35. Comparability of prognostic indices

Abstracts of the 4th Nottingham International Breast Cancer Conference lated for these small tumours. This has as its basis: not only tumour size but grade and lymph node status, these results are shown in Table 2. Table

2

Tumour

size < 10 mm

Actuarial lo-year survival

NPI

n

Good Moderate

246

90%

90 7

73% 42%

Poor

This shows that the index is a more powerful predictor of survival than size alone, even for small tumours. These findings show that biological factors are important in patients with small breast cancer and contradict the findings of Tabar.

34. Pathological and immunohistochemical prognostic factors in clinically node-negative breast cancer patients receiving adjuvant tamoxifen (BR3002 trial)

237

Nottingham City Hospital and University Hospital of South Manchester, UK Two similar prognostic indices (PI), which rely on measurement of prognostic factors at the time of operation, have been devised: both have been validated within their own centres on new data sets. The Nottingham Index (NPI) uses size, grade and lymph nodes status from axillary plus internal mammary sampling and uses node positivity, level, number involved. The Manchester Index (MPI) uses size, grade, menopausal status plus factors from axillary clearance - node positivity, level, number involved and ratio of number benign to number malignant. Both show steadily worse survival with each integer rise in index. Splitting both to five score levels, comparability of survival and of numbers selected to each level is seen: Score

NPI n = 2409

level

MPI n = 540

% selected % 10 year survival % selected % 10 year survival I

13 20 21 24 I5

88 80 64 46 13

13 12 31 24 13

93 82 69 46 6

P. R. Baker, R. A. Walker, J. A. Dunn, S. M. Pamell, P. S. Stonelake, C. M. Jevons, W. M. Gillespie, P. G. Baker, L. Dodson, G. D. Oates, D. J. Ellis, M. J. R. Lee, J. M. Morrison, D. Spooner

2 3 4 5

West Midlands Breast Group, UK

The accurate selection of patients with excellent long-term outlook (or) with poor short-term survival, have important implications regarding the prescription of adjuvant systemic therapy.

Although adjuvant tamoxifen reduces the rates of recurrence and death, many patients fail to benefit. Furthermore, the optimum duration of tamoxifen therapy is as yet unknown The availability of good prognostic markers might improve the selection of low and high risk subgroups. The patients studied were entered into the West Midlands Breast Group collaborative prospective randomized trial of radiotherapy and adjuvant tamoxifen (2 years or continuously) in the conservative management of clinically mode-negative early breast cancer. At a median follow-up of 4.2 years (IQ range, 2.8-5.8 years) there is no significant difference in relapse-free survival between the two durations of tamoxifen treatment. To date, primary tumours of 393 of the trial patients, of whom 104 developed recurrent disease, have been subjected to histopathology review and paraffin sections has been stained for nine proteins of prognostic interest. Log-rank analysis for relapse-free survival identified the following as significant prognostic factors: age [I 50 vs > 50 years; P = 0.0031, menopausal status [pre vs post; P = O.OOl], tumour pathological size (2 2 vs > 2 cm; P = 0.0031, tumour grade [l vs 2 vs 3; P < O.OOOl], vascular invasion [- vs +; P < 0.0151, ER-Hscore [< 100 vs > 100; P = 0.0004], PgR-% staining [< 20 vs > 20; P = 0.0002], pS2 [- vs +; P = 0.0261, c-erbB2 [- vs +; P < 0.003], ~53 [- vs +; P < 0.0021 and Ki-67 (MIBl) [- vs +; P = 0.003]. DCIS, cathepsin D, TGFcr and EGFR were not significantly associated with relapse-free survival. Applying Cox proportional hazards model to this subset of patients (303 with complete information; 80 recurrences) identified menopausal status [x2 = 15.8, P = O.OOOl], PgR% [xZ = 6.4, P = O.Ol], grade [x’ = 5.8, P < 0.021 and size [continuous: x’ = 3.9, P < 0.051 as independent prognostic factors; vascular invasion was of borderline significance [x? = 3.8, P = 0.051. This preliminary analysis suggests that several pathological and immunohistochemical factors have prognostic value in patients with clinically node-negative early breast cancer receiving adjuvant tamoxifen.

35. Comparability

of prognostic indices

A. Howell, E. Knox, R. Swindell, N. Bundred, M. Galea, R. W. Blarney, J. L. Haybittle, C. W. Elston. and I. 0. Ellis

36. Rates of local recurrence with neoadjuvant chemoendocrine therapy for primary breast cancer J. R. Benson, T. J. Powles, A. Makris, G. Knee, S Ashley,

A. G. Nash

Breast Unit, The Royal Marsden Hospital, London and Surrey, UK Breast conservation surgery for early breast cancer is associated with equivalent survival, but higher rates of local relapse than mastectomy. Pathological margin status and EIC are major determinants of relapse risk. In a randomized trial of chemoendocrine therapy administered either prior to or following primary surgery and radiotherapy, we have assessed neoadjuvant and adjuvant regimens on rates of loco-regional relapse. The first 200 patients (2 age 70 years) with operable stage I and II breast cancer diagnosed on fine needle aspirate have been analyzed. Those in the adjuvant group received 6 months of chemotherapy together with tamoxifen (continued for 5 years) whilst 3 months of chemoendocrine therapy before and after appropriate surgery and radiotherapy was given as a neoadjuvant schedule. Overall clinical response rates for the latter have been high (85%), and at a median follow-up of 28 months, only 4 patients have relapsed locally in either the breast (2 adjuvant. 1 neoadjuvant) or axilla (1 adjuvant). Lower rates of recurrence have occurred in the neoadjuvant group despite a greater proportion of positive pathological margins (28% neoadjuvant, 24% adjuvant). Tumour grade and extent of DCIS were similar for the two groups (P > 0.05) with significantly more tumours