Institut de cancérologie de l'Ouest, Saint-Herblain, France; CRCI2NA, INSERM UMR1307, CNRS-ERL6075, Université d'Angers, Université de Nantes, Nantes, France
Auto-generated profile, based on publicly available information
Ludovic Ferrer
Institut de cancérologie de l'Ouest, Saint-Herblain, France; CRCI2NA, INSERM UMR1307, CNRS-ERL6075, Université d'Angers, Université de Nantes, Nantes, France
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.
Radiomics have emerged as an exciting field of research over the past few years, with very wide potential applications in personalised and precision medicine of the future. Radiomics-based approaches are still however limited in daily clinical practice in oncology. This review focus on how radiomics could be incorporated into the radiation therapy pipeline, and globally help the radiation oncologist, from the tumour diagnosis to follow-up after treatment. Radiomics could impact on all steps of the treatment pipeline, once the limitations in terms of robustness and reproducibility are overcome. Major ongoing efforts should be made to collect and share data in the most standardised manner possible.
GATE figures among existing MC simulation codes for emission tomography applications. Although it allows working with voxellized activity distributions it does not enable the incorporation of physiological motion in a real time sense for such distributions. This restriction causes substantial drawbacks where dynamic phenomena such as respiratory or radiotracer motion are of interest. A new C++ class, named "GateRTPhantom" has been introduced in GATE to enable simulation of physiological motion in real time within voxellized phantoms. These objects are managed by a class named "GateRTPhantomMgr". The new approach has been assessed using the 4D NCAT phantom. The real time implementation approach (RTNCAT: Real Time NCAT) was compared with the classic implementation of dynamic processes through the individual simulation of a number of static frames. A total of 20 frames throughout a respiratory cycle were produced (0.250 ms each). The accuracy of the RTNCAT module was validated through a qualitative and quantitative comparison of 3D images containing the emission locations of single photons in the phantom. Parameters such as the total number, the mean value and the standard deviation of the number of emitted photons per voxel were evaluated. In all of these parameters a high level of agreement was observed supporting the developed approach for the real time modelling of physiological motion. In addition, practically identical times of execution were seen for the two approaches.
2022 Background: [F18]-2-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET) is commonly used in the diagnosis and staging of cancer. Clinical and radiological evaluation of response to neoadjuvant chemotherapy is presently based on change in tumor size, and the correlation with pathological response is often inaccurate. The aim of study is to prospectively assess FDG-PET sequential imaging as an early indicator of response to neoadjuvant chemotherapy in locally advanced breast cancer patients. Methods: A total of 250 attenuation-corrected scans has been obtained in 43 patients. Images were obtained with a CT/PET scanner (GE Discovery LS) following administration of FDG (5 MBq/kg) at the baseline and after the first, second, third and sixth course of chemotherapy. PET data were interpreted using standardized uptake value (SUV) measurements. The changes of SUV data with regard to baseline values were correlated with surgical/histopathological response. Results: Surgery and histopathology after 6 courses of chemotherapy revealed gross residual disease (GRD) in 18 patients and minimal residual disease (MDR) in 25 patients. Significant differences in FDG absolute uptake between responding and non responding patients were observed at the baseline (p<0.05) but this difference disappeared after the first course of chemotherapy. FDG uptake showed minimal change in non responding patients as documented by histopathological analysis but marked decrease up to the background level in most (19/25) responding patients. Considering a cut off value of 40% for the SUV decrease with regard to baseline, the sensitivity, specificity and accuracy for response to chemotherapy after one, two or three courses were respectively (60%, 100%, 77%), (88%, 100%, 93%), (96%, 100%, 98%). Conclusions: FDG-PET imaging of locally advanced breast cancer following two courses of neoadjuvant chemotherapy can accuratly predict pathological response to treatment. No significant financial relationships to disclose.
Internal vectorised therapies, particularly with [177Lu]-labelled agents, are increasingly used for metastatic prostate cancer and neuroendocrine tumours. However, routine dosimetry for organs-at-risk and tumours remains limited due to the complexity and time requirements of current protocols. We developed a Generative Adversarial Network (GAN) to transform rapid 6 s SPECT projections into synthetic 30 s-equivalent projections. SPECT data from twenty patients and phantom acquisitions were collected at multiple time-points. The GAN accurately predicted 30 s projections, enabling estimation of time-integrated activities in kidneys and liver with maximum errors below 6 % and 1 %, respectively, compared to standard acquisitions. For tumours and phantom spheres, results were more variable. On phantom data, GAN-inferred reconstructions showed lower biases for spheres of 20, 8, and 1 mL (8.2 %, 6.9 %, and 21.7 %) compared to direct 6 s acquisitions (12.4 %, 20.4 %, and 24.0 %). However, in patient lesions, 37 segmented tumours showed higher median discrepancies in cumulated activity for the GAN (15.4 %) than for the 6 s approach (4.1 %). Our preliminary results indicate that the GAN can provide reliable dosimetry for organs-at-risk, but further optimisation is needed for small lesion quantification. This approach could reduce SPECT acquisition time from 45 to 9 min for standard three-bed studies, potentially facilitating wider adoption of dosimetry in nuclear medicine and addressing challenges related to toxicity and cumulative absorbed doses in personalised radiopharmaceutical therapy.
15 September 2026 in Zeitschrift für Medizinische Physik