- Email: [email protected]
Prognostic value of cytological grading of fine-needle aspirates from breast carcinomas
Prognostic value of cytological grading of fine-needle aspirates from breast carcinomas
Summary
Introduction
Because
The value of histological typing1 and grading2-5 of breast carcinomas is well established. The histological characteristics used to assess grade in Elston’s modified Bloom and Richardson method are tubule formation, cellular pleiomorphism, and mitotic rate.6 With this method breast cancers can be scored in the range 3 to 9 and assigned to one of three histological grades. Grading of breast carcinoma in isolation has prognostic value,7 and grading has become routine in many laboratories. It is recommended that grade should form a part of all pathology
neoadjuvant therapy, including preoperative chemotherapy and tamoxifen, is becoming increasingly common for early breast cancer, it is desirable to grade tumours before surgery so that the most appropriate medical regimen can be selected. We have used a cytological grading system for ductal carcinoma of type not otherwise specified
(NOS).
Wet-fixed Papanicolaou-stained breast aspirates are examined for the extent of cell dissociation, cell size and uniformity, and the appearance of nucleoli, the nuclear margin, and chromatin. 377 invasive breast carcinomas were removed after preoperative diagnostic fine-needle aspiration cytology (FNAC) during the 25 months of the study. 286 tumours were ductal carcinomas NOS on histology. We established three cytological grades and found that cytological grade corresponded well with the established histological grades (Elston’s modified Bloom and Richardson method). All cytological features included in the score were equally
important on regression analysis. This study shows that grading of breast cancer on FNAC is feasible and reproducible. Cytological grade may substitute for histological grade, so a combination of FNAC and mammography can provide information on tumour type, grade, and size before surgery. We recommend this grading system to centres that use FNAC for the diagnosis of breast cancers. Lancet 1994; 343: 947-49
reports.88 Lymph-node metastasis is another important prognostic factor;9 nodal status is scored in the range 1 to 3 by a sampling or clearance procedure. Tumour diameter, measured from the histologically stained slide, also has an influence on survival. The Nottingham prognostic index for breast carcinoma combines these three features.3.10 This index was developed retrospectivelylO but has since been validated over 20 years in Nottingham3 and other centres." Because neoadjuvant therapy is becoming increasingly popular as primary medical treatment for breast cancer, there is much attraction in grading a tumour on fine-needle aspiration cytology (FNAC). Such grading would allow assessment of the tumour in situ, so the most suitable treatment could be selected immediately and the morbidity associated with overtreatment of low-grade tumours could be avoided. Diagnosis of breast carcinoma by FNAC is increasingly common. It is, however, impossible to differentiate reliably and reproducibly between invasive carcinoma and intraduct carcinoma on FNAC alone.12,13 This defect is not surprising, since the diagnosis of intraduct neoplasia requires the careful study of overall architecture and basement membrane integrity that only histopathology can provide. Clinical and mammographic features may suggest pure intraduct neoplasia. We have developed a protocol for the cytological grading of ductal carcinoma of type not otherwise specified (NOS). We have routinely typed and graded all carcinomas diagnosed on breast FNAC since late 1991 and present our ’
findings here. Patients and methods Department of Cytopathology at Royal Surrey County Hospital reports on breast FNAC in more than 2100 patients per year, from both women undergoing screening and symptomatic cases. Only women whose subsequently excised tissue was processed and examined in the histopathology department of this hospital were included in the study to ensure standardisation. During the 25 months of the study 366 women had positive breast FNAC and subsequent excision of their tumours. 10 patients had more than one neoplasm. 377 aspirates with subsequent biopsy were received. The
Departments of Cytopathology (I A Robinson MRCPI, G McKee FIAC), Histopathology (A Nicholson MB, P A Jackson MRCPath, M G Cook FRCPath), and Surgery (J D’Arcy MB, M W Kissin FRCSI), Royal Surrey County Hospital, Egerton Road, Guildford, Surrey GU2 5XX, UK
Correspondence to: Dr Grace
McKee
We differentiate on FNAC between lobular carcinoma, ductal carcinoma of special types, and ductal carcinoma NOS. In late 1991
947
Table 4:
RBC =red blood cell.
Table 1: Criteria for
cytological grading of breast carcinoma
protocol for the cytological grading of duct carcinoma NOS was developed by one of us (GM). This system uses six of the well-recognised criteria of malignant disease (table 1) and each is a
scored in the range 1 to 3. Both wet-fixed Papanicolaou-stained and air-dried MayGrunwald-Giemsa-stained smears are routinely used for diagnosis; the grading assessment is done on Papanicolaou-stained smears. This staining permits assessment of nucleoli, chromatin pattern, and nuclear membrane. 14 Nuclear size is measured by comparison with associated red blood cells.15 Grading is done at the time of cytological diagnosis and forms part of the cytology report. The scores allotted for each of the features are added to give a final score of between 6 and 18 for each sample. Reproducibility of cytological grading was assessed by the blind regrading of 20 randomly selected aspirates by the cytopathologists on two occasions separated by 3 months. The subsequent excisions were processed and reported routinely and were graded by the established histological method.6 The histopathologists who report on breast samples are not involved in the breast FNAC reporting service. Neither regression analysis nor correlation coefficients are appropriate tests to compare the measurement methods, since it is
highly unlikely that two methods designed to measure the same quality would not be related.16 Multiple regression analysis was used to assess the significance of each cytological feature and X2 tests were done to assess the importance of eliminating single features or combinations or giving different weights to features in an attempt to refine the grading system. Results Of the 377
examined, 18 were classified on histology as intraduct (non-invasive), 22 as ductal carcinomas of special type, 3 as non-epithelial in origin, and tumours
lobular carcinomas. Thus 286 were classified as ductal carcinomas; 5 of these were incorrectly classified on FNAC (4 as lobular and 1 mucinous carcinomas). 281 tumours classified as ductal carcinoma NOS were included in the 48
as
analysis.
Table 2:
Comparison of histological and cytological grades
Agreement within and between observers for two cytological gradings of same samples
Table 3:
948
Multiple regression analysis of cytological features
On histological grading 92 (32 ° ) tumours were grade 1, 120 (43%) grade 2, and 69 (25%) grade 3 (table 2). To produce similar proportions in the cytological grades, we established categories with scores of 6-11 for grade 1,12-14 for grade 2, and 15-18 for grade 3. 96 (34 %) of tumours were of cytological grade 1, 123 (44%) of grade 2, and 62 (22%) of grade 3. The mean histological scores of cytological grades 1, 2, and 3 differed significantly from each other-5-19 (SD 1-16), 6-29 (1-39), and 7-33 (1-23), respectively (p < 005, t test). There was good agreement within and between cytopathologists on grading the 20 randomly selected cases on two occasions (table 3). Every cytological feature had a significant relation to histological score on regression analysis (p < 0-001in each case), and each was related to the others. Multiple regression analysis showed that extent of cell dissociation and appearance of nucleoli were the most influential features (table 4). Removal of each cytological feature in turn and re-analysis of the data did not affect the X2 test results on the relation between the grading systems (X2 89-6, p < 0-001 for all six features); X2 ranged from 85-5 when cell uniformity was excluded to 67-1 when dissociation was excluded (p < 0-001 in each case). Empirical weighting (ie, doubling) of the value for dissociation and for nucleoli had no effect. =
Discussion The features for cytological grading of breast cancer are not well established and there have been only a few retrospective studies with differing conclusions.14,15,17-21 Two studies found cell clustering or dissociation useful19,20 but one found it difficult to assess.15 Nuclear size has long been established as a prognostic indicator .15,18-21 Graticules or micrometers have been used in most studies. These methods are precise but time-consuming and are not widely used. Nuclear size is variable, even within similar preparations because the degree of unavoidable air-drying varies. We overcame this problem by comparing cell size with that of adjacent red blood cells.15 Cell pleiomorphism may also be useful;15,19,20 it is straightforward to assess and correlates with histological pleiomorphism. 15 We assessed three components of cellular pleiomorphism separately: cellular uniformity (ie, similarity of cell appearance), nuclear outline, and chromatin pattern all correlated significantly with
histological score. The presence or absence of nucleoli has been used in other grading systems.19,20 Nucleoli are not used in assessing histological grade unless multiple nucleoli are present.6 Black et all did use the presence of nucleoli in their histological nuclear grading system, which includes the regularity of the nuclear outline, the delicacy of the chromatin strands, presence or absence of nucleoli, and the number of mitotic figures. This system is as valid as the original Bloom and Richardson method.23 Thomas et al20 found nucleoli unhelpful in determining grade, but we found the presence and character of nucleoli valuable in
grading
both
independently
and in combination with the
other cytological features. The aim of comparing one measurement system with another is to see whether the agreement is good enough for the other. Many such investigations use of correlation, but this practice is inappropriate. Since the two measurement methods (the standard and the one under investigation) measure the same quality, it would be remarkable if they did not agree. By contrast with previous studies, our cytological grades clearly distinguished between the three established histological grades. Another important factor in comparisons of measurement methods is variability between observers. For histopathological grading of breast neoplasms, agreement ranges from 90% for two pathologists working in the same centre6 to 26% for non-coordinators in the National Coordinating Group for Breast Screening Pathology (unpublished). Cytological grade does not correspond exactly to histological grade, but it is equally good, especially when the variability in histological grade is considered. Variability between cytological observers was similar to that between histological observers. Our finding that the cytological grading system did not improve when we excluded certain features or weighted others suggests that we are correct to ascribe equal importance to each feature. The features quantified were related to each other. The overview of all breast-cancer adjuvant-therapy trials published in 1992z4 showed a long-term survival benefit for patients receiving adjuvant therapy. The Nottingham prognostic index is a useful and sensitive guide for selecting adjuvant systemic therapy25 and there is no reason why such an index should not be developed for neoadjuvant treatment. Such an index might imitate the Nottingham prognostic index, since tumour diameter can be estimated by ultrasound, tumour grade by cytology, and lymph-node status by a staging lymphadenectomy while the primary tumour is left in situ. A new era of systemic adjuvant therapy has been heralded by studies in which treatment was given before surgery, either in conventional regimens26,27 or by cutaneous infusion of newer drugs .2" These studies have shown a high response rate with the tumours rapidly decreasing in size. It has been suggested that faster-growing tumours (grade 3) are more likely to respond to chemotherapy than low-grade slow-growing lesions, which may be better suited to pre-treatment with tamoxifen. Assessment of the biological aggressiveness of the cancer without removing it would therefore be valuable. FNAC grading allows such assessment and serial estimates to see how pre-treatment modulates tumour grade. Our cytological grading system has been in use for longer than 2 years. This prospective study shows that it is possible to grade invasive breast cancer on cytological aspirates and that cytological grade corresponds well with histological grade. The system we use is simple and takes little time. The information obtained is of clinical use and may be of prognostic value especially when combined with mammographic findings. one to
replace
coefficients
We thank Dr Martin Bland and Ms Caroline Ong for statistical advice and Dr Phillip McKee for constructive advice, suggestions, and encouragement
during preparation of the paper.
References
Broughton N, Locker A, Blamey R, Elston C. Pathological prognostic factors in breast cancer—II histological type: relationship with survival in a large study with a long term follow-up. Histopathology 1992; 20: 479-89. 2 Bloom H, Richardson W. Histological grading and prognosis in breast cancer. Br J Cancer 1957; 11: 359-77. 3 Elston C, Ellis I. Pathological prognostic factors in breast cancer-I the value of histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology 1991; 19: 403-10. 4 Fisher E, Redmond C, Fisher B. Histological grading of breast cancer. Pathology Annu 1980; 15: 239-51. 5 Page D, Anderson T, eds. Diagnostic histopathology of the breast. 1 st ed. Edinburgh: Churchill Livingstone, 1987. 6 Elston C. Grading of invasive carcinoma of the breast. In: Page D, Anderson T, ed. Diagnostic histopathology of the breast. Edinburgh: Churchill Livingstone, 1987: 300-11. 7 Doussal V, Tubiana M, Friedman S, Hacene K, Spyratos F, Brunet M. Prognostic value of histologic grade nuclear component 1
Ellis I, Galea M,
of Scarf-Bloom-Richardson. Cancer 1989; 64: 1914-21. National Institute of Health Consensus Development. Statement on the treatment of early stage breast cancer. Oncology 1991; 5: 120-24. 9 Fisher E. The impact of pathology on the biologic, diagnostic, prognostic and therapeutic considerations in breast cancer. Surg Clin North Am 1984; 64: 361-66. 10 Haybittle J, Blamey R, Elston C. A prognostic index in primary breast cancer. Br J Cancer 1982; 45: 361-66. 11 Todd J, Dowle C, Williams M, et al. Confirmation of a prognostic index in primary breast cancer. Br J Cancer 1987; 56: 489-92. 12 Sneige N, White VA, Katz RL, Troncoso P, Libshitz HI. Ductal carcinoma-in-situ of the breast: fine-needle aspiration cytology of 12 8
cases. Diagnost Cytopathol 1989; 5: 371-77. Zajicek J. Aspiration biopsy cytology: cytology of supra diaphragmatic organs. New York: Karger, 1974: 166-88. 14 Schulte E, Wittekind C. The influence of wet fixed Papanicolaou and the air dried Giemsa technique on nuclear parameters in breast cancer cytology: a cytomorphometric analysis. Diagnost Cytopathol 1987; 3:
13
256-61. 15 Hunt CM, Ellis
IO, Elston CW, Locker A, Pearson D, Blarney RW. Cytological grading of breast carcinoma-a feasible proposition? Cytopathology 1990; 1: 287-95. 16 Bland M, Altman D. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; i: 307-10. 17 Ciatto S, Bonardi R, Herd-Smith A, Cariaggi P, Confortini M, Bulgaresi G. Prognostic value of breast cancer grading: a retrospective study of 213 cases. Diagnost Cytopathol 1993; 9: 160-63. 18 Mossler J, McCarthy K, Woodward B, Mitchner L, Johnston E. Correlation of mean nuclear area with estrogen receptor content in aspiration cytology of breast carcinoma. Acta Cytol 1982; 26: 417-21. 19
Mouriquand J, Gozlan-Fior M, Villemain D, et al. Value of cytoprognostic classification in breast carcinomas. J Clin Pathol 1986; 39: 489-96.
20 Thomas J, Mallon E, George W. needle aspirates from benign and Pathol 1989; 42: 28-34.
Semiquantitative analysis of fine malignant breast lesions. J Clin
Zajdela A, DeLaRiva L, Ghossein N. The relation of prognosis to the nuclear diameter of breast cancer cells obtained by cytologic aspiration. Acta Cytol 1984; 23: 75-80. 22 Black M, Opler S, Speer F. Survival in breast cancer cases in relation to structure of the primary tumour and regional lymph nodes. Surg Gynecol Obstet 1955; 100: 543-51. 23 Eichner W, Lemmon H, Friedell G. Tumour grade in the prognosis of breast cancer. Nebraska Med J 1970; 55: 405-09. 24 Early Breast Cancer Trialists’ Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy. Lancet 1992; 339: 1-15. 25 Morgan D, Sibbering D, Galea M, Ellis I, Elston C, Blarney R. Selection for adjuvant therapy using the Nottingham prognostic index. Breast 1993; 2: 206. 26 Zurrida S, Greco M, Veronesi U. Surgical aspects of the conservative treatment of large size breast cancers after primary chemotherapy. Breast 1993: 2: 187. 27 Powles T. A randomised trial of adjuvant versus neoadjuvant endocrine chemotherapy of operable breast cancer. Breast 1993; 2: 187. 28 Smith I, Jones A, O’Brien M, McKinner J, Baum M. Primary medical (neoadjuvant) chemotherapy for operable breast cancer. Eur J Cancer 1993; 29: 1796-99. 21
949
Summary
Introduction
Because
The value of histological typing1 and grading2-5 of breast carcinomas is well established. The histological characteristics used to assess grade in Elston’s modified Bloom and Richardson method are tubule formation, cellular pleiomorphism, and mitotic rate.6 With this method breast cancers can be scored in the range 3 to 9 and assigned to one of three histological grades. Grading of breast carcinoma in isolation has prognostic value,7 and grading has become routine in many laboratories. It is recommended that grade should form a part of all pathology
neoadjuvant therapy, including preoperative chemotherapy and tamoxifen, is becoming increasingly common for early breast cancer, it is desirable to grade tumours before surgery so that the most appropriate medical regimen can be selected. We have used a cytological grading system for ductal carcinoma of type not otherwise specified
(NOS).
Wet-fixed Papanicolaou-stained breast aspirates are examined for the extent of cell dissociation, cell size and uniformity, and the appearance of nucleoli, the nuclear margin, and chromatin. 377 invasive breast carcinomas were removed after preoperative diagnostic fine-needle aspiration cytology (FNAC) during the 25 months of the study. 286 tumours were ductal carcinomas NOS on histology. We established three cytological grades and found that cytological grade corresponded well with the established histological grades (Elston’s modified Bloom and Richardson method). All cytological features included in the score were equally
important on regression analysis. This study shows that grading of breast cancer on FNAC is feasible and reproducible. Cytological grade may substitute for histological grade, so a combination of FNAC and mammography can provide information on tumour type, grade, and size before surgery. We recommend this grading system to centres that use FNAC for the diagnosis of breast cancers. Lancet 1994; 343: 947-49
reports.88 Lymph-node metastasis is another important prognostic factor;9 nodal status is scored in the range 1 to 3 by a sampling or clearance procedure. Tumour diameter, measured from the histologically stained slide, also has an influence on survival. The Nottingham prognostic index for breast carcinoma combines these three features.3.10 This index was developed retrospectivelylO but has since been validated over 20 years in Nottingham3 and other centres." Because neoadjuvant therapy is becoming increasingly popular as primary medical treatment for breast cancer, there is much attraction in grading a tumour on fine-needle aspiration cytology (FNAC). Such grading would allow assessment of the tumour in situ, so the most suitable treatment could be selected immediately and the morbidity associated with overtreatment of low-grade tumours could be avoided. Diagnosis of breast carcinoma by FNAC is increasingly common. It is, however, impossible to differentiate reliably and reproducibly between invasive carcinoma and intraduct carcinoma on FNAC alone.12,13 This defect is not surprising, since the diagnosis of intraduct neoplasia requires the careful study of overall architecture and basement membrane integrity that only histopathology can provide. Clinical and mammographic features may suggest pure intraduct neoplasia. We have developed a protocol for the cytological grading of ductal carcinoma of type not otherwise specified (NOS). We have routinely typed and graded all carcinomas diagnosed on breast FNAC since late 1991 and present our ’
findings here. Patients and methods Department of Cytopathology at Royal Surrey County Hospital reports on breast FNAC in more than 2100 patients per year, from both women undergoing screening and symptomatic cases. Only women whose subsequently excised tissue was processed and examined in the histopathology department of this hospital were included in the study to ensure standardisation. During the 25 months of the study 366 women had positive breast FNAC and subsequent excision of their tumours. 10 patients had more than one neoplasm. 377 aspirates with subsequent biopsy were received. The
Departments of Cytopathology (I A Robinson MRCPI, G McKee FIAC), Histopathology (A Nicholson MB, P A Jackson MRCPath, M G Cook FRCPath), and Surgery (J D’Arcy MB, M W Kissin FRCSI), Royal Surrey County Hospital, Egerton Road, Guildford, Surrey GU2 5XX, UK
Correspondence to: Dr Grace
McKee
We differentiate on FNAC between lobular carcinoma, ductal carcinoma of special types, and ductal carcinoma NOS. In late 1991
947
Table 4:
RBC =red blood cell.
Table 1: Criteria for
cytological grading of breast carcinoma
protocol for the cytological grading of duct carcinoma NOS was developed by one of us (GM). This system uses six of the well-recognised criteria of malignant disease (table 1) and each is a
scored in the range 1 to 3. Both wet-fixed Papanicolaou-stained and air-dried MayGrunwald-Giemsa-stained smears are routinely used for diagnosis; the grading assessment is done on Papanicolaou-stained smears. This staining permits assessment of nucleoli, chromatin pattern, and nuclear membrane. 14 Nuclear size is measured by comparison with associated red blood cells.15 Grading is done at the time of cytological diagnosis and forms part of the cytology report. The scores allotted for each of the features are added to give a final score of between 6 and 18 for each sample. Reproducibility of cytological grading was assessed by the blind regrading of 20 randomly selected aspirates by the cytopathologists on two occasions separated by 3 months. The subsequent excisions were processed and reported routinely and were graded by the established histological method.6 The histopathologists who report on breast samples are not involved in the breast FNAC reporting service. Neither regression analysis nor correlation coefficients are appropriate tests to compare the measurement methods, since it is
highly unlikely that two methods designed to measure the same quality would not be related.16 Multiple regression analysis was used to assess the significance of each cytological feature and X2 tests were done to assess the importance of eliminating single features or combinations or giving different weights to features in an attempt to refine the grading system. Results Of the 377
examined, 18 were classified on histology as intraduct (non-invasive), 22 as ductal carcinomas of special type, 3 as non-epithelial in origin, and tumours
lobular carcinomas. Thus 286 were classified as ductal carcinomas; 5 of these were incorrectly classified on FNAC (4 as lobular and 1 mucinous carcinomas). 281 tumours classified as ductal carcinoma NOS were included in the 48
as
analysis.
Table 2:
Comparison of histological and cytological grades
Agreement within and between observers for two cytological gradings of same samples
Table 3:
948
Multiple regression analysis of cytological features
On histological grading 92 (32 ° ) tumours were grade 1, 120 (43%) grade 2, and 69 (25%) grade 3 (table 2). To produce similar proportions in the cytological grades, we established categories with scores of 6-11 for grade 1,12-14 for grade 2, and 15-18 for grade 3. 96 (34 %) of tumours were of cytological grade 1, 123 (44%) of grade 2, and 62 (22%) of grade 3. The mean histological scores of cytological grades 1, 2, and 3 differed significantly from each other-5-19 (SD 1-16), 6-29 (1-39), and 7-33 (1-23), respectively (p < 005, t test). There was good agreement within and between cytopathologists on grading the 20 randomly selected cases on two occasions (table 3). Every cytological feature had a significant relation to histological score on regression analysis (p < 0-001in each case), and each was related to the others. Multiple regression analysis showed that extent of cell dissociation and appearance of nucleoli were the most influential features (table 4). Removal of each cytological feature in turn and re-analysis of the data did not affect the X2 test results on the relation between the grading systems (X2 89-6, p < 0-001 for all six features); X2 ranged from 85-5 when cell uniformity was excluded to 67-1 when dissociation was excluded (p < 0-001 in each case). Empirical weighting (ie, doubling) of the value for dissociation and for nucleoli had no effect. =
Discussion The features for cytological grading of breast cancer are not well established and there have been only a few retrospective studies with differing conclusions.14,15,17-21 Two studies found cell clustering or dissociation useful19,20 but one found it difficult to assess.15 Nuclear size has long been established as a prognostic indicator .15,18-21 Graticules or micrometers have been used in most studies. These methods are precise but time-consuming and are not widely used. Nuclear size is variable, even within similar preparations because the degree of unavoidable air-drying varies. We overcame this problem by comparing cell size with that of adjacent red blood cells.15 Cell pleiomorphism may also be useful;15,19,20 it is straightforward to assess and correlates with histological pleiomorphism. 15 We assessed three components of cellular pleiomorphism separately: cellular uniformity (ie, similarity of cell appearance), nuclear outline, and chromatin pattern all correlated significantly with
histological score. The presence or absence of nucleoli has been used in other grading systems.19,20 Nucleoli are not used in assessing histological grade unless multiple nucleoli are present.6 Black et all did use the presence of nucleoli in their histological nuclear grading system, which includes the regularity of the nuclear outline, the delicacy of the chromatin strands, presence or absence of nucleoli, and the number of mitotic figures. This system is as valid as the original Bloom and Richardson method.23 Thomas et al20 found nucleoli unhelpful in determining grade, but we found the presence and character of nucleoli valuable in
grading
both
independently
and in combination with the
other cytological features. The aim of comparing one measurement system with another is to see whether the agreement is good enough for the other. Many such investigations use of correlation, but this practice is inappropriate. Since the two measurement methods (the standard and the one under investigation) measure the same quality, it would be remarkable if they did not agree. By contrast with previous studies, our cytological grades clearly distinguished between the three established histological grades. Another important factor in comparisons of measurement methods is variability between observers. For histopathological grading of breast neoplasms, agreement ranges from 90% for two pathologists working in the same centre6 to 26% for non-coordinators in the National Coordinating Group for Breast Screening Pathology (unpublished). Cytological grade does not correspond exactly to histological grade, but it is equally good, especially when the variability in histological grade is considered. Variability between cytological observers was similar to that between histological observers. Our finding that the cytological grading system did not improve when we excluded certain features or weighted others suggests that we are correct to ascribe equal importance to each feature. The features quantified were related to each other. The overview of all breast-cancer adjuvant-therapy trials published in 1992z4 showed a long-term survival benefit for patients receiving adjuvant therapy. The Nottingham prognostic index is a useful and sensitive guide for selecting adjuvant systemic therapy25 and there is no reason why such an index should not be developed for neoadjuvant treatment. Such an index might imitate the Nottingham prognostic index, since tumour diameter can be estimated by ultrasound, tumour grade by cytology, and lymph-node status by a staging lymphadenectomy while the primary tumour is left in situ. A new era of systemic adjuvant therapy has been heralded by studies in which treatment was given before surgery, either in conventional regimens26,27 or by cutaneous infusion of newer drugs .2" These studies have shown a high response rate with the tumours rapidly decreasing in size. It has been suggested that faster-growing tumours (grade 3) are more likely to respond to chemotherapy than low-grade slow-growing lesions, which may be better suited to pre-treatment with tamoxifen. Assessment of the biological aggressiveness of the cancer without removing it would therefore be valuable. FNAC grading allows such assessment and serial estimates to see how pre-treatment modulates tumour grade. Our cytological grading system has been in use for longer than 2 years. This prospective study shows that it is possible to grade invasive breast cancer on cytological aspirates and that cytological grade corresponds well with histological grade. The system we use is simple and takes little time. The information obtained is of clinical use and may be of prognostic value especially when combined with mammographic findings. one to
replace
coefficients
We thank Dr Martin Bland and Ms Caroline Ong for statistical advice and Dr Phillip McKee for constructive advice, suggestions, and encouragement
during preparation of the paper.
References
Broughton N, Locker A, Blamey R, Elston C. Pathological prognostic factors in breast cancer—II histological type: relationship with survival in a large study with a long term follow-up. Histopathology 1992; 20: 479-89. 2 Bloom H, Richardson W. Histological grading and prognosis in breast cancer. Br J Cancer 1957; 11: 359-77. 3 Elston C, Ellis I. Pathological prognostic factors in breast cancer-I the value of histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology 1991; 19: 403-10. 4 Fisher E, Redmond C, Fisher B. Histological grading of breast cancer. Pathology Annu 1980; 15: 239-51. 5 Page D, Anderson T, eds. Diagnostic histopathology of the breast. 1 st ed. Edinburgh: Churchill Livingstone, 1987. 6 Elston C. Grading of invasive carcinoma of the breast. In: Page D, Anderson T, ed. Diagnostic histopathology of the breast. Edinburgh: Churchill Livingstone, 1987: 300-11. 7 Doussal V, Tubiana M, Friedman S, Hacene K, Spyratos F, Brunet M. Prognostic value of histologic grade nuclear component 1
Ellis I, Galea M,
of Scarf-Bloom-Richardson. Cancer 1989; 64: 1914-21. National Institute of Health Consensus Development. Statement on the treatment of early stage breast cancer. Oncology 1991; 5: 120-24. 9 Fisher E. The impact of pathology on the biologic, diagnostic, prognostic and therapeutic considerations in breast cancer. Surg Clin North Am 1984; 64: 361-66. 10 Haybittle J, Blamey R, Elston C. A prognostic index in primary breast cancer. Br J Cancer 1982; 45: 361-66. 11 Todd J, Dowle C, Williams M, et al. Confirmation of a prognostic index in primary breast cancer. Br J Cancer 1987; 56: 489-92. 12 Sneige N, White VA, Katz RL, Troncoso P, Libshitz HI. Ductal carcinoma-in-situ of the breast: fine-needle aspiration cytology of 12 8
cases. Diagnost Cytopathol 1989; 5: 371-77. Zajicek J. Aspiration biopsy cytology: cytology of supra diaphragmatic organs. New York: Karger, 1974: 166-88. 14 Schulte E, Wittekind C. The influence of wet fixed Papanicolaou and the air dried Giemsa technique on nuclear parameters in breast cancer cytology: a cytomorphometric analysis. Diagnost Cytopathol 1987; 3:
13
256-61. 15 Hunt CM, Ellis
IO, Elston CW, Locker A, Pearson D, Blarney RW. Cytological grading of breast carcinoma-a feasible proposition? Cytopathology 1990; 1: 287-95. 16 Bland M, Altman D. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; i: 307-10. 17 Ciatto S, Bonardi R, Herd-Smith A, Cariaggi P, Confortini M, Bulgaresi G. Prognostic value of breast cancer grading: a retrospective study of 213 cases. Diagnost Cytopathol 1993; 9: 160-63. 18 Mossler J, McCarthy K, Woodward B, Mitchner L, Johnston E. Correlation of mean nuclear area with estrogen receptor content in aspiration cytology of breast carcinoma. Acta Cytol 1982; 26: 417-21. 19
Mouriquand J, Gozlan-Fior M, Villemain D, et al. Value of cytoprognostic classification in breast carcinomas. J Clin Pathol 1986; 39: 489-96.
20 Thomas J, Mallon E, George W. needle aspirates from benign and Pathol 1989; 42: 28-34.
Semiquantitative analysis of fine malignant breast lesions. J Clin
Zajdela A, DeLaRiva L, Ghossein N. The relation of prognosis to the nuclear diameter of breast cancer cells obtained by cytologic aspiration. Acta Cytol 1984; 23: 75-80. 22 Black M, Opler S, Speer F. Survival in breast cancer cases in relation to structure of the primary tumour and regional lymph nodes. Surg Gynecol Obstet 1955; 100: 543-51. 23 Eichner W, Lemmon H, Friedell G. Tumour grade in the prognosis of breast cancer. Nebraska Med J 1970; 55: 405-09. 24 Early Breast Cancer Trialists’ Collaborative Group. Systemic treatment of early breast cancer by hormonal, cytotoxic, or immune therapy. Lancet 1992; 339: 1-15. 25 Morgan D, Sibbering D, Galea M, Ellis I, Elston C, Blarney R. Selection for adjuvant therapy using the Nottingham prognostic index. Breast 1993; 2: 206. 26 Zurrida S, Greco M, Veronesi U. Surgical aspects of the conservative treatment of large size breast cancers after primary chemotherapy. Breast 1993: 2: 187. 27 Powles T. A randomised trial of adjuvant versus neoadjuvant endocrine chemotherapy of operable breast cancer. Breast 1993; 2: 187. 28 Smith I, Jones A, O’Brien M, McKinner J, Baum M. Primary medical (neoadjuvant) chemotherapy for operable breast cancer. Eur J Cancer 1993; 29: 1796-99. 21
949