To report the influence of the posttreatment PSA nadir response at 2 years after high-dose external beam radiotherapy on distant metastases (DM) and cause-specific mortality (CSM). Between 1988 and 2000, 845 patients with clinically localized prostate cancer were treated with 3-dimensional conformal radiotherapy (3D-CRT) and intensity modulated radiotherapy (IMRT). All patients treated with neo-adjuvant or concurrent androgen deprivation therapy were excluded. Patients were included if >5 serial PSA measurements during the follow-up period were available. The median prescribed radiation dose level was 75.6 Gy (range, 66 Gy-86.4 Gy). The median follow-up time was 7.7 years (range, 2-16 years). A PSA relapse was defined according to the nadir plus 2 ng/mL definition. A fixed landmark (LM) time-point at 2-years was used to assess the influence of the nadir PSA as a time dependent variable on the long-term survival outcomes. The 5-year PSA relapse-free survival rates from the 2-year landmark time point for patients with nadir PSA ≤1.5 ng/mL and >1.5 ng/mL were 71% versus 12%, respectively (p < 0.0001). Multivariate analysis demonstrated that the nadir PSA of 1.5 ng/mL at the landmark time-point was the strongest predictor of long-term PSA relapse-free survival. The 5-year cumulative incidence of DM was 9% in patients whose nadir PSA levels were ≤1.5 ng/mL at the 2-year LM, and was twice as high as patients whose nadir PSA values were >1.5 ng/mL at this time point. Multivariate competing risk regression model showed that the nadir PSA reached at 2-year (p = 0.002) as well as Gleason score (p < 0.001) and higher clinical stage (p = 0.04) were predictors of distant metastases-free survival. The 5-year CSM was 5.7% in patients whose nadir PSA values were ≤1.5 ng/mL and 10.3% among patients whose nadir were >1.5 ng/mL at the 2-year LM (p < 0.001). Multivariate competing risk regression model showed that higher Gleason scores (p = 0.002) and higher pre-RT PSA values (p = 0.03) were independent risk factors associated with CSS after adjusted for T-stage and nadir PSA. A trend was observed for increased prostate cancer mortality among those with higher PSA nadir levels at the 2-year landmark (p = 0.08). In a stratified competing risk analysis, nadir PSA was significantly associated with prostate cancer mortality (p = 0.02). The PSA nadir levels of ≤1.5 ng/mL at 2-years after external beam radiotherapy for prostate cancer predicts for long-term distant metastases-free survival and cause-specific mortality outcomes. Patients with higher absolute nadir-levels at 2-years after treatment should be evaluated for the presence of non-responding disease and earlier salvage treatment interventions should be considered in these patients.

Induction therapy provides a unique opportunity to evaluate therapies in patients with NSCLC to personalize care and facilitate drug research. We designed this 50 patient prospective trial to correlate response with gefitinib measured by CT with mutations in EGFR exons 19 and 21 and to simultaneously assess the agent’s ability to induce regressions preoperatively in persons with NSCLC. To do this, we enriched the group studied to select individuals with tumors more likely to harbor EGFR mutations. Patients with stage I or II NSCLC have a baseline chest CT and a core-needle biopsy to detect EGFR mutations. All participants smoked cigarettes 10 pack years and/or had tumors with bronchioloalveolar features. All receive gefitinib 250 mg daily. After 21 days, the CT is repeated and resection follows. Surgical specimens are again assayed for EGFR mutations and KRAS mutations. Patients with mutation and/or 25% regression (WHO) are given gefitinib for 2 years. Among 34 patients enrolled to date, EGFR mutations were detected in 12(35%); results were identical in 17/17 matched core biopsy and resection specimens. Thirteen (38%) had a 25% bidimensional tumor reduction after 21 days. The mutation rate was 77% in responding patients and 10% in patients without response, P=0.0001 for the primary study endpoint. The regression observed after 21 days of gefitinib and the correlation with EGFR mutation for each patient is shown in the figure. 0/3 patients with KRAS mutations had response or EGFR mutation. No increase in perioperative complications occurred. 15 patients received gefitinib postoperatively. In this ongoing trial, 1) The rate of EGFR mutation was significantly higher in patients with response to gefitinib. 2) Our enrichment strategy resulted in the detection of three times the expected number of EGFR mutations in USA patients with NSCLC. 3) Gefitinib sensitivity can be assessed by CT in 21 days. 4) Results of mutation detection were identical in pre- and post-treatment specimens. Support: CA05826, CA113653.

7686 Background: Genomic analyses of lung adenocarcinoma have yielded striking advances that are already impacting on clinical management. Further advances in understanding the biological heterogeneity of this disease will require integration of multiple types of genomic data. To this end, we have assembled a large integrated genomics dataset of lung adenocarcinomas. Here, we highlight one of the novel findings emerging from its initial analysis. Methods: 173 primary lung adenocarcinomas were included in this analysis. Profiling of genomic gains and losses was done by array comparative genomic hybridization (aCGH) on Agilent 44K arrays. Expression profiling was based on Affymetrix U133A arrays. The dataset was annotated for EGFR mutations (exon 19 deletion and L858R) by sensitive PCR-based assays and for KRAS mutations by sequencing. Results: By unsupervised analysis of the aCGH data, the 173 tumors clustered robustly into two or three patterns of co-occurring gains and losses. One aCGH cluster was strongly associated with EGFR mutation (p<10−4) and was characterized by 7p gains (in the EGFR region) and 8p losses. Remarkably, by expression profiling the most consistently underexpressed gene (p<10−9) in EGFR mutant cases compared to EGFR wild type cases was a MAPK phosphatase gene at 8p12, DUSP4 (MKP-2). The DUSP4 region showed genomic loss in 27/35 EGFR mutant cases vs 47/138 non-mutated cases (p<10−4). A limited screen (n=11) has so far revealed no DUSP4 mutations. Western blotting shows low DUSP4 in most EGFR mutant lines, compared to KRAS mutant lines. Conclusions: EGFR mutations in lung adenocarcinomas are strongly associated with genomic loss and low expression of DUSP4. DUSPs are known to be transcriptionally upregulated by MAPK signaling as a negative feedback mechanism and DUSP family members are emerging as putative tumor suppressors in other cancers. We hypothesize that DUSP4 loss cooperates with EGFR mutation to allow full oncogenic activation of the MAPK pathway. Functional studies of DUSP4 in lung adenocarcinoma cell lines are in progress. Our data highlight the value of large, integrated, highly annotated genomic datasets in generating novel insights and hypotheses. No significant financial relationships to disclose.

In recent years numerous investigators have conducted genetic studies of pairs of tumor specimens from the same patient to determine whether the tumors share a clonal origin. These studies have the potential to be of considerable clinical significance, especially in clinical settings where the distinction of a new primary cancer and metastatic spread of a previous cancer would lead to radically different indications for treatment. Studies of clonality have typically involved comparison of the patterns of somatic mutations in the tumors at candidate genetic loci to see if the patterns are sufficiently similar to indicate a clonal origin. More recently, some investigators have explored the use of array CGH for this purpose. Standard clustering approaches have been used to analyze the data, but these existing statistical methods are not suited to this problem due to the paired nature of the data, and the fact that there exists no “gold standard” diagnosis to provide a definitive determination of which pairs are clonal and which pairs are of independent origin. In this article we propose a new statistical method that focuses on the individual allelic gains or losses that have been identified in both tumors, and a statistical test is developed that assesses the degree of matching of the locations of the markers that indicate the endpoints of the allelic change. The validity and statistical power of the test is evaluated, and it is shown to be a promising approach for establishing clonality in tumor samples.

15070 Background: Our recent analyses (JCO, in press) showed that residual nodal disease but not T-stage predicted survival after chemo-radiotherapy (CRT) and surgery for esophageal adenocarcinoma (AC). In this study, we investigated prognostic factors for esophageal squamous cell carcinoma (SCC) after CRT. Methods: Retrospective review of patients with esophageal SCC who had CRT and esophagectomy. Data collected: demographics, CRT details, pathologic findings, and survival. Statistical methods included recursive partitioning (RP) and Kaplan-Meier (KM) analyses. Results: Patients included in the study were treated between 1/1996 and 2/2006. Follow up was thru 5/06. 91 patients were appropriate for analysis. There were 56 men (61.5%) and 35 women (38.5%). 72 (79.1%) patients had clinical regional disease prior to treatment, while the rest had locally advanced disease. Median radiation dose was 5040 cGy, and 78 (85.7%) patients received cisplatin based chemotherapy. 49 (53.8%) patients had a complete local pathologic response (pCR), including 10/91 (10.9%) who had a pCR with residual nodal disease. 42 (46.2%) patients had residual local disease. RP analysis identified 3 prognostic groups: a) Group 1 (n=52), patients with minimal residual local disease (pCR & T1- regardless of nodal status), b) Group 2 (n=28), patients with residual T2 disease (N0 and N1) as well as patients with T3–4N0 disease, and c) Group 3 (n=11), patients with residual T3–4N1 disease. 3-year survival by KM analysis was 68.4% in group 1, 45.6% in group 2, and 0 % in group 3 (p<0.001). Conclusions: Unlike adenocarcinoma of the esophagus where residual nodal disease after CRT is the most significant predictor of survival, in SCC of the esophagus, the presence of minimal residual local disease after CRT, regardless of nodal status, predicts the best survival. The implications of these findings might include establishing different endpoints to assess response to treatment and prognostic criteria. No significant financial relationships to disclose.

MOTIVATION Array CGH technologies enable the simultaneous measurement of DNA copy number for thousands of sites on a genome. We developed the circular binary segmentation (CBS) algorithm to divide the genome into regions of equal copy number. The algorithm tests for change-points using a maximal t-statistic with a permutation reference distribution to obtain the corresponding P-value. The number of computations required for the maximal test statistic is O(N2), where N is the number of markers. This makes the full permutation approach computationally prohibitive for the newer arrays that contain tens of thousands markers and highlights the need for a faster algorithm. RESULTS We present a hybrid approach to obtain the P-value of the test statistic in linear time. We also introduce a rule for stopping early when there is strong evidence for the presence of a change. We show through simulations that the hybrid approach provides a substantial gain in speed with only a negligible loss in accuracy and that the stopping rule further increases speed. We also present the analyses of array CGH data from breast cancer cell lines to show the impact of the new approaches on the analysis of real data. AVAILABILITY An R version of the CBS algorithm has been implemented in the "DNAcopy" package of the Bioconductor project. The proposed hybrid method for the P-value is available in version 1.2.1 or higher and the stopping rule for declaring a change early is available in version 1.5.1 or higher.

This phase I study sought to determine the toxicity profile, pharmacokinetics, and antitumor activity of giving carboplatin every 3 weeks and paclitaxel weekly in patients with relapsed ovarian cancer. Eligible patients with relapsed epithelial ovarian cancer and prior treatment with platinum- and paclitaxel-based therapy were treated with an escalating regimen of carboplatin (day 1) at an area under the curve (AUC) of 4–6 and 1-h infusions of paclitaxel (days 1, 8, and 15) at 50–80 mg/m 2 cycled at 3-week intervals. Pharmacokinetic studies were performed on the first day of cycles 1 and 2. All patients had a platinum-free interval of greater than 6 months from the most recent platinum treatment. A total of 77 cycles were administered to 16 patients, with a similar median number of cycles per patient at each dose level varying from 4.6 to 5.3. Febrile neutropenia and grade 4 thrombocytopenia were the dose-limiting toxicities at dose levels 3 and 4 after the third cycle, with no mucositis, nausea, vomiting, or peripheral neuropathy observed greater than grade 2. The maximum tolerated dose of carboplatin was an AUC of 5 and 80 mg/m 2 for paclitaxel. Pharmacokinetic analysis showed a marginal statistical difference with regard to reduced systemic paclitaxel concentration after cycle 2 compared with cycle 1 ( P = 0.06). Of nine patients evaluable for a radiographic response, the response rate was 66.6% with a complete response of 33.3%. All five patients with nonmeasurable disease achieved a biochemical response. The combination of carboplatin given every 3 weeks at an AUC of 5 and 1-h weekly paclitaxel at 80 mg/m 2 is a feasible and reasonably well-tolerated regimen and may have significant antitumor activity in relapsed ovarian cancer patients.

PurposeLung adenocarcinomas with mutations in exons 19 and 21 of the epidermal growth factor receptor gene (EGFR) demonstrate sensitivity to gefitinib or erlotinib. Investigators have reported an association between EGFR mutations and the amount and duration of cigarette smoking, with the highest incidence of mutations seen in never smokers.MethodsEGFR exon 19 and 21 mutation status was determined in 265 tumor samples using direct sequencing, polymerase chain reaction (PCR), or PCR-based restriction fragment length polymorphism analysis. A detailed smoking history was obtained. Patients were categorized as never smokers (< 100 lifetime cigarettes), former smokers (quit ≥ 1 year ago), or current smokers (quit < 1 year ago).ResultsWe detected EGFR mutations in 34 (51%) of 67 never smokers (95% CI, 38% to 64%), 29 (19%) of 151 former smokers (95% CI, 13% to 27%), and two (4%) of 47 current smokers (95% CI, 1% to 16%). Significantly fewer EGFR mutations were found in people who smoked for more than 15 pack-years (P < .001) or stopped smoking less than 25 years ago (P < .02) compared with individuals who never smoked. The number of smoking pack-years and smoke-free years predicted the prevalence of EGFR mutations (areas under receiver operating characteristic curve = 0.78 and 0.77, respectively).ConclusionThe likelihood of EGFR mutations in exons 19 and 21 decreases as the number of pack-years increases. Mutations were less common in people who smoked for more than 15 pack-years or who stopped smoking cigarettes less than 25 years ago. These data can assist clinicians in assessing the likelihood of exon 19 and 21 EGFR mutations in patients with lung adenocarcinoma when mutational analysis is not feasible.

5554 Background: RAI-refractory thyroid cancer carries a poor prognosis, and no effective systemic therapy exists. DEP is a potent histone deacetylase inhibitor with broad in vitro and phase I anti-tumor activity, and has restored sodium-iodide symporter expression and RAI avidity in thyroid cancer cell lines. Methods: Eligible pts have confirmed papillary, follicular, or Hürthle cell carcinoma, progressive measurable disease, and adequate hepatic/renal function. All pts are RAI-refractory by diagnostic RAI whole body scan (WBS). Prior chemotherapy in the recurrent/metastatic setting, and history of significant cardiac disease are exclusions. DEP 13 mg/m2 IV is given on day 1, 8, and 15, every 28 days. A novel Simon 2-stage design specifies an initial accrual of 21 pts, with expansion to 41 total pts if 2 RECIST responses or 6 pts with PFS at 6 months are observed. Primary outcome is RECIST response rate. Change in RAI avidity (by diagnostic RAI WBS) is a secondary outcome. Results: To date, 14 pts have been enrolled: female 50%, median age 62 (42–78), median Karnofsky 90 (70–100), papillary/follicular/Hürthle histology (7/1/6). Grade 3–5 toxicities possibly related to drug: grade 5 sudden death (1); no grade 4 toxicity; grade 3 toxicity - fatigue (1), dysphagia (1), dyspnea (1). No grade 3 thrombocytopenia (phase I DLT) has been observed. Grade 1 (6) and grade 2 (4) fatigue have been common. As a result of the grade 5 event, protocol accrual and treatment were suspended temporarily. After NCI-CTEP review, more restrictive cardiac exclusions were added and protocol accrual was resumed. Response at 2 months: stable disease (4); progression (2); inevaluable (6, due to protocol suspension); withdrew consent (1); too early (1). In a proof-of-principle event, clinically significant restoration of RAI avidity was observed in a 78 year-old woman with papillary carcinoma. Per protocol, she was taken off study and given treatment-dose RAI; post-therapy WBS confirmed dramatic RAI avidity. Conclusions: Preliminary signs of in vivo reversal of RAI resistance have been observed. Accrual continues with more restrictive cardiac exclusions, and close surveillance for cardiac events. No significant financial relationships to disclose.

OBJECTIVES:To determine the incidence and prognostic implications of positive mesorectal lymph nodes in patients undergoing total pelvic exenteration for recurrent gynecologic malignancies.METHODS:We performed a retrospective chart review of all patients who had undergone total pelvic exenteration for a gynecologic malignancy between July 1992 and December 2003. Patient charts were reviewed for information regarding demographics, site of cancer, histology, pathology report, and time to recurrence.RESULTS:Fifty-eight women had undergone total pelvic exenteration for recurrent gynecologic malignancies during the study period and 57 were available for analysis. Primary cancer site was as follows: cervix, 37 (65%); vagina, 8 (14%); vulva, 5 (9%); and uterine corpus, 7 (12%). In 30 patients (53%), the mesorectal lymph node status was pathologically evaluated. Of these 30 patients, 3 (10%) had positive mesorectal lymph nodes at the time of total pelvic exenteration. All 3 patients had rectal wall involvement (rectal submucosa, 2; rectal mucosa, 1), and all 3 patients recurred within 4 months of pelvic exenteration. The median time to recurrence after surgery was 2.4 months in those patients with positive mesorectal lymph nodes compared with 7.3 months in those with negative mesorectal lymph nodes (P = 0.005). When individually adjusted for other prognostic variables, such as margin status, tumor grade, lymphovascular space involvement, primary cancer site, and histologic type, a finding of positive mesorectal lymph nodes was associated with a shorter time to recurrence of disease (all P < 0.05).CONCLUSIONS:Mesorectal lymph node involvement is a common finding at total pelvic exenteration, particularly in patients with rectal wall involvement. Patients with positive mesorectal lymph nodes appear to have a worse outcome with a shorter time to recurrence of disease.

Background. Chest wall resections are associated with significant morbidity, with respiratory failure in as many as 27% of patients. We hypothesized that our selective use of a rigid prosthesis for reconstruction reduces respiratory complications.Methods. The records of all patients undergoing chest wall resection and reconstruction were reviewed. Patient demographics, use of preoperative therapy, the location and size of the chest wall defect, performance of lung resection if any, the type of prosthesis, and postoperative complications were recorded. Predictor of complications were identified by chi(2) and logistic regression analyses.Results. From January 1, 1995, to July 1, 2003, 262 patients (median age, 60 years) underwent chest wall resection for tumor in 251 (96%), radiation necrosis in 7 (2.7%); and infection in 4 patients (1.3%). The median defect size was 80 cm(2) (range, 2.7 to 1,200 cm(2)) and the median number of ribs resected was 3 (range, 1 to 8). Major lung resection was performed in 85 patients (34%). Prosthetic reconstruction was rigid (polypropylene mesh/methylmethacrylate composite) in 112 (42.7%), nonrigid (polytetrafluoroethylene or polypropylene mesh) in 97 (37%), and none in 53 patients. Postoperatively, 10 patients died (3.8%), 4 of whom had pneumonectomy plus chest wall resection. Respiratory failure occurred in 8 patients (3.1%). By multivariate analysis, the size of the chest wall defect was the most significant predictor of complications.Conclusions. Our incidence of respiratory failure is lower than previously reported and may relate to our use of rigid repair for defects likely to cause a flail segment. Pneumonectomy plus chest wall resection should be performed only in highly selected patients.

In recent years health services researchers have conducted 'volume-outcome' studies to evaluate whether providers (hospitals or surgeons) who treat many patients for a specialized condition have better outcomes than those that treat few patients. These studies and the inherent clustering of events by provider present an unusual statistical problem. The volume-outcome setting is unique in that 'volume' reflects both the primary factor under study and also the cluster size. Consequently, the assumptions inherent in the use of available methods that correct for clustering might be violated in this setting. To address this issue, we investigate via simulation the properties of three estimation procedures for the analysis of cluster correlated data, specifically in the context of volume-outcome studies. We examine and compare the validity and efficiency of widely-available statistical techniques that have been used in the context of volume-outcome studies: generalized estimating equations (GEE) using both the independence and exchangeable correlation structures; random effects models; and the weighted GEE approach proposed by Williamson et al. (Biometrics 2003; 59:36-42) to account for informative clustering. Using data generated either from an underlying true random effects model or a cluster correlated model we show that both the random effects and the GEE with an exchangeable correlation structure have generally good properties, with relatively low bias for estimating the volume parameter and its variance. By contrast, the cluster weighted GEE method is inefficient.

A variety of agents have emerged to treat patients with recurrent epithelial ovarian cancer (EOC). Most patients receive both topotecan (T) and liposomal doxorubicin (D); however, there are no data regarding the benefit of a sequence-D followed by T (DT) or T followed by D (TD). We identified 89 consecutive patients with recurrent EOC, who received both D and T from January 1994 to January 2004 at Memorial Hospital. We compared the duration of treatment, toxicity, and overall survival (OS) for patients who received either DT or TD. Sixty-four patients received DT, and 25 patients received TD. The groups were balanced regarding age, stage, surgical debulking, platinum sensitivity, prior therapy, and intervening drugs between D and T. Median numbers of cycles on DT and TD were seven and six, respectively (P= 0.61); there was no difference in duration based on platinum sensitivity. Removal from therapy for toxicity was similar, DT (22%) and TD (36%) (P= 0.18). Finally, there was no difference in median OS based on sequence, DT (18.28 months) and TD (17.75 months) (P= NS). Platinum sensitivity did not affect median OS based on sequence. Based on duration, toxicity, and OS there is no advantage of one sequence of D and T when treating patients with recurrent EOC.

7021 Background: Induction therapy provides a unique opportunity for the objective assessment of therapies in patients with NSCLC to facilitate care and advance research. We designed this 40 patient prospective trial to measure gefitinib's ability to induce regressions preoperatively in individuals with NSCLC and to simultaneously correlate response with mutations in EGFR exons 19 and 21. Patients with 'in-vivo' gefitinib sensitivity and/or mutations receive gefitinib postoperatively as well. To facilitate our goals, we enriched the population studied to select individuals with tumors more likely to harbor mutations in EGFR. METHODS At diagnosis, patients with stage I or II NSCLC had baseline chest CT imaging and a core-needle biopsy to detect EGFR mutations. All participants smoked cigarettes ≤ 10 pack years and/or had tumors with bronchioloalveolar features. All received gefitinib 250 mg daily. After 21 days, CT imaging was repeated and resection followed. Surgical specimens were agained assayed for EGFR mutations. Patients with mutation and/or a ≥ 25% regression (WHO) were given gefitinib 250 mg for 2 years. RESULTS 20 patients enrolled to date. EGFR mutations were detected in both pre and post gefitinib specimens in 4/19 tested (21%). Five of 19 (26%) had a ≥ 25% bidimensional tumor reduction after 3 weeks. Regressions ≥ 25% were seen in 2/4 (50%) with exon 19 and 21 EGFR mutations and 3/15 (20%) with WT EGFR. The other 2 patients with mutations had 11% and 23% tumor regressions. The 6 patients who had either a ≥ 25% lesion reduction and/or mutation received gefitinib postoperatively. No perioperative complications related to gefitinib occurred. All patients are relapse free. CONCLUSIONS In this ongoing trial, 1) Our 'enrichment strategy' resulted in the detection of twice the expected number of EGFR mutations in USA patients with NSCLC. 2) Gefitinib sensitivity can be assessed by CT in 3 weeks. 3) Results of mutation detection were identical in pre- and post-treatment specimens. 4) The rate of response is numerically higher in patients with EGFR mutations in exons 19 and 21. 5) We are using the surgical specimens to elucidate determinants of gefitinib sensitivity in the absence of EGFR mutation. Support: CA05826, CA113653. [Table: see text].

Background. As the general population ages, it becomes increasingly important to understand the potential contribution of chronologic age to mortality after esophagectomy. Because this risk is poorly defined, we sought to determine whether extreme age (> 80 years) is an independent risk factor after esophagectomy.Methods. We analyzed a prospectively maintained, single-institution database of 858 consecutive patients who underwent esophagectomy between January 1996 and May 2005. Data evaluated included patient demographics, medical comorbidity, types of resections performed, length of stay, postoperative adverse events, and overall survival. We used univariate, multivariate, and Kaplan - Meier analysis to determine the influence of age on postoperative morbidity, in-hospital survival, and overall survival.Results. Of 858 patients, 31 (10 female, 21 male) were older than 80 years of age. Preliminary analysis indicated that patients younger than 50 years (n = 107) had significantly fewer comorbidities; these were excluded from analysis. In the remaining 751 patients, the age older than 80 cohort was compared with patients aged 50 to 79. Patients aged 50 to 79 were grouped because of similar characteristics (length of stay, hospital death). There were no significant differences in comorbidities, types of resections, or postoperative complication type or severity between the two groups. Postoperative death, length of stay, and survival, however, were significantly worse in patients older than 80. In a logistic regression model controlling for comorbidity, age older than 80 was significantly associated with increased perioperative mortality (hazard-ratio, 3.9; p < 0.01).Conclusions. Patients older than 80 years have increased mortality risk after esophagectomy, independent of comorbidity. Octogenarian status should be a consideration in the management of these patients.

Purpose: In patients with non-small cell lung cancer (NSCLC), mutations in the epidermal growth factor receptor (EGFR) tyrosine kinase domain have been associated with sensitivity to erlotinib and gefitinib. We undertook this study to explore the relationship between EGFR mutation type and clinical variables, including treatment with gefitinib and erlotinib.Experimental Design: In patients with NSCLC, EGFR exon 19 deletion mutations and EGFR L858R point mutations were analyzed by nonsequencing PCR-based methods from paraffin blocks of tissue obtained before treatment. The results were correlated with clinical information (sex, pathologic subtype, race/ethnicity, treatment, and overall survival).Results: The two most common EGFR mutations were identified in 24% (70 of 291; 95% confidence interval, 26%-38%) of tumors from patients with NSCLC. EGFR mutation was associated with Asian ethnicity (P = 0.0023) and being a "never smoker" (P = 0.0001). Among patients with EGFR mutations, 39% (27 of 70) had EGFR L858R, whereas 61% (43 of 70) had an EGFR exon 19 deletion. After treatment with erlotinib (n = 12) or gefitinib (n = 22), patients with EGFR mutations had a median overall survival of 20 months. After treatment with erlotinib or gefitinib, patients with EGFR exon 19 deletions had significantly longer median survival than patients with EGFR L858R (34 versus 8 months; log-rank P = 0.01).Conclusions: EGFR mutations in exons 19 or 21 are correlated with clinical factors predictive of response to gefitinib and erlotinib. Those with EGFR exon 19 deletion mutations had a longer median survival than patients with EGFR L858R point mutation. These observations warrant confirmation in a prospective study and exploration of the biological mechanisms of the differences between the two major EGFR mutations.

7123 Background: Measuring DNA in plasma using quantitative RT-PCR is a promising method for surveillance of NSCLC. This strategy has the potential to assess response more accurately or quickly than radiologic tests, or to confirm the success of surgical resection and select patients for adjuvant therapy. Methylated genes may be more specific than total DNA for the presence of cancer. Methods: We are conducting two prospective trials measuring total DNA, and 10 commonly hypermethylated genes in the plasma of NSCLC patients before and after therapy. For patients undergoing surgery, plasma samples are collected before, and at three timepoints after surgery (3–8 d, 2–5 w, 2–4 m). For patients on chemotherapy, collections are before, and coincident with up to 3 radiologic response assessments. DNA is extracted from a fixed volume of plasma, and measured using quantitative PCR of the β-Actin gene. After treating DNA with sodium metabisulfite, methylation-specific RT-PCR is performed to amplify the promoter region of APC, p16, MGMT, GSTP1, DAPK, CDH1, RASSF1A, WIF-1, SOCS-3, METH-2, assuming the promoter is fully methylated. Gene levels are normalized to β-Actin in modified DNA. Results: To date, plasma analysis has been completed on 50 patients with advanced NSCLC undergoing chemotherapy (320 planned), and 60 patients with early-stage NSCLC undergoing surgery (400 planned). Pretreatment levels of total DNA are higher in metastatic NSCLC (median 9.2, range 2.2–1020 ng/ml) compared to early-stage (median 3.6, range 0–1280 ng/ml), p < 0.001 by Wilcoxon rank sum test. At least one methylated gene is detectable in pretreatment plasma in 36% of patients with metastatic NSCLC, and 20% of patients with early-stage NSCLC. Conclusions: Methylated genes are rarely found in the pre-treatment plasma of NSCLC patients in this experiment, despite testing a panel of 10 genes. Plasma total DNA levels correlate with NSCLC stage. Correlation of baseline levels of total DNA, and temporal trends during treatment, to overall survival and radiologic response are planned. Supported by R01-CA092315. [Table: see text]

BACKGROUND: The benefit of cytoreductive surgery for patients with recurrent epithelial ovarian cancer has not been defined clearly. The objective of this study was to identify prognostic factors for survival in patients who underwent secondary cytoreduction for recurrent, platinum-sensitive epithelial ovarian cancer and to establish generally applicable guidelines and selection criteria.METHODS: The authors reviewed all patients who underwent secondary cytoreduction for recurrent epithelial ovarian cancer from 1987 to 2001. Potential prognostic factors were evaluated in univariate and multivariate analyses.RESULTS: in total, 157 patients underwent secondary cytoreduction, and 153 of those patients were evaluable. After secondary cytoreduction, the median follow-up was 36.9 months (range, 0.2-125.6 months), and the median survival was 41.7 months (95% confidence interval, 36.0-47.2 months). For patients who had a disease-free interval prior to recurrence of between 6 months and 12 months, the median survival was 30 months compared with 39 months for patients who had a disease-free interval between 13 months and 30 months and 51 months for patients who had a disease-free interval > 30 months (P = .005). For patients who had a single site of recurrence, the median survival was 60 months compared with 42 months for patients who had multiple sites of recurrence and 28 months for patients who had carcinomatosis (P < .001). The median survival for patients who had residual disease that measured <= 0.5 cm was 56 months compared with 27 months for patients who had residual disease that measured > 0.5 cm (P < .001). On multivariate analysis, disease-free interval (P = .004), the number of recurrence sites (P = .01), and residual disease (P < .001) were significant prognostic factors.CONCLUSIONS: In the authors' analysis of secondary cytoreduction for recurrent epithelial ovarian cancer, a significant survival benefit was demonstrated for residual disease that measured <= 0.5 cm. The disease-free interval and the number of recurrence sites should be used as selection criteria for offering secondary cytoreduction.