SCIENTIFIC AND MEDICAL NEWS

Discover the latest advances in healthcare, ongoing research, and technological innovations that are transforming medical practices. Through this page, we share news, studies, and publications that are shaping the future of the sector.

AI for pancreatic cancer detection using deep learning applied to multiparametric MRI

Project led by the Harvard Medical School and MIT team – 2023

Medical context

Pancreatic cancer is one of the deadliest cancers, with a 5-year survival rate below 10%. Early diagnosis is extremely difficult, as symptoms are subtle and lesions are often invisible in standard examinations during the initial stages. However, multiparametric MRI (diffusion, T1/T2, perfusion) is beginning to reveal subtle biomarkers, undetectable to the human eye but exploitable by algorithms.

Technological innovation

In 2023, a joint team from Massachusetts General Hospital, Harvard Medical School, and MIT’s Computer Science & AI Lab (CSAIL) developed a multimodal deep learning model capable of early pancreatic cancer detection on multiparametric MRI, even before clinical signs appear.

The Model: PanCan-Net

  • Architecture: 3D CNN + multi-head attention, trained separately on each MRI channel, then fused via a cross-channel attention module inspired by Transformers.
  • Training: Over 30,000 abdominal MRIs collected over 10 years, including 800 confirmed positive cases.
  • Preprocessing: Inter-site harmonization, anatomical z-score normalization, automatic masking of non-pancreatic organs.
  • Results: AUC of 0.92 for precancerous lesions (PanIN), several months before clinical diagnosis.

Team and leadership

  • Dr. Ferenc A. Jolesz (posthumously) – pioneer of interventional MRI, whose work inspired the approach.
  • Dr. Leo Grady – Project PI, expert in medical image analysis using graph theory (formerly at HeartFlow, now at Paige).
  • Dr. Regina Barzilay (MIT) – Model architect, expert in deep learning models for oncology.

Clinical implications

  • Early Diagnosis: Detection up to 6 months before visible signs in traditional radiology.
  • Inclusion in High-Risk Cohorts: The model enables stratification of asymptomatic patients for closer monitoring.
  • Anticipated Mortality Reduction: According to clinical modeling, this type of detection could triple the 5-year survival rate in high-risk populations.

Economic and industrial impact

  • Pilot adoption by Mayo Clinic and Brigham & Women’s Hospital in Boston.
  • Ongoing discussions with GE Healthcare to integrate PanCan-Net into next-generation MRI consoles (SIGNA Hero).
  • A spin-off startup, EarlyPan.AI, has raised $25M to industrialize the model and obtain FDA certification.

Limitations and future perspectives

  • The model remains dependent on MRI sequence quality, particularly diffusion-weighted imaging (DWI).
  • Potential biases have been identified in certain ethnic subgroups, currently being corrected through new multi-site data.
  • An extension is underway to cross-reference imaging with genetic data using a multimodal transformer, in collaboration with Google DeepMind.

Patient digital twins generated by diffusion models for clinical trajectory modeling

Project led by INSERM & Inria in collaboration with the Health Data Hub – 2024

Medical context

Patient trajectory modeling is a key challenge in predictive medicine: anticipating a patient’s evolution based on their history allows for earlier treatment adjustments and even prevention of clinical deterioration. However, in practice, these trajectories are often incomplete, noisy, or biased (loss of follow-up, variable interventions, etc.). This is where digital twins become relevant: artificial yet realistic patients evolving in a temporally coherent space aligned with real-world cases.

Technological innovation

In 2024, a Franco-European consortium led by INSERM U1138, Inria Saclay, and the Health Data Hub developed a pipeline based on temporal diffusion models, tailored for generating longitudinal clinical trajectories.
The Model: TwinDiff

  • Technology: Stochastic temporal diffusion model, inspired by DDPM (Denoising Diffusion Probabilistic Models), combined with a structured encoder (GRU-type) to integrate patient history.
  • Data: 1.2 million patient journeys extracted from the SNDS (French National Health Data System), covering chronic diseases (diabetes, COPD, heart failure).
  • Output: Digital twins reproducing the sequence of hospitalizations, prescriptions, biological test results, and critical events.

Team and leadership

  • Dr. Thomas Walter (Inria) – expert in deep learning for medical time series, scientific lead of the project.
  • Prof. Valérie Serre (INSERM / Université Paris Cité) – clinical hematologist, ensuring the medical validity of synthetic trajectories.
  • Morgane Rousseau (Health Data Hub) – ethics and regulatory lead of the project, overseeing GDPR compliance and validation framework.

Clinical implications

  • Simulation of virtual clinical trials: creation of artificial control groups in studies where comparator arms are difficult to establish.
  • Testing hypothetical interventions: simulating the impact of a therapeutic modification on a given patient.
  • Enhancing predictive model training through enriched, diversified, and GDPR-compliant datasets.

Market and regulatory impact

  • The model is being tested as a proof of concept by two industry partners: Servier (oncology) and Biogen (neurology).
  • Discussions underway with the EMA for regulatory validation of digital twins in post-marketing authorization studies.
  • Publication expected in 2025 in Nature Digital Medicine.

Challenges and future perspectives

  • Validation remains challenging: how can we prove that an artificial twin is “accurate” if no observed future exists?
  • Integration of unstructured modalities (clinical notes, images) is underway via a multimodal model like Flamingo.
  • An open-source tool, TwinDiffBench, is planned to enable other hospitals to assess the quality of digital twins using their own data.

Real-world clinical AI agents: the example of Hippocratic AI for postoperative monitoring

Launched in 2023 in pilot hospitals in the United States (Cedars-Sinai, UCSF)

Medical context

Postoperative monitoring is a critical point in the patient journey. Poor management of early complications or side effects can lead to avoidable readmissions. However, caregivers often lack the time to provide intensive follow-up at home. It is in this context that medical conversational agents are emerging, particularly with the advent of specialized healthcare LLMs, capable of conversing with patients, detecting weak signals, and alerting medical teams.

Technological innovation

Hippocratic AI, a startup founded by former executives from Google Health, Amazon Alexa, and Epic Systems, developed in 2023 a HIPAA-certified voice AI agent for postoperative telephone monitoring.
The Agent: Hippocratic Care Agent v1.3

  • Technology: Proprietary LLM fine-tuned on structured medical corpora + patient dialogues, combined with a real-time voice-to-intent system.
  • Functions: Automated daily calls to patients, symptom triage, alerts to caregivers in case of deterioration, secure recording.
  • Covered specialties: Initially bariatric, orthopedic, and gynecological surgery.

Clinical implementation

  • Pilot hospitals: Cedars-Sinai (Los Angeles), UCSF Medical Center, Houston Methodist
  • Supervision: Each AI agent is overseen by a coordinating nurse, with human validation before any critical escalation.
  • Trial duration: 6 months, over 10,000 patients monitored

Results and feedback

  • 23% reduction in 30-day postoperative readmissions
  • 92% patient satisfaction rate, particularly regarding clarity of exchanges and speed of responses
  • Currently being adopted in 12 other hospitals, including the Mayo Clinic

Industrial dimension

  • Hippocratic AI raised $65M in Series A funding, with participation from General Catalyst, Andreessen Horowitz, and Index Ventures.
  • Advanced discussions with Epic to integrate the AI agent into MyChart and with CVS Health for deployment in home care services.

Critical points

  • The model remains limited to simple and codified interactions, even though it understands natural language.
  • Strong trust, explainability, and accountability issues: any alert signal triggers human verification.
  • Ongoing work to adapt the model for vulnerable populations (elderly, non-native speakers, cognitively impaired individuals).

PubMedGPT-X: The medical language model that enters the hospital

A LLM Designed for Clinicians, Not Just Demos
While large general-purpose language models like GPT-4 or Claude demonstrate their power across various tasks, their use in medical environments remains challenging: hallucinations, lack of traceability, and absence of clinical validation make them risky to use. To bridge this gap, the MIT, in collaboration with Mass General Brigham and Harvard Medical School, developed PubMedGPT-X, a language model specifically trained for hospital applications, with one priority: reliability.
This open-source LLM is based on a Transformer-type architecture optimized for a rich biomedical vocabulary. It has been trained on tens of millions of documents: scientific articles from PubMed, clinical guidelines, anonymized hospital records, consultation notes, and therapeutic protocols.
Unlike its general-purpose counterparts, PubMedGPT-X is designed to produce texts aligned with medical standards—structured, contextualized, and verifiable.

Real-world experimentation in hospital departments

In 2024, the model was deployed in real-world conditions in three university hospitals in Boston, within internal medicine, oncology, and geriatrics departments.
Clinicians use it directly through an interface integrated into the Electronic Health Record (EHR). Three use cases were tested:

  1. Generating discharge letters from raw notes;
  2. Automatically summarizing complex hospital stays, including treatments, comorbidities, and medical decisions;
  3. Preparing for Multidisciplinary Team Meetings (MDT) by extracting key elements from patient records.

Early feedback is significant: clinicians save an average of 30 to 40% of time on document writing while producing more standardized, readable, and compliant content according to regulatory requirements.

A medically controlled technology

The project team has implemented an integrated factual verification system: each model output is cross-checked with a medical evidence database (guidelines, referenced publications) and flagged if inconsistencies appear. A confidence score is provided for each response, facilitating human review.
The project is led by Dr. Regina Barzilay, an AI and cancer medicine specialist, in collaboration with Prof. Mark Sendak (Duke Health) for clinical integration. The model is currently being validated for broader deployment in several U.S. public hospitals.

A key component for the augmented hospital

PubMedGPT-X could become a backbone for documentation in healthcare institutions: automatically structuring information, supporting care continuity, and improving traceability in overloaded environments. By opening its code and training data, the team also aims to create an ethical and sovereign standard against closed LLMs dominated by Big Tech.

Multimodal AI in healthcare: towards integrated clinical intelligence

Understanding a patient as a whole
One of the major challenges in medicine is not only analyzing an image, a biological result, or a clinical note separately, but interpreting them together. Medical practice is inherently multimodal: a diagnosis relies on images, medical history, symptoms, lab results, and sometimes also genetic information or social context. Until now, AI has processed these elements in silos. Today, a new generation of models aims to merge them into a unified representation. This is the ambition of the Med-Flamingo project, led by Google DeepMind in collaboration with the University of Cambridge and King’s College Hospital. The goal: to design a multimodal medical model capable of reading a radiological image, understanding its clinical context, and formulating a hypothesis or response to a medical question in natural language.

An AI that sees, reads, and reason

Med-Flamingo is based on an architecture derived from Flamingo, DeepMind’s general-purpose multimodal model. Adapted to medicine, it has been trained on a corpus of over 500,000 text-image pairs, including MRIs, X-rays, CT scans annotated with their reports, as well as doctor-patient dialogues.
The model receives an image as input, accompanied by free text (e.g., “62-year-old patient, history of smoking, acute chest pain”) and must produce a response: a diagnostic hypothesis, a therapeutic direction, or an explanation for educational use. In early tests, the model demonstrated an impressive ability to generalize: detecting a pneumothorax on an X-ray, justifying a specific anticoagulant dose, or formulating a response to the patient in simple language.

Towards an augmented clinical interface

The team has initiated an experiment in two London hospitals. The model is used as a co-pilot in emergency departments to assist in rapid decision-making. For example, after analyzing an image and the patient’s file, the AI provides a prioritized summary of possible diagnoses, with a confidence level and links to corresponding guidelines. The model is not autonomous: it only intervenes at the clinician’s request and does not issue orders or prescriptions, ensuring compliance with regulatory frameworks. Each response is logged and can be audited later.

Challenges to overcome

The potential is vast, but several obstacles remain. The model needs to learn to manage uncertainty, gaps in the file, variations in image quality, and contradictions between sources. It also needs to adapt to local contexts: the same symptom does not always have the same meaning depending on the country, practices, or populations. Discussions are ongoing with the UK’s NHS to expand the model’s use to general medicine and with the EMA to initiate a regulatory validation framework for multimodal AIs.

A new era of medical AI

With this type of model, we enter a paradigm of global patient understanding, where AI becomes a transversal clinical assistant, capable of contextualizing, prioritizing, and explaining. It is no longer just a technical tool but a reasoning partner, which could transform the way doctors navigate the complexity of modern clinical cases.