Ensemble based in transfer learning for cytological classification in pleural fluid - Takeaways - MDSpire

Ensemble based in transfer learning for cytological classification in pleural fluid

  • By

  • Frida López-Córdova

  • Hugo Vega-Huerta

  • Gisella Luisa Elena Maquen-Niño

  • Jaime Cáceres-Pizarro

  • Ciro Rodriguez

  • David Calderón

  • Juan Gamarra-Moreno

  • Percy De-la-Cruz-VdV

  • Luis Guerra-Grados

  • Santiago Moquillaza-Henríquez

  • Oscar Benito-Pacheco

  • Ivan Adrianzén-Olano

  • Mario Chauca

  • June 10, 2026

  • 0 min

Share

  • 1

    The study develops an ensemble deep learning framework for classifying pleural cytology images into malignant and non-malignant categories.

  • 2

    Three CNN architectures—ResNet50V2, DenseNet121, and InceptionV3—were combined using transfer learning to enhance diagnostic accuracy.

  • 3

    Data augmentation significantly improved model performance, with the best accuracy achieved at 96.2% on the local dataset using soft voting.

  • 4

    The ensemble approach demonstrated superior performance compared to individual models, addressing the challenges of manual cytological interpretation.

  • 5

    This automated solution has potential applications in low-resource healthcare settings, improving access to accurate cancer diagnosis.

Original Source(s)

Ensemble based in transfer learning for cytological classification in pleural fluid

  • by Frida López-Córdova, Hugo Vega-Huerta, Gisella Luisa Elena Maquen-Niño, Jaime Cáceres-Pizarro, Ciro Rodriguez, David Calderón, Juan Gamarra-Moreno, Percy De-la-Cruz-VdV, Luis Guerra-Grados, Santiago Moquillaza-Henríquez, Oscar Benito-Pacheco, Ivan Adrianzén-Olano, Mario Chauca

  • June 10, 2026

Related Content

Reduced DPP9 levels sensitize experimental breast tumors to combinatory treatment with irradiation and Olaparib

BCOR-CCNB3 fusion–positive sarcoma treated with chemotherapy and carbon-ion radiotherapy: a case report of long-term disease control of a cervical spine case with a 12-year follow-up

Long-read sequencing for cancer liquid biopsy: advancing precision oncology