Saving a Thyroid! The goal is twofold: to improve diagnosis of thyroid nodules classified as “indeterminate for malignancy” and to reduce the number of unnecessary thyroidectomies, often performed on young patients, mostly women aged between 20 and 40. This is the aim of “Let’s Save a Thyroid!”, a project developed within the BiUniCrowd crowdfunding programme of the University of Milano-Bicocca, already supported by Fondazione Cariplo and Thales Alenia Space.
Thyroid diseases are widespread and on the rise, with estimates suggesting that about half of the population will develop at least one thyroid nodule in their lifetime. Distinguishing between benign and malignant nodules—and treating them accordingly—remains a challenge. In 15–30% of cases, nodules are classified as “indeterminate,” leading to potentially avoidable surgeries. With the help of new biological biomarkers and artificial intelligence, the research team from Milano-Bicocca (and partners) aims to reduce diagnostic uncertainty, improve patients' quality of life, and optimise healthcare resources.
Specifically, by using artificial intelligence, it will be possible to automatically identify different cellular entities within a biopsy, thus assisting pathologists in evaluating cytological samples from thyroid nodules and reducing the number of indeterminate diagnoses.
To make this possible, the project is seeking support through a crowdfunding campaign on Ideaginger.it, the most successful fundraising platform in Italy. The initial goal is to raise €6,000, which will be used to strengthen computational infrastructure, develop an app to support clinicians, and organise scientific outreach events.
The “Let’s Save a Thyroid!” team includes pathologists, biotechnologists, statisticians, and computer scientists, led by Giulia Capitoli, researcher in medical statistics.
“Our aim is to improve patients’ quality of life by avoiding unnecessary thyroidectomies and reducing anxiety and costs linked to uncertain diagnoses,” says Capitoli. “At the same time, we want to make research more accessible by actively involving the public.”
Outreach is a key part of the project. A dedicated public event is planned for autumn 2025, during which participants will explore the team’s work from 2016 to the present through talks and hands-on activities. Visitors will take part in a diagnostic journey via an escape room, simulating the real work of researchers—from clinical evaluation with pathologists, to biomarker extraction with biochemists, image analysis with IT specialists, and the creation of predictive models with statisticians.
Alongside Capitoli, the team includes Vincenzo L’Imperio, Giorgio Cazzaniga, and Antonio Maria Alviano, physicians and researchers in pathological anatomy responsible for the clinical part of the project. Supporting them are Vanna Denti and Lisa Pagani, biotechnologists specialised in analysing biological samples to identify disease biomarkers. The resulting data will be processed by statisticians Giulia Capitoli, Giulia Risca, Francesco Denti, and Maria Francesca Marino, who will use predictive models to link biological features with clinical outcomes and integrate this into an app to support diagnosis.
Vasco Coelho, a computer scientist, is developing neural networks to aid clinicians in the detection, segmentation, and classification of thyroid tissue samples using AI-based image analysis. These tools will be integrated into the app. Finally, Sofia Martinelli, a communication specialist, will lead the promotion of the project and coordinate the crowdfunding campaign.