Abstract P2-16-01: Relationship between overall survival and surrogate measures in patients with metastatic breast cancer treated with chemotherapy

Background: While metastatic breast cancer (MBC) is considered an incurable disease, nearly one-fifth of patients live longer than five years following diagnosis. In an effort to identify novel agents earlier, surrogate end points of overall survival (OS) such as response or progression are often employed in randomized controlled trials (RCTs). The goal of this analysis is to evaluate patterns of outcome reporting and the relationship of OS with surrogate measures in RCTs of patients with MBC.

Methods: The analysis was based on data from a systematic review of patients with MBC evaluating the clinical impact of chemotherapy intensity on survival. Reports of phase 2-3 RCTs published between 1990-2013 comparing more intense chemotherapy regimens (higher dose intensity or use of additional agents) with less intense were identified. For each RCT, clinical, treatment, demographic and outcome data were extracted. Outcomes evaluated included OS, progression free survival (PFS), and time to progression (TTP) with a focus on median survival and hazard ratios (HRs) as measures of treatment effect. Survival post progression (SPP) was calculated as the difference between median survival and median progression free time. The relations between various outcome measures were estimated utilizing weighted Pearson correlation coefficient (CORR) adjusted by Fisher's transformation. Weights were assigned proportionally to the sample size of individual RCTs.

Results: The review identified 70 eligible RCTs including 15,043 patients with MBC. Average median OS, PFS, and TTP were 19.2, 6.9, and 8.1 months reported in 96%, 60%, and 43% of studies, respectively. Progression could be determined in 66 studies, while 6 RCTs provided both outcomes. TTP was more often utilized in earlier studies (65% in 1990-2000, 35% in 2001-2008, and 33% in 2009-2013) and it was superseded by PFS in later years (20%, 70% and 81%, respectively). Only 37%, 33%, and 11% of RCTs reported HRs for OS, PFS, and TTP, respectively. HRs were more often available in recent publications (20% in 1990-2000, 22% in 2001-2008, 63% in 2009-2013 provided HR for OS). The correlation between reported HR and HR estimated by the ratio of arm-specific median survival times was high for OS (CORR=0.87, 95%CI: 0.73-0.94) and TTP (CORR=0.92, 95%CI: 0.61-0.99) and slightly lower for PFS (CORR=0.72, 95%CI: 0.44-0.87). The relationship between OS and surrogate measures (PFS, TTP) was weaker. The correlation between HR for OS and PFS was 0.49 (95%CI: 0.21-0.69) and for OS and TTP it was 0.26 (95%CI: -0.13-0.58). Survival time following progression was dependent on treatment type and was longer in less intense arms than more intense (mean SPP: 12.4 months vs. 11.4 months, P=0.0155).

Conclusions: In RCTs of patients with MBC treated with chemotherapy, when HR is not reported and if necessary statistical conditions are met, the HR approximated by ratio of median survival times may be a suitable proxy estimate. In agreement with other reports, neither PFS nor TTP are acceptable surrogate outcomes for OS in MBC, as survival following progression may be substantial. In these patients, crossover and post-trial treatments may influence the relationship between OS and surrogate measures.

Citation Format: Culakova E, Poniewierski MS, Crawford J, Dale DC, Lyman GH. Relationship between overall survival and surrogate measures in patients with metastatic breast cancer treated with chemotherapy [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P2-16-01.

The impact of the granulocyte colony-stimulating factor on chemotherapy dose intensity and cancer survival: a systematic review and meta-analysis of randomized controlled trials

BACKGROUND

The granulocyte colony-stimulating factor (G-CSF) is utilized to reduce neutropenic complications in patients receiving cancer chemotherapy. This study represents a systematic review and evidence summary of the impact of G-CSF support on chemotherapy dose intensity and overall mortality.

MATERIALS AND METHODS

All randomized controlled trials (RCTs) comparing chemotherapy with or without G-CSF support and reporting all-cause mortality with at least 2 years of follow-up were sought. Dual-blind data abstraction of disease, treatment, patient and outcome study results with conflict resolution by third party was carried out.

RESULTS

The search revealed 61 randomized comparisons of chemotherapy with or without initial G-CSF support. Death was reported in 4251 patients randomized to G-CSFs and in 5188 controls. Relative risk (RR) with G-CSF support for all-cause mortality was 0.93 (95% confidence interval: 0.90-0.96; P < 0.001). RR for mortality varied by intended chemotherapy dose and schedule: same dose and schedule (RR = 0.96; P = 0.060), dose dense (RR = 0.89; P < 0.001), dose escalation (RR = 0.92; P = 0.019) and drug substitution or addition (RR = 0.94; P = 0.003). Greater RR reduction was observed among studies with longer follow-up (P = 0.02), where treatment was for curative intent (RR = 0.91; P < 0.001), and where survival was the primary outcome (RR = 0.91; P < 0.001).

CONCLUSIONS

All-cause mortality is reduced in patients receiving chemotherapy with primary G-CSF support. The greatest impact was observed in RCTs in patients receiving dose-dense schedules.

Abstract P3-06-07: Ki67 as a Predictive Marker of Response to Neoadjuvant Chemotherapy in Patients with Early-Stage Breast Cancer (ESBC): A Systematic Review and Evidence Summary

Background: Immunohistochemical (IHC) assessment of the proportion of cells staining for the KI67 nuclear antigen is being increasing utilized in the management of patients with early-stage breast cancer (ESBC). A comprehensive systematic review and evidence synthesis of biomarkers potentially predictive of response to systemic therapy was initiated as a part of an NCI-funded comparative effectiveness research program.

Methods: Studies of chemotherapy response prediction based on baseline IHC assessment of Ki67 in patients with ESBC receiving neoadjuvant systemic therapy were identified. Response was specified as pathologic complete response (pCR) or clinical response (ClinR). Assay predictive performance for response was assessed on the basis of sensitivity, specificity, predictive value and predictive odds ratio (POR±95%CLs) utilizing mixed effects models. Study results were fitted in an ROC analysis based on the method of DerSimonian and Laird. Publication bias was evaluated on the basis of funnel plot asymmetry assessed by Egger's regression intercept and Begg and Mazumdar's rank correlation.

Results: Of 469 potentially eligible studies, dual blind full text review identified 42 eligible studies reporting 44 independent cohorts with 6,716 patients (21–979). While Ki67 cutpoints varied considerably, they were most commonly between 10%–30% (median 20%, range 1–50%). The analysis prsented here is limited to the 30 studies of ESBC patients (N = 3,343) receiving neoadjuvant therapy of which 14 reported fewer than 100 patients. The proportion of patients with elevated Ki67 across studies ranged from 0.20–0.92 (median = 0.54). Sensitivity and specificity for treatment response in patients with high vs. low baseline Ki67 was 0.65 [0.61, 0.68] and 0.52 [0.50, 0.54], respectively. Estimated response rates across studies in patients with high vs. low Ki67 were 31% [29%, 34%] and 19% [17%, 21%], respectively. The estimated POR for response across studies was 2.82 [2.14, 3.72; P &lt; .001].

Abstract P1-15-04: The relationship of relative dose intensity and supportive care to febrile neutropenia rates in patients with early stage breast cancer receiving chemotherapy: a prospective cohort study of chemotherapy-associated toxicity

Background: Febrile neutropenia (FN) represents a major dose-limiting toxicity of cancer chemotherapy resulting in considerable morbidity, mortality, and cost. Patients have the highest risk of the initial neutropenic event in cycle 1 when most patients receive full dose chemotherapy. This study evaluates time course of neutropenic events in patients receiving chemotherapy for early-stage breast cancer (ESBC) and supportive care interventions that modify FN risk in ESBC patients treated in actual oncology practice.

Methods: A prospective cohort study of adult cancer patients with solid tumors or lymphoma starting a chemotherapy regimen was conducted at 115 U.S. sites. Toxicities associated with chemotherapy were recorded in up to 4 cycles including severe neutropenia (SN), FN, and infection. Documented clinical interventions included reductions in chemotherapy relative dose intensity (RDI), the use of colony-stimulating factors (CSFs), and antibiotics.

Results: A total of 1202 ESBC patients starting chemotherapy were analyzed, of which 1154, 1099, and 896 reached the midcycle of cycles 2, 3, and 4, respectively. While the majority of neutropenic and infection events occurred in cycle 1, decreasing rates of FN and infection in later cycles correlated with increasing reductions in dose intensity and increased use of CSFs and antibiotics.

P5-20-03: Personalizing Supportive Care: A Clinical Prediction Model for Neutropenic Complications in Patients with Early-Stage Breast Cancer (ESBC) Receiving Intermediate Risk Chemotherapy

Background: Neutropenic complications including severe and febrile neutropenia (FN) represent major dose-limiting toxicities of cancer chemotherapy. A general risk model for neutropenic complications across major solid tumors has been developed and validated (Lyman et al. Cancer 2011). Current guidelines recommend consideration of primary prophylaxis with a colony-stimulating factor (CSF) in patients at &gt;20% risk of FN. The decision for primary CSF prophylaxis in patients on intermediate risk chemotherapy (10-20%) is based on physician assessment of individual patient risk factors for FN. This study assesses the ability of this general FN risk model to identify ESBC patients on intermediate risk chemotherapy who are at a personal high risk for developing a neutropenic complication.

Methods: A prospective cohort study accrued 4458 consenting patients starting a new chemotherapy regimen at 115 randomly selected community oncology practices throughout the United States from 2002–2006. The risk of severe or febrile neutropenia (SNFN) in cycle 1 and across 4 cycles was estimated [±95% CI] utilizing logistic regression analysis and adjusting for key clinical factors including among others: age, prior chemotherapy, abnormal hepatic or renal function, low pretreatment white blood count, immunosuppressive medications, CSF prophylaxis, and planned relative dose intensity as well as major chemotherapeutic agents. The cumulative risk of FN across 4 cycles was also estimated by the product limit method of Kaplan and Meier.

Results: Among 1224 patients with ESBC, 822 received intermediate risk chemotherapy based on National Comprehensive Cancer Network guidelines. Among these patients, cycle 1 SNFN occurred in 37%, at least one episode of FN over 4 cycles of chemotherapy in 17%, with 15% receiving primary CSF prophylaxis. The predicted risk of cycle 1 SNFN ranged from 1%-79%, with mean (median) risk of 33.8% (39.0%). Model performance was good with a c-statistic of 0.73 [0.69−0.76]. Based on this general FN risk model, cycle 1 SNFN occurred in 47% of predicted high risk ESBC patients [42 — 52%] compared to 13% [8-17%] of low risk patients. One or more FN events over 4 cycles occurred in 20% [17-24%] of predicted high risk versus 10% [6-14%] in low risk patients. The cumulative risk of FN by Kaplan-Meier estimation was 23% in high risk and 10% in low risk patients. Model sensitivity and specificity for FN were 83% and 33%, respectively. The majority of SNFN (76%) and FN (58%) events among high risk patients occurred in cycle 1. 50% of high risk patients who did not receive primary CSF prophylaxis went on to receive CSF during subsequent cycles.

Conclusions: Based on good test performance characteristics, this clinical FN prediction model also identifies ESBC patients receiving intermediate risk chemotherapy at high personal risk for FN (FN &gt;20%) over the first 4 cycles of chemotherapy. Half of predicted high risk patients without primary CSF prophylaxis will be given CSF in subsequent cycles after the occurrence of a neutropenic complication. This also confirms previous clinical trial findings that the majority of febrile neutropenic events occur in the first cycle.

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P5-20-03.

P5-13-17: Multigene Signature Assays in Patients with Early-Stage Breast Cancer (ESBC) Receiving Neoadjuvant Chemotherapy: An NCI-Funded Systematic Review and Evidence Summary of Predictive Performance

Background: A comprehensive literature search and evidence synthesis of multigene signatures predictive of response to systemic chemotherapy in patients with breast cancer was initiated as a part of an NCI-funded program on Comparative Effectiveness Research.

Methods: Validation studies were sought of multigene signatures for prediction of chemotherapy response (favorable vs unfavorable) in ESBC patient cohorts different from those used for signature development. Pooled estimates [±95% CI] of assay performance for predicting clinical outcome included sensitivity, specificity, likelihood ratio, predictive value (PV) and predictive odds ratio (POR) utilizing mixed effects models based on the method of Mantel-Haenszel. Exploratory metaregression analyses on log (POR) were also performed. Studies were classified by validation type including cell lines to patients, independent internal sample, random split sample, or external validation. Evidence for publication bias was assessed by Egger's regression intercept and Begg and Mazumdar's rank correction. Results: Dual-blind review of abstracts identified 33 studies of neoadjuvant chemotherapy response of which 29 stratified treatment response by signature classifier category. Classifier development was based on tumor response prediction in 20 studies, prognosis in 5, and molecular classification in 4. The Table shows assay performance measures overall and by study validation type. Assay performance based on the POR was positively associated with overall study quality (P=.015) and journal impact factor (P=.020). However, strong evidence for publication bias was observed based on both regression intercept (P&lt;.001) and rank correlation (P=.005). No significant differences in assay performance were noted for assays originally developed for response prediction (POR=5.3), prognosis (POR=6.6) or molecular classification (P=6.9) (P=.770).

P3-14-04: Assessment of Genomic Prognostic Signatures as Predictors of Response to Neoadjuvant Chemotherapy in Patients with Early Stage Breast Cancer

Background: Based on results from randomized clinical trials, adjuvant and neoadjuvant chemotherapy (NCT) strategies in early stage breast cancer patients (ESBC) achieve comparable long term results. Recently, a number of genomic signatures have been reported, distinguishing patients with low versus high risk of recurrence. While developed primarily as prognostic assays, these classifiers have also been proposed to be predictive of benefit from systemic chemotherapy. Neoadjuvant studies provide an opportunity to evaluate their predictive value for response to NCT.

Methods: A systematic review of gene expression profile studies in ESBC patients receiving chemotherapy was conducted. Medline search of original research articles of human studies published between January 2000 and February 2011 was based on key words and MeSH heading terms. Publications presenting outcomes for chemotherapy treated patients in groups stratified by multi-gene array signatures and utilizing a new independent cohort of patients compared to the original development cohort were selected. Information from eligible studies was extracted by dual abstraction. Reported results were synthesized into combined diagnostic odds ratio (DOR) using method of Mantel-Haenszel. This analysis is restricted to neoadjuvant studies investigating the association of genomic signature prognostic categories with objective tumor response to chemotherapy. Results: A total of 42 articles were eligible for data abstraction. Out of these, 6 publications evaluated response to NCT in good (low risk of recurrence) versus poor prognosis groups based on genomic prediction. Since two of the studies analyzed the same signature on a cohort with large overlap, only 5 studies were included in the final analysis, accounting for n=918 patients. Response consisted of pathologic complete response (pCR) in 3 studies, pCR or minimal residual disease (1 study), and clinical complete response (1 study). Prognostic genomic assays included Oncotype DX (1), MammaPrint (1), Genomic Grade Index (2) and PAM50 Risk of Relapse Score (1). Eight different chemotherapy regimens were utilized. The most common drugs were cyclophosphamide, anthracyclines, taxanes, and 5-fluorouracil. Across all genomic signatures, good prognosis patients, as defined by gene expression data, demonstrated consistently low rates of response to chemotherapy (median 3%, range 0–12%) compared to patients with less favorable prognosis (median 32%, range 19–43%). Odds ratio for response in poor versus good prognosis patients ranged from 3.9 to 21.7 with combined DOR= 6.6 (95% CI 3.9−11.3, P&lt;0.0001). No heterogeneity was determined across studies (P=0.4). The C-statistic estimating assay discriminatory ability was reported in 3 studies ranged from 0.72 to 0.78.

Conclusions: Across all genomic prognostic signatures reported, only a very small proportion of patients with signature predicted good prognosis achieved complete response to NCT. This suggests low sensitivity to chemotherapy among good prognosis patients, as determined by the prognostic genomic signatures. This further confirms the association between poor prognosis tumors and higher responsiveness to chemotherapy.

Funding: NCI: UC2CA14041-01

Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P3-14-04.

Risk factors for in-hospital mortality and prolonged length of stay in older patients with solid tumors

9550

Background: Little is known about risk factors that contribute to prolonged hospitalization and mortality in older patients with cancer. Methods: Cancer patients ≥65 years of age hospitalized between 1995 and 2003 at 133 academic medical centers were evaluated using the University Health System Consortium discharge database. This study identified 386,377 older hospitalized patients with various solid tumors. Multivariate analyses were performed to determine variables independently associated with the primary endpoints: length of stay (LOS) ≥10 days and in-hospital mortality (IHM). Results: Average LOS was 7.5 days with 23% hospitalized ≥10 days. A significant improvement in LOS was observed over the study timeframe (p<.0001). Patients with gastric cancer had the greatest risk of prolonged LOS while those with breast cancer had the lowest risk. Additional risk factors for prolonged LOS included infection, venous thromboembolism and red blood cell transfusion (RBCT). The overall rate of IHM was 7.3% with a significant improvement in risk over the study timeframe (p<.0001). IHM was strongly associated with prolonged LOS (p<.0001). Older patients with primary central nervous system malignancies had the highest rates of IHM (OR=1.81; 95% CI: 1.59–2.07), followed by esophageal and lung cancer. Male gender was a risk factor for both IHM and prolonged LOS (p<.0001). Older African American cancer patients were more likely to experience prolonged LOS and IHM compared with Caucasian patients (p<.0001) after adjustment for cancer type and comorbidities. Additional risk factors associated with IHM included metastatic disease, active infection, neutropenia, renal disease, lung disease, arterial and venous thromboembolism, congestive heart failure, hepatic disease, and RBCT. Conclusions: Improving trends in LOS and IHM for older patients with solid tumors were observed over time in this study. Risk factors associated with IHM such as infection, neutropenia and RBCT, when modified, could potentially further reduce rates of prolonged LOS and IHM in older cancer patients.

[Table: see text]

Acute myeloid leukemia or myelodysplastic syndrome (AML/MDS) and overall mortality with chemotherapy (CT) and granulocyte colony-stimulating factor (G-CSF): A meta-analysis of randomized controlled trials (RCTs)

9524

Background: To evaluate the risk of AML/MDS and overall mortality in patients receiving CT ± G-CSF, a meta-analysis of RCTs were conducted. Methods: Electronic databases searched through August 2008 identified 3,794 articles for initial screening. Eligibility included RCTs of solid tumor or lymphoma patients randomized to CT ± primary G-CSF support, ≥2 years follow-up and reporting AML/MDS or all second malignancies. Pre-specified study categories included: a)same dose/schedule, b)dose-dense or c)dose-escalated CT. Primary outcomes were AML/MDS and mortality. Dual blinded data extraction was performed. Relative risk (RR) and absolute risk difference (ARD) were estimated by Mantel-Haenszel. Results: Median follow-up was 54 months. 12,642 patients were randomized to CT ± primary G-CSF support. Second malignancies were reported in 3.3% and 3.2% with and without G-CSF, respectively (P=.942). RR for AML/MDS with CT+G-CSF compared to control was 1.92 [P=.006] with ARD increase of 0.4% [P=.008]. RR for AML/MDS in study categories to receive the same, dose-dense or dose-escalated CT+G-CSF were 1.95 [P=.346], 1.20 [P=.666] and 2.47 [P=.006], respectively. RR for mortality with CT+G-CSF was 0.898 [P<.0001] with ARD decrease of 3.3% [P<.0001]. RR for mortality in study categories to receive the same, dose-dense or dose-escalated CT+G-CSF were 0.95 [P=.140], 0.84 [P<.001] and 0.91 [P=.019], respectively. Delivered relative dose intensities (RDI) were 1.18, 1.46 and 1.23 in studies planned to receive the same, dose-dense or dose-escalated CT, respectively. A significant association was observed between delivered RDI and reduced mortality [P=.013]. No differences in estimates of AML/MDS or mortality were observed between industry and non-industry-funded studies. Conclusions: Risk of AML/MDS is increased with dose escalated CT+G-CSF. Nevertheless, CT+G-CSF is associated with greater reductions in overall mortality. Dose-dense regimens are associated with the greatest RR reduction in mortality and lowest risk of AML/MDS. Further research is needed to differentiate any impact of G-CSF on the risk of AML/MDS from that due to increased CT intensity.

[Table: see text]

A validated risk model for early neutropenic events in older cancer patients receiving systemic chemotherapy

9036

Background: A prospective, nationwide study was undertaken to develop and validate a risk model for early neutropenic events (NE) in older cancer patients undergoing chemotherapy. Methods: 1,386 patients =65 years of age with lung, breast, colorectal, ovarian cancer or lymphoma were prospectively registered at 117 randomly selected sites. Data on up to 4 cycles were collected upon initiation of chemotherapy. A logistic regression model for cycle 1 NE consisting of febrile neutropenia (FN; fever/infection and absolute neutrophil count nadir <1x109/L) or severe neutropenia (SN; neutrophils <.5x109/L) was derived on 1,378 patients with available data. Validation was performed using a split sample random selection process. Results: No significant differences in distribution of NE or predictive factors were observed between derivation dataset (n=922) and validation dataset (n=464). Major independent baseline clinical risk factors for cycle 1 NE in the derivation model (DM) included: anthracycline based regimens (p<.001), non-chemotherapy immune-modulatory agents (p=.003), elevated bilirubin (p=.016), reduced glomerular filtration rate (p<.001), cancer type (p=.02), planned relative dose intensity =85% (p=.027), and regimens containing cyclophosphamide (p<.001), etoposide (p=.002) or ifosfamide (p=.032). Reduced risk of cycle 1 NE was associated with myeloid growth factor (MGF) prophylaxis (p<.001). DM R2 was 0.478 and c-statistic 0.88 [95% CI 0.86–0.91; p<.001]. At median predicted risk of cycle 1 NE of 7%, model test performance (MTP) showed: sensitivity 90%; specificity 59%; and predictive value positive and negative of 32% and 97%, respectively. Cycle 1–4 FN risk in the DM was 16.6% and 3.3% among high and low risk patients, respectively. The validation model (VM) R2 was 0.508 and c-statistic 0.89 [95% CI: 0.86–0.93; p<.001]. MTP in the VM demonstrated: sensitivity 90%; specificity 65%; predictive value positive and negative of 36% and 97%, respectively. Cycle 1–4 FN risk in the VM was 16.8% and 1.6% in high and low risk patients, respectively. Conclusions: This validated risk model demonstrated good discrimination between older cancer patients at decreased risk for NE, and those at increased risk who may benefit from targeted prophylaxis with MGF.

No significant financial relationships to disclose.

Risk factors for hospitalization in elderly cancer patients receiving chemotherapy

9035

Background: Chemotherapy-associated hospitalization is a major source of morbidity and cost in cancer care, particularly for elderly (age ≥ 65) cancer patients. Hospitalization in the elderly often leads to an irreversible decline in functional status unrelated to the acute event that prompted hospital admission. Currently, little is known about the risk factors that may lead to increased risk of hospitalization in elderly patients receiving chemotherapy (CTX). Methods: 871 patients with solid tumors or lymphoma initiating a new CTX regimen were prospectively enrolled at 60 randomly selected US community oncology sites between 8/2004 and 10/2005. Of these, 361 elderly patients aged 65–91 were identified and followed. Primary endpoint of this investigation was hematologic toxicity and hospitalization was secondary. Pre- CTX patient data were analyzed for increased risk of hospitalization in univariate analysis using the chi-square test. Results: A total of 155 (18%) patients were hospitalized resulting in 215 hospitalizations. Median time to first hospitalization was the second cycle of CTX. 81/361 (22%) of elderly patients were hospitalized compared to only 74/510 (15%) of younger patients (p=0.003). The rate of hospitalization increased in a linear fashion between ages 65–80. Reasons for hospitalization in the elderly included infection, fever, or febrile neutropenia (36%), cardiopulmonary disease (CPD) (12%), vomiting or dehydration (13%), other gastrointestinal (11%), transfusion (8%), thrombosis (4%), CTX administration (4%), and other (13%). Major independent pre-CTX factors that predicted hospitalization in the elderly included male gender (p=0.0004), hemoglobin <11 g/dL (p=0.02), abnormal platelet count (<150k or >350k) (p=0.05), CPD (p=0.03), creatinine >1.5 mg/dL (p=0.05), and ≥ 2 concomitant medications (p=0.0008). Elderly patients with lung cancer (p=0.001) and lymphoma (p=0.05) had significantly higher rates of hospitalization when compared to other solid tumors. Conclusions: These data suggest that the risk of hospitalization increases in elderly cancer patients with age and that pre-CTX factors may be useful in identifying a subpopulation at increased risk for hospitalization.

No significant financial relationships to disclose.

Neutropenic complications in ovarian cancer patients receiving chemotherapy: Results of a prospective observational study

19610

Background: A nationwide, prospective cohort study of patients receiving cancer chemotherapy was undertaken to better define treatment-associated toxicities and delivered dose intensity. Methods: Following informed consent, 3,670 patients initiating a new chemotherapy regimen with lymphoma (n=547) or carcinomas of the colon (n=521), breast (n=1,473), lung (n=907) or ovary (n=312) were prospectively registered at 117 randomly selected US oncology practices between March 2002 and December 2005. Pretreatment characteristics including clinical and sociodemographic factors and practice setting associated with reductions in chemotherapy dose intensity over 4 cycles were evaluated. Results: Chemotherapy regimen standards and intended dose and schedule were defined in 97% and 93% of patients, respectively. Average initial (planned) relative dose intensity (RDI) was 90% while the mean RDI actually delivered over 4 cycles was 83%. Reductions in RDI =15% of standard were planned in 23% of patients and actually occurred over 4 cycles in 38%. Body surface area (BSA) calculated by the method of Mosteller exceeded the BSA utilized by the oncology practices in two-thirds of patients most often due to capping at 2 m2. Obese patients with a body mass index (BMI) =30 were more likely to receive planned RDI reductions =15% (30%) than patients with BMI <30 (21%) (P<.0001). Other factors associated with planned reductions in RDI in multivariate analysis included: older age, male gender, lower education, absence of myeloid growth factor prophylaxis and cancers of the lung or ovary as well as the practice setting, size and geographic location. Factors associated with unplanned (subsequent) reductions in RDI included: older age, male gender, poorer Charlson comorbidity index, prior chemotherapy, lung cancer, low baseline hematocrit, white blood or platelet count, poor renal or hepatic function and low serum albumin as well as smaller practice and rural setting. Conclusions: A substantial proportion of patients receiving cancer chemotherapy experience both planned and unplanned reductions in RDI. Further understanding of factors associated with such reductions may provide opportunities for improving the quality of cancer care and clinical outcomes.

[Table: see text]

Factors associated with reductions in chemotherapy dose intensity: Impact of clinical, sociodemographic and practice setting

6550

Background: A nationwide, prospective cohort study of patients receiving cancer chemotherapy was undertaken to better define treatment-associated toxicities and delivered dose intensity. Methods: Following informed consent, 3,670 patients initiating a new chemotherapy regimen with lymphoma (n=547) or carcinomas of the colon (n=521), breast (n=1,473), lung (n=907) or ovary (n=312) were prospectively registered at 117 randomly selected US oncology practices between March 2002 and December 2005. Pretreatment characteristics including clinical and sociodemographic factors and practice setting associated with reductions in chemotherapy dose intensity over 4 cycles were evaluated. Results: Chemotherapy regimen standards and intended dose and schedule were defined in 97% and 93% of patients, respectively. Average initial (planned) relative dose intensity (RDI) was 90% while the mean RDI actually delivered over 4 cycles was 83%. Reductions in RDI =15% of standard were planned in 23% of patients and actually occurred over 4 cycles in 38%. Body surface area (BSA) calculated by the method of Mosteller exceeded the BSA utilized by the oncology practices in two-thirds of patients most often due to capping at 2 m2. Obese patients with a body mass index (BMI) =30 were more likely to receive planned RDI reductions =15% (30%) than patients with BMI <30 (21%) (P<.0001). Other factors associated with planned reductions in RDI in multivariate analysis included: older age, male gender, lower education, absence of myeloid growth factor prophylaxis and cancers of the lung or ovary as well as the practice setting, size and geographic location. Factors associated with unplanned (subsequent) reductions in RDI included: older age, male gender, poorer Charlson comorbidity index, prior chemotherapy, lung cancer, low baseline hematocrit, white blood or platelet count, poor renal or hepatic function and low serum albumin as well as smaller practice and rural setting. Conclusions: A substantial proportion of patients receiving cancer chemotherapy experience both planned and unplanned reductions in RDI. Further understanding of factors associated with such reductions may provide opportunities for improving the quality of cancer care and clinical outcomes.

No significant financial relationships to disclose.

Prospective validation of a risk model for first cycle neutropenic complications in patients receiving cancer chemotherapy

8561

Background: A nationwide, prospective cohort study was undertaken to develop and validate a risk model for neutropenic complications (NC) in cancer patients receiving chemotherapy. Methods: 3,596 patients initiating a new chemotherapy regimen with solid tumors or lymphoma were registered at 115 randomly selected sites. Data on at least 1 cycle of chemotherapy were available on 3,468. A logistic regression model for cycle 1 NC was derived and then validated using a split sample random selection process. Results: The risk of cycle 1 NC ranged from 5.5%-30.2%, averaging 18.5% across tumor types. No significant differences in distribution of NC or predictive factors were observed between the derivation dataset (n=2,592) or the validation dataset (n=876). Major independent baseline clinical risk factors for cycle 1 NC in the derivation model include: prior chemotherapy (P=.044), number of myelosuppressive agents (P<.0001), anthracycline-based regimens (P<.0001), planned delivery >85% of standard (P<.0001), cancer type (P<.0001), concurrent antibiotics (P=.023) or phenothiazines (P=.006), abnormal alkaline phosphatase (P=.002), elevated bilirubin (P=.031), low platelets (P=.004), elevated glucose (P=.023) and reduced glomerular filtration rate (P=.013). Reduced risk of cycle 1 NC was associated with primary prophylaxis with a myeloid growth factor (P<.0001). Model R2 was 0.273 and c-statistic 0.80 [95% CI: 0.78–0.82; P<.0001]. At the median predicted risk of cycle 1 NC of 11%, model test performance consisted of: sensitivity 84%; specificity 57% and diagnostic odds ratio (DOR) 7.2 while cycle 1 NC risk was 31% and 6% among high risk and low risk half, respectively. The model performed well in the smaller validation dataset with a model R2 of 0.354 and c-statistic of 0.84 [95% CI: 0.81–0.87, P<.0001]. Test performance of the model in the validation sample included: sensitivity 90%; specificity 62%; DOR 14.1 and risks of 35% and 4% in high risk and low risk patients, respectively. Conclusions: Validation in a randomly selected patient sample suggests that this model has general applicability in identifying patients at increased risk for NC. Further validation in other independent cancer patient populations receiving chemotherapy is planned.

[Table: see text]

Economic analysis of prophylactic granulocyte colony-stimulating factor (G-CSF) use based on a risk model for neutropenic complications in breast cancer patients receiving adjuvant chemotherapy

6107

Background: Although recent economic analyses of prophylactic G-CSF provide cost saving febrile neutropenia (FN) risk estimates of approximately 20%, many regimens have reported rates <20%. A prospective nationwide cohort study was undertaken to develop risk models for neutropenic complications (NC) including severe and febrile neutropenia in patients receiving cancer chemotherapy (Lyman ASCO 2005). A cost-effectiveness model is presented to evaluate the economic impact of G-CSF prophylaxis based on the model. Methods: Data on 974 consecutive breast cancer patients receiving adjuvant chemotherapy at 115 randomly selected practice sites were analyzed. The clinical and cost impact of G-CSF prophylaxis in high-risk patients based on the model was compared with: 1) no G-CSF; 2) primary prophylaxis; and 3) secondary prophylaxis. Pegfilgrastim costs were based on Medicare pricing while hospitalization costs and mortality on national hospitalization data. Results: Independent predictors of first cycle NC included: type and schedule of chemotherapy, diabetes, elevated bilirubin, planned RDI >85%, low glomerular filtration rate and low neutrophil count. Prophylactic G-CSF was associated with a decreased risk. Model R2=0.327 and c-statistic=0.80 [95% CI: 0.78–0.83; P<.001]. At a baseline FN risk of 8.4% per cycle, the expected costs over four cycles of chemotherapy were: no pegfilgrastim: $1,285; primary prophylaxis: $2,573; secondary prophylaxis: $2,040 and model-targeted G-CSF: $1,527. Expected cost varied with FN risk and model performance. Primary prophylaxis was associated with lower cost than no prophylaxis at FN risk >18%, while the model outperformed both strategies at an FN risk >10%. At a baseline cycle risk of FN of 8.4%, model-guided G-CSF was associated with an expected cost of $44,980 per life saved. Cost savings increased as model discrimination increased. The model was consistently associated with lower cost compared to secondary prophylaxis. Conclusions: A risk model for NC has been developed in breast cancer patients receiving adjuvant chemotherapy. Use of the model to guide G-CSF support appears to be cost-effective at an overall FN risk of 10%.

[Table: see text]

Variation in the cost of treatment for colorectal cancer

3625

Background: Recent publications have drawn attention to the escalating costs of drugs utilized for the treatment and supportive care of colorectal cancer (CRC) patients. As more advanced chemotherapeutic and supportive agents are approved, the cost of treatment will likely continue to increase. The goal of this analysis was to estimate costs of frequently used CRC chemotherapeutics and other commonly prescribed agents. Methods: Costs of drugs in CRC regimens for patients participating in a prospective, observational study from 115 community oncology practices between March 2002 and March 2005 were calculated. Standard dosages and schedules were determined through literature review and an expert CRC oncologist. Trimmed mean (90%) costs were computed using average wholesale prices (AWP) retrieved from the Lexicon Database Drug Product (January 2005). All prices noted are 95% AWP. All chemotherapeutic drugs were combined to arrive at regimen specific prices. Other regimens were defined as regimens with less than four patients. Price per cycle was estimated based on a standard patient BSA of 2.0 m2 and standard regimen schedules. Growth factor costs were derived from expected per cycle utilization, while prices for other supportive drugs were calculated using dosages for single events. Results: Patients were categorized into the following regimens ( table below). The price of a commonly used monoclonal antibody is: Bevacizumab ($2613), based on every 2 week dosing. Growth factor prices (per cycle) were estimated: Filgrastim ($2499), Pegfilgrastim ($2928), Epoetin ($552), and Darbepoetin Alfa ($1002). Prices per event for other commonly used agents were: Atropine ($245), Loperamide ($3), Ondansetron ($931), and Granisetron ($232). Conclusions: Prices of CRC regimens vary considerably, with newer agents and supportive drugs adding substantially to costs, particularly in late stage disease. Research on improved outcomes or treatment benefits associated with high cost treatment is warranted.

[Table: see text]

[Table: see text]

A prediction model for chemotherapy-associated thrombocytopenia in cancer patients

8616

Background: Thrombocytopenia (TP) can lead to serious complications, however, little is known about the incidence and risk factors for chemotherapy-associated TP. A prospective, nationwide cohort study was undertaken to better define the impact of TP in cancer treatment. Methods: 2,842 patients with cancer of the breast, lung, colon, ovary or lymphoma initiating a new chemotherapy regimen have been prospectively enrolled at 115 randomly selected US community oncology practices between 2002 and 2005. Risk factors for chemotherapy-associated TP were identified, a multivariate logistic regression model based on pretreatment characteristics was developed, and test performance characteristics were estimated. Results: Over a median of 3 cycles of chemotherapy, minimum recorded platelet counts were: ≥150K in 53% of patients; 100–150K in 26%; 75–100K in 8%; 50–75K in 6% and <50K in 7%. Significant independent predictive factors for platelets <75K include type of cancer (P<.0001), type of chemotherapy including gemcitabine-based (P<.0001), anthracycline-based (P<.0001) and platinum-based (P<.0001) regimens, prior chemotherapy (P<.0001) or surgery (P=.005), age (P=.015), Caucasian ethnicity (P=.022), body surface area (P=.0001), planned relative dose intensity ≥85% (P=.082), diabetes (P=.018), pulmonary disease (P=.011), abnormal baseline platelets (P<.0001), hematocrit (P=0.030), alkaline phosphatase (P=.072) or albumin (P=.017). Model fit was good (Chi-square, P<.0001), R2 = 0.735 and c-statistic = 0.816 [95% CI: 0.792–0.840, P<.0001]. Model test performance characteristics [95% CI] at a ≥20% risk of TP include: sensitivity 56% [51–61]; specificity 88% [87–89]; likelihood ratio positive 4.63 [4.02–5.33]; likelihood ratio negative 0.50 [0.45–0.57]; and diagnostic odds ratio 9.22 [7.23–11.75]. Validation of the model is underway. Conclusions: This prediction model based on pretreatment factors identifies with high specificity patients at risk for clinically important chemotherapy-associated thrombocytopenia early in the treatment course. It may provide a valuable tool for guiding chemotherapy and new supportive care measures.

[Table: see text]

Social and racial disparities in the use of non-standard breast cancer adjuvant chemotherapy regimens

6036

Background: Economically disadvantaged and black women have worse stage-specific breast cancer outcomes than other women, even after controlling for tumor histologic features. Disparities quality of chemotherapy may contribute to differences in outcome. The purpose of this study was to investigate the use of non-standard breast cancer adjuvant chemotherapy regimens in black women and those of lower socioeconomic status (SES). Methods: Detailed information on patient, disease, and treatment factors was collected prospectively on 1,006 patients receiving adjuvant chemotherapy for early-stage breast cancer in 115 oncology practices throughout the US. All patients signed informed consent. Regimens included in the guidelines of the National Comprehensive Cancer Network were considered standard regimens; all others were considered non-standard. Receipt of non-standard regimens was examined according to clinical and non-clinical factors. Differences between groups were assessed using a chi-square test. Multivariate logistic regression was used to identify factors associated with use of non-standard regimens. Results: Non-standard regimens were used in the treatment of 136 (14%) of the participants. Black patients were twice as likely to receive a non-standard regimen as whites (23% vs. 12%, p = .0014). Patients with less than a high school education were twice as likely to receive a non-standard regimen compared with those with a college education (21% vs. 8%, p = 0.0011). Other factors associated with non-standard chemotherapy regimens were past chemotherapy exposure (p < .0001), higher stage disease (p < .0001) and geographic location (p = 0.0059). Age, comorbidity, body mass index, type of insurance, and employment status were not associated with receipt of non-standard chemotherapy. In multivariate analysis, all variables that were significant in the bivariate analysis remained independently associated with receipt of non-standard chemotherapy. Conclusions: The more frequent use of non-guideline concordant adjuvant chemotherapy regimens in black women and women with lower SES may contribute to their less favorable outcomes. These findings offer an opportunity to improve patient care and perhaps cancer outcomes.

[Table: see text]

Non-clinical factors associated with intentionally reduced breast cancer chemotherapy doses

6020

Background: Breast cancer survival and case-fatality rates are lower in women of lower socioeconomic status (SES) compared to those of higher SES. Disparities in the quality of chemotherapy may contribute to poorer outcomes among women of lower SES. The purpose of this study was to investigate the relation between SES (as measured by both individual and area-level characteristics) and the use of intentionally reduced breast cancer adjuvant chemotherapy doses. Methods: The cohort consisted of 764 breast cancer patients treated with a standard adjuvant chemotherapy regimen for stages I, II, or III enrolled in an IRB approved prospective registry. Information was collected on patient, disease, and treatment characteristics, including chemotherapy doses. Zip code level data on median household income (MHI), and proportions of residents with a high school (HS) or college degree were obtained from the US 2000 Census. Doses for cycle 1 of chemotherapy < 85% of standard were considered to be reduced. Multivariate logistic regression analysis was performed to identify factors associated with the use of reduced first cycle doses. Results: In bivariate analysis, body mass index (BMI) (P=.003), education (P=.002), and region (P < .001) were associated with the use of reduced doses. Race was borderline significant (P=.056). Reduced doses were more commonly administered to patients who lived in zip codes with lower MHI (P < .001), lower education (P=.006), and higher levels of poverty (P < .001). Age, stage, comorbidity, hormone receptor status, insurance, occupation, employment and marital status were not associated with use of reduced doses. In multivariate analysis without practice effects, high BMI (P < .001), education less than HS (P=.009), and geographic region (P < .001) were independently associated with reduced doses. The effect of BMI (P < .001) stayed significant after adjustment for practice site, but education (P=.082) and geographic region (P=.127) became less so. Conclusions: Non-clinical factors, such as educational attainment and area level income, are associated with intentionally reduced doses of adjuvant chemotherapy. Interventions directed at improving quality of chemotherapy may improve breast cancer outcomes in patients of lower SES.

No significant financial relationships to disclose.