Isolated Tumor Cells in Bone Marrow Predict Poor Outcome in Many Breast Ca Patients

SAN ANTONIO—The question of how much weight to give isolated bone marrow micrometastases in making breast cancer treatment decisions was the focus of three presentations at the 26th San Antonio Breast Cancer Symposium. Two of these studies confirmed that occult micrometastases are a warning sign of poor prognosis in many breast cancer patients. The third found that isolated tumor cells in the sentinel lymph node are not a major problem in patients with ductal carcinoma in situ (DCIS).

Stephan Braun, MD, reported on behalf of European researchers in the Collaborative Group on Bone Marrow Mi-crometastasis that occult metastatic cells in the bone marrow are associated with worse overall survival for breast cancer patients as a whole and particularly for node-negative patients who have not received adjuvant therapy (abstract 7). He suggested that patients who test positive for occult metastatic cells may be candidates for more aggressive treatment strategies and that adjuvant therapy might not be needed in some patients who do not have such cells in their bone marrow.

Dr. Braun, of the University Clinic for Obstetrics and Gynecology, Innsbruck, Austria, reported a pooled analysis of 10-year survival for 4,268 breast cancer patients from eight recently published studies addressing long-term outcome of patients with and without occult metastatic cells. Patients received chemotherapy that was standard at the time of the study, but Dr. Braun pointed out that the study covered quite a long time, over which regimens changed.

The primary study endpoint was overall survival. Patients were free of distant metastases at primary surgery, 90% were stage T1/T2; 58% were node negative; 70% had received adjuvant therapy. Occult metastatic cells were detected in the bone marrow of 1,259 patients (30%). The proportion of patients with occult cells increased significantly (P < .001) across strata of tumor size, nodal status, and grading.

The pooled analysis showed significantly shorter disease-free survival in patients with occult metastatic cells (hazard ratio [HR] 2.10, P < .001). Multivariate analysis that included bone marrow positivity, lymph node positivity, tumor size, tumor grade, and estrogen-receptor (ER) positivity supported the importance of occult metastatic cells and showed a significant decrease in overall survival in patients positive for occult cells (HR 2.28, P < .001).

Dr. Braun said that the presence of occult metastatic cells is an independent predictive factor of survival for patients with stage I-III breast cancer.

The analysis included 10 years of follow-up with median follow-up of 58 months, during which 763 patients (18%) died. Death from any cause was significantly more likely in occult-cell-positive patients (HR 2.33, P < .001).

In the subgroup of 289 node-negative, occult-cell-positive patients who had not received adjuvant therapy (23% of the total), both disease-free and overall survival were significantly worse than in similar patients who did not have occult metastatic cells (P < .001).

"In pathologic node-negative breast cancer patients who have not had adjuvant therapy, this points to a role for occult metastatic cells in the progression of breast cancer. Occult cells should be used for risk stratification of patients in future clinical trials," Dr. Braun said. "Cells may leave the tumor early in the process and form a pool of cells that leads to metastasis later; thus, cells in the micrometastases may be different from cells in the primary tumor."

During the discussion period, one attendee pointed out that almost 70% of patients who had occult metastatic cells did not display overt tumors at 10 years of follow-up, which shows that not all micrometastases are clinically important. He suggested that cells that are more differentiated have limited pathologic capacity, and that looking for stem cell markers on micrometastases might provide more definitive information. Dr. Braun agreed that there is an urgent need to refine the assays for occult metastatic cells, which he described as "crude at present."

Norwegian Study

Bjorn Naume, MD, of the Norwegian Radium Hospital, Oslo, reported that the presence of isolated tumor cells in the bone marrow 3 years after diagnosis in disease-free breast cancer patients predicts an unfavorable outcome (abstract 8). "Bone marrow analysis can be used in the follow-up of patients for detecting early disseminating disease. Bone marrow positivity at 3 years might help select patients for secondary adjuvant therapy," he said.

This research group had previously found that isolated tumor cells in bone marrow were an independent prognostic factor for both distant disease-free survival and breast-cancer-specific survival in 817 breast cancer patients analyzed at the time of operation. Dr. Naume reported an analysis of outcomes correlated with results of a second bone marrow exam done at 3 years’ follow-up in 356 initially disease-free patients.

Bone marrow aspirated from the posterior iliac crest bilaterally was stained for immunocytochemical analysis with anticytokeratin mAbs (AE1/AE3) and alkaline phosphatase/antialkaline phosphatase reaction. Based on criteria from the European Working Group for Standardization of Tumor Cell Detection, any bone marrow with 1 or more cells was scored as positive.

At a median of 25.3 months’ follow-up after the second bone marrow aspiration (65.8 months since diagnosis), 32 patients had relapsed (12 local, 20 systemic). Relapse was significantly more common in patients who had isolated tumor cells in their bone marrow (21% vs 7%, P = .001).

Ten patients died of breast cancer during the observation period. Survival analyses revealed that isolated tumor cells in bone marrow predicted reduced distant disease-free and breast-cancer-specific survival (P < .001), and were an independent prognostic factor for both, Dr. Naume said. Patients with a positive bone marrow both at diagnosis and after 3 years had an especially poor prognosis and a systemic relapse rate of 29.4%.

DCIS Sentinel Node Micromets

Jan Wijsman, MD, of the Netherlands Cancer Institute/Antoni van Leeuwen-hoekhuis, Amsterdam, reported reassuring data showing that micrometastases in sentinel axillary lymph nodes of patients with ductal carcinoma in situ (DCIS) or small invasive carcinomas of the breast are not a harbinger of poor survival and should not trigger complete axillary lymph node dissection (ALND) in such patients (abstract 9).

"We undertook this study because the finding of a positive cell in patients with a nonmetastasizing disease raised some questions for us," Dr. Wijsman said.

The researchers resectioned all archived lymph node samples from a consecutive series of patients with DCIS or with invasive ductal or lobular carcinoma (IDC/ILC) of 5 mm or less who had undergone ALND as part of their treatment.

They compared the incidence of lymph node metastases after routine examination with that after re-evaluation of new sections of all archived lymph nodes immunostained for the epithelial marker CAM 5.2. The presence of macro- or micrometastases or solitary cells (ie, clusters of 10 to 20 cells) was scored and related to survival.

Immunostaining of all retrieved lymph nodes (mean, 14 per patient) showed metastases present in 11% of patients with pure DCIS and in 27% of patients with DCIS and microinvasion (see Table). "Most of these were single tumor cells or micrometastases. We found tumor cells in a single lymph node in five patients and in two nodes in two patients. We never found tumor cells in more than two nodes. All node-positive patients have remained free of disease at a median follow-up of 102 months, without adjuvant therapy."

Routine examination showed one macrometastasis in 71 patients with pure DCIS, no metastases in 12 patients with DCIS with microinvasion, and 1 macro-metastasis in 18 patients with IDC/ILC of 5 mm or less.

"Since no more than two nodes were affected, and survival was excellent, and previous studies have shown excellent prognosis without ALND in these low-risk tumors, we advise not to perform an ALND when a micrometastasis is found in the sentinel node of DCIS. ALND may also be omitted in T1a tumors in the absence of other negative prognostic factors," Dr. Wijsman said.