Introduction: Cervical cancer is one of the leading causes of female mortality worldwide. Early diagnosis by Pap smear is essential, but it still suffers from limitations in regions with few resources and subjectivity in the interpretation of results. Thus, Artificial Intelligence (AI) has stood out for its ability to analyze data accurately, being an alternative method and aiding in the early identification of precancerous and cancerous lesions. Objective: To develop an artificial intelligence capable of performing automated reading and assisted diagnosis of cytopathological slides, with an emphasis on detecting cellular changes associated with cervical cancer in the city of Belém, state of Pará. Methodology: Thirty images of Pap smears, labeled as positive or negative diagnoses, extracted from a clinical database were used. The images underwent preprocessing for standardization, and data augmentation was performed dynamically with the ImageDataGenerator tool (Keras). Then, the Convolutional Neural Network model was trained and evaluated with metrics such as accuracy, sensitivity, F1-score, in addition to the analysis of theROC curve and AUC. Results: This section presents the results obtained by applying the two proposed approaches: the simple CNN built from scratch and themodel based on transfer learning with MobileNetV2. The evaluation metrics were calculated based on the testset, which comprises 20% of the total images (6 images in total, 3 positive and 3 negative). The second approach evaluated the application of transfer learning, using the MobileNetV2 architecture pre-trained on theImageNet base. Only the upper layers were adjusted for the binary classification task. This model presented better overall performance, with less variation in themetrics and stability in the training process, which was completed in 22 epochs. Conclusion: The AI was successfully assembled to read cytopathology slides to aid in the early diagnosis of cervical cancer. Thus, this approach has the potential to revolutionize the early detection of neoplasms. The expectation is that AI will become an increasingly valuable tool to increase the chances of effective treatment, increasing diagnostic accuracy, reducing analysis time and expanding access to screening.
It is with great enthusiasm that we present the Annals of the Oncology International Symposium 2025, an event that continues to solidify its significance in the oncology landscape of northern Brazil. Held in Belém, Pará, Oncology 2025 centered around the theme "The cancer control challenge: better knowing it to best facing it," dedicating itself to exploring the latest frontiers in cancer treatment and prevention.
This year, the symposium provided a deep dive into the essential role of knowledge in the fight against cancer, presenting new perspectives and scientific advancements across various areas of oncology. Renowned global experts gathered to share their most recent research and innovative approaches, offering participants a comprehensive view of the challenges faced by healthcare professionals and patients worldwide.
Presentations and discussions during the event focused on critical topics such as the use of new technologies, advancements in personalized therapies, and more effective prevention strategies. Additionally, particular attention was given to the unique challenges faced by the Amazon region, with efforts aimed at developing region-specific solutions to meet local needs.
Beyond being a high-caliber academic event, Oncology 2025 stood out as a moment for integration and professional networking, with the warm hospitality of the city of Belém offering participants a unique experience. This event became a platform for exchanging ideas, where science, culture, and humanity came together in pursuit of a common goal: to improve cancer control both in Brazil and globally.
This collection of abstracts and articles presented during the event reflects the ongoing dedication to research and the development of innovative solutions, highlighting the importance of collaboration and shared knowledge in the fight against cancer.
General Submission Guidelines:
The presenting author, who does not have to be the first author, must be registered for Oncology 2025.
Each abstract may have up to 10 authors, including the main author and co-authors.
Only original, unpublished work will be accepted.
Submissions must be related to oncology. However, project descriptions, work proposals, experience reports, and literature reviews will not be considered.
Clinical case reports are allowed, provided the abstract addresses scientific questions, details clinical observations, and includes primary scientific data.
The abstract must be written in English, but presentations may be given in Portuguese.
Abstracts must be between 300 and 500 words.
Comissão Organizadora
Comissão Científica
See Annals of Oncology 2023 at:
https://www.even3.com.br/anais/oncology-2023-international-symposium/
Complete tool for submission, evaluation and publication of proceedings for scientific events.
