P2-12-27: Simply Adding Together the Diameters of Tumor Foci in Patients with Multicentric or Multifocal Disease Does Not Add Any Additional Prognostic Information: An Analysis from NCIC CTG MA.12 Randomized Placebo-Controlled Trial of Tamoxifen after Adjuvant Chemotherapy in Pre-Menopausal Women with Early Breast Cancer

Background: A common clinical conundrum in breast cancer management is whether pathologic T stage in women with multicentric or multifocal disease should be taken as the diameter of the largest focus or as the sum of all foci in the breast. Most staging systems, such as the American Joint Committee on Cancer (AJCC), simply use the largest tumor focus for staging. We examine here the impact of alternate methods of estimating tumour size including measures of total tumor size, volume and surface area.

Materials & Methods: NCIC CTG MA.12 is a randomized placebo-controlled trial of tamoxifen after adjuvant chemotherapy for pre-menopausal women with early breast cancer. Median follow up is 9.7 years. Pathologically reported patient tumor dimensions for up to 3 foci were utilized to examine the effects of tumor size on Breast-Cancer-Free-Interval (BCFI), defined as the time from randomization until recurrence (defined as first local, regional, distant, or contralateral invasive tumor or DCIS). Tumor size was estimated as 1) pathologic T stage as per AJCC criteria; 2) largest dimension of largest tumor focus (cm); 3) sum of largest dimension(s) of tumor foci (cm); 4) sum of surface area(s) of tumor foci (cm2), and 5) sum of volume of tumor foci (cm3). Step-wise forward unstratified Cox regression was used to assess the different effects of tumor size. Results: This study accrued 672 patients, 43% with T1 tumors, 51% with T2 tumors, and 6% with T3/T4 tumors; 25% were node negative and 56% had 1–3 positive lymph nodes. 75% were locally determined to have hormone receptor positive tumors. A higher number of involved lymph nodes was associated with significantly shorter BCFI (p<0.0001). None of pathologic T stage (p=0.14), largest dimension of largest tumor size (p=0.14), sum of largest dimensions of tumor foci (p=0.24), sum of surface area (p=0.38), and sum of volume of foci (p=0.51) were significantly associated with BCFI. Likewise, lymphovascular invasion (p=0.08), grade (p=0.14), nor administration of anthracycline therapy (p=0.08) were associated with BCFI.

Discussion: In the MA.12 population of pre-menopausal women randomized to either tamoxifen or placebo, the sole factor significantly associated with BCFI was nodal status. No measure of tumor size in unifocal or multicentric/multifocal tumors impacted BCFI. The findings of this mature data set suggest that simply adding together the diameters of tumors in patients with multicentric or multifocal disease did not add any additional prognostic information.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P2-12-27.

P2-12-23: How Should We Assess Tumour Size (T Stage) in Patients with Multicentric/Multifocal Breast Cancer? Results from the NCIC CTG MA.5 Randomized Trial of CEF vs. CMF in Pre-Menopausal Women with Node Positive Breast Cancer

Background

A common clinical conundrum in breast cancer management is whether pathologic T stage in women with multicentric or multifocal disease should be taken as the diameter of the largest focus or as the sum of all foci in the breast. Most staging systems, such as the American Joint Committee on Cancer (AJCC), simply use the largest tumour focus for staging. We examine here the impact of alternate methods of estimating tumour size including measures of total tumour size, volume and surface area.

Methods: NCIC CTG MA.5 is a randomized trial of CEF versus CMF in pre-menopausal women with node positive breast cancer.

Median follow up is 10 years. Pathologically reported patient tumour dimensions for up to 3 foci were utilized to examine the effects of tumour size on Breast-Cancer-Free-Interval (BCFI). BCFI is defined as the time from randomization until recurrence: first local invasive or DCIS, regional, distant, contralateral invasive or DCIS. Tumour size was estimated as 1) pathologic T stage as per AJCC criteria; 2) largest dimension of largest tumour focus (cm); 3) sum of largest dimension(s) of tumour foci (cm); 4) sum of surface area(s) of tumour foci (cm2), and 5) sum of volume of tumour foci (cm3). Step-wise forward unstratified Cox regression was used to assess the different effects of tumour size.

Results: This study accrued 710 patients, 37% with T1 tumours, 52% with T2 tumours and 9% with T3 tumours; 61% had 1 to 3 positive lymph nodes. 59% hormone receptor positive. Higher pathologic T stage (p=0.001) and greater surface area (p=0.02) were associated with shorter BCFI, as was lymphovascular invasion (p=0.03), and # of lymph nodes involved (p<0.0001). Administration of anthracycline therapy led to significantly longer BCFI (0.003). The sum of largest tumour sizes (p=0.33) and sum of tumour volume (p=0.34) were not significantly associated with BCFI. Additionally, when the less complete locally reported tumour grade data were included, higher tumour grade was associated with shorter BCFI (p<0.0001).

Conclusions: Consideration of multicentric and multifocal disease was an important adjunct to standard pathologic tumour size as was estimation of tumour surface area in this chemotherapy trial of node positive premenopausal women. However, simply adding together the diameters of tumours in patients with multicentric or multifocal disease did not add any additional prognostic information in this high risk patient population.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P2-12-23.

P4-12-05: Impact of the Lifestyle Intervention Study in Adjuvant Treatment of Early Breast Cancer (LISA) Weight Loss Intervention upon Physical Activity

Observational evidence shows a relationship between obesity and poor prognosis in breast cancer (BC). Physical activity (PA) is an important component of weight loss and maintenance, but most large-scale interventions in BC patients have produced only modest improvements in activity. We sought to evaluate changes in PA in women participating in LISA, a randomized trial coordinated by the Ontario Clinical Oncology Group which was designed to examine the impact of a telephone-based weight loss intervention (WLI) upon disease free survival in BC patients.

Methods: Participants were randomized 1:1 to the WLI or educational control (EC) group. Eligibility included diagnosis of Stage I-III BC, BMI ≥24 kg/m2, and treatment with letrozole. The WLI, based on the Diabetes Prevention Program, focused on weight reduction through calorie restriction and increased physical activity. Delivery involved 19 calls, mailings and a participant manual. The PA goal was 150 minutes/week. PA was measured using the International Physical Activity Questionnaire (Short Form) at baseline, 6, 12, and 18 months. Changes in time (minutes/week) spent sitting and engaging in moderate, vigorous, and walking activities were compared between groups.

Results 338 women were randomized to WLI (n=171) or EC (n=167) from 20 centers in Canada and the USA. The study was discontinued due to loss of funding. The WLI arm lost significantly more weight than EC arm, with mean weight loss of 6.1% of body weight vs. 0.6% at 12 months (p<0.001). Activity data are presented in Table 1. At baseline, participants were inactive; median vigorous activity was zero minutes/week and median time spent sitting was more than 35 hours/week. WLI participants reported significantly higher participation in vigorous, moderate and walking activities, and lower levels of sedentary behavior, compared to controls at 6 and 12 months. Higher levels of PA were significantly associated with increased weight loss at all time periods. Factors associated with increases in activity included higher baseline BMI (p=0.014), lower baseline activity (p<0.001) and assignment to the WLI arm (p=0.02). Women assigned to WLI increased their PA during the intervention period (p=0.017) even after adjusting for other significant baseline factors.

Conclusion Participants in the LISA WLI reported significantly higher levels of PA compared with controls. Activity increased most in women who were heavier and less active at baseline. Further study of this WLI as a potential means to improve BC outcomes is warranted.

Maximum smoothed likelihood for multivariate mixtures

We introduce an algorithm for estimating the parameters in a finite mixture of completely unspecified multivariate components in at least three dimensions under the assumption of conditionally independent coordinate dimensions. We prove that this algorithm, based on a majorization-minimization idea, possesses a desirable descent property just as any em algorithm does. We discuss the similarities between our algorithm and a related one, the so-called nonlinearly smoothed em algorithm for the non-mixture setting. We also demonstrate via simulation studies that the new algorithm gives very similar results to another algorithm that has been shown empirically to be effective but that does not satisfy any descent property. We provide code for implementing the new algorithm in a publicly available R package.

The chemotherapy metabolite acrolein upregulates thrombin generation and impairs the protein C anticoagulant pathway in animal-based and cell-based models

BACKGROUND

Thrombosis is a common complication in cancer patients receiving chemotherapy regimens that include cyclophosphamide. However, the mechanisms by which these agents increase this risk are largely uncharacterized.

OBJECTIVES

To examine the effects of cyclophosphamide and its metabolite acrolein on procoagulant and anticoagulant pathways in both cell-based and animal-based models.

METHODS

Thrombin and activated protein C (APC) generation were measured in defibrinated plasma exposed to acrolein-treated endothelial and smooth muscle cells. Tissue factor (TF) activity was measured on acrolein-treated cells. Cell surface levels of phosphatidylserine, TF, endothelial protein C receptor and thrombomodulin were measured. Healthy BALB/c mice received injections of saline (control), acrolein, or cyclophosphamide; blood was collected, and plasma thrombin-antithrombin (TAT) complex, protein C and APC levels were analyzed.

RESULTS

Exposure of acrolein-treated endothelial and smooth muscle cells to defibrinated plasma increased thrombin generation in the plasma. This was associated with enhanced phosphatidylserine exposure and/or increased TF activity on acrolein-treated cells. Despite elevated levels of thrombin generation, plasma APC levels were not elevated. In vivo, treatment of mice with cyclophosphamide and acrolein resulted in elevations of plasma TAT complex levels, whereas APC levels remained low.

CONCLUSIONS

This is the first study to examine thrombin generation and the APC pathway in chemotherapy-treated mice. Cyclophosphamide and acrolein appear to upregulate procoagulant pathways, while impairing endogenous anticoagulant pathways. This may explain, in part, the increased risk of thrombosis observed in cancer patients receiving cyclophosphamide-containing chemotherapy.

Abstract P5-10-18: Echocardiography on Neonates Exposed to Anthracycline Therapy In Utero for Maternal Breast Cancer

Background: Anthracycline therapy carries a risk for cardiac toxicity in pediatric and adult populations. It is unknown if this risk also applies to fetuses exposed in utero. The objective of this study is to evaluate the cardiac function of neonates exposed to anthracyclines in utero. Materials and Methods: This was an observational study. Pregnant women diagnosed with breast cancer who underwent anthracycline based chemotherapy and delivered a liveborn infant after 24 weeks, were offered enrollment. Patients provided written consent to obtain doses and dates of chemotherapy treatment. Echocardiograms were offered to all newborns exposed as fetuses to anthracyclines. Women chose to perform this echocardiogram locally or at the institution of the primary investigator. Baseline fetal cardiac evaluation was reportedly normal on ultrasound prior to starting chemotherapy.

Chemotherapy Regimens:

Adriamycin/Cytoxan (n=20).

5-Fluorouracil/Adriamycin/Cytoxan (n=3)

There were no cases of exposure to epirubicin or idarubicin.

Mean dosage of Adriamycin was 57/m2 +/−8.0.

Results: Twenty three neonates participated.

Neonatal echocardiograms done at a mean age of 15 +/−26 months, (range 3 days of life to 9.75 years). RV wall, IVS and LV posterior wall diastolic and systolic thickness measurements were within normal limits for age and BSA. Mean shortening fraction was 41+/−10%. (normal 33 +/− 3%). In all 23, AV and semilunar valves were normal. Great arteries were normal in 22/23, with one case of a small pulmonary artery fistula. At 2 years of age, this was not hemodynamically significant. Patent foramen ovale was noted in 4 cases, all in children less than 2.0 months of age. Two children ages 3 days to 1.4 months had a PDA. One child at 1 year of age had an ASD. No structural or functional heart defects were diagnosed in any of these children.

Discussion:

Cardiac function appears to be preserved in neonates exposed to adriamycin chemotherapy in utero.

Any toxic effects on fetal myocytes was not evident in postnatal echocardiograms.

Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P5-10-18.