FBFrançoise Beltjens

Françoise Beltjens

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Françoise Beltjens has a total of 67 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Age-Related Distribution of Homologous Recombination Deficiency in Advanced Ovarian Carcinoma: A Large Real-World French Cohort.

    by Gertych Witold, Trecourt Alexis, Morineau Faustine, Beacco Pauline, Brasselet Astrid, Carnet Marion, Chevriaux Angélique, Chevrier Sandy, Dupasquier Morgane, Gibeaud Anne, et al.

  • Chemo-prAIdict Breast: A deep learning solution for predicting residual disease on biopsies of breast cancer patients treated with neoadjuvant chemotherapy

    by Natalia Fernanda Valderrama, Louis-Oscar Morel, Daniel Tshokola Mweze, Valentin Derangère, Isabelle Desmoulins, Didier Mayeur, Courèche Kaderbhai, Silvia Ilie, Audrey Hennequin, Nicolas Roussot, et al.


  • Open Access

    Cas no 2. Le cystadénocarcinome mucineux mammaire : à propos d’un cas

    by Françoise Beltjens, Nicolas Hamant, Anthony Bergeron, Céline Charon-Barra, Laurent Arnould

  • Open Access

    Deep Multiple Instance Learning Model to Predict Outcome of Pancreatic Cancer Following Surgery

    by Caroline Truntzer, Dina Ouahbi, Titouan Huppé, David Rageot, Alis Ilie, Chloe Molimard, Françoise Beltjens, Anthony Bergeron, Angelique Vienot, Christophe Borg, et al.

    Background/Objectives: Pancreatic ductal adenocarcinoma (PDAC) is a cancer with very poor prognosis despite early surgical management. To date, only clinical variables are used to predict outcome for decision-making about adjuvant therapy. We sought to generate a deep learning approach based on hematoxylin and eosin (H&E) or hematoxylin, eosin and saffron (HES) whole slides to predict patients’ outcome, compare these new entities with known molecular subtypes and question their biological significance; Methods: We used as a training set a retrospective private cohort of 206 patients treated by surgery for PDAC cancer and a validation cohort of 166 non-metastatic patients from The Cancer Genome Atlas (TCGA) PDAC project. We estimated a multi-instance learning survival model to predict relapse in the training set and evaluated its performance in the validation set. RNAseq and exome data from the TCGA PDAC database were used to describe the transcriptomic and genomic features associated with deep learning classification; Results: Based on the estimation of an attention-based multi-instance learning survival model, we identified two groups of patients with a distinct prognosis. There was a significant difference in progression-free survival (PFS) between these two groups in the training set (hazard ratio HR = 0.72 [0.54;0.96]; p = 0.03) and in the validation set (HR = 0.63 [0.42;0.94]; p = 0.01). Transcriptomic and genomic features revealed that the poor prognosis group was associated with a squamous phenotype. Conclusions: Our study demonstrates that deep learning could be used to predict PDAC prognosis and offer assistance in better choosing adjuvant treatment.


  • ER−/PR+ breast cancer: A distinct entity, which is morphologically and molecularly close to triple‐negative breast cancer

    by Françoise Beltjens, Damien Molly, Aurélie Bertaut, Corentin Richard, Isabelle Desmoulins, Catherine Loustalot, Céline Charon‐Barra, Emilie Courcet, Anthony Bergeron, Sylvain Ladoire, et al.

    Determining the status of steroid hormone receptors [oestrogen (ER) and progesterone receptors (PR)] is a crucial part of the breast cancer workup. Thereby, breast cancers can be classified into four subtypes. However, the existence of ER−/PR+ tumours, often reported to be ill‐classified due to technical errors, remains controversial. In order to address this controversy, we reviewed the hormone receptor status of 49 breast tumours previously classified as ER−/PR+ by immunohistochemistry, and compared clinical, pathological and molecular characteristics of confirmed ER−/PR+ tumours with those of ER+ and triple‐negative tumours. We unequivocally confirmed the ER−/PR+ status in 27 of 49 tumours (0.3% of all breast cancers diagnosed in our institution between 2000 and 2014). We found that ER−/PR+ were morphologically and histologically similar to triple‐negative tumours, but very distinct from ER+ tumours, with more aggressive phenotypes and more frequent basal marker expression than the latter. On the molecular level, RNA sequencing revealed different gene expression profiles between the three groups. Of particular interest, several genes controlled by the suppressor of zest 12 (SUZ12) were upregulated in ER−/PR+ tumours. Overall, our results confirm that ER−/PR+ breast cancers are an extremely rare but ‘real’ tumour subtype that requires careful diagnosis and has distinct features warranting different responsiveness to therapies and different clinical outcomes. Studies on larger cohorts are needed to further characterise these tumours. The likely involvement of SUZ12 in their biology is an interesting finding which may – in a long run – give rise to the development of new therapeutic alternatives.


  • Open Access

    Trends in molecular subtypes of breast cancer: description of incidence rates between 2007 and 2012 from three French registries

    by Marion Cortet, Aurélie Bertaut, Florence Molinié, Simona Bara, Françoise Beltjens, Charles Coutant, Patrick Arveux

    The incidence and incidence trends of breast cancer according to molecular subtype are unknown at a population level in France. The registry data enables this study and may give this information, that is crucial to describe and understand breast cancer epidemiology. We estimated the incidence rates of breast cancer for each molecular subtype using data from three cancer registries in France for the period from 2007 to 2012. Molecular subtypes were defined with immunohistochemical data. Poisson models were estimated to modelize the course of breast cancer incidence and to test the trends. The study included 12,040 patients diagnosed between 2007 and 2012 in the three administrative areas covered by the registries. There was no significant trends in the proportion of each molecular subtype year by year. The age distribution of incident cases was different depending on the molecular subtypes (p < 0.001). The course of incidence between 2007 and 2012 was also different depending on molecular subtype according to the multivariate Poisson model (p < 0.001). The description of incident cases of breast cancer according to molecular subtype at a population level showed differences in trends. The trends in incidence differed according to molecular subtype, and this should improve our understanding of overall changes in incidence. This analysis is important to plan screening and treatment resources at a population level.