Skip to Main content Skip to Navigation
Master Thesis

Radiomique de l’imagerie par résonnance magnétique (IRM) multiparamétrique pour la prédiction de la réponse pathologique complète chez les femmes traitées par chimiothérapie néoadjuvante d’un cancer du sein triple négatif

Abstract : Purpose: We studied the performance of pre-therapeutic radiomic MRI analysis in order to predict the full pathological response in women treated with neoadjuvant chemotherapy for triple negative breast cancer. Materials and methods : That monocentric retrospective study concerned 76 female patients that benefited from MRI in our centre before neo-adjuvant chemotherapy, between January 2007 and June 2017. The analysis of tumoral and ipsilateral mammary parenchymal was realized on T1, T2, diffusion and dynamic T1 sequences after injection of gadolinium. The initial radiom was made of 342 characteristics describing the shape, the size, the greyscale distribution and the texture of the segmented region. For each tumoral lesion, the tumor and the parenchyma were segmented independently and were analyzed with a radiomic analysis. Moreover, several combinations of radiomic analysis were realized based on the 4 different sequences (T1 weighting, T2 weighting, diffusion and dynamic with gadolinium injection weighting). The main purpose was to predict the full pathological response on the excised tissues. Results: Radiomic MRI could predict the response to chemotherapy with an average of AUC value of 0.75. Several models were analyzed and the best result was the one with T1, T2 and dynamic sequences centered on the tumor. The addition of tumoral diffusion did not produce better results. Conclusion: That study demonstrates that pre-therapeutic radiomic MRI analysis can be useful as a biomarker in order to predict the full pathological response amongst female patients diagnosed with triple negative breast cancer and treated with neo-adjuvant chemotherapy.
Document type :
Master Thesis
Complete list of metadatas

Cited literature [85 references]  Display  Hide  Download

https://dumas.ccsd.cnrs.fr/dumas-02747062
Contributor : Jean-Hugues Morneau <>
Submitted on : Wednesday, June 3, 2020 - 11:17:32 AM
Last modification on : Wednesday, July 15, 2020 - 8:58:04 AM

File

2019GREA5158_chaudet_pierre(1)...
Files produced by the author(s)

Identifiers

  • HAL Id : dumas-02747062, version 1

Citation

Pierre Chaudet. Radiomique de l’imagerie par résonnance magnétique (IRM) multiparamétrique pour la prédiction de la réponse pathologique complète chez les femmes traitées par chimiothérapie néoadjuvante d’un cancer du sein triple négatif. Médecine humaine et pathologie. 2019. ⟨dumas-02747062⟩

Share

Metrics

Record views

57

Files downloads

78