Frail Pts Research Articles

Importance and Determinants of Comorbidities, Functional Limitations and Multiple Myeloma (MM)-Specific Risk Factors: Further Development of an Improved and Weighted MM-Risk Score (Freiburg Comorbidity Index [FCI

Introduction: In addition to disease-specific and age-related factors, type and severity of comorbidities play a relevant role, influence the tolerance of anti-MM-treatment and overall survival (OS). We have identified an impaired Karnofsky Performance Status (KPS), lung and renal impairment as significant risk factors for inferior outcome (Kleber,…Engelhardt. BCJ 2011, Kleber,…Engelhardt. CLML 2013, Engelhardt et al. Haematologica 2014). These variables were combined in a comorbidity score (initial Freiburg Comorbidity Index [iFCI]). The objectives of this analysis were to refine the iFCI ('revised FCI’ [rFCI]) by adding host- and disease-specific risk factors, cytogenetics, physical function and quality of life. Moreover, we assessed the benefit of a possible weighting of the rFCI, and compared the rFCI to well-established comorbidity indices, namely Charlson Comorbidity Index (CCI), Hematopoietic cell transplantation-specific comorbidity index (HCT-CI) and Kaplan-Feinstein (KF).Methods: We assessed 803 consecutive patients (pts) treated at our institution between 1997-2012, determining comorbidities as weighted renal, lung, heart, liver, gastrointestinal diseases, KPS, disability, frailty, infection, pain, secondary malignancies, peripheral neuropathy, thrombosis and disease parameters (e.g. cytogenetics). We divided our cohort into a training (n=553) and validation set (n=250) and performed a multivariate analysis via backward selection. Regression coefficients were used to derive weights for the score. Apart from scoring both iFCI and rFCI, we also assessed the CCI, HCT-CI and KF.Results: Our pts showed a typical median age for a tertiary referral center of 63 years (range: 21-93). 26% revealed less favorable cytogenetics, defined as del(17p13), del(13q14), t(4;14), t(14;16) and chromosome 1 abnormalities. Each half of the pts had received either standard chemotherapy or stem cell transplantation. Frequent comorbidities (>30%) were KPS, heart, renal, lung impairment, disability and frailty. Univariate analysis revealed age, renal, lung and heart disease, KPS, disability, frailty, pain and infections as significant. Multivariate risks proved to be advanced age (>70 years), renal, lung, KPS impairment, frailty and cytogenetics with hazard ratios (HR) of 2.2, 1.8, 1.3, 3.2, 1.9 and 1.5, respectively. The rFCI allowed to distinguish low-, intermediate- and high-risk pts with largely different median OS of 11.2, 4.8 and 2.6 years, conclusively confirmed via validation analysis with distinct median OS differences of not reached, 6.5 or 1.4 years, respectively. Weighting of the single risk factors led to a score with maximum points of 39. In order to simplify this score, the single weights were divided by 2 and rounded, which led to a 20-point score (rFCI modified I [mod I]). A 2nd modification led to a 9-point score (rFCI mod II) which was obtained with single risk factors being scored based on their HR. These modified rFCI scores I and II allowed equally well to allocate MM pts in low-, intermediate- and high-risk groups as with the 39-point-rFCI, besides being simpler in their application. Compared to the CCI, HCT-CI and KF, the rFCI remained highly significant. For further comparison, all comorbidity indices in the training and the validation set were divided into two risk groups according to the cut-offs obtained from our initial analyses (BCJ 2011, CLML 2013). Regardless of whether we scored MM pts with the iFCI, rFCI, CCI, HCT-CI or KF, ‘low-risk’ pts had longer median survival than ‘high-risk’ pts. The comparison via median comorbidity indices showed superiority of the rFCI and CCI in the training set and of the rFCI and HCT-CI in the validation set. A further univariate analysis and comparison by dividing the different scores in risk groups based on 25% and 75% quantiles, revealed the highest HR for the rFCI both in the training and the validation set.Conclusions: As comorbidities in MM are frequent, a detailed comorbidity assessment, including host- and disease-specific risk factors, allows an improved risk evaluation in often frail pts. Age, renal, lung, KPS impairment, frailty and cytogenetics are relevant risk factors included in our rFCI. Advantages of the rFCI vs. iFCI are the inclusion of MM-specific risks including cytogenetics, the more accurate assessment of pts’ physical conditions, lower prediction errors and its simple clinical applicability. DisclosuresNo relevant conflicts of interest to declare.

A prospective study to evaluate the Vulnerable Elders Survey-13 (VES-13) as a tool to identify frail older cancer patients (pts).

9546 Background: A major challenge in the care of older cancer pts is identification of who might benefit from adjuvant chemotherapy (CT). Frail pts (FP) typically tolerate CT poorly and/or die fro...

Phase II study of first-line combined chemotherapy bevacizumab with modified RPMI regimen for elderly or frail patients with unresectable or metastatic colorectal cancer (OGSG0802) update analysis.

e14601 Background: The first-line combined chemotherapy RPMI regimen with bevacizumab (Bmab) in AVF2192g tri al were active. According to the results of efficacy and safety, a fluoropyrimidine (FU) + Bmab regimen is regarded as one of treatment options for 1st-line chemotherapy in many guidelines. We planned a phase II study of modified RPMI regimen with Bmab especially for elderly or frail Pts. Methods: Pts with confirmed unresectable/metastatic colorectal cancer without previous chemotherapy, and not suitable for intensive chemotherapy were enrolled. Pts received modified RPMI regimen (5-FU 600 mg/m2 and l-leucovorin 200 mg/m2 bolus day 1, 8, 15) and Bmab 5 mg/kg day 1, 15, q4w) until disease progression or study withdrawal. The primary endpoint was overall response rate (ORR), and the secondary endpoints were PFS, OS and safety. Results: 41 Pts were enrolled from 13 institutions. Pts characteristics were as follows; median age 76 (range 56-90); male/female, 18/23; ECOG performance status 0/1/2, 21/19/1. The ORR, the rate of best response, the disease control rate (CR+PR+SD) were 36.6%, 56.1%, 85.4%, respectively. Median follow-up period was 14.4 months(Updated results will be presented). 28 Pts (68%) had objective progression and a patient (2.4%) died without progression. The median PFS and OS were 9.0 months (95%CI, 7.5–19.6) and 24.0 months (95%CI, 20.1–NR). The incidences of grade 3 or 4 adverse events were: leukopenia (7%), neutropenia (24%), thrombocytopenia (2%), diarrhea (5%), anorexia (10%), fatigue (5%), stomatitis (7%) and hypertension (5%). Grade 3 febrile neutropenia and grade 4 pulmonary embolism was observed in one pt. Five pts (12%) discontinued the treatment due to severe or uncontrollable adverse event. Conclusions: The modified RPMI regimen with Bmab showed promising activity, and was well tolerated for elderly or frail Pts. ORR and the median PFS of this regimen were similar to historical data with FU + Bmab. This regimen may be a good option for patients with poor PS or elderly with advantages of not requiring percutaneous port placement nor compliance or oral agents. Clinical trial information: UMIN000002182.

Phase II study of first-line combined chemotherapy bevacizumab with modified RPMI regimen for elderly or frail patients with unresectable or metastatic colorectal cancer (OGSG0802).

573 Background: The first-line combined chemotherapy RPMI regimen with bevacizumab (Bmab) in AVF2192g trial for elderly or frail patients (Pts) could be active to progression free survival (PFS), the secondary endpoint, although it may not prolong overall survival (OS), the primary endpoint. According to the results of efficacy and safety studies, a fluoropyrimidine (FU) + Bmab regimen is regarded as an option for 1st-line chemotherapy. We planned a phase II study of modified RPMI regimen with Bmab for elderly or frail Pts. Methods: Pts with confirmed unresectable/metastatic colorectal cancer without previous chemotherapy, and not suitable for intensive chemotherapy were enrolled. Pts received modified RPMI regimen (5-FU 600 mg/m2 and l-leucovorin 200 mg/m2 bolus day 1, 8, 15) and Bmab 5 mg/kg day 1, 15, q4w) until disease progression or study withdrawal. The primary endpoint was overall response rate (ORR), and the secondary endpoints were PFS, OS and safety. Results: 41 Pts were enrolled from 13 institutions. Pts characteristics were; median age 76 (range 56-90); male/female, 18/23; ECOG performance status 0/1/2, 21/19/1. The ORR, the rate of best response, the disease control rate (CR+PR+SD) were 36.6%, 56.1%, 85.4%, respectively. Median follow-up period was 14.4 months. 28 Pts (68%) had objective progression and a patient (2.4%) died without progression. The median PFS and OS were 9.0 months (95%CI, 7.5–19.6) and 24.0 months (95%CI, 20.1–NR). The incidences of grade 3 or 4 adverse events were: leukopenia (7%), neutropenia (24%), thrombocytopenia (2%), diarrhea (5%), anorexia (10%), fatigue (5%), stomatitis (7%) and hypertension (5%). Grade 3 febrile neutropenia and grade 4 pulmonary embolism was observed in one pt. Five pts (12%) discontinued the treatment due to severe or uncontrollable adverse event. Conclusions: The modified RPMI regimen with Bmab showed activity, and was well tolerated for elderly or frail Pts. ORR and the median PFS of this regimen were similar to historical data with FU + Bmab. This regimen may be a good option for not requiring percutaneous port placement Pts. Clinical trial information: UMIN000002182.

Treatment of Diffuse Large B-Cell Lymphoma (DLBCL) Patients (pts) Age 80 and Older: Analysis of the Veterans Health Administration (VHA) National Database

AbstractAbstract ▪968▪This icon denotes a clinically relevant abstract Introduction:The incidence of DLBCL increases with age and the population ≥80 is the fastest growing segment of the elderly population. Optimal treatment of these very elderly DLBCL pts remains unclear. While chemo-immunotherapy can be curative, benefits of treatment must be balanced against toxicities in older, frail pts. Previous studies have suggested that doxorubicin treatment is associated with better survival in the very elderly population, though only on univariate analyses. To better understand the relationship between treatment and survival in pts ≥80, we performed a retrospective analysis of veterans to examine the relationship between treatment and overall survival (OS). Methods:We identified 523 DLBCL pts ≥80 diagnosed between 1998 and 2008 in the VHA Central Cancer Registry. We excluded pts with primary cutaneous or central nervous system DLBCL and pts treated with unknown agents outside the VHA. Data on stage, LDH, co-morbidities, date of death, treatment and supportive care medications were obtained for each pt. Treatment related mortality (TRM) was defined as death in the 30 days following any cycle of therapy. OS was defined as time from diagnosis to death. ECOG Performance Status (PS) at diagnosis was either obtained from pt charts or assessed from physician, nursing, and physical therapy notes available for each patient. Results:Of the 476 pts that met inclusion criteria, 193 received at least one dose of doxorubicin therapy (DT), 92 were treated without doxorubicin (non-DT) and 191 received no system therapy (NST). Median OS times were 30.2 months, 12.3 months, and 1.9 months in the DT, non-DT, and NST groups, respectively (Log-rank p<0.001). Baseline characteristics of 285 treated patients presented in Table 1. Among the 273 pts in whom treatment dates were available, 48 (18%) experienced TRM. Of these 48 deaths, 32 (67%) were associated with the first treatment cycle. On multivariate analysis, rituximab significantly reduced the risk of death (HR .62, CI .44 - .87), while GCSF use demonstrated a trend towards decreased death (HR .76, CI .57 – 1.03). Comorbidities (HR 1.1, CI 1.02 – 1.19) and PS were associated with increased risk of death. PS of 2–4 was associated with a marked increased risk of death within the first 45 days after diagnosis (HR 19.7, CI 2.6– 146) and a lesser risk after 45 days (HR 1.9, CI 1.4 –2.6). After controlling for other variables, doxorubicin use was no longer significantly associated with OS (HR .87, CI .64 – 1.17), suggesting the OS difference noted between the DT and non-DT groups was primarily due to differences in baseline patient characteristics. Conclusions:While we acknowledge the limitations of analyses of observational data, the observed 18% rate of TRM (compared to historical TRM rates of 4–6% seen in trials of R-CHOP) suggests that standard treatment paradigms are too toxic for many patients in this age group. The lack of association between doxorubicin and OS calls into question the importance of this drug in this patient population. In the absence of better risk stratification, treatment with less toxic regimens- either through attenuated doses or different drugs- may be appropriate in this population.Table 1:Demographic and clinical characteristics of US veterans with DLBCL aged 80 and older at time of diagnosis, stratified by treatment group (N=285)Demographic and Clinical CharacteristicsDoxorubicin n=193Non-Doxorubicin n=193p-valueAge at DX (mean years)83.183.60.086Stage (%)0.153Stage I28.522.8Stage II15.515.2Stage III23.819.6Stage IV32.142.4LDH (%)0.978Elevated49.750.0Not Elevated42.542.4Unknown7.87.6Performance status (%)0.0210–152.338.02–446.160.9Unknown1.61.1Age adjusted IPI (%)0.103015.57.6126.430.4231.627.2318.127.2Unknown8.37.6Co-morbidity (Charlson score) (Mean)1.82.30.028B-symptom( %)0.042Yes42.055.4No55.443.5Unknown2.61.1CSF use (%)78.847.2<.001Rituximab use (%)80.876.10.356TX DrugCHOP/R-CHOP190-CNOP/R-CNOP-5CVP/R-CVP-53R-EPOCH1-Rituximab-21Other213 Disclosures:Carson:Genentech: Consultancy, Honoraria, Speakers Bureau. Nabhan:Genentech: Research Funding, Speakers Bureau.

573P - Phase II Study of 1st-Line Combined Chemotherapy Bevacizumab with Modified Rpmi Regimen for Elderly or Frail Patients with Unresectable or Metastatic Colorectal Cancer (OGSG0802)

ABSTRACT Background/ Objectives The first-line combined chemotherapy RPMI regimen with bevacizumab (Bmab) in AVF2192g trial for elderly or frail patients (Pts) could be active to progression free survival (PFS), the secondary endpoint, although it may not prolong overall survival (OS), the primary endpoint. According to the results of efficacy and safety studies, a fluoropyrimidine (FU) + Bmab regimen is regarded as an option for 1st-line chemotherapy. We planned a phase II study of modified RPMI regimen with Bmab for elderly or frail Pts. Methods Pts with confirmed unresectable/metastatic colorectal cancer without previous chemotherapy, and fulfilling at least one of the following characteristics were enrolled: Age ≧65 years, ECOG performance status 1 or 2, serum albumin ≦ 3.5g/dl, incompatible with receiving oxaliplatin or irinotecan, or prior abdominal/pelvic radiotherapy. Pts received modified RPMI regimen (5-FU 600 mg/m2 and l-leucovorin 200 mg/m2 bolus day 1, 8, 15) and Bmab 5 mg/kg day 1, 15, q4w) until disease progression or study withdrawal. The primary endpoint was overall response rate (ORR), and the secondary endpoints were PFS, OS, safety and treatment completion rate of treatment. Results 41 Pts (mean age 76 years [range 56–90]; 55% male) were enrolled to this trial from 13 institutions. 28 Pts (68%) had objective progression and a patient (2.4%) died without progression. The ORR, the primary endpoint, the rate of best response, the disease control rate (CR + PR + SD) were 36.6%, 56.1%, 85.4%, respectively. The median PFS was 9.0 months (95%CI, 7.5–19.6) by independent central review. Median OS was 30.2 months (95%CI, 23.8–Not achieved). Mean 5-FU and Bmab dose intensity were 86.9% and 83.6%, respectively. The incidences of all grade/grade 3-4 adverse events: leukopenia (66/7), neutropenia (58/24), thrombocytopenia (39/2), nausea (29/0), diarrhoea (34/5), anorexia (32/10), fatigue (49/5), stomatitis (29/7) and hypertension (24/5). Grade 3 febrile neutropenia and grade 4 pulmonary embolism was observed in one patient. Conclusions The modified RPMI regimen with Bmab showed activity, and was well tolerated by elderly or frail Pts. ORR and the median PFS of this regimen were similar to historical data with FU + Bmab. This regimen may be a good option not requiring percutaneous port placement for elderly or frail Pts. . Disclosure M. Yoshida: The author received few payments for lectures from Chugai Pharmaceutical Co., Ltd. Grants for research were also provided to the Chemotherapy Center of Osaka Medical College by this pharmaceutical companies. All other authors have declared no conflicts of interest.

1037P - Docetaxel, Carboplatin and Fluoro-Uracil (TCF) Induction Therapy in Locally Advanced Head and Neck Squamous Cell Carcinoma (HNSCC) Patients with Contraindication for Cisplatin Based Combination (TPF)

ABSTRACT Background TPF regimen is a standard induction combination for fit locally advanced HNSCC patients (pts). TCF with carboplatin may be an option in frail pts with contraindication for cisplatin, but data are missing about safety and efficacy of such an approach. Method Retrospective study of monocentric cohort of HNSCC patients treated from 2007 to 2011 with TCF induction regimen at Gustave Roussy Institute. TCF regimen combined Docetaxel 75mg/m2, Carboplatin AUC4, and 5FU: 750 mg/m2/day for 5 days. Results Population included 34 pts, median age 59 years (44-78), all with Charlson Comorbidity Index score ≥ 2, with oropharynx, hypopharynx, oral cavity, larynx and paranasal sinuses HNSCC localization (32.3, 23.5, 17.6, 14.7 and 11.8% respectively). TNM stage were: T1-2 : 4 pts (11.7%), T3-4: 27 pts (79.4%), rTx: 3 pts (8.8%); N0: 8 pts (23.5%), N1: 4 pts (11.8%), N2a-N2b: 6 pts (17.6%), N2c-N3: 13 pts (38.2%); M1: 1 pt (2.9%). TPF contraindications were respiratory disease (di.) (11 pts: 32.3%), heart di. (6 pts: 17.6%), liver di. (3 pts: 8.8%), alcohol related neuropathy (3 pts: 8.8%), hearing di. (5 pts: 14.7%), renal di. (5 pts: 14.7%), other co-morbidities (3 pts: 8.8%).Patients received 2 to 4 cycles of TCF. GCSF was delivered in 31 pts (91.2%). Cumulated incidence of Grade 3-4 toxicity was 37.2%, and one toxic death occurred at C2. Five pts (14.7%) stopped treatment because of toxicity. RECIST evaluation found complete response in 1 patient (2.9%), partial response in 23 pts (67.6%), stable disease in 6 pts (17.7%), and progression disease in 1 pt (2.9%). Subsequent treatment was surgery followed by radiotherapy (9 pts, 26.5%), radiotherapy or chemoradiotherapy (22 pts, 64.7%), chemotherapy (3 pts, 8.8%). Response after RT was: CR: 16 pts (64%), PR: 4 pts (16 %), SD: 3 pts (12 %), PD: 2 pts (8 %). Median recurrence-free survival was 8.7 months, median overall survival (OS): 22.9 months. Personal history of cancer (HR = 4.6, p = 0.02), tobacco (HR = 37.5, p = 0.04), alcohol intake (HR = 0.12, p = 0.01), hypertension ( HR = 4.4, p = 0.01), T4 stage ( HR = 0.1, p = 0.02), N2c-N3 stage (HR = 18.5, p = 0.003), BMI Conclusion TCF induction regimen was manageable and provided objective responses in 70% of frail HNSCC patients. Disclosure All authors have declared no conflicts of interest.

Phase II trial of sorafenib and capecitabine for hepatocellular carcinoma in cirrhotic patients.

341 Background: Sorafenib (sorf.) has become the standard treatment for Hepatocellular Carcinoma (HCC), demonstrating an increase in patients (pts) median survival from 7.9 to 10.7 months (m) (Llovett, NEJM 2008). Abou-Alfa added doxorubicin to sorf. observing a median survival of 13.7 m vs 6.7 m. for doxorubicin + placebo, (JAMA. 2010). Since most of our HCC pts were cirrhotic and 8/12 HCV +, and 2/12 HBV only +, with cirrhosis and thrombopenia we chose to add capecitabine (cap) to sorf in a phase II trial, anticipating better tolerance. Primary objectives were to assess safety and tolerability of the combination while secondary objectives included estimations of PFS, OST and documentation of disease control rate (DCR = CR+PR+SD). With relaxation of eligibility criteria C-P B-7 pts and ≥ 50,000 platelets were allowed. Methods: Sorf dose was 400 mg BID with dose adjustments for sorf related hand foot syndrome (HFS). Starting cap dose was 850mg/m2 BID x 7 Q 14 and adjusted as tolerated to 650mg/m2 BID or to 500 mg/m2 BID x 7 Q 14.. The lowest dose used was 250 mg/m2; BID (500 mg BID) for 5 days Q14. Patients: Only 12 of about 30 assessed patients met eligibility criteria, and were recruited between Nov. 2, 2009 and Sep. 13, 2011.The protocol was reviewed and approved by the WIRB and all participants signed an IRB approved consent form. Results: Preliminary Results: Eight of 12 patients are alive, survivals ranging from 2.5 to 23 m. One patient has an unmaintained radiological CR continuing for 23 months with an AFP decrease from 139 to 2.9. Another patient had a PR and an AFP decrease from 355 to 3.2 ng/mL. She died in PR at 13 months from an UGI bleed. Grade 3 and 4 toxicities were limited to the skin, and included HFS, in 4/12, skin ulcers at sites of prior osteomyelitis (1/12); requiring treatment cessation, and fatigue in 4 /12. Conclusions: With meticulous attention to cap dose adjustments as dictated by HFS this combination can be tolerated longitudinally in these frail pts that are not candidates for adding more aggressive agents to sorf. The objective radiological responses corroborated by AFP decreases in this very small cohort suggest that the findings are real, and warrant expansion of the trial.

A Study of Outcomes of Diffuse Large B Cell Lymphoma in the Elderly

Abstract 4950 Introduction:Despite improvements in outcome using chemoimmunotherapy, treating elderly patients with diffuse large B cell lymphoma (DLBCL) remains challenging. Reliable risk stratification beyond age is lacking and consequently elderly patients are often treated with reduced-dose therapeutic regimen with palliative intent. We retrospectively analyzed outcomes of DLBCL patients aged 65 and older treated at the University of Pennsylvania. Patients and method:We identified 41 patients (pts) with diagnosis of DLBCL and age >65 years. Median age was 74 years (range: 65–86). Eight pts (20%) were age 80 or older. There were no differences in IPI, elevated LDH, or bcl-2 expression, whereas bcl-6 expression was more common in patients >80 years with 75% (6/8) vs. 27% (24/33) (p=0.01).Overall, 31 pts were treated with R-CHOP, 2 pts with R-HyperCVAD and one pt with R-CVP. Seven patients with a median age of 77 years were considered too frail for standard chemotherapy and were treated with a “split R-CHOP” regimen consisting of: rituximab 375 mg/m2 day 1, cyclophosphamide 375 mg/m2 day 1 and 15, adriamycin 25 mg/m2 day 1 and 15, vincristine 1 mg day 1 and 15 and Prednisone 50 mg day 1–5 and day 15–19. Results:Overall, the complete remission (CR) rate was 56% with seven treatment related deaths (17%). For patients between 65 to 80 years of age and deemed fit for standard chemotherapy (n=28), CR rate was 57% with four treatment related deaths (14%). Among patients over 80 and deemed fit for standard chemotherapy (n=6), the CR rate was 50% with 2 treatment related deaths (33%). Frail patients treated with the “split R-CHOP regimen” (n=7) had a CR rate of 57% with one (14%) treatment related death.For all patients, the median progression-free survival (PFS) was 1 year with a median overall survival (OS) of 2 years. The median PFS for pts between 65 and 80 years of age treated with standard chemotherapy was 16 months. Median PFS in pts >80 years of age treated with standard chemotherapy was 7 months, whereas median PFS in frail pts treated with “split R-CHOP regimen” was 11.7 months. Conclusions:Our data reveal interesting findings about elderly pts treated for DLBCL. PFS and OS in general are poor as has been reported by others. In our data set, pts>80 years considered fit for standard chemotherapy, had a shorter PFS then fit patients between 65 to 80 years.Intriguingly, PFS among frail elderly patients treated with a “split R-CHOP” regimen appears to be superior to that of elderly deemed more robust who were treated with standard dose RCHOP-21. Though limited by small numbers, our institutional data suggest that frail patients can tolerate a modified R-CHOP regimen although survival remains short. [Display omitted] Disclosures:No relevant conflicts of interest to declare.

Safety of adjuvant trastuzumab (T) in elderly patients with breast cancer.

282 Background: T is a mainstay in adjuvant therapy for HER2+ breast cancer (BC) patients (pts). Safety and efficacy of T in elderly patients are largely unknown. In HERA trial, NSABP B-31, NCCTG N9831 only 16% of pts were older than 60 years. Risk factors for T related cardiotoxicity are age (&gt;50 y/o), hypertension, baseline LVEF (left ventricular ejection fraction &lt;55%), previous antracycline therapy and BMI. Methods: Charts of pts &gt;65 y/o with early HER2+ BC treated with T as adjuvant or neoadjuvant therapy at our institution were retrospectively reviewed. Primary endpoint was the evaluation of T cardiac toxicity and safety. Results: 22 elderly out of 172 pts (12%) were identified: 19 pts were treated only with surgery and adjuvant chemotherapy with concomitant or sequential T, 3 more pts also received neoadjuvant chemotherapy concomitant with T. According to Balducci’s criteria, fit, vulnerable and frail pts were 20, 2, 0 respectively. Median age was 69 y/o (range 65-76). Hormonal status was negative in 10/22 (45%). 21/22 were histologic grade 3. Median follow-up was 33 months. Baseline comorbidities were the following: hypertension (G2-3) in 17 pts, diabetes mellitus in 1, supra/infraventricular arrhythmia (G1-2) in 3 and 1 pts. Antracyclines were administered in 16 pts (liposomal-doxorubicin in 5 pts), a sequential taxane-regimen was used in 3 more pts. Neoadjuvant weekly Paclitaxel and concomitant T was used in 3 pts. Median basal LVEF was 65% (range 59-74%). 2 pts developed an asymptomatic 10% LVEF drop from baseline (left ventricular systolic dysfunction G1) during T treatment. Known cardiac risk factors were hypertension in 1 pt and previous antracycline based chemotherapy in both. They recovered within 9 months. One minor adverse event was atrial fibrillation (G2) during T treatment. Conclusions: T was well tolerated in elderly pts. More data are needed in order to understand the correlations between T related toxicity and cardiovascular risk factors. Long term safety of T treatment should verify the reversibility of cardiac T related toxicity on elderly pts.

Prognostic value of multidimensional geriatric assessment (MGA) on survival of a prospective cohort of 880 elderly cancer patients (ECP).

9065 Background: Estimated tumor-independent life expectancy is a key element for treatment choices in ECP. MGA allows to subdivide them into three categories of Fit, Vulnerable and Frail pts. Whether MGA is able to predict survival independently from tumor type, stage, and treatment is still an open question. Methods: All consecutive pts ≥70 years admitted to our Geriatric Oncology Program underwent MGA (Basso CROH 2009) and were prospectively followed. All-cause survival was compared according to their condition of fitness, vulnerability or frailty. Type of tumor, disease status at first visit (neoadjuvant/adjuvant vs advanced), administration of any treatment (cytotoxic/endocrine/targeted therapy) were assessed. Kaplan-Meier survival method was used with multivariate (Cox) analysis. Results: From 09-2003 to 10-2010, 880 ECP were enrolled, median 77 yrs (range 70-96), 28% males, with breast (54.5%), colorectal (14.1%), lung (8.2%) or other sites (23.2%) of cancer. Seventy-one percent were seen for neoadjuvant/adjuvant treatment; 79.8% were prescribed some oncological treatment. After a median follow-up of 26 mo (1-77), 334 have died (38%). Status at MGA predicted overall survival (p<0.001). Significance was maintained in the subgroups with adjuvant/neoadjuvant vs advanced disease, treated vs untreated pts, and in pts with breast cancer or other sites, but was lost when only colon or lung cancer were considered (Table). At multivariate analysis, Fitness maintained a strong impact on survival compared to Frailty (HR=0.41, CI:0.31–0.54) as well as Vulnerability vs Frailty (HR= 0.59, CI: 0.45 – 0.77). Conclusions: MGA predicts significantly different survival probability in our large cohort of ECP, independently of stage and treatment. Lack of significance in lung or colorectal cancer may be due to lower patient numbers. Survival (%) at 24 mo (univariate log rank test). Fit 338 pts Vulnerable 288 pts Frail 254 pts P value Overall 83 70 60 <0.001 Disease status Adjuvant/neoadjuvant 93 86 79 <0.001 Advanced 52 33 37 0.009 Disease site Breast 94 90 78 <0.001 Colorectal 75 56 51 0.185 Lung 33 17 17 0.241 Other 68 45 41 0.001 Oncological treatment Yes 83 72 65 <0.001 None 87 74 51 <0.001

Biweekly liposomal pegylated doxorubicin in elderly women with advanced breast cancer: A prospective multicenter trial focusing on tolerability and cardiotoxicity.

1138 Background: To assess tolerability and activity of liposomal pegylated doxorubicin (PLD) in women older than 70 years with locally advanced or metastatic breast cancer within a multicenter prospective trial. Methods: All pts underwent Multidimensional Geriatric Assessment (MGA) and frail pts (dependence in Activities of Daily Living; three or more severe comorbidities, geriatric syndromes) were excluded. Normal cardiac function was required at baseline. Bi-weekly schedule of 20 mg/m2 was adopted. More than 12 doses were prescribed only to responsive pts. Results: From Feb 2006 to Sep 2009, 28 pts were enrolled, median age 77 yrs (range 70-93), 10.7% had locally advanced disease, 10.7% only bone metastases, 78.6% visceral involvement; 28.6% of pts had been pre-treated with one or more lines of chemotherapy for advanced disease. After MGA, 35.7% were deemed fit and 64.3% vulnerable. A mean of 7.2 doses were delivered (range 1 to 20), with an actual dose intensity of 8.9 mg/m2/week (range 5.2-10). Worst toxicity experienced by pts is outlined in table. No cardiac events were registered. Only 1 pt received both G-CSF and erythropoietin. Causes of treatment interruption were progression 46.4%, refusal/loss at follow-up 21.4%, toxicities 10.7%, other or ongoing 10.7%. RECIST response was obtained in 34.8% of 23 evaluable pts, 30.4% remained stable; with a median TTP of 4.4 mo (11 censored), and a median OS of 15.4 mo. Receiving full dose intensity vs <9 mg/m2/week did not impact on response (p=0.26) and progression (p=0.6), as well as MGA status (vulnerable vs FIT, p=0.65 for response, p=0.9 for progression). Conclusions: Biweekly PLD monochemotherapy is well tolerated in terms of hematological and cardiac safety both by fit and vulnerable pts, with an apparently fairly good response rate. Close follow-up of pts is recommended in order to monitor skin toxicity and avoid early refusal/loss at follow-up. Accrual is still ongoing. Toxicity according to CTCAE v3 Grade 1-2 Grade 3-4 Neutropenia 23% 0 Thrombocytopenia 5% 0 Anemia 18% 9% Palmar-plantar erythrodysesthesia 41% 9% Nausea/vomiting 32% 0 Mucositis 18% 9% Infection 0 5% No significant financial relationships to disclose.

A multicenter phase II randomized study of gemcitabine (G) weekly followed by erlotinib (E) after progression versus E followed by G after progression in advanced non-small cell lung cancer (NSCLC) in vulnerable elderly patients selected with a comprehensive geriatric assessment (CGA) (GFPC*0505).

e18002 Background: Elderly patients (pts) are a heterogeneous population in which anticancer therapeutic decision is often difficult. The objective of this study was to evaluate the feasibility and activity of weekly G followed by E after progression versus E following by G after progression in vulnerable elderly advanced NSCLC patients selected on the basis of a CGA and Charlson score (CS). Methods: This multicenter phase II study randomized chemonaive pts with stage IV and IIIB with pleural effusion received G 1,250mg/m2 days 1, 8, for a maximum of 18 weeks or until progression, followed by E until second progression (arm A) versus E in first line followed by G until second progression (arm B).Only vulnerable pts (without dependence for ADL and geriatric syndrome) selected by CGA using a software (EGSK) included in a palm assistant with 10 items: socio-economic conditions, cognitive assessment, depression scale, nutritional risk, quality of life(QOL), ADL, IADL (Lawton), incontinence, falls, pain. Major endpoint was time to second progression (TTP2), secondary endpoints were response, time to first progression (TTP1), safety, QOL. Results: 95 pts (67 available today) were enrolled from Aug 2006 to Dec 2009. Patients: Male/Female 54/13, median age 78.1; PS 0/1/2:22/33/10, median co-morbidity (CS) 1.85 (1-3), median IADL score 3.48 (1-4), median global geriatric index: 17.7/20. Respectively for arm A and B, partial response was 18% and 7%, stable disease 18% and 27%, progressive disease 64% and 67%,TTP1 81 and 63 days, median survival 4.3 and 3.9 months. The main WHO grade 3/4 toxicity was anemia 12.5% and fatigue 12.5% in arm A and fatigue 28.6% and anemia 5.7% in B. Conclusions: Preliminary analysis suggests that either the G followed by E or E followed by G appears to have acceptable toxicity but moderate activity, comparable in terms of efficacy to prior data in this vulnerable elderly NSCLC pts population with significant co-morbidities. Our trial confirms the necessity to identify the better tools in CGA to distinguish frail pts. Author Disclosure Employment or Leadership Position Consultant or Advisory Role Stock Ownership Honoraria Research Funding Expert Testimony Other Remuneration Amgen, Boehringer Ingelheim, Lilly, Roche Amgen, Lilly, Roche Amgen, Lilly, Roche

Allogeneic Transplantation with a Treosulfan-Fludarabine Conditioning Regimen Is Feasible and Effective in Elderly Patients with Refractory or Relapsed Acute Myeloid Leukaemia (AML) and Myelodysplastic Syndrome (MDS); Preliminary Results From a Single Center Experience.

Abstract 4318 Backgroundthe majority of AML and poor risk MDS cases in the elderly (>60 years) are resistant to induction chemotherapy or relapse after initial response; thereafter, most of these patients (pts) receive only supportive treatments, with an expected survival of few weeks. Allogeneic (allo) stem cell transplantation (SCT) could be the only potentially salvage strategy; actually, for the high probability of letal toxicites due to the conditioning regimen, acute graft-versus-host disease (aGvHD) and infections, these pts are rarely offered this procedure.Since 2005, we have transplanted 12 pts, age>60, with refractory or relapsed AML/MDS, from an allogeneic donor. Data on feasibility and outcome are here reported. Aimto retrospectively evaluate data on alloSCT in elderly pts with refractory or relapsed AML and MDS, transplanted at our Institute. Patients and Methodsperiod 9/2005 to 4/2009, 12 pts, median age 65 (61-72). Performance status (Karnofsky): median 90% (80-100), HCT-CI: median score 1 (0-4). Diagnosis (WHO): AML 3, AMLMD 4, RAEB2 2, MDS/MPD 2, therapy-related AML 1. Median number of chemo cycles before SCT: 2 (1-7), all pts received at least 1 cycle with intermediate or high-dose cytarabine, 2 pts also received an autologous SCT. Donors: matched sibling 2, matched unrelated 2, related haploidentical 7, cord blood 1. Disease status at SCT: primary refractory 7, relapsed 5. Conditioning regimen: treosulfan (TREO) 14 g/sqm for 3 days, fludarabine (FLU) 30 mg/sqm for 5 days, ATG 10 mg/kg for 3 days for SCT from alternative donors. GvHD prophylaxis: T-cell depletion in 2 cases, cyclosporine (CSA)/methotrexate in 7, rapamycin/mycophenolate mofetil (MMF) in 3. Results11/12 (92%) pts engrafted and were in CR at day +30, with 95-100% donor chimerism (VNTR). Deaths within day 30 and day 100 were 0 and 3 (25%), respectively. Acute GvHD: 6 pts (55%), grade II (3) or III (3), only of skin in 3 cases, skin+liver in 1, only intestinal in 2. Relapses were 4 (36%). Overall TRM was 25% (3 pts), with all deaths due to infections during immunosuppression for aGVHD; 5 more pts died because of disease progression. Median OS from SCT of all 12 pts was 180 (56-1150) days, DFS (11 pts) was 128 (24-1114) days. At last follow-up (FU) 4 pts (33%) are alive, all in CR with a median FU of 406 (128-1150) days. ConclusionsalloSCT proved to be feasible in our elderly pts with AML/MDS refractory to induction or relapsed after initial complete remission. Early letal toxicities, due to the conditioning treatment, were absent; aGVHD, infections and relapses were the principal causes of mortality. Up to now, prolonged survival (>1 year) free from disease has been obtained in 25% of pts; of the 4 pts alive in CR, only 1 is receiving immunosuppression (CSA), for chronic GVHD. In conclusion, the TREO-FLU association showed a reduced-toxicity profile in these frail pts and a substantial anti-leukemic effect. Better prevention of aGVHD should be obtained, expecially after SCT from alternative donors, possibly with the rapamycin/MMF combination, which we are currently investigating at our Institute. AlloSCT after TREO-FLU conditioning can be considered an effective option for AML/MDS elderly pts with active disease after failure of previous intensive treatments. Disclosures:bonini:MolMed S.p.A.: Consultancy.

Reduced Intensity Conditioning (RIC) Regimen Followed By Allogeneic Hematopoietic Stem Cell Transplantation (HSCT) In Adult Patients (PTS) With Acute Lymphoblastic Leukemia (ALL)

Background: RIC regimens are less myelosuppressive, but still allow for successful engraftment with acceptable treatment-related mortality (TRM) in frail pts who otherwise would not be suitable candidates for HSCT. This is particularly relevant in ALL, since pts often sustain toxicity from dose-intense regimens, or may be diagnosed in advanced age. The antitumor effect of this approach is not well-established in ALL. Methods: We evaluated 30 ALL pts (19 M/11 F) treated from 1996 to 2008 with fludarabine 120 mg/m2, melphalan 140 mg/m2 and unmanipulated stem cells. GVHD prophylaxis consisted of tacrolimus and mini-methotrexate. Anti-thymocyte-globulin was added to matched unrelated pts. Results: Median age was 44 years (range 23–64). ECOG status at time of HSCT was 0 (n = 16), 1 (n = 10) or 2 (n = 4) with median co-morbidity score of 3 (range 0–7) by Charlson Comorbidity Index. 24 pts had B-lineage and 6 had T-lineage disease. Cytogenetic data were available for 26 pts; 19 had high-risk cytogenetics, including 9 with Ph+ disease. Disease stage at time of study entry was CR1 (n = 5), ≥CR1 (n = 12), or primary or refractory relapse (n = 13), with median 2 prior chemotherapy regimens (range 1–4); five pts had a prior allogeneic HSCT. Donor type was matched related (n = 13) or matched unrelated (n = 17) and stem cell source was bone marrow (n = 14) or peripheral blood (n = 16). The median total nucleated cell dose and CD34+ cell dose were 3.80 × 108 cells (range 0.68 – 17.16) and 4.15 × 106 cells (range 1.78 – 12.03), respectively. Median time to ANC 0.5 × 109/L was 13 days (range 10–24). Median time to platelet count 20 × 109/L was 18 days (range 10–57). Eight pts were alive at a median follow up of 12 months from HSCT (range 3–59). OS and DFS were 32% and 29%, respectively, at 1 year. Of note, only 1 among 5 pts in CR1 had disease progression, compared to 8 among 13 with refractory disease at time of HSCT. The cumulative incidence of acute GVHD, grades II-IV and III-IV were 40% and 13%, respectively, and chronic GVHD was 22% (7% for extensive). The cumulative incidence of TRM at 100 days and 1 year were 17% and 33%, respectively. There were 22 deaths: recurrence 14, infection 4, GVHD 4. Conclusion: RIC HSCT can provide disease control in patients with ALL, and merits further evaluation. Alternative treatment strategies need to be explored in pts with advanced disease. The observed TRM rate is comparable to previously reported rates in in heavily pre-treated leukemia patients.

A phase I study of pegylated liposomal doxorubicin + gemcitabine in a fixed dose-rate infusion for the treatment of patients with recurrent ovarian cancer

5579 Background: Pegylated liposomal doxorubicin (PLD) constitutes the reference treatment for platinum-resistant (PR) ovarian cancer (OC). The combination of gemcitabine (G) + PLD seems to have additive activity in OC; however, it also has additive haematological toxicity, precluding administration of optimal dose of both drugs. Fixed dose-rate infusion (FDRI) of G improves haematological tolerance. Methods: We present data of a phase I trial using the combination of these two drugs, including G in this FDRI to ascertain whether we can enhance the therapeutic index of this combination. Inclusion criteria were: recurrent OC with a platinum-free interval <1 year, primary or secondary treatment with platinum, age < 80 years, Karnosky >60 and with normal organ function. The starting dose of G was 1500 mg/m2, 10 mg/m2 minute, q 2 weeks (± 250 mg G titration depending on toxicity) followed by PLD 35 mg/m2 q 4 weeks. Pharmacokinetic analyses were performed on days 1 and 15 of the first cycle. Results: From December 2005 to July 2007, 20 patients (pts) were enrolled. 5 were non-evaluable for toxicity due to early progression. In the first step, 4 (pts) were treated and 2 developed dose-limiting toxicity (DLT) consisting on grade 4 neutropenia and grade 3 stomatitis. In the next step G1250/PLD35, 5 pts have been included, 2 developed late grade 3 stomatitis and dermatitis. Heavily-pretreated pts (>6 cycles) and/or >70 years (frail pts) were treated with G1000/PLD35; to date 1 of 6 has presented grade 3 stomatitis. PLD did not affect the pharmacokinetic of G or its metabolites. Conclusions: In conclusion, the recommended dose for a phase II study in recurrent OC is G1250/PLD35, although dose reduction may be necessary due to late toxicity (being stomatitis and dermatitis the most common DLT). For frail pts, the recommended dose is G1000/PLD35. The study is ongoing in order to have more solid long-term toxicity data. Activity appears to be promising: 30.7% (4/13) objective response in PR OC (platinum-free interval < 6 months) and 85.8% (6/7) in pts with a platinum-free interval between 6–12 months. No significant financial relationships to disclose.

Docetaxel-based chemotherapy in elderly patients (≥ 75 years) with castration resistant prostate cancer (CRPC): A French National study of 175 patients

5145 Background: Although docetaxel-based chemotherapy is a standard in CRPC, only limited data is available in elderly patients (pts) with CRPC treated with chemotherapy. Methods: Data from 175 pts older than 75 years with metastatic CRPC treated with first-line docetaxel in 9 French cancer centers from 2000 to 2007 were analyzed. Results: Median age was 78 years. 95 pts (54%) received a “standard” docetaxel regimen (SR) (70–75 mg/m² d1, q3w) and 80 pts (46%) an “adapted” regimen (AR) (30–35 mg/m² weekly). Pts treated with an AR were older (≥ 80 years: 41.2% vs 25.3%, p=0.04), had a poorer performance status (PS≥ 2: 28% vs 14%; p=0.04) and had pain more often (63% vs 42%, p=0.01) than pts treated with a SR. Grade 3–4 adverse events occurred in 46% and were correlated with a poor PS and the presence of visceral metastases, but not with the regimen (SR vs AR). More pts in the SR group had febrile neutropenia (9.5% vs 1%, p=0.02) whereas more pts in the AR group had severe fatigue (20% vs 6%, p=0.01). Early discontinuation of chemotherapy due to toxicity occurred more frequently in the AR group (30% vs 8%, p=0.0005). Four deaths (2.3%) were related to treatment. The PSA response rate was 71% and 68% in the SR and in the AR groups, respectively. Median progression-free survival was 7.4 months (95% confidence interval (CI): 6.5–8.4). Median overall survival was 15 months (95%CI: 12.8–17.1). In multivariate analysis, PS and presence of visceral metastases, but not age, predicted for overall survival. Conclusions: Pts over 75 years considered fit for the SR had a similar safety and efficacy profile than younger pts. In contrast, frail pts treated with an AR experienced a poor tolerance. The optimal schedule of docetaxel in frail pts remains to be elucidated. Geriatric tools should be used to detect more accurately elderly CRPC pts unfit for chemotherapy. No significant financial relationships to disclose.

An attempt to correlate “comprehensive geriatric assessment” (CGA), treatment assignment and clinical outcome in elderly cancer patients: Results of a phase II open study

19592 Background and Methods: The design was a prospective Phase II open study. Elderly (age =65 years) cancer patients (pts) of different sites were assigned to 3 different CGA categories: “fit”, “intermediate” and “frail”. Therefore, an appropriate treatment was administered and the clinical outcome was assessed. “Fit” pts were assigned standard chemotherapy, “intermediate” pts tailored (chemo-) therapy, “frail” pts monochemotherapy (as “supportive” therapy) or only “supportive care”. The primary endpoint of the study was to correlate CGA with treatment and clinical outcome which was based on: objective clinical response (RECIST), ECOG PS, toxicity (NCI CTC v.3), survival. Patients who completed at least 3 months of treatment were evaluable. At December 2006, 136 pts were enrolled; 114 (mean age 73.6 years) 104 of whom stage IV were evaluable: 36 “fit”, 36 “intermediate” and 42 “frail”. At baseline no difference was found between the CGA categories as for clinical characteristics (M/F ratio, stage, tumor site) except for ECOG PS which was significantly higher in frail pts (ANOVA test). Results: At 3 months the objective clinical response to the assigned treatment was: 9 CR, 9 PR, 13 SD and 5 PD for “fit” pts; 9 PR, 14 SD and 13 PD for “intermediate” pts; 6 PR, 11 SD and 25 PD for “frail” pts. As for the correlation of CGA categories with treatment and clinical outcome: a significantly higher OR, particularly CR, in favor of fit pts was observed, ECOG PS was maintained significantly higher in frail pts, toxicity was the highest in intermediate pts, the dose intensity was highest in fit pts. At the multivariate regression analisys the independent predictive factors for the clinical response were initial CGA and dose intensity. Conclusions: The study confirms: 1) Fit pts should be treated with the most active regimens not differently from adult pts; 2) the “intermediate” group should be probably better assessed as for inclusion criteria; 3) frail pts need a more exhaustive assessment of the quality of life currently not included in the CGA. In conclusion, the CGA assessment is strongly recommended as an essential tool for the clinical evaluation of elderly pts. Large prospective clinical trials in this field are awaited. No significant financial relationships to disclose.

An attempt to correlate “comprehensive geriatric assessment” (CGA), treatment assignment and clinical outcome in elderly cancer patients: Preliminary results of a phase II open study

18599 Background: The aim of our study was to verify the feasibility of using the CGA as an effective instrumental tool for treatment assignment and verify its influence on clinical outcome in elderly cancer patients (pts). Methods: The study design was a prospective Phase II open study. Elderly (age ≥65 years) cancer pts were assigned to 3 different CGA categories (“Fit”, “Intermediate” and “Frail”): accordingly, an appropriate treatment was administered and the clinical outcome was assessed. “Fit” pts were assigned standard chemotherapy, “intermediate” pts tailored (chemo) therapy, “frail” pts monochemotherapy (as “supportive” therapy) or only “supportive” therapy. The primary endpoint of the study was to correlate CGA with treatment and clinical outcome which was based on: objective clinical response (RECIST), ECOG PS, toxicity (NCI CTC v.3), survival, quality of life. Patients who completed at least 3 months of treatment were evaluable. Results: At January 2006, 72 pts were enrolled (mean age 74.4 years, range 65–91, M/F 39/33), 35 of whom (34 stage IV) were evaluable: 5 were “fit”, 16 “intermediate” and 14 “frail”. At baseline no difference was found between the CGA categories in the clinical characteristics (M/F ratio, stage, tumor site) except for ECOG PS which was significantly higher in frail pts (ANOVA test). The objective clinical response to the assigned treatment was: 1 PR, 1 SD and 3 PD for “fit” pts; 1 PR, 9 SD and 6 PD for “intermediate” pts; 4 PR, 4 SD and 6 PD for “frail” pts. Comprehensively, the ORR was 17.1% and the median survival was 4.15 months; 25 out of 35 pts are alive. As for the correlation of CGA categories with treatment and clinical outcome, no difference was found in the clinical outcome variables after 3 months of treatment (ANOVA test). Our results are preliminary as the study is still in progress: the required accrual is at least 32 pts for each CGA category. Conclusions: The CGA assessment is strongly recommended as an essential component of the clinical evaluation of elderly pts. Large prospective clinical trials in this field are awaited. Work Supported by: MIUR, Rome, Italy: NRP No. 2004067078. No significant financial relationships to disclose.

An alternative schedule of oral vinorelbine tested as 2nd-line treatment on elderly patients with advanced non-small cell lung cancer (NSCLC): A phase I feasibility trial

7328 Background: Oral vinorelbine (O-VNR), administered weekly until PD, has shown good activity as 1st-line treatment in elderly pts (RR:11%, MST 8.2 mos), though G3–4 neutropenia and G1–2 diarrhea were reported in 50% and 33% of pts respectively (Gridelli, Eur J Cancer 2004). Therefore, in the attempt to reduce toxicity, we performed this phase I study to test an alternative schedule of O-VNR, administered 2 weeks out of 3, in 2nd-line setting of NSCLC elderly pts relapsed after front-line treatment. Methods: From Aug 2003 till Oct 2004, 18 consecutive elderly pts (>70y) with advanced NSCLC progressing after first-line therapy received O-VNR 60 mg/m2 dd 1,8 every 3 wks. Pts’ characteristics were as follows: median age 74 y (range 70–79), gender M/F 83%/17%, histology adeno/squamous/undiff. 33%/50%/17%, ECOG PS 0/1/2 5%/72%/22%. First-line chemotherapy: carboplatin-based doublets in fit pts (73%), single agent gemcitabine in frail pts (27%). Response was assessed after 3 cycles of therapy. Results: 47 cy...