Purpose: A double-blind randomized Phase II chemoprevention trial of α-difluoromethylornithine (DFMO) was conducted in a group of women at high risk for development of breast cancer. DFMO is an irreversible inhibitor of ornithine decarboxylase, the limiting enzyme of polyamine synthesis that is often up-regulated in breast cancer.

Experimental Design: Study entrants were required to have random periareolar fine-needle aspiration cytology prior to entry that exhibited hyperplasia or hyperplasia with atypia, as well as a mammogram and clinical breast exam judged as not suspicious for breast cancer and no clinical hearing loss. Subjects were randomized to 6 months of oral DFMO (0.5 g/m2/day) or placebo, followed by repeat fine-needle aspiration and biomarker assessment. The main study end point was an improvement in cytologic pattern.

Results: Of 119 subjects entered, 96% completed the study and were evaluable for the main study end point. A modest reduction (28%) in average total urine polyamines was obtained in the DFMO group, but there was no reduction in the spermidine:spermine ratio. There was no difference in cytologic improvement between DFMO and placebo. Likewise, there was no difference between DFMO and placebo for the secondary end points of breast molecular marker changes (immunocytochemical expression of proliferating cell nuclear antigen, p53, and epidermal growth factor receptor), mammographic breast density, serum insulin-like growth factor I: insulin-like growth factor binding protein 3 ratio, adverse events, quality of life indices, or subsequent cancer development.

Conclusions: DFMO at a dose level of 0.5 g/m2/day administered for 6 months does not modulate breast risk biomarkers tested in this study.

Although tamoxifen has been found to reduce the incidence of breast cancer by 50% in some high-risk cohorts, its use is associated with an increased risk of uterine cancer, thromboembolic phenomena, and perimenopausal symptoms (1). Furthermore, tamoxifen does not reduce the incidence of ER3 -negative cancers, and the risk of developing ER-positive cancer is reduced by only two-thirds (1). Development of additional preventive agents is needed with activity in ER-negative as well as tamoxifen-resistant precancerous disease.

DFMO is an irreversible inhibitor of ODC, the rate-limiting enzyme of polyamine synthesis (2). It is attractive as a potential chemopreventive agent because it has the potential to disrupt biological processes important for both ER-negative and ER-positive tumor formation (3, 4, 5), reduces the incidence of carcinogen-induced and spontaneous mammary cancers in animal models (6, 7, 8, 9), and has minimal side effects at low doses (10, 11, 12, 13, 14, 15).

Polyamines (putrescine, spermidine, and spermine) enhance proliferation and other changes associated with cell growth and transformation by increasing transcriptional activation at gene promotors including those for proteases, cell adhesion, and cytoskeletal proteins via hormone-dependent and -independent transcription factors (16, 17, 18, 19, 20, 21). This is accomplished through facilitation of coactivator binding to key receptors such as ER-α and nuclear factor κB and their corresponding response elements in DNA (5, 21, 22, 23, 24, 25).

After exposure to estradiol or peptide growth factors such as IGF-I or epidermal growth factor, intracellular polyamines are increased via activation of ODC as well as enhanced intracellular polyamine transport from extracellular sources (26, 27, 28, 29, 30, 31). Active transport of polyamines is tightly controlled in normal epithelial cells, but negative feedback signals are blunted in transformed and malignant tissues, resulting in increased intracellular transport despite elevated intracellular polyamine levels (6, 32, 33). Increased ODC expression and elevated polyamine levels are observed in breast cancer relative to normal tissue, with ODC expression detected by immunohistochemistry in 10–90% of malignant epithelial cells (34, 35, 36). ODC expression and polyamine levels appear to be highest in poorly differentiated rapidly proliferating tumors (37, 38), but no consistent correlations have been reported with the presence or absence of hormone receptors (38, 39, 40). ODC expression has also been detected in epithelial cells in benign breast disease (41).

Inhibition of ODC activity by DFMO results in major reductions in putrescine and spermidine but only small reductions in spermine (6). Furthermore, spermidine and putrescine can be replenished via spermine pools through the enzyme spermidine/spermine N1-acetyltransferase (see Fig. 1; Ref. 6). Cell growth is diminished only with depletion of spermidine levels (6, 11), so that the spermidine:spermine ratio is often used as a marker of DFMO biochemical activity (10) in chemoprevention trials.

Toxicity from DFMO is dose dependent. GI and hematological side effects as well as ototoxicity are common in doses of ≥3 g/m2/day and occasional in doses of 1–3 g/m2/day but rare in doses of ≤0.5 g/m2/day (10, 11, 13, 14, 15, 42, 43, 44, 45). Despite minimal side effects, low-dose DFMO (0.5 g/m2/day) is associated with a reduced colonic spermidine:spermine ratio (13). Therefore, a dose of 0.5 g/m2 was selected by the NCI Chemoprevention Branch as the optimal daily dose for Phase II prevention trials in breast as well as several other organ sites (46).

In 1997, we initiated a Phase II prevention trial in women at increased risk for breast cancer. The purpose of the study was to assess the ability of DFMO at a dose of 0.5 g/m2/day for 6 months to modulate several breast cancer risk biomarkers relative to placebo. We selected random periareolar FNA as our method for breast tissue sampling to obtain cells for cytology and proliferation assessments based on our prospective study (47), which demonstrated a high prevalence of hyperplasia (49%) and hyperplasia with atypia (21%) in women at increased risk for breast cancer undergoing this type of sampling. Furthermore, random FNA evidence of hyperplasia with atypia was strongly correlated with short-term risk of breast cancer independent of risk based on epidemiological factors (47). Importantly, this sampling method had been demonstrated to have good patient acceptance with 85% of 480 women returning for a second aspiration within 6–18 months (47).

Several risk biomarkers were selected as response indicators, including cytologic pattern, proliferation index, mammographic breast density area, and serum IGF-I level and the ratio of IGF-I level to its binding protein IGFBP-3 (48, 49, 50, 51, 52, 53). Modulation of cytologic pattern was selected as the primary end point. Breast tissue polyamines were not assessed as part of this trial due to concerns that there would be insufficient specimen for assay. Urine polyamines were monitored, however, as a measure of the biochemical activity of DFMO in the individual subject.

Eligibility.

To be eligible for a screening FNA, women were required to be at increased risk for breast cancer on the basis of family history (≥1 affected first-degree relative or multiple affected second-degree relatives), prior treated breast cancer or breast biopsy showing atypical hyperplasia or lobular carcinoma in situ, or mammogram exhibiting ≥50% density relative to the entire breast area. Women exhibiting one or more of these characteristics underwent a baseline periareolar random FNA (see below). Women younger than 30 years could only be aspirated if they were within 10 years of the age of onset of breast cancer in a close relative; women older than 55 years could only participate if they had prior evidence of generalized proliferative breast disease. Women were allowed to be on oral contraceptives or HRT, provided they had not changed their dose of these medications for at least 6 months prior to baseline aspiration.

To enroll in the study, women were required to have a baseline random periareolar FNA (C. J. F.) within 6 months of randomization demonstrating either hyperplasia with atypia or hyperplasia without atypia but with expression (≥2+ intensity of ≥10% of epithelial cells) of p53 and/or EGFR. Women must have had reasonable renal and liver function (creatinine < 1.5 mg/dl, albumin > 3.0 g/dl, bilirubin < 1.5 mg/dl, aspartate aminotransferase < 100 units/liter, alkaline phosphatase < 200 units/liter). Separate consents were signed for the screening FNA and blood sampling and for participation in the chemoprevention trial. Mammography must have been performed within 6 months of randomization, with an interpretation (C. H. J. C.) of not suspicious for breast cancer, i.e., an ACR score of I-III (54). Subjects were required to report no clinical hearing loss and to exhibit hearing thresholds on baseline audiometry (J. A. F.) of no greater than 25 dB for frequencies between 0.5 and 2.0 kHz and no greater than 40 dB for frequencies above 2 kHz. Premenopausal women were required to take birth control precautions throughout the study.

Study Design.

A schema of the study design is shown in Fig. 2. Subjects were stratified by menopause status and presence or absence of atypia in the baseline FNA and were randomized to take 0.5 g/m2 DFMO or placebo p.o. once per day for 6 months. DFMO was formulated as an aqueous solution containing DFMO at 200 mg/ml. Molecular weight of DFMO is 236.65. DFMO or matched placebo was to be taken with fruit juice once daily in the morning with breakfast. Subjects were contacted monthly by the study nurse (D. A. B.) and queried for adverse events using a formatted interview technique. Collection of urine for assessment of polyamines; collection of blood for measurement of DFMO levels as well as complete blood count, renal, and liver function tests; clinical breast exams; QOL surveys; and study medication compliance checks were performed every 2 months (i.e., 0, 2, 4, and 6 months). History and physical exam, mammography, audiometry, QOL surveys, FNA for breast biomarkers, and collection of blood for assessment of serum hormones and growth factors were repeated at the end of the 6-month study period. Optional germ-line genetic testing for BRCA1 and BRCA2 mutations was offered at study exit (55).

Random Periareolar FNA.

For premenopausal women, random periareolar FNAs were performed (C. J. F.) during day 1–14 of the menstrual cycle. Eight to 10 aspirations were performed per breast from two anesthetized sites in the upper outer and upper inner quadrant (47). For women without prior cancer, both breasts were aspirated. For women with a prior history of treated ductal carcinoma in situ or invasive cancer in one breast, only the contralateral breast was aspirated. Aspirated cells were immediately expressed into a 15-ml tube of cold RPMI 1640 and kept in an ice bath until processed. Cells from both breasts were pooled as had been done in our prior prospective study. Approximately one-third of the aspirate was used for cytology, and two-thirds were used for the remaining biomarker studies. After the aspiration, cold packs were applied to the breasts for 10 min, and then the breasts were bound in kerlex gauze for 12–24 h (47).

Cytologic Assessment.

Slides for cytology were prepared by filtration and pap staining and classified as nonproliferative, hyperplasia, or hyperplasia with atypia (56). Cytology preparations were also given a semiquantitative index score by assigning 1–4 points to each of six morphological characteristics (cell arrangement, pleomorphism, number of myoepithelial cells, anisonucleosis, nucleoli, and chromatin clumping) according to the method of Masood et al.(57). In general, nonproliferative cytology scored between 6 and 10, hyperplasia scored 11–14, and hyperplasia with atypia scored 15–18. Apocrine metaplasia does not lend itself to this scoring system because some nuclear features of apocrine metaplasia would result in samples scoring in the proliferative range. Because apocrine metaplasia is considered nonproliferative, samples with only apocrine metaplasia were arbitrarily assigned a score of 8, which is in the middle of the nonproliferative range. Morphological category and Masood scores were assigned separately, i.e., Masood scores were not used to determine morphological category assignment. A single cytopathologist (C. M. Z.) assigned cytologic category and semiquantitative Masood scores to all entry and exit specimens within 1 month of aspiration. These values were analyzed for the primary end point. Before unblinding of the study, all slides were comprehensively reviewed by the same cytopathologist (C. M. Z.) without knowledge of treatment assignment or whether the slide was a baseline or 6 month specimen. Slides were likewise reviewed in a blinded fashion by a second cytopathologist (S. M.). Slide sets for six subjects were not available for comprehensive review because either a baseline or repeat aspiration slide had been broken in transit.

Immunocytochemistry.

Slides were prepared for immunocytochemistry by cytospin technique (58). Slides for EGFR and PCNA were fixed in 10% neutral buffered formalin; acetone fixation was used for p53. EGFR was detected with F-7 monoclonal antibody (Sigma Pharmaceuticals), PCNA was detected with P10 antibody (Dako), and p53 was detected with Pab240 (Oncogene Science). PCNA was scored as positive only if strong or granular (replicon-associated) nuclear staining was observed because these patterns are better correlated with other proliferation markers as well as disease-free survival in cancer treatment trials (59, 60, 61, 62, 63, 64, 65, 66). Immunocytochemistry slides were scored by a weighted immunocytochemistry scoring system (67). For immunocytochemistry scoring, technicians selected the most intensely stained area on the slide for assessment, and 100 cells were manually counted. Biesterfeld et al.(66) have demonstrated that selecting the most intensely stained area for ER, progesterone receptor, MIB-1, and PCNA in breast cancer had a greater prognostic/predictive significance than using a mean value for all viewing fields. Each cell assessed was scored as exhibiting 0–4+ stain intensity. The weighted score was calculated from the proportion of cells out of 100 exhibiting each stain intensity. Thus, if 20% of cells counted exhibited 0 staining, 50% of cells counted exhibited 1+ staining, and 30% of cells counted exhibited 2+ staining, the weighted intensity score would be 1.1. Two technicians scored each slide independently. In cases of disagreement, scores were averaged. In general, specimens that exhibited ≥2+ intensity staining of ≥10% of epithelial cells were considered to show evidence of antigen expression.

Blood Hormones, IGF-I, and IGFBP-3.

Serum for assay of estradiol, estrone, FSH, progesterone, IGF-I, and IGFBP-3 was obtained at the time of baseline aspiration or at study entry. All hormone, IGF-I, and IGFBP-3 specimens were obtained during day 1–14 of the menstrual cycle for premenopausal women. FSH specimens were obtained between day 3 and 5 of the menstrual cycle for premenopausal women. Serum was transferred to plastic tubes and stored at −80°C until analysis by radioimmunoassay for hormones and by two-site immunoradiometric assay for IGF-I and IGFBP-3 at Midwest Research Institute.

Urine Polyamine Levels.

First morning urine was collected at 0, 2, 4, and 6 months (end of study) and frozen at −80°C until analysis (Dr. Lawrence Demers, Pennsylvania State University Medical Center, Hershey, PA) of polyamine levels by reverse-phase high-performance liquid chromatography with fluorescence detection after derivitization with dansyl chloride (68, 69, 70).

Plasma DFMO Levels.

Plasma was collected at 2, 4, and 6 months and frozen at −80°C until analysis by reverse-phase liquid chromatography using UV detection (71) at Midwest Research Institute.

ACR Scores of Mammograms.

Standard assessment of mammograms was performed by one radiologist (C. H. J. C.), who categorized each mammogram using the 5-point scoring system recommended by the ACR (65). Under this system, scores of I and II are assigned to nonsuspicious mammograms, a score of III is assigned to mammograms with suggested short interval follow-up, a score of IV denotes a suspicious abnormality that should be biopsied, and a score of V indicates a lesion highly suggestive of malignancy.

Mammographic Density.

For calibration purposes, all mammograms were performed on a rhodium target unit with a stepwedge in place. Films were digitized on a Lumiscan 85 film digitizer and then assessed (N. F. B.) for the percentage of area of increased density relative to the entire breast area in a blinded fashion as described previously (72).

Audiometry.

Standard audiometry evaluation was performed prestudy and poststudy and as clinically indicated after a complaint of hearing loss. Hearing thresholds at both air and bone conduction were determined for 0.5, 1, 2, 3, and 4 kHz; air only was determined at 6 and 8 kHz.

QOL.

Patient-rated QOL was assessed at 0, 2, 4, and 6 months with Health Survey SF-36 (73, 74), which is used frequently in clinical trials and contains two summary scales reflecting QOL related to physical health and mental health (75). The two scales are aggregates of factor score coefficients on the eight subscales of Health Survey SF-36 (i.e., physical functioning, role function-physical, bodily pain, social functioning, emotional well-being, role function-emotional, vitality, and general health perceptions) based on data representative of the general population of the United States. They are standardized to have a mean of 50 and a SD of 10 in the general population, with higher scores representing more positive QOL. General population norms are available, and psychometric adequacy has been demonstrated in a variety of medical populations (75).

Statistical Methods.

The primary end point of the study was a comparison between placebo and DFMO of improvement in cytology assessed by both cytology category and a semiquantitative cytology index score, using the original readings of the primary cytopathologist (C. M. Z.). We assumed a possible 30% improvement in cytology pattern in the placebo group between baseline and reaspiration 6 months later. Given this assumption, accrual of 100 evaluable subjects would allow us to detect a 60% (doubling) improvement in cytology category for the DFMO group with 80% power and a type I error rate of 0.05. This same number of subjects gives us 80% power to detect a 5-point reduction in mean cytology index in the DFMO group compared with no reduction in the placebo group, assuming a SD of 10 points with a type I error rate of 0.05. We targeted 120 eligible subjects to ensure 100 evaluable subjects for the main study end point, as specified by the NCI contract. Subjects were stratified by cytology (hyperplasia with versus without atypia) and menopause status (premenopausal versus postmenopausal). This resulted in four groups with the indicated number of subjects in each: (a) atypia/premenopausal (36); (b) without atypia/premenopausal (33); (c) atypia/postmenopausal (24); and (d) without atypia/postmenopausal (26). Subjects were randomized within each stratum to placebo or DFMO in a 1:1 ratio, with the preparation of the blinded randomization codes performed by an outside statistical consultant.

Quantitative baseline variables were summarized by means, and categorical baseline variables were summarized by frequencies. Intent-to-treat principles were used throughout, and for analysis of primary end points, no change was imputed for cases of missing data. For secondary end points and exploration of possible correlations, only available data were analyzed, and no adjustments were made for multiple comparisons. The primary end point, cytologic improvement from baseline, was summarized by the mean change in index score and by the proportion of subjects who exhibited a cytologic improvement by treatment group. For purposes of analysis, improvement was defined as a change in morphological category (e.g., atypia to hyperplasia or hyperplasia to nonproliferative) or a decrease in the Masood score of 3 or more points. The criterion of 3 points was based on our previous experience that a trained cytopathologist was unlikely to vary by more than 2 points when blindly rereading the same slides. Secondly, it was considered that a decrease of 3 points would typically be accompanied by a change in morphological category and would therefore be clinically relevant. Fisher’s exact test was used to compare the proportion of subjects who exhibited a cytologic improvement between the two groups. Change in Masood score was summarized by its mean for each treatment group, and this change was compared via a two-sample t test. The proportion of subjects in each group showing improvement in ACR score was compared by Fisher’s exact test. Changes in other quantitative measures including PCNA, EGFR, and p53 intensity score and breast density were summarized and compared. Side effects were summarized by the percentage of subjects who had a grade 2 or higher adverse event in each treatment group, by specific events. Fisher’s exact test was used to compare the proportion of grade 2 or higher toxicities between the two treatment groups. Polyamine levels were summarized by means over four time periods, 0, 2, 4, and 6 months. Mixed linear regression was used to compare these levels over time, assuming an autoregressive correlation structure, between the two treatment groups, adjusting for baseline levels. Similar analyses were performed on QOL measures.

Accrual To Study.

Four hundred and one women potentially willing to enter the study underwent random periareolar FNA. Of these, 217 women were potentially eligible on the basis of cytology and biomarkers. Ninety-eight were ineligible on the basis of other study criteria or declined participation. One hundred and nineteen subjects were randomized beginning in June 1997. Accrual took place over 23 months, and total time from entry of first subject to exit of last subject was 29 months.

Baseline Characteristics.

Baseline characteristics in both the placebo and DFMO groups are listed in Table 1. Mean age for the placebo and DFMO groups was similar at 46 and 47 years, respectively. Six minority, non-Caucasian women (5%) were enrolled in the study, and there was a slightly higher proportion of these women in the DFMO group. Fifty-eight percent of entrants were premenopausal, and 42% were postmenopausal. Sixty-two percent of postmenopausal subjects received hormone replacement during study, thus 84% of entrants had a high estrogen milieu at baseline, as reflected by a mean entry FSH of 17 mIU/ml (ranging from the limit of detection of 1.5 to 96.9 mIU/ml) and a mean estradiol level of 75 pg/ml (ranging from the limit of detection of 5 to 346 pg/ml) for the entire cohort. The mean area of the breast considered to be mammographically dense was 46% (range, 0–85%). Fifty percent of subjects had hyperplasia with atypia in their baseline FNA. Median 10-year calculated Gail risk was 3.9% and 4.6% for the placebo and DFMO subjects, respectively. There were no statistically significant differences between the placebo and the DFMO groups for any of the baseline variables shown in Table 1. Furthermore, there was no difference in baseline variables, with the exception of Masood score, between women with atypia and those without atypia. Nor was there a difference, with the exception of serum FSH levels, between premenopausal and postmenopausal women.

Retention on Study.

Five subjects did not complete the study. One had breast cancer detected while on study (randomized to DFMO), one requested to go off study because of toxicity (randomized to placebo), two were noncompliant regarding required clinic visits (one each from the placebo group and the DFMO group), and one never received study medication (randomized to placebo). The remaining 114 subjects completed the study and underwent repeat FNA. Thus, cytologic morphology prestudy and poststudy, our main study end point, was available for 96% of entered subjects.

Self-reported Study Medication Compliance.

With the exception of the five subjects mentioned above, compliance as indicated by self-reported consumption of study medication was excellent, with mean compliance values of 96% and 97% for the placebo and DFMO groups, respectively. Only one subject, randomized to DFMO, reported consuming <75% of study medication and thus was considered to be noncompliant.

Plasma DFMO Levels.

Because of the potential unreliability of self-reported compliance, plasma levels of DFMO were measured at 2, 4, and 6 months (for the DFMO group only) and used as a secondary measure of compliance. Mean plasma DFMO levels were 6.73 μg/ml (0.028 mm) at 2 months, 6.17 μg/ml at 4 months, and 6.53 μg/ml at 6 months. Values ranged from the limit of detection (0.5 μg/ml) to 18.3 μg/ml. For the three time points studied, there were 21 total instances in 16 subjects when no DFMO was detectable, including 8 subjects with no detectable DFMO at 6 months. However, five of these eight subjects had at least a 25% reduction in polyamines at 6 months. Thus, at most 3 subjects (of 60 subjects in the DFMO group that completed the study) could be considered as noncompliant on the basis of both DFMO level (below limit of detection at 6 months) and total polyamine levels.

Urine Polyamines.

Overall, means for total urine polyamines were lower in the DFMO group than the placebo group when adjusting for baseline level and time (P ≤ 0.0001). A 28% reduction in the average total polyamine level in the DFMO group versus an 8% reduction in the placebo group was observed by 2 months on study (Fig. 3). However, decreases were observed only in putrescine (29%) and spermidine (24%; P < 0.0001). No significant differences were noted between DFMO and placebo for spermine (P = 0.81) or spermidine:spermine ratio (P = 0.052) when adjusting for time and baseline values. There were no correlations between changes in urine polyamines and changes in biomarkers.

Breast Cytologic Change.

A mean baseline cytology index score of 13.4 was observed for both subjects randomized to DFMO and subjects randomized to placebo. Overall, there was little change in the cytology index score between the prestudy and poststudy FNAs. The mean index score change was −0.46 (SD = 2.76) for the placebo group and −0.47 (SD = 2.32) for the DFMO group (Table 2). There was likewise no difference between placebo and DFMO groups in the proportion of subjects who were judged to show cytologic improvement either by reduction in cytologic category or a drop of 3 or more Masood score index points, using an intent-to-treat analysis and imputing no change for missing data. The proportion of subjects exhibiting a reduction in cytologic category was 28% (95% confidence limits of 16–40%) for placebo and 27% (95% confidence limits of 16–39%) for DFMO. The proportion of subjects exhibiting a drop of 3 or more Masood score index points was 18% (95% confidence limits of 8–27%) for placebo and 16% (95% confidence limits of 7–25%) for DFMO. There was no correlation between improvement in cytology and reduction in urinary polyamines in the DFMO group.

For the blinded reassessment by the original cytopathologist (C. M. Z.) or the second reviewer (S. M.), there were also no differences between placebo and DFMO groups, although there was intra- and interobserver discordance in interpretations. Intraobserver (first reading versus review by C. M. Z.) discordances of 22% and 25% were obtained for the primary question of whether cytologic improvement was observed by change in category or by a 3-point change in the Masood score, respectively. Interobserver (review by C. M. Z. versus review by S. M.) discordances of 26% and 23% were similarly obtained. Interobserver discordance for benign cytology category in our study compares favorably with similar analyses reported by Sidawy et al.(76).

Change in Breast Proliferation and Other Molecular Markers.

There was no difference between placebo and DFMO groups in change in PCNA weighted intensity score between the prestudy and poststudy FNAs. The mean change in PCNA weighted intensity score was −0.12 (SD = 0.65) for the placebo group and 0.14 (SD = 0.70) for the DFMO group (Table 2). However, there was a significant difference (P = 0.017) between the placebo and DFMO groups in the change in the percentage of cells judged to have PCNA expression. The mean change in the percentage of cells expressing PCNA was −10% (SD = 28%) for the placebo group and 7% (SD = 38%) for the DFMO group, favoring the placebo group. Change in the percentage of cells expressing PCNA (but not weighted index score) was correlated with a change in mammographic breast density. No significant differences between placebo and DFMO groups were noted for change in EGFR or p53 expression as assessed by weighted intensity score (Table 2).

Mammographic Change.

With a mean baseline mammographic breast density area of 46.1% and 45.6% for placebo and DFMO groups, respectively, the mean change between the prestudy and poststudy mammograms in the area considered to be dense was −2.0% and −2.7% for placebo and DFMO groups, respectively, and was not different (Table 2). There was a significant difference in the percentage of subjects who showed improvement in categorical ACR scores, favoring the DFMO group (P = 0.017), using analysis of either all subjects, imputing no change for missing data [2 of 57 (4%) for placebo versus 11 of 62 (18%) for DFMO], or only those for whom prestudy and poststudy mammograms were available [2 of 54 (4%) for placebo versus 11 of 59 (19%) for DFMO]. No correlations were observed between change in individual polyamine levels and/or spermidine:spermine ratio and the change in breast density or change in ACR categorical scores.

Serum IGF-I/IGFBP-3.

For serum levels of IGF-I, IGFBP-3, or ratio of IGF-I to IGFBP-3, there was no difference between placebo and DFMO groups for prestudy values, poststudy values (Fig. 4), or change between the prestudy and poststudy values (Table 2). There was no correlation between individual changes in serum IGF-I or IGFBP-3 levels or IGF-I:IGFBP-3 ratios and changes in serum estradiol or urinary polyamine levels. Change in serum IGFBP-3 was negatively correlated with change in breast tissue PCNA (r = −0.23; P = 0.033).

Serum Hormones.

There were no statistically significant differences between placebo and DFMO groups for prestudy or poststudy serum levels of estradiol, estrone, FSH, and progesterone (Fig. 5). There was, however, a significant difference (P = 0.028) between placebo and DFMO groups for change in estradiol levels between the prestudy and poststudy values (Table 2), with mean values in the placebo group decreasing from 80.7 to 50.6 pg/ml, whereas those in the DFMO group increased slightly from 69.8 to 75.9 pg/ml (Fig. 5). A similar decrease was observed for estrone in the placebo group, but this was not statistically significantly different (P = 0.076) from the change in the DFMO group. There were no differences between the placebo and DFMO groups in prestudy and poststudy changes in FSH and progesterone levels (Table 2).

Adverse Events.

There were no differences in the incidence or intensity of adverse events in the placebo versus DFMO groups, despite a reduction in urine polyamines in the latter. Because administration of DFMO at high doses is known to be associated with an increased incidence of tinnitus, reversible hearing loss, and GI effects such as dyspepsia, nausea, diarrhea, and flatulence, these types of events were of particular interest.

Grade 2 or greater GI effects, reported tinnitus, or measured hearing loss were the same for placebo and DFMO groups (Tables 3 and 4). GI toxicity (nausea, dyspepsia, flatulence, and diarrhea) occurred in 32% of subjects randomized to placebo and 34% of subjects randomized to DFMO. Similarly, 25% of subjects randomized to placebo reported tinnitus or had a ≥10-dB hearing loss, as did 27% of subjects randomized to DFMO. Study agent was discontinued when grade 2–3 tinnitus occurred. This generally resolved in 2–3 days, and study agent was reinstituted. Only four subjects (6.5%) in the DFMO group and two subjects (3.5%) in the placebo group had a ≥15-dB shift in the poststudy bone threshold audiogram (Table 4). The 10- and 15-dB shifts in threshold are criteria proposed by Shotland et al.(77) for defining audiotoxicity in the context of a chemoprevention trial. There was no association between auditory toxicity and prestudy urinary polyamine levels, poststudy polyamine levels, change in polyamine levels, or plasma DFMO levels.

QOL.

There were no significant differences in QOL scores between the DFMO and placebo group. Means across the four assessment points on the Health Survey SF-36 Physical Component Summary scale ranged from 49.5 to 51.2 (SD = 9.0–9.5) for the DFMO group and from 50.7 to 52.1 (SD = 7.9–9.1) for the placebo group. On the Mental Component Summary scale, means ranged from 49.3 to 52.4 (SD = 8.3–11.5) for the DFMO group and from 48.6 to 51.3 (SD = 8.8–11.4) for the placebo group across the four assessment points. These means are comparable with those for the general population of women ages 45–54 years in the United States (Physical M = 49.0, Mental M = 50.1).

Mixed-model analyses of variance, controlling for baseline values on the dependent variables, were conducted on the two Health Survey SF-36 scales, with study arm and menopausal status as categorical independent variables and assessment point (time) as a within-subjects variable. No significant main effects emerged for study arm. A significant arm × time interaction emerged on the physical health scale (P = 0.018). Follow-up analyses revealed that women receiving DFMO reported lower QOL related to physical health than did women receiving the placebo, but only at the 4-month assessment (P = 0.028). Analysis of the mental health scale yielded a significant arm × menopausal status interaction (P = 0.0078). Mental health-related QOL was lower in the DFMO arm than in the placebo arm, but only for the postmenopausal subjects.

Cancer Development.

Including the one cancer that was detected while the subject was on study, four breast cancers have subsequently been detected in the DFMO group, and five breast cancers have subsequently been detected in the placebo group. Five subjects developing breast cancer had atypia (Masood scores of 16, 17, 15, 15, and 16) in their baseline FNA, and four had hyperplasia without atypia (Masood scores of 11, 10, 11, and 11) in their baseline FNA. Median follow-up time since randomization is 39 months. Subjects with baseline atypia developed cancer within 6–16 months of aspiration, and those without atypia developed cancer within 23–50 months of aspiration (Fig. 6). By Cox regression analysis, there was no significant association between subsequent cancer detection and baseline cytology category, polyamine levels, Masood score index, or 10-year Gail risk; however, too few cancers have developed for a meaningful analysis.

DFMO at a dose of 0.5 g/m2/day for 6 months produced modest decreases in total urine polyamines but did not significantly modulate any of the breast cancer risk biomarkers selected for study relative to placebo including breast hyperplasia with or without atypia (assessed by improvement in cytologic category or semiquantitative index score), breast epithelial cell proliferation (measured by PCNA expression), mammographic breast density area, or serum IGF-I and IGF-I:IGFBP-3 ratio.

The most probable reason for lack of a modulation of cytology and proliferation index was that a dose of DFMO sufficient to cause reduction in the spermidine:spermine ratio was not delivered. We did not measure breast DFMO levels, but the mean plasma DFMO level of 0.028 mm was below the 0.1 mm level consistently associated with reduced spermidine:spermine ratios and suppression of proliferation (3, 78). Furthermore, it is below the 1 mm level associated with reduction of spermine and enhanced apoptosis in in vitro studies (3, 79). Improvement in cytomorphology in women with benign breast disease is likely to require both a reduction in proliferation and an increase in apoptosis (80, 81). Exogenous polyamine availability from diet and colonic bacterial production, as well as high estrogen levels, may also counteract the ability of DFMO to suppress proliferation at low to moderate dose levels (82, 83, 84, 85). Dietary polyamine restriction and antibiotics are often used to enhance the effects of DFMO in animal studies (83, 84, 85), but we did not use such strategies. Furthermore, 84% of our subjects were either premenopausal or taking HRT.

Even today, the minimal dose level of DFMO required for effect in breast tissue is unknown. The 0.5 g/m2/day dose of DFMO used was that selected by the NCI Chemoprevention Branch based on its excellent safety profile after 6–12 months of administration (10, 13, 15) and its ability to reduce colonic putrescine and spermidine:spermine ratios (10, 13). Moreover, the minimal dose for biological effect may be tissue and disease specific (11). Subsequent to initiation of our study, Carbone et al.(86), using oral DFMO at an identical dose, reported no effects on skin spermidine levels, although putrescine was reduced. Likewise, Brenner et al.(87) found no reduction in esophageal polyamines or the proportion of cells expressing Ki-67/MIB-1 in a 6-month study of 0.5 g/m2/day DFMO versus placebo in subjects with Barrett’s esophagus. Mitchell et al.(88) found significant decreases in the cervical spermidine:spermine ratio after 1 month of DFMO at a dose of 1 g/m2/day but found no significant decreases at the 0.5 g/m2/day level. Because the current study was conducted under a Phase II NCI contract, no preliminary study was performed to document the effects of 0.5 g/m2/day DFMO on breast tissue polyamines or other biomarkers before initiation.

Likewise, the upper nontoxic dose limit for chronic DFMO administration is not firmly established. We did not observe adverse auditory, GI, or other events more frequently for subjects randomized to DFMO than for those randomized to placebo, similar to results reported by Meyskens et al.(10) using the same dose of 0.5 g/m2/day. Given the excellent tolerability of chronic DFMO in doses of 0.6–1.0 g/m2/day (15), doses in this range could be explored for effects on breast polyamines as well as risk biomarkers. DFMO has a short half-life of 3.5 h, and twice daily dosing would be another option for more consistent polyamine suppression (89).

Although inadequate dose is the most likely explanation for the lack of observed effect of DFMO on our main response indicators, another possible reason is that a portion of the study cohort may not have exhibited elevated ODC activity in breast tissue. A precedent is the situation in breast cancer treatment where antihormonal agents generally do not produce responses in the absence of ER-α expression (90). Subjects selected on the basis of elevated ODC expression in their hyperplastic epithelium might have been more responsive to DFMO, even at low doses. However, a specific antibody for detection of ODC expression by immunohistochemistry (36) was not generally available when our study was initiated in 1997, and other means of assessment of ODC activity or polyamine levels were not feasible with the small amounts of tissue available from FNA.

Although the biomarkers we selected as response indicators had previously been shown to be associated with short-term or longer-term risk, other biological processes affected by DFMO and potentially important for breast neoplastic promotion and progression were not specifically assessed. ODC activity is associated with tyrosine phosphorylation of key signaling proteins in the mitogen-activated protein kinase pathway and release of matrix metalloproteinases (30, 91). Measurement of phosphorylated activated mitogen-activated protein kinase (pERK1/2) and metalloproteinases such as matrix metalloproteinase 2 might be a more sensitive indicator of the biological activity of low-dose DFMO than proliferation markers (36, 88, 91, 92).

Finally, given the modest sample size used in this study, physiological, sampling, and assessment variability may have contributed to the negative outcome. These are common problems in trials in which biomarkers are used as the primary outcome measures (67, 93) and attest to the need to include a blinded placebo control in such studies. What measures could be taken to reduce these sources of variability?

Repeated sampling of the same area of the breast with core biopsies directed at a mammographically dense area and ductal lavage are two possibilities to reduce sampling variability. Historically, prevention trials with repeated core biopsies have not accrued well in the United States (94), although a small successful trial using core biopsies in high-risk women has recently been reported from the United Kingdom (95). There is little information on whether sampling or interpretive discordance would be less of a problem using this technique. Ductal lavage offers the theoretical advantage of repeated access to the same site with a less invasive procedure than core biopsy and has been reported as well tolerated (96). However, ductal lavage is generally performed only for those women producing nipple aspirate fluid, which might eliminate 15–40% of potential subjects (96, 97). There is little information on the success rate of repeated cannulation of the same duct after an interval 6 months or more. Given that occult hyperplasia with or without atypia is generally multifocal and multicentric (98, 99), it is possible that modulation of a field effect as assessed by random tissue sampling may have as much or more implication for risk reduction as elimination of a specific focus of atypia from a single duct.

It is difficult to determine how much of the observed variability in cytologic improvement was due to sampling versus interpretive discordance in our study. Interpretive variability is substantial when assessing either cytologic and histological preparations of breast tissue specimens (76, 100). Despite interpretive variability, hyperplasia ± atypia is an established risk factor for breast cancer (48), and tamoxifen, an approved agent for breast cancer risk reduction (1), was found to reduce the incidence of hyperplasia ± atypia in the National Surgical Adjuvant Breast and Bowel Project 1 trial (101). Use of Consensus Panel Criteria (benign, intermediate/atypia, and suspicious) developed to characterize diagnostic FNA biopsies has been suggested as a method of reducing interpretive variability and has the advantage of consolidating borderline hyperplasia/atypical hyperplasia into one category (102). Use of Consensus Panel Criteria has the disadvantage that both nonproliferative and most hyperplasia specimens without atypia are lumped under the benign grouping, resulting in only two groups and little room for change in response to an effective agent. Nuclear morphometry (103) and semiquantitative index scoring (57) such as that used in this study are also potential measures to reduce morphological interpretive variability, but none of these assessment techniques have been fully evaluated in prospective trials as risk biomarkers. Of interest, however, is that there was little variation in the cytology Masood index score in the placebo group (e.g., mean of −0.46 for the pre-poststudy change). If the Masood index score were to be assessed in a prospective trial as predictive of later cancer development/detection in a high-risk population, then change in cytology index score might be a better method for assessment of morphological change than improvement in cytologic category. Alternatively, improvement in cytologic category could continue to be used as a response end point if the number of evaluable subjects/arm was increased to offset the observed variability in the placebo group.

We observed substantial variation in the change in the proportion of epithelial cells expressing the proliferation marker PCNA in the placebo group. Use of Ki-67/MIB-1 has been associated with less interpretive variability than PCNA by some investigators (59, 62, 66), and expression of Ki-67/MIB-1 within hyperplastic foci has been reported as correlating with breast cancer development in a case-control study (49). Furthermore, early reduction in Ki-67/MIB-1 is associated with later clinical response to tamoxifen and other antihormonal agents in breast cancer treatment trials (104, 105, 106, 107). However, the mean proportion of cells staining positive for MIB-1 is only 3–5% in hyperplasia (49, 108), whereas PCNA staining is much more diffuse, with a higher proportion of cells staining positive for benign breast disease (109). Because we were concerned that the number of epithelial cells in the repeat FNA samples might be too low (e.g., <500) for adequate Ki-67/MIB-1 if DFMO was effective in reducing proliferation and/or normalizing morphological pattern, we chose PCNA for this study. However, for the majority of subjects, adequate numbers of cells would have been available for Ki-67/MIB-1 assessment at both baseline and completion of treatment.

Mammographic density is a reflection of the proportion of stroma and epithelial cells relative to fat (110, 111) and varies with the proliferative activity of ductal epithelium and stroma (112). We found little variation in area of density using a computer-based software system (113) for assessing the proportional area considered to be dense when imaging for premenopausal women was performed in the follicular phase of the menstrual cycle, and subjects were imaged on the same equipment at baseline and completion of study (114, 115). Given the importance of high breast density (>60%) as a risk factor (51, 53), low variation of mean area of density, and reduction of density with tamoxifen relative to placebo in small preliminary studies (116, 117), breast density deserves further study as a response end point in Phase II trials.

There was also little variation in serum IGF-I:IGFBP-3 ratios in the placebo group over the course of the study. High serum levels of IGF-I (e.g., >200 mm) and/or IGF-I:IGFBP-3 ratios in the upper tertiles of the normal range have been correlated with breast risk in premenopausal women (50). IGF-I is a powerful breast epithelial cell mitogen, and in premenopausal women, serum IGF-I:IGFBP-3 may be a surrogate for tissue IGF-I bioactivity (118). Although tamoxifen and several other potential preventive agents such as fenretinide reduce serum IGF-I:IGFBP-3 (119), reduction in cancer incidence is not necessarily correlated with drug-induced reductions in IGF-I:IGFBP-3 (120).

Despite the fact that DFMO at a dose of 0.5 g/m2/day did not modulate the risk biomarkers selected, the clinical model performed well in terms of (a) identification of subjects at high short-term risk for breast cancer, (b) subject accrual and retention, and (c) ability to sample tissue for biomarker assessment at both baseline and study completion. Selection of study subjects on the basis of epidemiological risk factors and random periareolar FNA evidence of hyperplasia ± atypia produced a cohort at very high short-term risk as evidenced by in situ or invasive breast cancer in 9 of the 119 subjects at a median follow-up time of 42 months from baseline aspiration (39 months from randomization). This is consistent with findings in our previously reported study, in which 17 of 337 high-risk women with random FNA evidence of hyperplasia with or without atypia had developed in situ or invasive breast cancer at a median follow-up time of 45 months from initial aspiration. Twenty-five percent of subjects screened by random periareolar FNA ultimately entered the study, with 95% of subjects completing the second FNA procedure, attesting to the excellent subject tolerance of repeated tissue sampling in a relatively short time frame using this technique. Accrual was accomplished in a single institution on schedule and is, to our knowledge, the first randomized placebo-controlled Phase II breast cancer chemoprevention study to be completed in the United States using breast tissue biomarkers as an end point. Finally, random periareolar FNA was able to produce sufficient material for cytology and three additional immunocytochemical tests at baseline and follow-up in the majority of subjects.

In conclusion, DFMO at a dose of 0.5 g/m2/day for 6 months is not associated with significant modulation of cytologic pattern, proliferation, mammographic density, or IGF-I relative to placebo. However, we believe that the random periareolar FNA approach to breast tissue acquisition provides a feasible means of obtaining tissue for initial assessment of potential chemopreventive efficacy in women at high risk of developing breast cancer. Modification of cytology index scores, proliferation indices, and mammographic breast density area would appear to be the most useful and clinically relevant markers to chose as response markers and main study end points. However, a final validation of these end points will only be obtained when a trial is performed with an agent that is in fact effective and a statistically significant improvement in these variables is obtained that is correlated with reduction in cancer incidence.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

Supported by contract NO1-CN-65124 from the Chemoprevention Branch, Cancer Prevention Research Program-Cancer Control, Division of Cancer Prevention of the National Cancer Institute, NIH.

3

The abbreviations used are: ER, estrogen receptor; DFMO, α-difluoromethylornithine; FNA, fine-needle aspiration; ODC, ornithine decarboxylase; PCNA, proliferating cell nuclear antigen; EGFR, epidermal growth factor receptor; IGF, insulin-like growth factor; IGFBP-3, IGF binding protein 3; ACR, American College of Radiology; HRT, hormone replacement therapy; FSH, follicle-stimulating hormone; QOL, quality of life; NCI, National Cancer Institute; GI, gastrointestinal; dB, decibel.

Fig. 1.

Schematic of the polyamine synthesis pathways, with blockade by DFMO. PAO, polyamine oxidase; SSAT, spermidine/spermine N1-acetyltransferase.

Fig. 1.

Schematic of the polyamine synthesis pathways, with blockade by DFMO. PAO, polyamine oxidase; SSAT, spermidine/spermine N1-acetyltransferase.

Close modal
Fig. 2.

Flow chart for study design.

Fig. 2.

Flow chart for study design.

Close modal
Fig. 3.

Mean values for urinary polyamines measured at baseline, 2 months, 4 months, and 6 months (poststudy). Total polyamine level is the sum of putrescine, spermidine, and spermine values. Placebo, open symbols; DFMO, closed symbols.

Fig. 3.

Mean values for urinary polyamines measured at baseline, 2 months, 4 months, and 6 months (poststudy). Total polyamine level is the sum of putrescine, spermidine, and spermine values. Placebo, open symbols; DFMO, closed symbols.

Close modal
Fig. 4.

Mean serum levels for IGF-I, IGFBP-3, and the IGF-I:IGFBP-3 ratio, measured at baseline and at completion of study. There were no differences between groups or between pre- and poststudy specimens. Error bars, SDs.

Fig. 4.

Mean serum levels for IGF-I, IGFBP-3, and the IGF-I:IGFBP-3 ratio, measured at baseline and at completion of study. There were no differences between groups or between pre- and poststudy specimens. Error bars, SDs.

Close modal
Fig. 5.

Mean serum levels of progesterone, estrone, estradiol, and FSH measured at baseline and at completion of study. Error bars, SDs.

Fig. 5.

Mean serum levels of progesterone, estrone, estradiol, and FSH measured at baseline and at completion of study. Error bars, SDs.

Close modal
Fig. 6.

Detection of breast cancer in subjects as a function of time after entry onto the study. Subjects with evidence of atypia in their baseline FNA demonstrated breast cancer at earlier times than did subjects with only hyperplasia in the baseline FNA. Triangles indicate censored subjects.

Fig. 6.

Detection of breast cancer in subjects as a function of time after entry onto the study. Subjects with evidence of atypia in their baseline FNA demonstrated breast cancer at earlier times than did subjects with only hyperplasia in the baseline FNA. Triangles indicate censored subjects.

Close modal
Table 1

Baseline characteristics for eligible subjects

Placebo (n = 57)DFMO (n = 62)
Age (yrs) (mean) 46.4 46.7 
Ethnicity (non-white) (frequency) 4% 6% 
Ten-year Gail risk (%) (mean) 4.6 3.9 
Menopausal status (frequency)   
 Premenopausal 61% 55% 
 Postmenopausal, on HRT 25% 27% 
 Postmenopausal, no HRT 14% 18% 
Atypia in baseline FNA (frequency) 51% 50% 
Masood score (mean) 13.4 13.4 
EGFR intensity score (mean) 1.83 1.90 
p53 intensity score (mean) 1.44 1.48 
PCNA intensity score (mean) 1.82 1.72 
PCNA (% of cells expressing) (mean) 32.5 26.7 
Serum estradiol (pg/ml) (mean) 81 70 
Serum FSH (mIU/ml) (mean) 15 19 
Serum IGF-I (ng/ml) (mean) 195 170 
Serum IGFBP-3 (ng/ml) (mean) 3315 3307 
Serum IGF-I:IGFBP-3 molar ratio (mean) 0.22 0.20 
Urine spermine (nmol/mg creatinine) (mean) 0.37 0.43 
Urine spermidine (nmol/mg creatinine) (mean) 7.5 7.9 
Urine putrescine (nmol/mg creatinine) (mean) 20 21 
Area of increased breast density (%) (mean) 46.1 45.6 
Mammogram ACR score (frequency)   
 I 25% 16% 
 II 68% 74% 
 III 7% 10% 
Placebo (n = 57)DFMO (n = 62)
Age (yrs) (mean) 46.4 46.7 
Ethnicity (non-white) (frequency) 4% 6% 
Ten-year Gail risk (%) (mean) 4.6 3.9 
Menopausal status (frequency)   
 Premenopausal 61% 55% 
 Postmenopausal, on HRT 25% 27% 
 Postmenopausal, no HRT 14% 18% 
Atypia in baseline FNA (frequency) 51% 50% 
Masood score (mean) 13.4 13.4 
EGFR intensity score (mean) 1.83 1.90 
p53 intensity score (mean) 1.44 1.48 
PCNA intensity score (mean) 1.82 1.72 
PCNA (% of cells expressing) (mean) 32.5 26.7 
Serum estradiol (pg/ml) (mean) 81 70 
Serum FSH (mIU/ml) (mean) 15 19 
Serum IGF-I (ng/ml) (mean) 195 170 
Serum IGFBP-3 (ng/ml) (mean) 3315 3307 
Serum IGF-I:IGFBP-3 molar ratio (mean) 0.22 0.20 
Urine spermine (nmol/mg creatinine) (mean) 0.37 0.43 
Urine spermidine (nmol/mg creatinine) (mean) 7.5 7.9 
Urine putrescine (nmol/mg creatinine) (mean) 20 21 
Area of increased breast density (%) (mean) 46.1 45.6 
Mammogram ACR score (frequency)   
 I 25% 16% 
 II 68% 74% 
 III 7% 10% 
Table 2

Comparison of quantitative variables as response biomarkers

Difference between prestudy and poststudy valuesP by t test
PlaceboDFMO
NMean (SD)NMean (SD)
Masood cytology index score 52 −0.46 (2.76) 60 −0.47 (2.32) 0.99 
PCNA percent labeled (intensity score ≥2+) cells 43 −10 (28) 46 +7 (38) 0.014 
PCNA expression (weighted intensity score) 43 −0.12 (0.65) 46 +0.14 (0.70) 0.071 
EGFR expression (weighted intensity score) 50 −0.38 (0.51) 52 −0.31 (0.66) 0.57 
p53 expression (weighted intensity score) 42 −0.11 (0.39) 51 −0.07 (0.54) 0.66 
Breast density (% of area considered dense) 50 −1.95 (8.57) 56 −2.73 (7.85) 0.63 
Serum IGF-I (ng/ml) 55 1.4 (79.35) 59 7.85 (59.9) 0.63 
Serum IGFBP-3 (ng/ml) 52 155 (579) 59 127 (627) 0.81 
Serum IGF-I:IGFBP-3 (molar) ratio 52 −0.00014 (0.088) 59 0.00036 (0.066) 0.87 
Serum estradiol (pg/ml) 55 −28.8 (71.0) 59 3.2 (81.9) 0.028 
Serum estrone (pg/ml) 55 −12.3 (81.2) 59 15.8 (85.7) 0.076 
Serum FSH (mIU/ml) 55 4.8 (12.4) 59 2.1 (13.4) 0.27 
Serum progesterone (ng/ml) 52 −0.44 (2.74) 58 −0.76 (3.05) 0.55 
Difference between prestudy and poststudy valuesP by t test
PlaceboDFMO
NMean (SD)NMean (SD)
Masood cytology index score 52 −0.46 (2.76) 60 −0.47 (2.32) 0.99 
PCNA percent labeled (intensity score ≥2+) cells 43 −10 (28) 46 +7 (38) 0.014 
PCNA expression (weighted intensity score) 43 −0.12 (0.65) 46 +0.14 (0.70) 0.071 
EGFR expression (weighted intensity score) 50 −0.38 (0.51) 52 −0.31 (0.66) 0.57 
p53 expression (weighted intensity score) 42 −0.11 (0.39) 51 −0.07 (0.54) 0.66 
Breast density (% of area considered dense) 50 −1.95 (8.57) 56 −2.73 (7.85) 0.63 
Serum IGF-I (ng/ml) 55 1.4 (79.35) 59 7.85 (59.9) 0.63 
Serum IGFBP-3 (ng/ml) 52 155 (579) 59 127 (627) 0.81 
Serum IGF-I:IGFBP-3 (molar) ratio 52 −0.00014 (0.088) 59 0.00036 (0.066) 0.87 
Serum estradiol (pg/ml) 55 −28.8 (71.0) 59 3.2 (81.9) 0.028 
Serum estrone (pg/ml) 55 −12.3 (81.2) 59 15.8 (85.7) 0.076 
Serum FSH (mIU/ml) 55 4.8 (12.4) 59 2.1 (13.4) 0.27 
Serum progesterone (ng/ml) 52 −0.44 (2.74) 58 −0.76 (3.05) 0.55 
Table 3

GI and auditory toxicity

Placebo (n = 57)DFMO (n = 62)
≥Grade 2 GI toxicity 32% 34% 
≥Grade 1 audiotoxicitya (tinnitus ± ≥15-dB shift) 26% 26% 
≥Grade 2 audiotoxicitya (≥30-dB shift) 
Placebo (n = 57)DFMO (n = 62)
≥Grade 2 GI toxicity 32% 34% 
≥Grade 1 audiotoxicitya (tinnitus ± ≥15-dB shift) 26% 26% 
≥Grade 2 audiotoxicitya (≥30-dB shift) 
a

Criteria for audiotoxicity proposed by Shotland et al.(77).

Table 4

Audiometry-measured hearing changes

Placebo (n = 57)DFMO (n = 62)P
Hearing loss ≥10 dB air or bone any frequency 70% 58% 0.17 
Hearing loss ≥10 dB bone conduction any frequency 35% 23% 0.13 
Hearing loss ≥15 dB bone conduction any frequency 3.5% 6.5% 0.68 
Placebo (n = 57)DFMO (n = 62)P
Hearing loss ≥10 dB air or bone any frequency 70% 58% 0.17 
Hearing loss ≥10 dB bone conduction any frequency 35% 23% 0.13 
Hearing loss ≥15 dB bone conduction any frequency 3.5% 6.5% 0.68 

The devotion and effort of many individuals in the University of Kansas Medical Center Breast Cancer Prevention Center who contributed to the successful implementation of this project are acknowledged with gratitude.

1
Fisher B., Costantino J. P., Wickerham D. L., et al Tamoxifen for prevention of breast cancer: report of the National Surgical Adjuvant Breast and Bowel Project P-1 Study.
J. Natl. Cancer Inst. (Bethesda)
,
90
:
1371
-1388,  
1998
.
2
Verma A. K. The enzyme-activated irreversible inhibitor of ornithine decarboxylase, dl-α-difluoromethylornithine: a chemopreventive agent.
Prev. Med.
,
18
:
646
-652,  
1989
.
3
Glikman P., Manni A., Demers L., Bartholomew M. Polyamine involvement in the growth of hormone-responsive and -resistant human breast cancer cells in culture.
Cancer Res.
,
49
:
1371
-1376,  
1989
.
4
Suh N., Glasebrook A. L., Palkowitz A. D., Bryant H. U., Burris L. L., Starling J. J., Pearce H. L., Williams C., Peer C., Wang Y., Sporn M. B. Arzoxifene, a new selective estrogen receptor modulator for chemoprevention of experimental breast cancer.
Cancer Res.
,
61
:
8412
-8415,  
2001
.
5
Hobbs C. A., Paul B. A., Gilmour S. K. Deregulation of polyamine biosynthesis alters intrinsic histone acetyltransferase and deacetylase activities in murine skin and tumors.
Cancer Res.
,
62
:
67
-74,  
2002
.
6
Marton L. J., Pegg A. E. Polyamines as targets for therapeutic intervention.
Annu. Rev. Pharmacol. Toxicol.
,
35
:
55
-91,  
1995
.
7
Thompson H. J., Ronan A. M. Effect of d,l-2-difluoromethylornithine and endocrine manipulation on the induction of mammary carcinogenesis by 1-methyl-1-nitrosourea.
Carcinogenesis (Lond.)
,
7
:
2003
-2006,  
1986
.
8
Ratko T. A., Detrisac C. J., Rao K. V., Thomas C. F., Kelloff G. J., Moon R. C. Interspecies analysis of the chemopreventive efficacy of dietary α-difluoromethylornithine.
Anticancer Res.
,
10
:
67
-72,  
1990
.
9
Green J. E., Shibata M. A., Shibata E., Moon R. C., Anver M. R., Kelloff G., Lubet R. 2-Difluoromethylornithine and dehydroepiandrosterone inhibit mammary tumor progression but not mammary or prostate tumor initiation in C3(1)/SV40 T/t-antigen transgenic mice.
Cancer Res.
,
61
:
7449
-7455,  
2001
.
10
Meyskens F. L., Gerner E. W., Emerson S., Pelot D., Durbin T., Doyle K., Lagerberg W. Effect of α-difluoromethylornithine on rectal mucosal levels of polyamines in a randomized, double-blinded trial for colon cancer prevention.
J. Natl. Cancer Inst. (Bethesda)
,
90
:
1212
-1218,  
1998
.
11
Meyskens F. L., Gerner E. W. Development of difluoromethylornithine (DFMO) as a chemoprevention agent.
Clin. Cancer Res.
,
5
:
945
-951,  
1999
.
12
Pegg A. E., Hu R. H. Effect of polyamine analogues and inhibition of polyamine oxidase on spermidine/spermine N1-acetyltransferase activity and cell proliferation.
Cancer Lett.
,
95
:
247
-252,  
1995
.
13
Love R. R., Jacoby R., Newton M. A., Tutsch K. D., Simon K., Pomplun M., Verma A. K. A randomized, placebo-controlled trial of low-dose α-difluoromethylornithine in individuals at risk for colorectal cancer.
Cancer Epidemiol. Biomark. Prev.
,
7
:
989
-992,  
1998
.
14
Creaven P. J., Pendyala L., Petrelli N. J. Evaluation of α-difluoromethylornithine as a potential chemopreventive agent: tolerance to daily oral administration in humans.
Cancer Epidemiol. Biomark. Prev.
,
2
:
243
-247,  
1993
.
15
Loprinzi C. L., Messing E. M., O’Fallon J. R., Poon M. A., Love R. R., Quella S. K., Trump D. L., Morton R. F., Novotny P. Toxicity evaluation of difluoromethylornithine: doses for chemoprevention trials.
Cancer Epidemiol. Biomark. Prev.
,
5
:
371
-374,  
1996
.
16
Wang J. Y., McCormack S. A., Viar M. J., Wang H., Tzen C. Y., Scott R. E., Johnson L. R. Decreased expression of protooncogenes c-fos, c-myc, and c-jun following polyamine depletion in IEC-6 cells.
Am. J. Physiol.
,
265
:
G331
-G338,  
1993
.
17
Bryans M., Harley E., Gilmour S. K. Elevated cellular polyamine levels enhance promoter activity in vivo.
Biochem. Biophys. Res. Commun.
,
226
:
618
-625,  
1996
.
18
Kubota S., Kiyosawa H., Nomura Y., et al Ornithine decarboxylase overexpression in mouse 10T1/2 fibroblasts: cellular transformation and invasion.
J. Natl. Cancer Inst. (Bethesda)
,
89
:
567
-571,  
1997
.
19
Auvinen M. Cell transformation, invasion, and angiogenesis: a regulatory role for ornithine decarboxylase and polyamines?.
J. Natl. Cancer Inst. (Bethesda)
,
89
:
533
-537,  
1997
.
20
Auvinen M., Paasinen A., Andersson L. C., Holtta E. Ornithine decarboxylase activity is critical for cell transformation.
Nature (Lond.)
,
360
:
355
-358,  
1992
.
21
Shah N., Thomas T. J., Lewis J. S., Klinge C. M., Shirahata A., Gelinas C., Thomas T. Regulation of estrogenic and nuclear factor κB functions by polyamines and their role in polyamine analog-induced apoptosis of breast cancer cells.
Oncogene
,
20
:
1715
-1729,  
2001
.
22
Pollard K. J., Samuels M. L., Crowley K. A., Hansen J. C., Peterson C. L. Functional interaction between GCN5 and polyamines: a new role for core histone acetylation.
EMBO J.
,
18
:
5622
-5633,  
1999
.
23
Chakravarti D., LaMorte V. J., Nelson M. C., Nakajima T., Schulman I. G., Juguilon H., Montminy M., Evans R. M. Role of CBP/P300 in nuclear receptor signalling.
Nature (Lond.)
,
383
:
99
-103,  
1996
.
24
Schiltz R. L., Mizzen C. A., Vassilev A., Cook R. G., Allis C. D., Nakatani Y. Overlapping but distinct patterns of histone acetylation by the human coactivators p300 and PCAF within nucleosomal substrates.
J. Biol. Chem.
,
274
:
1189
-1192,  
1999
.
25
Bannister A. J., Kouzarides T. The CBP co-activator is a histone acetyltransferase.
Nature (Lond.)
,
384
:
641
-643,  
1996
.
26
Glikman P. L., Manni A., Bartholomew M., Demers L. Polyamine involvement in basal and estradiol-stimulated insulin-like growth factor I secretion and action in breast cancer cells in culture.
J. Steroid Biochem. Mol. Biol.
,
37
:
1
-10,  
1990
.
27
Bello-Fernandez C., Packham G., Cleveland J. L. The ornithine decarboxylase gene is a transcriptional target of c-Myc.
Proc. Natl. Acad. Sci. USA
,
90
:
7804
-7808,  
1993
.
28
Huber M., Poulin R. Permissive role of polyamines in the cooperative action of estrogens and insulin or insulin-like growth factor I on human breast cancer cell growth.
J. Clin. Endocrinol. Metab.
,
81
:
113
-123,  
1996
.
29
Moshier J. A., Malecka-Panas E., Geng H., Dosescu J., Tureaud J., Skunca M., Majumdar A. P. Ornithine decarboxylase transformation of NIH/3T3 cells is mediated by altered epidermal growth factor receptor activity.
Cancer Res.
,
55
:
5358
-5365,  
1995
.
30
Auvinen M., Paasinen-Sohns A., Hirai H., Andersson L. C., Holtta E. Ornithine decarboxylase- and ras-induced cell transformations: reversal by protein tyrosine kinase inhibitors and role of pp130CAS.
Mol. Cell. Biol.
,
15
:
6513
-6525,  
1995
.
31
DisPasquale A., White D., McGuire J. Epidermal growth factor stimulates putrescine transport and ornithine decarboxylase activity in cultivated human fibroblasts.
Exp. Cell Res.
,
116
:
317
-323,  
1978
.
32
Seiler N., Delcros J. G., Moulinoux J. P. Polyamine transport in mammalian cells. An update.
Int. J. Biochem. Cell Biol.
,
28
:
843
-861,  
1996
.
33
Leveque J., Foucher F., Bansard J. Y., Havouis R., Grall J. Y., Moulinoux J. P. Polyamine profiles in tumor, normal tissue of the homologous breast, blood, and urine of breast cancer sufferers.
Breast Cancer Res. Treat.
,
60
:
99
-105,  
2000
.
34
Chanda R., Ganguly A. K. Polyamines in relation to human breast, rectal and squamous cell carcinoma.
Cancer Lett.
,
39
:
311
-318,  
1988
.
35
Persson I., Adami H. O., Bergkvist L., Lindgren A., Pettersson B., Hoover R., Schairer C. Risk of endometrial cancer after treatment with oestrogens alone or in conjunction with progestogens: results of a prospective study.
Br. Med. J.
,
298
:
147
-151,  
1989
.
36
Manni A., Astrow S. H., Gammon S., Thompson J., Mauger D., Washington S. Immunohistochemical detection of ornithine-decarboxylase in primary and metastatic human breast cancer specimens.
Breast Cancer Res. Treat.
,
67
:
147
-156,  
2001
.
37
Janne J., Poso H., Raina A. Polyamines in rapid growth and cancer.
Biochim. Biophys. Acta
,
473
:
241
-293,  
1978
.
38
Glikman P., Vegh I., Pollina M. A., Mosto A. H., Levy C. M. Ornithine decarboxylase activity, prolactin blood levels, and estradiol and progesterone receptors in human breast cancer.
Cancer (Phila.)
,
60
:
2237
-2243,  
1987
.
39
Manni A., Mauger D., Gimotty P., Badger B. Prognostic influence on survival of increased ornithine decarboxylase activity in human breast cancer.
Clin. Cancer Res.
,
2
:
1901
-1906,  
1996
.
40
Manni A., Grove R., Kunselman S., Aldaz M. Involvement of the polyamine pathway in breast cancer progression.
Cancer Lett.
,
92
:
49
-57,  
1995
.
41
Manni A., Trout D., Verderame M. F., Washington S., Mauger D., Demers L. Effect of α-difluoromethyl-ornithine on the expression and function of the epidermal growth factor receptor in human breast epithelial cells in culture.
Breast Cancer Res. Treat.
,
68
:
139
-146,  
2001
.
42
Abeloff M. D., Slavik M., Luk G. D., Griffin C. A., Hermann J., Blanc O., Sjoerdsma A., Baylin S. B. Phase I trial and pharmacokinetic studies of α-difluoromethylornithine: an inhibitor of polyamine biosynthesis.
J. Clin. Oncol.
,
2
:
124
-130,  
1984
.
43
Abeloff M. D., Rosen S. T., Luk G. D., Baylin S. B., Zeltzman M., Sjoerdsma A. Phase II trials of α-difluoromethylornithine, an inhibitor of polyamine synthesis, in advanced small cell lung cancer and colon cancer.
Cancer Treat. Rep.
,
70
:
843
-845,  
1986
.
44
Loprinzi C. L., Messing E. M. A prospective clinical trial of difluoromethylornithine (DFMO) in patients with resected superficial bladder cancer.
J. Cell. Biochem. Suppl.
,
161
:
153
-155,  
1992
.
45
Meyskens F. L., Kingsley E. M., Glattke T., Loescher L., Booth A. A Phase II study of α-difluoromethylornithine (DFMO) for the treatment of metastatic melanoma.
Investig. New Drugs
,
4
:
257
-262,  
1986
.
46
Kelloff G. J., Crowell J. A., Boone C. W., Steele V. E., Lubet R. A., Greenwald P., Alberts D. S., Covey J. M., Doody L. A., Knapp G. G., et al Clinical development plan: 2-difluoromethylornithine (DFMO).
J. Cell. Biochem. Suppl.
,
20
:
147
-165,  
1994
.
47
Fabian C. J., Kimler B. F., Zalles C. M., Klemp J. R., Kamel S., Zeiger S., Mayo M. S. Short-term breast cancer prediction by random periareolar fine-needle aspiration cytology and the Gail risk model.
J. Natl. Cancer Inst. (Bethesda)
,
92
:
1217
-1227,  
2000
.
48
Page D. L., Dupont W. D. Anatomic markers of human premalignancy and risk of breast cancer.
Cancer (Phila.)
,
66
:
1326
-1335,  
1990
.
49
Shaaban A. M., Sloane J. P., West C. R., Foster C. S. Breast cancer risk in usual ductal hyperplasia is defined by estrogen receptor-α and Ki-67 expression.
Am. J. Pathol.
,
160
:
597
-604,  
2002
.
50
Hankinson S. E., Willet W. C., Manson J. E., Colditz G. A., Hunter D. J., Spiegelman D., et al Plasma sex steroid levels and risk of breast cancer in postmenopausal women.
J. Natl. Cancer Inst. (Bethesda)
,
90
:
1292
-1299,  
1998
.
51
Byrne C., Schairer C., Wolfe J., et al Mammographic features and breast cancer risk: effects with time, age and menopause status.
J. Natl. Cancer Inst. (Bethesda)
,
87
:
1622
-1629,  
1995
.
52
Byrne C., Colditz G. A., Willet W. C., Speizer F. E., Pollack M., Hankinson S. E. Plasma insulin-like growth factor (IGF) I, IGF-binding protein 3, and mammographic density.
Cancer Res.
,
60
:
3744
-3748,  
2000
.
53
Boyd N. F., Lockwood G. A., Bying J. W., et al Mammographic densities and breast cancer risk.
Cancer Epidemiol. Biomark. Prev.
,
7
:
1133
-1144,  
1998
.
54
Baker J. A., Kornguth P. J., Floyd C. E., Jr. Breast imaging reporting and data system standardized mammography lexicon: observer variability in lesion description.
Am. J. Roentgenol.
,
166
:
773
-778,  
1996
.
55
Klemp J., Brady D., Kimler B., Fabian C. J. Incidence of BRCA1/2 germ line alterations in a high risk cohort participating in a Phase II chemoprevention trial.
Eur. J. Cancer
,
36
:
1209
-1214,  
2000
.
56
Zalles C., Kimler B. F., Kamel S., McKittrick R., Fabian C. J. Cytologic patterns in random aspirates from women at high and low risk for breast cancer.
Breast J.
,
1
:
343
-349,  
1995
.
57
Masood S., Frykberg E. R., McLellan G. L., Scalapino M. C., Mitchum D. G., Bullard J. B. Prospective evaluation of radiologically directed fine-needle aspiration biopsy of nonpalpable breast lesions.
Cancer (Phila.)
,
66
:
1480
-1487,  
1990
.
58
Fabian C. J., Kamel S., Zalles C. M., Kimler B. F. Identification of a chemoprevention cohort from a population of women at high risk for breast cancer.
J. Cell. Biochem. Suppl.
,
25
:
112
-122,  
1996
.
59
Keshgegian A. A., Cnaan A. Proliferation markers in breast carcinoma. Mitotic figure count, S-phase fraction, proliferating cell nuclear antigen, Ki-67 and MIB-1.
Am. J. Clin. Pathol.
,
104
:
42
-49,  
1995
.
60
van Dierendonck J. H., Keijzer R., van de Velde C. J., Cornelisse C. J. Clear distribution of the Ki-67 antigen during the cell cycle: comparison with growth fraction in human breast cancer cells.
Cancer Res.
,
49
:
2999
-3006,  
1989
.
61
van Dierendonck J. H., Wijsman J. H., Keijzer R., van de Velde C. J., Cornelisse C. J. Cell-cycle-related staining patterns of anti-proliferating cell nuclear antigen monoclonal antibodies. Comparison with BrdUrd labeling and Ki-67 staining.
Am. J. Pathol.
,
138
:
1165
-1172,  
1991
.
62
Tuccari G., Rizzo A., Muscara M., Giuffre G., Barresi G. PCNA/cyclin expression in breast carcinomas: its relationships with Ki-67, ER, PgR immunostaining and clinico-pathologic aspects.
Pathologica (Genoa)
,
85
:
47
-55,  
1993
.
63
Aaltomaa S., Lipponen P., Syrjanen K. Proliferating cell nuclear antigen (PCNA) immunolabeling as a prognostic factor in axillary lymph node negative breast cancer.
Anticancer Res.
,
13
:
533
-538,  
1993
.
64
Gee J. M., Douglas-Jones A., Hepburn P., Sharma A. K., McClelland R. A., Ellis I. O., Nicholson R. I. A cautionary note regarding the application of Ki-67 antibodies to paraffin-embedded breast cancers.
J. Pathol.
,
177
:
285
-293,  
1995
.
65
Yu C. C., Dublin E. A., Camplejohn R. S., Levison D. A. Optimization of immunohistochemical staining of proliferating cells in paraffin sections of breast carcinoma using antibodies to proliferating cell nuclear antigen and the Ki-67 antigen.
Anal. Cell Pathol.
,
9
:
45
-52,  
1995
.
66
Biesterfeld S., Kluppel D., Koch R., Schneider S., Steinhagen G., Mihalcea A. M., Schroder W. Rapid and prognostically valid quantification of immunohistochemical reactions by immunohistometry of the most positive tumour focus. A prospective follow-up study on breast cancer using antibodies against MIB-1, PCNA, ER, and PR.
J. Pathol.
,
185
:
25
-31,  
1998
.
67
Grizzle W. E., Meyers R. B., Oelschlager D. K. Prognostic biomarkers in breast cancer: factors affecting immunohistochemical evaluation.
Breast J.
,
1
:
243
-250,  
1995
.
68
Poyhonen M. J., Uusitalo U. M., Kari A., Takala J. A., Alakuijala L. A., Eloranta T. O. Urinary excretion of polyamines: importance of circadian rhythm, age, sex, menstrual cycle, weight, and creatinine excretion.
Am. J. Clin. Nutr.
,
52
:
746
-751,  
1990
.
69
Shinpo K., Fujita K., Maruta K., Teradaira R., Nagatsu T. Comparative measurements of urinary polyamines in early morning and 24-hour urine specimens.
Clin. Chim. Acta
,
131
:
143
-148,  
1983
.
70
Kabra P. M., Lee H. K., Lubich W. P., Marton L. J. Solid-phase extraction and determination of dansyl derivatives of unconjugated and acetylated polyamines by reversed-phase liquid chromatography: improved separation systems for polyamines in cerebrospinal fluid, urine and tissue.
J. Chromatogr.
,
380
:
19
-32,  
1986
.
71
Hu T., Zuo H., Riley C. M., Stobaugh J. F., Lunte S. M. Determination of α-difluoromethylornithine in blood by microdialysis sampling and capillary electrophoresis with UV detection.
J. Chromatogr. A.
,
716
:
381
-388,  
1995
.
72
Boyd N. F., Lockwood G. A., Marten L. J., Knight J. A., Jong R. A., Fishell E., et al Mammographic densities and risk of breast cancer among subjects with a family history of this disease.
J. Natl. Cancer Inst. (Bethesda)
,
91
:
1404
-1408,  
1999
.
73
Ware J. E., Sherbourne C. D. The MOS 36-item short form health survey (SF-36). I. Conceptual framework and item selection.
Med. Care (Phila.)
,
30
:
473
-483,  
1992
.
74
Ware J. E. .
SF-36 Health Survey: Manual and Interpretation Guide
, Health Assessment Lab, New England Medical Center Boston, MA  
1993
.
75
Ware J. E. .
SF-36 Physical and Mental Health Summary Scales: A User’s Manual
, Health Assessment Lab, New England Medical Center Boston, MA  
1994
.
76
Sidawy M. K., Stoler M. H., Frable W. J., Frost A. R., Masood S., Miller T. R., Silverberg S. G., Sneige N., Wang H. H. Interobserver variability in the classification of proliferative breast lesions by fine-needle aspiration: results of the Papanicolaou Society of Cytopathology Study.
Diagn. Cytopathol.
,
18
:
150
-165,  
1998
.
77
Shotland L. I., Ondrey F. G., Mayo K. A., Viner J. L. Recommendations for cancer prevention trials using potentially ototoxic test agents.
J. Clin. Oncol.
,
19
:
1658
-1663,  
2001
.
78
Huber M., Poulin R. Antiproliferative effect of spermine depletion by N-cyclohexyl-1,3-diaminopropane in human breast cancer cells.
Cancer Res.
,
55
:
934
-943,  
1995
.
79
Brune B., Hartzell P., Nicotera P., Orrenius S. Spermine prevents endonuclease activation and apoptosis in thymocytes.
Exp. Cell Res.
,
195
:
323
-329,  
1991
.
80
Mommers E. C., van Diest P. J., Leonhart A. M., et al Expression of proliferation and apoptosis-related proteins in usual ductal hyperplasia of the breast.
Hum. Pathol.
,
29
:
1539
-1545,  
1998
.
81
Mommers E. C., van Diest P. J., Leonhart A. M., Meijer C. J., Baak J. P. Balance of cell proliferation and apoptosis in breast carcinogenesis.
Breast Cancer Res. Treat.
,
58
:
163
-169,  
1999
.
82
Kendra K. L., Katzenellenbogen B. S. An evaluation of the involvement of polyamines in modulating MCF-7 human breast cancer cell proliferation and progesterone receptor levels by estrogen and antiestrogen.
J. Steroid Biochem.
,
28
:
123
-128,  
1987
.
83
Kremmer T., Palyi I., Daubner D., Boldizsar M., Vincze B., Paulik E., Sugar J., Pokorny E., Tury E. Comparative studies on the polyamine metabolism and DFMO treatment of MCF-7 and MDA-MB-231 breast cancer cell lines and xenografts.
Anticancer Res.
,
11
:
1807
-1813,  
1991
.
84
Manni A., Badger B., Martel J., Demers L. Role of polyamines in the growth of hormone-responsive and -resistant human breast cancer cells in nude mice.
Cancer Lett.
,
66
:
1
-9,  
1992
.
85
Leveque J., Burtin F., Catros-Quemener V., Havouis R., Moulinoux J. P. The gastrointestinal polyamine source depletion enhances DFMO induced polyamine depletion in MCF-7 human breast cancer cells in vivo.
Anticancer Res.
,
18
:
2663
-2668,  
1998
.
86
Carbone P. P., Pirsch J. D., Thomas J. P., Douglas J. A., Verma A. K., Larson P. O., Snow S., Tutsch K. D., Pauk D. Phase I chemoprevention study of difluoromethylornithine in subjects with organ transplants.
Cancer Epidemiol. Biomark. Prev.
,
10
:
657
-661,  
2001
.
87
Brenner D. E., Normolle D. P., Kresty L. A., et al Effect of difluoromethylornithine (DFMO) upon cellular proliferation and cyclin D expression in human metaplastic Barrett’s esophagus mucosa.
Proc. Am. Assoc. Cancer Res.
,
43
:
1007
2002
.
88
Mitchell M. F., Tortolero-Luna G., Lee J. J., Hittelman W. N., Lotan R., Wharton J. T., Hong W. K., Nishioka K. Phase I dose de-escalation trial of α-difluoromethylornithine in patients with grade 3 cervical intraepithelial neoplasia.
Clin. Cancer Res.
,
4
:
303
-310,  
1998
.
89
Love R. R., Carbone P. P., Verma A. K., Gilmore D., Carey P., Tutsch K. D., Pomplun M., Wilding G. Randomized Phase I chemoprevention dose-seeking study of α-difluoromethylornithine.
J. Natl. Cancer Inst. (Bethesda)
,
85
:
732
-737,  
1993
.
90
Osborne C. K., Zhao H., Fuqua S. A. Selective estrogen receptor modulators: structure, function, and clinical use.
J. Clin. Oncol.
,
18
:
3172
-3186,  
2000
.
91
Wallon U. M., Shassetz L. R., Cress A. E., Bowden G. T., Gerner E. W. Polyamine-dependent expression of the matrix metalloproteinase matrilysin in a human colon cancer-derived cell line.
Mol. Carcinog.
,
11
:
138
-144,  
1994
.
92
Boiko I. V., Mitchell M. F., Pandey D. K., White R. A., Hu W., Malpica A., Nishioka K., Boone C. W., Atkinson E. N., Hittelman W. N. DNA image cytometric measurement as a surrogate end point biomarker in a Phase I trial of α-difluoromethylornithine for cervical intraepithelial neoplasia.
Cancer Epidemiol. Biomark. Prev.
,
6
:
849
-855,  
1997
.
93
Urban D., Myers R., Manne U., Weiss H., Mohler J., Perkins D., Markiewicz M., Lieberman R., Kelloff G., Marshall M., Grizzle W. Evaluation of biomarker modulation by fenretinide in prostate cancer patients.
Eur. Urol.
,
35
:
429
-438,  
1999
.
94
Fabian C. J., Kimler B. F., Elledge R. M., Grizzle W. E., Beenken S. W., Ward J. H. Models for early chemoprevention trials in breast cancer.
Hematol. Oncol. Clin. N. Am.
,
12
:
993
-1017,  
1998
.
95
Harper-Wynne C., Ross G., Sacks N., et al A pilot prevention study of the aromatase inhibitor letrozole: effects on breast cell proliferation and bone/lipid indices in healthy postmenopausal women.
Breast Cancer Res. Treat.
,
65
:
225
2001
.
96
Dooley W. C., Ljung B. M., Veronesi U., et al Ductal lavage for detection of cellular atypia in women at high risk for breast cancer.
J. Natl. Cancer Inst. (Bethesda)
,
93
:
1624
-1632,  
2001
.
97
Wrensch M. R., Petrakis N. L., Miike R., King E. B., Chew K., Neuhaus J., Lee M. M., Rhys M. Breast cancer risk in women with abnormal cytology in nipple aspirates of beast fluid.
J. Natl. Cancer Inst. (Bethesda)
,
93
:
1791
-1798,  
2001
.
98
Nielsen M., Thomsen J. L., Primdahl S., et al Breast cancer and atypia among young and middle-aged women: a study of 110 medicolegal autopsies.
Br. J. Cancer
,
56
:
814
-819,  
1987
.
99
Bhathal P. S., Brown R. W., Lesueur G. C., et al Frequency of benign and malignant breast lesions in 207 consecutive autopsies in Australian women.
Br. J. Cancer
,
51
:
271
-278,  
1985
.
100
Schnitt S. J., Connolly J. L., Tavassoli F. A., Fechner R. E., Kempson R. L., Gelman R., Page D. L. Interobserver reproducibility in the diagnosis of ductal proliferative breast lesions using standardized criteria.
Am. J. Surg. Pathol.
,
16
:
1133
-1143,  
1992
.
101
Tan-Chiu E., Costantino J., Wang J., Paik S., Bulch C., Wickerham D. L., Wolmark N. The effect of tamoxifen on benign breast disease. Findings from the National Surgical Adjuvant Breast and Bowel Project (NSABP) Breast Cancer Prevention Trial.
Breast Cancer Res. Treat.
,
69
:
210
2001
.
102
Final version: the uniform approach to breast fine-needle aspiration biopsy.
Breast J.
,
3
:
149
-168,  
1997
.
103
Bacus J. W., Boone C. W., Bacus J. V., Follen M., Kelloff G. J., Kagan V., Lippman S. M. Image morphometric nuclear grading of intraepithelial neoplastic lesions with applications to cancer chemoprevention trials.
Cancer Epidemiol. Biomark. Prev.
,
8
:
1087
-1094,  
1999
.
104
Makris A., Powles T. J., Allred D. C., Ashley S. E., Trott P. A., Ormerod M. G., Titley J. C., Dowsett M. Quantitative changes in cytological molecular markers during primary medical treatment of breast cancer: a pilot study.
Breast Cancer Res. Treat.
,
53
:
51
-59,  
1999
.
105
Chang J., Powles T. J., Allred D. C., Ashley S. E., Makris A., Gregory R. K., Osborne C. K., Dowsett M. Prediction of clinical outcome from primary tamoxifen by expression of biologic markers in breast cancer patients.
Clin. Cancer Res.
,
6
:
616
-621,  
2000
.
106
Dowsett M., Dixon J. M., Horgan K., Salter J., Hills M., Harvey E. Antiproliferative effects of idoxifene in a placebo-controlled trial in primary human breast cancer.
Clin. Cancer Res.
,
6
:
2260
-2267,  
2000
.
107
Ellis M. J., Coop A., Singh B., Mauriac L., Llombart-Cussac A., Jaenicke F., Miller W. R., Evans D. B., Dugan M., Brady C., Quebe-Fehling E., Borgs M. Letrozole is more effective neoadjuvant endocrine therapy than tamoxifen for ErbB-1- and/or ErbB-2-positive, estrogen receptor-positive primary breast cancer: evidence from a Phase III randomized trial.
J. Clin. Oncol.
,
19
:
3808
-3816,  
2001
.
108
Allred D. C., Mohsin S. K., Fuqua S. A. Histological and biological evolution of human premalignant breast disease.
Endocr. Relat. Cancer
,
8
:
47
-61,  
2001
.
109
Shrestha P., Yamada K., Wada T., Maeda S., Watatani M., Yasutomi M., Takagi H., Mori M. Proliferating cell nuclear antigen in breast lesions: correlation of c-erbB-2 oncoprotein and EGF receptor and its clinicopathological significance in breast cancer.
Virchows Arch. A Pathol. Anat. Histopathol.
,
421
:
193
-202,  
1992
.
110
Oza A. M., Boyd N. F. Mammographic parenchymal patterns: a marker of breast cancer risk.
Epidemiol. Rev.
,
15
:
196
-208,  
1993
.
111
Gertig D. M., Stillman I. E., Byrne C., Spiegelman D., Schnitt S. J., Connolly J. L., Colditz G. A., Hunter D. J. Association of age and reproductive factors with benign breast tissue composition.
Cancer Epidemiol. Biomark. Prev.
,
8
:
873
-879,  
1999
.
112
Ursin G., Parisky Y. R., Pike M. C., Spicer D. V. Mammographic density changes during the menstrual cycle.
Cancer Epidemiol. Biomark. Prev.
,
10
:
141
-142,  
2001
.
113
Boyd N. F., Byng J. W., Jong R. A., Fishell E. K., Little L. E., Miller A. B., Lockwood G. A., Tritchler D. L., Yaffe M. J. Quantitative classification of mammographic densities and breast cancer risk: results from the Canadian National Breast Screening Study.
J. Natl. Cancer Inst. (Bethesda)
,
87
:
670
-675,  
1995
.
114
White E., Velentgas P., Mandelson M. T., Lehman C. D., Elmore J. G., Porter P., Yasui Y., Taplin S. H. Variation in mammographic breast density by time in menstrual cycle among women aged 40–49 years.
J. Natl. Cancer Inst. (Bethesda)
,
90
:
906
-910,  
1998
.
115
Persson I., Thurfjell E., Holmberg L. Effect of estrogen and estrogen-progestin replacement regimens on mammographic breast parenchymal density.
J. Clin. Oncol.
,
15
:
3201
-3207,  
1997
.
116
Atkinson C., Warren R., Bingham S. A., et al Mammographic patterns as a predictive biomarker of breast cancer risk: effect of tamoxifen.
Cancer Epidemiol. Biomark. Prev.
,
8
:
863
-866,  
1999
.
117
Brisson J., Brisson B., Cote G., Maunsell E., Berube S., Robert J. Tamoxifen and mammographic breast densities.
Cancer Epidemiol. Biomark. Prev.
,
9
:
911
-915,  
2000
.
118
Pollak M. Insulin-like growth factor physiology and cancer risk.
Eur. J. Cancer
,
36
:
1224
-1228,  
2000
.
119
Bonanni B., Johansson H., Gandini S., Guerrieri-Gonzaga A., Torrisi R., Sandri M. T., Cazzaniga M., Mora S., Robertson C., Lien E. A., Decensi A. Effect of low dose tamoxifen on the insulin-like growth factor system in healthy women.
Breast Cancer Res. Treat.
,
69
:
21
-27,  
2001
.
120
Decensi A., Mariani L., Johansson H., et al Role of plasma IGF-1 as a surrogate biomarker of second breast cancer in a prevention trial of fenretinide.
Proc. Am. Assoc. Cancer Res.
,
43
:
821
2002
.