- Email: [email protected]
Confirmation of a prognostic index for breast cancer
150
The Breast
In discriminant analysis for grade, Score predicts 72% of cases correctly. An objective measure of histological grade using flow and static cytometry is described. This can be calculated by technical officers. Score predicts grade correctly in over 70% of cases; this reproducible cytological grading system should be of value in clinical practice.
The index derived by Haybittle et al produces prognostic groupings with as-good discriminatory powers as a more complex index derived from the data. The index predicts both survival and time to general recurrence (ie. death from disease), and the separation between the three groups is still apparent at 10 years. This analysis shows both the general applicability of the Nottingham prognostic index, and also the maintenance of the difference in the group’s survival after 10 years’ follow-up. Comparison
CONFIRMATION OF A PROGNOSTIC INDEX FOR BREAST CANCER
Prognostic
MH Galea, CW Elston, IO Ellis, RW Blarney City Hospital, Nottingham, UK A prognostic index, derived initially from a group of 387 patients with primary breast cancer and follow-up of l-6 years has been reported (1). This index has been recalculated for the same 387 patients now with a minimum follow-up of 11 years. The three factors used in the index, size. made and lvmnh node staee have remained indenendent 18 predictors of prognos~ and their c&tribution to the index remains unchanged with time. First analysis Factors Menopausal status Tumour size Lymph node stage Tumour grade ER status
Further analysis
Z
Beta
Z
Beta
1.5 2.92 5.29 4.56 -1.72
0.5 0.17 0.76 0.82 -0.34
0.18 2.6 1 8.64 5.48 -1.32
0.03 0.13 0.78 0.62 -0.20
The index has been applied prospectively to a further 1168 patients with a minimum follow-up of 2 years and shown to be equally effective in identifying patients with either a good prognosis (NPI ~3.4). n=372(31%), 10 year survival 82% and unlikely to benefit from adjuvant therapies, or a poor prognosis (NPI >5.4), n=l74(15%), 10 year survival 17% in whom adjuvant therapies would be of value. This prognostic index, based on tumour size, grade and lymph node stage is a powerful method of predicting prognosis and defining patient management. This study has shown that the index holds its power to at least 15 years of follow-up. Reference 1. Haybittle JL, Blarney RW, Elston CW et al. A prognostic index in primary breast cancer. BR J Cancer 1982; 45: 301-66
17 ANALYSIS OF LONG-TERM PROGNOSTIC FACTORS IN BREAST CANCER M Jones’ and E A Benson’ ‘Yorkshire Regional Cancer Organisation 2Yorkshire Breast Cancer Group, UK The purpose of this analysis was to identify long-term prognostic factors on an unselected group of patients, and to compare the resulting prognostic index with that derived by Haybittle, Blarney et al, 1982. Between 1975 and 1981, 1186 cases of non-metastatic breast cancer were registered by fifteen consultants from the Yorkshire Breast Cancer Group. (YBCG) Multivariate survival analysis was performed, using Cox proportional hazards model. This gave a prognostic index using the following variables: clinical tumour size, tumour grade (vs. 2+3) nodal involvement (path assessment) and age at diagnosis * [* young age being associated with a poorer prognosis] Patients were categorised into Good, Moderate and Poor prognostic groups on the basis of the inter-quartile ranges of their covariate score. The general recurrence-free survival rates for the three groups at five and ten years are given below. Patients were then separately categorised into the Good, Moderate and Poor prognostic groups defined by Haybittle, Blarney et al, 1982. These were based on tumour size, nodal involvement, and Bloom and Richardson grade. Again, five and ten-year groups are given below.
of 5- and IO-year general recurrence
rates
group
5-year rate (%) Nottingham YBCG
IO-year rate (%) Nottingham YBCG
Good Moderate Poor
88 68 44
80 57 39
87 67 42
80 58 38
OF INTEGRlN AND STROMAL PROTEIN EXPRESSION - A MARKER OF PRE-MALIGNANT CHANGE?
ALTERATION
J. Louise Jones and Rosemary A. Walker Department of Pathology, Leicester Royal Infirmary,
UK
Interaction between epithelia and the extracellular matrix are known to play a central role in the modulation of cell growth, differentiation and motility. Any alterations in these interactions may result in loss of normal control mechanisms and development of the malignant phenotype. The interactions are mediated in part by the cell adhesion molecules called integrins. This study examines the expression of integrin subunits IX*,cts, B, and B4 and the stromal protein tenascin in 53 breast carcinomas, noninvolved breast tissue from 21 of these cases and 32 normal/benign cases. Frozen tissue and an indirect immunoperoxidase technique was used throughout. Linear staining in relation to the basement membrane was seen for all integrins in the normal/benign cases. The carcinomas showed complete loss of reactivity in 65% of cases for ctz, 80% for CL,and B4 and 90% for pt. Those showing reactivity displayed a diffuse cytoplasmic type of staining. The non-involved breast tissue showed linear basement membrane staining with a2 and Br, but for a6 and p4, 66% of cases displayed reactivity identical to that of the corresponding tumour. For tenascin, band-like staining around ducts was seen in normal/benign cases, with a diffuse coarse reactivity in all carcinomas. The altered a& integrin expression in non-involved tissue in cancerous breast may be an early event in the neoplastic process, and as such be of use as a marker of pre-malignant change. The modulation of a& integrin expression is currently under investigation.
19 IMPLICATIONS OF UNIFOCAL BREAST. CANCER FOR CONSERVING THERAPY R. Holland, Department
D. Faverly of Pathology,
Nijmegem,
Netherlands
Conventional breast-conserving therapy includes local radiotherapy of the breast after excision of the tumor. This radiotherapy is aimed at eradicating residual foci of a multifocal tumor. In a previous study we showed that as many as 40% of the breast cancers are non-multifocal. If these tumors were recognized in the surgical biopsy, post-operative local radiotherapy could be omitted. In the present study we assessed the percentage of unifocal tumors in a series of 149 patients with invasive breast cancers with sizes up to 4 cm treated by mastectomy. Tumor unifocality and multifocality in the mastectomy specimens were assessed by Egan’s correlated radiologic/histopathologic technique. The results show that in this series 52% of the tumors were unifocal defined as having no tumor foci (invasive, ductal in situ or intralymphatic) beyond 1 cm from the edge of the dominant mass. We further studied if the unifocal tumors could have been distinguished from the multifocal tumors in an excisional biopsy by
The Breast
In discriminant analysis for grade, Score predicts 72% of cases correctly. An objective measure of histological grade using flow and static cytometry is described. This can be calculated by technical officers. Score predicts grade correctly in over 70% of cases; this reproducible cytological grading system should be of value in clinical practice.
The index derived by Haybittle et al produces prognostic groupings with as-good discriminatory powers as a more complex index derived from the data. The index predicts both survival and time to general recurrence (ie. death from disease), and the separation between the three groups is still apparent at 10 years. This analysis shows both the general applicability of the Nottingham prognostic index, and also the maintenance of the difference in the group’s survival after 10 years’ follow-up. Comparison
CONFIRMATION OF A PROGNOSTIC INDEX FOR BREAST CANCER
Prognostic
MH Galea, CW Elston, IO Ellis, RW Blarney City Hospital, Nottingham, UK A prognostic index, derived initially from a group of 387 patients with primary breast cancer and follow-up of l-6 years has been reported (1). This index has been recalculated for the same 387 patients now with a minimum follow-up of 11 years. The three factors used in the index, size. made and lvmnh node staee have remained indenendent 18 predictors of prognos~ and their c&tribution to the index remains unchanged with time. First analysis Factors Menopausal status Tumour size Lymph node stage Tumour grade ER status
Further analysis
Z
Beta
Z
Beta
1.5 2.92 5.29 4.56 -1.72
0.5 0.17 0.76 0.82 -0.34
0.18 2.6 1 8.64 5.48 -1.32
0.03 0.13 0.78 0.62 -0.20
The index has been applied prospectively to a further 1168 patients with a minimum follow-up of 2 years and shown to be equally effective in identifying patients with either a good prognosis (NPI ~3.4). n=372(31%), 10 year survival 82% and unlikely to benefit from adjuvant therapies, or a poor prognosis (NPI >5.4), n=l74(15%), 10 year survival 17% in whom adjuvant therapies would be of value. This prognostic index, based on tumour size, grade and lymph node stage is a powerful method of predicting prognosis and defining patient management. This study has shown that the index holds its power to at least 15 years of follow-up. Reference 1. Haybittle JL, Blarney RW, Elston CW et al. A prognostic index in primary breast cancer. BR J Cancer 1982; 45: 301-66
17 ANALYSIS OF LONG-TERM PROGNOSTIC FACTORS IN BREAST CANCER M Jones’ and E A Benson’ ‘Yorkshire Regional Cancer Organisation 2Yorkshire Breast Cancer Group, UK The purpose of this analysis was to identify long-term prognostic factors on an unselected group of patients, and to compare the resulting prognostic index with that derived by Haybittle, Blarney et al, 1982. Between 1975 and 1981, 1186 cases of non-metastatic breast cancer were registered by fifteen consultants from the Yorkshire Breast Cancer Group. (YBCG) Multivariate survival analysis was performed, using Cox proportional hazards model. This gave a prognostic index using the following variables: clinical tumour size, tumour grade (vs. 2+3) nodal involvement (path assessment) and age at diagnosis * [* young age being associated with a poorer prognosis] Patients were categorised into Good, Moderate and Poor prognostic groups on the basis of the inter-quartile ranges of their covariate score. The general recurrence-free survival rates for the three groups at five and ten years are given below. Patients were then separately categorised into the Good, Moderate and Poor prognostic groups defined by Haybittle, Blarney et al, 1982. These were based on tumour size, nodal involvement, and Bloom and Richardson grade. Again, five and ten-year groups are given below.
of 5- and IO-year general recurrence
rates
group
5-year rate (%) Nottingham YBCG
IO-year rate (%) Nottingham YBCG
Good Moderate Poor
88 68 44
80 57 39
87 67 42
80 58 38
OF INTEGRlN AND STROMAL PROTEIN EXPRESSION - A MARKER OF PRE-MALIGNANT CHANGE?
ALTERATION
J. Louise Jones and Rosemary A. Walker Department of Pathology, Leicester Royal Infirmary,
UK
Interaction between epithelia and the extracellular matrix are known to play a central role in the modulation of cell growth, differentiation and motility. Any alterations in these interactions may result in loss of normal control mechanisms and development of the malignant phenotype. The interactions are mediated in part by the cell adhesion molecules called integrins. This study examines the expression of integrin subunits IX*,cts, B, and B4 and the stromal protein tenascin in 53 breast carcinomas, noninvolved breast tissue from 21 of these cases and 32 normal/benign cases. Frozen tissue and an indirect immunoperoxidase technique was used throughout. Linear staining in relation to the basement membrane was seen for all integrins in the normal/benign cases. The carcinomas showed complete loss of reactivity in 65% of cases for ctz, 80% for CL,and B4 and 90% for pt. Those showing reactivity displayed a diffuse cytoplasmic type of staining. The non-involved breast tissue showed linear basement membrane staining with a2 and Br, but for a6 and p4, 66% of cases displayed reactivity identical to that of the corresponding tumour. For tenascin, band-like staining around ducts was seen in normal/benign cases, with a diffuse coarse reactivity in all carcinomas. The altered a& integrin expression in non-involved tissue in cancerous breast may be an early event in the neoplastic process, and as such be of use as a marker of pre-malignant change. The modulation of a& integrin expression is currently under investigation.
19 IMPLICATIONS OF UNIFOCAL BREAST. CANCER FOR CONSERVING THERAPY R. Holland, Department
D. Faverly of Pathology,
Nijmegem,
Netherlands
Conventional breast-conserving therapy includes local radiotherapy of the breast after excision of the tumor. This radiotherapy is aimed at eradicating residual foci of a multifocal tumor. In a previous study we showed that as many as 40% of the breast cancers are non-multifocal. If these tumors were recognized in the surgical biopsy, post-operative local radiotherapy could be omitted. In the present study we assessed the percentage of unifocal tumors in a series of 149 patients with invasive breast cancers with sizes up to 4 cm treated by mastectomy. Tumor unifocality and multifocality in the mastectomy specimens were assessed by Egan’s correlated radiologic/histopathologic technique. The results show that in this series 52% of the tumors were unifocal defined as having no tumor foci (invasive, ductal in situ or intralymphatic) beyond 1 cm from the edge of the dominant mass. We further studied if the unifocal tumors could have been distinguished from the multifocal tumors in an excisional biopsy by