Transcript Ki-67 <= 14% (n)

Prognostic impact of Ki-67 in Croatian women with early breast cancer (single-institution prospective
observational study)
Ivan Bilić, Natalija Dedić Plavetić, Marko Kralik*, Paula Podolski, Ana Koši Kunac, Damir Vrbanec
Department of oncology, University Hospital Center Zagreb, Croatia
*Department of radiology, University Hospital Center Zagreb, Croatia
MATERIALS AND METHODS
Final total number of patients (pts.) was 209 pts (215 patients started; 6 were excluded being
initially metastatic, or lost to follow-up). Median age at diagnosis was 56 years (min 30, max
87). Median follow up is 9.38 years (min 8 months, max 10.4 years). 135 pts. were menopausal
at diagnosis. All pts. were diagnosed by pathohistological examination of extirpated tumor
including immunohistochemistry (IHC) for estrogen (ER) and progesterone (PR) receptor, as
well as Her-2 and Ki-67 determination. Her-2/neu confirming methods were FISH or CISH for
suspected amplification (++ by IHC). Adjuvant therapy followed according to TNM stage and
accepted ESMO guidelines, with exception of adjuvant trastuzumab, which was not reimbursed
by health insurance at that time.
Pts. were grouped according to Ki-67 expression (cut-off value 14% - percentage of positive
nuclei per hundred tumor cells), grade, stage and ER/PR/Her-2 profiles (five breast cancer
subtypes according to St. Gallen recommendations). Kaplan-Meier curve with log-rank test, or
Kruskal-Wallis, Mann-Whitney test with Bonferroni correction were used for univariate, and
Cox regresion analysis for multivariate models.
RESULTS
Overall survival (OS) significantly differs in Ki-67 positive vs. negative group (positive >14%,
OS=77.9 vs. 62.2, p=0.023). Accordingly, pts. that died during follow-up have significantly
higher Ki-67 value (16 vs. 9.67, p=0.003). DFS was not significantly affected by Ki67
positivity (data not shown).
Ki-67 value correlated with histologic grade (10 vs. 16.8 vs. 38.9% in grades I to III,
respectively – p<0.001, χ2=15.204). Kruskal-Wallis analysis confirmed non-equity of Ki-67
among grade-groups, but non-parametric test between I and II grade failed to reach statistic
significance.
Tumor grade had prognostic impact on OS (after 10 years, 90.1vs.77.7 vs. 57.4 in grades I to
III, respectively – p=0.006) and on disease free survival (DFS) as well.
Stratification according to five molecular intrinsic subtypes of breast cancer, using ER/HR/Her2 and Ki-67 (luminal A (1) and B (2), Her-2 enriched luminal (3) and non-luminal (4) and (5)
triple-negative) revealed significantly lower Ki-67 expression in luminal A-like and
significantly higher in luminal B-like group than Her-2 enriched and triple negative-like groups
(P<0.005). In multivariate regression analysis Ki-67 expression did not reveal significant
impact on OS and DFS of patients stratified in aforementioned subtype-like groups.
Table 1. Description of the patient cohort (N=209)
Age (yr)
Menopause (n)
Tumor type (n)
Hystological grade (n)
ER positive (n)
PR positive (n)
Her-2 positive (n)
Ki-67 (%)
Ki-67 <= 14% (n)
Follow-up (yr)
Died (n)
Relapse (n)
Died without relapse (n)
Relapse site (n)
Picture 1. Overall survival is significantly lower in Ki67 positive group.
Mean 57,63; St.dev. 12,089; Median 56; Min 30; Max 87
135
Ductal - 129, Lobular - 44, Mixed - 11, Other - 31
I - 33, II - 110, III - 72
135
110
35
Mean 11,4; St. dev. 12,88; Median 7,2; Min 2; Max 64
160
Mean 8,127, St.dev. 2,756 Median 9,38, Min 0,397, Max 10,412
51
61
14
Local 16, Distant 41, Both 2, Contralateral 2
Picture 2. Ki67 is significantly higher among patients who died during follow-up period
18
16
(mean)
16
Ki67 (%) - mean value
INTRODUCTION
Ki-67 is increasingly gaining attention as important proliferation marker in breast cancer. It is
considered to be relatively easy assayed by commercialy available kit for immunohistological
annalysis of tumor tissue. However, despite it is widely adopted, it is still under scrutiny
concerning the role in therapeutic decision-making.
We found data on prognostic factors in Croatian women suffering from early breast cancer
scarcely published. Here we present single-institution data on proliferation marker Ki-67 in
patient cohort with primary operable breast cancer treated in Breast cancer unit of University
Hospital Center Zagreb during years 2002 and 2003.
Picture 3. Higher grade tumors are more frequently Ki67 positive
14
12
10
9,67
(mean)
Median
11,1
8
6
Median
6,4
4
2
0
Alive
Dead
Picture 4. Overall survival is significantly lower among patients with higer tumor-grade
Picture 5. Relapse is more frequent when tumor-grade is higher
CONCLUSIONS
• Ki-67 alone had prognostic value for OS at 10 years follow-up in our cohort.
• Ki-67 level and tumor grade showed correlation, mostly significant, but tumor grade appeared to have stronger prognostic value than Ki-67.
• Grouping in molecular intrinsic-like subtypes of breast cancer according to IHC measurements of ER/PR/Her-2/Ki-67, showed Ki-67 difference between luminal A and B-like and the rest of subtypes, but
failed to show prognostic impact concerning OS and DFS between groups.