Storage of Histopathology data: Traditional & New Ways
The records for histopath will comprise of the request form, the cell specimen itself, the blocks tissue blocks taken and the report.
The request form is normally kept for 20 years, the tissue is kept for a minimum of six weeks so that gives the clinician the time to look at the report and check if he disagrees or he wants a re-test so that we can take more samples. The block is kept for another 20 years and the slides fade within a year or two depending upon the stains. The written report is kept for at least 20 years.
This varies by jurisdiction as well countries, sometimes it might be less, sometimes more since there may be more conditions. Now because of digital storage, it can be kept for a perpetually amount of time.
How to analyse histopathology data?
So because I am a histopathologist and a statistician. I am very often involved in research. The big problem we face is we cannot retrieve all the records pertaining to a particular disease. If this happens, we are going to / we will continue to have a very lopsided view of the disease. We can only learn from whatever we’ve retrieved.
Things we have missed in the past will continue to be missed again if this is the case. I believe this is something a digital solution or software must resolve. With a system, I must be able to search reports by using code words. Not just limited to a test, but these code words must be relevant to the body of the reports; making search and research hassle-free.
For example, if I want to look for something like Schiller Duval body and where all it has been used, I should be able to get all the reports that have this term in their body using the system. Similarly, all the reports have positive or negative findings. So this is a software issue/enhancement that needs to be leveraged.
CrelioHealth offers a boolean search feature to improve research and studies for histopathologists. Know how it can be beneficial for your lab here.
“This is one of my chief complaints as a statistician because we have to do this research and I often decline because we can’t retrieve all the cases which are likely to belong to this group then you are going to have very important cases left out you can’t learn from them so there is no point in doing that research.”
This is very important to us because sometimes as pathologists we debate between two different diagnoses. I will use a particular word to say that it’s not this case – I will say there are no Schiller Duval bodies so that means that that particular case can enter into the differential diagnosis. We need to pull out that case as well do its final coding may be different. You need to be able to pull out cases using multiple keywords in the body.
Can mRNA be used for cancer treatment?
I’m not sure but it looks like the technology works if you want to deliver a short set of nucleotides to the cell but whether if you can cure any cancer using this particular technology is yet to be proven. I am hopeful that they will be able to use this technology not just for cancer but other kinds of diseases as well. For example, cellular level of diseases.
What are the trends in histopathology?
The H&E part of the histo is fairly static now over the years. Of course, we have classified and re-classified the diseases, number of times but now it’s all at the molecular level. So I think in the future whether it’s training pathologists or setting up labs, molecular is going to play a very very important role. For diagnostic, prognostic and predictive testing.
For example, in lung cancer, you look to have to look at sections, we classify into groups, to see if a particular “picture’ will respond to a particular immune therapy, then we would need molecular testing. The basic H&E workhorse will remain the same for a long time.
Can AI take over the job of Histopathologist?
When I first entered the training in 1996, one of my bosses told me that by the time I graduate, AI will have taken over my job but I am still here there is no sign of AI replacing me.
But I think molecular diagnosis will definitely play a very big role, H&E (Hematoxylin and Eosin) will still be central but molecular will play a big peripheral role. Mostly it will be interdependent.