RELATIONSHIP OF DNA PLOIDY AND S-PHASE FRACTION TO SURVIVAL AFTER FIRST RECURRENCE OF BREAST CANCER

Transcription

1 ~ Act0 Oncologica Vol. 33, No. 4, pp , 1994 RELATIONSHIP OF DNA PLOIDY AND S-PHASE FRACTION TO SURVIVAL AFTER FIRST RECURRENCE OF BREAST CANCER OLLE STAL, JOHN M. CARSTENSEN, STEN WINGREN, LARS ERIK RUTQVIST, LAMBERT SKOOG, CLAES KLINTENBERG and Bo NORDENSKJOLD Flow cytometry was performed on frozen specimens from the primary tumour of 184 women with recurrent breast cancer. No significant association was seen between DNA ploidy and the other prognostic factors investigated. Patients with a high S-phase fraction had more often a negative estrogen receptor (ER) status and a short disease-free interval. A shorter survival after disease recurrence was seen both in patients with DNA aneuploid tumours and among those with a high S-phase fraction. Patients with DNA tetraploid tumours showed the longest survival after recurrence. In this subgroup, half of the patients survived more than 3 years after recurrence and the estimated survival rate at years was 17%. In a Cox s regression analysis including 116 patients, site of recurrence, number of positive nodes at time of primary operation, size and ER content of the primary tumour as well as DNA ploidy showed additional prognostic value. The prognosis after recurrence of breast cancer has been shown to depend, in part, on site of recurrence, time passed since initial treatment (1, 2), presenting clinical stage (2), and primary estrogen receptor (ER) content (3-5). However, there is a large unexplained variation in survival time between patients with recurrent breast cancer. Several studies have shown that measurements of DNA ploidy and/or the fraction of S-phase cells contribute prognostic information in addition to that of nodal status, tumour size, and ER content in primary breast cancer (6-18). On the other hand, most studies concerning the relationship between these cellular variables and survival after relapse have been either inconclusive or negative (6, 9,, 19-23), though some studies have shown a significantly shorter survival Accepted 30 March From the Department of Oncology (0. Stil, S. Wingren, C. Klintenberg, B. Nordenskjold), the Faculty of Health Sciences and Department of Health and Society (J. M. Carstensen), Linkoping University, Linkoping, Sweden and the Oncologic Centre (L.E. Rutqvist) and Division of Cytology (L. Skoog), Karolinksa Hospital, Stockholm, Sweden. Correspondence to: Dr Olle StBI, Department of Oncology, Faculty of Health Sciences, Linkoping University, S Linkoping, Sweden. time in patients with a high proportion of cells in S-phase (, ). In a previous paper we have reported on the results from an analysis of the relationships of DNA ploidy and S- phase fraction to recurrence-free interval and survival after operation of the primary breast tumour (13). In this report, we describe our findings for the same series of patients when DNA-ploidy and S-phase fraction of the primary tumour were related to postrecurrence survival in an analysis also including site of recurrence, recurrencefree interval, nodal status at primary operation, tumour size and ER content. Material and Methods Patients. The patients included in the present study originate from a cohort of 472 women, diagnosed for primary breast cancer in the Stockholm Health Care Region during February 1978 through December 1981 (13). The cohort consisted of women with unilateral breast cancer, no previous history of cancer, no preoperative therapy, and no evidence of distant dissemination at diagnosis. The patients were operated with a modified radical mastectomy (94%) or a partial mastectomy with axillary dissection. Twenty-three per cent received adjuvant sys- 0 Scandinavian University Press ISSN X 423

2 4 0. STAL ET AL. temic treatment; either CMF, tamoxifen or a combination of both. Of the patients, 31% received postmastectomy radiation to the chest wall and regional nodes, and all patients treated with a partial mastectomy were given postoperative radiotherapy to the breast. Follow-up visits took place once every 3 months during the first 2 years, every 6 months during 2-5 years and thereafter yearly. Routinely these visits only included a physical examination. Chest x-rays, bone scans, bloodtests, etc., were performed if clinical signs or symptoms indicated possible relapse. Disease recurrence was confirmed when possible by biopsy. and visceral metastases, however, were sometimes established on unequivocal radiological evidence. Patients who did not come to a scheduled follow-up visit were investigated using official death registers, a computerized register of in-patient care which covers all admissions into hospitals in Stockholm County, or were contacted by letter. Recurrence was dated from the first evidence of relapse based on physical, histologic or imaging data. A total of 186 patients developed a loco-regional and/or distant recurrence before January 1, Except for two patients with their first metastasis established at autopsy, these patients constitute the study group in this report. Of the 184 women in the study group, 17% experienced their recurrence within the first year after diagnosis, 30% during the second year, and 22% during the fifth year or later. Table 1 shows the distribution of patients according to the dominant site of first recurrence. Following recurrence, treatment was decided individually for each patient by the responsible clinician, and we have not attempted to account for the effects of treatment differences in the present study. DNA jlow cytometry and estrogen receptor analysis. All specimens were collected from fresh surgical resections of the primary tumours and stored below -70 C until analysed for DNA pattern and estrogen receptors. For flow cytometry, small pieces of the tumours were minced in 0.6 ml citrate buffer, and trout and chicken red blood cells were added as internal marker cells. A suspension of isolated nuclei was prepared as described by Vindelov et al. () using Nonidet P40 as detergent, treatment with trypsin, and stabilization by sperminetetrahydrochloride. After addition of RNAase, the suspension was filtered through a 41 pm nylon mesh and stained with propidium iodide. Imprints from the tissues were examined to ascertain the presence of tumour cells in the sample. Cell suspensions were analysed using a Leitz MPV FLOW flow cytometer interfaced to a Monroe OC8888 microcomputer. The software used for data acquisition and multi-parameter analysis was developed in our laboratory. Illumination from a high-pressure mercury lamp was used with light filtered through an AL interference filter with peak transmission at 546nm and 20nm bandwidth. Emission was recorded through a dichroic mirror TK 580 and a 590 nm long-pass filter. DNA histograms, usually with cells, were recorded. The DNA index was calculated assuming that chicken and trout erythrocytes show 35% and 80% respectively of the fluorescence of human diploid cells stained with propidium iodide (27). The coefficient of variation (CV) of DNA peaks was estimated from the width of peak at half-maximum peak height. Mean CV of the tumour DNA peaks was 4.1% (range 2.1-8%). The percentage of S-phase cells was estimated using a rectangular model. The number of S-phase cells was calculated by multiplying the number of channels between the Golr and GJM peaks by the mean number of cells in channels in the S-phase area interactively selected by the operator. Four DNA ploidy types were identified (Table 2). The tumours were classified as non-diploid if two or more peaks in addition to the reference peaks could be distinguished, otherwise as DNA diploid. In 116 of 184 DNA histograms the S-phase fraction was estimated. Mean and standard deviation (SD) of the S-phase fraction in relation to DNA ploidy are shown in Table 2. In most cases of hypodiploid and multiploid tumours, the S-phase level was not estimated since the histograms showed overlapping populations. Furthermore, the fraction of S-phase cells was not assessed if the number of tumour cells was small compared to that of other cells and registrations derived from debris. In the survival analysis S-phase fraction was categorized by our standard cut-off levels of 5% and %. Cytosol receptors for estradiol were measured as described Table 1 Number of patients by dominant site of first recurrence Dominant site of recurrence n % Lung Pleura Liver Brain Other viscera Table 2 Number of patients and mean S-phase fraction by DNA ploidy S-phase fraction Range of DNA ploidy type DNA Index n ' n Mean SD Diploid Tetraploid Hyperdiploid Other aneuploid' <1.0, ~ ' Including DNA multiploid tumours with two or more DNA non-diploid cell populations irrespective of DNA Index.

3 DNA PLOIDY AND S-PHASE FRACTION IN RECURRENT BREAST CANCER 4 by Wrange et al. (28) using incubation with 5 nm 3Hestradiol and isoelectric focusing of the receptor in polyacrylamide gel. The amount of receptor was related to the tumour DNA content. Statistical methods. Patient survival was calculated from the date of first recurrence until the date of death ( 171 patients) or the closing date of January 1, Those alive at the closing date were treated as censored observations. No patient was lost to follow-up. The cumulative probability of survival was estimated using the productlimit method (29). Relative death rates and two-tailed tests of the statistical significance of relationships between study variables and survival were computed using Cox s proportional hazards method (30). * P-0.m o a 72 so 1211 XOtttk. Fig. 1. Survival after first recurrence by DNA ploidy of the primary breast tumour. Results As shown in Table 3, there was no significant association between DNA ploidy and other prognostic variables. Patients with lung metastases had the highest S-phase levels, and those with bone metastases had the lowest levels, but there was no clear trend when the site categories were ordered according to prognosis. There was a significant Table 3 DNA ploidy and S-phase fraction in relation to other variables DNA Ploidy correlation between S-phase fraction and disease-free interval as well as ER content. The survival curves for different ploidy types are shown in Fig. 1. The DNA aneuploid categories ( hyperdiploid and other aneuploid ) showed significantly shorter survival than the DNA euploid types (Table 4). There was a tendency to longer survival in DNA tetraploid cases com- S-phase percentage Variable Percent Test for Test for Category n aneuploid trend n Mean SD trend Dominant site of recurrence Lung, pleura Other Disease-free interval 0-1 years 1-2 years 2-4 years >4 years Size <20 mm mm >31 mm No. of positive nodes at time of primary diagnosis >4 ER content (fmol/pg DNA) < > 1.00 Of the primary breast tumour p = 0. p = 0.27 p = 0.32 p = 0.37 p = p = 0. p = p = 0.18 p = 0.57 p = The diploid and tetraploid tumours were treated as one group and the rest of the tumours were regarded as DNA aneuploid.

4 4 0. STAL ET AL. Table 4 Univariate survival analyses O78, Median Log-rank Variable survival test for Category n (months) trend DNA ploidy Diploid Tetraploid Hyperdiploid Other aneuploid S-phase fraction < 5% 5-% > % Dominant site of recurrence Lung, pleura Other Disease-free interval 0-1 years 1-2 years 2-4 years >4 years Size <20 mm ~ R I >31 mm No. of positive nodes at time of primary diagnosis >4 ER content (fmol/mg DNA) <O.lO > 1.00 Of the primary tumour P= : 0.0 p = p < p = 0.22 p = p = p = pared to diploid, but the difference was not statistically significant. The proportion of cells in S-phase was significantly related to survival, with shorter survival in patients with higher S-phase fraction (Fig. 2 and Table 4). The site of recurrence was the single most important predictor of survival time (Table 4). Patients with visceral metastases ( lung, pleura and other ) had a medium survival time one-third that of patients with the recurrence located to soft tissue or bone. The disease-free interval after operation was not significantly related to survival after recurrence, though a tendency to longer survival among patients with later recurrences could be seen. An ER-positive primary tumour was associated with a longer survival after recurrence, whereas a larger primary tumour or a large number of positive nodes were related to shorter survival. Of all categories, the patients with DNA tetraploid tumows had the longest survival with a median of more than OB 120 Yontha Fig. 2. Survival after first recurrence by S-phase fraction of the primary breast tumour. 3 years, whereas more than half of the patients with highly proliferative tumours died within the first year after recurrence (Figs 1, 2). More than 90% of the patients died during the follow-up period. However, the estimated year survival rate after recurrence was over % for patients with small primary tumours ( <20 mm), for those with low or moderate S- phase fraction ( c lo%), for node-negative patients and for those whose dominant site of recurrence was soft tissue. For patients with DNA tetraploid tumours the survival rate was 17%. In Table 5, the results from the fit of a Cox s multiple regression model are presented. Only the 1 16 patients with S-phase measurements were included in the analysis. For comparison, results from univariate Cox analyses of the 116 patients are presented in the same table. Except for S-phase fraction and disease-free interval, all variables were significantly associated with postrecurrence survival time. The patients with DNA tetraploid tumours lived longer than diploid cases (p = 0.0). In the multivariate analysis, we also tested all possible two-way interactions. The 21 interaction terms were entered in a stepwise Cox s regression analysis in which all main effects were included from the first step. The first, and only convincing, interaction term entered was that between primary nodal status and tumour size (p = ). This indicated that nodal status was less important as predictor when the primary tumour was large and, at the same time, that tumour size was of less prognostic value when there were several positive nodes. The relationships between the older variables and postrecurrence survival were not substantially altered when this interaction was accounted for in the multivariate Cox s model. Discussion In the present study of recurrent breast cancer, DNA ploidy and S-phase fraction determined on the primary

5 DNA PLOIDY AND S-PHASE FRACTION IN RECURRENT BREAST CANCER 421 Table 5 Survival analyses using Cox's model Variable Category Multivariate analysis Univariate death rate Death rate 95% confidence n ratio ratio interval Test for trend DNA ploidy' Diploid Tetraploid Hyperdiploid Other aneuploid S-phase fraction' < 5% 5-% > % Dominant site of recurrence Lung, pleura Other Disease-free interval 0-1 years 1-2 years 2-4 years >4 years Size' t20 mm mm >31 mm No. of positive nodes at time of primary diagnosis >4 ER content' (fmol/mg DNA) to. 0.- o > 1.00 ' Of the primary tumour tumour were clearly related to postrecurrence survival, Moreover, DNA ploidy had significant, additional prognostic value when adjusted for various other characteristics. The multivariate analysis showed also that site of recurrence, number of positive nodes at time of primary operation, size, and ER content of the primary tumour, but not disease-free interval, were significantly prognostic factors. In the studies of Hedley et al. (19), Stuart-Harris et al. (21) and Harvey et al. (lo), patients with DNA diploid tumours seemed to have shorter survival times after recurrence than those with aneuploid primary tumours. The number of patients were, however, small and consequently the estimates were connected with large standard errors. Cornelisse et al. (9) investigated the relationship between DNA ploidy and survival in 158 patients with distant O p = p=0.17 p = p = 0.29 p < p = p = recurrence of their breast cancer, and found a slightly better survival in DNA diploid than in aneuploid cases (p = 0.1). A similar trend was observed in a Finnish study including 2 patients with known DNA ploidy status (23). In a study of 136 patients, Baildam et al. (22) observed a somewhat longer survival in patients with DNA diploid or tetraploid tumours compared to those with aneuploid tumours (p = 0.08). In contrast to the studies mentioned, we have divided non-diploid tumours into three categories. As in the study of Baildarn et al. (22), the DNA tetraploid group seemed to have the best prognosis. The relationship between the fraction of tumour cells in S phase, measured by thymidine labelling, and postrecurrence survival has been studied by several authors. Meyer & Lee () made their measurements on 48 relapsed breast

6 STAL ET AL. carcinomas and found a significantly shorter survival among those with high S-phase levels. In a material of 54 relapsing patients, Silvestrini et al. (20) did not find any difference in postrecurrence survival between slowly and rapidly proliferating primary tumours, whereas data on 63 relapsed patients reported by Tubiana et al. (6) suggested that patients with a high or a median labelling index died sooner than those with a low labelling index of the primary tumour. In a study using static cytometry, Hatschek et al. () found a similar result. Like Tubiana et al. (6) and Hatschek et al. (), we divided the patients in three groups by S-phase level. Silvestrini et al. (20) used only two categories, and this may have influenced their negative results. The variation in results between the studies can, however, also be due to the small number of patients included in some of them. Our finding of a prognostic value in recurrent breast cancer of primary ER content is in accordance with other large studies (3-5). Similar to the results of Koenders et al. (5), we found lymph node status, but not disease-free interval, to be of prognostic value. Earlier reports are more conflicting concerning the importance of primary nodal status as a predictor of postrecurrence survival (4). The decreasing prognostic importance of nodal status in patients with larger primary tumours indicated by the present study may in part explain the deviance between previous results. In conclusion, our results demonstrate that several of the most important predictors of recurrence-free interval and post-operation survival, including DNA ploidy and S-phase fraction, also may be used as predictors of postrecurrence survival. ACKNOWLEDGEMENTS This work was supported by grants from the Swedish Cancer Society and the King Gustav V Jubileefund. REFERENCES 1. Cutler SJ, Asire AJ, Taylor SG 111. Classification of patients with disseminated cancer of the breast. Cancer 1969; : Devitt JE. The enigmatic behavior of breast cancer. Cancer 1971; 27: Clark GM, Sledge GW Jr, Osborne CK, et al. Survival from first recurrence: Relative importance of prognostic factors in 15 breast cancer patients. J Clin Oncol 1987; 5: Shek LLM, Godolphin W, Spinelli JJ. Oestrogen receptors, nodes and stage as predictors of post-recurrence survival in 457 breast cancer patients. Br J Cancer 1987; 56: Koenders PG, Beex LVAM, Kloppenborg PWC, Smals AGH, Benraad ThJ, the Breast Cancer Study Group. Human breast cancer: survival from first metastasis. Breast Cancer Res Treat 1992; 21: Tubiana M, Pejovic MH, Chavaudra N, et al. The long-term progriostic significance of the thymidine labelling index in breast cancer. Int J Cancer 1984; 33: Klintenberg C, Stll 0, Nordenskjold B, et al. Proliferative index, cytosol estrogen receptor and axillary node status as prognostic predictors in human mammary carcinoma. Breast Cancer Res Treat 1986; 7 (Suppl): Silvestrini R, Daidone MG, Di Fronzo G, et al. Prognostic implication of labeling index versus estrogen receptors and tumor size in node-negative breast cancer. Breast Cancer Res Treat 1986; 7: Cornelisse CJ, van de Velde CJH, Caspers RJC, et al. DNA ploidy and survival in breast cancer patients, Cytometry 1987; 8: 2-.. Harvey J, de Klerk N, Berryman I, et al. Nuclear DNA content and prognosis in human breast cancer: a static cytophotometric study. Breast Cancer Res Treat 1987; 9: Kallioniemi 0-P, Blanco G, Alavikko M, et al. Tumour DNA ploidy as an independent prognostic factor in breast cancer. Br J Cancer 1987; 56: Meyer JS, Province M. Proliferative index of breast carcinoma by thymidine labelling: prognostic power independent of stage, estrogen and progesterone receptors. Breast Cancer Res Treat 1988; 12: Stil 0, Wingren S, Carstensen J, et al. Prognostic value of DNA ploidy and S-phase fraction in relation to estrogen receptor content and clinicopoathological variables in primary breast cancer. Eur J Cancer Clin Oncol 1989; : von Rosen A, Rutqvist LE, Carstensen J, et al. Prognostic value o! nuclear DNA content in breast cancer in relation to tumour size, nodal status and estrogen receptor content. Breast Cancer Res Treat 1989; 13: Hatschek T, Fagerberg G, Stil 0, et al. Cytometriccharacterization and clinical course of breast cancer diagnosed in a population based screening program. Cancer 1989; 64: Sigurdsson H, Baldetorp B, Borg A, et al. Indicators of prognosis in node-negative breast cancer. N Engl J Med 1990; 322: Ferno M, Baldetorp B, Borg A, Olsson H, Sigurdsson H, Killander K. Flow cytometric DNA index and S-phase fraction in breast cancer in relation to other prognostic variables and to clinical outcome. Acta Oncol 1992; 31: Stll 0, Carstensen J, Hatschek T, Nordenskjold. Significance of S-phase fraction and hormone receptor content in the management of young breast cancer patients. Br J Cancer 1992; 66: Hedley DW, Rugg CA, Ng ABP, et al. Influence of cellular DNA content on disease-free survival of stage I1 breast cancer patients. Cancer Res 1984; 44: Silvestrini R, Daidone MG, Gasparini G. Cell kinetics as a prognostic marker in node-negative breast cancer. Cancer 1985; 56: Stuart-Harris R, Hedley DW, Taylor IW, et al. Tumour ploidy, response and survival in patients receiving endocrine therapy for advanced breast cancer. Br J Cancer 1985; 51: Baildam AD, Zaloudik J, Howell A, et al. DNA analysis by flow cytometry, response to endocrine treatment and prognosis in advanced carcinoma of the breast. Br J Cancer 1987; 55: Blanco G, Holli K, Heikkinen M, Kallioniemi 0-P, Taskinen P. Prognostic factors in recurrent breast cancer: relationships to site of recurrence, disease-free interval, female sex steroid receptor.s, ploidy and histological malignancy grading. Br J Cancer 1990; 62: Meyer JS, Lee JY. Relationships of S-phase fraction of breast carcinoma in relapse to duration of remission, estrogen receptor content, therapeutic responsiveness, and duration of survival. Cancer Res 1980; 40:

7 DNA PLOIDY AND S-PHASE FRACTION IN RECURRENT BREAST CANCER 429. Hatschek T, Carstensen J, Fagerberg G, Stll 0, Grontoft 0, Nordenskjold B. Influence of S-phase fraction on metastatic pattern and post-recurrence survival in a randomized mammography screening trial. Breast Cancer Res and Treat 1989; 14: Vindelov LL, Christensen IBJ, Nissen NI. A detergent-trypsin method for the preparation of nuclei for flow cytometric DNA analysis. Cytometry 1983; 3: Vindelov LL, Christensen IBJ, Nissen NI. Standardization of high resolution flow cytometric DNA analysis by the simulta- neous use of chicken and trout red blood cells as internal standards. Cytometry 1983; 3: Wrange 0, Nordenskjold B, Gustafsson J-A. Cytosol estradiol receptor in human mammary carcinoma: an assay on isoelectric focussing in polyacrylamide gel. Anal Biochem 1978; 85: Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: Cox DR. Regression models and life tables (with discussion). J R Statist SOC B 1972; :