistochemical techniques. J Pathol 1990;160:174A. 35. Galea MH, Athanassiou E, Bell J, Dilks B, Robertson JF, Elston CW, et al. Occult regional lymph node metastases from breast carcinoma. Immunohistochemical detection with antibodies CAM 5.2 and NCRC-11. J Pathol 1991;165(3):221–7. 36. Chen Z-L, Wen D-L, Coulson WF, Giuliano AE, Cochran AJ. Occult metastases in the axillary lymph nodes of patients with breast cancer node negative by clinical and histological examination and conventional histology. Dis Markers 1991; 9(5):239–48. 37. Byrne J, Horgan P, England SS, Callaghan J, Given HF. A preliminary report on the usefulness of monoclonal antibodies to CA 15-3 and MCA in the detection of micrometastases in axillary lymph nodes draining primary breast cancer. Eur J Surg 1992;28:658–60. 38. de Mascarel I, Binichon F, Coindre JM, Trojani M. Prognostic significance of breast cancer axillary node micrometastases assessed by two special techniques: re-evaluation with longer follow-up. Br J Cancer 1992;66(3):523–7. 39. Clayton F, Hopkins CL. Pathologic correlates of prognosis in lymph node positive breast carcinomas. Cancer 1993; 71(5):1780–90. 40. Hainsworth PJ, Tjandra J, Stillwell RG, Machet D, Henderson MA, Rennie GC, et al. Detection and significance of occult metastases in node-negative breast cancer. Br J Surg 1993; 80(4):459–63. 41. Elson CE, Kulfe D, Johnson WW. Immunohistochemical detection and significance of axillary lymph node micrometastases in breast carcinoma. Anal Quant Cytol Histol 1993;15(3):171–8. 42. Nasser IA, Lee A, Bosari S, Saganich R, Heatley G, Silverman

/ 7b71$$1343

09-10-97 10:15:57

cana

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

1197

ML. Occult axillary lymph node metastases in ‘‘node-negative’’ breast carcinoma. Hum Pathol 1993;24(9):950–7. McGuckin MA, Cummings MC, Walsh MD, Hohn BG, Bennett IC, Wright RG. Occult axillary node metastases in breast cancer: their detection and prognostic significance. Br J Cancer 1996;73(1):88–95. Molland JG, Barraclough BH, Gebski V, Miliken J, Bilous M. Prognostic significance of c-erbB-2 oncogene in nodenegative breast cancer. Aust N Z J Surg 1996;66(2):64–70. Schoenfeld A, Luqmani Y, Smith D, O’Reilly S, Shousha S, Sinnett HD, et al. Detection of breast cancer micrometastases in axillary lymph nodes using polymerase chain reaction. Cancer Res 1994;54(11):2986–90. Mori M, Momori K, Inoue H, Barnard GF, Tsuji K, Nanbara S, et al. Detection of cancer micrometastases in lymph nodes by reverse transcriptase-polymerase chain reaction. Cancer Res 1995;55(15):3417–20. Noguchi S, Aihara T, Motomura K, Inaji H, Imaoka S, Koyama H. Detection of breast cancer micrometastases in axillary lymph nodes by means of reverse transcriptase-polymerase chain reaction. Am J Pathol 1996;148(2):649–56. Trojani M, de Mascarel I, Coindre JM, Delsol G. Micrometastases to axillary lymph nodes from invasive lobular carcinoma of the breast: detection by immunohistochemistry and prognostic significance. Br J Cancer 1987;56:838–9. Rosen PP, Saigo P, Braun DW, Weathers E, Kinne DW. Prognosis in stage II (T1N1M0) breast cancer. Ann Surg 1981; 194(5):576–84. Morton D, Wen D-R, Wong JH, Economou JS, Cagle LA, Storm FK, et al. Technical details of intraoperative lymphatic mapping for early stage melanoma. Arch Surg 1992;127(4): 392–9. Krag DN, Weaver D, Alex JC, Fairbank JT. Surgical resection and radiolocalization of sentinel lymph node in breast cancer using a gamma probe. Surg Oncol 1993;2:335–9. Giuliano AE, Kirgan D, Guenther JM. Lymphatic mapping and sentinel lymphadenectomy for breast cancer. Ann Surg 1994;220:391–401. Giuliano AE, Dale P, Turner RR, Morton DL, Evans SW, Krasne DL. Improved axillary staging of breast cancer with sentinel lymphadenectomy. Ann Surg 1995;222(3):394–401. Chadha M, Chabon AB, Friedmann P, Vikram B. Predictors of axillary lymph node metastases in patients with T1 breast cancer: a multivariate analysis. Cancer 1994;73:350–3. Bland KI, Konstadoulakis MM, Vezeridis MP, Wanebo HJ. Oncogene protein coexpression: value of Ha-ras, c-myc, cfos, and p53 as prognostic discriminates for breast carcinoma. Ann Surg 1996;221(6):706–20.

W: Cancer