- Email: [email protected]
Morphologic Effects of Neoadjuvant Chemotherapy in Locally Advanced Breast Cancer
PATHOLOGY RESEARCH AND PRACTICE © Gustav Fischer Verlag
Morphologic Effects of Neoadjuvant Chemotherapy in Locally Advanced Breast Cancer U. M. Moll and J. Chumas Department of Pathology, State University of New York at Stony Brook, Stony Brook, New York
Summary
The use of chemotherapy as primary treatment in early and locally advanced breast cancer is rising. As a result, many resected tumors were exposed to cytotoxic drugs in vivo. To study resulting histopathologic changes, we examined 61 patients with locally advanced stage III breast cancer who had been treated with a standardized neoadjuvant polychemotherapy regimen before undergoing surgical resection 3 months later. Matched pairs of pre- and posttherapy breast tissue were evaluated for morphologic changes in the residual malignant and benign breast tissue compartment. A potential correlation between changes and the original p53 immunophenotype was examined as well. In 11 cases (18%), complete pathologic remission with no residual tumor in the mastectomy specimen was achieved. This response was not correlated to the original p53 status. The remaining 50 cases showed residual tumor. The most prominent histologic change was an increase in nuclear atypia of tumor cells (51 % of the cases). This effect was independent of the presence or absence of nuclear p53 accumulation in the pre-treatment specimens. Nuclear atypia was frequently accompanied by tumor cell enlargement (in 49% of the cases). Most commonly, a tumor with relatively small cells presented with large epithelioid apocrine features after treatment. In 6 cases (13%), the mitotic rate decreased significantly, while in 12 cases (26%) the mitotic rate increased after chemotherapy. Elston histogrades remained unchanged in 70% of the cases but increased in 17% and decreased in 13%, mainly due to changes in mitotic rates. Extensive tumor cell vacuolization, a common change seen after radiotherapy, was a minor finding but was seen focally. Within the non-malignant compartment, lobular atrophy with hyalinization and minimal epithelial atypia of lobules and ducts were common. We conclude that changes in residual tumor Pathol. Res. Pract. 193: 187-196 (1997)
and normal breast are common following systemic cytotoxic therapy. As neoadjuvant chemotherapy becomes mainstream management for locally advanced breast cancer, pathologists are required to recognize treatment induced changes. For correct histopathologic assessment, therapy induced morphologic alterations need to be distinguished from tumor-intrinsic morphologic features.
Key words: Chemotherapy - Breast cancer - Neoadjuvant therapy - Histopathology
Introduction
Over the last few years a rapidly increasing trend has been seen to treat locally advanced breast cancer with primary (neoadjuvant) chemotherapy, typically followed by tumor resection and/or radiotherapy [5]. Historically, this treatment started to develop in the 1970's on the background of dismal results that had previously been obtained with local and regional therapies [8, 9, 14, 15]. The neoadjuvant protocol has a strong theoretical rationale: shrinking the tumor mass, decreasing the possibility of facilitated dissemination of metastatic cells during surgical manipulation and starting systemic treatment immediately when micrometastasis are at their smallest. Compared to post-operative chemotherapy, the demonstrated advantages of neoadjuvant chemotherapy are the following: 1. In a randomized trial of 272 stage III patients, overall survival was better when the same chemotherapeutic regimen was given preoperatively rather than postoperatively [21 J. 2. A very high objective response rate after induction is achieved, resulting in substantial mass reduction of the primary tumor and tumor bearing regional lymph 0344-0338/97/0193-0187$5.0010
188 . U. M. Moll and 1. Chumas
nodes. This leads to "down-staging" and a lower TNMscore in 60-70% of all patients after chemotherapy. In practice, this means that unresectable tumors such as stage III B or inflammatory breast cancers become resectable [5, 9, 14, 17,25,26]. 3. Aside from the gained operability, some patients become candidates for breast conservation therapy [10, 14, 18,20,21]. Complete pathologic remissions, i.e. no detectable tumor cells in the resection specimen after extensive sampling, however, are only seen in 10-22% of the patients [12, 13, 16,28]. Variations of the neoadjuvant approach exist. Some centers employ neoadjuvant chemotherapy followed by radiotherapy (external and interstitial), followed again by adjuvant chemotherapy [18, 20, 31]. In this regimen, surgical resection is reserved for local failures only. Careful long-term follow-up of these variations are still forthcoming and will determine the optimal modality and sequence of its use. Naturally, the emphasis on virtually all neoadjuvant studies so far has been on clinical response and outcome. However, the drug-induced morphologic changes that may occur in the neoplastic as well as in the non-neoplastic tissue compartment are of considerable importance. For correct histopathologic assessment, pathologists need to be able to account for therapy induced morphologic alterations and distinguish them from tumor-intrinsic morphologic features. The goal of this study was to help further define these druginduced histologic changes. Material and Methods Tissue: Sixty-one matched pairs of breast cancer tissue came from patients with stage III disease who presented since 1989 to the Institut Gustave Roussy at Villejuif, France. They included premenopausal and postmenopausal women. Patients were selected for the following criteria: less than 70 years of age, clinical size of tumor >3 cm in largest diameter but absence of distant metastasis, no inflammatory type tumor and no previous cancer treatment. All but 2 tumors were infiltrating ductal carcinomas; the remainders were of the infiltrative lobular type. Before treatment, tru-cut needle biopsies were obtained, embedded in paraffin and used for pathologic diagnosis and immunocytochemistry. Each patient then received one monthly cycle of doxorubicin (50 mg/m 2 day I), vincristin (1 mg/m 2 - day 1), cyclophosphamide (200 mg/m 2 - day 2-4), methotrexate (10 mg/m 2 - day 2-4) and 5-fluorouracil (300 mg/m 2 - day 2-4) for 3 consecutive months. Four weeks later, patients underwent surgical resection of the residual tumor consisting of mastectomy (>3 cm) or lumpectomy «3 cm) depending on tumor size. Patients had not received radiotherapy prior to mastectomy. The first 32 of this 61 patient cohort were previously analyzed by us for drug induced changes of the p53 immunophenotype [23}. Morphologic evaluation: One representative H&E slide each of all pre- and posttreatment tissue pairs was available
for our study. Slides were compared by first evaluating the pre-treatment biopsy followed by the resection specimen. In evaluating histologic features of the neoplastic element, carcinomas were assessed using the WHO classification [2}. Histologic grade was assigned by the Elston system after evaluating nuclear grade (pleomorphism), mitotic rate and tubule formation {l0}. For nuclear grade, tumors composed of nuclei with minimal variation in size and shape were given 1 point, moderate pleomorphism was given 2 points and marked variation (including nuclear enlargement, vacuolization, hyperchromatism, bizarre nuclei and multinucleation) were given 3 points. For mitotic rates (MI), 10 consecutive high power fields (HPF) (400x) from each pretreatment sample were counted whenever possible; in cases of less than 10 HPF, the number of counted mitosis was extrapolated to 10 fields. For the posttreatment sample, 10 HPF from a central zone of the tumor, more closely simulating the origin of the needle biopsy, was counted. Tumors with an MI of 0-5 were given I point, those with an MI of 6-1 0 were given 2 points and those with an MI > 10 were given 3 points. For tubule formation, tumors with >75% of the invasive component composed of tubules were given 1 point, 10-75% were given 2 points and <10% were given 3 points. An overall Elston score was assigned by adding points from the individual components. P53 immunocytochemistry: Thirty-two of the 61 tissue pairs were previously analyzed by us for p53-mediated tumor cell response [23}. The remaining pairs were stained exactly like we did previously [23}. Briefly, 4-5 /lm sections were dewaxed and microwave processed in 10 mM citric acid buffer, pH 6 for 2x5 min before blocking with 10% normal serum. Monoclonal antibodies DO-I (Santa Cruz Biotechnologies), DO-7 (1: 100) (Vector) and polyclonal CM-I (Vector) diluted in 2% BSA/PBS were applied overnight at 4 0c. After washing, sections were incubated in biotinylated secondary IgG followed by streptavidin-coupled horseradish-peroxidase/ DAB (Zymed, San Francisco) and alkaline phosphatase/Fast Red (Biogenex, San Ramon, CA) for detection. In all runs, color development was strictly standardized for time and substrate concentration in order to allow a meaningful assessment of staining intensity. Normal mouse IgG at 500 ng/m\ served as negative control. Statistics: The InStat program for Macintosh was used to calculate P-values. P < 0.05 was considered significant.
Results Histologicfindings in the neoplastic breast
Eleven cases (18%) showed complete pathological remission with no residual tumor after thorough sampling of the mastectomy specimen (Table 1). Complete remission was not correlated with the original p53 staTable I. Tumor response after chemotherapy Number of cases (%) Complete pathologic remission Residual tumor after therapy
11/6\ (18%) 50/61 (82%)
Chemotherapy Effect in Breast Tissue . 189 Table 2. Histological features of tumors before and after chemotherapy Histology Prechemotherapy
Histology Postchemotherapy
Number of Cases
Infiltrating ductal carcinoma Infiltrating ductal carcinoma Infiltrating ductal carcinoma Infiltrating ductal carcinoma, no DCIS Infiltrating ductal carcinoma + DCIS Infiltrating lobular carcinoma
no residual carcinoma Infiltrating ductal carcinoma Infiltrating ductal carcinoma + DCIS DCIS only Infiltrating ductal carcinoma + DCIS Infiltrating lobular carcinoma
11 32* 9** 2
3*** 2
2 cases had crush artifact in the core biopsy, precluding precise histological classification. * I case with extensive lymphatic invasion after chemotherapy; I case with no tubules in the biopsy but 75% tubule formation in the resection specimen. ** I case with mixed high grade and low grade DCIS, I case with intermediate grade DCIS and the remainder with high grade DCIS. *** DCIS was of high grade in all cases; I case exhibited no tubules in the core biopsy but 50% tubule formation in the resection specimen; I case showed focal mucinous differentiation in the resection specimen only.
Table 3. Histologic alterations in the neoplastic breast after neoadjuvant polychemotherapy* Higher Nuclear grade Mitotic rate Elston histograde
Lower
25/49 (51 %)** 1/49 (2%) 12/47 (26%) 6/47 (13%) 8/47 (17%) 6/47 (13%)
Prominent tumor cell 23/46 (49%) enlargement Vacuolated cytoplasm, focal 20/47 (43%) Histiocytic-type changes, focal 3/47 (6%) Apocrine differentiation 17/47 (36%) Change in inflammatory 17/47 (36%) response Change in growth pattern 8/47 (17%) * For each criterion, the corresponding pretreatment sample of each pair served as control. ** The mean nuclear grade increased by I point in the Elston scheme.
Table 4. Changes in nuclear grade following chemotherapy Pretreatment Nuclear Grade
Posttreatment Nuclear Grade
Number of Cases
I II III I
II III
3/25 11/25 10125 1125
IW III
25/49 (51 %) tumors had a higher nuclear grade after treatment. IW Bizarre nuclear forms, multinucleation and/or anaplasia was present to a significantly higher degree than before treatment.
tus as determined by immunocytochemistry, since 6 of the 11 cases were p53 positive, 4 were p53 negative and 1 case could not be determined. Likewise, no other histologic parameter could be identified that would predict complete remission. The remaining 50 cases with residual tumor all had some reduction of tumor mass, whose extent however was not quantitated clinically. Fifty-nine of the original 61 biopsy specimens showed infiltrating ductual carcinoma and 2 showed infiltrating lobular carcinoma (Table 2). The corresponding posttreatment histologies are listed in Table 2. No change in histologic type was found. The majority of the cases (46 patients) were of the infiltrating ductual type with or without DCIS. In 9 cases, DCIS was only present in the resection specimen, probably due to sampling error in the core biopsy. Interestingly, 2 cases that originally showed invasive carcinoma but no DCIS on biopsy, exhibited only DCIS after treatment. Whether this reflects drug-mediated killing of the invasive component is unclear with only 2 cases at hand. Several histological changes were consistently noted in the tumor cell population after chemotherapy. Generally, the histopathological appearance of the tumor worsened. The most prominent change was an increase in nuclear grade in 25 out of 49 (51 %) of the cases (Table 3). It included nuclear enlargement, prominent nucleoli, multinucleation, nuclear vacuolization and hyperchromatism (e.g. Fig. 1B). Originally, 34 tumors exhibited nuclear grade III, 18 tumors were grade II, 6 tumors were grade I and 3 cases could not be evaluated due to some crush artifact in the needle biopsy. After chemotherapy, the mean nuclear grade increased by 1 point in the Elston scheme (Table 4). Again, no correlation was found between pre-treatment p53 status and increase or decrease of nuclear atypia after treatment (Table 5). The median mitotic index for pre- and posttreatment samples was 1 (on a scale of 1-3). In 26% of
190 . U. M. Moll and J. Chumas
Table 5. Lack or correlation between abnormal p53 expression before therapy and nuclear atypia after therapy in 46 cases* Pre-treatment p53 status
Number of Cases with Post-treatment Nuclear Atypia Increase
p53 positive** p53 negative**
Decrease
No Change
15 0 10 8 I 12 P-value: not significant***
*
Out of 50 cases with residual post-treatment tumor, p53 expression could not be determined in 4 cases. ** p53 positive: abnormal expression of p53 in 2:25% of tumor cells by immunocytochemistry. ** p53 negative: no detectable p53 protein in tumor cells. *** P-value is 0.187 by Chi-square Test and 0.375 by Fisher's Exact Test.
Table 6. Changes in histograde and its correlation with associated mitotic rate and nuclear grade after treatment Increased Mitotic Rate
Increased Nuclear Grade
6 cases
5 cases
Decreased Elston Grade
Decreased Mitotic Rate
Decreased Nuclear Grade
6 cases
5 cases
I case
Increased Elston Grade 8 cases
~
~
the cases the increased atypia was accompanied by a higher mitotic index. In those cases, the rate typically increased by 5-7 mitoses per 10 HPF but 3 cases showed a dramatic rise, Le., from 11 to 32 mitoses per 10 HPF in case 15, from 3 to 22 mitoses in case 41 and from 4 to 25 mitoses in case 50. Incidentally, 2 of the 3 tumors exhibited nuclear p53 overexpression before treatment. Overall, the Elston histograde worsened in 8 out of 47 tumors (17%) and improved in 6 out of 47 tumors (13%) (Table 3). Thirty-three out of 47 tumors (70%) retained their Elston score. The increase or decrease in the Elston histograde was mainly due to corresponding changes in the mitotic index rather than in nuclear grade (Table 6). Interestingly, only 1 in 49 tumors (2%) exhibited a lower nuclear grade (Table 3). Also, 13% of the cases showed lower mitotic indices that translated to a lower Elston score in 13% of the cases. In 2 tumors that previously showed no tubules in the needle biopsy, greater than 50% and 75% tubule formation, respectively, was seen after treatment. Again, some caution has to be exerted concerning this seemingly treatment-induced
tumor differentiation, since needle biopsies are prone to some sampling bias with respect to architectural features such as tubule formation in addition to cytologic variation and mitotic count. The nuclear changes were often concomitant with considerable tumor cell enlargement (49% of the cases) from relatively small tumor cells before treatment to large epithelioid-type cells after treatment (Table 3 and compare Figs. lC, 2A, 2B and Fig. 3). Extensive apocrine changes consisting of enlarged epithelioid tumor cells with abundant eosinophilic cytoplasm and prominent cell borders were common after treatment (36% of the cases) (Table 3 and Fig. IC and 2A). Focally, the apocrine change became so prominent as to resemble "squamoid differentiation" (Fig. 1C and 2A). Vacuolization was often found interspersed in single tumor cells (43% of the cases), but was only rarely seen as an extensive change as shown in Fig. lA. Histiocyte-like tumor cell vacuolization was present very focally in 6% of the tumors (Fig. 10). Marked changes in growth patterns were found in 17% of the tumors but were divergent in nature. Notably, one case changed from a diffuse infiltrative pattern to a subtle hyalinizing pattern (Fig. 2C), while the reverse was true for another case. A third tumor converted from a diffuse infiltrative pattern to a mixed pattern of subtle hyalinization and nested growth after treatment. Changes in the inflammatory response after therapy occurred in 36% of the cases and consisted of quantitative increases or decreases in the tumor-associated infiltrate. Typically, the infiltrate was predominantly lymphocytic with plasmacellular dominance in some cases. Drug-induced morphologic changes in malignant epithelial cells of ductal carcinoma in situ (DCIS) were the same as we had seen in invasive tumor cells. They consisted of prominent tumor cell enlargement with vacuolization as well as increased atypia. Histologic findings in the non-neoplastic breast
Although case controlled comparison was hindered by a scarcity of normal breast tissue in the pretreatment biopsies, several histologic changes were consistently noted in residual noncarcinomatous breast tissue within lobules, large and small ducts and the stromal compartment (Table 7). The most prevalent lobular change consisted of diffuse lobular atrophy (48% of the cases) accompanied by a spectrum of intralobular fibrosis varying from basement membrane thickening (39%) (Fig. 4A) to hyalizination (48%) with occasional lobular calcification (9%) and inflammatory lobular stromal infiltrates (4%). The diagnosis of lobular atrophy was based on the histology of individual lobules rather than on the reduced number of lobular units per low power field due to the limited size of the biopsy. Drug-induced mild epithelial atypia, either focal or widespread, was found in 30% of lobules (Fig. 4B) and
Chemotherapy Effect in Breast Tissue . 191
A
B
c
D Fig. I. Morphologic changes in the residual neoplastic compartment after chemotherapy. On the left are the pre-treatment specimens and on the right are the matching post-treatment specimens. A) After chemotherapy, residual tumor cells show extensive vacuolization as the prominent feature. Increased nuclear atypia is present as well. Case 4. B) A significantly worse histologic appearance is present after therapy including increased atypia, anaplasia, pleomorphism and giant cell formation. Case 20. C) Paradoxical "squamoid differentiation" is seen afterwards. Case 15. D) Anaplastic giant cell formation with focal histiocytic change (see arrow). Case 8. Original magnification is x400 for all.
192 . U. M. Moll and J. Chumas
Fig. 2. Morphologic changes in the residual neoplastic compartment after chemotherapy. On the left are the pre-treatment specimens and on the right are the matching post-treatment specimens. A) Enlarged tumor cells (49% of the cases), often with apocrine features (36% of the cases) were prominent findings. Slightly increased nuclear atypia is present as well. Case 19. B) "Glassy cell" change. Case 49. C) A diffusely infiltrating, florid tumor becomes hyalinized and is characterized by rare malignant cells subtlely invading dense fibrous stroma. Case 35. Original magnification is x400 for all.
45% of larger ducts. Typically, however, atypia was not seen in cases that had undergone widespread atrophy. Ductal epithelial vacuolization, sometimes associated with pigment-laden foamy macrophages in the lumen, could be seen in some tumors. Aside from fibrosis and a tumor-induced desmoplastic reaction which were virtually universal, the stromal compartment showed the lowest frequency of drug-induced changes (Table 7). Some cases showed elastosis and interstitial calcification. Three tumors exhibited a peculiar
radial central necrosis of the stroma associated with mild lymphocytic infiltration. These reactions closely resembled a nodular fasciitis-like scar. Three tumors showed no change within the non-neoplastic compartment.
Discussion Several studies have previously analyzed the histopathologic effects of neoadjuvant polychemotherapy on
Chemotherapy Effect in Breast Tissue . 193 Table 7. Histologic findings in non-neoplastic tissue Number of cases (%)*
Fig. 3. After therapy, only ductal hyperplasia was present in the residual tissue of case 27. Marked atypia is present in the luminal epithelial cell layer while the myoepithelial cells remain unaffected. Original magnification is x400.
Lobular atrophy/hyalinization Basement membrane thickening Lobular atypia Lobular calcification Lobular stromal infiltration
21/44 (48%) 17/44 (39%) 13/44 (30%) 4/44 (9%) 2/44 (4,5%)
Ductal atypia Basement membrane thickening Ductal cytoplasmic vacuolization Macrophages, pigmented and foamy Ductal calcification Ductal epithelial necrosis
20/44 (45%) 13/44 (30%) 9/44 (20%) 5/44 (11 %) 4/44 (9%) 1/44 (2%)
3/48 (6%) 2/48 (4%)
No change in any compartment
3/48 (6%)
*
A
B Fig. 4. Morphologic changes in the residual normal breast tissue after chemotherapy. A) Lobular atrophy with basement membrane thickening. Case 19. The changes resemble those after radiotherapy. B) Mild atypia in benign lobules. Case 11. (A, B) original magnification x400.
4/48 (8%) 3/48 (6%)
Elastosis Stromal nodular fasciitis-like reaction with central necrosis Stromal calcification Edema
Data is not controlled for corresponding pretreatment status of each pair (see Results); In cases where the denominator is less than expected, the criteria could not be evaluated in the remaining specimens.
residual malignant and normal breast tissue [1, 6, 7, 12, 13, 19, 24, 28, 30]. Our study identified several morphologic findings that are in agreement with previously described changes including nuclear and cytoplasmic changes in residual tumor cells and regressive changes in benign breast tissue. Our study encompasses a large homogeneous patient group with locally advanced disease. All 61 patients were under 70 years of age, they presented with stage III disease and harbored a tumor over 3 cm in size. All patients underwent identical neoadjuvant polychemotherapy consisting of three 5drug cycles, followed by surgical resection 3 months later. Drug-induced morphologic alterations were identified by comparing matched pairs of the initial needle biopsy specimens with the final mastectomy specimens. Eighteen percent of the patients in our cohort responded with complete pathologic remission. This figure is in agreement with the experience of others which report a rate of 10-22% [12, 13, 16,28]. We found the most common histologic effect in residual carcinoma to be an increase in nuclear grade (51 % of the cases) due to nuclear enlargement, vesiculation and hyperchromasia, followed by tumor cell enlargement (49%), apocrine features (36%) and altered inflammatory reaction (36%) (Tables 3 and 4). Most likely, these changes are transient and degenerative in nature, especially since all our posttreatment samples were removed 4 weeks after 3 monthly cycles of chemotherapy. Close et al. [7] similarly identified increased nuclear grade as the most
194 . U. M. Moll and J. Chumas
common change. Furthermore, in a cytopuncture study of 35 matched samples before and after chemotherapy, Brifford et al. [6] found increased nuclear atypia in 20 of 35 cases. It is not possible to ascribe a particular morphologic change to a specific drug, since all studies (including ours) involve a multidrug regimen of alkylating agents, topoisomerase inhibitors, spindel inhibitors etc, which, in addition, might have overlapping effects. One study using prolonged tamoxifenlcytotoxic drug combinations found a prominent histiocyte-like tumor cell appearance after treatment that caused some difficulty in distinguishing from benign cells [19]. In contrast, we saw such cells only very focally and only in a few cases (6%). Reasons for this discrepancy could lie in the higher number of cycles (on average 7) and/or the addition of tamoxifen in the Kennedy study. Some treatment-related effects presented diagnostic problems as shown in case 27. After therapy, this tumor showed multiple ducts with widespread intraductal proliferation. This included focally highly atypical cells within the luminal epithelial layer superimposed on an otherwise non-atypical epithelial proliferation (Fig. 3). No invasive element was present. Notably, though, every duct showed an intact circumferential myoepithelial cell layer without any atypia. These features support intraductal hyperplasia with superimposed chemotherapeutic effects rather than a diagnosis of DCIS. Likewise, nuclear atypia in benign lobules, occasionally with binucleation, was present in 30% of the post-treatment tumors, leading to patterns that mimicked intrinsic lobular atypia (Fig. 4B). Some of these cases might have presented diagnostic difficulties had they been evaluated without knowledge of prior cytotoxic drug exposure. Overall, the histopathologic appearance of the tumors changed significantly enough that we agree with the cautious view of Rasbridge et al. [28], who recommended against grading a tumor after treatment. The presence of p53 in breast cancer has been associated with a poor longterm prognosis [3]. However, we were unable to correlate the initial p53 status of the tumor with short term responses such as pathologic remission or drug-induced nuclear atypia, possibly due to the small number of cases studied. In an earlier study involving the first 32 of the 61 patients, we found a treatment-associated change of the p53 immunophenotype in 12/31 (39%) of the tumors [23]. This phenomenon most likely indicates clonal selection. Interestingly, wild type p53 tumors were less likely to undergo clonal selection (3/12) than mutant p53 tumors (9/12). This clinical study suggests that wild-type p53 exerts a "guardian of genome" function not only in normal cells but also in breast cancer cells in vivo. Widespread tumor cell death as it occurs in treatmentassociated complete pathologic remission could be mediated by a number of mechanisms including p53-dependent apoptosis, p53-independent apoptosis or non-
apoptotic mitotic arrest. It is currently unclear, if wildtype p53 mediated apoptosis (rather than G1 arrest) in response to chemotherapy occurs in breast cancer. Currently available studies on breast cell lines with functional wild type p53 report either resistance to apoptosis [11] or only limited apoptotic response after DNA damaging drugs [11, 22]. Technically, two potential pitfalls must be considered when comparing the morphology of breast carcinomas between a needle biopsy and a mastectomy specimen. First, the small biopsy could have an inherent sampling error with respect to architectural and cytological features and mitotic count [4]. Second, the slower fixation of the larger tissue slice from the mastectomy might produce cytological artifacts. Sampling errors, however, are statistically counteracted by analyzing large numbers of cases. Taken together, the Close study [7] and our study suggest that the observed increase in nuclear grade is unlikely due to sampling error. Concerning potential fixation artifacts, Close et al. [7] also looked at 10 needle biopsies and corresponding mastectomies from patients who did not receive any preoperative chemotherapy. Because these mastectomy specimens lacked any of the nuclear and cytoplasmic features seen in the treated specimens, the possibility of an artifact was excluded. Our experience concerning the morphology of untreated mastectomies is identical to that of Close et al. [7] and Sharkey et al. [30]. Largely driven by the newer concept that breast cancer is a systemic disease from the onset, the management of invasive disease has undergone major changes. Lately, a strong trend is seen towards neoadjuvant polychemotherapy as the first line of treatment in locally advanced breast cancer as opposed to postsurgical chemotherapy [27]. Several large studies start to show an overall survival benefit from neoadjuvant versus adjuvant therapy [21, 29]. It is conceivable that this treatment will also become a tool in early breast cancer. Therefore, it will be increasingly important for pathologists to recognize drug-induced histologic changes when diagnosing such breast cancer specimens. Acknowledgement: The authors thank Drs. Guy Riou and Marie-Christine Mathieu, Institut Gustave Roussy, 94805 Villejuif, France for making tissue available. This work was supported by an Emil C. Voll Breast Cancer Research Grant to UMM.
References I. Anderson EDC, Forrest APM, Hawkins RA et al. (1991) Primary systemic therapy for operable breast cancer. Br J Cancer 63,561-566 2. Azzopardi JG, Chepick OF, Hartmann WH et al. (1981) International Histological Classification of Tumours. No.
Chemotherapy Effect in Breast Tissue . 195
3.
4. 5.
6. 7. 8. 9.
10.
11.
12.
13.
14.
15.
16. 17.
2. Histological typing of breast tumours. Second Edition, World Health Organization, Geneva Barnes DB, Dublin EA, Fisher CF, Levison DA, Millis RR (1993) Immunohistochemical detection of p53 protein in mammary carcinoma: an important new independent indicator of prognosis? Hum Pathol24, 469--476 Biesterfeld S, Noll I, Noll E, Wohltmann D, Bocking A (1995) Mitotic frequency as a prognostic factor in breast cancer. Hum Pathol26, 47-52 Bonadonna G, Veronesi V, Brambilla C, Ferrari L, Luini A, Greco M, Bartoli C, Coopmans De Yolda G, Zucali R, Rilke F, Andreola C, Silvestrini R, Di Fronzo G, Valagussa P (1990) Primary chemotherapy to avoid mastectomy in tumors with diameters of three centimeters or more. Nat J Cancer Inst 82, 1539-1545 Brifford M, Spyratos F, Tubiana-Huhn M et al. (1989) Sequential cytopunctures during preoperative chemotherapy for primary breast cancer. Cancer 63,631-637 Close PM, Tiltman AJ, Hacking EA (1987) The effect of chemotherapy on the morphology of human breast carcinoma. S Afr Med J 72, 103-106 Davila E and Vogel CL (1984) Management of locally advanced breast cancer (stage III): a review. Int Adv Surg Oncol 7,297-327 DeLena M, Zucali R, Viganotti G, Valagussa P, Bonadonna G (1978) Combined chemotherapy-radiotherapy approach in locally advanced (T3 b-T4) breast cancer. Cancer Chemother Pharmacoll, 53-57 Elston CW, Ellis 10 (1991) Pathological prognostic factors in breast cancer. I. The value of histologic grade in breast cancer. Experience from a large study with longterm follow-up. Histopathology 19,403--410 Fan S, Smith ML, Rivet DJ 2nd, Duba D, Zhan Q, Kohn KW, Fornace AJ Jr, O'Connor PM (1995) Disruption of p53 function sensitizes breast cancer MCF-7 cells to cisplatin and pentoxyfylline. Cancer Res 55, 1649-1654 Frierson HF Jr, Fechner RE (1994) Histologic grade of locally advanced infiltrating ductal carcinoma after treatment with induction chemotherapy. Am J Clin Pathol 102, 154-157 Honkoop AH, Pinedo HM, De Jong JS, Verheul HMW, Linn SC, Hoekman K, Wagstaff J, Van Diest PJ (1997) Effects of chemotherapy on pathologic and biologic characteristics of locally advanced breast cancer. Am J Clin Pathol 107, 211-218 Hortobagyi GN, Spanos W, Montague ED, Buzdar AU, Yap H-Y, Blumenschein GR (1983) Treatment of locoregionally advanced breast cancer with surgery, radiotherapy, and combination chemoimmunotherapy. Int J Radiat Oncol BioI Phys 9, 643-650 Hortobagyi GN, Buzdar AU (1991) Locally advanced breast cancer: a review including M.D. Anderson experience. In: J Ragaz, 1M Ariel (Eds) High Risk Breast Cancer, pp 382--415. Springer Verlag, Berlin Hortobagyi GN, Buzdar AU, Strom EA, Ames FC; Singletary SE (1995) Primary chemotherapy for early and advanced breast cancer. Cancer Letters 90, 103-109 Israel L, Breau J-L, Morene JF et al. (1986) Two years of high dose cyclophosphamide and 5-fluoruracil followed by surgery after 3 months for acute inflammatory breast carcinoma. Cancer 57, 24-28
18. Jacquilat C, Baillet F, Weil M, Auclerc G, Housset M, Auclerc MF, Sellami M, Jindani A, Thill I, Soubrane C, Khayat D (1988) Results of conservative treatment combining induction (neoadjuvant) and consolidation chemotherapy, hormonotherapy, and external and interstitial irradiation in 98 patients with locally advanced breast cancer (IIIA-IIIB). Cancer 61, 1977-1982 19. Kennedy S, Merino MJ, Swain SM, Lippman ME (1990) The effects of hormonal and chemotherapy on tumoral and nonneoplastic breast tissue. Hum Pathol 21, 192-198 20. Lippman ME, Sorace RA, Bagley CS, Danforth DW, Lichter A, Wesley MN (1986) Treatment of locally advanced breast cancer using primary induction chemotherapy with hormonal synchronization followed by radiation therapy with or without debulking surgery. NCI Monogr 1, 153-159 21. Mauriac L, Durand M, Avril A, Dilhuydy J-M (1991) Effects of primary chemotherapy in conservative treatment of breast cancer patients with operable tumors larger than 3 em: results of a randomized trial in a single center. Ann OncoI2,347-354 22. Merlo GR, Basolo F, Fiore L, Duboc L, Hynes NE (1995) P53-dependent and p53-independent activation of apoptosis in mammary epithelial cells reveals a survival function of EGF and insulin. J Cell Bioi 128, 1185-1196 23. Moll UM, Ostermeyer AG, Ahomadegbe J-C, Mathieu M-C, Riou G (1995) P53-mediated tumor cell response to chemotherapeutic DNA damage: a preliminary study in matched pairs of breast cancer biopsies. Human Pathology 26, 1293-1301 24. Morrow M, Braverman A, Thelmo W et al. (1986) Multimodal therapy for locally advanced breast cancer. Arch Surg 121,1291-1296 25. Olson JE, Gray R, Sponzo RW, Damsker 11, Tormey DC, Cummings FJ (1986) Management of nonresectable locally advanced (Stage III) breast cancer: an ECOG trial. Breast Cancer Res Treat 8 (Abst. 124): 109 26. Perloff M, Lesnick GJ, Korzun A, Chu F, Holland JF, Thirlwell MP, Ellison RR, Carey RW, Leone L, Weinberg V, Rice MA, Wood WC (1988) Combination chemotherapy with mastectomy or radiotherapy for stage III breast carcinoma: A cancer and leukemia group B study. J Clin Oncol 6, 261-269 27. Powles TJ, Hikish TF, Makris A, Ashley SE, O'Brien ME, Tidy VA, Casey S et al. (1995) Randomized trial of chemoendocrine therapy started before or after surgery for treatment of primary breast cancer. J Clin Oncol 13, 547-552 28. Rasbridge SA, Gillett CE, Seymour AM, Patel K, Richards MA, Rubens RD, Millis RR (1994) The effects of chemotherapy on morphology, cellular proliferation, apoptosis and oncoprotein expression in primary breast carcinoma. British J Cancer 70, 335-341 29. Scholl SM, Fourquet A, Asselain B, Pierga JY, Vilcoq JR, Durand JC, Dorval T, Palangie T, Jouve M, Beuzeboc Pet al. (1994) Neoadjuvant versus adjuvant chemotherapy in premenopausal patients with tumors considered too large for breast conserving surgery: preliminary results of a randomized trial: S6. Eur J Cancer 30A (5), 645 30. Sharkey FE, Addington SL, Fowler LJ, Page CP, Cruz AB (1996) Effects of preoperative chemotherapy on the mor-
196 . U. M. Moll and J. Chumas phology of resectable breast carcinoma. Mod Pathol 9, 893-900 31. Williams MR, Gilson D, Marsh L, Morgan DAL, Nicholson RL, Elston CW, Griffiths K, Blarney RW (1988) The early results from a randomised study of radiotherapy versus Nolvadex (Tamoxiden) as initial treatment for stage III breast cancer. Eur J Surg Oncol 14,235-240
Received: October 21, 1996 Accepted in revised form: April I, 1997 Address for correspondence: Dr. U. M. Moll, Department of Pathology, SUNY Stony Brook Medical School, Stony Brook NY N.Y, 11794-8691, USA. Tel.: (516) 444 2459; e-mail: [email protected]
Morphologic Effects of Neoadjuvant Chemotherapy in Locally Advanced Breast Cancer U. M. Moll and J. Chumas Department of Pathology, State University of New York at Stony Brook, Stony Brook, New York
Summary
The use of chemotherapy as primary treatment in early and locally advanced breast cancer is rising. As a result, many resected tumors were exposed to cytotoxic drugs in vivo. To study resulting histopathologic changes, we examined 61 patients with locally advanced stage III breast cancer who had been treated with a standardized neoadjuvant polychemotherapy regimen before undergoing surgical resection 3 months later. Matched pairs of pre- and posttherapy breast tissue were evaluated for morphologic changes in the residual malignant and benign breast tissue compartment. A potential correlation between changes and the original p53 immunophenotype was examined as well. In 11 cases (18%), complete pathologic remission with no residual tumor in the mastectomy specimen was achieved. This response was not correlated to the original p53 status. The remaining 50 cases showed residual tumor. The most prominent histologic change was an increase in nuclear atypia of tumor cells (51 % of the cases). This effect was independent of the presence or absence of nuclear p53 accumulation in the pre-treatment specimens. Nuclear atypia was frequently accompanied by tumor cell enlargement (in 49% of the cases). Most commonly, a tumor with relatively small cells presented with large epithelioid apocrine features after treatment. In 6 cases (13%), the mitotic rate decreased significantly, while in 12 cases (26%) the mitotic rate increased after chemotherapy. Elston histogrades remained unchanged in 70% of the cases but increased in 17% and decreased in 13%, mainly due to changes in mitotic rates. Extensive tumor cell vacuolization, a common change seen after radiotherapy, was a minor finding but was seen focally. Within the non-malignant compartment, lobular atrophy with hyalinization and minimal epithelial atypia of lobules and ducts were common. We conclude that changes in residual tumor Pathol. Res. Pract. 193: 187-196 (1997)
and normal breast are common following systemic cytotoxic therapy. As neoadjuvant chemotherapy becomes mainstream management for locally advanced breast cancer, pathologists are required to recognize treatment induced changes. For correct histopathologic assessment, therapy induced morphologic alterations need to be distinguished from tumor-intrinsic morphologic features.
Key words: Chemotherapy - Breast cancer - Neoadjuvant therapy - Histopathology
Introduction
Over the last few years a rapidly increasing trend has been seen to treat locally advanced breast cancer with primary (neoadjuvant) chemotherapy, typically followed by tumor resection and/or radiotherapy [5]. Historically, this treatment started to develop in the 1970's on the background of dismal results that had previously been obtained with local and regional therapies [8, 9, 14, 15]. The neoadjuvant protocol has a strong theoretical rationale: shrinking the tumor mass, decreasing the possibility of facilitated dissemination of metastatic cells during surgical manipulation and starting systemic treatment immediately when micrometastasis are at their smallest. Compared to post-operative chemotherapy, the demonstrated advantages of neoadjuvant chemotherapy are the following: 1. In a randomized trial of 272 stage III patients, overall survival was better when the same chemotherapeutic regimen was given preoperatively rather than postoperatively [21 J. 2. A very high objective response rate after induction is achieved, resulting in substantial mass reduction of the primary tumor and tumor bearing regional lymph 0344-0338/97/0193-0187$5.0010
188 . U. M. Moll and 1. Chumas
nodes. This leads to "down-staging" and a lower TNMscore in 60-70% of all patients after chemotherapy. In practice, this means that unresectable tumors such as stage III B or inflammatory breast cancers become resectable [5, 9, 14, 17,25,26]. 3. Aside from the gained operability, some patients become candidates for breast conservation therapy [10, 14, 18,20,21]. Complete pathologic remissions, i.e. no detectable tumor cells in the resection specimen after extensive sampling, however, are only seen in 10-22% of the patients [12, 13, 16,28]. Variations of the neoadjuvant approach exist. Some centers employ neoadjuvant chemotherapy followed by radiotherapy (external and interstitial), followed again by adjuvant chemotherapy [18, 20, 31]. In this regimen, surgical resection is reserved for local failures only. Careful long-term follow-up of these variations are still forthcoming and will determine the optimal modality and sequence of its use. Naturally, the emphasis on virtually all neoadjuvant studies so far has been on clinical response and outcome. However, the drug-induced morphologic changes that may occur in the neoplastic as well as in the non-neoplastic tissue compartment are of considerable importance. For correct histopathologic assessment, pathologists need to be able to account for therapy induced morphologic alterations and distinguish them from tumor-intrinsic morphologic features. The goal of this study was to help further define these druginduced histologic changes. Material and Methods Tissue: Sixty-one matched pairs of breast cancer tissue came from patients with stage III disease who presented since 1989 to the Institut Gustave Roussy at Villejuif, France. They included premenopausal and postmenopausal women. Patients were selected for the following criteria: less than 70 years of age, clinical size of tumor >3 cm in largest diameter but absence of distant metastasis, no inflammatory type tumor and no previous cancer treatment. All but 2 tumors were infiltrating ductal carcinomas; the remainders were of the infiltrative lobular type. Before treatment, tru-cut needle biopsies were obtained, embedded in paraffin and used for pathologic diagnosis and immunocytochemistry. Each patient then received one monthly cycle of doxorubicin (50 mg/m 2 day I), vincristin (1 mg/m 2 - day 1), cyclophosphamide (200 mg/m 2 - day 2-4), methotrexate (10 mg/m 2 - day 2-4) and 5-fluorouracil (300 mg/m 2 - day 2-4) for 3 consecutive months. Four weeks later, patients underwent surgical resection of the residual tumor consisting of mastectomy (>3 cm) or lumpectomy «3 cm) depending on tumor size. Patients had not received radiotherapy prior to mastectomy. The first 32 of this 61 patient cohort were previously analyzed by us for drug induced changes of the p53 immunophenotype [23}. Morphologic evaluation: One representative H&E slide each of all pre- and posttreatment tissue pairs was available
for our study. Slides were compared by first evaluating the pre-treatment biopsy followed by the resection specimen. In evaluating histologic features of the neoplastic element, carcinomas were assessed using the WHO classification [2}. Histologic grade was assigned by the Elston system after evaluating nuclear grade (pleomorphism), mitotic rate and tubule formation {l0}. For nuclear grade, tumors composed of nuclei with minimal variation in size and shape were given 1 point, moderate pleomorphism was given 2 points and marked variation (including nuclear enlargement, vacuolization, hyperchromatism, bizarre nuclei and multinucleation) were given 3 points. For mitotic rates (MI), 10 consecutive high power fields (HPF) (400x) from each pretreatment sample were counted whenever possible; in cases of less than 10 HPF, the number of counted mitosis was extrapolated to 10 fields. For the posttreatment sample, 10 HPF from a central zone of the tumor, more closely simulating the origin of the needle biopsy, was counted. Tumors with an MI of 0-5 were given I point, those with an MI of 6-1 0 were given 2 points and those with an MI > 10 were given 3 points. For tubule formation, tumors with >75% of the invasive component composed of tubules were given 1 point, 10-75% were given 2 points and <10% were given 3 points. An overall Elston score was assigned by adding points from the individual components. P53 immunocytochemistry: Thirty-two of the 61 tissue pairs were previously analyzed by us for p53-mediated tumor cell response [23}. The remaining pairs were stained exactly like we did previously [23}. Briefly, 4-5 /lm sections were dewaxed and microwave processed in 10 mM citric acid buffer, pH 6 for 2x5 min before blocking with 10% normal serum. Monoclonal antibodies DO-I (Santa Cruz Biotechnologies), DO-7 (1: 100) (Vector) and polyclonal CM-I (Vector) diluted in 2% BSA/PBS were applied overnight at 4 0c. After washing, sections were incubated in biotinylated secondary IgG followed by streptavidin-coupled horseradish-peroxidase/ DAB (Zymed, San Francisco) and alkaline phosphatase/Fast Red (Biogenex, San Ramon, CA) for detection. In all runs, color development was strictly standardized for time and substrate concentration in order to allow a meaningful assessment of staining intensity. Normal mouse IgG at 500 ng/m\ served as negative control. Statistics: The InStat program for Macintosh was used to calculate P-values. P < 0.05 was considered significant.
Results Histologicfindings in the neoplastic breast
Eleven cases (18%) showed complete pathological remission with no residual tumor after thorough sampling of the mastectomy specimen (Table 1). Complete remission was not correlated with the original p53 staTable I. Tumor response after chemotherapy Number of cases (%) Complete pathologic remission Residual tumor after therapy
11/6\ (18%) 50/61 (82%)
Chemotherapy Effect in Breast Tissue . 189 Table 2. Histological features of tumors before and after chemotherapy Histology Prechemotherapy
Histology Postchemotherapy
Number of Cases
Infiltrating ductal carcinoma Infiltrating ductal carcinoma Infiltrating ductal carcinoma Infiltrating ductal carcinoma, no DCIS Infiltrating ductal carcinoma + DCIS Infiltrating lobular carcinoma
no residual carcinoma Infiltrating ductal carcinoma Infiltrating ductal carcinoma + DCIS DCIS only Infiltrating ductal carcinoma + DCIS Infiltrating lobular carcinoma
11 32* 9** 2
3*** 2
2 cases had crush artifact in the core biopsy, precluding precise histological classification. * I case with extensive lymphatic invasion after chemotherapy; I case with no tubules in the biopsy but 75% tubule formation in the resection specimen. ** I case with mixed high grade and low grade DCIS, I case with intermediate grade DCIS and the remainder with high grade DCIS. *** DCIS was of high grade in all cases; I case exhibited no tubules in the core biopsy but 50% tubule formation in the resection specimen; I case showed focal mucinous differentiation in the resection specimen only.
Table 3. Histologic alterations in the neoplastic breast after neoadjuvant polychemotherapy* Higher Nuclear grade Mitotic rate Elston histograde
Lower
25/49 (51 %)** 1/49 (2%) 12/47 (26%) 6/47 (13%) 8/47 (17%) 6/47 (13%)
Prominent tumor cell 23/46 (49%) enlargement Vacuolated cytoplasm, focal 20/47 (43%) Histiocytic-type changes, focal 3/47 (6%) Apocrine differentiation 17/47 (36%) Change in inflammatory 17/47 (36%) response Change in growth pattern 8/47 (17%) * For each criterion, the corresponding pretreatment sample of each pair served as control. ** The mean nuclear grade increased by I point in the Elston scheme.
Table 4. Changes in nuclear grade following chemotherapy Pretreatment Nuclear Grade
Posttreatment Nuclear Grade
Number of Cases
I II III I
II III
3/25 11/25 10125 1125
IW III
25/49 (51 %) tumors had a higher nuclear grade after treatment. IW Bizarre nuclear forms, multinucleation and/or anaplasia was present to a significantly higher degree than before treatment.
tus as determined by immunocytochemistry, since 6 of the 11 cases were p53 positive, 4 were p53 negative and 1 case could not be determined. Likewise, no other histologic parameter could be identified that would predict complete remission. The remaining 50 cases with residual tumor all had some reduction of tumor mass, whose extent however was not quantitated clinically. Fifty-nine of the original 61 biopsy specimens showed infiltrating ductual carcinoma and 2 showed infiltrating lobular carcinoma (Table 2). The corresponding posttreatment histologies are listed in Table 2. No change in histologic type was found. The majority of the cases (46 patients) were of the infiltrating ductual type with or without DCIS. In 9 cases, DCIS was only present in the resection specimen, probably due to sampling error in the core biopsy. Interestingly, 2 cases that originally showed invasive carcinoma but no DCIS on biopsy, exhibited only DCIS after treatment. Whether this reflects drug-mediated killing of the invasive component is unclear with only 2 cases at hand. Several histological changes were consistently noted in the tumor cell population after chemotherapy. Generally, the histopathological appearance of the tumor worsened. The most prominent change was an increase in nuclear grade in 25 out of 49 (51 %) of the cases (Table 3). It included nuclear enlargement, prominent nucleoli, multinucleation, nuclear vacuolization and hyperchromatism (e.g. Fig. 1B). Originally, 34 tumors exhibited nuclear grade III, 18 tumors were grade II, 6 tumors were grade I and 3 cases could not be evaluated due to some crush artifact in the needle biopsy. After chemotherapy, the mean nuclear grade increased by 1 point in the Elston scheme (Table 4). Again, no correlation was found between pre-treatment p53 status and increase or decrease of nuclear atypia after treatment (Table 5). The median mitotic index for pre- and posttreatment samples was 1 (on a scale of 1-3). In 26% of
190 . U. M. Moll and J. Chumas
Table 5. Lack or correlation between abnormal p53 expression before therapy and nuclear atypia after therapy in 46 cases* Pre-treatment p53 status
Number of Cases with Post-treatment Nuclear Atypia Increase
p53 positive** p53 negative**
Decrease
No Change
15 0 10 8 I 12 P-value: not significant***
*
Out of 50 cases with residual post-treatment tumor, p53 expression could not be determined in 4 cases. ** p53 positive: abnormal expression of p53 in 2:25% of tumor cells by immunocytochemistry. ** p53 negative: no detectable p53 protein in tumor cells. *** P-value is 0.187 by Chi-square Test and 0.375 by Fisher's Exact Test.
Table 6. Changes in histograde and its correlation with associated mitotic rate and nuclear grade after treatment Increased Mitotic Rate
Increased Nuclear Grade
6 cases
5 cases
Decreased Elston Grade
Decreased Mitotic Rate
Decreased Nuclear Grade
6 cases
5 cases
I case
Increased Elston Grade 8 cases
~
~
the cases the increased atypia was accompanied by a higher mitotic index. In those cases, the rate typically increased by 5-7 mitoses per 10 HPF but 3 cases showed a dramatic rise, Le., from 11 to 32 mitoses per 10 HPF in case 15, from 3 to 22 mitoses in case 41 and from 4 to 25 mitoses in case 50. Incidentally, 2 of the 3 tumors exhibited nuclear p53 overexpression before treatment. Overall, the Elston histograde worsened in 8 out of 47 tumors (17%) and improved in 6 out of 47 tumors (13%) (Table 3). Thirty-three out of 47 tumors (70%) retained their Elston score. The increase or decrease in the Elston histograde was mainly due to corresponding changes in the mitotic index rather than in nuclear grade (Table 6). Interestingly, only 1 in 49 tumors (2%) exhibited a lower nuclear grade (Table 3). Also, 13% of the cases showed lower mitotic indices that translated to a lower Elston score in 13% of the cases. In 2 tumors that previously showed no tubules in the needle biopsy, greater than 50% and 75% tubule formation, respectively, was seen after treatment. Again, some caution has to be exerted concerning this seemingly treatment-induced
tumor differentiation, since needle biopsies are prone to some sampling bias with respect to architectural features such as tubule formation in addition to cytologic variation and mitotic count. The nuclear changes were often concomitant with considerable tumor cell enlargement (49% of the cases) from relatively small tumor cells before treatment to large epithelioid-type cells after treatment (Table 3 and compare Figs. lC, 2A, 2B and Fig. 3). Extensive apocrine changes consisting of enlarged epithelioid tumor cells with abundant eosinophilic cytoplasm and prominent cell borders were common after treatment (36% of the cases) (Table 3 and Fig. IC and 2A). Focally, the apocrine change became so prominent as to resemble "squamoid differentiation" (Fig. 1C and 2A). Vacuolization was often found interspersed in single tumor cells (43% of the cases), but was only rarely seen as an extensive change as shown in Fig. lA. Histiocyte-like tumor cell vacuolization was present very focally in 6% of the tumors (Fig. 10). Marked changes in growth patterns were found in 17% of the tumors but were divergent in nature. Notably, one case changed from a diffuse infiltrative pattern to a subtle hyalinizing pattern (Fig. 2C), while the reverse was true for another case. A third tumor converted from a diffuse infiltrative pattern to a mixed pattern of subtle hyalinization and nested growth after treatment. Changes in the inflammatory response after therapy occurred in 36% of the cases and consisted of quantitative increases or decreases in the tumor-associated infiltrate. Typically, the infiltrate was predominantly lymphocytic with plasmacellular dominance in some cases. Drug-induced morphologic changes in malignant epithelial cells of ductal carcinoma in situ (DCIS) were the same as we had seen in invasive tumor cells. They consisted of prominent tumor cell enlargement with vacuolization as well as increased atypia. Histologic findings in the non-neoplastic breast
Although case controlled comparison was hindered by a scarcity of normal breast tissue in the pretreatment biopsies, several histologic changes were consistently noted in residual noncarcinomatous breast tissue within lobules, large and small ducts and the stromal compartment (Table 7). The most prevalent lobular change consisted of diffuse lobular atrophy (48% of the cases) accompanied by a spectrum of intralobular fibrosis varying from basement membrane thickening (39%) (Fig. 4A) to hyalizination (48%) with occasional lobular calcification (9%) and inflammatory lobular stromal infiltrates (4%). The diagnosis of lobular atrophy was based on the histology of individual lobules rather than on the reduced number of lobular units per low power field due to the limited size of the biopsy. Drug-induced mild epithelial atypia, either focal or widespread, was found in 30% of lobules (Fig. 4B) and
Chemotherapy Effect in Breast Tissue . 191
A
B
c
D Fig. I. Morphologic changes in the residual neoplastic compartment after chemotherapy. On the left are the pre-treatment specimens and on the right are the matching post-treatment specimens. A) After chemotherapy, residual tumor cells show extensive vacuolization as the prominent feature. Increased nuclear atypia is present as well. Case 4. B) A significantly worse histologic appearance is present after therapy including increased atypia, anaplasia, pleomorphism and giant cell formation. Case 20. C) Paradoxical "squamoid differentiation" is seen afterwards. Case 15. D) Anaplastic giant cell formation with focal histiocytic change (see arrow). Case 8. Original magnification is x400 for all.
192 . U. M. Moll and J. Chumas
Fig. 2. Morphologic changes in the residual neoplastic compartment after chemotherapy. On the left are the pre-treatment specimens and on the right are the matching post-treatment specimens. A) Enlarged tumor cells (49% of the cases), often with apocrine features (36% of the cases) were prominent findings. Slightly increased nuclear atypia is present as well. Case 19. B) "Glassy cell" change. Case 49. C) A diffusely infiltrating, florid tumor becomes hyalinized and is characterized by rare malignant cells subtlely invading dense fibrous stroma. Case 35. Original magnification is x400 for all.
45% of larger ducts. Typically, however, atypia was not seen in cases that had undergone widespread atrophy. Ductal epithelial vacuolization, sometimes associated with pigment-laden foamy macrophages in the lumen, could be seen in some tumors. Aside from fibrosis and a tumor-induced desmoplastic reaction which were virtually universal, the stromal compartment showed the lowest frequency of drug-induced changes (Table 7). Some cases showed elastosis and interstitial calcification. Three tumors exhibited a peculiar
radial central necrosis of the stroma associated with mild lymphocytic infiltration. These reactions closely resembled a nodular fasciitis-like scar. Three tumors showed no change within the non-neoplastic compartment.
Discussion Several studies have previously analyzed the histopathologic effects of neoadjuvant polychemotherapy on
Chemotherapy Effect in Breast Tissue . 193 Table 7. Histologic findings in non-neoplastic tissue Number of cases (%)*
Fig. 3. After therapy, only ductal hyperplasia was present in the residual tissue of case 27. Marked atypia is present in the luminal epithelial cell layer while the myoepithelial cells remain unaffected. Original magnification is x400.
Lobular atrophy/hyalinization Basement membrane thickening Lobular atypia Lobular calcification Lobular stromal infiltration
21/44 (48%) 17/44 (39%) 13/44 (30%) 4/44 (9%) 2/44 (4,5%)
Ductal atypia Basement membrane thickening Ductal cytoplasmic vacuolization Macrophages, pigmented and foamy Ductal calcification Ductal epithelial necrosis
20/44 (45%) 13/44 (30%) 9/44 (20%) 5/44 (11 %) 4/44 (9%) 1/44 (2%)
3/48 (6%) 2/48 (4%)
No change in any compartment
3/48 (6%)
*
A
B Fig. 4. Morphologic changes in the residual normal breast tissue after chemotherapy. A) Lobular atrophy with basement membrane thickening. Case 19. The changes resemble those after radiotherapy. B) Mild atypia in benign lobules. Case 11. (A, B) original magnification x400.
4/48 (8%) 3/48 (6%)
Elastosis Stromal nodular fasciitis-like reaction with central necrosis Stromal calcification Edema
Data is not controlled for corresponding pretreatment status of each pair (see Results); In cases where the denominator is less than expected, the criteria could not be evaluated in the remaining specimens.
residual malignant and normal breast tissue [1, 6, 7, 12, 13, 19, 24, 28, 30]. Our study identified several morphologic findings that are in agreement with previously described changes including nuclear and cytoplasmic changes in residual tumor cells and regressive changes in benign breast tissue. Our study encompasses a large homogeneous patient group with locally advanced disease. All 61 patients were under 70 years of age, they presented with stage III disease and harbored a tumor over 3 cm in size. All patients underwent identical neoadjuvant polychemotherapy consisting of three 5drug cycles, followed by surgical resection 3 months later. Drug-induced morphologic alterations were identified by comparing matched pairs of the initial needle biopsy specimens with the final mastectomy specimens. Eighteen percent of the patients in our cohort responded with complete pathologic remission. This figure is in agreement with the experience of others which report a rate of 10-22% [12, 13, 16,28]. We found the most common histologic effect in residual carcinoma to be an increase in nuclear grade (51 % of the cases) due to nuclear enlargement, vesiculation and hyperchromasia, followed by tumor cell enlargement (49%), apocrine features (36%) and altered inflammatory reaction (36%) (Tables 3 and 4). Most likely, these changes are transient and degenerative in nature, especially since all our posttreatment samples were removed 4 weeks after 3 monthly cycles of chemotherapy. Close et al. [7] similarly identified increased nuclear grade as the most
194 . U. M. Moll and J. Chumas
common change. Furthermore, in a cytopuncture study of 35 matched samples before and after chemotherapy, Brifford et al. [6] found increased nuclear atypia in 20 of 35 cases. It is not possible to ascribe a particular morphologic change to a specific drug, since all studies (including ours) involve a multidrug regimen of alkylating agents, topoisomerase inhibitors, spindel inhibitors etc, which, in addition, might have overlapping effects. One study using prolonged tamoxifenlcytotoxic drug combinations found a prominent histiocyte-like tumor cell appearance after treatment that caused some difficulty in distinguishing from benign cells [19]. In contrast, we saw such cells only very focally and only in a few cases (6%). Reasons for this discrepancy could lie in the higher number of cycles (on average 7) and/or the addition of tamoxifen in the Kennedy study. Some treatment-related effects presented diagnostic problems as shown in case 27. After therapy, this tumor showed multiple ducts with widespread intraductal proliferation. This included focally highly atypical cells within the luminal epithelial layer superimposed on an otherwise non-atypical epithelial proliferation (Fig. 3). No invasive element was present. Notably, though, every duct showed an intact circumferential myoepithelial cell layer without any atypia. These features support intraductal hyperplasia with superimposed chemotherapeutic effects rather than a diagnosis of DCIS. Likewise, nuclear atypia in benign lobules, occasionally with binucleation, was present in 30% of the post-treatment tumors, leading to patterns that mimicked intrinsic lobular atypia (Fig. 4B). Some of these cases might have presented diagnostic difficulties had they been evaluated without knowledge of prior cytotoxic drug exposure. Overall, the histopathologic appearance of the tumors changed significantly enough that we agree with the cautious view of Rasbridge et al. [28], who recommended against grading a tumor after treatment. The presence of p53 in breast cancer has been associated with a poor longterm prognosis [3]. However, we were unable to correlate the initial p53 status of the tumor with short term responses such as pathologic remission or drug-induced nuclear atypia, possibly due to the small number of cases studied. In an earlier study involving the first 32 of the 61 patients, we found a treatment-associated change of the p53 immunophenotype in 12/31 (39%) of the tumors [23]. This phenomenon most likely indicates clonal selection. Interestingly, wild type p53 tumors were less likely to undergo clonal selection (3/12) than mutant p53 tumors (9/12). This clinical study suggests that wild-type p53 exerts a "guardian of genome" function not only in normal cells but also in breast cancer cells in vivo. Widespread tumor cell death as it occurs in treatmentassociated complete pathologic remission could be mediated by a number of mechanisms including p53-dependent apoptosis, p53-independent apoptosis or non-
apoptotic mitotic arrest. It is currently unclear, if wildtype p53 mediated apoptosis (rather than G1 arrest) in response to chemotherapy occurs in breast cancer. Currently available studies on breast cell lines with functional wild type p53 report either resistance to apoptosis [11] or only limited apoptotic response after DNA damaging drugs [11, 22]. Technically, two potential pitfalls must be considered when comparing the morphology of breast carcinomas between a needle biopsy and a mastectomy specimen. First, the small biopsy could have an inherent sampling error with respect to architectural and cytological features and mitotic count [4]. Second, the slower fixation of the larger tissue slice from the mastectomy might produce cytological artifacts. Sampling errors, however, are statistically counteracted by analyzing large numbers of cases. Taken together, the Close study [7] and our study suggest that the observed increase in nuclear grade is unlikely due to sampling error. Concerning potential fixation artifacts, Close et al. [7] also looked at 10 needle biopsies and corresponding mastectomies from patients who did not receive any preoperative chemotherapy. Because these mastectomy specimens lacked any of the nuclear and cytoplasmic features seen in the treated specimens, the possibility of an artifact was excluded. Our experience concerning the morphology of untreated mastectomies is identical to that of Close et al. [7] and Sharkey et al. [30]. Largely driven by the newer concept that breast cancer is a systemic disease from the onset, the management of invasive disease has undergone major changes. Lately, a strong trend is seen towards neoadjuvant polychemotherapy as the first line of treatment in locally advanced breast cancer as opposed to postsurgical chemotherapy [27]. Several large studies start to show an overall survival benefit from neoadjuvant versus adjuvant therapy [21, 29]. It is conceivable that this treatment will also become a tool in early breast cancer. Therefore, it will be increasingly important for pathologists to recognize drug-induced histologic changes when diagnosing such breast cancer specimens. Acknowledgement: The authors thank Drs. Guy Riou and Marie-Christine Mathieu, Institut Gustave Roussy, 94805 Villejuif, France for making tissue available. This work was supported by an Emil C. Voll Breast Cancer Research Grant to UMM.
References I. Anderson EDC, Forrest APM, Hawkins RA et al. (1991) Primary systemic therapy for operable breast cancer. Br J Cancer 63,561-566 2. Azzopardi JG, Chepick OF, Hartmann WH et al. (1981) International Histological Classification of Tumours. No.
Chemotherapy Effect in Breast Tissue . 195
3.
4. 5.
6. 7. 8. 9.
10.
11.
12.
13.
14.
15.
16. 17.
2. Histological typing of breast tumours. Second Edition, World Health Organization, Geneva Barnes DB, Dublin EA, Fisher CF, Levison DA, Millis RR (1993) Immunohistochemical detection of p53 protein in mammary carcinoma: an important new independent indicator of prognosis? Hum Pathol24, 469--476 Biesterfeld S, Noll I, Noll E, Wohltmann D, Bocking A (1995) Mitotic frequency as a prognostic factor in breast cancer. Hum Pathol26, 47-52 Bonadonna G, Veronesi V, Brambilla C, Ferrari L, Luini A, Greco M, Bartoli C, Coopmans De Yolda G, Zucali R, Rilke F, Andreola C, Silvestrini R, Di Fronzo G, Valagussa P (1990) Primary chemotherapy to avoid mastectomy in tumors with diameters of three centimeters or more. Nat J Cancer Inst 82, 1539-1545 Brifford M, Spyratos F, Tubiana-Huhn M et al. (1989) Sequential cytopunctures during preoperative chemotherapy for primary breast cancer. Cancer 63,631-637 Close PM, Tiltman AJ, Hacking EA (1987) The effect of chemotherapy on the morphology of human breast carcinoma. S Afr Med J 72, 103-106 Davila E and Vogel CL (1984) Management of locally advanced breast cancer (stage III): a review. Int Adv Surg Oncol 7,297-327 DeLena M, Zucali R, Viganotti G, Valagussa P, Bonadonna G (1978) Combined chemotherapy-radiotherapy approach in locally advanced (T3 b-T4) breast cancer. Cancer Chemother Pharmacoll, 53-57 Elston CW, Ellis 10 (1991) Pathological prognostic factors in breast cancer. I. The value of histologic grade in breast cancer. Experience from a large study with longterm follow-up. Histopathology 19,403--410 Fan S, Smith ML, Rivet DJ 2nd, Duba D, Zhan Q, Kohn KW, Fornace AJ Jr, O'Connor PM (1995) Disruption of p53 function sensitizes breast cancer MCF-7 cells to cisplatin and pentoxyfylline. Cancer Res 55, 1649-1654 Frierson HF Jr, Fechner RE (1994) Histologic grade of locally advanced infiltrating ductal carcinoma after treatment with induction chemotherapy. Am J Clin Pathol 102, 154-157 Honkoop AH, Pinedo HM, De Jong JS, Verheul HMW, Linn SC, Hoekman K, Wagstaff J, Van Diest PJ (1997) Effects of chemotherapy on pathologic and biologic characteristics of locally advanced breast cancer. Am J Clin Pathol 107, 211-218 Hortobagyi GN, Spanos W, Montague ED, Buzdar AU, Yap H-Y, Blumenschein GR (1983) Treatment of locoregionally advanced breast cancer with surgery, radiotherapy, and combination chemoimmunotherapy. Int J Radiat Oncol BioI Phys 9, 643-650 Hortobagyi GN, Buzdar AU (1991) Locally advanced breast cancer: a review including M.D. Anderson experience. In: J Ragaz, 1M Ariel (Eds) High Risk Breast Cancer, pp 382--415. Springer Verlag, Berlin Hortobagyi GN, Buzdar AU, Strom EA, Ames FC; Singletary SE (1995) Primary chemotherapy for early and advanced breast cancer. Cancer Letters 90, 103-109 Israel L, Breau J-L, Morene JF et al. (1986) Two years of high dose cyclophosphamide and 5-fluoruracil followed by surgery after 3 months for acute inflammatory breast carcinoma. Cancer 57, 24-28
18. Jacquilat C, Baillet F, Weil M, Auclerc G, Housset M, Auclerc MF, Sellami M, Jindani A, Thill I, Soubrane C, Khayat D (1988) Results of conservative treatment combining induction (neoadjuvant) and consolidation chemotherapy, hormonotherapy, and external and interstitial irradiation in 98 patients with locally advanced breast cancer (IIIA-IIIB). Cancer 61, 1977-1982 19. Kennedy S, Merino MJ, Swain SM, Lippman ME (1990) The effects of hormonal and chemotherapy on tumoral and nonneoplastic breast tissue. Hum Pathol 21, 192-198 20. Lippman ME, Sorace RA, Bagley CS, Danforth DW, Lichter A, Wesley MN (1986) Treatment of locally advanced breast cancer using primary induction chemotherapy with hormonal synchronization followed by radiation therapy with or without debulking surgery. NCI Monogr 1, 153-159 21. Mauriac L, Durand M, Avril A, Dilhuydy J-M (1991) Effects of primary chemotherapy in conservative treatment of breast cancer patients with operable tumors larger than 3 em: results of a randomized trial in a single center. Ann OncoI2,347-354 22. Merlo GR, Basolo F, Fiore L, Duboc L, Hynes NE (1995) P53-dependent and p53-independent activation of apoptosis in mammary epithelial cells reveals a survival function of EGF and insulin. J Cell Bioi 128, 1185-1196 23. Moll UM, Ostermeyer AG, Ahomadegbe J-C, Mathieu M-C, Riou G (1995) P53-mediated tumor cell response to chemotherapeutic DNA damage: a preliminary study in matched pairs of breast cancer biopsies. Human Pathology 26, 1293-1301 24. Morrow M, Braverman A, Thelmo W et al. (1986) Multimodal therapy for locally advanced breast cancer. Arch Surg 121,1291-1296 25. Olson JE, Gray R, Sponzo RW, Damsker 11, Tormey DC, Cummings FJ (1986) Management of nonresectable locally advanced (Stage III) breast cancer: an ECOG trial. Breast Cancer Res Treat 8 (Abst. 124): 109 26. Perloff M, Lesnick GJ, Korzun A, Chu F, Holland JF, Thirlwell MP, Ellison RR, Carey RW, Leone L, Weinberg V, Rice MA, Wood WC (1988) Combination chemotherapy with mastectomy or radiotherapy for stage III breast carcinoma: A cancer and leukemia group B study. J Clin Oncol 6, 261-269 27. Powles TJ, Hikish TF, Makris A, Ashley SE, O'Brien ME, Tidy VA, Casey S et al. (1995) Randomized trial of chemoendocrine therapy started before or after surgery for treatment of primary breast cancer. J Clin Oncol 13, 547-552 28. Rasbridge SA, Gillett CE, Seymour AM, Patel K, Richards MA, Rubens RD, Millis RR (1994) The effects of chemotherapy on morphology, cellular proliferation, apoptosis and oncoprotein expression in primary breast carcinoma. British J Cancer 70, 335-341 29. Scholl SM, Fourquet A, Asselain B, Pierga JY, Vilcoq JR, Durand JC, Dorval T, Palangie T, Jouve M, Beuzeboc Pet al. (1994) Neoadjuvant versus adjuvant chemotherapy in premenopausal patients with tumors considered too large for breast conserving surgery: preliminary results of a randomized trial: S6. Eur J Cancer 30A (5), 645 30. Sharkey FE, Addington SL, Fowler LJ, Page CP, Cruz AB (1996) Effects of preoperative chemotherapy on the mor-
196 . U. M. Moll and J. Chumas phology of resectable breast carcinoma. Mod Pathol 9, 893-900 31. Williams MR, Gilson D, Marsh L, Morgan DAL, Nicholson RL, Elston CW, Griffiths K, Blarney RW (1988) The early results from a randomised study of radiotherapy versus Nolvadex (Tamoxiden) as initial treatment for stage III breast cancer. Eur J Surg Oncol 14,235-240
Received: October 21, 1996 Accepted in revised form: April I, 1997 Address for correspondence: Dr. U. M. Moll, Department of Pathology, SUNY Stony Brook Medical School, Stony Brook NY N.Y, 11794-8691, USA. Tel.: (516) 444 2459; e-mail: [email protected]