 |
| Image Credit: MIT |
That’s key, MIT notes in a blog post, because black women are more than 42 percent more likely than white women to die from breast cancer, and one contributing factor could be that they aren’t as well-served by current early detection techniques. MIT says that its work in developing this technique was aimed specifically at making the assessment of health risks of this nature more accurate for minorities, who are often not well represented in development of deep learning models. The issue of algorithmic bias is a focus of a lot of industry research and even newer products forthcoming from technology companies working on deploying AI in the field.
This MIT tool, which is trained on mammograms and patient outcomes (eventual development of cancer being the key one) from over 60,000 patients (with over 90,000 mammograms total) from the Massachusetts General Hospital, starts from the data and uses deep learning to identify patters that would not be apparent or even observable by human clinicians. Because it’s not based on existing assumptions or received knowledge about risk factors, which are at best a suggestive framework, the results have so far shown to be far more accurate, especially at predictive, pre-diagnosis discovery.
 |
| “Neural network” models of AI process signals by sending them through a network of nodes analogous to neurons. Signals pass from node to node along links, analogs of the synaptic junctions between neurons. “Learning” improves the outcome by adjusting the weights that amplify or damp the signals each link carries. Nodes are typically arranged in a series of layers that are roughly analogous to different processing centers in the cortex. Today's computers can handle “deep-learning” networks with dozens of layers. Image credit: Lucy Reading-Ikkanda (artist). |
Worldwide, the cancer of the cervix (lower portion of the uterus) is the fourth most common cancer. It is also one of the most common causes of deaths due to cancer in women.
Most of my patients that participated in my public health project had either dementia, Alzheimer's or were frail and sometimes immobile. They would forget their surroundings, spouse name and even getting a regular medical checkup was a challenge. These women, when asked to go for cervical cancer diagnosis, opted out and never showed up. Most of these tests are widely available but are uncomfortable and invasive. Patients are also not keen to go for them unless indicated.
1 in 5 cancer patients across the world experience delay in diagnosis and, it holds true for cervical cancer as well. Cervical cancer is diagnosed more frequently at more advanced stages.
The human papillomavirus (HPV) infection is responsible for 90 percent cases. However, all women infected with this virus will not develop cervical cancer. Of the 150-300 known strains of HPV, 15 are classified as high risk for causing cervical cancer. Other risk factors include a weak immune system as due to HIV infection, malnutrition, having sex from an early age, multiple sex partners, multiple pregnancies, and smoking. Oral contraceptive pill use for a long time has been associated with increased risk of cervical cancer.
Training models (Deep Learning) is just the first of many steps in translating exciting research into a real product. A pathologist’s report after reviewing a patient’s biological tissue samples is often the gold standard in the diagnosis of many diseases. For cancer, in particular, a pathologist’s diagnosis has a profound impact on a patient’s therapy. The reviewing of pathology slides is a very complex task, requiring years of training to gain the expertise and experience to do well. Even with this extensive training, there can be substantial variability in the diagnoses given by different pathologists for the same patient, which can lead to misdiagnoses.
Studying and reviewing all the biological tissues visible on a slide is done by a Pathologist. There can be multiple slides per patient, and at 40X magnification each is more than 10+ gigapixels when digitized. There is a lot of data to cover, and often time is limited.
To address these issues of limited time and diagnostic variability, we are evaluating ways deep learning methods like automated detection algorithm can be applied to digital pathology, helping improve diagnosis and also complement pathologists’ workflow.
In various instances and research on the incidence of cervical cancer and ineffective diagnosis, the following results were found. Patients were either:
1) Adequately screened with normal results
2) Inadequately tested with normal results
3) Unscreened
4) Low-grade abnormality
5) High-grade abnormality
The microstructure of normal tissue is uniform but as the disease progresses the tissue microstructure becomes complex and different. Based on this correlation, Scientists have created a novel light scattering-based method to identify these unique microstructures for detecting cancer progression.
The morphology of healthy and precancerous cervical tissue sites are entirely different, and light that gets scattered from these tissues varies accordingly. It is difficult to evaluate with naked eyes the subtle differences in the scattered light characteristics of the normal and precancerous tissue. The AI (Artificial intelligence) identifies precancerous tissue, and also the stage of progression in minutes.
Cervical cancer diagnoses and deaths are predicted to rise among women over the age of 50. However, deaths from the disease among the young who have been vaccinated are likely to be almost eradicated thanks to school vaccination programs for HPV.
One such company to mention that we worked with at TechEmerge is MobileODT. MobileODT turns mobile technologies into intelligent visual diagnostic tools that enable any health provider, anywhere in the world, to conduct visual inspections at the level of an expert practitioner.
AI and deep learning methods give a promising start to discover the unknown, but there is a lot that still needs to be done. AI Algorithms from some of the most innovative companies across the world are now available for $1/scan and are targetted towards developing countries to improve diagnosis and save time.
Recommended reads:
https://www.digitaltrends.com/cool-tech/screening-ai-cervical-cancer/
https://www.omicsonline.org/open-access/novel-benchmark-database-of-digitized-and-calibrated-cervical-cells-forartificial-intelligence-based-screening-of-cervical-cancer-ccoa-1000105.php?aid=68453
https://www.cdc.gov/cancer/cervical/statistics/index.htm
https://www.mobileodt.com
https://www.pnas.org/content/116/4/1074
http://news.mit.edu/2019/using-ai-predict-breast-cancer-and-personalize-care-0507