¹⁸ FDG PET/CT imaging is commonly used in diagnosis and follow-up of metastatic breast cancer, but its quantitative analysis is complicated by the number and location heterogeneity of metastatic lesions. Considering that bones are the most common location among metastatic sites, this work aims to compare different approaches to segment the bones and bone metastatic lesions in breast cancer.Two deep learning methods based on U-Net were developed and trained to segment either both bones and bone lesions or bone lesions alone on PET/CT images. These methods were cross-validated on 24 patients from the prospective EPICURE seinmeta metastatic breast cancer study and were evaluated using recall and precision to measure lesion detection, as well as the Dice score to assess bones and bone lesions segmentation accuracy.Results show that taking into account bone information in the training process allows to improve the precision of the lesions detection as well as the Dice score of the segmented lesions. Moreover, using the obtained bone and bone lesion masks, we were able to compute a PET bone index (PBI) inspired by the recognized Bone Scan Index (BSI). This automatically computed PBI globally agrees with the one calculated from ground truth delineations.Clinical relevance— We propose a completely automatic deep learning based method to detect and segment bones and bone lesions on ¹⁸ FDG PET/CT in the context of metastatic breast cancer. We also introduce an automatic PET bone index which could be incorporated in the monitoring and decision process.

Metastatic breast cancer patients receive lifelong medication and are regularly monitored for disease progression. The aim of this work was to (1) propose networks to segment breast cancer metastatic lesions on longitudinal whole-body PET/CT and (2) extract imaging biomarkers from the segmentations and evaluate their potential to determine treatment response. Baseline and follow-up PET/CT images of 60 patients from the EPICURE${}_{seinmeta}$ study were used to train two deep-learning models to segment breast cancer metastatic lesions: One for baseline images and one for follow-up images. From the automatic segmentations, four imaging biomarkers were computed and evaluated: SUL${}_{peak}$, Total Lesion Glycolysis (TLG), PET Bone Index (PBI) and PET Liver Index (PLI). The first network obtained a mean Dice score of 0.66 on baseline acquisitions. The second network obtained a mean Dice score of 0.58 on follow-up acquisitions. SUL${}_{peak}$, with a 32% decrease between baseline and follow-up, was the biomarker best able to assess patients’ response (sensitivity 87%, specificity 87%), followed by TLG (43% decrease, sensitivity 73%, specificity 81%) and PBI (8% decrease, sensitivity 69%, specificity 69%). Our networks constitute promising tools for the automatic segmentation of lesions in patients with metastatic breast cancer allowing treatment response assessment with several biomarkers.

Les patientes atteintes d’un cancer du sein métastatique reçoivent un traitement tout au long de leur vie et sont suivies très régulièrement afin d’estimer l’efficacité de leurs traitements. Pour évaluer la réponse au traitement, les médecins utilisent des critères d’interprétation des images comme RECIST ou PERCIST. Ces critères se limitent à l’évaluation quantitative d’une partie des tumeurs. La segmentation de toutes les tumeurs et leur caractérisation grâce à plusieurs paramètres (ou biomarqueurs) permettraient l’évaluation plus précise de la réponse au traitement des patientes. Cela permettrait aussi de prédire différentes informations sur les lésions (sous-type moléculaire) ou sur les patientes (survie sans progression, survie globale). Au cours de cette thèse, nous avons développé une méthode permettant la segmentation sur plusieurs acquisitions de toutes les lésions chez des patientes atteintes d’un cancer du sein métastatique. Les segmentations obtenues ont permis l’extraction de différents biomarqueurs que nous avons ensuite utilisés pour évaluer automatiquement la réponse thérapeutique des patientes. Enfin, nous avons utilisé l’apprentissage profond pour prédire le sous-type moléculaire des lésions métastatiques.