LUNG CANCER CELL CLASSIFICATION

FROM HISTOPATHOLOGICAL IMAGES

USING DEEP LEARNING
Presented by Presented To
NAME ID
Dr. Md. Ashraful Alam
Sharmin 24266010 Associate Professor
Akther Rima Dept of CSE
Md Shorif 24273005 Brac University
hossain
Nafisa 24273006
Tasnim
Abstract
Lung cancer classification is essential for determining treatment, as different types like
NSCLC and SCLC require unique approaches. However, variations in tumor appearance on
medical images make this challenging. Advanced methods, including machine learning, are
transforming diagnostics by identifying subtle patterns in imaging data. Our research
explores a novel approach that integrates multimodal imaging and AI to improve
classification accuracy, aiming to enhance personalized treatment strategies. This
presentation highlights current challenges, cutting-edge solutions, and findings from our
experiments, offering insights into future advancements in lung cancer diagnostics.
Literature Review [1/2]
Reference Algorithm Dataset Contribution Limitation
[1] CNN CT scan Accurate lung cancer Difficulties included
detection and comparisons, a small,
categorization are varied sample.
aided by deep CNN.
[2] DCNN microscopic images From cytological The precision obtained
pictures, Deep CNN is quite poor.
automates the
categorization of lung
cancer.
[3] VGG and ResNet Histopathological Autonomous DL The accuracy of the
images system detects cancer classification required
cell more work.
Literature Review [2/2]
Reference Algorithm Dataset Contribution Limitation
[4] GLCM, CCSA, PNN CT scan Improved Efficiency of
classification suggested classifier is
precision through hindered by outliers.
CCSA algorithm.
[5] CNN LC25000 dataset CNN accurately
Adding more
differentiated lung
elements could
cancer cell. enhance cancer
detection.
[6] CNN LC25000 dataset Used CNNs for Extensive training is
automating hampered by limited
inspection of biopsy computer.
images.
 Research goal
 To develop an accurate and efficient system for the classification and detection of lung cancer.
 To utilize deep learning models for improved diagnostic performance.
 To optimize model performance through hyperparameter tuning.
 To achieve high accuracy in detecting lung cancer, supporting early diagnosis and treatment.
Methodology [1/2]
The process of the recommended framework is shown in below Fig . The suggested system
covers –
■ Data preprocessing
■ Model selection (Deep learning)
■ Model training
■ Lung classification
■ And results for a deep neural network that classifies medical images.
Methodology [2/2]
Conclusion
This research will focus on developing advanced methods for lung cancer classification
using machine learning and multimodal data integration. Future work aims to address
current diagnostic challenges, enhance accuracy, and contribute to personalized treatment
strategies. The anticipated findings could provide valuable insights for improving outcomes
in lung cancer care.
References
1. Primakov, S.P., Ibrahim, A., van Timmeren, J.E. et al. Automated detection and segmentation of non-small cell lung cancer
computed tomography images. Nat Commun 13, 3423 (2022). https://doi.org/10.1038/s41467-022-30841-3
2. Baek, S., He, Y., Allen, B.G. et al. Deep segmentation networks predict survival of non-small cell lung cancer. Sci Rep 9,
17286 (2019). https://doi.org/10.1038/s41598-019-53461-2
3. Chao, H., Shan, H., Homayounieh, F. et al. Deep learning predicts cardiovascular disease risks from lung cancer screening
low dose computed tomography. Nat Commun 12, 2963 (2021). https://doi.org/10.1038/s41467-021-23235-4
4. Primakov, S.P., Ibrahim, A., van Timmeren, J.E. et al. Automated detection and segmentation of non-small cell lung cancer
computed tomography images. Nat Commun 13, 3423 (2022). https://doi.org/10.1038/s41467-022-30841-3
5. Shimazaki, A., Ueda, D., Choppin, A. et al. Deep learning-based algorithm for lung cancer detection on chest radiographs
using the segmentation method. Sci Rep 12, 727 (2022). https://doi.org/10.1038/s41598-021-04667-w
6. Baek, S., He, Y., Allen, B.G. et al. Deep segmentation networks predict survival of non-small cell lung cancer. Sci Rep 9,
17286 (2019). https://doi.org/10.1038/s41598-019-53461-2
7. Chaunzwa, T.L., Hosny, A., Xu, Y. et al. Deep learning classification of lung cancer histology using CT images. Sci Rep 11,
5471 (2021). https://doi.org/10.1038/s41598-021-84630-x