review on the research review committee at MBU tirupati

research

1. Mohan Babu University Sree Sainath Nagar, Tirupati - 517102 By: KURNUTALA NARESH 22202R010070 Research Scholar (Part-Time) Computer Science and Engineering School of Computing Guide: Dr. D.GANESH Associate Professor Computer Science and Engineering School of Computing Enhancing Pancreatic Cancer Detection with PanVCFS -Net: Integrating VGG19 and Chaotic Crow Search Optimization in CT Imaging

2. CONTENTS ABSTRACT INTRODUCTION LITERATURE REVIEW PROBLEM STATEMENT RESEARCH OBJECTIVES PROPOSED METHODOLOGY CONCLUSION AND FUTURE WORK REFERENCES 2

3. ABSTRACT we present PanVCFS -Net, a cutting-edge deep learning model designed to improve the identification and diagnosis of pancreatic cancer using computed tomography (CT) scans. The model makes use of the well-known VGG19 architecture for its powerful feature extraction capabilities, which are critical for detecting complicated patterns linked with pancreatic cancer. The Chaotic Crow Search optimization algorithm refines feature selection and model parameters to improve performance. A fully connected Softmax layer completes the classification process, predicting the likelihood of pancreatic cancer with great accuracy. The model processes a large dataset of 18,942 DICOM (Digital Imaging and Communications in Medicine) images from The Cancer Imaging Archive (TCIA), using advanced image processing techniques like Dualistic Sub-Image Histogram Equalization (DSIHE) to improve contrast and Grey-Level Co-occurrence Matrix (GLCM) to extract detailed texture features. 3

4. ABSTRACT These processes are critical for making an accurate distinction between normal and cancerous tissues. PanVCFS -Net, which was trained and validated using a strict 70-30 split, has demonstrated exceptional diagnostic accuracy (99.7%), sensitivity (95.0%), and specificity (94.5%), outperforming existing diagnostic models and establishing itself as a highly effective tool in the early detection and precise diagnosis of pancreatic cancer, significantly improving patient outcomes through timely medical intervention. 4

5. INTRODUCTION The pancreas, a critical organ in the abdomen, is essential for digestion and hormone production [1-2]. Pancreatic cancer, which originates in the pancreas' tissues, is a particularly aggressive type of cancer that frequently remains unnoticed in its early stages due to its modest symptoms [3-5]. This form of cancer is renowned for its quick progression and poor prognosis, thus early detection and efficient treatment options are important to improve patient outcomes [6]. Pancreatic cancer poses considerable health hazards because of its high fatality rate and difficult treatment [7]. According to global data, pancreatic cancer is the twelfth most frequent disease worldwide, but it is also the seventh leading cause of cancer mortality [8-11]. This disparity emphasizes the disease's fatal nature, as it frequently remains undiagnosed until advanced stages, making timely management difficult [12-14]. Pancreatic cancer accounts for around 3% of all malignancies and 7% of all cancer deaths in the United States, highlighting the critical need for more effective diagnostic and therapeutic techniques [15]. 5

6. INTRODUCTION Recent advances in medical imaging and machine learning have begun to alter the landscape of cancer detection and therapy, particularly pancreatic cancer [16-17].Innovative diagnostic strategies, such as the use of deep learning models and advanced imaging techniques, have showed promise in detecting the disease at an earlier stage than traditional methods [18-20].These technologies can evaluate massive amounts of medical imaging data with great precision and speed, providing fresh hope for detecting tiny signals of early-stage pancreatic cancer that would otherwise go unreported [21]. This work introduces PanVCFS-Net, a unique deep learning model that aims to increase the accuracy and efficiency of pancreatic cancer diagnosis using computed tomography (CT) scans. PanVCFS-Net aspires to be a valuable tool in the battle against pancreatic cancer by combining advanced image processing techniques and powerful machine learning algorithms. This model uses the VGG19 architecture with Chaotic Crow Search optimization to improve feature extraction and classification accuracy, resulting in higher performance in preliminary testing. 6

7. LITERATURE SURVEY A residual 3D U-Net model was proposed by Ramaekers et al. [22] in 2024 for the identification of pancreatic cancer. The model achieved a remarkable specificity of 86%, accuracy of 98%, and an F1 score of 98%. By utilizing residual connections in a 3D U-Net architecture, this model effectively captures intricate spatial correlations in medical imaging data, essential for precise identification of pancreatic cancers. In that same year, Supriya and colleagues [23] investigated the use of conventional deep neural networks for the identification of pancreatic cancer. They achieved an accuracy rate of 98% and precision scores of 0.98. Although these metrics demonstrate strong performance, conventional deep neural networks sometimes encounter difficulties in generalizing, particularly when trained on small datasets, which is commonly the situation in medical imaging. Furthermore, the deep neural network models are prone to overfitting and may need the use of substantial hyperparameter tuning and data augmentation methods to achieve optimal performance. 7

8. LITERATURE SURVEY In 2023, Pradip et al. [24] presented a specialized lightweight deep learning model designed to achieve optimal performance in the detection of pancreatic cancer. While the paper emphasizes the advantages of a less computing burden and quicker inference times, lightweight models may occasionally sacrifice the depth of feature extraction, which could impact the reliability of cancer detection, particularly in situations with minor imaging differences. Furthermore, Gandikota et al. [25] integrated the tunicate swarm technique with a deep learning model, resulting in a remarkable level of accuracy, sensitivity, and specificity of 99.55%. The present methodology incorporates an optimization algorithm to improve the feature selection process of the model, so substantially boosting the performance of detection. 8

9. PROBLEM IDENTIFICATION The automatic detection and classification of pancreatic cancer diseases, there is a pressing need to develop robust deep learning models. The study seeks to accomplish four primary objectives: first, to construct a convolutional neural network (CNN) model for the automatic detection and classification of pancreatic cancer. Second, to comprehensively evaluate the performance of diverse deep learning models, including recurrent neural networks (RNNs) and transfer learning models such as AlexNet , VGG, Inception, ResNet , DenseNet , and Region-based faster R-CNN, in the detection and classification of pancreatic cancer diseases. Third, to extend the application of deep learning models by developing a vision transformer ( ViT ) for the automatic detection and classification of pancreatic cancer diseases, utilizing Grad-CAM for interpretability. Lastly, to investigate the influence of factors such as image resolution, network architecture, and training parameters on the proposed models' performance, comparing and analyzing their effectiveness using metrics such as accuracy, precision, and recall. 9

10. PROBLEM IDENTIFICATION 10

11. CONCLUSION 11 To summarize, PanVCFS-Net represents a tremendous achievement in the realm of medical imaging, notably in the early identification and diagnosis of pancreatic cancer using computed tomography. This model has performed admirably by leveraging the sophisticated VGG19 architecture for feature extraction and enhancing it with the Chaotic Crow Search method for optimal feature selection and parameter adjustment. The use of modern image processing techniques, combined with a thorough examination of a large DICOM image dataset, has allowed PanVCFS-Net to attain outstanding accuracy, sensitivity, and specificity values. The findings highlight the model's potential as a transformative diagnostic tool, capable of considerably improving clinical outcomes via early detection of malignant alterations in pancreatic tissue.

12. CONCLUSION The goal of this research is to broaden PanVCFS -Net's capabilities to include the detection of multiple types of pancreatic cancer, perhaps providing a more comprehensive diagnostic tool that can distinguish between different tumor types based on their imaging properties. Furthermore, applying this model to additional abdominal tumors could capitalize on its robust design, potentially leading to comprehensive diagnostic solutions that improve patient care across a broader range of gastrointestinal disorders. This future work will try to not only improve the model's accuracy and efficiency, but also to investigate its scalability and adaptability in dealing with various and complex cancer diagnoses. 12

13. REFERENCES 1.Rahib L, Smith BD, Aizenberg R, Rosenzweig AB, Fleshman JM, Matrisian LM. Projecting cancer incidence and deaths to 2030: the unexpected burden of thyroid, liver, and pancreas cancers in the United States. Cancer Res. 2014;74:2913–21. 2.Siegel RL, Miller KD, Fuchs HE, Jemal A. Cancer statistics, 2022. CA Cancer J Clin . 2022;72:7–33. 3.Agarwal B, Correa AM, Ho L. Survival in pancreatic carcinoma based on tumor size. Pancreas. 2008;36:e15-20. 4. De La Cruz, M.S.D.; Young, A.P.; Ruffin, M.T. Diagnosis and management of pancreatic cancer.  Am. Fam. Physician   2014 ,  89 , 626–632. 5.van der Geest , L.G.M.; Lemmens , V.E.P.P.; de Hingh , I.H.J.T.; van Laarhoven , C.J.H.M.; Bollen , T.L.; Nio , C.Y.; van Eijck , C.H.J.; Busch, O.R.C.; Besselink , M.G. Dutch Pancreatic Cancer Group Nationwide Outcomes in Patients Undergoing Surgical Exploration without Resection for Pancreatic Cancer.  Br. J. Surg.   2017 ,  104 , 1568–1577. 5. M. Griffin, S. N. Duggan, R. Ryan, R. McDermott, J. Geoghegan , and K. C. Conlon, ―Characterising the impact of body composition change during neoadjuvant chemotherapy for pancreatic cancer,‖ Pancreatology , vol. 19, no. 6, pp. 850–857, Sep. 2019, doi : 10.1016/j.pan.2019.07.039. 6.Treadwell, J.R.; Zafar, H.M.; Mitchell, M.D.; Tipton, K.; Teitelbaum , U.; Jue , J. Imaging Tests for the Diagnosis and Staging of Pancreatic Adenocarcinoma: A Meta-Analysis.  Pancreas   2016 ,  45 , 789–795. 13

14. 8.Gandikota HP, S. A, M. SK (2023) CT scan pancreatic cancer segmentation and classification using deep learning and the tunicate swarm algorithm. PLoS ONE 18(11): e0292785. https://doi.org/10.1371/journal.pone.0292785 9.Viriyasaranon T, Chun JW, Koh YH, Cho JH, Jung MK, Kim SH, Kim HJ, Lee WJ, Choi JH, Woo SM. Annotation-Efficient Deep Learning Model for Pancreatic Cancer Diagnosis and Classification Using CT Images: A Retrospective Diagnostic Study. Cancers (Basel). 2023 Jun 28;15(13):3392. doi : 10.3390/cancers15133392. PMID: 37444502; PMCID: PMC10340780. 10.Khdhir, Radhia , AymenBelghith , and SalwaOthmen . "Pancreatic Cancer Segmentation and Classification in CT Imaging Using Antlion Optimization and Deep Learning Mechanism." International Journal of Advanced Computer Science and Applications, vol. 14, no. 3, 2023, pp. 50-59 11.Chang, D., Chen, PT., Wang, P. et al. Detection of pancreatic cancer with two- and three-dimensional radiomic analysis in a nationwide population-based real-world dataset. BMC Cancer 23, 58 (2023). https://doi.org/10.1186/s12885-023-10536-8 12.Chen, Po-Ting, et al. "Pancreatic Cancer Detection on CT Scans with Deep Learning: A Nationwide Population-based Study." Radiology, vol. 306, no. 1, 13 Sept. 2022, https://doi.org/10.1148/radiol.220152. 13.Roth, H., Farag , A., Turkbey , E. B., Lu, L., Liu, J., &Summers, R. M. (2016). Data From Pancreas-CT (Version 2) [Data set]. The Cancer Imaging Archive. https://doi.org/10.7937/K9/TCIA.2016.tNB1kqBU S 14

15. PAPERS SUBMITTED FOR PUBLICATION 15 International Journal: “Leveraging high resolution remote sensing images for vehicle type detection using sparrow search optimization with deep learning ” submitted in “MULTIMEDIA TOOLS AND APPLICATIONS”, SPRINGER PUBLISHERS.

16. 16