Have you heard about how Artificial Intelligence is changing healthcare and patient outcomes? It's fascinating! Machine learning and Deep-learning algorithms played a significant role when I started developing solutions to provide breast health examination, cervical screening, and accurate diagnosis to women in remote areas and underserved communities. 

Breast cancer and Cervical cancer statistics

Breast cancer is the most common cancer among women in India. Some of the statistics related to breast cancer could be scary and of great concern.

1. 1 out of 3 female cancer patients suffers from this Breast cancer in India. 
2. The incidence of breast cancer is increasing at an annual rate of around 2.5%, with 162,468 new cases and 87,090 deaths reported in 2020. 
3. The burden of breast cancer is higher in urban areas, with 1 in 22 women, while 1 in 60 women are affected in rural areas. 
4. Unfortunately, 70% of breast cancer cases in India are diagnosed at an advanced stage, increasing the cost of treatment to 10-16 times and lowering the likelihood of survival by 25%. 

The statistics related to cervical cancer are equally grim 

1. In India, 16 in 1000 women are at risk of acquiring cervical cancer, while 1 in 3 affected women die yearly. Globally, from 2018 to 2030. 
2. Deaths due to cervical cancer are expected to rise from 311,000 to 400,000. Cervical cancer is highly preventable, with early diagnosis improving survival rates. 
3. However, due to a lack of resources, this disease is vicious, and the cost of treatment burdens society considerably. 

Challenges for early screening and diagnosis of cancer

Implementing strategies for an early diagnosis like self-breast examination, clinical breast examination, or mammography for breast cancer, vaccination, and screening programs for cervical cancer detection are challenging for several reasons. 

• Awareness about these cancers and the importance of early detection is lacking among the women
• Rural and remote areas of India lack access to proper screening facilities or programs
• In India, there is only one radiologist available for every 100,000 people, and gynecologists are not enough to cater to the population
• Pain during advanced screening biopsy and radiation or chemotherapy treatment limits women from opting for advanced treatment
  The sensitivity of screening tests is limited due to technology constraints
• Cultural barriers pose a challenge in rural areas as women have reservations about testing
• The high cost of histopathology tests for cervical cancer and mammography for breast cancer leads to many rural women avoiding this screening procedure

Developing AI-based solutions to address these challenges

Keeping these challenges in mind, me and the team started developing a solution that is painless, accurate, easy to use, accessible to many women, and portable to remote areas. Successfully, we were able to develop pre-screening modalities for breast cancer and cervical cancer using AI technologies. I witnessed the benefits of AI technology at the grass-root level during its implementation for large-scale screening programs. The positive impact it had on women's lives was something that overjoyed me.

The technology made breast health examination cervical screening more streamlined, efficient, and patient-friendly. The utilization of machine learning and deep-learning algorithms was crucial in this context. For example, the breast cancer screening modality improved the accuracy of breast lump detection by 19% compared to clinical breast examination.

Fig. 1. Examples of breast cancer are detected with the aid of

A. Mammograms in a 47-year-old female with invasive ductal carcinoma. B. Heatmap and abnormality score are shown as in the viewer of the AI-based software. C. The patient was recalled by three of 10 radiologists when reading without AI assistance and by nine of 10 radiologists using AI-based software for support. AI = artificial intelligence. Adopted from Lee JH, Kim KH, Lee EH, et al. Improving the Performance of Radiologists Using Artificial Intelligence-Based Detection Support Software for Mammography: A Multi-Reader Study. Korean J Radiol. 2022;23(5):505-516. doi:10.3348/kjr.2021.0476

In Fig. 1, we can see the application of the trained deep learning-based software that was trained. The software provides an abnormality score and displays a heatmap indicating the abnormal region's location.

In this article, we will explore how Automated Visual Evaluation (AVE) enhances healthcare's clinical aspect and brings care to underserved regions. And in the following article will explore how AI is also revolutionizing how the healthcare industry operates.

Cervical cancer Diagnostics using Automated Visual Evaluation (AVE)

Developing the Automated Visual Evaluation (AVE) algorithm was a significant breakthrough in cervical cancer screening. AVE is an advanced image processing software that provides an objective and automated evaluation of cervical images as either control or cases. To develop the training model for AVE,  I looked at a vast dataset of images with their metadata and screening results. These datasets were used to create a proof-of-concept using the Faster R-CNN approach to deep learning object detection. 

By automating the cervical cancer screening process using AVE, medical professionals could efficiently and accurately triage women who required medical intervention and thus reduce the number of deaths caused by this disease.

Figure 2. The system architecture of the automated visual evaluation algorithm. Two models are trained: a cervix locator (top) and the automated visual evaluation detection algorithm (bottom). The final validation algorithm incorporated both the cervix locator and automated visual evaluation. Adopted from Liming Hu et al. 2019. An Observational Study of Deep Learning and Automated Evaluation of Cervical Images for Cancer Screening 

AI tools can analyze vast amounts of medical data and generate insights that would be difficult, if not impossible, for humans to identify by leveraging advanced algorithms and machine learning techniques. The application of AI in healthcare is still in its early stages, but the potential benefits are significant. As AI technologies continue to evolve and mature, we can expect to see even more significant improvements in clinical decision-making, patient care, surgical accuracy, and better patient outcomes.

To learn more about such use cases of how big data is being analyzed and implemented across industries, sign up with us to receive more such blogs. You may write to us and share your ideas, questions, and thoughts at mitra@setuschool.com

References

1. Lee JH, Kim KH, Lee EH, et al. Improving the Performance of Radiologists Using Artificial Intelligence-Based Detection Support Software for Mammography: A Multi-Reader Study. Korean J Radiol. 2022;23(5):505-516. doi:10.3348/kjr.2021.0476
2. Afsaneh E, Sharifdini A, Ghazzaghi H, Ghobadi MZ. Recent applications of machine learning and deep learning models in the prediction, diagnosis, and management of diabetes: a comprehensive review. Diabetol Metab Syndr. 2022;14(1):196. Published 2022 Dec 27. doi:10.1186/s13098-022-00969-9
3. Bhattacharya, S., Varshney, S., Heidler, P., & Tripathi, S. K. (2022). Expanding the horizon for breast cancer screening in India through artificial intelligent technologies -A mini-review. Frontiers in Digital Health, 4. https://doi.org/10.3389/fdgth.2022.1082884
4. Ginsburg O, Yip CH, Brooks A, et al. Breast cancer early detection: A phased approach to implementation. Cancer. 2020;126 Suppl 10(Suppl 10):2379-2393. doi:10.1002/cncr.32887
5. Singh M, Jha RP, Shri N, Bhattacharyya K, Patel P, Dhamnetiya D. Secular trends in incidence and mortality of cervical cancer in India and its states, 1990-2019: data from the Global Burden of Disease 2019 Study. BMC Cancer. 2022;22(1):149. Published 2022 Feb 7. doi:10.1186/s12885-022-09232-w
6. Liming Hu, David Bell, Sameer Antani, Zhiyun Xue, Kai Yu, Matthew P Horning, Noni Gachuhi, Benjamin Wilson, Mayoore S Jaiswal, Brian Befano, L Rodney Long, Rolando Herrero, Mark H Einstein, Robert D Burk, Maria Demarco, Julia C Gage, Ana Cecilia Rodriguez, Nicolas Wentzensen, Mark Schiffman, An Observational Study of Deep Learning and Automated Evaluation of Cervical Images for Cancer Screening, JNCI: Journal of the National Cancer Institute, Volume 111, Issue 9, September 2019, Pages 923–932, https://doi.org/10.1093/jnci/djy225