NUMERO: #20240589

Título: DEEP LEARNING MAY DIFFERENTIATE PLASMA CELL NEOPLASMS

Nome do Apresentador: Lucas Lacerda de SOUZA

Categoria do Trabalho: Painel de pesquisa científica (PPC)

Área Temática: Patologia Oral

Resumo: Objective: The study aims to assess the effectiveness of a deep learning-based model in assisting pathologists to differentiate between plasmablastic lymphoma (PL) and multiple myeloma (MM), challenging diagnoses in the pathology field.
Study Design: Using whole slide images from 18 patients (8 PL, 10 MM), the study annotated, segmented, and broke down the images into 35,577 patches of 299x299 pixels. These patches trained a deep learning model using the InceptionV3 pre-trained model and other machine learning and image processing open-source libraries.
Results: The InceptionV3 model achieved an accuracy of 93.59%, with a training split of 50%. Performance metrics included an F1-score of 0.93, precision of 92.96%, recall of 93.09%, and Cohen's Kappa of 86.06%, indicating substantial agreement. ROC curve analysis highlighted InceptionV3s excellent class separation ability with a score of 93.09%.
Conclusion: This study highlights the potential of incorporating advanced deep learning algorithms into the diagnostic process for plasma cell neoplasms. By enhancing accuracy in distinguishing PL from MM, the study showcases artificial intelligences role in improving pathology diagnostics, suggesting that future efforts should aim at dataset expansion and model refinement for better clinical application.

Autor 1:  Lucas Lacerda de SOUZA

E-mail 1:  [email protected]

Autor 2:   Alan Roger SANTOS-SILVA

E-mail 2:  [email protected]

Autor 3 :  Felipe Paiva FONSECA

E-mail 3:  [email protected]

Autor 4:  Ayyub ALZAHEM

E-mail 4:  [email protected]

Autor 5:  Marcio Ajudarte LOPES

E-mail 5:  [email protected]

Autor 6:  Ahmed HAGAG

E-mail 6:  [email protected]

Autor 7:  Pablo Agustin VARGAS

E-mail 7:  [email protected]