Malnutrition affects up to 75% of cancer patients and results from a combination of anorexia and metabolic dysregulation. Metabolic and nutritional abnormalities in cancer patients can lead to cachexia, a multifactorial syndrome characterized by involuntary loss of skeletal muscle mass, systemic inflammation and increased protein catabolism. Cancer cachexia negatively affects patients' outcomes, response to anticancer treatments, quality of life, and survival. However, risk of malnutrition, and cachexia are still under-recognized in cancer patients. The Prevalence of Malnutrition in Oncology (PreMiO) study revealed that 51% of patients already had nutritional deficiencies at their first medical oncology visit. Here, we report the results of the subsequent retrospective, observational NUTRItional status at first medical oncology visit ON Clinical Outcomes (NUTRIONCO) study, aimed at assessing the impact of baseline nutritional and non-nutritional variables collected in the PreMiO study on the clinical outcomes of the same patients followed up from August 2019 to October 2021. We have highlighted a statistically significant association between baseline variables and patient death, rehospitalization, treatment toxicity, and disease progression at follow-up. We found a higher overall survival probability in the well-nourished general study population vs. malnourished patients (p < 0.001). Of major interest is the fact that patient stratification revealed that malnutrition decreased survival probability in non-metastatic patients but not in metastatic patients (p < 0.001). Multivariate analysis confirmed that baseline malnutrition (p = 0.004) and VAS score for appetite loss (p = 0.0104), in addition to albumin < 35 g/L (p < 0.0001) and neutrophil/lymphocyte ratio > 3 (p = 0.0007), were independently associated with the death of non-metastatic patients at follow-up. These findings highlight the importance of proactive, early management of malnutrition and cachexia in cancer patients, and in particular, in non-metastatic patients, from the perspective of a substantial improvement of their clinical outcomes.

S. Lopatriello*, D. Amoroso, S. Donati, O. Alabiso, L. Forti, A. Fornasiero, A. Smergo, A. Lalli, C. Iacono, A. Lucenti, L. D’Alonzo, C. Negrini PBE Consulting, 37121 Verona, Italy Medical Oncology Unit, Versilia Hospital and Tuscany Tumor Institute, Lido di Camaiore, Lucca, Italy Medical Oncology Unit, Azienda Ospedaliera-Universitaria Maggiore della Carita‘, Novara, Italy Medical Oncology Unit, Ospedale Immacolata Concezione, Piove di Sacco, Padua, Italy Medical Oncology Unit, Ospedale Maria SS d. Splendore, Giulianova, Teramo, Italy Medical Oncology Unit, Azienda Ospedaliera ‘Civile-Maria Paterno Arezzo’, Ragusa, Italy CM Pharma, CM Sistemi, Milan, Italy

Aim: To describe the healthcare resource consumption of metastatic colorectal cancer (MCRC) patients in the Italian healthcare setting.Methods: A retrospective chart analysis estimating direct medical costs of first-line infusional 5-Fluorouracil (5-FU) or oral Capecitabine (CAP), associated or not with other chemotherapies, from the Italian Healthcare Service (IHCS) and Hospital (H) perspectives.Results: 202 subjects were analysed. CAP patients (N = 66) were older, with a more compromised clinical status and received less chemotherapy agents in association than 5-FU patients (N = 136). From the IHCS perspective, mean total costs per patient were is an element of 12,029 and is an element of 5,781 in the 5-FU and CAP arms respectively; is an element of 7,338 and is an element of 4,688 from the H perspective. The infusional administration route of 5-FU was a cost driver from both perspectives. Sensitivity analyses found the results to be robust to variations in base case parameters.Conclusions: Management of MCRC by oral chemotherapies may be an economically advantageous option to both IHCS and hospitals. (C) 2008 Elsevier Ltd. All rights reserved.

11036 Background: Clinico-pathological patterns, management and outcomes of HR+ vs HR- have been evaluated in a retrospective study. Methods: A series of 2686 BC patients (pts), treated at our institution from 1988 to 2000 with known immunohistochemical receptor status was investigated. Correlation between HR status and tumour size (T), nodal status (N), grading (G), mib-1, HER-2 and p53 and was evaluated by χ square test. Statistical analysis was performed to test the interaction between HR status and Disease Free Survival (DFS) and with specific and overall survival, using the log-rank test. Results: Among 2686 patients, 467 HR- (17%) and 2219 (83%) HR+ (364 pre and 1855 postmenopausal) were observed, with 1586 ER+PgR+, 555 ER+PgR-, 78 ER-PgR+. Pts with positive HR status showed a statistically significant correlation with lower T, G, mib-1, HER-2 and p53 expression.Chemotherapy was administered to 681 HR+BC (31% ) and to 301 HR-BC (65%), according to stage, age, and G. Tamoxifen based endocrine therapy was administered to 1167 HR+BC (53%) (associated to chemotherapy in 428 and alone in 739 pts). At a median follow up of 108 months, 499 relapses occurred in HR+ (22%) , and 141 in HR- pts (30%), with a 5,10, 15y DFS of 83%,72%, 62% in HR+ and 72%,66%, 57% respectively in HR- pts. More frequent sites of relapse were soft tissue (46%) and bone (35%) in HR+ and soft tissue (50%) and visceral (30%) in HR- pts. Seven hundred sixty pts have died: 576 HR+ (26%) and 184 HR- (39%), with a 5, 10, 15y specific survival of 90%, 84%, 75% which was significantly superior to the 78%, 73%, 69% respectively of HR-. However, considering that 230 HR+ and 44 HR- pts died without relapse, the 5,10,15y overall survival was 90%, 75% 61% for HR+ and 78%, 67%, 60% respectively for HR-. It appears therefore that the specific survival advantage in early years in favour of HR+ is progressively reduced by non specific mortality in late years which results in similar 15 y OS rates among HR+ and HR- pts. Conclusions: Our series appear to confirm the relationship between HR+ status and other more favourable clinical-pathological features, although long term OS is not apparently affected. No significant financial relationships to disclose.

10669 Background: A consecutive series of br.ca. pts, treated between Jan 1st 1990 to Dec 31st 1999 in our Department, is the basis of our retrospective study, aimed to create a data base on routinary clinical management of early br.ca. pts, to which compare similar series and literature data. Methods: All Clinical Records were reviewed and computerized. Disease free and overall survival were estimated using the product-limit method of Kaplan and Meier. The log-rank test was used to compare prognosis between different subgroups. Results: Among 2924 consecutive br.ca. pts, 836 were younger than 50 years (med. age 44) and 2088 older (med. age 63). Regional nodes were negative (N−) in 1754, positive (N+) in 1027 and unknown in the remaining pts. So, 2593 pts were stage I-II and 301 stage IIIA-B. Hormonal Receptor status (available on 2560 pts) was positive for Estrogen (ER+) in 2021 pts and for Progesterone (PgR+) in 1649 pts. Moreover, 1571 pts were ER+Pgr+, 539 ER-PgR−, 78 ER-PgR+ and 461 ER+PgR−. HER2 was overexpressed in 262/1426 (18%) pts. Tumor grading (available on 2176 cases) was G1–2 in 1411 and G3–4 in 765 cases. After surgery, 731 pts received adjuvant Tamoxifen, 507 pts CMF ± Antracyclines chemotherapy, 434 pts both chemotherapy and Tamoxifen and 958 pts none. (no therapy data are available for the remaining 334 pts). At a median f.up of 9.8 years, 993/2924 pts (33.9%) have recurred, (med. DFS 137 mos) with a 5, 10 and 15 y probability of recurrence of 26, 44 and 63% respectively. Corresponding figures of recurrence for N− pts were 14, 30 and 50% (med. DFS 168 mos), while for N+ pts were 41, 61 and 77% (med DFS 81 mos). For younger N+ pts treated with chemotherapy, the 5 years probability of recurrence was 34% while it was 24% for older ER+ pts treated with hormonal therapy. So far, 794/2924 (27.5%) pts have died, with a 5, 10 and 15 y probability of death of 13, 27 and 41%. This was 5, 16 and 28% for N- pts and 22, 41 and 56% for N+ pts. For younger N+ pts treated with chemotherapy, the 5 y probability of death was 14%, as it was for older ER+ pts treated with Tamoxifen. Conclusions: Although this data are not yet conclusive, it appears that large part of the clinical improvements reported in clinical trials may be achieved in the routine management of breast cancer pts. No significant financial relationships to disclose.

BACKGROUNDTo determine if protracted low-dose oral idarubicin (IDA), feasible in a previous dose-finding study, would result in similar activity and a better toxicity profile in patients with metastatic breast cancer.PATIENTS AND METHODSElderly women (> or=65 years) with metastatic breast carcinoma were treated with 7.5 mg/day for 21 consecutive days, every 4 weeks. After the first fourteen patients, due to excessive toxicity, the protocol was amended to 5 mg/day. IDA and Idarubicinol (IDOL) plasma concentrations (C(trough)) were investigated in all patients.RESULTSBetween April 1999 and June 2004, 47 elderly patients were accrued in this two-part study (14 and 33 patients respectively). The median age was 74 and 75 years respectively. Visceral involvement was present in most patients. A partial response was noted in 7/31 patients (22%; 95% CI, 9.6-41.1%). Eleven patients had stable disease (33%). At the dose of 5 mg/day the treatment was well tolerated. Neutropenia grade 4 was present in only 6% of patients; alopecia > grade 1 and cardiotoxicity did not occur. The median time to progression was 3 months and the median overall survival was 17 months. IDA C(trough) and IDOL C(trough) levels were significantly associated with haematologic toxicity.CONCLUSIONThis study shows that idarubicin at the dose of 5 mg/day for 21 consecutive days is feasible and effective in elderly breast cancer patients but do not demonstrate an improvement in efficacy. A determination of the IDA and IDOL plasma levels (C(trough)) is predictive for toxicity.

727 Background: Anthracyclines are among the most active agents in the treatment of MBC. There is a concern in their use in elderly pts due to possible cardiotoxicity Methods: Pts with MBC, measurable or evaluable lesions, aged 65 or more, PS 0–2, no prior anthracyclines, were treated with IDA initially 7.5 mg/day and after the first 14 pts with 5 mg/day for 21 consecutive days q 4 weeks. The primary aim of the study was the evaluation of activity and tolerability, secondary aim was to correlate the toxicity profile with the plasmatic pharmacokinetics of the metabolite. IDA and IDOL plasma concentrations were investigated on day 0, 7, 14, 21. Results: Between April 1999 and June 2004, 47 pts (14 treated with 7.5 mg and 33 treated with 5 mg/day) were accrued. Median age was 74 and 75 years respectively. Visceral involvement was present in 18/33 (55%). Median number of metastatic sites was 1. Toxicity was low with 5 mg/day (neutropenia grade 4 in 6%). Alopecia grade > 1 and cardiotoxicity did not occur. The table below gives the results of the overall study. IDA and IDOL plasma levels will be presented. Median TTP was 6.7 months, median OS was 22.5 months. Conclusions: This schedule of IDA is feasible and gives elderly pts an opportunity to be treated at home with an active drug whose main toxicity is neutropenia. No significant financial relationships to disclose.

5071 Background: . Gemcitabine (G) and Antracyclines are active as single agent in resistant or refractory ovarian cancer (OC). The aim of this study was to assess the efficacy and toxicity of G and Epirubicin (E) combination in patients (pts) with recurrent OC. Methods: Pts with recurrent epithelial OC or peritoneal carcinoma or carcinoma of the tuba, failing platinum- regimens (free-interval ≤ 12 months), measurable tumor, WHO PS ≤ 2, age ≥18 and ≤ 75, no previous antracycline were included. Treatment consisted of G 1000 mg/m2(day 1, 8) and E 60 mg/m2 (day 1) every 3 weeks for 3 to 6 cycles (in pts with objective response or stable disease after 3 cycles). Between june 02 and october 04, 26 pts were included. Median age: 59 yrs (43–74). Median PS: 1 (0–2). Figo: Stage III 12 and stage IV 14. Serous: 18/26; endometrioid 1/26; unclassified adenocarcinoma: 5/26; tuba 1 and peritoneal carcinoma 1. Median platinum free interval was 6 months (range 0–12). Median of previous regimens was 1 (range 1–4). Results: 132 cycles were administered (median 6 for patient). Two pts are not evaluable for response (one too early, the other for early PS worsening and PD). 24 pts are evaluable for response, according to Recist criteria. Overall response rate (ORR) was 45,8% (1 CR+ 10 PR). 8 pts (33,3%) had SD and 5 (20,8%)PD. Median time to response was 10 wks (range 8–20) and median duration of response was 8 mos (range 5–12). Kaplan-Meier estimate of 8 mos OS was 81% and the PFS 42,8%. Toxicity (NCI criteria) in the 26 pts evaluable was: Grade 3–4: leucopenia 30%, neutropenia 57%, thrombocytopenia 4%, anemia 11,5%, liver 15%, mucositis 7,6%; 8 pts 30% required G-CSF and 3 pts 11,5% blood transfusion. The median interval between cycles was 21 days, only 13 treatment delays for toxicity. No life threatening events occurred. Conclusions: In our study the combination of G and E has shown in these platinum resistant/refractory recurrent OC pts encouraging activity with a 45% ORR and 33% SD, and manageable toxicity. No significant financial relationships to disclose.

686 Background: The tissue microarray (TMA) technique allows to assemble multiple tumor samples in the same paraffin block to evaluate multiple markers using immunohistochemistry. Methods: we have constructed TMAs of 293 breast carcinomas with complete follow-up (median 132 mos). Four tumor cores per sample (i.e.: 4 replica) were included in TMA blocks. We evaluated the expression of ER, PgR, Her2/neu (Herceptest), p53, MIB1, Bcl2, p27, p63, HMW cytokeratin (HMWCK), and FITH. The 4 different replica of each case were separately analysed by two pathologists. Her2/neu was evaluated according to FDA score. For all other markers we evaluated cellular compartimentalization, intensity (from 0 to 3) and percentage of reacting cells. Results: from the original 1172 cores in the TMAs, 78% were evaluable for each marker because of lack of cells or loss of cores during the TMA sections processing. The most relevant prognostic markers were Her2/neu and p53: they retained prognostic value even after adjusting for N status (p<0.01 for OS and DFS). MIB1 was of borderline prognostic significance. ER, PgR, Bcl2, p27, FITH, p63 and HMWCK did not provide prognostic information. Conclusions: Our results confirm that TMA may be a reliable tool in the evaluation of prognostic/predictive markers, with the advantage of sparing time and tissues. The most reliable prognostic marker was Her2/neu. These data further support that Her2/neu, beside being predictive of therapeutic response to Herceptin, is a reliable prognostic marker. Unexpected negative results for ER,PgR and Bcl2 are more difficult to explain, and further studies on larger TMA are necessary.Loss of p27 or FITH, and the identification of markers of the so-called “basal phenotype” (p63 and HMWCK) of breast carcinomas were not prognostic predictors in our series, in keeping with their known controversial prognostic role No significant financial relationships to disclose.

5405 Gemcitabine is a fluorine deoxycitidine analog that has shown clinical activity against several solid tumor including ovarian cancer, pancreatic cancer and in non-small-cell lung cancer (NSCLC) when combined with cisplatin (DDP). Gemcitabine is a pro-drug that requires for its activity, intracellular activation to triphosphate form (dFdCTP). Gemcitabine is rapidly deactivated in plasma by cytidine deaminase to 2’,2’-difluorodeoxycitidine (dFdU). Previous studies demonstrated that the achievement of a threshold value of plasma concentration of Gemcitabine superior to 10 μM is crucial to obtain intracellular accumulation of dFdCTP. Prolonged infusion of Gemcitabine at fixed dose rate (FDR) of 10 mg/m2/min compared with standard 30min infusion are able to maintain for a longer period the above mentioned critical plasma concentration, increasing at maximal levels the rate of accumulation of the active metabolite dFdCTP. In order to evaluate the pharmacological profile of Gemcitabine administered as FDR infusion in NSCLC patients treated in combination with DDP, we designed a dose-finding and pharmacokinetic study. Gemcitabine was administered at FDR of 10 mg/m2/min at doses ranging between 600 and 1200 mg/m2 on day 1 and 8, with 75 mg/m2 of DDP on day 8. Blood samples were collected on day 1 and 8 during the infusions and up to 24 hours. Drug and dFdU plasma levels were determined by HPLC/MS/MS (LOQ 10 ng/ml). Pharmacokinetic results obtained in 22 patients revealed that DDP does not affect Gemcitabine pharmacokinetics. The target plasma concentration of 10μM was achieved in 70% of the 12 patients treated with 600-900 mg/m2 (range 10.1-17.5 μM), but surprisingly in only 40% of the 10 patients treated with 1000-1200 mg/m2 (range 7.5-12.7 μM). These data were in accord with the observation made after 3 cycles, that the majority of the responses (8/13) were obtained with the dose 600-900 mg/m2 and that the treatment was well tolerated even at the higher doses of 1000-1200 mg/m2. Further pharmacokinetic studies that involve also the measure of dFdCTP are in progress to define the optimal schedule of Gemcitabine administration as FDR infusion. (Partially supported by Eli Lilly Italia).

9600 Background: We investigated the prognostic significance of Her2/neu expression using FDA approved Herceptest® and the tissue microarray (TMA) technique. We used our in-house developed WATMAS, and investigated the most appropriate scoring system to be applied to the analysis of multiple replica samples from the same tumors. Methods: WATMAS includes Beecher TMA arrayer, robotized microscope, relational database and internet technology through DHTML, XML, JavaScript and VBScript. WATMAS allowed remote analysis of the TMA slides. 436 breast cancers were analysed; complete follow-up was obtained for 185 patients (120m median f-up). Four 1 mm cores per sample (i.e.: 4 replica) were included in TMA blocks; sections were immunostained with Herceptest®. For each core we evaluated using the FDA approved score, and data were stored in the database through web interface. Scores of the 4 different replica of each case were separately analysed in order to obtain the most clinically representative value. Results: The maximum score (Maxscore) of the 4 replicas pertaining to the single case was selected as the most representative. Maxscores were distributed as follows: 0= 60%, 1= 20%, 2=12% and 3= 8%. High Maxscore was associated with high grade (p=0.03) and with poor overall and relapse free survival in the whole series of cases, and in N1/2 subgroup. Prognostic value was retained after multivariate analysis. Conclusions: Our study is one of the largest published studies of Her2/neu expression in relation to prognosis using TMA and the FDA approved Herceptest®. Our results confirm the value of TMA in the evaluation of prognostic/predictive markers in breast cancer and the need of an appropriately tailored scoring system. Our WATMAS improved reliability and ease in the evaluation of TMA and allowed remote cooperation between researchers. No significant financial relationships to disclose.