Banca de QUALIFICAÇÃO: VITÓRIA TAVARES DE CASTRO

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
STUDENT : VITÓRIA TAVARES DE CASTRO
DATE: 12/08/2025
TIME: 14:00
LOCAL: Plataforma Teams
TITLE:

Adverse effects of radiotherapy in head and neck: imaging evidence and prediction using Machine Learning

KEY WORDS:

Head and Neck Cancer; Radiotherapy; Adverse Effects; Bone Changes; Osteoradionecrosis; Machine Learning

PAGES: 100
BIG AREA: Ciências da Saúde
AREA: Odontologia
SUMMARY:

This project aims to investigate diagnostic and predictive approaches for adverse effects of radiotherapy in the head and neck region, with a focus on bone changes and related toxicities. The methodology was divided into three chapters: (1) a systematic review of maxillomandibular bone changes post-radiotherapy, analyzing studies using panoramic imaging and computed tomography; (2) a systematic review and meta-analysis of the accuracy of computational models, especially Machine Learning (ML), for predicting osteoradionecrosis (ORN); (3) the development and validation of a predictive model based on ML, using clinical data from patients treated at a public hospital. In chapter 1, 30 studies with 2,441 patients showed that bone changes, including ORN (7% of cases), are common after radiotherapy. Chapter 2 analysed 10 studies with 5,255 patients, indicating that models such as Random Forest demonstrated good predictive accuracy for ORN (0.76), with traditional algorithms outperforming Deep Learning, although limited by heterogeneity. In chapter 3, preliminary data from 34 patients were collected, with plans to build predictive models for toxicities using multiple algorithms and performance metrics. The results indicate that advanced imaging techniques and ML hold great potential to improve early detection and prediction of radiotherapy-related toxicities in the head and neck region. In order for these tools to be effectively incorporated into clinical practice, further advances are still needed in methodological standardization, rigorous external validation, and the development of interpretable models that ensure greater reliability and applicability across diverse contexts. These innovations are expected to contribute to more personalized and effective treatment strategies, positively impacting patients’ quality of life and healthcare management.

COMMITTEE MEMBERS:
Externa à Instituição - ANA CAROLINA PRADO RIBEIRO E SILVA - UNICAMP
Interno - 3437282 - ANDRE FERREIRA LEITE
Presidente - 1278451 - ELIETE NEVES DA SILVA GUERRA
Externa à Instituição - FABIANA TOLENTINO DE ALMEIDA MARQUES - UNIALBERTA