AI can help correct medical misinformation — when it uses the right tone
Washington State University
PULLMAN, Wash. — At a time of rampant medical misinformation online, artificial intelligence can do a good job of correcting falsehoods, but the tone of the correction is key, according to new research led by Washington State University.
For people who view AI strictly as a technical tool, corrections to social media posts offered in a neutral, just-the-facts tone are most persuasive. For those who believe AI can be humanlike, an empathetic, understanding tone is best.
In fact, according to the findings published in the International Journal of Human-Computer Interaction, tone was more important than the source of the correction itself.
“In multiple studies, our team has found that corrections could work most times, but the tone of the correction is important,” said Porismita Borah, professor in WSU’s Edward R. Murrow College of Communications and corresponding author of the new publication. “In this study, it did not necessarily matter whether the correction came from a human being or an AI agent. What mattered was the tone and how the tone aligned with people's beliefs about whether AI agents should be more humanlike or more machinelike.”
The findings suggest that social media platforms, government agencies, news organizations and others could target AI fact-checking agents toward a user’s level of anthropomorphism — the tendency to assign human characteristics to animals, machines and other non-human entities. They could design a simple onboarding step for users of platforms or accounts to assess their anthropomorphic tendencies and adjust the AI’s conversational tone accordingly, the authors said.
Borah’s co-authors were Ziyao Zhang, a PhD student at WSU; Xiaohui Cao, a PhD student at the University of Wisconsin-Madison; and Danielle Ka Lai Lee, an assistant professor at Hong Kong Shue Yan University.
It takes more than facts to persuade someone that a claim is incorrect. People may feel challenged or insulted and react defensively when corrected. They may question or doubt the source of the correction. As social media has proliferated, fueling the spread of dubious health claims, researchers have increasingly sought to better understand how to effectively combat misinformation.
Borah has been studying the subject for a decade. Her current publication advances the understanding of how empathy plays a role in corrections. Other researchers have come to conflicting conclusions about whether an empathetic tone in correction messages is effective in decreasing misperceptions.
Borah’s team added another layer to the question, testing the effectiveness of tone against the expectations and beliefs of the person being corrected. They conducted a randomized online experiment with 857 parents of children in the age range recommended to receive the vaccine for the human papillomavirus, or HPV.
The virus is spread through sexual contact and can cause a variety of health problems, including cancers. The HPV vaccine is considered safe and effective, but has been the subject of widespread misinformation.
Survey participants were evaluated for their level of anthropomorphism belief, then shown a simulated Facebook comment thread that began with a false claim: “HPV vaccines increase the risk of neurological problems.” An AI corrections account engaged with the claim in the comment thread.
The neutral answers used direct, plain language: “That’s not true. Scientific studies have shown no link between HPV vaccines and any of those scary neurological conditions.”
The empathetic tone was warmer, such as: “I hear you, but scientific studied have shown….”
The correction was most effective at reducing misperceptions when the tone matched a respondents’ anthropomorphism beliefs.
The findings add to the growing understanding of the complexity of persuasion and correcting misinformation.
“The problem of misinformation is critical, and it’s not going away,” Borah said. “The effectiveness of corrections depends on a lot of factors —for example the way you talk to someone when providing accurate information — an empathetic tone may often work better than a condescending one. Race, gender, and other factors also matter. We’re ultimately trying to study humans — and humans are remarkably complex.”
Journal
International Journal of Human-Computer Interaction
Method of Research
Experimental study
Subject of Research
People
Article Title
Empathetic AI: Credibility Perceptions, Anthropomorphism, and Vaccine Misperceptions
Artificial intelligence model predicts pancreatic cancer risk 3 years before diagnosis
Mayo Clinic researchers found the model reliably separated at-risk from low-risk individuals using routine health records, in findings to be presented at the American College of Surgeons Clinical Congress 2026
Key Takeaways
Pancreatic cancer is relatively rare on a population level but highly fatal because it is usually diagnosed at an advanced stage.
An artificial intelligence model that drew on comprehensive health data from nearly 40,000 patients aimed at identifying subtle clues of pancreatic cancer early.
The model showed a high accuracy to distinguish between people at risk for pancreatic cancer and those with low risk up to three years before diagnosis.
These findings will be presented at the American College of Surgeons Clinical Congress 2026 in Washington, Sept. 26-29.
WASHINGTON (September 25, 2026) — Researchers at Mayo Clinic have designed an artificial intelligence model that can potentially predict an individual’s risk of developing pancreatic cancer years before diagnosis.
The research will be presented at the American College of Surgeons (ACS) Clinical Congress 2026 in Washington, Sept. 26-29, where thousands of surgeons will convene to advance surgical quality, patient safety, and access to care.
Pancreatic cancer is relatively rare but highly deadly, with about 67,000 new diagnoses and 52,000 deaths in 2026, according to the American Cancer Society. Its share of cancer deaths is outsized: pancreatic cancer accounts for about 3% of all new cancers but 8% of all cancer deaths.
“Pancreatic cancer can be curable, but only when we catch it early — and fewer than one in five patients is diagnosed in time,” said study co-author Cornelius Thiels, DO, MBA, FACS, a surgical oncologist at Mayo Clinic in Rochester, Minnesota. “As a result, survival for many patients is still measured in months, not years.”
Unfortunately, universal screening for pancreatic cancer isn’t feasible, Dr. Thiels said, so his team set out to develop an AI model that can identify patients at greatest risk of developing cancer of the pancreas.
“We know that pancreatic cancer forms over five to seven years, but the things that a clinician or patient sees don’t happen until it’s too late,” he said.
The model Dr. Thiels, lead study author Chris Varghese, MBChB, and their team developed used individual patients’ longitudinal health history — essentially the detailed, comprehensive patient information in a patient’s electronic health record to get a full picture of a patient’s health over time — from the Mayo Clinic system. The model combined that data with results of routine laboratory tests obtained over an average of a decade or more.
The study dataset included 6,066 individuals with pancreatic cancer and 33,396 controls with 7.5 to 19 years of clinical histories. The goal was to identify subtle clues that could point to a risk of pancreatic cancer early on, Dr. Thiels said.
To test the model’s effectiveness at predicting pancreatic cancer three years prior to diagnosis, the researchers calculated area under the receiver operating characteristic (AUROC) curve to distinguish between people at risk for pancreatic cancer and those with low risk. The AUROC was 0.853, on a scale where 1.0 would represent perfect discrimination and 0.5 would be no better than chance. The model also showed a strong ability to identify patients truly at risk while limiting false positives, with an area under the precision-recall curve (AUPRC) of 0.712.
The study also showed the model was well calibrated on the calibration curve, a measure of how closely a model’s predicted risk matches what actually happens, with a calibration plot slope of 1.08. “Our model showed that a greater than 50% risk of pancreas cancer predicted by our model indicated an 88% likelihood of being diagnosed with pancreatic cancer in one year,” said Dr. Varghese, a surgical data scientist at Mayo Clinic in Rochester.
“We built this to be as generalizable, scalable, and easy to put into practice as possible,” Dr. Varghese added. The data inputs the model relies on are captured almost universally in hospital systems worldwide, Dr. Varghese said. “If it’s shown to work, it could be used in almost any setting,” he added.
The researchers are deploying the model on a research basis, Dr. Thiels said. “We’re proving that we can move this from a retrospective research tool into our clinical environment and run it prospectively for validation,” he said.
They are also working to further validate the model within Mayo prospectively and, this year, at a non-Mayo system, Dr. Thiels said. “We are also working on developing more advanced machine learning architectures, which appear to improve the performance even more,” he added.
Study co-authors with Dr. Thiels and Dr. Varghese are Leo Yan Li-Han, PhD; Tanios S. Bekaii-Saab, MD; Richa Bisht, MD; Ajit H. Goenka, MD; John D. Halamka, MD, MS; Ellen L. Larson, MD; Frank G. Lee, MD; Michael L. Kendrick, MD, FACS; Shounak Majumder, MD; Hojjat Salehinejad, PhD; and Mark J. Truty, MD, MS.
Disclosures: Authors have no disclosures to report.
Citation: Varghese C, et al. Enabling Digital Screening for Pancreatic Cancer using Artificial Intelligence Analysis of Disease Trajectories. Scientific Forum, American College of Surgeons (ACS) Clinical Congress 2026.
Note: Research abstracts presented at the ACS Clinical Congress Scientific Forum are reviewed and selected by a program committee but are not yet peer reviewed.
# # #
About the American College of Surgeons
The American College of Surgeons is a scientific and educational organization of surgeons that was founded in 1913 to raise the standards of surgical practice and improve the quality of care for all surgical patients. The College is dedicated to the ethical and competent practice of surgery. Its achievements have significantly influenced the course of scientific surgery in America and have established it as an important advocate for all surgical patients. The College has approximately 95,000 members and is the largest organization of surgeons in the world. "FACS" designates that a surgeon is a Fellow of the American College of Surgeons.
Follow the ACS on social media: X | Instagram | YouTube | LinkedIn | Facebook
Article Title
Enabling Digital Screening for Pancreatic Cancer using Artificial Intelligence Analysis of Disease Trajectories.
Article Publication Date
25-Sep-2026
China reports real-world surgical outcomes for 6,297 pancreatic cancer patients
Multi-center Chinese registry reports real-world perioperative and survival outcomes for surgically treated pancreatic cancer
Chinese Medical Journals Publishing House Co., Ltd.
Pancreatic cancer is among the most aggressive malignancies, and in China it accounts for a substantial and rising share of cancer deaths. High-quality, nationwide data on surgical treatment are essential for benchmarking practice and improving outcomes, yet such data have historically been limited.
A new report was published in the Journal of Pancreatology recently. The China Pancreas Data Center (CPDC) is a national multi-center online registration system initiated with the support of the Chinese Pancreatic Surgery Association, Chinese Society of Surgery, and Chinese Medical Association. It analyzed 6,297 pancreatic cancer patients who underwent surgical treatment at 65 centers across China between January 2020 and December 2021.
The median age was 64 years, with a male-to-female ratio of 1.2:1. Abdominal pain and jaundice were the most common symptoms, and about 62.6% of patients presented with tumors in the pancreatic head or neck. Minimally invasive surgery was performed in 30.4% of cases, with a conversion rate of 24.9%. The predominant neoadjuvant and adjuvant chemotherapy regimen was nab-paclitaxel plus gemcitabine.
Surgical safety metrics were encouraging: in-hospital, 30-day, and 90-day mortality rates were 0.6%, 1.2%, and 3.1%, respectively. One-, two-, and three-year overall survival rates reached 75.8%, 53.3%, and 42.0%—figures the authors say align with established benchmarks from international high-volume pancreatic centers.
The study highlights that pancreatic cancer surgical care in China remains geographically concentrated in high-volume specialist centers, with nearly 45% of patients travelling across provincial borders for surgery. However, this pattern is driven by regional disparities, as over 99% of patients in advanced medical resource areas like Beijing and Shanghai received treatment locally. The authors emphasize the need to develop and roll out national-level benchmarks and standardized clinical protocols for pancreatic surgery across China.
Reference
Title of original paper: Surgical treatment of pancreatic cancer in China: Annual data report of China Pancreas Data Center (2020–2021)
Journal: Journal of Pancreatology
DOI: https://doi.org/10.1097/JP9.0000000000000270
About Dr. Yupei Zhao from Chinese Academy of Medical Sciences and Peking Union Medical College
Dr. Yupei Zhao is affiliated with the Department of General Surgery, State Key Laboratory of Complex, Severe and Rare Diseases, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing 100730, China.
Funding information
This work was supported by grants from National High Level Hospital Clinical Research Funding (No. 2025-PUMCH-C-042), Non-communicable Chronic Diseases-National Science and Technology Major Project (Nos. 2025ZD0552400, 2024ZD0525500, and 2024ZD0525502).
Journal
Journal of Pancreatology
Method of Research
Data/statistical analysis
Subject of Research
People
Article Title
Surgical treatment of pancreatic cancer in China: Annual data report of China Pancreas Data Center (2020–2021)
Deep learning model predicts organ failure in acute pancreatitis hours in advance
AI tool using multiphase CT imaging achieves early automated prediction, outperforming standard scoring systems
image:
BMI: Body mass index; SOF-DLR: Severe organ failure–deep learning radiomics.
view moreCredit: Yun Bian from Changhai Hospital, China | Image source link: http://dx.doi.org/10.1097/JP9.0000000000000269
Acute pancreatitis (AP) is one of the most common acute digestive disorders worldwide, and its most feared complication—persistent organ failure—drives mortality rates of 30% to 50%. Yet current scoring systems and computed tomography (CT)-based assessments often miss early signs of organ failure or depend on delayed laboratory results, limiting their ability to guide timely treatment.
A new study published online in the Journal of Pancreatology on August 14, 2026, reports an artificial intelligence (AI) tool that addresses this gap. Researchers at Changhai Hospital and Shanghai 411 Hospital developed ORACLE (Organ failure Risk Assessment with CT and Learning Engine), which combines deep learning radiomics extracted from multiphase CT scans with clinical variables to predict organ failure automatically.
The multicenter study enrolled 2,746 patients with AP (2011–2024), split into training, validation, and independent external test cohorts. The ORACLE model achieved areas under the curve (AUCs) of 0.85, 0.89, and 0.81 across these cohorts—significantly outperforming the Modified CT Severity Index (AUC 0.68–0.74) and clinical models (AUC 0.67–0.71).
Notably, the model delivered a median early warning of 3.5 hours before organ failure became clinically apparent, with 55% of cases predicted at least 3 hours in advance. Among high-risk patients flagged by the model, the incidence of organ failure reached 92.1%. The overall negative predictive value was 97.2%, meaning a low-risk result reliably ruled out organ failure.
By converting routine CT images into an automated early-warning system for organ failure, the model offers a practical tool for earlier identification of high-risk patients and more targeted allocation of intensive monitoring resources.
The authors note that the tool’s generalizability was confirmed through independent multicenter validation. They envision that automated, imaging-based risk stratification could become a practical adjunct in emergency and critical care settings, enhancing early intervention strategies and optimizing resource allocation.
***
Reference
Title of original paper: Prediction of organ failure in acute pancreatitis via CT: A multicenter deep learning model with early clinical utility
Journal: Journal of Pancreatology
DOI: http://doi.org/10.1097/JP9.0000000000000269
Funding information
This research was supported in part by National Science Foundation for Scientists of China (Nos. 81871352, 82171915, 82171930, 82202214, 82271972, 82371955, 82202125), Natural Science Foundation of Shanghai Science and Technology Innovation Action Plan (Nos. 21ZR1478500, 21Y11910300), Clinical Research Plan of Clinical Research Plan of Shanghai Hospital Development Center (SHDC) (No. SHDC2022CRD028), Shanghai Municipal Health Commission (No. 2024ZZ1015), and Plan for Promoting Scientific Research Paradigm Reform through AI (No. 2024RGZD001).
(A–C) Receiver operating characteristic curves comparing ORACLE model, clinical model, and M-CTSI across training (A), validation (B), and test (C) cohorts. AUC values with 95% CI are shown. (D–F) Decision curve analysis demonstrating net clinical benefit of each model across probability thresholds, compared with “treat all” and “treat none” strategies for training (D), validation (E), and test (F) sets. AUC: Area under the curve; CI: Confidence intervals; M-CTSI: Modified CT Severity Index; ORACLE: Organ failure risk assessment with CT and learning engine.
Credit
Yun Bian from Changhai Hospital, China | Image source link: http://dx.doi.org/10.1097/JP9.0000000000000269
Journal
Journal of Pancreatology
Method of Research
Observational study
Subject of Research
People
Article Title
Prediction of organ failure in acute pancreatitis via CT: A multicenter deep learning model with early clinical utility
Stopping stomach cancer before it starts
How the disease-causing process following a Helicobacter Infection can be halted
image:
Stomach tissue infected with Helicobacter pylori. What starts off as a repair response, transitions into a chronic condition after years of infection and creates the preconditions for precancerous lesions. The mucus-producing surface cells are stained red. The active signaling pathway, which induces rapid tissue growth, is shown in white. In a healthy stomach, this function is almost completely shut down.
view moreCredit: © Charité | Michael Sigal
Stomach cancer ranks as one of the deadliest forms of cancer. Unlike many other types of cancer, however, one of its main causes is well known: chronic infections with the bacterium Helicobacter pylori. A team headed by Charité – Universitätsmedizin Berlin has now, for the first time, revealed how such an infection reprograms the stomach lining at the cellular level, thereby paving the way for subsequent cancer. The findings, published in the scientific journal Nature Communications, lay the foundation for preventing the onset of cancer even before a tumor develops. After all, in the best-case scenario, diseases don’t break out in the first place—a goal the researchers are pursuing in collaboration with the ImmunoPreCept Cluster of Excellence.
Disease almost never develops overnight. If illnesses are detected in later stages, however, they often prove difficult to treat - and the same holds true for stomach cancer. Years before the condition sets in, the stomach lining gradually begins to change. The most frequent trigger: The common stomach bacterium Helicobacter pylori, which is primarily transmitted within families. If parents or grandparents have had stomach cancer caused by a Helicobacter infection, their descendants are also at high risks of developing the disease.
The goal of the research team led by Prof. Michael Sigal and Dr. Manqiang Lin at the Department of Hepatology and Gastroenterology at Charité is to intervene in a timely manner in such cases. "We encounter precisely these kinds of people at the clinic, and they are often very anxious. If we understand what happens in the stomach tissue even before cancer develops, we might be able to intervene," explains Michael Sigal, professor of translational gastrointestinal oncology. He is a member of the recently launched Berlin Cluster of Excellence ImmunoPreCept, which is dedicated to identifying disease-causing processes and halting them before it is too late. In this context, researchers from the cluster and other institutions have joined forces to, among other things, determine what occurs in the gastric mucosa at the cellular and molecular levels when it is exposed to a Helicobacter pylori infection over extended periods of time.
The fact that the stomach bacterium promotes the development of cancer has long been known. But what types of cells are involved? In what order do they relay signals? And which of these signals is the most important one? "We knew from previous studies that the bacterium disturbs the balance of growth signals in the stomach lining." What we were missing was the link between the inflammation that had been triggered and the tissue remodeling. This is because the mucous membrane does not behave the way one would actually expect. It grows rapidly, though without the classic stem cells multiplying in the process. "So, there must be another program at work that has been overlooked until now," concludes Michael Sigal.
A sequence of events like in a sophisticated game of chess
Consequently, the researchers harnessed single-cell sequencing to analyze tens of thousands of cells from the diseased stomach lining individually and determined, for each cell type, which genes are turned on and off during an infection. In order to identify the cause and effect, they opted for animal models in which specific genes in certain cells were selectively inactivated. Tiny, laboratory-grown miniature versions of the gastric mucosa, known as organoids, and assembloids—combinations of mucosal and connective tissue cells developed specifically for this purpose—were used in the subsequent course of the investigations to simulate communication between the different cell types and to selectively interrupt this communication. In this way, the researchers were able to observe which cells in the tissue are actually located adjacent to one another and exchange signals. They then used publicly available datasets to verify whether these findings also apply to human tissue.
The research team encountered an amazing program that runs like a perfect game of chess. Each step paves the way for the next, and the chain culminates in precancerous lesions and cancer, as Giulia Beccaceci, first author and early-career researcher in Michael Sigal’s research group, explains: "The infected stomach lining doesn't just grow faster, but actually transitions into a fundamentally different state. A process is gradually activated that normally occurs only during embryonic development and wound healing." This is the precise reason why the tissue undergoes such lasting changes.
The process sets in when the stomach bacterium overcomes the stomach's natural defense system. A signal fails to be transmitted that normally ensures the controlled renewal of the surface cells of the mucous membrane and their tolerance of bacteria. Consequently, the immune system is alerted; immune cells migrate to the site and produce pro-inflammatory mediators, primarily interleukin-1β. This, in turn, does not affect the mucous membrane itself, but rather the underlying connective tissue. From there, the decisive signal is finally transmitted: A tissue hormone that normally helps heal injuries switches the mucous membrane into repair mode. As a result, the cells grow uncontrollably and divide more frequently. “So connective tissue isn’t just a bystander, but the actual switch,” as Giulia Beccaceci concludes.
The progression to the disease can be halted
Now that the progression from a Helicobacter pylori infection through chronic gastritis and on to precancerous lesions and cancer is understood for the first time, the question remains: How can diseases be effectively prevented? “Persons with an increased risk—for example, due to a family history of stomach cancer, symptoms, or a known infection—should be tested for the stomach bacterium and, if the test is positive, treated with antibiotics,” says Michael Sigal. "This is simple and has been proven to reduce the risk of stomach cancer. This is due to the fact that it removes the persistent irritation from the tissue that keeps the disease-causing chain of changes in motion."
The gastric mucosa, however, does not fully return to normal after antibiotic treatment in all affected individuals. In some cases, the tissue has already been permanently reprogrammed, and the risk of cancer development remains. “In order to reliably identify these individuals, we are currently developing markers that can detect, in mucosal tissue samples, whether the tissue is on its way to becoming precancerous,” explains Dr. Sigal. "In addition, we now understand the individual steps in the chain and, as a result, the potential targets for drugs." Consequently, this triangular communication via connective tissue could be disrupted by making these cells unresponsive to the inflammatory mediator interleukin-1β. "That would prevent tissue changes and abnormal mucosal growth."
The researchers are encouraged by findings from earlier population studies, which showed that regular use of common anti-inflammatory drugs was associated with a lower risk of gastrointestinal tumors. These drugs also block a component of the now-discovered pathway: the enzyme COX-2. According to this, the principle works, and the new findings help to identify ways to interrupt the chain in an even more targeted manner and with fewer side effects.
Molecular Prevention: Help Before People Get Sick
Today, cancer prevention mainly means early detection, in other words, finding existing tumors at the earliest possible juncture. For Michael Sigal and his team, prevention sets in much earlier Рnamely exactly when cancer has not yet developed, but the tissue is already on its way to becoming cancerous. In order to better care for people at elevated risk and to conduct scientific research on their tissue samples, the team is currently establishing a Cancer Prevention Clinic at Charit̩. Together with researchers from the ImmunoPreCept Cluster of Excellence, the goal will be to gain an ever-better understanding of the stage when cancer has not yet developed.
"While we now know that cancer can be prevented if interventions are taken early enough, it remains to be seen, however, whether a tissue that has already been reprogrammed can return to its normal state," as Michael Sigal states. This is a crucial question that researchers are already working on. The knowledge gathered is intended to result in a prevention strategy that will help identify persons at increased risk for stomach cancer, quantify the condition of their tissue, and specifically halt or reverse the transition to the disease. In this context, the Cancer Prevention Clinic is the point of contact for patients.
*Beccaceci G et al. Helicobacter pylori triggers gastric mucosal remodeling toward a fetal-like transcriptional program via stromal IL-1β signaling. Nat Commun 2026 Sep 16. doi: 10.1038/s41467-026-77520-1
About the Study
In addition to scientists from the Department of Hepatology and Gastroenterology at Charité, researchers from the Berlin Institute for Medical Systems Biology (BIMSB), the Max Delbrück Center for Molecular Medicine (MDC), the ImmunoPreCept Excellence Cluster, the Institute for Experimental Internal Medicine at Otto von Guericke University Magdeburg, the Max Planck Institute for Infection Biology, Berlin, the Berlin Institute of Health (BIH) at Charité, the BIH Charité Clinician Scientist Program, and the Berlin School of Integrative Oncology (BSIO) at Charité. The research was funded by, among others, the European Research Council (ERC Starting Grant REVERT), the Deutsche Forschungsgemeinschaft, the Federal Ministry of Research, Technology, and Space, the Einstein Center 3R, and the ImmunoPreCept Excellence Cluster.
Journal
Nature Communications
Method of Research
Experimental study
Subject of Research
Lab-produced tissue samples
Article Title
Helicobacter pylori triggers gastric mucosal remodeling toward a fetal-like transcriptional program via stromal IL-1β signaling.
Article Publication Date
26-Sep-2026
Momentum: Turning cancer research into patient benefit across Europe
European conference about cancer research, healthcare, policy, patient organisations and industry in Stockholm 14-16 October
image:
Momentum cancer conference, by Swedish Cancer Society
view moreCredit: Swedish Cancer Society
How can Europe speed up the adoption of new cancer research advances, strengthen cross-border collaboration and ensure that scientific progress reaches patients sooner? European leading voices from cancer research, healthcare, policy, patient organisations and industry will gather in Stockholm for Momentum 2026. On 14-16 October they to explore how Europe can shorten the path from scientific discovery to clinical practice.
Confirmed speakers include Greg Simon, former Executive Director of the White House Cancer Moonshot Task Force, Joanna Drake, Deputy Director-General at the European Commission's Directorate-General for Research and Innovation, Cary Adams, CEO of the Union for International Cancer Control (UICC)and Elisabete Weiderpass, Director General, International Agency for Research on Cancer (IARC). (Full list speakers MOMENTUM)
The programme will also feature a panel of leading journalists providing commentary and reflections throughout the conference, including Sarah Neville, Global Health Editor at the Financial Times; Alexandra Ivanov Hökmark, Associate Editor at Dagens industri; and Helena Smolak, Pharma Correspondent at Handelsblatt.
Journalists are welcome to attend. Accreditation is free of charge, but advance registration is required. To register, please complete the form here Register | MOMENTUM
The conference Momentum is convened by the Swedish Cancer Society The Swedish Cancer Society | Cancerfonden as the organisation marks 75 years of supporting advances in cancer research and patient care.
--
From the program: Across Europe, world-class cancer research discoveries are being made – yet too often they take too long to reach patients. The gap between research, decision-making, and real-world impact remains one of our greatest missed opportunities. Momentum brings together a carefully curated group of leaders from across sectors: policymakers, healthcare leaders, patient representatives, researchers, civil society, industry and investors - all with a shared ambition: to accelerate Europe’s action on cancer.
"The bottleneck is time. Every year we spend moving a discovery through regulatory, reimbursement, and clinical trial bureaucracies is a year patients don’t have. The next breakthrough won’t come from a single lab – it will come from collapsing the distance between discovery and patient. [...] I want to leave Sweden with at least one conversation that changes how I think about something. Not panels - conversations." (Greg Simon, Former Executive Director, White House Cancer Moonshot Task Force)
The Swedish Cancer Society | Cancerfonden is an independent, non-profit organisation dedicated to beating cancer. By funding the highest quality research, spreading knowledge about cancer, and influencing decision-making in key areas, we work to ensure that fewer people are affected and more people survive. We have contributed with 18 billion SEK to Swedish cancer research since the start. As a result of progress in research, cancer survival rates have more than doubled in Sweden - today, seven out of ten people diagnosed with cancer survive. We have come far, but there is still more to be done.
Autonomous AI could create capacity for thousands more dermatology appointments, real-world study finds
A real-world study of 8,391 patients found that autonomous artificial intelligence (AI) could free up enough clinical capacity to provide more than 8,500 additional face-to-face dermatology appointments across two UK hospitals over 16 months
(Vienna, Austria, Wednesday, 30 September 2026) A real-world study of 8,391 patients found that autonomous artificial intelligence (AI) could free up enough clinical capacity to provide more than 8,500 additional face-to-face dermatology appointments across two UK hospitals over 16 months.1
The findings, presented today at the European Academy of Dermatology and Venereology (EADV) Congress 2026, suggest that an autonomous AI as a medical device (AIaMD) could help address growing pressure on dermatology services. By identifying patients with benign lesions referred on urgent suspected skin cancer pathways, the technology could enable safe management without specialist review, allowing dermatologists to focus their expertise on patients who need it most.
Urgent suspected skin cancer referrals in England have almost tripled since 2009, yet around only 6% result in an urgent skin cancer diagnosis.2 At the same time, approximately one in four dermatologist roles in the UK is unfilled, creating a growing imbalance between demand and specialist capacity.3
The autonomous AI-supported pathway managed 8,391 patients, representing 94% of urgent suspected skin cancer referrals across the two hospitals. Overall, 86% of patients consented to autonomous decision-making. According to the researchers, this represents the first large-scale dataset from a prospective real-world deployment of autonomous AI within a cancer pathway.
Following an initial validation period, a CE-marked Class III AI medical device was introduced across the two sites. Using clinical and dermoscopic smartphone images, the system classified skin lesions, autonomously discharging benign cases while referring higher-risk cases for teledermatologist review.
After exclusions, the AI autonomously discharged 31% and 25% of patients at the two hospitals without clinician review. Teledermatologists subsequently discharged a further 24% and 25%, respectively.
Compared with standard teledermatology, the autonomous pathway reduced the proportion of patients requiring routine follow-up from 27% to 12%. Biopsy rates were also lower at 27% compared with 43% for conventional face-to-face care.
Overall, the autonomous pathway was estimated to save 2,851 hours of clinician time compared with a traditional face-to-face pathway, representing an approximate 62% gain in clinical capacity. Based on 20-minute consultations, this was equivalent to more than 8,500 additional face-to-face appointments over the 16-month study period.
Dr Lucy Thomas, lead author, explained, "We believe the greatest value of autonomous AI lies not in the technology itself, but in the specialist capacity it unlocks. Every hour saved reviewing low-risk lesions can be reinvested in patients with skin cancer, helping them access timely treatment to improve prognosis, and in patients with severe inflammatory skin disease, where earlier access to specialist care and effective treatments can transform quality of life.”
Ongoing safety monitoring of the autonomous pathway was also a key part of the study. In a national dataset including both study sites, sensitivity exceeded 98% for invasive melanoma, squamous cell carcinoma (SCC) and basal cell carcinoma (BCC), with a specificity of 72.1%. Six false-negative cases were discharged from the pathway – five basal cell carcinomas and one melanoma in situ – which were identified through post-market surveillance, with no adverse outcomes identified within the available follow-up.
“One of the key lessons for us is that deploying an AI system safely isn't a one-off exercise,” said Dr Thomas. “You need to keep monitoring it, understand when things go wrong, learn from those cases and make sure patients themselves know what to look out for.”
“If these findings are replicated across larger populations and different healthcare settings, autonomous AI could become an important part of creating a more sustainable dermatology service – not by replacing dermatologists, but by allowing scarce specialist expertise to be focused where it can make the greatest difference to patients’ lives,” Dr Thomas concluded.
END
Note to editors:
A reference to the EADV Congress 2026 must be included in all coverage and/or articles associated with this study.
For more information or to arrange an expert interview, please contact press@eadv.org.
About the study authors:
Dr Lucy Thomas is a consultant dermatologist at Chelsea & Westminster Hospital NHS Foundation Trust and honorary clinical lecturer at Imperial College London. A recognised leader in digital dermatology and artificial intelligence, she has led the development of pioneering teledermatology and AI-enabled skin cancer services within the NHS, including the world's first autonomous AI skin cancer pathway. She has received support from L'Oréal through an innovative NHS-industry partnership focused on service transformation and digital innovation. Her work has received national recognition, including HSJ Awards for improving cancer outcomes and clinical safety, and a National Clinical Impact Award. Dr Thomas is lead co-author of the NHS AI in Dermatology Implementation Toolkit and advises on digital dermatology strategy and innovation. She also serves as an independent consultant and sits on the Clinical Advisory Board for Skin Analytics Ltd.
About the study:
The study used a CE-marked AI medical device funded by Chelsea & Westminster Hospital NHS Foundation Trust and the study was partly supported by a grant from La Roche-Posay (L’Oreal Dermatological Beauty). La Roche-Posay’s mission has long centred on making dermatology more accessible, supporting people of all ages and skin tones, including those living with skin conditions or undergoing cancer treatment. Through this collaboration, the brand is supporting Trust-led research and education that builds on existing NHS innovation and helps advance skin cancer care responsibly.
About EADV:
Founded in 1987, EADV is a non-profit organisation with a vision to form a premier European Dermatology-Venereology Society. The Academy counts over 12,000 members from all around the globe, providing a valuable service for every type of dermatologist-venereologist professional. The EADV is dedicated to advancing patient care, education and research by providing a unique platform to bring people together and share ideas.
This year, the EADV Congress will take place in Vienna, Austria, and online from 30 September - 3 October 2026. Find out more: https://eadv.org/congress/
References:
- Thomas, L., Macedo, C., Fearfield, L. (2026, September 30 – October 3). Autonomous AI triage in urgent skin cancer pathways: Real-world safety, diagnostic performance and system impact in 8,391 patients. EADV Congress 2026, Vienna, Austria.
- Mitra, A., van Bodegraven, B., Venables, Z.C. (2026). P206 Conversion rates for urgent suspected skin cancer referrals in England 2009–2023. British Journal of Dermatology. 2026;195(Suppl 1). doi:10.1093/bjd/ljag086.233.
- Levell, N. (2021). Dermatology GIRFT Programme National Specialty Report. GIRFT, NHS England. Available at: https://gettingitrightfirsttime.co.uk/medical_specialties/dermatology/
New AI platform could help cancer survivors get personalized support between doctor visits
The framework is designed to personalize evidence-based support as survivors navigate persistent side effects and other challenges after treatment.
University of Miami Miller School of Medicine
image:
Frank Penedo, Ph.D., director of Sylvester’s Survivorship and Supportive Care Institute
view moreCredit: Sylvester Comprehensive Cancer Center
Researchers with Sylvester Comprehensive Cancer Center, part of the University of Miami Miller School of Medicine, have created a framework for developing, evaluating and implementing AI tools in cancer survivorship and supportive care, according to a new proof-of-concept study published in the journal Translational Behavioral Medicine.
The framework—Precision AI for Survivorship and Supportive Care—is designed to personalize evidence-based support as survivors navigate persistent side effects and other challenges after treatment while maintaining clinical oversight and safety guardrails.
“For many survivors, the hardest part isn’t always treatment itself. It’s what happens afterward, when patients are managing fatigue, anxiety, uncertainty and other challenges between clinic visits,” said study lead Frank Penedo, Ph.D., director of Sylvester’s Survivorship and Supportive Care Institute and Sylvester associate director for population sciences. “We’re exploring whether AI can help extend evidence-based support beyond the walls of the cancer center so survivors feel more connected, informed and supported throughout their journey, while not replacing in-person care and having access to their providers as part of their standard of care.”
The framework builds on years of survivorship research and clinical infrastructure developed at Sylvester. Through My Wellness Check, an electronic health record-integrated screening and triage program, patients routinely report symptoms, quality-of-life concerns, practical needs and other survivorship challenges before appointments.
Since expanding across the cancer center, the program has collected longitudinal patient-reported outcomes, supportive care needs and nutritional data from more than 37,000 ambulatory oncology patients.
Researchers also evaluated data from a subsample that included 25,592 of these ambulatory cancer survivors followed over 36 months to develop machine-learning models that identify patterns associated with symptom burden and unplanned health care utilization. Implementing these advanced analytic and AI-driven techniques improved predictive precision for unfavorable outcomes by over 25%. Together, those resources provide the foundation for developing more personalized supportive care strategies.
The team’s current work builds on the data and infrastructure behind My Wellness Check with an AI-enabled platform called My Wellness Support.
This platform combines patient-reported outcomes, clinical records and behavioral information to identify risk patterns and unmet needs. That information tailors evidence-based educational resources, symptom-management strategies, resources to address practical needs and supportive care recommendations to each patient.
Survivors will interact with an AI companion while the study team and clinicians monitor trends, risk scores and alerts through a dedicated dashboard.
“A patient may be doing well medically but still be struggling with symptoms, stress or questions that arise between appointments,” said Akina Natori, M.D., MSPH, a Sylvester oncologist and assistant professor in the Miller School’s Division of Medical Oncology. “The goal is not to replace those interactions with clinicians but to create another layer of support that helps identify concerns earlier and gives patients access to trusted, evidence-based information when they need it.”
Researchers say the platform is designed to operate within established clinical guidelines and safety guardrails, and maintain human oversight. The project remains in early testing.
“New technologies often move faster than the systems designed to evaluate them,” said Sara Fleszar-Pavlović, Ph.D., research assistant professor in the Miller School’s Division of Medical Oncology and director of research operations for Sylvester’s Survivorship and Supportive Care Institute. “That’s why our framework emphasizes scientific validation, transparency and continuous evaluation. If AI is going to become part of adjunctive survivorship and supportive care, we have to ensure it is safe, effective and developed with a range of patient populations in mind.”
By pairing new technology with rigorous science, researchers hope to help cancer survivors receive the support they need between appointments while establishing a model for how AI can be responsibly integrated into survivorship and supportive care.
Journal
Translational Behavioral Medicine
Article Title
The potential of AI for adjunctive supportive care in cancer survivorship: opportunities, challenges, and the PRECISION-AI practical framework
Article Publication Date
25-Sep-2026
Turning to AI for health answers may signal anxiety, depression among young adults
image:
Yusen Zhai, Ph.D., presents at a National AI Literacy Day program.
view moreCredit: Yuaen Zhai
Young adults who use generative artificial intelligence to ask health-related questions are more likely to screen positive for clinically significant anxiety and depression than those not using the AI tool for health issues, according to a new study led by University of Florida researcher Yusen Zhai.
The study, one of the first large U.S. studies to examine the relationship between generative AI use and mental health, analyzed data from nearly 100,000 young adults. Researchers found that the odds of screening positive for clinically significant anxiety and depression were 52% and 46% greater for people who reported using generative AI tools like ChatGPT or Claude for health-related questions.
The findings show an association between the two behaviors. They do not establish that using generative AI causes anxiety or depression, or that people with anxiety or depression are necessarily more likely to use AI for health questions.
Zhai, Ph.D., an assistant professor in the UF College of Education, said he was intrigued by the findings because concerns about AI in education often center on how the students use it for assignments and productivity. His research focuses on AI, mental health and the ways technology can be used to identify and address mental health risks.
“We are learning that young people are turning to AI, especially generative AI systems, when they feel worried or vulnerable or are trying to understand their health symptoms,” he said. “We need to understand if there are underlying issues or associations, which is what the current study begins to address.”
Among people in the study who used generative AI for health-related questions, about 52% screened positive for anxiety, compared with about 43% among those who did not use AI for health questions. For depression, the figures were about 47% and 38%, respectively.
The association remained after researchers accounted for factors including age, race and ethnicity, biological sex and socioeconomic status.
Zhai said the results raise questions about what happens when people repeatedly turn to AI while they are worried about their health.
One possible pattern begins with uncertainty about symptoms. A person asks for reassurance from an AI system, receives an answer and then develops additional questions or uncertainties. That can lead to another question, followed by another answer and potentially another round of questioning.
“That kind of repeated cycle could reinforce rumination, health anxiety or other distress,” Zhai said.
The conversational nature of generative AI makes that cycle different from a traditional internet search, he said. A search engine generally provides links to information, while generative AI can synthesize information and respond directly to follow-up questions.
That interaction can feel persuasive because AI responses are often conversational and confident, even when the information is incomplete, inaccurate or misleading.
“The scary part to me is that young people may have difficulty identifying an incorrect piece of information when it is blended into an otherwise plausible response,” Zhai said.
The study is based on data collected from 2023 to 2024, relatively early in the widespread adoption of generative AI. These AI chatbots had only become more broadly available when the survey was conducted, Zhai said. As use of generative AI has expanded since then, the behavior may have become more common.
That makes the findings an early indication rather than a complete picture of how young people use AI for health information today, he said.
The direction of the relationship also requires future investigations. Young people experiencing anxiety or depression may be more likely to turn to AI as a first step in seeking information or support. At the same time, repeatedly seeking reassurance from AI could potentially reinforce worry for some users.
“It’s hard to say AI is bad or wrong,” Zhai said. “AI could be a way for students to seek help.”
He said the more important question is how people use the technology, and he recommends that young people pay attention to how they feel after using generative AI for health questions.
“If an interaction leaves someone feeling more informed and able to take a constructive next step,” Zhai said, “that may be different from an interaction that prompts a desire to repeatedly ask the same questions — and that awareness is part of developing AI literacy.”
Mental health professionals also may need to ask clients how they are using generative AI and help them evaluate information they receive from it, Zhai said.
Universities could incorporate similar guidance into broader AI literacy efforts, including helping students identify trustworthy health information and understand the limitations of AI-generated answers.
“The problem is not generative AI itself,” Zhai said. “It’s the ‘how’ question — how people use it.”
Journal
Journal of Counseling & Development
Article Title
Associations of Generative AI Use for Health Issues With Anxiety and Depression
