Pathology is in the Midst of a Digital Transformation
Its Successful Implementation will have far Reaching Implications in the World of Healthcare
Since the start of the millennial generation, advances in technology have helped take the world digital. The widespread adoption of personal computers, the internet and most recently, the smartphone have been the big watershed events that have completely transformed the way we live and work. There are few aspects of our life that are yet to radically reinvent themselves and maintain the digital status quo. So far, pathology has lagged, but not for long.
Pathology is the branch of medicine that deals with the laboratory examination of body tissue for diagnostic or forensic purposes. The gateway to analysing each cell within our bodies, and the interconnections between them has always been the microscope. A nifty tool, but one that has been largely unchanged for the last 150 years. Due to its intricate nature and extreme computational requirements, technologists have so far been unable to transform this field of medicine. However, rapid innovations in cloud computing and AI mean we are now ready to revolutionise the field.
The biggest benefactor of this revolution in its early days has been Cancer diagnostics. The timing could not be better as practitioners are pushed to the brink in trying to provide better patient outcomes.
Some of the burning needs for digital pathology in Cancer diagnostics today:
- Growing demand and reducing practice
- Improving reproducibility and accuracy
- Democratization of specialization
Growing Demand and Reducing Practice
In the US, there has been an ~20% decline in practicing pathologists from 2007 to 2017 with a subsequent ~40% increase in workload. The National Cancer Institute has indicated that 1 in 2 American males and 1 in 3 American females are at risk of developing Cancer in their lifetime.
Doctors are stretched thin for resources and this problem isn’t going to fix itself.
Improving Reproducibility and Accuracy
A pathologist’s job is very tough to do. Cancer is tough to diagnose as it does not follow any rules. The disease continues to morph in new ways that are tough to keep a track of, making diagnosis a subjective and highly specialised practice.
These difficulties are amplified multi fold when considering the large volume of data that needs to be processed and analysed in order to make a diagnosis.
Consider the image below. The pathologist needs to identify the cells that are cancerous (if any) and proceed to make various calculations such as the number of cancerous cells on a given slide.
Try counting the number of cells in this image. Definitely not impossible, but it would probably take a while. Additionally, the count each individual gets will be slightly differ when compared to the next. The count an individual would get were they to analyse the same slide twice would probably be a bit different too.
Now consider that this is just 1/20th (or 5%) of the actual slide, there are multiple slides to be analysed for a given patient, the count of cells is just one of the calculations or ‘morphometrics’ that need to be performed and the pathologist has no more than a minute or two to spend on a given slide.
AI solutions using Deep Learning algorithms aim to automate such tasks. Thus providing reproducible results coupled with a significant reduction in the time spent analysing a given slide.
Democratisation of Specialization
As mentioned earlier, Cancer rapidly mutates and continues to do so in ways that have not been seen before. Each organ (Lung, Oral, Breast, Skin) has many sub-types of cancer and discerning between two different types could mean entirely different forms of treatment for the patient.
Many a time, pathologists seek second opinions from their peers who may be in a different city or country. In order to get their opinion, the patient has to physically ship glass slides to the expert which is expensive, time consuming, often breaks in transit and faces its own host of other challenges.
Digital Pathology aims to mitigate this by providing digital copies of glass slides that can be shared, annotated and stored for future use.
***
As can be seen above, the various challenges being faced by the industry can be mitigated by implementing digital pathology. This being said, the digitization journey has so far been long and arduous involving government entities, labs and private companies coming together to innovate.
These innovations are an amalgamation of technological advancements in the fields of data storage, cloud computing, AI, camera lenses and imaging software.
Pathology is at the cusp of its digital transformation and its successful implementation would have far reaching implications in the world of healthcare.
