Search

International audience; BackgroundNETSARC (netsarc.org) is a network of 26 sarcoma reference centers with specialized multidisciplinary tumor boards (MDTB) aiming to improve the outcome of sarcoma patients. Since 2010, presentation to an MDTB and expert pathological review are mandatory for sarcoma patients nationwide. In the present work, the impact of surgery in a reference center on the survival of sarcoma patients investigated using this national NETSARC registry.Patients and methodsPatients’ characteristics and follow-up are prospectively collected and data monitored. Descriptive, uni- and multivariate analysis of prognostic factors were conducted in the entire series (N = 35 784) and in the subgroup of incident patient population (N = 29 497).ResultsAmong the 35 784 patients, 155 different histological subtypes were reported. 4310 (11.6%) patients were metastatic at diagnosis. Previous cancer, previous radiotherapy, neurofibromatosis type 1 (NF1), and Li–Fraumeni syndrome were reported in 12.5%, 3.6%, 0.7%, and 0.1% of patients respectively. Among the 29 497 incident patients, 25 851 (87.6%) patients had surgical removal of the sarcoma, including 9949 (33.7%) operated in a NETSARC center. Location, grade, age, size, depth, histotypes, gender, NF1, and surgery outside a NETSARC center all correlated to overall survival (OS), local relapse free survival (LRFS), and event-free survival (EFS) in the incident patient population. NF1 history was one of the strongest adverse prognostic factors for LRFS, EFS, and OS. Presentation to an MDTB was associated with an improved LRFS and EFS, but was an adverse prognostic factor for OS if surgery was not carried out in a reference center. In multivariate analysis, surgery in a NETSARC center was positively correlated with LRFS, EFS, and OS [P

The natural history of localized high-grade sarcomas of the digestive tract (SDT) excluding GIST has been rarely considered owing to their low incidence and heterogeneity. We describe the histoclinical characteristics of SDT and correlate them with patients' outcomes.We retrospectively collected medical files from a European database covering connective tissue tumors listed in Europe for about twenty years. Only untreated localized primary high-grade SDT were included. A central histological review was performed for each case. Patients' characteristics were compared and correlated with clinical outcomes.A total of 45 patients were identified. Leiomyosarcomas (LMS) and undifferentiated sarcomas (UDS) were predominant, the former having better overall survival (OS) and progressionfree survival (PFS) while the latter having a worse outcome than the other histological types. Complete remission was obtained in 34 patients (75%) and was associated with male sex, age over 40 years and monofocal tumor. Complete surgery and LMS histology were associated with a better prognosis without any significant difference in baseline characteristics or in treatment modalities.Complete surgery and histological type seem to be prognostic indicators of SDT. These results suggest the importance of treating these patients in a reference center.

Background NETSARC (netsarc.org) is a network of 26 sarcoma reference centers with specialized multidisciplinary tumor boards (MDTB) aiming to improve the outcome of sarcoma patients. Since 2010, presentation to an MDTB and expert pathological review are mandatory for sarcoma patients nationwide. In the present work, the impact of surgery in a reference center on the survival of sarcoma patients investigated using this national NETSARC registry. Patients and methods Patients’ characteristics and follow-up are prospectively collected and data monitored. Descriptive, uni- and multivariate analysis of prognostic factors were conducted in the entire series (N = 35 784) and in the subgroup of incident patient population (N = 29 497). Results Among the 35 784 patients, 155 different histological subtypes were reported. 4310 (11.6%) patients were metastatic at diagnosis. Previous cancer, previous radiotherapy, neurofibromatosis type 1 (NF1), and Li–Fraumeni syndrome were reported in 12.5%, 3.6%, 0.7%, and 0.1% of patients respectively. Among the 29 497 incident patients, 25 851 (87.6%) patients had surgical removal of the sarcoma, including 9949 (33.7%) operated in a NETSARC center. Location, grade, age, size, depth, histotypes, gender, NF1, and surgery outside a NETSARC center all correlated to overall survival (OS), local relapse free survival (LRFS), and event-free survival (EFS) in the incident patient population. NF1 history was one of the strongest adverse prognostic factors for LRFS, EFS, and OS. Presentation to an MDTB was associated with an improved LRFS and EFS, but was an adverse prognostic factor for OS if surgery was not carried out in a reference center. In multivariate analysis, surgery in a NETSARC center was positively correlated with LRFS, EFS, and OS [P

Limited information is available on clinical management of Flat Bone Osteosarcomas (FBOS). We retrospectively analysed prognostic factors and outcome. Twenty-eight patients were treated in our institution. Survival curves were obtained by the Kaplan-Meier method and compared with the log-rank test. The overall survival (OS) rates at 5 and 10 years were 52.4% and 45.8% respectively. The event-free survival (EFS) rates at 5 and 10 years were 41.5%. The factors influencing EFS in univariate analysis were location, metastatic disease at diagnosis, effect of neoadjuvant chemotherapy, histological response and adequate local tumour control. Location, metastatic disease at diagnosis, effect of neoadjuvant chemotherapy, histological response and local recurrence were statistically correlated with OS. Multivariate analysis retained metastatic disease at diagnosis as prognostic factors of EFS and OS. Our results suggest a more favourable outcome of FBOS as the use of a treatment scheme based on the protocols for long bone osteosarcomas. However, an adequate local treatment is essential to ensure a better outcome.

Les therapeutiques ciblees sont en train d'enrichir largement l'arsenal therapeutique en cancerologie. L'eventail des cibles moleculaires possibles et les proprietes de ces nouvelles therapeutiques ciblees amenent a se poser plusieurs questions: selon quelle approche ces therapeutiques devraient-elles etre utilisees? Therapeutique ciblee sur une seule ou sur plusieurs cibles moleculaires? Selon quelles modalites devraientelles etre utilisees? En monotherapie dirigee contre plusieurs cibles ou en combinant plusieurs therapeutiques moleculaires ciblees? Dans quel contexte les utiliser? Pour bloquer simultanement des cibles moleculaires «precoces» et «plus tardives», pour bloquer simultanement des cibles moleculaires «tardives ou secondaires» ou surtout pour bloquer les voies de resistance. Cet article expose les proprietes des therapeutiques uni et multicibles avant d'envisager les moyens de bloquer les voies de resistance en prenant notamment deux exemples de therapeutiques multicibles utilisees apres echec de therapeutique plus selective: dans les GIST resistants a l'imatinib et dans les adenocarcinomes mammaires resistants au trastuzumab.

Summary Objective: A Bayesian population pharmacokinetics study of data from routine therapeutic drug monitoring cisplatin during a 5-day infusion of cisplatin. Methods: A total of 95 kinetics data sets (58 patients) were available to perform this study. Individual pharmacokinetic parameters, estimated from 20 courses of treatment in 18 patients, were used to calculate the population parameters (cl: 0·175 ± 0·034 L/h; t½: 327 ± 91 h). The accuracy of Bayesian forecasting was tested by comparing in 40 other patients the clearance values calculated either from a complete kinetics profile (eight sampling times) or from three early samples and the new population parameters. Finally, drug monitoring efficacy was assessed by comparing the target Cmax values with the Cmax obtained after dose adjustment based upon early Bayesian estimation of the individual pharmacokinetic parameters. Results: No significant difference was found between Bayesian and experimental clearances. Besides, dose-individualization proved to successfully adjust Cmax values around their respective target. Conclusion: The new reference pharmacokinetic population parameters lead to accurate estimation of individual pharmacokinetic parameters from a limited number of samples, thus allowing efficient therapeutic drug monitoring during 5-day infusion regimens of cisplatin.

Resume Les adenocarcinomes oesocardiaux representent actuellement en France au moins 20 a 40 % des cancers de l’œsophage. IIS presentent un risque eleve de recidive ganglionnaire et metastatique. Seulement 50 a 70 % peuvent beneficier d’une exerese a visee curative au moment du diagnostic et parmi ceux-ci plus de la moitie recidivent. Le traitement de reference des formes metastatiques est devenu la chimiotherapie. Trois etudes controlees ont montre que la chimiotherapie, par rapport au simple traitement symptomatique, permet d’allonger la duree mediane de survie de 3–4 mois a 10–12 mois et d’ameliorer la qualite de vie des patients. Le protocole epirubicine-cisplatine-5-fluoro-uracile en perfusion continue (ECF) est le plus efficace actuellement mais s’avere tres astreignant pour les patients. En France, l’association la plus utilisee reste encore 5-fluoro-uracilecisplatine. Les nouvelles molecules (docetaxel, CPT11, oxaliplatine) utilisees seules ou en association, notamment avec le 5-fluoro-uracile, semblent prometteuses. Le traitement privilegie des recidives locoregionales des adenocarcinomes œsocardiaux est la chimioradiotherapie. Les resultats actuels des etudes de chimioradiotherapie concomitante, a base de 5-fluoro-uracile seul, 5-fluoro-uracile-cisplatine ou 5-fluoro-uracile-mitomycine, dans les cancers de l’œsophage, en situation pre-operatoire ou en traitement exclusif, justifient l’utilisation d’une chimioradiotherapie en cas de recidive locoregionale apres chirurgie d’exerese. Cette association est efficace sur la dysphagie et permet parfois l’obtention de reponses tumorales completes. Le schema optimal de chimioradiotherapie reste encore a definir (dose et etalement de la radiotherapie, choix des medicaments anticancereux).

We investigated whether head-and-neck cancers are associated with an increased micronucleated cell rates (MN cell rates) and whether risk factors for these cancers are associated with alterations in micronucleated lymphocytes. MN cell rates were assessed in cytokinesis-blocked lymphocytes of 57 head-and-neck cancer patients (CP) before any anticancer treatment and of 198 male and female healthy subjects (HS). In the HS group, only smoking status significantly affect MN cell rates. In CP group age, sex, tobacco status, alcohol status, tumor stage, family history of cancer had no significant effect. For the non-smokers, the comparison between MN cell rates in HS and CP adjusted for age and sex showed a significant difference. The increase of MN cell rates in non-smokers patients may be attributable to cancer status. For the smokers, the comparison of MN cell rates in HS and CP matched for age and sex showed no significant difference. Pathological status could mask the smoking effect on peripheral blood lymphocytes in patients. Moreover, it probably could partly explain why MN cell rates in matched-CP smokers and HS smokers were similar. The authors do not recommend the CBMN assay in this present form to study smoking DNA-damage effects in peripheral blood lymphocytes of cancer patients, especially for patients with upper aero-digestive tract cancers or lung cancers for which tobacco is the major risk factor.

Paclitaxel is a recent chemotherapeutic agent that inhibits tubulin depolymerization in tumoral cells. Despite its increasing use against various human cancers, the genotoxicity of paclitaxel has never been studied on normal human cells. The in vitro genotoxic effects of the drug were evaluated with two complementary mutagenesis tests on human T-lymphocytes: (1) the cytokinesis-blocked micronuclei assay (CBMN) in combination with fluorescent in situ hybridization (FISH) of nonspecific centromeric probes and (2) the comet assay performed in three ways: on stimulated lymphocytes as in the CBMN, and on freshly isolated lymphocytes at both 4 and 37 degrees C. A slight cytotoxicity of 2.5 to 10 nM paclitaxel was found in the CBMN and a significant increase in the binucleated micronucleated cell rates was observed, with a concentration-dependent manner. In the FISH analysis, more than 85% of the micronuclei (MN) were centromere positive, and a ratio of 72. 2 to 78.6% of these MN contained more than one centromere. Moreover, at 10 nM of paclitaxel, 35.6% of the cells are multimicronucleated lymphocytes. Unexpectedly, paclitaxel induced single-strand breaks on proliferating lymphocytes at 5 and 7.5 nM but not in resting cells, even at 5 to 15 microM. These in vitro results showed that (1) paclitaxel does not present any direct DNA action in resting cells, (2) DNA damage detected in stimulated lymphocytes may be linked either to a high frequency of cells in the S-phase cell cycle or to a direct DNA damaging effect on replicating cells, and (3) paclitaxel is a strong in vitro aneugenic drug on human normal cells, at clinically relevant concentrations.

Resume Introduction Le syndrome de Maffucci-Kast est une maladie congenitale, non hereditaire, tres proche de la maladie de Ollier,associant des hemangiomes cutanes multiples, une dyschondroplasie et souvent des enchondromes. Exegese Nous rapportons le cas unique d'une double localisation synchrone d'un chondrosarcome des fosses nasales et de la paroi thoracique anterieure revelant un syndrome de Maffucci-Kast. Conclusion Une localisation chondrosarcomateuse atypique apparemment isolee doit faire rechercher l'existence d'une enchondromatose multiple.

Introduction: Because of the increased use of im plantable access systems, the incidence of blood stream and catheter infections associated with these systems has concomitantly increased. Classically, he parin-lock flush solutions were used to prevent thrombosis; more recently, vancomycin was added to the solution to prevent infections caused by Gram- positive bacteria, particularly coagulase-negative Staphylococci. Disorders due to Gram-negative organ isms have now appeared in oncologic patients. We therefore tested the addition of colistin to heparin- vancomycin solutions. Colistin was chosen for its good activity against Gram-negative bacteria (98% susceptibility in our hospital), its good tolerance due to low systemic passage, and its low cost. Methods: We developed formulations contain ing heparin (100 IU/mL) and various concentrations of vancomycin (10-500 μg/mL) and colistin (10-100 μg/mL) in 0.9% NaCl. Each sterile solution was tested for physical and chemical compatibility (spectropho tometry, nuclear magnetic resonance, and pH mea surements) and its antibacterial activity (against ox acillin-resistant Staphylococcus aureus, Enterococcus faecium, Klebsiella pneumoniae exhibiting broad- spectrum betalactamase (BSBL), imipenem-resistant Pseudomonas aeruginosa) for 2 months at 4°C and at room temperature. Results: The most suitable combination of drugs is heparin (100 IU/mL), vancomycin (100 μg/mL), and colistin (100 μg/mL). This flush solution main tains activity when stored at 4°C for up to 1 month. Conclusions: We feel that the combination of heparin, vancomycin, and colistin can be used as a flush solution for indwelling catheters.

Micronucleated cell rates were assessed in cytokinesis-blocked lymphocytes of 198 male and female healthysubjects (HS) not occupational^ exposed to genotoxic risksand of 70 male and female cancer patients (CP) prior toany anticancer treatment. In the HS group, spontaneousmicronucleated cell rates (MN cell rates) were 9.7 ± 2.8per 1000 binucleated lymphocytes and 9.8 ± 3.1 for malesand females respectively. In the CP group, spontaneousMN cell rates were 21.1 ± 153 per 1000 binucleatedlymphocytes and 19.1 ± 11.2 for males and femalesrespectively. Moreover, they were shown to have a largeinter-individual variability in the two groups. The study ofinter-individual variation factors showed that only tobaccocould affect MN cell rate in HS whereas age and sexapparently had no significant effect. In the CP group, onlyage significantly affected MN cell rate, whereas sex, tobacco,alcohol, imaging techniques and tumour stage had nosignificant effect There was no significant difference in thedistribution of gender between HS and CP, whereas therewas a significant difference in the distribution of age andtobacco between the two groups. The comparison of MNcell rates in 54 HS and 54 CP matched for age and sexshowed a statistically significant difference. SpontaneousMN cell rates of these two populations reflect environmentalexposure. Moreover, for CP it most probably refers tovarious cellular lesions and genetic damage.IntroductionThe survey of human genetic integrity has become of specialinterest since the development of industrial activities andconsequent exposure to chemical and physical genotoxins.Moreover, other factors such as lifestyle, various medicaltherapies and climatic changes may promote genetic damage.Human genetic damage can be assessed by using differentin vitro and in vivo assays (Ishidate, 1988).The validity of the cytokinesis-blocked micronucleus assay(CBMN assay) as a biomarker of chromosome damage individing mammalian cells is well established (O'Donovan,1990). This assay offers an easier and less tedious alternativeto metaphase chromosome analysis, with the advantage thatexposure to both clastogens and aneugens may be detected(Arlett et ai, 1989; Fenech, 1990). This method has led toincreased interest by some cytogeneticists in the applicationof the CBMN assay for population monitoring (Di Georgioet ai, 1994; Duffaud et ai, 1996; Gutierrez et ai, 1995;Klemans et ai, 1995; Norppa et ai, 1993; Villani etai, 1995),although some authors do not agree on this point (VanHummelen et ai, 1994). The occurrence of spontaneousmicronuclei (MNi) in healthy subjects (HS) is well established(Fenech, 1993). Most authors report spontaneous micronucle-ated (MN) cell rates in HS with mean value of ~10 MNi per1000 binucleated lymphocytes (range 4-15) (Hogstedt, 1984;Fenech, 1993).hi order to determine baseline data in healthy subjects theCBMN assay was applied to the lymphocytes of 198 male andfemale subjects. The effects of age, sex and tobacco onspontaneous MN cell rates were studied. Then, the baselineMN cell rate in 70 cancer patients (CP) was determined byapplication of the same assay, and the effects of sex, age,tobacco and alcohol consumption, imaging techniques andtumour stage on MN cell rates were studied. Moreover, thevariation factors distribution in the two groups was studied inorder to compare spontaneous MN cell rates in HS with thoseinCP.Materials and methods

Resume Depuis sa decouverte en 1948, le methotrexate (MTX) a vu son champ d'application s'etendre et c'est en cherchant a vaincre les resistances a ce medicament que le concept de MTX a haute dose a ete propose. Afin de prescrire une posologie optimale du medicament, pour obtenir le meilleur effet therapeutique avec une toxicite moindre, on a realise, parallelement aux etudes de pharmacocinetique, un ≪ sauvetage ≫ par l'acide folinique et un dosage rapide du MTX. La maitrise des variations interet intra-individuelles a beneficie de l'adaptation individuelle de posologie a partir, soit de la dose test, soit de la pharmacocinetique de population avec la methode bayesienne d'estimation des parametres et la possibilite d'adaptation de la dose en temps reel, lors de la perfusion du MTX haute dose (MTX-HD). Apres avoir analyse les differentes methodes, nous presentons des resultats de pratique courante portant sur 778 cures de MTX-HD en separant les 238 cures des osteosarcomes qui constituent un groupe homogene de patients. A partir du choix par le clinicien de la concentration plasmatique maximale theorique et de la duree de perfusion, le biologiste assure le guidage pharmacocinetique avec adaptation individuelle de posologie a la 6e heure de perfusion de MTX. Le traitement se deroule en toute securite pour le patient qui recoit une dose efficace optimale sans apparition d'effets secondaires graves (aucune toxicite de grade 4 de la codification OMS). De plus, nous realisons en moyenne une intensification de posologie au cours des traitements des osteosarcomes puisque nous pouvons donner sans risque 65 % de MTX de plus que la dose preconisee chez l'adulte (8 g/m2) et meme 10 % de plus que la dose proposee chez l'enfant (12 g/m2).

Anticancer drugs dosage are currently adjusted to body surface area. This measure is supposed to be the best morphometric parameter to adjust anticancer drug doses. However, dose adjustment to body surface area has not historically been rigorously demonstrated. We propose a method to objectively test this parameter utility. The statistical justification of drug adjustment to body surface area can use mathematical equations to be expressed. We can demonstrate that body surface area and plasmatic total clearance of a drug should be correlated to adjust dose anticancer drug to body surface area. When we test the hypothesis of body surface area and plasmatic clearance correlation for cytarabine and adriamycin we did not find any significant correlation. For these anticancer drugs, dose adjustment to body surface area increase their pharmacokinetic and efficiency variabilities. The concept of dose-intensity is probably the best justification of individual dose adjustment from plasmatic drug samples, and from pharmacokinetic and pharmacodynamic studies. The determination of the "maximal tolerable exposition" and of the "minimal effective exposition" should reduce the overexpression of toxic risks and avoid the ineffective underexpositions. However, it is difficult to precisely define these two expositions and to research the most relevant pharmacokinetic parameters to their measure. Area under curve appears to be their most appropriate expression.

BACKGROUND Chemotherapy failure that is due to cellular drug resistance remains a major problem in most cancer patients. One type of drug resistance that has been characterized is the multidrug resistance phenomenon, which demonstrates a reduced ability of cancer cells to accumulate drugs as a result of the effects of an energy-dependent unidirectional drug efflux pump with a broad substrate specificity. This drug pump is composed of a 170-kd transmembrane glycoprotein referred to as the P-glycoprotein (P-gp) that uses energy in the form of adenosine triphosphate to transport drugs through a channel formed by transmembrane segments. PURPOSE Our purpose was to detect the levels of P-gp expression in frozen untreated breast carcinomas by immunocytochemical assays and to correlate these levels to current prognostic indicators and, in a few cases, to MDR1 (also known as PGY1) mRNA expression by polymerase chain reaction (PCR). METHODS The immunocytochemical expression of the multidrug resistance gene, P-gp, was investigated using a specific monoclonal antibody (JSB1) against P-gp in 5-microns frozen sequential sections of breast carcinomas obtained from 213 patients. Microscopic images of immunostained preparations were evaluated by image analysis and were compared with MDR1 transcription (mRNA) assessed by PCR in 16 patients. Quantitative P-gp immunocytochemical assays were correlated to histoprognostic factors and immunocytochemical indicators. RESULTS Among the 213 breast carcinomas tested, 113 (53%) were P-gp positive, but in 28% of the tumors, the immunostained surface accounted for less than 5% of the total area stained. Quantitative immunocytochemistry reflecting the amount of intracellular P-gp antigen strongly correlated (r = 0.865; two-sided, P < .0001; Pearson's test) with the quantitative evaluation of the scanner analysis of mRNA transcripts. The P-gp expression was significantly (two-sided, P < .001) correlated with p53 expression in tumors, to cathepsin D and Ki67 (two-sided, P < .01) immunoreactivity, and to a lesser extent, the detection of estrogen receptor antigenic sites (two-sided, P = .019). P-gp expression was found to be independent of expression of progesterone receptor and pS2, pHER-2/neu, and CD31 in tumors and from patient age, tumor size, histologic types, grades and Nottingham prognostic index, and nodal status. CONCLUSIONS The quantitative immunocytochemical assays of P-gp are correlated to PCR analysis of MDR1 expression, and such correlations can be useful in evaluating potential multidrug resistance in breast cancer. However, the clinical significance of P-gp immunodetections remains to be further determined.

Angiogenesis is an essential prerequisite in the growth and dissemination of soft tissue sarcoma. The understanding of the VEGF and VEGFR pathway implication in the development of non GIST soft tissue sarcoma and the discovery of the antitumoral activity of drugs that inhibit this pathway in other solid tumors, led to the development of antiangiogenic drugs anti VEGF in soft tissue sarcoma, as monoclonal antibody (bevacuzimab) or as small molecules (tyrosine-kinase inhibitors anti VEGFR). This manuscript presents the results of the first clinical trials that have evaluated the efficacy and safety of some angiogenesis inhibitors in soft tissue sarcomas.

Limited information is available on clinical management of Flat Bone Osteosarcomas (FBOS). We retrospectively analysed prognostic factors and outcome. Twenty-eight patients were treated in our institution. Survival curves were obtained by the Kaplan-Meier method and compared with the log-rank test. The overall survival (OS) rates at 5 and 10 years were 52.4% and 45.8% respectively. The event-free survival (EFS) rates at 5 and 10 years were 41.5%. The factors influencing EFS in univariate analysis were location, metastatic disease at diagnosis, effect of neoadjuvant chemotherapy, histological response and adequate local tumour control. Location, metastatic disease at diagnosis, effect of neoadjuvant chemotherapy, histological response and local recurrence were statistically correlated with OS. Multivariate analysis retained metastatic disease at diagnosis as prognostic factors of EFS and OS. Our results suggest a more favourable outcome of FBOS as the use of a treatment scheme based on the protocols for long bone osteosarcomas. However, an adequate local treatment is essential to ensure a better outcome.

This research studies social representations of cancer and chemotherapy with patients (N = 62) and of oncologists (N = 26) in a medical oncology unit. The collection of the data was made from free association tasks (concerning cancer and chemotherapy) and attitude questions concerning chemotherapy. Patients also had to produce representations "for" oncologists and conversely. Results indicate a variety of representations marked by the status of sample (patients versus oncologists). The stressful component of the disease is very present. Contrary to the patients, oncologists seize relatively well the representation, which the patients have of concerning cancer and chemotherapy. Patients' representations are connected with phenomena of belief that give evidence for some process of anticipation and expectation linked to the experience of illness and testify the emotional charge related to it. The study of the representations is particularly relevant to highlight the psychosocial stakes associated with the therapeutic situation.

The purpose of this study was to evaluate the activity and safety of OSI-211, the liposomal form of lurtotecan, in patients ineligible for curative surgery or radiotherapy and with metastatic/locoregional recurrent squamous cell carcinoma of the head and neck (SCCHN) and target lesions either within a previously irradiated field ("within") or outside a previously irradiated field ("outside"). OSI-211 was given intravenously over 30 min on days 1 and 8 at 2.4 mg/m2/day, repeated every 21 days (1 cycle). From July 2001 to March 2002, 32 patients from 14 institutions were enrolled in the "within" arm and 18 in the "outside" arm. In the "within" arm, two patients were ineligible because their tumour site was not allowed in the protocol (nasopharynx, skin) and two other patients never started treatment. Of the 46 eligible patients who started treatment, there was one objective response (response rate: 2.2% (95% Confidence Interval (CI): [0-11.5%]). Twelve patients in the "within" arm and 6 in the "outside" arm had stable disease, with a median duration of 18 weeks, 95% CI (12.7-25.7). The median time to progression was 6 weeks (95%CI: [5.9-12.7] weeks). Haematological toxicity was moderate in both arms. The most common haematological toxicity was grade 1-2 anaemia in 79% of patients. Non-haematological toxicity was mild in both arms. The most common grade 3-4 non-haematological toxicity was infection in 8.5% of patients. OSI-211 administered on d1 and d8, every 3 weeks, is well tolerated, but shows only minimal activity in locally advanced/metastatic SCCHN.

A Bayesian population pharmacokinetics study of data from routine therapeutic drug monitoring cisplatin during a 5-day infusion of cisplatin.A total of 95 kinetics data sets (58 patients) were available to perform this study. Individual pharmacokinetic parameters, estimated from 20 courses of treatment in 18 patients, were used to calculate the population parameters (cl: 0.175 +/- 0.034 L/h; t(1/2): 327 +/- 91 h). The accuracy of Bayesian forecasting was tested by comparing in 40 other patients the clearance values calculated either from a complete kinetics profile (eight sampling times) or from three early samples and the new population parameters. Finally, drug monitoring efficacy was assessed by comparing the target Cmax values with the Cmax obtained after dose adjustment based upon early Bayesian estimation of the individual pharmacokinetic parameters.: No significant difference was found between Bayesian and experimental clearances. Besides, dose-individualization proved to successfully adjust Cmax values around their respective target.The new reference pharmacokinetic population parameters lead to accurate estimation of individual pharmacokinetic parameters from a limited number of samples, thus allowing efficient therapeutic drug monitoring during 5-day infusion regimens of cisplatin.

Increased micronucleated cell rates, dicentric chromosomes, and other chromosomal damages have been reported in lymphocytes of cancer patients prior to the initiation of chemotherapy, and/or radiotherapy. The cause of these chromosomal damages in these lymphocytes remains unclear. In the present work, we investigated whether these micronuclei mainly reflect structural or numerical chromosomal aberrations by applying the cytokinesis-blocked micronucleus (CBMN) assay in combination with fluorescent in situ hybridization (FISH) of a DNA centromeric probe on blood samples of 10 untreated cancer patients (UCPs), and 10 healthy subjects (HSs). Micronucleated binucleated lymphocyte rate was significantly increased in patients (mean+/-S.D.: 19.0 per thousand +/-14.1 versus 9.2 per thousand +/-4.6 in controls). Trinucleated cytokinesis-blocked cells were not significantly higher in patients than in controls. Acentromeric, centromeric, and multicentromeric micronucleus levels were two-fold higher in patients than in controls, but the difference was significant only with acentromeric micronuclei. The percentage of micronuclei containing one or more centromeres averaged 69.2, and 71.5% in patients, and controls, respectively. The percentage of micronuclei containing several centromeres was 44.7% in patients, and 54.6% in controls. Among centromere-positive micronuclei, the percentage of micronuclei containing several centromeres averaged 59.7% in patients, and 75.4% in controls. These results indicate that genetic instability in peripheral blood lymphocytes of UCPs occurs because of enhanced chromosome breakage. However, a substantial proportion of this genetic instability occurs because of defects in chromosome segregation.

Adenocarcinomas of esophagus and cardia represent in France approximately 20 to 40% of the esophagus cancers. They have a high risk to develop lymph nodes metastases and liver metastases. Currently, only 50 to 70% of patients may benefit from surgical curative resection at diagnosis, but more than 50% of them will recur. The standard of treatment of these metastatic adenocarcinomas is chemotherapy. Three large randomized comparative studies, between chemotherapy and supportive care, showed that chemotherapy significantly extends the median of survival (from 3-4 months to 10-12 months) and improves the quality of life. Currently, the combination of epirubicin-cisplatin-continuous 5FU (ECF) is the most effective regimen but it is difficult to administer and tolerate because of the long continuous 5FU infusion. In France, the most commonly used combination regimen still associates 5FU and cisplatin. New drugs (such as docetaxel, CPT11, oxaliplatin) used alone or in combination, especially with 5U, are very promising. Radio-chemotherapy is the preferred treatment for locoregional recurrences, because it improves dyphagia and enables to obtain complete tumor responses. Current results from concomitant radio-chemotherapy studies for esophagus cancer, based on 5FU alone, 5FU-cisplatin or 5FU-mitomycin, given as preoperative treatment or as exclusive treatment, support to use radio-chemotherapy for the treatment of loco-regional recurrences after surgical resection. Nevertheless, the optimal radio-chemotherapy schedule still remain to be defined (dose, duration, splitting of radiotherapy, choice of anticancer drugs).

XR 5000 is one of a series of tricyclic carboxamide-based cytotoxic agents. It binds to DNA by intercalation and stimulates DNA cleavage by inhibition of both topoisomerase I and II, thus possibly overcoming the resistance resulting from downregulation of either enzyme. Twenty patients with advanced or metastatic colorectal cancer, unpretreated for metastatic disease, received XR 5000 at the dose of 3010 mg/m(2) in a 120-h central intravenous (i.v.) infusion every 3 weeks. Response was evaluated every two cycles. No complete (CR) or partial responses (PR) were observed in eligible patients (response rate, 0 of 19, 0%; 95% confidence interval (CI): 0-18%). 5 patients had stable disease, which lasted from 79 to 157 days. Haematological toxicity was low, since only one grade 4 neutropenia and two grade 3 anaemia were observed. Other treatment-related grade 3-4 toxicities were: deep venous thrombosis (2 cases), liver toxicity, diarrhoea, anorexia, dyspnoea, chest pain, infection (1 case each). Despite the good toxicity profile, these results do not support further trials with XR 5000 in metastatic colorectal cancer.

The authors describe an original experience with 3 years of a nutritional workshop for cancer patients. This intervention combine an information about nutritional aspects of cancer with psychosocial support, to buffer psychological and nutritional consequences of cancer. The workshop, leaded by two specialized teams, one in medical oncology, the other in public health, is proposed to patients during and after a specific treatment. In one day, it provided information about nutrition and cancer, diet education and psychosocial support with supportive-expressive group. At this day, the evaluation of this intervention is only subjective. Fifty-six patients participated in at least one workshop, with majority of women (91%). Nineteen workshops were leaded with average participant number of 7 per workshop the third year. The authors believe that nutritional workshops are of great help for cancer patients, by enhancing social reinsertion, giving opportunity of emotional expression and humanizing the treatment. Our experience show it is possible to propose psychosocial intervention in institution in the context of Mediterranean country. We are leading currently a study that will permit a more systematic evaluation of the effects of this intervention.

Anticancer agents go through a process by which their antitumor activity, on the basis of the amount of tumor shrinkage they could generate, has been investigated. In the late 1970s, the International Union Against Cancer and the World Health Organization (WHO) introduced specific criteria for the codification of tumor response evaluation. In 1994, several organizations involved in clinical cancer research joined together to undertake the review of these response evaluation criteria on the basis of their experience and knowledge. After several years of intensive discussions, new guidelines are ready and will replace the previous WHO criteria. In parallel to this initiative, one of the participating groups developed a model by which response rates could be derived from unidimensional measurement of tumor lesions instead of the usual bidimensional approach. This new concept has been largely validated by the Response Evaluation Criteria in Solid Tumors (Recist) Group and integrated into the present guidelines. This special article provides some philosophic background to clarify the various purposes of response evaluation. It proposes a model by which a combined assessment of all existing lesions, characterized by target lesion (to be measured) and nontarget lesions, is used to extrapolate an overall response to treatment. Methods of assessing tumor lesions are better codified. All other aspects of response evaluation have been discussed, reviewed, and amended whenever suitable.

Between March and September 1988, 74 patients with progressive ovarian cancer after prior platinum-based therapy were treated with the luteinizing hormone-releasing hormone (LHRH) agonist Triptorelin (Decapeptyl(R)). Treatment consisted of i.m. injection of 3.75 mg of microencapsulated Triptorelin on days 1, 8 and 28 followed by 4-weekly injections until tumor progression. No objective responses were observed. Eleven out of 68 evaluable patients (16%) had stable disease. The median progression-free survival was 5 months in patients with disease stabilization and 2 months for all evaluable patients. The median survival for patients with disease stabilization was 17 months, whereas for all patients it was 4 months. The treatment was well tolerated; the only reported adverse events were incidental hot flushes. This study showed that the LHRH agonist Triptorelin has only modest efficacy in patients pretreated with platinum-containing chemotherapy. [(C) 2001 Lippincott Williams & Wilkins.].

Paclitaxel is a recent chemotherapeutic agent that inhibits tubulin depolymerization in tumoral cells. Despite its increasing use against various human cancers, the genotoxicity of paclitaxel has never been studied on normal human cells. The in vitro genotoxic effects of the drug were evaluated with two complementary mutagenesis tests on human T-lymphocytes: (1) the cytokinesis-blocked micronuclei assay (CBMN) in combination with fluorescent in situ hybridization (FISH) of nonspecific centromeric probes and (2) the comet assay performed in three ways: on stimulated lymphocytes as in the CBMN, and on freshly isolated lymphocytes at both 4 and 37 degrees C. A slight cytotoxicity of 2.5 to 10 nM paclitaxel was found in the CBMN and a significant increase in the binucleated micronucleated cell rates was observed, with a concentration-dependent manner. In the FISH analysis, more than 85% of the micronuclei (MN) were centromere positive, and a ratio of 72. 2 to 78.6% of these MN contained more than one centromere. Moreover, at 10 nM of paclitaxel, 35.6% of the cells are multimicronucleated lymphocytes. Unexpectedly, paclitaxel induced single-strand breaks on proliferating lymphocytes at 5 and 7.5 nM but not in resting cells, even at 5 to 15 microM. These in vitro results showed that (1) paclitaxel does not present any direct DNA action in resting cells, (2) DNA damage detected in stimulated lymphocytes may be linked either to a high frequency of cells in the S-phase cell cycle or to a direct DNA damaging effect on replicating cells, and (3) paclitaxel is a strong in vitro aneugenic drug on human normal cells, at clinically relevant concentrations.

Maffucci's syndrome is a congenital non-hereditary disease very similar to Ollier's disease and associates multiple cutaneous hemangiomas, dyschondroplasia and often enchondromas.We report a unique case involving synchronous localization of chondrosarcoma in nasal fossae and anterior chest wall, disclosing Maffucci's syndrome.Atypical chondrosarcoma localization must lead to further investigation of potential multiple enchondromatosis.

Doxorubicin and cisplatin are two major anticancer drugs, and are also known to be mutagen. Using short term mutagenesis tests, the cytokinesis-block micronucleus test and chromosome aberrations test, a study of the cytotoxicity and the mutagenicity of these two drugs has been aimed to determine a genotoxic of reference for these tests. Cisplatin and doxorubicin were genotoxic and gave positive results with the two tests. Since cisplatin was more cytotoxic than doxorubicin for a same genotoxicity, doxorubicin has been selected as a positive control for these two short-term mutagenesis tests. A study of the individual variability in the response to in vitro doxorubicin exposure was made using the cytokinesis-block micronucleus test, applied to cultured T lymphocytes form healthy subjects and cancer patients. Micronucleated cell rate before (T0) and after in vitro exposure to doxorubicin (T1) were determined in the two groups of subjects. Micronucleated cell rates T1 were significantly higher than T0 for healthy subjects and cancer patients. A calculated sensitivity index [(T1)-(T0)] is proposed to evaluate the individual sensitivity to the positive control doxorubicin.

The cytokinesis-blocked micronucleus assay (CBMN) is a short-term mutagenesis test which offers an easier and less tedious alternative to metaphase chromosome analysis, with the advantage that exposure to both clastogens and aneugens may be detected. The CBMN assay has been used in evaluating the genotoxic consequences of exposures to environmental and occupational mutagens and carcinogens. Micronucleated cell rates (MN cell rates) were assessed in cytokinesis-blocked lymphocytes of 70 male and female cancer patients prior to any anticancer treatment. The study of interindividual variation factors showed that only age significantly affect MN cell rate, whereas sex, tobacco, alcohol, imaging techniques and tumour stage had no significant effect. The comparison of micronucleated cell rates in 198 healthy subjects and 70 cancer patients matched for age and sex showed a statistically significant difference. Spontaneous elevated MN cell rates of cancer patients refer to previous exposition of genotoxic or mutagenic environmental agents. Moreover, the MN cell rates in cancer patients most probably refers to various cellular lesions and genetic damages.

With an anti-infectious and an antithrombotic prophylaxis aims, a locked flush solution including heparin and vancomycin, was used systematically for implantable venous access system for each patient, from january to april 1995. Since the 6th of april 1995, in order to widen the antibiotic spectrum on Gram negative bacteriae, we added colimycin at the flush solution. In 1995, 342 hospitalised patients held this type of venous access and received chemotherapy and/or radiochemotherapy for cancer. Two thousand six hundred thirty three manipulations were done, 575 with the first flush solution, 2058 with the second. During the year, 15 implantable access system (4.4%) were considered as infected, only 3 (0.9%) were removed, in the first period. THe infectious rate seemed to be stable, but the bacterial assessment to be modified between the two periods. The Gram negative bacterial infections seemed to decrease with colimycin addition (33% versus 50%). These results must be confirmed by a long term and/or randomized study.

A pilot study was conducted to determine whether any relationship exists between micronucleated cell rates of female uterine cervical epithelium and current smoking status. Cervical uterine cells obtained from 118 pre-menopausal females, seen between September 1994 and June 1995 at the Occupational Medecine Interprofessional Association (AIMT), were tested for micronucleated cells by the micronucleus assay. Of the 68 subjects taken from this population that provided a least a 1000 analysed cells, 36 were non-smokers and 32 were smokers. Age distributions were different between smokers and non-smokers but results showed that age had no effect on micronucleated cell levels. Micronucleated cervical cell rates reached 0.66% in non-smokers and 1.11% in smokers: these two levels were not statistically different. No association was noted between number of cigarettes smoked and micronucleated cell levels. Results suggested that consuming 5-20 cigarettes per day was not enough to show a smoking effect on cervical micronucleated cells. A test with a greater number of female subjects would be necessary to confirm this.

We report a case of primary osteogenic sarcoma of the third lumbar vertebra, detailing the neuroradiologic and therapeutic aspects. The clinical presentation was limited to low back pain which radiated to the left thigh for 5 months. Lumbosacral spine roentgenograms revealed a sclerotic lesion of the left part of the body of the third lumbar vertebra. Treatment consisted of total vertebrectomy, chemotherapy completed with radiotherapy. Fourteen months after a complex combined treatment no recurrence was observed. A review of the literature highlighted the rarity of this tumor. Usually, patients with vertebral osteogenic sarcoma do poorly. Today, the therapeutic approach for these spinal tumors should use techniques developed in the treatment of osteosarcoma of the extremities because of their encouraging results.

The rarity of primary osteosarcoma of the spine led us to index the 66 reported cases published in literature. From this analysis a difference was found between spinal osteosarcoma and osteosarcoma of the extremities. Tumors of the spine appear to be two times more frequent in the male population in their thirties. The average period between the beginning of the symptoms and the first consultation is seven months. Back pain is permanent and localized to the affected vertebra. In 80 percent of the cases, neurological symptoms already exist at the stage of the diagnosis. Magnetic resonance imaging (MRI), computed tomography and standard X-ray remain complementary in the morphological analysis of this tumor. All the aspects from the lytic to sclerotic forms are noted, although the lytic form is common. Among spinal osteosarcoma, the lumbar vertebrae are the most frequently affected. Diagnosis can only be established by pathology, even though this may also lead to some errors. In all the reported cases surgery is used, but carcinological methodology is not possible and a complete removal of affected tissue is difficult, with this being achieved in only a quarter of the cases. Radiation therapy, when used, requires doses of 70 Gy to 80 Gy without any certitude of controlling the tumour and with high risks of post-radiation complications. Chemotherapy on its own, despite the use of high-dose methotrexate, only has a temporary effect due to partial action on the primary center. Twenty years ago, only twenty percent of all patients suffering from osteosarcoma lived beyond two years, with worse prognosis for spinal osteogenic sarcoma. Today, the therapeutic approach for spinal tumors uses techniques developed in the treatment of osteosarcoma of the extremities, which can now expect more than seventy percent of all patients to live beyond five years. Present day methods recommend a rapid confirmation of the diagnosis, and then a neoadjuvant chemotherapy followed by surgery to remove all the affected area. This strategy allows an evaluation of the tumor chemosensitivity and to adapt the treatment in consequence. The latest results of this treatment on spinal osteosarcoma appear to be encouraging.

Anticancer drugs dosage are currently adjusted to body surface area. This measure is supposed to be the best morphometric parameter to adjust anticancer drug doses. However, dose adjustment to body surface area has not historically been rigorously demonstrated. We propose a method to objectively test this parameter utility. The statistical justification of drug adjustment to body surface area can use mathematical equations to be expressed. We can demonstrate that body surface area and plasmatic total clearance of a drug should be correlated to adjust dose anticancer drug to body surface area. When we test the hypothesis of body surface area and plasmatic clearance correlation for cytarabine and adriamycin we did not find any significant correlation. For these anticancer drugs, dose adjustment to body surface area increase their pharmacokinetic and efficiency variabilities. The concept of dose-intensity is probably the best justification of individual dose adjustment from plasmatic drug samples, and from pharmacokinetic and pharmacodynamic studies. The determination of the "maximal tolerable exposition" and of the "minimal effective exposition" should reduce the overexpression of toxic risks and avoid the ineffective underexpositions. However, it is difficult to precisely define these two expositions and to research the most relevant pharmacokinetic parameters to their measure. Area under curve appears to be their most appropriate expression.

Since its discovery in 1948 the clinical applications of methotrexate have widened; and in order to overcome resistances to methotrexate, the concept of high-dose methotrexate has been proposed. The use of rescue by folinic acid, as well as rapid dosage of MTX coupled with pharmacokinetic studies, have permitted us to administer an optimum dose of drug, with maximum therapeutic effects, but with reduced toxicity. Individual adaptation of posology, calculated using the test dose or according to population pharmacokinetic with a Bayesian method of parameter estimation (which allows us to adjust the dose of high-dose methotrexate during its infusion) permits control of inter and intra-individual variations of this drug. After analysis of the different methods proposed, we now present the results of 778 courses of treatment by high-dose methotrexate (while separating 238 courses for osteosarcoma as these formed a homogeneous group of patients). Theoretical maximum concentration and length of infusion were decided by physicians, followed by individual adaptation of posology by pharmacologists at the sixth hour of infusion of methotrexate. This treatment unites maximum security for the patient with no serious side effects (no grade 4 toxicity according to WHO classification), while receiving an optimum dose of methotrexate. In courses of MTX for osteosarcoma, the dose of MTX can be further intensified without risk, by administering on average 65% more than the usual dose in adults (8 g/m2) and 10% more than the usual dose in children (12 g/m2).

In spite of the development of new automated methods proposed for scoring of micronuclei, manual analysis of micronucleated cells by light microscopy is still largely employed. The purpose of this study was to define the minimal number of mononucleated and binucleated cells representative of the total number of cells of the side, to determine the division rate. The second aim was to define the minimal number of binucleated cells representative of the total number of the binucleated cells on the same slide, to determine the rate of binucleated micronucleated cells. Total mononucleated, binucleated and polynucleated cells were scored in each slide for determining the 'theoretical division rate'. Among total binucleated lymphocytes observed in each slide, micronucleated cells containing one, two or more micronuclei were scored to determine the 'theoretical micronuclei rate'. Then, 'experimental division rates' were calculated by dividing microscope slides into random areas of 500 lymphocytes, and recording mononucleated, binucleated cells. 'Experimental micronuclei rates' were determined by dividing microscope slides into random areas of 500 binucleated cells, and recording micronucleated cells containing one or more micronuclei. 'Experimental division rates' obtained by examining sets of 4000 cells were shown to be not representative of the corresponding 'theoretical rate' (statistical difference by the chi-squared test, p0.05)). 'Experimental micronuclei rates' obtained by examining sets of 500 binucleated cells were shown to be representative of the corresponding 'theoretical micronuclei rate' (no statistical difference by the chi-squared test, p0.05).

F Duffaud, S Salas, T Huyn, JL DevilleLa Timone University Hospital, Marseille, FranceAbstract: Mutations of the KIT gene are the molecular hallmark of most gastrointestinal stromal tumors (GISTs). GIST has become a model for targeted treatment of solid tumors, imatinib becoming the standard first-line treatment of these tumors in the advanced/metastatic phase. Because of the efficacy of imatinib treatment in the advanced setting, its role following resection of a primary non-metastatic GIST was investigated. The recently published phase III, doubleblind, placebo-controlled, multicenter ACOSOG Z9001 study showed that adjuvant therapy is safe, and significantly improves recurrence-free survival compared to placebo when given after resection. To what extent imatinib will improve overall survival has yet to be answered. What is clear is that high-risk GIST patients definitely need adjuvant therapy, and that 1 year of imatinib is not enough for the patients who do need it. The questions of optimal duration of imatinib treatment in the adjuvant setting, adequate selection of risk patients and effect of imatinib on overall survival are currently being studied.Keywords: GIST, gastrointestinal stromal tumors, imatinib, adjuvant therapy

10580 Background: The role of metastasectomy and front-line chemotherapy modalities remain debated in the metastastatic soft tissue sarcoma setting. Methods: Data from 147 M-LMS patients (1988–2008) reviewed in 6 French centers were retrospectively analyzed. Prognostic factors for progression-free and overall survivals (PFS and OS) were identified using log-rank tests and Cox multivariate analysis. The respective impact of treatment modalities on PFS and OS were assessed after adjustment to prognostic factors. Results: This database included 46 (31%) uterus and 101 soft-tissues (69%) M-LMS. All patients received doxorubicin as front-line regimen associated with ifosfamide in 78 cases (53%) or dacarbazine in 56 cases (38%). The planned dose of doxorubicin was > 60 mg/m2/3 weeks in 24 patients (16%). After front-line chemotherapy, 36 patients with lung metastasis (24%) underwent subsequent complete metastasectomy. The median PFS was 6 months. The univariate analysis identified the following prognostic factors for PFS: performance status (PS), grade, presence of liver or lung metastases. But, the multivariate analysis did not retain independent prognostic factor for PFS. Only one treatment parameter was associated with better PFS: planned doxorubicin dose superior to 60 mg/m2/3 weeks (HR=7.57 [1.32–10.40], p=0.023). The median OS was 14 months (1–115). The univariate analysis identified the following prognostic factors: PS, time interval between diagnosis and metastatic relapse, local relapse and grade. Under multivariate analysis, there was only one good prognostic factor for PFS: interval time between initial diagnosis and metastasis > 12 months (p=0.006). After adjustment to this factor, multivariate analysis shown that complete metastasectomy improved the PFS (HR=0.52, [0.38–0.87], p=0.012) and addition of ifosfamide was associated with worst outcome (HR=1.42 [1.05–2.10], p=0.028). Other chemotherapy parameters did not significantly modify OS. Conclusions: Doxorubicin dose and metastasectomy remain the cornerstone of the optimal treatment of M-LMS. Addition of Ifosfamide seemed to be associated with worst outcome. Dacarbazine seemed to have no significant impact. [Table: see text]

10016 Background: IM the first-line treatment for advanced GIST, must be given continuously until disease progression (PD) or intolerance (Blay et al, ASCO 2004). The impact on long term overall survival (OS) of IM discontinuation in responding patients (pts) and its re-introduction at progression is unknown. Methods: This prospective multicenter BFR14 study was initiated in June 2002. After 1 year of IM 400mg/day, 58 pts free from progression were randomly offered to continue or interrupt IM until PD. Pts allocated to the interruption (I) arm could restart IM (same dose) in case of PD. Primary endpoint was progression-free survival (PFS); secondary endpoints were OS, secondary response after IM re-introduction, identification of molecular determinants of response. Survival data were compared using the log-rank test. Results: Pt characteristics were well balanced between the two arms. Current median follow- up after randomization is 37 months (range 0–42): 29/32 pts (91%) in arm I vs. 16/26 pts (62%) in continuous (C) arm experienced PD. IM interruption was significantly associated with reduced PFS (p-4) with a median of 6 months (95% CI, 3.5–9.5) for arm I. Among the 26 pts of I arm who restarted IM after first progression, 24 (92%) achieved tumor control (OR or SD), one progressed, one died from an unrelated cause. The 2-yr OS rates were 87% and 92% for arms I and C, respectively (P = 0.78). Conclusions: Despite a significant increase of PD in the experimental arm, IM re-introduction is safe and allows a tumor control in the majority of pts. No unfavourable impact of IM interruption on OS was reported. Treatment interruption could be a therapeutic option in advanced GIST pts exhibiting intolerance to IM. Updated results including mutational analysis will be presented at the meeting. No significant financial relationships to disclose.