Des micrométastases ou cellules isolées tumorales peuvent être détectées en analyse anatomopathologique avec éventuellement l'aide de coloration par immuno-histochimie au niveau des ganglions, de la moelle osseuse et du sang. Ces atteintes de toute petite taille ont-elles une réalité clinique, c'est-à-dire un impact pronostique et en conséquence un impact sur les décisions thérapeutiques ? L'impact de la détection de micrométastase des ganglions sentinelles est discuté avec deux éléments : 1) le taux de ganglions non sentinelles (GNS) envahis au curage axillaire complémentaire est-il différent de celui des GS pN0 et de celui des GS pN1a ? 2) les survies sont-elles significativement différentes en fonction de cette atteinte des GS ? Le risque d'atteinte des GNS en cas de micrométastase ou de cellules isolées est différent de celui des cas où les GS sont indemnes (risque de 7 à 8 %) et de celui des cas où les GS sont envahis par une macrométastase (risque de 30 à 50 %). La présence de micrométastase a un impact pour les cancers du sein sur le risque d'envahissement d'autres ganglions du creux axillaire, possiblement sur la survie globale ou sans récidive ainsi que sur l'indication de traitement adjuvant de chimiothérapie et/ou de radiothérapie (sur les champs de traitement des aires ganglionnaires régionales). Il s'agit donc bien d'une réalité clinique, même si la question du bénéfice du curage axillaire complémentaire reste posée.Micrometastases or sub-micrometastases can be detected by standard histopathological method sometimes associated with immunohistochemistry in lymph nodes, bone marrow and blood. The consequence of these small size involvement may be prognostic and therapeutic. Two factors are necessary to assess this kind of involvement: the rate of involvement of non-sentinel lymph node after axillary lymph node dissection and significative difference of survivals. The rate of involvement of non-sentinel lymph node in case of micrometastases or sub-micrometastases is different from the rate of involvement in case of no lymph node metastases (7 to 8%) or in case of macrometases (30 to 50%). Micrometastase is an important factor to determine the rate of involvement of non-sentinel lymph node, the overall or disease free survival and to assess the need of radiotherapy and chemotherapy. In conclusion, micrometastases and sub-micrometastases have a clinical impact even if complementary axillary lymph node dissection is still discussed.

BACKGROUND:Few population-based studies have reported jointly analyses of relative survival according to the following prognostic factors: tumour-node-metastasis (TNM) stage, age, number of examined and positive nodes, hormonal status, histological Scarff, Bloom and Richardson (SBR) grade, tumour extension, hormone receptor status and tumour multifocal status.PATIENTS AND METHODS:Data on female invasive breast cancer were provided by the Cote d'Or breast cancer registry. The Kaplan-Meier method and log-rank test were used to estimate and compare the survival probability at 1, 5, 10 and 15 years. The effect of prognostic factors on survival was assessed with crude and relative multivariate survival analyses.RESULTS:Crude survival seemed to be worse in patients aged >60 years compared with those aged 45-60 (P > 0.0001), whereas relative survival did not differ. TNM stage, histological SBR grade, progesterone receptor status, tumour multifocal status, locoregional extension and the period of diagnosis were independent prognostic factors of crude and relative survival.CONCLUSION:Breast cancer is influenced by many factors. Despite the absence of any association between the number of examined nodes and overall survival in this study, the number of nodes removed, in conjunction with other prognostic factors, may be useful in selecting node-negative patients for systemic therapy.

Survival data on female invasive breast cancer with 9-year follow-up from five French cancer registries were analysed by logistic regression for prognostic factors of cancer stage. The Kaplan–Meier method and log-rank test were used to estimate and compare the overall survival probability at 5 and 7 years, and at the endpoint. The Cox regression model was used for multivariate analysis. County of residence, age group, occupational status, mammographic surveillance, gynaecological prevention consultations and the diagnosis mammography, whether within a screening framework or not, were independent prognostic factors of survival. Moreover, for the same age group, and only for cancers T2 and/or N+ (whether 1, 2 or 3) and M0, the prognosis was significantly better when the diagnosis mammography was done within the framework of screening. Socio-economic and surveillance characteristics are independent prognostic factors of both breast cancer stage at diagnosis and of survival. Screening mammography is an independent prognostic factor of survival.

BACKGROUND. Breast carcinoma survival rates were found to be higher in the U.S. than in Europe.METHODS. Multiple regression analysis of breast carcinoma survival rates among women diagnosed between 1990 and 1992 was performed using clinical data from population-based case series from the Surveillance, Epidemiogy, and End Results (SEER) program (13,172 women) and the European Concerted Action on survival and Care of Cancer Patients (EUROCARE) project (4478 women).RESULTS. Early-stage tumors (T1NOMO) were more frequent in the SEER data (41% of cases) than in the EUROCARE data (29%). In the SEER data, early tumors were more frequent in women age greater than or equal to 65 years (43%) than in younger women (38%), whereas the reverse was true in the European data (25% vs. 31%). In both case series, > 90% of women underwent surgery and 81-82% underwent lymphadenectomy, but the number of axillary lymph nodes evaluated was higher in the SEER data than in the EUROCARE data. The 5-year survival rate was higher in the U.S. case series (89%) than in the European series (79%). This differential was observed for each stage category evaluated: early (T1N0M0), large lymph node-negative (T2-3N0M0), lymph node-positive (T1-3N+M0), locally advanced (T4M0), and metastatic (M1) tumors. The overall relative excess risk (RER) of death was significantly higher (RER, 1.37; 95% confidence interval [95% CI], 1.25-1.50) among European women compared with U.S. women (referent group). Adjustment for stage, age, surgery, and the number of lymph nodes evaluated explained most of the excess risk (RER, 1.07; 95% CI, 0.98-1.17).CONCLUSIONS. Transatlantic differences in the 5-year survival rates for women diagnosed with breast carcinoma between 1990 and 1992 were attributable mainly to differences in stage of disease. Resources should be invested to achieve earlier diagnosis of breast carcinoma in Europe, especially for elderly women. (C) 2003 American Cancer Society.

We analysed the 5-year relative survival among 4473 breast cancer cases diagnosed in 1990–1992 from cancer registries in Estonia, France, Italy, Spain, the Netherlands and the UK. Among eight categories based on ICD-O codes (infiltrating ductal carcinoma, lobular plus mixed carcinoma, comedocarcinoma, ‘special types’, medullary carcinoma, not otherwise specified (NOS) carcinoma, other carcinoma and cancer without microscopic confirmation), the 5-year relative survival ranged from 66% (95% CI 61–71) for NOS carcinoma to 95% (95% CI 90–100) for special types (tubular, apocrine, cribriform, papillary, mucinous and signet ring cell); 27% (95% CI 18–36) for cases without microscopic confirmation. Differences in 5-year relative survival by tumor morphology and hormone receptor status were modelled using a multiple regression approach based on generalised linear models. Morphology and hormone receptor status were confirmed as significant survival predictors in this population-based study, even after adjusting for age and stage at diagnosis.

The aim of our study was to evaluate and compare in thyroid cancer patients the predictive value for disease progression of thyroglobulin (Tg) levels measured under thyroid-stimulating hormone (TSH) stimulation, in the postoperative period just before I-131 ablative therapy and at the time of control 6-12 mo later. Methods: Two-hundred twelve consecutive patients treated for a well-differentiated thyroid carcinoma (184 papillary, 28 follicular) with no initial distant metastases were retrospectively studied. All patients had a total or near-total thyroidectomy followed by ablation with 3.7 GBq I-131. Tg levels were determined just before ablative therapy (Tg1) and 6-12 mo later (Tg2). Thresholds of 30 and 10 ng/mL were used for Tg1 and Tg2, respectively. Univariate and multivariate analyses were performed to assess the predictive value for disease progression of the 2 Tg determinations. Results: Thirty patients had a Tg1 level > 30 ng/mL. Six to 12 mo later, 30 patients had a Tg2 level > 10 ng/mL, 19 of whom had initially a Tg1 level > 30 ng/mL. Disease progression was reported in 20 patients (9%). Progression-free survival rates were significantly lower in patients with a low Tg1 or Tg2 level but the difference was more important with Tg2. With univariate analysis, 5 variables were significantly associated with disease progression: Tg2, Tg1, node invasion, extrathyroidal extension, and tumor size. With multivariate analysis, only Tg2 (odds ratio [OR] = 16.4; 95% confidence interval [95% Cl] = 5.7-47.4; P < 0.001) and node invasion (OR = 2.7; 95% Cl = 1.0-7.2; P = 0.04) had an independent prognostic value. When only initial parameters were considered, Tg1 and node invasion were the 2 independent prognostic factors. The OR decreased for Tg1 (OR = 10.1; 95% Cl = 4.0-25.7; P < 0.001) but increased for node invasion (OR = 4.4; 95% Cl = 1.7-11.2; P = 0.002). Conclusion: Among all clinical and tumoral variables, lymph node invasion and serum Tg level are 2 important parameters to define the risk of disease progression. Although Tg2 appears more significant than Tg1, both Tg levels measured under TSH stimulation, in the postoperative period and a few months after ablative therapy, have a predictive value. In clinical practice, patients at risk can be selected as soon as the initial lymph node status and Tg1 level are known.

Purpose: To quantify the risk of acute leukemia after adjuvant therapy, especially chemotherapy with topoisomerase II inhibitors.Patients and Methods: We performed a population-based study in a cohort of 3,093 women younger than 85 years who resided in the French administrative area of the Cote d'Or, who were given a first diagnosis of primary breast cancer between 1982 and 1996, and who received a curative treatment. Information about therapy and follow-up events war obtained from records of cancer registries their covered this area.Results: Until December 1998, 10 cases of acute leukemia, including nonlymphoid acute leukemia and refractory anemia with excess of blasts, occurred in patients before any local or distant recurrence, All cases developed in the first 4 years of follow-up. Compared with the general female population, the incidence rate of leukemia was significantly increased in women who received radiotherapy and chemotherapy (standardized incidence ratio, 28.5; P < .0001). A dose-dependent increase in the risk of leukemia was observed in women treated with mitoxantrone. Cox regression analysis showed that the risk of leukemia was significantly lower in patients treated with anthracyclines than in those treated with mitoxantrone at cumulative doses greater than or equal to 13 mg/m(2).Conclusion: The combination of adjuvant radiotherapy and chemotherapy with mitoxantrone induces a high risk of acute leukemia in patients with breast cancer. A leukemogenic effect of chemotherapy with anthracyclines cannot be ruled out with certainty. However, there are some suggestions that these topoisomerase II inhibitors might be lass leukemogenic than mitoxantrone and could be preferred in an adjuvant setting. J Clin Oncol 18:2836-2842. (C) 2000 by American Society of Clinical Oncology.

Our objectives were to evaluate the effectiveness of cervical cancer screening outside organized programs in the prevention of cervical carcinoma in situ (CIS) and to enhance the way in which case control studies avoid some common biases. In our case-control study, we assessed all incident, histologically verified cases of CIS registered from 1987 to 1997 in the population-based cancer registry of C te-d'Or, France (N = 104) and 208 controls randomly selected from the screened population and matched for age, date of last screening, residence, and pathology laboratory results. We considered as appropriate for controls screened women who had had at least one Papanicolaou smear in the 3 years preceding the diagnosis or similar period. Screening for controls was higher (67.8%) than for cases (41.4%; P < .001), with a relative protection against CIS of 3.09 (95% confidence interval, 1.83-5.22) and a prevented fraction in the screened population of 45% to 50%. These findings suggest a protective advantage for CIS even in the absence of organized screening. The methodologic approach has advantages as compared to previous types of case-control studies. Although further refinements still are warranted, learning about the protective effect of screening for CIS provides information that may be useful in assessing the impact of a screening policy on women actually at risk of invasive cervical cancer.

Purpose: The aim of the study was to determine the predictive factors of complications, to evaluate the impact of customized treatment planning on late normal tissue effects per stage, and to report disease-free survival (DFS) and local control (LC) rates.Methods and Materials: From 1970 to 1994, 642 patients were treated with radiotherapy alone for carcinoma of the intact uterine cervix. According to the International Federation of Gynecology and Obstetrics (FIGO) substaging, 34% were Stage I, 39% Stage II, and 27% Stage III. The analysis was divided into three periods: 1970-1978 (use of standard prescriptions),1979-1984 (implementation of individual adjustments), 1985-1994 (systematic individual adjustments). Five-year DFS, LC, and complications rates were calculated using the Kaplan-Meier method. Predictive factors of complications were determined by univariate analysis using frequency tables and nonparametric t-tests. Multivariate analysis consisted of a polychotomous stepwise regression.Results: The comparison of the three time periods showed a significant reduction of the external radiation dose (dose above 40 Gy in 47% of patients before 1979 vs. 36% after 1984), of the use of parametrial boost (55 % vs. 39%), of the use of vaginal cylinder (28% vs. 11.5%), and of the HWT volume (combined intracavitary and external irradiation) (842 cc vs. 503 cc on average). The total sequelae/complications rate, all toxicity grades, all stages, all organs was 51%. Five-year actuarial rate per toxicity grade was: G1, 42%; G2, 23.5%; G3, 10%; G4, 3%. The three main predictive factors for rectal and bladder sequelae/complications (all toxicity grades) taking into account time period were: the increase of external radiation dose, the high dose rate at reference points, and the whole vagina brachytherapy. No G4 occurred in the third period. The rate of G3 complications dropped from 16% to 6% over time: from 5% during the first period to 0% during the third period in Stage I, from 8% to 6% in Stage II, and from 23% to 12% in Stage III. G3 currently describes a variety of clinical situations with a different impact on quality of life which justifies further refinements of definitions of late effects. In our experience the severity of G3 markedly decreased: leas than one-third of G3 had a real impact on quality of life in the last period compared to more than two-thirds in the first period. Meanwhile, 5-year LC rates remained stable in Stages I and II, 91% and 85% respectively. Conversely they fell from 75% to 55% in Stage III, thus raising the problem of underdosage and/or more accurate staging with time.Conclusions: Customized treatment planning eradicated lethal complications and provided a significant decrease of G3 in all stages while maintaining high cure rates in early stages. Dose reduction should be considered with caution in Stage III. (C) 2000 Elsevier Science Inc.

Objectives: To conduct a survey of the angiosarcomas developing after breast conservation for carcinoma in the French Cancer Centers, to study the evolution of these cases in detail, and to review literature in an attempt to propose an optimal treatment scheme.

Using the 2,208 non metastatic breast cancer diagnosed women who underwent breast surgery from the 2,432 first breast cancers recorded by the French département of Côte-d'Or cancer registry from 1982 to 1992, we described in this well-defined population, the trend in breast-cancer adjuvant treatment, and related practices with recommendations according to risk groups. Adjuvant treatment was received by 44.1% of the 2,208 women. Inflammatory tumors were systematically treated with chemotherapy. For the non-inflammatory M0 breast-cancer (2,167 women), the adjuvant treatment probability was mainly determined by the nodal involvement. After adjustment on the tumor- and host-characteristics, time period was associated with increased probability of adjuvant treatment in the whole group of patients and within each pN subgroup. In the group without nodal involvement, this increase was also associated with the SBR histologic grade. There was strong evidence of large changes in breast-cancer adjuvant treatment. Clinical practices in the Côte-d'Or region have paralleled the NIH recommendations. For node-negative breast cancers, these trends appeared despite persistent uncertainty in the definition of subgroups to treat. Over time, the SBR histologic grade became an apparent factor of treatment. This use as a treatment indicator was done without validation in any adjuvant treatment trial. The simultaneous recommendation to treat with the absence of guidelines could lead to the treating of an increasingly large group by extending the "high-risk" definition. This could be a non-optimal management of risk while putting strain on health care resources.

STUDY OBJECTIVE: In many countries, cancer registries cover only a small part of the national population. Cancer incidence for the rest of the country has therefore to be estimated. This can be done from mortality data using the relation between incidence and mortality observed in the cancer registry areas. Such an approach was used to study geographical variation and trend of colorectal and breast cancer incidence in France where 10% of the national population is covered by cancer registries. DESIGN: This study applies the incidence/mortality ratios of cancer registry areas to regional mortality data to obtain an estimation of cancer incidence at a given point in time. Age and period effects are included in the statistical models. MAIN RESULTS: The incidence estimations are given for 21 administrative regions and three time points (1985, 1990, 1995). The European standardised incidence rates for breast cancer ranged from 86.8 to 128.8. For colorectal cancer, these rates ranged from 48.2 to 79.6 for men, and from 32.5 to 48.8 for women. Breast cancer incidence has increased considerably between 1985 and 1995 with a higher increase in the north than in the south of France. The incidence of colorectal cancer has also increased, albeit to a lesser extent. CONCLUSION: The incidence estimation method proposed leads to regional incidence rates that are useful for planning health care services on a regional basis and may also be used to study regional differences in incidence. This method is useful when only partial incidence data are available.

PURPOSE:The aim of this study was to describe the implication of the different health care structures in the treatment of breast cancer.METHODS:In Côte-d'Or, from 1982 to 1992, there were 2432 cases of breast cancer. Surgery came first as treatment for 93% of the patients, radiotherapy came second (77%). The department is subdivided in several geographic areas (ZPIU):--Dijon, equipped with university hospital (UH) and with private hospitals (PH),--cities with general hospitals (GH)--and areas without hospitals. Demographic, geographic and clinical variables were studied in order to explain the patient distribution between the various hospitals.RESULTS:52% of the cases were operated in PH, 37% in UH and 11% in GH. The main users of the GH were women who lived nearby. Age over 75 was associated with a treatment in GH. Women with clinical signs of severity were twice as often operated in UH rather than PH. Post-operative radiotherapy was done in 95% of the cases in the same structure where surgery was done.CONCLUSION:No matter how popular university and private hospitals were in our regional capital, general hospitals played a proximity role.

From 1970 to 1994, 642 patients with carcinomas of intact uterine cervix were treated with radiotherapy alone Univariate and multivariate analysis was carried out of predictive factors for: 1) pelvic control and survival rates; 2) complications using French-Italian Syllabus, combined with an evaluation of the impact of customized treatment planning policy (CTP) on G3-G4. According to Figo substaging 30% of patients were stage I, 42% stage II and 28% stage III/IV. Diameter of cervical disease was 3-5 cm in 38% of cases and > 5 cm in 15%. Nodal involvement from lymphangiogram was 21%. The distribution of sequelae and complications was: G1 23%, G2 18%, G3 6%, G4 2.5%. The distribution of G3-G4 per organ was: genitalia 6% (no G4), rectum 4%, colon 1.5%, bladder 1.2%, soft tissues 1%, small bowel 0.5%. Stage (RR ranging from 1.5 for stage IIb to 5 for stage III/IV), tumor size (RR = 1.5), nodal involvement (RR = 2) were significant predictive factors for survival and pelvic control rates (p < 0.0001). In univariate analysis the main factors influencing the risk of G3-G4 complications were: Figo substaging, external radiation dose over 40 Gy (ED), parametrium boost (PB), use of brachytherapy vaginal cylinders applicator (CA), high HWT and mean rectal dose rate for rectal complications. In multivariate analysis, CA remained the only predictive factor for G3-G4 bladder events (odds ratio OR = 10.8) while the increase of mean dose rate (OR = 1.1), use of CA (OR = 4.2) and ED > 40 Gy (OR = 4.4) were predictive of severe rectal sequelae. Prevention of complications based upon individual changes of treatment planning according to dosimetry parameters led to a sharp decrease in severe complications with time. No G4 occurred after 1983. G3 rates dropped from 5% before 1978 to 0% after 1983 in stage I, from 10% to 6% in stage II and from 23% to 12% in stages III/IV. Meanwhile 5-year LC rates remained stable in early stages, about 91% in stage I and 85% in stage II, conversely they fell from 75% to 55% in stages III/IV, thus raising the problem of underdosage and/or more reliable staging with time. It is concluded that radiotherapy prescriptions based upon tumor diameter per stage and delivered using CTP led to an eradication of lethal complications and provided significant decrease of G3 in all cases while maintaining high cure rates in early stages. Dose reduction should be considered with caution in stages III/IV.

L’estimation de l’incidence des cancers pour la France metropolitaine est l’un des objectifs du Reseau francais des registres de cancers (Francim). Cette etude presente les estimations etablies pour les cancers genitaux et mammaires feminins a partir des donnees observees dans huit departements francais. L’observation a porte sur la periode 1983-1987 et 9,1 % de la population francaise. La qualite de l’enregistrement effectue par les huit registres inclus (Calvados, Cote-d’Or, Doubs, Herault, Isere, Bas-Rhin, Somme, Tarn) a ete evaluee par le Comite national des registres. Les taux departementaux d’incidence, bruts et standardises selon la population mondiale, ont ete etablis pour chaque localisation tumorale. Le taux national d’incidence a ete estime en modelisant les taux departementaux specifiques pour l’âge comme une fonction des taux de mortalite correspondants. Le nombre estime de nouveaux cas par an de cancer du sein etait de 25 277, celui des cancers genitaux pelviens etait de 13 856. Le risque de cancer, estime par le taux cumule 0-74 ans, etait de 7,1 % pour le sein, de 1,2 % pour le col de l’uterus, de 1,4 % pour le corps de l’uterus et de 1,1 % pour l’ovaire. Ces cancers genitaux et mammaires representaient 49 % de l’ensemble des cancers des femmes. L’incidence observee du cancer du sein etait stable d’un departement a l’autre. D’importantes variations interdepartementales d’incidence du cancer du col de l’uterus etaient observees et certains departements francais presentaient des taux voisins des maxima europeens. Les variations interdepartementales d’incidence des autres cancers genitaux pelviens etaient faibles. Les estimations fournies permettent de mesurer l’importance des cancers genitaux et mammaires feminins dans la sante publique en France. L’objectif du reseau Francim est de decrire l’evolution du risque de cancer sur une longue periode et de construire des projections fiables pour les annees 2000 et 2005 en combinant les presentes donnees et celles de la periode 1988-1992 actuellement en cours d’exploitation.

The aim of the study was to assess the incident number of female breast and genital tract cancers for the whole of France. The study focused on the 1983-1987 period and on 9.1% of the French population. The incident number of female breast and genital tract cancers was estimated for each site and for each of eight French administrative regions covered by a cancer registry qualified through the National Committee of Registries (Calvados, Côte-d'Or, Doubs, Hérault, Isère, Bas-Rhin, Somme, Tarn). Information on mortality rates was available at a regional level as well as at a nationwide level. The method estimated the national incidence rate modelizing the regional age-specific incidence rate as a function of corresponding mortality rate. Breast cancer was the leading site with 25,277 new cases per year while female genital tract cancers affected about 13,856 women. The cancer risk, estimated in using cumulative rate 0-74 years, was assessed at 7.1% for breast and at 1.2%, 1.4% and 1.1% for cervix uteri, corpus uteri and ovary respectively. Breast and genital tract cancers constituted 49% of the whole of cancers in women. Observed breast incidence rate was stable through French regions. There were pronounced contrasts in cervix uteri cancer risk, and some french regions displayed a high risk close to the observed European maxima. Interregional contrasts in risk of the other genital tract cancers were less striking. This study emphasized the importance of female breast and genital tract cancers for public health in France. The main aim of the French Cancer Registries Network is to provide a comprehensive description of cancer risk in France and to produce pertinent projection to 2005 horizon, combining the present data and the already accumulated 1988-1992 data.

TP53 abnormalities have been reported as an early event in the process of cellular transformation of human breast cancers, and involved in mammary-tumor evolution, from in situ to invasive disease. In this study, node-negative (N-) tumors were examined for TP53 allelic loss in relation to different genetic instability events, including allelic loss at chromosome 17p13.3 and c-H-ras-1 loci, as well as alteration of the c-myc and c-erbB-2/neu oncogenes. TP53 allelic loss was analyzed to determine whether such an abnormality was the more important, among other genetic events, in the N- tumors, whether it appeared independently of these genetic events, and whether accumulation of genetic events arises in this group of breast tumors. Clinicopathological parameters were also examined. Loss of heterozygosity (LOH) at the TP53 gene appears the most frequent alteration detected (26% vs. 13%, 8%, 9% and 3% for LOH at D17S30 and c-H-ras-1 loci, and amplification of c-myc and c-erbB-2/neu respectively). There was no association between LOH at the TP53 locus and other genetic events. Among clinicopathological parameters, significant associations were observed only with estrogen-receptor-negative tumors (p = 0.05). Our results demonstrate that LOH at TP53 arises more frequently in the N- breast cancer, thus supporting earlier findings suggesting that TP53 abnormality has a role early in the pathogenesis of breast lesions. Moreover, the data indicate that accumulation of many genetic events occurs at a low level in N- breast tumors, and that TP53 abnormality occurs independently of these genetic events.

From 1981 to 1985, 428 patients presenting with an epidermoid carcinoma of the hypopharynx and/or larynx were treated with a curative intent by surgical resection and postoperative irradiation. Two-thirds of the tumours were T3 and 60% of patients presented with a clinical node involvement. The rates of local failure were 8%, 18% and 13%, respectively, for cancers of the larynx, of the piriform sinus and of the posterior wall; the rates of regional failure were 8%, 23% and 13%, respectively. There is no head and neck site with either a high or low risk of recurrence after resection, but the capsular rupture remains a factor of poor prognosis. The survival rate at 5 years of the whole series is 38%, for laryngeal localisation it reaches 62%. The risk of metastases is related to the node involvement and the interval between surgery and irradiation.