BEGIN:VCALENDAR VERSION:2.0 PRODID:-//Talks.cam//talks.cam.ac.uk// X-WR-CALNAME:Talks.cam BEGIN:VEVENT SUMMARY:Cambridge MedAI Seminar - February 2026 - Margherita Favali and Ti ago Assis DTSTART:20260224T114500Z DTEND:20260224T130000Z UID:TALK245038@talks.cam.ac.uk CONTACT:Hannah Clayton DESCRIPTION:Sign-up on Eventbrite: https://medai-feb2026.eventbrite.co.uk\ n\nJoin us for the *Cambridge AI in Medicine Seminar Series*\, hosted by t he *Cancer Research UK Cambridge Centre* and the *Department of Radiology at Addenbrooke’s*. This series brings together leading experts to explor e cutting-edge AI applications in healthcare – from disease diagnosis to drug discovery. It’s a unique opportunity for researchers\, practitione rs\, and students to stay at the forefront of AI innovations and engage in discussions shaping the future of AI in healthcare.\n\nThis month’s sem inar will be held on *Tuesday 24 February 2026\, 12-1pm at the Jeffrey Che ah Biomedical Centre (Main Lecture Theatre)\, University of Cambridge* and *streamed online via Zoom*. A light lunch from Aromi will be served from 11:45. The event will feature the following talks:\n\n*_CT-Based Deep Lear ning Model for Predicting Immunotherapy Efficacy in Non-Small Cell Lung Ca ncer Patients_ - Margherita Favali\, PhD student\, Department of Electroni cs Information and Bioengineering\, Politecnico Milano\, Italy*\n\nMargher ita Favali is a PhD student in Bioengineering at Politecnico di Milano. Sh e is currently working on the I3Lung project\, focusing on developing Arti ficial Intelligence models to predict Immunotherapy response in NSCLC pati ents using CT scans\, with an emphasis on applying explainability techniqu es to make these models transparent and understandable.\n\n*Abstract*: Non -small cell lung cancer (NSCLC) accounts for the majority of all lung canc ers and\, despite advances in systemic therapies\, prognosis remains poor. Immunotherapy (IO) has transformed the treatment landscape for advanced N SCLC\; however\, outcome heterogeneity persists\, underscoring the need fo r reliable tools for early risk stratification. While deep learning models show promising results\, trustworthy clinical deployment requires evaluat ion beyond performance\, incorporating fairness and explainability. This w ork presents a structured framework for the trustworthy assessment of medi cal imaging models applied to six-month overall survival (OS6) prediction in IO-treated NSCLC patients. Using 483 patients with baseline CT scans fr om the APOLLO11 observational study\, we used a 2D ResNet50 architecture p retrained on RadImageNet\, and systematically analyze the impact of oversa mpling\, undersampling\, and class-weighted loss on performance. We evalua te model along three dimensions: (1) predictive performance\, (2) equalise d-odds fairness across sex- and age-sensitive attributes\, and (3) explain ability with Grad-CAM\, assessed quantitatively through randomization and perturbation-based faithfulness tests. The 2D ResNet50 trained with class- weighted loss achieved AUC equal to 0.68 and satisfied fairness criteria a cross subgroups under the equalized odds metric. Explainability analysis s howed modest but meaningful faithfulness.\n\n*_Biomechanically-Guided Deep Learning for Brain Tumour Surgery_ - Tiago Assis\, Machine Learning Resea rcher\, Faculty of Sciences\, University of Lisbon\, Portugal*\n\nTiago As sis is a Machine Learning Researcher and holds an MSc in Data Science from the Faculty of Sciences\, University of Lisbon\, where he specialized in deep learning for medical imaging. His research centers on data-driven met hods for image-guided surgery and multimodal continual learning under miss ing data and distribution shift. He recently shared his work at MICCAI 202 5 in South Korea\, where he combined biomechanical brain modeling with dee p neural networks to achieve physically consistent brain shift compensatio n for image-guided neurosurgery. Prior to transitioning into medical AI\, Tiago completed a BSc in Biochemistry at NOVA University Lisbon and worked as a lab tech in Clinical Pathology.\n\n*Abstract*: Brain shift remains a major limitation in image-guided neurosurgery\, reducing the accuracy of neuronavigation as resection progresses. While biomechanical models can co mpensate for this deformation\, their computational cost limits intraopera tive use. Keypoint-based registration methods offer faster alternatives bu t rely on geometric interpolators that can produce physically unrealistic outputs. This talk presents a deep learning framework for brain shift comp ensation that achieves biomechanical accuracy at clinical inference speeds . We constructed a large-scale dataset of patient-specific brain deformati ons using biomechanical simulations\, then trained a deep neural network t o generate physically plausible deformation fields from sparse intraoperat ive data. By supervising the network with these simulations\, we implicitl y encode constitutive properties of brain tissue biomechanics into its lea rned representations. Our experiments demonstrate that this work provides a practical compromise between physical fidelity and computational efficie ncy for intraoperative brain shift compensation.\n\nThis is a hybrid event so you can also join via Zoom: https://zoom.us/j/99050467573?pwd=UE5OdFdT SFdZeUtIcU1DbXpmdlNGZz09\n\nMeeting ID: 990 5046 7573 and Passcode: 617729 \n\nWe look forward to your participation! If you are interested in gettin g involved and presenting your work\, please email Ines Machado at im549@c am.ac.uk\n\nFor more information about this seminar series\, see: https:// www.integratedcancermedicine.org/research/cambridge-medai-seminar-series/ LOCATION:Jeffrey Cheah Biomedical Centre (Main Lecture Theatre)\, Universi ty of Cambridge END:VEVENT END:VCALENDAR