Beyond Pink Ribbons: Rethinking Breast Cancer Awareness in the Age of AI

Breast cancer, notorious as the most common cancer globally, has also seen a sharp increase (39%) in India in the last two decades. ¹ It has leapfrogged to being the single most prevalent form of cancer, accounting for 13.6% of all cancers and responsible for around 10% of overall mortality. ² While early detection and implementation of definitive treatment are recognized as means to combat the disease burden globally, unique challenges are encountered in the Indian setup. General lack of awareness, poor penetration of screening strategies, economic challenges and paucity of resources leads to delayed initial presentation, incomplete or highly fragmented treatment, poor compliance and non-adherence to treatment protocols with high attrition rates.

While early detection remains key, the initial symptom more often than not is a painless lump, which is easily missed or overlooked, leading to significant delay in seeking treatment. Worldwide, screening programs have been constituted targeting women in the 40-50 years age group.³ Mammography is the single most important tool in this regard. However, in India, paucity of skilled resources and facilities makes this a herculean task.

Artificial Intelligence has been revolutionizing healthcare as a whole. While the conventional Computer aided Detection (CAD) in digital mammograms was a form of machine learning that could pick up specific patterns like masses or calcifications, it was meant more as a “spellcheck” for the radiologist and is as time-consuming as a normal mammogram. However, it is the application of Deep Learning (DL), a subset of machine learning that mimics human learning, and is able to analyze and interpret vast data sets, that is piquing researchers’ interest. The availability of a huge quantum of retrospective data allows these programs to be trained quickly to interpret images accurately. While millions of neurons function in the human brain by creating “networks”, Artificial Neural Networks (ANN) function similarly by creating multi-layered connection maps, and with the addition of a “pooling layer” for the information to be analyzed, allowing for self-reasoning and abstraction and interpretation of datasets.  Research today has shown that a combined human and AI analysis can increase the sensitivity and specificity of breast cancer detection.  It can reduce the workload of radiologists by picking up “routine lesions” and allowing them to focus on more complex ones. It can also be used to detect subtle and interval lesions. ⁴ The amalgamation of AI into mammography as a part of a larger public health initiative could thus pave the way for the future of mass screening programs and revolutionize breast cancer detection.

Tackling breast cancer is however a multi-faceted problem. In India, most patients present when the lump is significantly large in size or ulcerated, by which time the disease staging is much more advanced, which contributes to our poor survival rates compared to the global norms.⁵ There is also significant delay in initiation of treatment, probably due to the high patient load impacting an overburdened healthcare system. Add to this the worries of social isolation and embarrassment, taboos associated with disease in “private areas” especially in rural areas and the significant economic burden of completing a cancer treatment, we face a requirement of tailor-made solutions for our ecosystem.

A study in Mumbai noted that just a simple breast examination by a primary health care worker once in two years could pick up more cases of breast cancer.⁶ In reality, the first point of contact for a patient may not be the oncologist in the city, but the local Anganwadi worker. As the world moves to a risk-based screening strategy, the need of the hour is to identify risk factors unique to the Indian setup and to equip the healthcare system to recognize them, right up to the grassroots level. AI could be the key in studying risk factors and drafting easily accessible patient recognition protocols and pathways to complete treatment.

The World Health Organization, on March 9, 2021, introduced a Global Breast Cancer Initiative to reduce global breast mortality by 2.5% by 2040.⁷ Artificial Intelligence may just be the shot in the arm that could empower the Indian healthcare system to combat the challenges of breast cancer in a proactive manner.

REFERENCES

1. India State-Level Disease Burden Initiative Cancer Collaborators. The burden of cancers and their variations across the states of India: the Global Burden of Disease Study 1990-2016. Lancet Oncol. 2018;19:1289–1306. [PMC free article] [PubMed] [Google Scholar]

2. International Agency for Research on Cancer. India Source: Globocan 2020.. Available from: https://gco.iarc.fr/today/data/factsheets/populations/356-india-fact-sheets.pdf .

3. Lehman CD, Arao RF, et al. National performance benchmarks for modern screening digital     mammography: update from the Breast Cancer Surveillance Consortium. Radiology 2017; 283: 49.

4. Chan HP, Hadjiiski LM, Samala RK.  Computer‐aided diagnosis in the era of deep learning. Med Phys 2020; 47: e218–e27

5. Maurya AP, Brahmachari S. Current Status of Breast Cancer Management in India. Indian J Surg. 2020

6. Mittra I, Mishra GA, Dikshit RP, Gupta S, Kulkarni VY, Shaikh HKA, Shastri SS, Hawaldar R, Pramesh CS, Badwe RA. Effect of screening by clinical breast examination on breast cancer incidence and mortality after 20 years: prospective, cluster randomised controlled trial in Mumbai. BMJ. 2021;372:n256

7. DownToEarth WHO launches renewed efforts to reduce breast cancer mortality [Internet]. 2021. Available from: https://www.downtoearth.org.in/news/health/who-launches-renewed-efforts-to-reduce-breast-cancer-mortality-75839 .