25 June 2025 in Computers in Biology and Medicine
Sarcopenia, defined as the progressive loss of skeletal muscle mass and function, has been associated with poor prognosis in patients with pancreatic cancer, particularly those with borderline resectable pancreatic cancer (BRPC). Although body composition can be extracted from routine CT imaging, sarcopenia assessment remains underused in clinical practice. Recent advances in artificial intelligence (AI) offer the potential to automate and standardize this process, but their clinical translation remains limited. This narrative review aims to critically evaluate (1) the clinical impact of CT-defined sarcopenia in BRPC, and (2) the performance and maturity of AI-based methods for automated muscle and fat segmentation on CT images. A dual-axis literature search was conducted to identify clinical studies assessing the prognostic role of sarcopenia in BRPC, and technical studies developing AI-based segmentation models for body composition analysis. Structured data extraction was applied to 13 clinical and 71 technical studies. A PRISMA-inspired flow diagram was included to ensure methodological transparency. Sarcopenia was consistently associated with worse survival and treatment tolerance in BRPC, yet clinical definitions and cut-offs varied widely. AI models-mostly 2D U-Nets trained on L3-level CT slices-achieved high segmentation accuracy (mean DSC >0.93), but external validation and standardization were often lacking. CT-based AI assessment of sarcopenia holds promise for improving patient stratification in BRPC. However, its clinical adoption will require standardization, integration into decision-support frameworks, and prospective validation across diverse populations.
1 February 2023 in Radiotherapy and Oncology
VBValérie Bernier-Chastagner
Valérie Bernier-Chastagner
+1
26 May 2024 in Radiotherapy and Oncology
VBValérie Bernier-Chastagner
Valérie Bernier-Chastagner
To identify dosimetric predictive factors of sensorineural hearing loss (SNHL) in children after cranial radiation therapy (RT) in a single institution using dosimetric data from the French National Registry PediaRT. Complete audiological follow-up data were available for 44 children treated with cranial RT between 2014 and 2021 at our institution. The median age at the time of RT initiation was 9 years (range: 2-17 years). No children presented with hearing loss prior to treatment. SNHL was defined as a Chang ototoxicity grade ≥ 1a or higher. Median audiometric follow-up duration was 51 months. Seven children (16 %) developed SNHL with a median time to occurrence of 33 months (range, 18-46 months). The estimated SNHL cumulative rate at 2 years post-RT was 4,5% ± 3,1% and at 5 years was 21 % ± 7.2 %. Multiple Cox regression models showed that the association of the age at radiotherapy and the dosimetric values to the inner ear canal and cochlea were the most significant predictive factors of SNHL occurrence. No child who received less than 35 Gy on average to both cochleae (n = 26) suffered from SNHL, whereas the 5-year SNHL cumulative incidence for the children who received greater than or equal to 35 Gy on average to either cochlea (n = 18) was 51.8 % ± 15.1 %. Doses received by the inner ear canal and cochlea, associated with the age at RT initiation, are the main predictive factors for radiation-induced SNHL. A median dose to either cochlea over 35 Gy significantly increases the risk of SNHL and justify close audiometric monitoring to detect and equip hearing loss at an early stage.
24 November 2025 in International Journal of Gynecological Cancer
28 July 2026 in npj Digital Medicine
The EU AI Act subjects high-risk clinical AI systems to deployer obligations whose institutional operationalization remains undefined. Drawing on experience at a French comprehensive cancer center, we present a governance framework combining three novel components: a dual-axis model for institutional risk stratification, a five-category AI-related adverse event taxonomy inspired by pharmacovigilance, and structured proficiency testing against simulated AI errors. Successful AI integration depends on institutional readiness, not algorithmic performance.
3 October 2026 in BMC Medical Education
Accurate delineation of organs at risk (OARs) remains a cornerstone of radiotherapy planning. As AI-assisted auto-contouring enters routine practice, the p