Fax: (011) 91 112658 8663
Touch imprint cytology of axillary lymph nodes after neoadjuvant chemotherapy in patients with breast carcinoma
Article first published online: 29 OCT 2003
Copyright © 2003 American Cancer Society
Volume 99, Issue 6, pages 346–351, 25 December 2003
How to Cite
Jain, P., Kumar, R., Anand, M., Asthana, S., Deo, S. V. S., Gupta, R., Bhutani, M., Karak, A. K. and Shukla, N. K. (2003), Touch imprint cytology of axillary lymph nodes after neoadjuvant chemotherapy in patients with breast carcinoma. Cancer, 99: 346–351. doi: 10.1002/cncr.11825
- Issue published online: 12 DEC 2003
- Article first published online: 29 OCT 2003
- Manuscript Accepted: 1 SEP 2003
- Manuscript Revised: 25 AUG 2003
- Manuscript Received: 7 FEB 2003
- Terry Fox Cancer Foundation, Canada
- breast carcinoma;
- imprint cytology;
- neoadjuvant chemotherapy;
- axillary lymph node metastasis
Neoadjuvant chemotherapy (NC) reportedly downstages axilla in approximately one-third of patients with locally advanced breast carcinoma (LABC). Postchemotherapy axillary lymph node status is an important prognostic factor. In the current study, the authors evaluated the reliability of touch imprint cytology (TIC) in detecting axillary lymph node metastasis after NC and identified chemotherapy-induced changes that may influence this assessment.
Thirty-three patients with LABC were studied. Seventeen patients had received chemotherapy before surgery (NC group) and 16 patients had not (non-NC group). Touch imprints were made from either the largest axillary lymph node (in 13 patients from the NC group and 16 patients from the non-NC) or the sentinel lymph node (in 4 patients from the NC group). Imprints from the NC group were evaluated for metastasis and were correlated with histopathology. Touch imprints from both groups were compared for cellularity, tumor load, necrosis/degeneration, and histiocytes.
Cytologic evaluation for metastasis was 100% concordant with histopathology in all 17 patients from the NC group (9 positive results [53%] and 8 negative results [47%]). The presence of few tumor cells in sparsely cellular imprints that exhibited necrosis (two patients) and the presence of only degenerating/necrotic tumor cells (two patients) were two cytologic patterns unique to post-NC imprints that may have influenced their accurate assessment.
TIC was found to be reliable for the intraoperative evaluation of axillary lymph node metastasis after NC. However, a careful examination is warranted in sparsely cellular imprints, because there is the possibility of overlooking a small group of tumor cells. To the authors' knowledge, the significance of finding extensive necrosis in axillary lymph nodes after NC is not known and may be investigated. Cancer (Cancer Cytopathol) 2003;99:346–51. © 2003 American Cancer Society.
Neoadjuvant chemotherapy (NC) is used increasingly in the treatment of patients with locally advanced breast carcinoma (LABC). In addition to providing breast conservation, it induces axillary lymph node remission in approximately one-third of patients.1, 2 A negative axilla after chemotherapy is reported to not only predict higher survival rates in these patients but also makes them potentially suitable for nonsurgical management of the axilla, particularly with sentinel lymph node (SLN) biopsy emerging as a reliable tool for determining axillary lymph node status.2–7 Information regarding the postchemotherapy axillary lymph node status therefore is vital and, if available intraoperatively, would facilitate a decision regarding axillary lymph node dissection at the time of initial surgery itself. Frozen sections may be used for this purpose, but the procedure is time-consuming, results in significant tissue loss, and requires considerable investment.8 Touch imprint cytology (TIC), in contrast, is quicker, cheaper, provides clear cytologic details, and has a reported accuracy of approximately 95% for the intraoperative identification of axillary lymph node metastasis in patients with breast carcinoma.9, 10 However, to our knowledge there is a paucity of literature on the effect of chemotherapy on axillary lymph node cytology. After chemotherapy, the draining lymph node may exhibit histologic changes such as fibrosis and necrosis,11 which may influence the cellularity of a touch imprint adversely and mar the cytologic details. In the current study, we compared cytologic features of postchemotherapy lymph nodes with TIC in lymph nodes from chemotherapy-naïve patients, evaluated the accuracy of TIC in identifying axillary lymph node metastasis after NC, and identified postchemotherapy changes that may influence this prognostically and therapeutically important assessment.
MATERIALS AND METHODS
Thirty-three patients with LABC were studied. Seventeen patients had received chemotherapy prior to surgery (the NC group), and 16 patients had not (the non-NC group). In NC group, 16 patients received 3 cycles of an anthracycline-based regime (cyclophosphamide, epirubicin, and 5-fluorouracil) as induction chemotherapy, and 1 patient received 3 cycles of methotrexate, cyclophosphamide, and 5-fluorouracil chemotherapy. Four to 6 weeks after the last chemotherapy cycle, 15 patients underwent modified radical mastectomy, and 2 patients underwent breast-conservation surgery. All 16 patients in the non-NC group had undergone modified radical mastectomy. At surgery, the largest axillary lymph node (LAN) encountered during axillary lymph node dissection was identified for TIC examination in 29 patients (all 16 patients in the non-NC group and 13 patients in the NC group). Four patients in the NC group who were seen during later part of the study underwent SLN biopsy using the 1% isosulfan blue-dye method12 as part of a validation study on SLN mapping in patients with breast carcinoma at our center. Patients with LABC who were treated with NC had not undergone SLN mapping during the initial phase of SLN biopsy validation. Informed consent was obtained from all patients.
The lymph node (LAN or SLN) was sectioned into 5-mm slices, and a minimum of 2 imprints per cut surface were made on a slide in the operating room by the surgeon. The lymph node, which was labeled separately, was submitted along with the rest of the complete axillary lymph node dissection specimen for routine pathologic examination. If it was determined that a lymph node (LAN or SLN) from the NC group was tumor free on examination of the initial hematoxylin and eosin (H & E) section, then 3 deeper H & E sections at 50-μm intervals were examined, and an additional section at a third level was immunostained with a pan-cytokeratin monoclonal antibody (MNF-116; Dakopatts, Glostrup, Denmark).
On average, 7 imprints (range, 4–12 imprints) were studied per lymph node. The touch imprint smears were air dried, fixed in methanol, and subjected to a rapid Jenner–Giemsa staining procedure in which slides were stained for 2 minutes by 0.3% Jenner stain followed by counterstaining for 7 minutes with Giemsa stain diluted 1 in 3 with phosphate-buffered saline (pH 6.8). The imprint slides were examined independently by at least two oncopathologists for metastatic tumor deposits and were reported as positive, suspicious, or negative for malignancy. The average time taken for staining and reporting was 15 minutes. TIC results were compared with the histopathology results, which were taken as the gold standard. The touch imprint smears also were assessed for cellularity, tumor load, histiocytes, and necrosis/degeneration. The cellularity was evaluated subjectively as “adequate” if tumor cells and/or lymphoid cells occupied > 20% of the area in a low-power field (LPF; magnification, × 10) of any touch imprint or as “sparse” if tumor cells occupied < 20% of the area in an LPF of all touch imprints. Histiocytes were recorded as either none to occasional (0–1 histiocytes per LPF) or conspicuous (an average of ≥ 2 histiocytes per LPF). Necrosis/degenerative changes were recorded as either absent or present. Chi-square and Fisher exact tests (when cell counts were fewer than five) were used to assess differences in the cytologic features between the two groups.
Patients in the NC group were younger (mean age, 48 years; range, 29–60 years) than the patients in the non-NC group (mean age, 54 years; range, 35–75 years). On clinical examination, all 33 patients had TNM Stage III disease13 (T3N1M0, T3N2M0, or T4anyNM0 [i.e., tumor measuring > 5 cm in greatest dimension (T3), tumor any size with direct extension to the chest wall and skin (T4), metastasis to movable ipsilateral axillary lymph node(s) (N1), metastasis to ipsilateral axillary lymph node(s) fixed to one another or to other structures (N2), and no distant metastasis (M0)]). Of the 15 patients with clinically involved axilla prior to NC, 8 patients (53%) achieved a complete clinical axillary response, and 4 patients (27%) achieved a pathologic axillary disease remission after chemotherapy. On histologic examination, all patients had infiltrating ductal carcinoma. A mean of 15.1 lymph nodes (range, 7–28 lymph nodes) were dissected from the axillary specimen in the NC group, and a mean of 19.4 lymph nodes (range, 10–39 lymph nodes) were removed in the non-NC group. Fewer axillary lymph nodes were found to be positive for metastasis in the NC group (mean, 4.3 lymph nodes; range, 0–15 lymph nodes) compared with the non-NC group (mean, 9.7 lymph nodes; range, 0–33 lymph nodes).
Of 17 patients in the NC group, touch imprints were positive for malignancy in 9 patients (53%) and negative in 8 patients (47%). Although histologic examination of all 9 lymph nodes with positive touch imprints revealed macrometastasis (metastatic focus measuring > 2 mm in greatest dimension) on the initial H & E section, metastasis was not found even on additional levels and immunohistochemical evaluation of the 8 lymph nodes that had negative touch imprints. Thus, there was a 100% correlation (sensitivity, 100% and specificity, 100%) between TIC and histologic evaluation of lymph nodes for metastasis. TIC accurately predicted axillary lymph node status in all four patients (two positive and two negative) for whom SLN biopsy specimens were studied.
TIC findings in 17 patients from the NC group are shown in Table 1, and a comparison of cytologic features between the 2 groups has been summarized in Table 2. Tumor load in the positive touch imprints ranged from the presence of only a few tumor cells, to a variable number of tumor clusters, to sheets of tumor cells (Figs. 1, 2). Imprints from 13 patients in the NC group and from all 16 patients in the non-NC group were adequately cellular. Of the four patients in the NC group who had sparsely cellular touch imprints, three patients were positive for metastasis. Two of these patients (Patients 5 and 12) had only an occasional cluster or group of 2–3 tumor cells, which were missed by 1 of the 2 pathologists. Histiocytes were conspicuous in six patients from the NC group and in seven patients from the non-NC group. Necrosis and degenerative changes were noted only in the positive lymph nodes and were seen more frequently in the NC group (seven patients) compared with the non-NC group (three patients). Degenerated tumor cells had a relatively more condensed chromatin and had granular debris in the cytoplasm (Fig. 3). Necrotic tissue had an amorphous, eosinophilic, granular appearance; and necrotic tumor cells were identified by their pyknotic nuclei and lack of cytoplasm (Fig. 4). Although viable tumor cells always appeared to accompany necrosis in the non-NC group, two patients in the NC group exhibited only nonviable tumor cells on touch imprints. In Patient 5, only necrotic fragments and 2 partially degenerated tumor cells were seen and, in Patient 9, all tumor clusters were necrotic. Although histologic examination of the corresponding lymph nodes revealed extensive necrosis in both patients, viable metastasis was noted in one of them. In both patients, however, at least one other axillary lymph node had viable tumor cells on histologic examination.
|Patient no.||Cellularity||Tumor clusters||Histiocytesa||Necrosisb||Metastatic status|
|Imprint||H & E, IHC|
|Feature||Neoadjuvant chemotherapy||No neoadjuvant chemotherapy||P value|
|No. of tumor clusters|
|None to occasional||11||9|
The usefulness of TIC for the rapid intraoperative diagnosis of axillary lymph node metastasis in patients with breast carcinoma has been evaluated in many studies. Quill et al.9 and Anastasiadis et al.10 reported diagnostic accuracy of 95% and 94%, respectively, using Giemsa stain in a sampling of axillary lymph nodes. Compared with conventional H & E paraffin sections of SLNs in chemotherapy-naïve patients, touch imprint results have been reported with a sensitivity and accuracy of 82–97% and 95–98%, respectively.8, 14–16 To the best of our knowledge, the literature concerning post-NC assessment of SLN by TIC is limited to only 1 study in which touch imprints were falsely negative in 10% of patients.17 The majority of the false-negative events on TIC have been attributed to micrometastasis (metastatic focus measuring ≤ 2 mm in greatest dimension), and reliable intraoperative identification of micrometastasis remains a problem area.8, 16–18 Although the clinical significance of micrometastasis and its impact on patient survival is controversial, its presence in an SLN may warrant a complete axillary lymph node dissection.6, 19, 20 In the current series, TIC had 100% sensitivity and specificity for identifying lymph node metastasis. The high sensitivity and absence of any false-negative imprint results in the current study may be because of the fact that none of the lymph nodes we studied had micrometastasis. Thus, the data from the current study suggest that the accuracy of TIC in identifying lymph node metastasis (at least macrometastasis) does not appear to be influenced by the administration of chemotherapy prior to surgery. In addition, in all four patients who had SLN biopsy specimens studied, TIC of the SLN correctly predicted the final axillary lymph node status. With increasing experience, as more patients with LABC undergo SLN assessment after NC, TIC may be useful for rapid intraoperative evaluation of axillary status after chemotherapy.
When we compared cytologic features between the two groups, not unexpectedly, we observed sparser cellularity, lesser tumor load, and more necrosis/degeneration in touch imprints after NC. Although none of these cytologic differences reached a level of statistical significance, possibly because of the small sample size, we noted two cytologic patterns that were unique to post-NC touch imprints and reflected significant chemotherapy-induced changes in the lymph node. The presence of only a few tumor cells in a sparsely cellular touch imprint that exhibited necrosis comprised one such cytologic pattern. In the current study, this pattern was observed in two patients who had tumor cells that were missed initially, and TIC was interpreted falsely as negative by one of the pathologists. In a study of 31 axillary SLN after NC, Cohen et al. found few atypical tumor cells on review of the original H & E sections from 3 false-negative SLNs (1 negative SLN and 2 positive non-SLNs) with significant chemotherapy-induced changes.11 Thus, it appears that, after chemotherapy, a high index of suspicion and careful screening are warranted when examining touch imprints that exhibit sparse cellularity and necrosis lest a small group of tumor cells is overlooked.
The presence of only degenerating/necrotic tumor cells, the second cytologic pattern unique to the post-NC setting, was noted in two patients in our study. The fact that viable metastasis was observed on histologic examination of at least one other axillary lymph node in both of these patients suggests that the exclusive presence of necrosis in one axillary lymph node does not rule out viable tumor in others. It may be particularly important to bear this observation in mind while examining SLN biopsy specimens for predicting the axillary lymph node status. It has been found that extensive necrosis after chemotherapy is of prognostic importance in some solid tumors. In patients with osteogenic sarcoma who are treated with chemotherapy, the presence of predominant areas of necrosis with only scattered foci of histologically viable tumor cells has been reported to be correlated with better disease-free survival.21 In a series of patients with metastatic nonseminomatous germ cell tumor of the testis who were treated with NC, it was found that, compared with patients whose surgery showed only necrotic debris, patients who had residual tumor after chemotherapy were at much greater risk of recurrence.22 The clinical significance of finding extensive necrosis after NC in patients with breast carcinoma is not known. It may be of interest to investigate whether extensive necrosis after chemotherapy in the axillary lymph nodes of patients with lymph node-positive breast carcinoma identifies a subset of patients with a better prognosis. In addition, the presence of necrosis may be of diagnostic value: In the current study, it was seen specifically only in the lymph nodes that were positive for metastasis.
Despite the small number of patients, the results of the current preliminary study indicate that TIC is reliable and is a potentially useful method for the intraoperative evaluation of axillary lymph nodes after chemotherapy. A high index of suspicion and careful screening are warranted when examining touch imprints with sparse cellularity and necrosis in a post-NC setting. Further studies may be undertaken to evaluate the clinical significance of finding extensive necrosis in the axillary lymph nodes of patients with breast carcinoma who are treated with NC.
- 16Intraoperative assessment of sentinel lymph nodes in patients with breast carcinoma: accuracy of rapid imprint cytology compared with definitive histologic workup. Cancer (Cancer Cytopathol). 2002; 96: 150–156., , , et al.
- 20The significance of micrometastasis. In: CodyHS, editor. Sentinel lymph node biopsy. London: Martin Dunitz Ltd., 2002: 311–319..