ABSTRACT
BACKGROUND: Linking independent sources of data describing the same individuals enable innovative epidemiological and health studies but require a robust record linkage approach. We describe a hybrid record linkage process to link databases from two independent ongoing French national studies, GEMO (Genetic Modifiers of BRCA1 and BRCA2), which focuses on the identification of genetic factors modifying cancer risk of BRCA1 and BRCA2 mutation carriers, and GENEPSO (prospective cohort of BRCAx mutation carriers), which focuses on environmental and lifestyle risk factors. METHODS: To identify as many as possible of the individuals participating in the two studies but not registered by a shared identifier, we combined probabilistic record linkage (PRL) and supervised machine learning (ML). This approach (named "PRL + ML") combined together the candidate matches identified by both approaches. We built the ML model using the gold standard on a first version of the two databases as a training dataset. This gold standard was obtained from PRL-derived matches verified by an exhaustive manual review. Results The Random Forest (RF) algorithm showed a highest recall (0.985) among six widely used ML algorithms: RF, Bagged trees, AdaBoost, Support Vector Machine, Neural Network. Therefore, RF was selected to build the ML model since our goal was to identify the maximum number of true matches. Our combined linkage PRL + ML showed a higher recall (range 0.988-0.992) than either PRL (range 0.916-0.991) or ML (0.981) alone. It identified 1995 individuals participating in both GEMO (6375 participants) and GENEPSO (4925 participants). CONCLUSIONS: Our hybrid linkage process represents an efficient tool for linking GEMO and GENEPSO. It may be generalizable to other epidemiological studies involving other databases and registries.
Subject(s)
Breast Neoplasms , BRCA1 Protein/genetics , BRCA2 Protein/genetics , Cohort Studies , Databases, Factual , Female , Genetic Predisposition to Disease , Humans , Mutation , Prospective Studies , RiskABSTRACT
PURPOSE: Among metastatic breast cancer (MBC) patients, those with a triple-negative breast cancer phenotype (mTNBC) have the worst prognosis, but the benefit of chemotherapy beyond second line on outcome remains uncertain. The purpose of this study was to identify predictive factors of outcome after third- or fourth-line chemotherapy. METHODS: The ESME-MBC database is a French prospective real-life cohort with homogeneous data collection, including patients who initiated first-line treatment for MBC (2008-2016) in 18 cancer centers. After selection of mTNBC cases, we searched for independent predictive factors (Cox proportional-hazards regression models) for overall survival (OS) on third- and fourth-line chemotherapy (OS3, OS4). We built prognostic nomograms based on the main prognostic factors identified. RESULTS: Of the 22,266 MBC cases in the ESME cohort, 2903 were mTNBC, 1074 (37%) and 598 (20%) of which had received at least 3 or 4 lines of chemotherapy. PFS after first- and second-line chemotherapy (PFS1, PFS2) and number of metastatic sites ≥3 at baseline were identified by multivariate analysis as prognostic factors for both OS3 (HR = 0.76 95%CI[0.66-0.88], HR = 0.55 95%CI[0.46-0.65], HR = 1.36 95%CI[1.14-1.62], respectively), and OS4 (HR = 0.76 95%CI[0.63-0.91], HR = 0.56 95%CI[0.45-0.7], HR = 1.37 95%CI[1.07-1.74]), respectively. In addition, metastasis-free interval was identified as a prognostic factor for OS3 (p = 0.01), while PFS3 influenced OS4 (HR = 0.75 95%CI[0.57-0.98]). Nomograms predicting OS3 and OS4 achieved a C-index of 0.62 and 0.61, respectively. CONCLUSION: The duration of each previous PFS is a major prognostic factor for OS in mTNBC patients receiving third- or fourth-line chemotherapy. The clinical utility of nomograms including this information was not demonstrated.
Subject(s)
Antineoplastic Combined Chemotherapy Protocols/therapeutic use , Triple Negative Breast Neoplasms/drug therapy , Adult , Aged , Aged, 80 and over , Breast Neoplasms/drug therapy , Cohort Studies , Disease-Free Survival , Female , France , Humans , Middle Aged , Prognosis , Prospective Studies , Treatment Outcome , Triple Negative Breast Neoplasms/pathologyABSTRACT
BACKGROUND: For luminal metastatic breast cancer (MBC), endocrine therapy (ET) is the recommended initial treatment before chemotherapy. Our objective was to evaluate the efficacy of multiple ET lines in a real-life study. METHODS: The Breast Cancer Epidemiological Strategy and Medical Economics (ESME) project analysed data from all patients with systemic treatment for MBC initiated between 2008 and 2014 in one of the 18 French Comprehensive Cancer Centres. The primary end-point was the successive progression-free survival (PFS) evaluation. RESULTS: The ESME research programme included 9921 patients with hormone receptor-positive (HR+)/human epidermal growth factor receptor 2 (HER2) negative (HER2-) MBC. Before any chemotherapy, 4195 (43.4%), 1252 (29.8%) and 279 (6.6%) patients received one, two or three ET ± targeted therapy, respectively. The median PFS for first-, second- and third-line ET ± targeted therapy was 11.5 (95% confidence interval [CI], 10.8-12.1), 5.8 (95% CI, 5.3-6.1) and 5.5 (95% CI, 4.6-6.3) months, respectively. In a multivariate analysis, time from diagnosis to metastatic recurrence (P < 0.0001), presence of symptoms at metastatic relapse (P = 0.01), number of metastatic sites (P = 0.0003) and their localisation (P < 0.0001) were prognostic factors for PFS1. Duration of previous PFS was the only prognostic factor for subsequent PFS (10% threshold). Ten percent of the patients showed long-term response to ET, with a total treatment duration before chemotherapy ≥43.6 months. CONCLUSIONS: Median PFS in our HR+/HER2- real-life cohort is similar to median first-line PFS reported in clinical trials, regardless of ET used as second- and third-line treatment. Despite the international consensus on early initiation of ET, the latter is not prescribed in most of the cases. Patients with a low tumour burden may achieve prolonged response on ET.
Subject(s)
Antineoplastic Agents, Hormonal/therapeutic use , Biomarkers, Tumor/analysis , Breast Neoplasms/drug therapy , Receptor, ErbB-2/analysis , Receptors, Estrogen/analysis , Receptors, Progesterone/analysis , Adult , Aged , Aged, 80 and over , Antineoplastic Agents, Hormonal/adverse effects , Breast Neoplasms/enzymology , Breast Neoplasms/mortality , Breast Neoplasms/pathology , Disease Progression , Female , France , Humans , Middle Aged , Neoplasm Metastasis , Progression-Free Survival , Retrospective Studies , Risk Assessment , Risk Factors , Time Factors , Young AdultABSTRACT
BACKGROUND: Long-term survivors of non-Hodgkin lymphoma (NHL) must cope with treatment complications and late toxicities that affect their health-related quality of life. Little is known about the risk-to-benefit ratio of new agents like rituximab. The impact of treatment regimens and health disorders on long-term fatigue levels was investigated in a cross-sectional study. METHODS: Two self-administered questionnaires, the 20-item Multidimensional Fatigue Inventory (MFI-20) and a Life Situation Questionnaire, were mailed in 2015 to NHL survivors enrolled onto 12 successive clinical studies (1993-2010) conducted by the Lymphoma Study Association. Private addresses were obtained for 3317 survivors, of whom 1671 (50%) returned the questionnaires. Severe fatigue was defined as MFI-20 scores ≥60 on dimension scales scored from 0 to 100. Linear regression models were used to assess factors that were linked to increased fatigue levels. RESULTS: The study population included 906 men and 765 women, and the median age was 64 years (age range, 24-95 years). Overall, 811 survivors had received cyclophosphamide, doxorubicin, vincristine, prednisone (CHOP)-like chemotherapy, 518 had received high-dose CHOP, and 342 had undergone upfront autologous stem cell transplantation; 829 survivors also had received rituximab. In total, 1100 survivors (66%) reported 1 or more late health disorders. Severe fatigue was reported by 602 survivors (37%). Increased fatigue levels were associated (P < .001) with increased age, obesity, and the presence of health disorders, but not with initial treatment or rituximab. CONCLUSIONS: The survey confirms that high proportions long-term NHL survivors have severe fatigue. The results suggest that initial treatment and the receipt of rituximab have no influence on the development of long-term fatigue.
Subject(s)
Antineoplastic Combined Chemotherapy Protocols/adverse effects , Cancer Survivors/statistics & numerical data , Fatigue/etiology , Hematopoietic Stem Cell Transplantation/adverse effects , Lymphoma, Non-Hodgkin/therapy , Quality of Life , Adolescent , Adult , Aged , Aged, 80 and over , Combined Modality Therapy , Cross-Sectional Studies , Cyclophosphamide/administration & dosage , Doxorubicin/administration & dosage , Female , Follow-Up Studies , Humans , Male , Middle Aged , Prednisone/administration & dosage , Prognosis , Rituximab/administration & dosage , Time Factors , Transplantation, Autologous , Vincristine/administration & dosage , Young AdultABSTRACT
PURPOSE: The currently ongoing Epidemiological Strategy and Medical Economics (ESME) research programme aims at centralising real-life data on oncology care for epidemiological research purposes. We draw on results from the metastatic breast cancer (MBC) cohort to illustrate the methodology used for data collection in the ESME research programme. PARTICIPANTS: All consecutive ≥18 years patients with MBC treatment initiated between 2008 and 2014 in one of the 18 French Comprehensive Cancer Centres were selected. Diagnostic, therapeutic and follow-up data (demographics, primary tumour, metastatic disease, treatment patterns and vital status) were collected through the course of the disease. Data collection is updated annually. FINDING TO DATE: With a recruitment target of 30 000 patients with MBC by 2019, we currently screened a total of 45 329 patients, and >16 700 patients with a metastatic disease treatment initiated after 2008 have been selected. 20.7% of patients had an hormone receptor (HR)-negative MBC, 73.7% had a HER2-negative MBC and 13.9% were classified as triple-negative BC (ie, HER2 and HR status both negative). Median follow-up duration from MBC diagnosis was 48.55 months for the whole cohort. FUTURE PLANS: These real-world data will help standardise the management of MBC and improve patient care. A dozen of ancillary research projects have been conducted and some of them are already accepted for publication or ready to be issued. The ESME research programme is expanding to ovarian cancer and advanced/metastatic lung cancer. Our ultimate goal is to achieve a continuous link to the data of the cohort to the French national Health Data System for centralising data on healthcare reimbursement (drugs, medical procedures), inpatient/outpatient stays and visits in primary/secondary care settings. TRIAL REGISTRATION NUMBER: NCT03275311; Pre-results.