Search

PURPOSE This study documents the creation of automated, longitudinal, and prospective data and analytics platform for breast cancer at a regional cancer center. This platform combines principles of data warehousing with natural language processing (NLP) to provide the integrated, timely, meaningful, high-quality, and actionable data required to establish a learning health system. METHODS Data from six hospital information systems and one external data source were integrated on a nightly basis by automated extract/transform/load jobs. Free-text clinical documentation was processed using a commercial NLP engine. RESULTS The platform contains 141 data elements of 7,019 patients with newly diagnosed breast cancer who received care at our regional cancer center from January 1, 2014, to June 3, 2022. Daily updating of the database takes an average of 56 minutes. Evaluation of the tuning of NLP jobs found overall high performance, with an F1 of 1.0 for 19 variables, with a further 16 variables with an F1 of > 0.95. CONCLUSION This study describes how data warehousing combined with NLP can be used to create a prospective data and analytics platform to enable a learning health system. Although upfront time investment required to create the platform was considerable, now that it has been developed, daily data processing is completed automatically in less than an hour.

PURPOSE Clinicians need accurate and timely information on the impact of treatments on patient outcomes. The electronic health record (EHR) offers the potential for insight into real-world patient experiences and outcomes, but it is difficult to tap into. Our goal was to apply artificial intelligence technology to the EHR to characterize the clinical course of patients with stage III breast cancer. PATIENTS AND METHODS Data from patients with stage III breast cancer who presented between 2013 and 2015 were extracted from the EHR, de-identified, and imported into the IBM Cloud. Specialized natural language processing (NLP) annotators were developed to extract medical concepts from unstructured clinical text and transform them to structured attributes. In the validation phase, these annotators were applied to 19 additional patients with stage III breast cancer from the same period. The resulting data were compared with that in the medical chart (gold standard) for nine key indicators. RESULTS Information was extracted for 50 patients, including tumor stage (94% stage IIIA, 6% stage IIIB), age (28% 50 years or younger, 52% between 51 and 70 years, and 24% older than 70 years), receptor status (84% estrogen receptor positive, 74% progesterone receptor positive), and first treatment (72% surgery, 26% chemotherapy, 2% endocrine). Events in the patient’s journey were compiled to create a timeline. For 171 data elements, NLP and the chart disagreed for 41 (24%; 95% CI, 17.8% to 31.1%). With additional manipulation using simple logic, the disagreement was reduced to six elements (3.5%; 95% CI, 1.3% to 7.5%; F1 statistic, 0.9694). CONCLUSION It is possible to extract, read, and combine data from the EHR to view the patient journey. The agreement between NLP and the gold standard was high, which supports validity.

Purpose: To evaluate whether positron emission tomography (PET) combined with computed tomography (PET-CT) is cost saving, or cost neutral, compared with conventional imaging in management of patients with resectable colorectal cancer liver metastases. Methods: Cost evaluation of a randomized trial that compared the effect of PET-CT on surgical management of patients with resectable colorectal cancer liver metastases. Health care use data ≤ 1 year after random assignment was obtained from administrative databases. Cost analysis was undertaken from the perspective of a third-party payer (ie, Ministry of Health). Mean costs with 95% credible intervals (CrI) were estimated by using a Bayesian approach. Results: The estimated mean cost per patient in the 263 patients who underwent PET-CT was $45,454 CAD (range, $1,340 to $181,420) and in the 134 control patients, $40,859 CAD (range, $279 to $293,558), with a net difference of $4,327 CAD (95% CrI, −$2,207 to $10,614). The primary cost driver was hospitalization for liver surgery (difference of $2,997 CAD for PET-CT; 95% CrI, −$2,144 to $8,010), which was mainly a result of a longer length of hospital stay for the PET-CT arm (median, 7 v 6 days; P = .03) and a higher postoperative complication rate (20% v 10%; P = .01). Baseline characteristics were similar between groups, including the number of liver segments involved with cancer, number of segments resected, and type of liver resection performed. No difference in survival was detected between arms. Conclusion: PET-CT was associated with limited clinical benefit and a nonsignificant increased cost. Universal funding of PET-CT in the management of patients with resectable colorectal cancer liver metastases does not seem justified.

TPS761 Background: Randomized data from the interferon era demonstrated modest survival benefits of cytoreductive nephrectomy (CN) in patients with advanced renal cell carcinoma (aRCC). Results from SURTIME and CARMENA, conducted in the VEGF-targeted therapy era, have challenged the routine use of upfront CN especially in IMDC intermediate and poor risk patients. Furthermore, the treatment landscape in aRCC now includes first-line combination immunotherapy. Data from the Checkmate-214 trial showed that intermediate/poor risk patients have improved overall survival and objective response rate with ipilimumab and nivolumab (I/N) compared to sunitinib. Stereotactic body radiation therapy (SBRT) provides a convenient method for cytoreduction of the primary kidney lesion and may induce an ‘abscopal effect’, leading to enhanced systemic anti-tumour immune response. We hypothesize that SBRT to the primary kidney mass in aRCC patients will enhance the efficacy of I/N compared to standard of care I/N alone. Methods: This phase II trial randomizes untreated aRCC patients in a 2:1 fashion to I/N plus SBRT (30-40 Gy in 5 fractions) to the primary kidney mass between cycles 1 and 2 (experimental arm, E), versus standard of care I/N alone (standard arm, S). Eligible patients have biopsy-proven aRCC (any histology), IMDC intermediate/poor risk disease, and who decline or are unsuitable for CN. Patients with a primary kidney lesion ≥ 20cm, previous abdominal radiation precluding SBRT, or who have a contraindication to I/N are excluded. The primary objective is to compare the efficacy of I/N plus SBRT versus I/N alone, as determined by the hazard ratio for progression free survival (PFS). Secondary objectives include evaluation of safety, overall survival, objective response rate, and health-related quality of life. Exploratory analyses include blood immune signatures and stool microbiome. Up to 78 patients will be enrolled under the assumption of an improved 12-month PFS from 50% (S) to 75% (E), using a two-sided α=0.1, power=80%, and accounting for loss-to-follow-up and stratification using IMDC criteria 1-2 vs 3-6. Clinical trial information: NCT04090710.

TPS679 Background: AA has re-emerged as a promising anti-cancer agent based on recent knowledge of pharmacokinetics, discovery of unexpected mechanisms of action, and early phase trials with IV AA.(Shenoy et al, Cancer Cell 34: 2018, in press) We hypothesized that ccRCC would be particularly susceptible to anti-cancer effects of IV AA due to: a. TET dependent demethylation of the hypermethylated genome of ccRCC causing re-expression of tumor suppressors (Shenoy et al, AACR 2018 Targeting DNA methylation conf. A11; Hu et al, Clin Cancer Res 20:4349-60, 2014) b. H2O2 production causing intra-tumoral oxidative damage, hypothesized to be enhanced by high iron content in RCC microenvironment c. Intracellular accumulation of dehydroascorbic acid secondary to high HIF activity in ccRCC. Animal data and case reports support the hypothesis. Methods: Trial design: Patients (pts) with newly diagnosed metastatic/ unresectable ccRCC are randomized 1:1 to arm A (pazopanib 800 mg/d plus IV AA 1g/kg 3 times/week) or arm B (pazopanib 800 mg/d). Protocol treatment is for 10 cycles (unless PD, unacceptable AE, alternative therapy, or pt refusal), each cycle being 28 days. Primary endpoint is Treatment Failure-Free rate at 40 weeks (TFF40). Treatment Failure is defined as: Radiographic disease progression, off-protocol treatment due to AE, alternative therapy initiation (except metastasectomy post clinical benefit), or death. Secondary endpoints include OS, PFS, ORR and AE. Statistical methods: 82 eligible pts (41 in each in arm) will provide 81% power to detect a 19% increase of TFF40 from 45% in arm B to 64% in arm A assuming a one-sided type I error rate of 0.19 (EAST 6.4). Correlatives: Epigenetic mechanism: 5mC, 5hmC and H3K27me3 IHC, MeDIP/ hMeDIP seq, RNA seq. H2O2 mechanism: tumor microenvironment iron, tumor catalase IHC. Dehydroascorbic acid mechanism: HIF-1 alpha, HIF-2 alpha, GLUT-1 IHC. Key exclusion criteria: G6PD deficiency, renal disease (Cockcroft Gault CrCl < 55 ml/min). ClinicalTrials.gov Identifier: NCT03334409 Status: Open for accrual in 9/10 planned sites. Funding Source: Foundation. Clinical trial information: NCT03334409.

This Special Issue of Journal of Clinical Oncology provides a broad, in-depth look at the methodologic field that has become known as comparative effectiveness research (CER). The focus is on CER and its past, current, and future impact on cancer research, clinical oncology practice, and national health care policy. An extraordinary assemblage of authorities on a broad range of related topics has been gathered. The authors address issues ranging from those very familiar to oncologists and cancer researchers, such as randomized controlled trials and meta-analyses of such trials, to somewhat less familiar methodologies such as observational studies, decision modeling, and economic analyses of health care policy issues. The importance of our rapidly evolving understanding of the human genome and the molecular and genetic basis of new diagnostic and therapeutic tools will be discussed, along with the multiple challenges that they bring to CER. Important methodologic issues will be addressed, such as capturing patient-reported outcomes, conducting health technology assessments, and developing clinical guidelines to improve the quality of care. Finally, the manner in which CER has affected health care delivery and its cost, both in the United States and other developed nations such as Canada and the United Kingdom, will be discussed. The issue will end on a cautionary but positive note, reconciling the many varied approaches and expectations of CER as we strive to provide the most sophisticated and efficacious cancer care to all patients affected by this disease.

Purpose Many oncology clinical trials departments (CTDs) are in serious fiscal deficit and their sustainability is in jeopardy. This study investigates whether the payment models used to fund industry versus cooperative group trials contribute to the fiscal deficit of a CTD. Methods We examined the lifetime costs of all cooperative group and industry trials activated in the CTD of a cancer center between 2007 and 2011. A trial's lifetime is defined as being from the date the first patient was accrued until the last patient's actual or projected final follow-up visit. For each trial, we calculated the lifetime monthly net income, which was defined as monthly revenue minus monthly costs. Data sources included study protocols, trial budgets, and accrual data. Results Of the 97 trials analyzed, 64 (66%) were cooperative group trials. The pattern of lifetime net income for cooperative group trials has a positive peak during patient accrual followed by a negative trough during follow-up. In contrast, the pattern for industry trials resembled an “l” shape. The patterns reflect the differing payment models: upfront lump-sum payments (cooperative group) versus milestone payments (industry). Conclusion The negative trough in the lifetime net income of a cooperative group trial occurs because follow-up costs are typically not funded or are underfunded. CTDs accrue more patients in new trials to offset that deficit. The CTD uses revenue from accrual to existing trials to cross-subsidize past trials in follow-up. As the number of patients on follow-up increases, the fiscal deficit grows larger each year, perpetuating the cycle.

Purpose 2-[18F]fluorodeoxyglucose (FDG) positron emission tomography (PET) is potentially useful in assessing lymph nodes and detecting distant metastases in women with primary breast cancer. Patients and Methods Women diagnosed with operable breast cancer within 3 months underwent FDG-PET at one of five Ontario study centers followed by axillary lymph node assessment (ALNA) consisting of sentinel lymph node biopsy (SLNB) alone if sentinel lymph nodes (SLNs) were negative, SLNB with axillary lymph node dissection (ALND) if SLNB or PET was positive, or ALND alone if SLNs were not identified. Results Between January 2005 and March 2007, 325 analyzable women entered this study. Sentinel nodes were found for 312 (96%) of 325 women and were positive for tumor in 90 (29%) of 312. ALND was positive in seven additional women. Using ALNA as the gold standard, sensitivity for PET was 23.7% (95% CI, 15.9% to 33.6%), specificity was 99.6% (95% CI, 97.2% to 99.9%), positive predictive value was 95.8% (95% CI, 76.9% to 99.8%), negative predictive value was 75.4% (95% CI, 70.1% to 80.1%), and prevalence was 29.8% (95% CI, 25.0% to 35.2%). Using logistic regression, tumor size was predictive for prevalence of tumor in the axilla and for PET sensitivity. PET scan was suspicious for distant metastases in 13 patients; three (0.9%) were confirmed as metastatic disease and 10 (3.0%) were false positive. Conclusion FDG-PET is not sufficiently sensitive to detect positive axillary lymph nodes, nor is it sufficiently specific to appropriately identify distant metastases. However, the very high positive predictive value (96%) suggests that PET when positive is indicative of disease in axillary nodes, which may influence surgical care.

Purpose An Institute of Medicine report recommends that patients with cancer receive a survivorship care plan (SCP). The trial objective was to determine if an SCP for breast cancer survivors improves patient-reported outcomes. Patients and Methods Women with early-stage breast cancer who completed primary treatment at least 3 months previously were eligible. Consenting patients were allocated within two strata: less than 24 months and ≥ 24 months since diagnosis. All patients were transferred to their own primary care physician (PCP) for follow-up. In addition to a discharge visit, the intervention group received an SCP, which was reviewed during a 30-minute educational session with a nurse, and their PCP received the SCP and guideline on follow-up. The primary outcome was cancer-related distress at 12 months, assessed by the Impact of Event Scale (IES). Secondary outcomes included quality of life, patient satisfaction, continuity/coordination of care, and health service measures. Results Overall, 408 survivors were enrolled through nine tertiary cancer centers. There were no differences between groups on cancer-related distress or on any of the patient-reported secondary outcomes, and there were no differences when the two strata were analyzed separately. More patients in the intervention than control group correctly identify their PCP as primarily responsible for follow-up (98.7% v 89.1%; difference, 9.6%; 95% CI, 3.9 to 15.9; P = .005). Conclusion The results do not support the hypothesis that SCPs are beneficial for improving patient-reported outcomes. Transferring follow-up to PCPs is considered an important strategy to meet the demand for scarce oncology resources. SCPs were no better than a standard discharge visit with the oncologist to facilitate transfer.

3527Background: The PETCAM randomized trial evaluated the effect of preoperative PET-CT (vs. no PET-CT) on surgical management in patients with colorectal cancer liver metastases. This trial result...

562 Background: The PETCAM randomized trial evaluated the effect of preoperative PET-CT (vs. no PET-CT) on surgical management in patients with colorectal cancer liver metastases. In this study, 8% of patients had a change in surgical management, including a higher proportion of major liver resections in the PET-CT arm. The current study compares the intervention groups for 5-year disease free (DFS) and overall survival (OS), and evaluated their long-term clinical course, i.e. sites of recurrence and management of disease recurrence. Methods: Recruitment to the trial occurred between 2005-2010, with last follow-up in 2013. Data on recurrence, management of recurrence and mortality from 2013-2017 was collected from patient’s charts. Recurrences according to site and management were described. Cox proportional Hazard Models were used to calculate the risk for recurrence and death. OS was calculated with Kaplan-Meir method and compared with log-rank test. Results: At 5 years, 157 of 404 (39%) patients were still alive and 19 patients were lost to follow-up. Median follow-up is 4.2 years. There were no differences in DFS (HR: 1.12, 95%CI: 0.88-1.42) or OS (HR: 0.97, 95%CI: 0.74-1.28) between groups. The median DFS for the 372 patients who had surgery was 17 months, 95%CI: 14.7-19.4. Risks factors for recurrence were: extrahepatic disease, liver tumour size, and nodal stage. The median OS for all patients was 50 months, 95%CI: 43.5-64.3. Risks factors for death also included age and prior use of chemotherapy. During the follow-up period, 287/404, 71% patients recurred (mostly liver and lung); 137 (48%) were treated solely with chemotherapy and 35% were treated with surgery with curative intent. Of these, the majority recurred (109/116, 94%). The median OS following first recurrence was 27.5 months, 95%CI: 23-30. Conclusions: PET-CT did not improve DFS or OS. Survival following liver resection is similar to previous reports, however most patients experience disease recurrence. A substantial proportion of patients who recur undergo surgery, however it is likely that they will recur again.

Purpose The uptake of new health care technologies is usually driven by industry promotion, physician interest, patient demand, and institutional ability to acquire the technology. The introduction of positron emission tomography (PET) scanning in the province of Ontario, Canada, followed a different path. Methods The Ontario provincial government, through its Ministry of Health and Long-Term Care, commissioned a systematic review of the literature. When this found only weak evidence that PET has a positive impact on clinical outcomes, the Ministry introduced a provincial PET evaluation program to close the evidence gap. Results This article describes the challenges encountered establishing the PET evaluation program. These included the design and conduct of the initial clinical trials, the establishment of a PET cancer registry, standardizing how PET scans were performed and reported, and gaining acceptance by health professionals for the evaluative program. Conclusion The proliferation of health technologies is a key driver of increasing health care costs. The Ontario approach to the introduction of PET is a model worth consideration by health systems seeking to ensure that they receive value for money based on a strong evidentiary base when introducing new health technologies.

Purpose Previous trials have suggested a quality-of-life (QOL) improvement for anemic cancer patients treated with erythropoietin, but few used QOL as the primary outcome. We designed a trial to investigate the effects of epoetin alfa therapy on the QOL of anemic patients with advanced non–small-cell carcinoma of the lung (NSCLC). Patients and Methods A multicenter, randomized, double-blind, placebo-controlled trial was conducted. The proposed sample size was 300 patients. Eligible patients were required to have NSCLC unsuitable for curative therapy and baseline hemoglobin (Hgb) levels less than 121 g/L. Patients were assigned to 12 weekly injections of subcutaneous epoetin alpha or placebo, targeting Hgb levels between 120 and 140 g/L. The primary outcome was the difference in the change in Functional Assessment of Cancer Therapy–Anemia scores between baseline and 12 weeks. Results Reports of thrombotic events in other epoetin trials prompted an unplanned safety analysis after 70 patients had been randomly assigned (33 to the active arm and 37 to the placebo arm). This revealed a significant difference in the median survival in favor of the patients on the placebo arm of the trial (63 v 129 days; hazard ratio, 1.84; P = .04). The Steering Committee closed the trial. Patient numbers compromised the interpretation of the QOL analysis, but a positive Hgb response was noted with epoetin alfa treatment. Conclusion An unplanned safety analysis suggested decreased overall survival in patients with advanced NSCLC treated with epoetin alfa. Although infrequent, other similar reports highlight the need for ongoing trials evaluating erythropoietin receptor agonists to ensure that overall survival is monitored closely.

Purpose Most women with breast cancer are diagnosed at an early stage and more than 80% will be long-term survivors. Routine follow-up marks the transition from intensive treatment to survivorship. It is usual practice for routine follow-up to take place in specialist clinics. This study tested the hypothesis that follow-up by the patient's family physician is a safe and acceptable alternative to specialist follow-up. Patients and Methods A multicenter, randomized, controlled trial was conducted involving 968 patients with early-stage breast cancer who had completed adjuvant treatment, were disease free, and were between 9 and 15 months after diagnosis. Patients may have continued receiving adjuvant hormonal therapy. Patients were randomly allocated to follow-up in the cancer center according to usual practice (CC group) or follow-up from their own family physician (FP group). The primary outcome was the rate of recurrence-related serious clinical events (SCEs). The secondary outcome was health-related quality of life (HRQL). Results In the FP group, there were 54 recurrences (11.2%) and 29 deaths (6.0%). In the CC group, there were 64 recurrences (13.2%) and 30 deaths (6.2%). In the FP group, 17 patients (3.5%) compared with 18 patients (3.7%) in the CC group experienced an SCE (0.19% difference; 95% CI, −2.26% to 2.65%). No statistically significant differences (P < .05) were detected between groups on any of the HRQL questionnaires. Conclusion Breast cancer patients can be offered follow-up by their family physician without concern that important recurrence-related SCEs will occur more frequently or that HRQL will be negatively affected.

Purpose To investigate the therapeutic impact of chemotherapy-induced amenorrhea in premenopausal patients with breast cancer. Patients and Methods We conducted a retrospective cohort study of a National Cancer Institute of Canada Clinical Trials Group phase III trial involving premenopausal patients randomized to receive cyclophosphamide, methotrexate, and fluorouracil (CMF), versus intensive cyclophosphamide, epirubicin, and fluorouracil (CEF). The objectives of our study were to describe the incidence of amenorrhea at 6 and 12 months post-random assignment and to determine the association of amenorrhea with relapse-free and overall survival. Results Data on 442 patients were used in our analyses. Despite the higher cumulative dose of cyclophosphamide in the CMF treatment arm, at 6 months post–random assignment, the rate of amenorrhea was higher in the CEF group (relative risk, 1.2; 95% CI, 1.0 to 1.3), with no difference at 12 months. In the receptor-positive subgroup, 6-month amenorrhea rates were not associated with prognosis. In contrast, amenorrhea at 12 months was significantly associated with relapse-free survival (hazard ratio, 0.51; 95% CI, 0.32 to 0.82; P = .005) and overall survival (hazard ratio, 0.40; 95% CI, 0.22 to 0.72; P = .002). Conclusion Late chemotherapy-induced amenorrhea seems to be associated with improved outcome in patients with premenopausal, receptor-positive breast cancer.

Purpose To determine if iridium implant (IM) and external-beam radiation therapy (EBRT) is better than standard EBRT in locally advanced prostate cancer. Methods Patients with T2 and T3 prostate cancer with no evidence of metastatic disease were randomly assigned to EBRT of 66 Gy in 33 fractions during 6.5 weeks or to IM of 35 Gy delivered to the prostate during 48 hours plus EBRT of 40 Gy in 20 fractions during 4 weeks. The primary outcome consisted of biochemical or clinical failure (BCF). BCF was defined by biochemical failure, clinical failure, or death as a result of prostate cancer. Secondary outcomes included 2-year postradiation biopsy positivity, toxicity, and survival. Results Between 1992 and 1997, 51 patients were randomly assigned to receive IM plus EBRT, and 53 patients were randomly assigned to receive EBRT alone. The median follow-up was 8.2 years. In the IM plus EBRT arm, 17 patients (29%) experienced BCF compared with 33 patients (61%) in the EBRT arm (hazard ratio, 0.42; P = .0024). Eighty-seven patients (84%) had a postradiation biopsy; 10 (24%) of 42 in the IM plus EBRT arm had biopsy positivity compared with 23 (51%) of 45 in the EBRT arm (odds ratio, 0.30; P = .015). Overall survival was 94% in the IM plus EBRT arm versus 92% in the EBRT arm. Conclusion The combination of IM plus EBRT was superior to EBRT alone for BCF and postradiation biopsy. This trial provides evidence that higher doses of radiation delivered in a shorter duration result in better local as well as biochemical control in locally advanced prostrate cancer.

Randomized controlled trials have had an important impact on the treatment of women with early breast cancer. Many women routinely receive breast-conserving surgery plus radiation therapy, adjuvant chemotherapy, or hormonal therapy based on the results of randomized trials. However, despite the extensive use of combinations of chemotherapy, hormonal therapy, and radiation therapy, the optimal sequencing of these different adjuvant treatments for patients with early breast cancer remains unclear. Unfortunately, there have been few randomized trials addressing this important issue. In a randomized trial by Recht et al, 1 the optimal sequencing of radiation and chemotherapy in patients treated with breast-conserving surgery was evaluated. Of 244 patients, 122 were allocated to receive radiation therapy before anthracycline-based chemotherapy and 122 were allocated to receive radiation therapy after the same chemotherapy. The median length of follow-up was 58 months. Although the study was relatively small by today’s standards, there was a trend for an increased risk of distant recurrence for patients treated with radiation therapy first (36% v 25%; hazard ratio [HR], 1.62; 95% CI, 1.01 to 2.62;P .05). As a result, breast irradiation is now usually given after anthracycline-based chemotherapy. In early 1990s, investigators from the North American Breast Intergroup conducted a randomized trial in postmenopausal women with node-positive, hormone receptor–positive breast cancer, which compared tamoxifen, administered concurrently with cyclophosphamide, doxorubicin, and fluorouracil (n 550) with tamoxifen given after the completion of the same chemotherapy (n 566). 2 This was an important clinical question because there was some concern that the cytostatic effect of concurrent tamoxifen might interfere with the cytotoxic effects of chemotherapy on cancer cells. Preclinical data had demonstrated that human breast tumor cells in culture pretreated with tamoxifen were protected from the cytotoxicity of chemotherapy. 3 This concern was not shared by all investigators; several cooperative groups recommended the concurrent use of chemotherapy and tamoxifen in trials evaluating these agents. Furthermore, it had already been observed that combined chemotherapy and tamoxifen increased the risk of thromboembolic complications in patients. 4 With a median follow-up of 8.5 years, the results of the Intergroup study were presented at the 2000 American Society of Clinical Oncology Annual Meeting. The 8-year disease-free survival estimates were 67% for patients treated with sequential chemotherapy and tamoxifen compared with 62% for patients treated with concurrent chemotherapy and tamoxifen (P .05). 2 This study led to important

Purpose Few interventions have been designed and tested to improve recruitment to clinical trials in oncology. The multiple factors influencing patients' decisions have made the prioritization of specific interventions challenging. The present study was undertaken to identify the independent predictors of a cancer patient's decision to enter a randomized clinical trial. Methods A list of factors from the medical literature was augmented with a series of focus groups involving cancer patients, physicians, and clinical research associates (CRAs). A series of questionnaires was developed with items based on these factors and were administered concurrently to 189 cancer patients, their physicians, and CRAs following the patient's decision regarding trial entry. Forward logistic regression modeling was performed using the items significantly correlated (by univariate analysis) with the decision to enter a clinical trial. Results A number of items were significantly correlated with the patient's decision. In the multivariate logistic regression model, the patient's perception of personal benefit was the most important, with an odds ratio (OR) of 3.08 (P < .05). CRA-related items involving supportive aspects of the decision-making process were also important. These included whether the CRA helped with the decision (OR = 1.71; P < .05), and whether the decision was hard for the patient to make (OR = 0.52; P < .05). Conclusion Strategies that better address the potential benefits of trial entry may result in improved accrual. Interventions or aids that focus on the supportive aspects of the decision-making process while respecting the need for information and patient autonomy may also lead to meaningful improvements in accrual.

Purpose In experimental systems, interference with coagulation can affect tumor biology. Furthermore, it has been suggested that low molecular weight heparin therapy may prolong survival in patients with cancer. The primary aim of this study was to assess survival at 1 year of patients with advanced cancer. Patients and Methods Patients with advanced malignancy (N = 385) were randomly assigned to receive either a once-daily subcutaneous injection of dalteparin (5,000 IU), a low molecular weight heparin, or placebo for 1 year. Results The Kaplan-Meier survival estimates at 1, 2, and 3 years after randomization for patients receiving dalteparin were 46%, 27%, and 21%, respectively, compared with 41%, 18%, and 12%, respectively, for patients receiving placebo (P = .19). In an analysis not specified a priori, survival was examined in a subgroup of patients (dalteparin, n = 55; and placebo, n = 47) who had a better prognosis and who were alive 17 months after randomization. In these patients, Kaplan-Meier survival estimates at 2 and 3 years from randomization were significantly improved for patients receiving dalteparin versus placebo (78% v 55% and 60% v 36%, respectively, P = .03). The rates of symptomatic venous thromboembolism were 2.4% and 3.3% for dalteparin and placebo, respectively, with bleeding rates of 4.7% and 2.7%, respectively. Conclusion Dalteparin administration did not significantly improve 1-year survival rates in patients with advanced malignancy. However, the observed improved survival in a subgroup of patients with a better prognosis suggests a potential modifying effect of dalteparin on tumor biology.

478 Background: The recommendation to administer extended-duration (28 days) venous thromboembolic events (VTE) prophylaxis with low molecular weight heparin in patients undergoing abdominal and pelvic surgery for cancer has not been widely implemented mainly because most studies focus on asymptomatic events with unknown clinical significance. The objective of this study was to determine the post-hospital discharge VTE incidence in these patients who do not receive post-hospital discharge prophylaxis. Methods: Prospective cohort study of patients undergoing abdominal and pelvic operations for cancer within the gastrointestinal tract, hepatobiliary (HPB) system or gynecological organs, with surgery lasting > 1 hour, post-operative stay < 28 days, and not undergoing anticoagulant therapy. Patients were evaluated at 1, 3 and 6 months from index operation for the presence of VTE by means of a screening ultrasound at 28 days and a questionnaire at each follow-up. The proportion with 95% confidence interval (CI) of VTE was calculated. Multivariable logistical regression was performed. Results: Of 284 patients, there were 79 (28%) colorectal, 97 (35%) HPB and 100 (35%) gynecology. All patients received pre- and post-operative in-patient prophylaxis. The proportion of VTE at 6 months was 7%, 95% CI 4.4-10.7 (20 events). Most events occurred between 3-6 months, 4.6%, 95% CI 2.46 – 7.7. Only one event occurred at one month after surgery (0.35%, 95% CI 0.06-1.97). 50% of the cohort had screening ultrasound, all of which were negative. Events were evenly distributed according to the type of surgery. The proportion of patients who died was 6.6%, 95% CI 3.5 – 9.4 (17 patients, 2 of which had a VTE-related death). In the multivariable analysis, post-operative chemotherapy was significantly associated with VTE, odds ratio (OR) 2.74, 95% CI 1.07-6.99. Caprini score was also associated with VTE but was not significant when included in the multivariable analysis, OR 1.19, 95% CI 0.99-1.42. Conclusions: Incidence of VTE following abdominal cancer surgery is low. Most events occur between 3-6 months from surgery. Post-operative chemotherapy is significantly associated with post-hospital discharge VTE.

For many years, surgery, radiation therapy, chemotherapy, and endocrine therapy have been the mainstays of cancer treatment. These were the modalities available when I first started clinical practice in 1981. In the early 1990s, I began to notice that some patients were choosing alternatives to the mainstream or conventional treatments. These so-called alternative therapies included a range of interventions, such as dietary and behavioral interventions, vitamin supplements and herbs, and traditional systems such as Chinese and homeopathic medicines. 1 Most of these were not supported by sound scientific methods. Alternative therapies appealed to certain patients who were exasperated with the conventional treatments they had received, ter- rified of the adverse effects, or desperate for miracles because they had been told their cancers were incurable. It was not long before the term complementary medicine appeared. This referred to therapies or prac- tices used alongside conventional medicine, whereas alternative med- icine was used in place of conventional medicine. 2 Complementary and alternative medicines are often bundled together under the acronym CAM, the goals of which are to increase the efficacy of conventional cancer treatment programs, reduce symp- toms, and improve quality of life for patients with cancer. CAM is grouped into four domains to account for the wide diversity in thera- pies: mind-body medicine, biologically based practices, manipulative and body-based practices, and energy therapies. 3 Cutting across all four domains is the category of whole medical systems. These repre- sent comprehensive systems of theory and practice developed before— or parallel to— biomedicine. These systems usually have his- tories of effectiveness ascertained through many years of observation within specific cultural contexts (eg, traditional Chinese medicine). Journal of Clinical Oncology (JCO) recently published two studies evaluating different CAM therapies: acupuncture for hot flashes in women with breast cancer 4 and enzymes for patients with pancre- atic cancer. 5

There is much excitement in the oncology community with the recent emergence of molecular targeted therapies, such as rituxumab, trastuzumab, and bevacizumab. Such therapy, however, comes at considerable cost, and we know that there are many more new agents in the pipeline. All countries, irrespective of their health care system, are struggling with how to cope with the existing and further increasing financial burden caused by cancer and its treatment. Oncologists find themselves in a difficult position. On the one hand, as prescribers they can be put in a position of gatekeepers; on the other, they are advocates for their patients. Given that these competing roles occur on a daily basis, it is important that the Journal of Clinical Oncology publish studies on economic evaluation for the readership. Since 2000, the JCO has published approximately five articles per year on economic-related issues. As editors, we have considered—and sometimes struggled with—a number of issues when assessing economic articles. We would briefly like to review some of these issues, to provide some guidance to researchers on the type of manuscript in which we are interested. First, the question should be important and relevant. From time-to-time, we receive an economic analysis for a treatment that has been used for a long time and is based on high quality evidence, such as radiation after breast-conservation surgery. The results of the economic analysis are used to justify the use of the therapy (ie, it’s cheap so it is worth giving). This type of economic research is of low priority for the JCO because the treatment will be used anyways. In contrast, a research article examining resource allocation if such therapy was discontinued, thus freeing up resources for other treatments would be of interest. Second, the design of the study needs to be considered. The least rigorous in the design hierarchy is an article that merely describes the cost of an intervention without a comparator. This type of study is also of low priority for the JCO. As we move up the hierarchy we go from cost minimization, to cost effectiveness (C/ E), and to cost utility analysis. These three designs involve a comparison between two interventions— usually something new compared with a standard that provides some form of benchmark. We sometimes receive articles where decision analysis is used to compare different therapeutic strategies. We would prefer to see data from a randomized trial that addresses the question, but there are situations where a randomized trial will not be done and a decision analysis using the best available evidence can be helpful. A decision analysis can also be used to describe in a very analytic fashion the comparison of alternative therapies where there is a trade-off between efficacy and toxicity and the answer is not clear cut. Although a decision analysis can help one think through a complex problem, it might be argued that an experienced clinician presents the information on benefits and risks of treatment to a patient in every day practice to elicit a patient preference. The third issue to consider is the quality of the data used in the economic analysis. The best outcome data come from randomized trials. Usually the publications of randomized trials provide data on efficacy (mortality and recurrence) and toxicity. However, economic data (costs and consequences as measured depending on the method used; for example, utilities) are seldom collected prospectively in clinical trials because of the added cost. We would encourage that researchers try and collect economic data prospectively in trials. However, often costs and consequences are obtained from other sources. Although this may be a necessity, it is not optimal. Assumptions are often made to convert toxicity and quality of life into outcome measures such as utilities. This process needs to be data driven, but is still a best guess. For example, the limitations of utility values based on the opinions of volunteers or experts who have actually not experienced the intervention need to be recognized. Economic evaluation that uses efficacy data from nonrandomized studies can be problematic because of the potential for bias. The reader needs to be aware that the economic analysis output, such as C/E ratio, is only as good as the data that is put into the analysis. If the JCO has already published the main results on efficacy of a trial and a good companion economic study is done, we would be interested in publishing it. We also receive what we call “What if?” articles. The data on efficacy and toxicity entered into the economic analysis are based on a randomized trial with relatively short follow-up. However, the base case analysis is done by letting the model (often a Markov model) run into the future 10 or 15 years, assuming the benefit of the new therapy continues into the future. If the base case analysis was performed with the actual short follow-up, the C/E ratio is often very high and not acceptable. If the initial, very promising results hold, then with longer follow-up the C/E ratio is favorable. JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 25 NUMBER 6 FEBRUARY 2

PURPOSE To determine the relative efficacy of an intensive cyclophosphamide, epirubicin, and fluorouracil (CEF) adjuvant chemotherapy regimen compared with cyclophosphamide, methotrexate, and fluorouracil (CMF) in node-positive breast cancer. PATIENTS AND METHODS Premenopausal women with node-positive breast cancer were randomly allocated to receive either cyclophosphamide 100 mg/m2 orally days 1 through 14; methotrexate 40 mg/m2 intravenously (i.v.) days 1 and 8; and fluorouracil 600 mg/m2 i.v. days 1 and 8 or cyclophosphomide 75 mg/m2 orally days 1 through 14; epirubicin 60 mg/m2 i.v. days 1 and 8; and fluorouracil 500 mg/m2 i.v. days 1 and 8. Each cycle was administered monthly for 6 months. Patients administered CEF received antibiotic prophylaxis with cotrimoxazole two tablets twice a day for the duration of chemotherapy. RESULTS The median follow-up was 59 months. One hundred sixty-nine of the 359 CMF patients developed recurrence compared with 132 of the 351 CEF patients. The corresponding 5-year relapse-free survival rates were 53% and 63%, respectively (P = .009). One hundred seven CMF patients died compared with 85 CEF patients. The corresponding 5-year actuarial survival rates were 70% and 77%, respectively (P = .03). The rate of hospitalization for febrile neutropenia was 1.1% in the CMF group compared with 8.5% in the CEF group. There was one case of congestive heart failure in a patient who received CMF compared with none in the CEF group. Acute leukemia occurred in five patients in the CEF group. CONCLUSION The results of this trial show the superiority of CEF over CMF in terms of both disease-free and overall survival in premenopausal women with axillary node-positive breast cancer.

James R. Wright, Juravinski Cancer Centre at Hamilton Health Sciences, and Department of Medicine, McMaster University, Hamilton, ON, Canada Sarah Bouma, Juravinski Cancer Centre at Hamilton Health Sciences, Hamilton, ON, Canada Ian Dayes and Jonathan Sussman, Juravinski Cancer Centre at Hamilton Health Sciences, and Department of Medicine, McMaster University, Hamilton, ON, Canada Marko R. Simunovic, Juravinski Cancer Centre at Hamilton Health Sciences, and Department of Surgery, McMaster University, Hamilton, ON, Canada Mark N. Levine, Juravinski Cancer Centre at Hamilton Health Sciences, and Departments of Medicine and Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, ON, Canada Tim J. Whelan, Juravinski Cancer Centre at Hamilton Health Sciences, and Department of Medicine, McMaster University, Hamilton, ON, Canada

PURPOSE To determine the potential economic and policy implications that result from incorporating paclitaxel into first-line therapy for stage 3 and 4 ovarian cancer patients in the province of Ontario, Canada. METHODS A cost-effectiveness analysis was conducted to compare cisplatin/cyclophosphamide (CC), a standard therapy, with cisplatin/paclitaxel (CT). Based on survival curves from a clinical trial, mean costs and survival were calculated. Sensitivity analyses were conducted based on altering the duration of paclitaxel infusion, discount rates, and efficacy of paclitaxel. RESULTS The mean survival duration is prolonged from 2.06 years with the standard therapy to 2.44 years with the paclitaxel combination. The paclitaxel therapy is more expensive, with a mean cost of $17,469 (Canadian) per patient treated with CT compared with $5,228 per patient with CC. The incremental cost-effectiveness ratio is $32,213 per year gained. Sensitivity analyses show that the conclusions remain unchanged. The use of CT as first-line treatment for advanced ovarian cancer patients in Ontario requires an additional $9 million per year over and above the present costs to treat this patient population. CONCLUSION Although paclitaxel-based therapy prolongs survival, it comes at an increased cost. It may not be possible to fund paclitaxel treatment using resources presently allocated to first-line chemotherapy for advanced ovarian cancer. The policy implications for absorbing the cost of paclitaxel in the context of a publicly funded health care system are discussed.

Long-term indwelling central venous catheters are commonlyusedincancerpatientsfortheadministrationof chemotherapy, parenteral nutrition, and blood products and for the facilitation of the drawing of blood. Despite their important role in the delivery of quality patient care, central venous catheters can be associated with morbidity. One of the most common complications is catheter-related venous thrombosis. The extent of thrombosis associated with a central vein catheter can involve the catheter tip (ball-valve clot), the length of the catheter (fibrin sheath), orthecatheterizedvesselintheupperlimb,withorwithout the central vasculature of the neck or mediastinum. 1 In the 1990s, with the increasing use of central venous catheters in cancer patients, there was a clinical impression that catheter thrombosis was a common occurrence, associated with considerable morbidity. The results of two randomized trials provided evidence to support the use of prophylactic antithrombotic agents to prevent catheterassociated thrombosis. In a trial conducted by Monreal et al, 2 patients with a long-term subclavian venous catheter were randomly assigned either to 2,500 U of dalteparin low molecular weight heparin subcutaneously once daily for 90 days or to no prophylaxis. Upper limb venography was performed at 90 days or sooner if symptoms appeared. The trial was stopped early, after upper extremity thrombosis developed in one (6%) of 16 patients treated with low molecularweightheparincomparedwitheight(62%)of13 control patients (P.002). Six of the 29 thrombotic events were symptomatic. There was no increased risk of bleeding associated with the low molecular weight heparin. Bern et al 3 randomly assigned cancer patients with indwelling central venous catheters to either 1 mg of warfarin for 90 days or no treatment. This dosage regimen was based on the demonstrated efficacy of postoperative prophylaxis with 1 mg of warfarin in patients undergoing gynecologic surgery. 4 Venography was performed at 90 days. Onehundredtwenty-onepatientswererandomlyassigned, but only 80 patients completed the trial. Fifteen (37.5%) of 40 control patients developed thrombosis compared with four (9.5%) of 42 patients on warfarin (P .001). Seventeenofthethromboticepisodesweresymptomatic.Inmost cases, the prothrombin time was not prolonged, and there was no increase in bleeding associated with warfarin. After the publication of these studies, although there was lack of consensus on whether patients with central vein catheters should routinely receive thromboprophylaxis, low-dose warfarin was used in some clinical practices. In some patients, the prothrombin time was prolonged excessively because of interactions with chemotherapy drugs. 5 This issue contains two well-conducted randomized controlledtrialsbyVersoetal 6 andCoubanetal 7 evaluating thromboprophylaxis in cancer patients with central venous catheters. Both of these trials were double blind and, hence, provide higher quality evidence than the earlier published randomized trials that were not double blind. The two trials in this issue used different approaches in the choice of outcome for their studies. Verso et al 6 used mandatory venography of the upper limb at 42 days, whether the patient had symptoms or not. The use of this screening technique for venous thrombosis was adopted from prophylactic studies in patients undergoing orthopedic surgery, 8 with venography of the lower limbs at 7 to 10 days after surgery as the standard outcome measure used in clinical trials evaluating antithrombotics in this setting. Asymptomatic thrombosis in the deep veins of the lower limbs is considered a reasonable surrogate for the development of symptomatic venous thromboembolism because 20% of thrombi in the calf veins extend to the proximal deep veins and 50% of patients with proximal deep vein thrombosis have associated pulmonary embolism. Unfortunately, the natural history of catheter-associated thrombosis is not nearly as well defined as postoperative

Patients undergoing rectal cancer surgery may also receive radiation therapy to reduce local tumor recurrence and improve survival [1,2]. In parts of Europe, a 5-day course of preoperative radiotherapy (PRT) alone is recommended for most patients, while other European and North American countries utilize a 5-week course of postoperative chemoradiotherapy for patients with stage II or III tumors [3]. Total mesorectal excision (TME) stresses sharp and complete dissection of the mesorectum, the lymph node– bearing portion of the rectum [4]. Compared with traditional blunt rectal cancer surgery, TME seems to lead to superior rates of sphincter preservation, local tumor control, and survival [4]. The Dutch Colorectal Cancer Group has conducted a randomized trial that compared outcomes of TME, with and without PRT, for patients with resectable rectal cancer [5]. At a median follow-up of 2 years, local recurrence rates were 2.4% and 8.2% for radiated and nonradiated patients, respectively, with no significant difference in overall survival [5]. These results may change with further follow-up. In this issue of the Journal of Clinical Oncology, van den

1016Background: Anthracycline containing chemotherapy (An-CT) has led to improved clinical outcomes in women with early breast cancer (EBC); however cardiotoxicity (CTox) can occur acutely or years...

PURPOSE A recent double-blind, randomized trial demonstrated that very low-dose warfarin (VLDW) reduced the incidence of venous thromboembolism (VTE) without increasing the rate of bleeding in women with metastatic breast cancer receiving chemotherapy. We have evaluated the economic impact on the health care system of using VLDW in such patients. METHODS The records of patients entered onto the trial and a simultaneous, fully allocated, costing model for a tertiary care hospital in Hamilton, Canada were used to determine the difference in costs associated with the care of patients with and without VLDW. RESULTS The cost of providing VLDW was $ 21,854 (Canadian dollars) per 100 patients. This therapy led to a reduction in costs of $ 24,297 per 100 patients, thus saving the health care system $ 2,443 per 100 patients. In the sensitivity analysis, VLDW prophylaxis still did not increase health care costs unless the cost of VLDW was greatly increased, the cost of treating thromboembolic episodes was markedly reduced, or the incidence of either VTE or bleeding with VLDW was increased above the rates observed in the trial. CONCLUSION We conclude that for women receiving chemotherapy for metastatic breast cancer, the benefits of VLDW can be realized without increased health care costs.

296 Background: PETCAM was a randomized trial evaluating the effect of PET-CT compared to conventional imaging (control) on the surgical management of patients with resectable colorectal cancer liver metastases (CRLM). It concluded that PET-CT did not result in frequent change in surgical management (8·0%, 21/263) with only 2·7% (7/263) avoidance of liver resections. In this study we conducted a cost analysis of these two arms up to one year following randomization. Methods: Health care utilization was collected for all study participants. Unit costs for hospitalization, physician services, chemotherapy and outpatient radiological and endoscopic procedures were obtained from administrative databases. Cost analysis was undertaken from the perspective of a third-party payer (i.e., Ministry of Health). Mean cost with its 95% credible interval was estimated using a Bayesian approach. Results: The estimated mean cost per patient in the PET-CT arm was CAN $45,454 (min-max: 1,340-181,420) and in the control arm, CAN $40,859 (min-max: 279-293,558), with a net difference of CAN $4,327, 95% credible interval -2,207 to 10,614. The primary cost driver was cost of hospitalization for liver surgery (+ $2,997 CAN for the PET-CT arm), mainly due to a longer length of hospital stay for the PET-CT arm compared to control (median 7 days vs. 6 days, P= 0·034) and a higher rate of postoperative complications (52/255, 20% vs. 13/128, 10%, P = 0·014). Baseline characteristics were similar between groups, including a similar number of liver segments involved with cancer, number of segments resected and type of liver resection performed. Conclusions: PET-CT does not appear to provide a significant clinical benefit in the surgical management of patients with resectable CRLM and it is not cost saving compared to control.

PURPOSE The purpose of this study was to develop and validate an instrument for clinical trials to measure radiation-related acute morbidity and quality of life from the perspective of patients with head and neck cancer (HNC) treated with radiotherapy. METHODS The Head and Neck Radiotherapy Questionnaire (HNRQ) was developed by a panel of health care workers and patients, was pretested in a pilot study of HNC patients, and was validated in a randomized double-blind trial of concomitant fluorouracil (FUra) infusional therapy (1.2 g/m2 per 24 hours) or saline placebo administered for 72 hours in the first and third weeks of a 6 1/2-week course of radiation therapy. The HNRQ was validated against existing toxicity and performance status indices, all of which were measured weekly for the 6 1/2 weeks of treatment and for 4 weeks posttreatment. RESULTS There were three a priori constructs: (1) that the HNRQ scores would conform to a shallow U-shaped pattern to reflect declining quality of life (increasing morbidity) during radiation and recovery posttreatment; (2) that the HNRQ would correlate with existing toxicity indices (World Health Organization [WHO] stomatitis, Byfield stomatitis, WHO skin toxicity, Eastern Cooperative Oncology Group [ECOG] and Karnofsky performance status); and (3) that the HNRQ would discriminate between FUra and placebo groups. The HNRQ and its domain scores all showed a change from baseline reflecting increased morbidity during radiation (analysis of variance [ANOVA], P < .00001). The HNRQ correlated well with all other indices (r > or = .60), and domain scores correlated best with other indices that assess the same symptom complex (eg, HNRQ skin domain and WHO skin toxicity index, r = .77). There was a significant difference in HNRQ scores between the FUra and placebo groups during radiation (ANOVA, P = .0007), and all HNRQ domains also discriminated between the treatment groups. CONCLUSION The HNRQ is a valid measure of acute morbidity due to radiation therapy in patients with locally advanced HNC, and may be useful as an outcome measure for future clinical trials of radiation treatment strategies.

5523 Background: Women with locally advanced cervical cancer (LACC) are treated with concurrent chemotherapy and radiation therapy (CRT) with curative intent. Pre-treatment staging is important to ...

Accurate staging is important in the clinical management of nasopharyngeal carcinoma (NPC), which is endemic in southern China. The usual staging consists of magnetic resonance imaging of the head and neck followed by conventional imaging (CI) with chest x-ray, bone scan, and abdominal ultrasound to look for distant metastases. The potential for []F-fluorodeoxyglucose (FDG) positron emission tomography and computed tomography (PET-CT) to detect distant metastases in patients with negative CI tests has been reported in other jurisdictions. PET-CT is expensive and likely a limited resource in China. Tang et al describe a diagnostic testing strategy based on an approach that stratifies patients into different risk groups and then applies PET-CT selectively, with the goal of providing the greatest benefit for the most patients. In a prospective cohort study, patients with NPC underwent CI and PET-CT. Any abnormality suspicious for metastases was confirmed by biopsy, additional imaging, or if neither were done, 12 months of metastases-free follow-up. Any patient with a negative battery of tests who developed a distant metastasis within 12 months was considered to have had false-negative imaging tests. Imaging tests were read by experienced physicians without knowledge of the clinical course. The optimal design would have been a randomized trial in which patients with NPC were allocated to CI or CI plus PET-CT, after which an unbiased assessment of the added value of PET-CT to CI could have been obtained. The limitation of the cohort design is that there is no independent reference standard; the tests under study are used to define the outcome. Because PET-CT diagnosed more than twice as many patients with distant metastases as did CI, it is likely that the added value of PET-CT is overestimated. In the cohort of 583 patients, 86 (15%) were found to have distant metastases: 41 by PET-CT alone, one by CI alone, 30 by PET-CT and CI, and 14 by other tests (ie, 71 [83%] by PET-CT and 31 [36%] by CI). The sensitivity, specificity, positive predictive value, and negative predictive value of PET-CT for distant metastases were 83%, 98%, 88%, and 97%, respectively. The corresponding values for CI were 36%, 98%, 80%, and 90%. In developing their case, the investigators focus on “the positive yield” as the number of true-positive cases of the diagnostic test as a proportion of the total number of patients who underwent imaging (ie, 12.2% for PET-CT v 5.3% for CI). They present data to support that both nodal stage and Epstein-Barr virus (EBV) titer are associated with distant metastases, according to the following criteria: presence of N2 or N3 disease, 76% of patients with distant metastases and EBV DNA 4000 copies/mL, 80% with distant metastases. They then use both of these factors to define patient populations with higher prevalence of distant metastases in whom the positive yield with PET-CT is relatively high. For example, in patients with N2 or N3 disease, the positive yield becomes 62% compared with 27% with CI, and in patients with high EBV DNA levels, it is 22% versus 9% with CI. Finally, they combine both nodal status and EBV titer to define three risk groups (very low, low, and intermediate) and show that there is little to be gained over CI by performing PET-CT on the very low risk group. This approach works best when the false-positive fraction is low, but there will likely be a trade-off in terms of an increased falsenegative fraction. In 2006, Journal of Clinical Oncology published an editorial entitled, “Economic Evaluation in the Journal of Clinical Oncology: Past, Present, and Future,” which described a hierarchy of economic analyses from cost minimization (weakest) to cost effectiveness (strongest), as well as the type of economics articles the journal editors would be most interested in. A cost minimization study assumes that the outcomes of two interventions are the same and the difference in the cost of the two alternatives is all that matters. Presumably, the cheapest intervention would be chosen. The weakness with this design is that it ignores the downstream costs of an intervention. In a cost effectiveness study, the costs of both the interventions and what happens to the patient after the intervention are collected for both interventions. The ratio of the difference in costs between therapies divided by the difference in effects (eg, survival) is called the “incremental cost effectiveness ratio.” The economic analysis in the study by Tang can be classified as cost minimization. The costs were incomplete as they were limited to the cost of the imaging tests. The details of all the consequences of the imaging tests and associated costs were lacking (eg, standard therapy for NPC if no distant metastases detected, type of therapy if metastases detected on staging, and extra imaging tests and biopsies required for both true positives and false positives). On the basis of such limitations, care must be taken in making any inferences in terms of any economic advantage of restricting PET-CT to high-prevalence strata. There is no doubt that the economic comparison of CI plus PET-CT versus CI alone is complex. One way of handling such a comparison in an analytic sense would be to begin with a Markov model, which would lay out all the interventions, consequences, and outcomes. Then probabilities and costs are added to the model. JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 31 NUMBER 23 AUGUST 1

583 Background: The OncotypeDX 21-gene assay (ODX) is used to aid in decision-making for chemotherapy (CT) in patients with breast cancer (BC). The province of Ontario, with a population of approxi...

8551 Background: Alemtuzumab is an anti CD52 monoclonal antibody with single agent activity in relapsed peripheral T cell lymphomas (PTCL), but the optimal dose/schedule to combine with chemotherap...

11019 Background: Breast cancer intrinsic subtypes defined by the 50-gene PAM50 test were reported to be prognostic and predictive in endocrine therapy trials. MA21 randomized 2,104 patients to rec...

574 Background: Biomarkers of bone turnover, including urine N-telopeptide (uNTx) and serum C-telopeptide (sCTx) reflect tumor-related bone breakdown and have been used as surrogate measures of response to therapy in trials. Vascular endothelial growth factor (VEGF) levels correlate with extent of bone metastases (BM). We assessed whether vandetanib, an inhibitor of VEGF, epidermal growth factor receptor and RET signalling, improved uNTx response when added to fulvestrant (F) in patients with BM. Methods: Postmenopausal patients with bone only, or bone predominant, hormone receptor positive MBC were randomised to F (500mg IM day 1, 14, 28, then monthly) with either vandetanib (100mg PO OD) (FV) or placebo (FP) until progression. The primary objective was uNTx response (>30% reduction from baseline). uNTx was collected at baseline, weekly to wk 4, at wk 12 and then every 12 wks. Secondary objectives included PFS, OS, RECIST response, pain and toxicity. Results: 61 patients were allocated to FV and 68 to FP. Median age was 59. 18% had received 1 prior chemotherapy regime and 73% prior endocrine therapy for MBC. uNTx response (n=124 pts) was 64% for FV vs. 52% for FP (p=0.20). No difference was detected between groups for median PFS; 6 months for FV vs. 4.8 months for FP, HR=0.93 (95% CI: 0.64 to 1.36). 16 patients died in FV arm vs. 21 in the FP arm, HR=0.71 (95% CI: 0.37 to 1.36). For those patients with measurable disease, clinical benefit rates were 41% and 43%, respectively, p=1.00. Serious adverse events were similar, 3.3% for FV vs. 5.9% for FP. Elevated baseline uNTx (>65 nM BCE/mmol Cr) was prognostic for PFS, HR=1.62 (95% CI: 1.08 to 2.43 and for mortality, HR= 2.4 (95% CI: 1.2 to 4.6). In an exploratory analysis uNTx was predictive of responsiveness to FV for PFS, HR=0.60 when uNTx >65 vs. HR=1.37 when uNTx

6049 Background: The clinical utility of PET/CT in patients with suspected cancer recurrence remains unclear. The aim of this multi-center, prospective, comparative effectiveness study is to assess the impact of PET/CT on clinical management of patients with suspected cancer recurrence. Methods: Patients were eligible if cancer recurrence (non-small cell lung, breast, head and neck, ovarian, esophageal, Hodgkin’s or non-Hodgkin’s lymphoma) was clinically suspected, and if conventional imaging (e.g. X-ray, ultrasound, CT, or MRI) was non-diagnostic. As a pre-requisite to PET/CT booking, clinicians were asked at enrolment to indicate their planned management if PET/CT were not available. Patients then underwent 18FDG-PET/CT. Clinicians were then asked to indicate their management plan based on PET/CT findings. Patients were followed up once at 3 months. The primary outcome was change in planned management after PET/CT and was assessed independently and in duplicate by external outcome adjudicators using all available source documents. Results: 101 patients (mean age 64 y, 45% male, median 1.3 y since last treatment) were enrolled from 4 centers in Ontario, Canada between April 2009 and June 2011. Distribution of tumor types was: non-small cell lung (55%), breast (19%), ovarian (10%), esophageal (6%), lymphoma (6%), head and neck (4%). 8 patients did not complete the study (non-adherence to protocol, 2; death, 5; disease progression prior to PET/CT, 1), of whom 2 did not receive PET/CT. PET/CT changed planned management in 52 (53%) patients (Table). At 3 months, planned management was carried out in 46/52 (88%) patients. Conclusions: In patients with suspected cancer recurrence, PET/CT changes planned management from non-treatment to treatment for approximately 1 in every 3 patients (“number needed to scan” = 3) and contributes importantly to clinical management. [Table: see text]

390 Background: An increasing proportion of patients with liver CAM are undergoing hepatic resection with curative intent. Detection of occult metastastic disease is important in this setting, and PET/CT is used to identify patients with either inoperable, or limited resectable extrahepatic disease not identified with conventional imaging. We recently reported the results related to the primary objective of PETCAM, a randomized trial for patients with resectable liver CAM, where 7.6% had a change in management based on PET/CT (ASCO 2011). As a secondary objective, we examined whether patients undergoing PET/CT derived a survival benefit, as they may be considered to be better selected. Methods: PETCAM was a multicenter trial with 404 subjects randomized 2:1 to receive PET/CT or no PET/CT once they were assessed by a hepatobiliary surgeon to have resectable CAM. Subjects were followed over 4 years for overall survival (OS). As well as the intervention, we considered the SUV, Fong Score and other baseline factors as predictors for OS. Results: After a median 2.8 years of follow-up, 107 of the 404 (26%) study subjects had died. The 270 PET/CT subjects [and the 245 of these who underwent surgery] showed no statistically significant survival advantage over the 134 No PET/CT subjects [123 who underwent surgery] with a hazard ratio (HR) of 0.85, 95% confidence interval (CI): 0.57 to 1.3; p=0.41 [HR=0.81, 95% CI: 0.52 to 1.3; p=0.34]. Fong score was a strong predictor of OS for all patients (HR=1.4, 95% CI: 1.1 to 1.6), and for those who had surgery (HR=1.4, 95% CI, 1.1 to 1.7). In the PET/CT arm, SUV is strongly predictive of OS [HR per unit SUV increase =1.11, 95% CI: 1.05 to 1.18; p=0.0007. Conclusions: The addition of PET/CT had no effect on improving overall survival for patients who did or did not have hepatic resection for CAM. SUV is strongly predictive of survival for patients who had PET/CT, and Fong score is predictive of survival for all patients. The overall role of PET/CT in patients with resectable CAM appears limited and should be reevaluated in the current era of comparative effectiveness research and health care cost containment.