Enhancing Transfer Learning Model Performances using Cohen's Kappa Score at Predicting Breast Cancer
Keywords:
Cancer statistics, Deep Learning, Transfer Learning, Image Classification, EfficientNetB7Abstract
Over the past twenty years, cancer has been responsible for a significant number of deaths across the globe. A recent survey was conducted on cancer deaths from 1990 to 2018, where breast cancer (BC) found as the highest death rate, making it a major concern. Thus, minimizing the death ratio can only happen when predicting with a good model at the initial stage. To address this issue, a novel technique using the deep learning-based transfer learning (TL) method has been proposed here to predict this cancer in minimal time as this model has been pre-trained with a lot of images from the ImageNet dataset. Additionally, in this study, we have taken nine different TL models for cancer classification into either benign or malignant classes using 40 epochs at the training phase and compared their performances using predefined performance metrics. Furthermore, to increase the computational time and increase the accuracy of TL models, we have implemented the ‘Adam’ optimizer during the training phase. The result shows that the EfficientNetB7 model outperforms with a maximum accuracy of 98%, precision of 97.45%, f1-score of 97.42%, and a recall of 97.42% than other models. However, to make better decisions regarding the performance of each model, we further evaluated Cohen's kappa score (CKS) statistical method which specifically indicates how accurately our model can identify the cancer. These experiments have been carried out using Python 3.8.3 programming software on the Jupyter 6.4.3 Notebook application in the Windows 10 operating system.
Downloads
References
https://www.who.int/news-room/fact-sheets/detail/breast-cancer
S. Suhani et al., “An audit of over 1000 breast cancer patients from a tertiary care center of Northern India,” PubMed Breast Disease, Vol. 39(2), pp 91-99, 2020. DOI: 10.3233/BD-190435, PMID: 32310154.
S. Prusty et al., “SEMeL-LR: An improvised modeling approach using a meta-learning algorithm to classify breast cancer,” Engineering Applications of Artificial Intelligence, 129, 107630, 2024.
A. S. İlgün and V. Özmen, “The Impact of the COVID-19 Pandemic on Breast Cancer Patients,” European Journal Breast Health, Vol. 18(1), pp 85-90, 2021. DOI: 10.4274/ejbh.galenos.2021.2021-11-5, PMID: 35059596, PMCID: PMC8734528.
G. Vanni et al., “Delay in Breast Cancer Treatments During the First COVID-19 Lockdown. A Multicentric Analysis of 432 Patients,” Anticancer Research, Vol. 40(12), pp 7119-7125, 2020. DOI: 10.21873/anticanres.14741. Epub 2020 Dec 7. PMID: 33288611.
V. Ozmen et al., “Factors affecting breast cancer treatment delay in Turkey: a study from Turkish Federation of Breast Diseases Societies,” European Journal Public Health, Vol. 25(1), pp 9-14, 2015. DOI: 10.1093/eurpub/cku086, PMID: 25096257, PMCID: PMC4304375.
M. D A. Bashar and N. Begam, “Impact of COVID-19 Pandemic on Cancer Screening in India: Current Situation, Challenges and Way Forwards,” Asian Pacific Journal of Cancer Care, Vol. 6(1), pp 145-150, 2021. DOI: 10.31557/apjcc.2021.6.S1.
BJS Commission Team. BJS commission on surgery and perioperative care post-COVID-19. Br J Surg 2021;108:1162–1180.
R. J Bleicher et al., “Time to surgery and breast cancer survival in the United States,” JAMA Oncol 2016; 2:330-9. Back to cited text no. 4.
M Mateo et al., “Time to surgery and the impact of delay on triple negative breast cancers and other phenotypes,” J Clin Oncol 2018; 36 15 Suppl: e12606.
S. Prusty, S. Patnaik, S. K. Dash, “SKCV: Stratified K-fold cross-validation on ML classifiers for predicting cervical cancer,” Frontiers in Nanotechnology, 2022, 4, 972421.
W. Karl, M. Taghi and W. DingDing, “A survey of transfer learning,” Journal of Big Data, Vol. 3(9), 2016, DOI: 10.1186/s40537-016-0043-6.
S. Khan, N. Islam, Z. Jan, I. U. Din, and J. P. C Rodrigues, “A novel deep learning based framework for the detection and classification of breast cancer using transfer learning,” Pattern Recognition Letters, Vol. 125, pp 1-6, ISSN 0167-8655, 2019. DOI: 10.1016/j.patrec.2019.03.022.
R. Mehra, “Breast cancer histology images classification: Training from scratch or transfer learning?,” ICT Express, 4(4), 247-254, 2018.
W. Nawaz, S. Ahmed, M. Tahir and H. Khan, “Classification of Breast Cancer Histology Images Using ALEXNET,” Image Analysis and Recognition, pp 869-876, 2020. DOI: 10.1007/978-3-319-93000-8_99.
S. Prusty, P. Das, S. K. Dash, S. Patnaik, and S. G. P. Prusty, “Prediction of Breast cancer using integrated machine learning-fuzzy and dimension reduction techniques,” Journal of Intelligent & Fuzzy Systems, vol. 45, no. 1, pp. 1633-1652, 2023. DOI: 10.3233/JIFS-223265
G. S Bhattacharyya, D. C Doval, C. J Desai, H. Chaturvedi, S. Sharma, and S. P. Somashekhar, “Overview of breast cancer and implications of overtreatment of early-stage breast cancer: An Indian perspective,” JCO Glob Oncol; 6:789-98, 2020.
American Cancer Society, Cancer Facts & Figures 2020, no. 4, American Cancer Society, Atlanta, 2020.
H. Sung et al., “Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries,” CA Cancer J Clin 2021; 71: 209–49.
A. Gupta, B. S. Yadav, V. K. V. Gade, D. Rana, D. Bharti, and D. Dahiya, “Delays and disruptions in breast cancer care due to coronavirus disease-2019 pandemic: Real-World data from India,” Annals of Oncology Research and Therapy, 2022, 2(1), 23.
Indian Council of Medical Research. Advisory on Strategy for COVID-19 Testing in India (Version VI, dated 4th September 2020. Accessed February 04, 2021, at:
https://www.icmr.gov.in/pdf/covid/strategy/Testing_Strategy_v6_04092020.pdf
S. Prusty, S. K. Dash, S. Patnaik, et al., “EPD: an integrated modeling technique to classify BC,” In 2023 International Conference in Advances in Power, Signal, and Information Technology (APSIT) (pp. 651-655). IEEE.
S. Prusty, S. K. Dash, and S. Patnaik, “A novel transfer learning technique for detecting breast cancer mammograms using VGG16 bottleneck feature,” ECS Transactions, 107(1), 733, 2022.
K. Djunaidi, H. B. Agtriadi, D. Kuswardani and Y. S. Purwanto, “Gray level co-occurrence matrix feature extraction and histogram in breast cancer classification with ultrasonographic imagery,” Indones. J. Electr. Eng. Comput. Sci, 22(2), 187-192, 2020.
M. Tan and Q. Le, “Efficientnet: Rethinking model scaling for convolutional neural networks,” In International conference on machine learning (pp. 6105-6114), 2020, PMLR. https://arxiv.org/pdf/1905.11946v5.pdf.
S. Prusty, S. Patnaik and S. K. Dash, “EfficientNetB7: To predict lung and colon cancer at early stages,” In AIP Conference Proceedings (Vol. 2819, No. 1). AIP Publishing.
P. Zhang, Z. Ma, Y. Zhang, X. Chen, and G. Wang, “Improved Inception V3 method and its effect on radiologists’ performance of tumor classification with automated breast ultrasound system,” Gland Surgery, 10(7), 2232, 2021, doi: 10.21037/gs-21-328.
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.
