Shahid Akhter, editor, ETHealthworld, spoke to Prof. Debarka Sengupta, Associate Professor (CB, CSE), IIIT-Delhi, about his most recent work on cancer diagnosis and precision therapy that makes it more affordable and accurate as well.
Cancer diagnosis: Current scenario
The world of cancer biology and cancer detection has evolved over the years, and the way cancer was diagnosed was primarily based on imaging techniques. But nowadays cancer can be diagnosed much early at its onset in different methods. For example, we can use liquid biopsy-based technologies, which allow you to look for cancer biomarkers in the blood such as the circulating tumour DNA on the subcutaneous tumour cells, but they are extremely rare, so it is very hard to find them in the blood. So you typically miss many patients who actually have cancer. However, nowadays, there are better ways of doing it where you can actually look for things that are abundant in blood, like tumour-integrated platelets, which are basically your blood players that come in close proximity with the cancer cells. And then they learn about the signals of the cancer cells, which can be tracked down from the blood, and thereby you can accurately diagnose the cancers in the patients.
Cancer diagnosis: Challenges
The major challenge in cancer diagnosis is the lack of markers that are easy to access from the patient’s blood. For example, circulating tumour cells are extremely rare in the patient’s blood, so they cannot be accessed so easily. In our country, it’s even more difficult to do it because we don’t have sufficient infrastructure to do these activities. Another example is circulating tumour DNA. But again, circulating tumour DNA is also rare. Imaging-based technologies such as PET scanning, etc. are extremely expensive, so they cannot be used for regular screening purposes. So our group does research on how to make inexpensive approaches to detect cancers from the patient’s blood that are much more affordable as well as accurate.
Cancer diagnosis: Research of IIIT-Delhi
So our research in cancer biology can be segmented into two different parts. Number one is how to diagnose cancer, and number two is how to give the right therapy to the right patient. In the first case, to diagnose cancers, we have developed a method by which, instead of looking at the rare biomarkers in the blood, we can look at tumour-infiltrating platelets, which are present in the blood in abundance and therefore can be easily isolated in the common COVID laboratory. and you can extract the iron. And then, based on that, you can use an artificial intelligence-based model to predict the existence of cancer. It is very inexpensive and available everywhere because it can be implemented using a simple RTPCR machine. So this has been transferred as a technology to a company called Care On COVIDITY Private Limited which is doing clinical validation for commercializing the same.
When it comes to deciding the right therapy for cancer patients, it’s of paramount importance because, in the case of cancer, when you administer a therapy to the patient, you might see a great response in the initial few days. But in the end, the cancer cells become resistant. As such, they stop the therapy. They don’t respond to the therapy, and the patient’s situation deteriorates slowly until he succumbs. We have developed very recently an approach by which you can use the gene expression data of the tumor to make predictions about which drug would work best on a given cancer patient. It’s also known as “personalised therapy” or “precision therapy” in cancer, which is extremely widespread these days in the USA or other advanced countries.
In India, precision therapy is still in a very nascent phase. Our approach and our research are working, and slowly other people are also getting into this field and trying to come up with innovative ways of identifying which drugs work best. And they are actively using artificial intelligence to model this by looking at the genetic and molecular data. So I’m extremely hopeful that with the help of all this, we will be able to resolve this. We have recently published a story on this idea in Nature Communications, where we discussed how the drug prediction for a given patient can be made extremely accurately and then how it improves the patient’s prognosis. This has attracted a company called Gender Critter Incorporation, which is a Texas-based company. And we are developing a more expansive method for all cancers using different kinds of molecular markers, not only in expression but other types of markers as well, to use AI and predict the right therapy for the cancer patient. Which we hope will be there in the market in the next two or three years time, which will make the prognosis of the cancer patients learn better. And all this work has been done out of IIIT-Delhi, where I’m an Associate Professor, with the help of other collaborators from across different continents.
IIIT-Delhi
IIIT-Delhi is a very unique place. It is a research-backed teaching institute. So although it is a university where we teach BTECH, MTECH, and PhD programmes, all different kinds of programs, we have a significant amount of stress based on research. More recently, with the inception of this department of Computational Biology, we have nearly ten to twelve faculty members who are actively working on different areas of healthcare research, including structural biology, chemical biology, mathematical biology, Food Related research, how AI can be used in food, how AI can be used in ICU for scanning the patient data and to predict any wrong signal ahead of time. And all this research is happening here.
More recently we have come with an experimental laboratory facility, which is very unlikely to find any IT focused institution which IIIT-Delhi is and there we are now able to validate a lot of hypotheses which is generated using computational methods. And a lot of people here are doing research on topics related to the patient’s health. Very recently we had a paper coming out in a collaboration between Gaurav Ahuja’s lab and My lab, where we could create an AI Based model to predict a molecule, whether it is carcinogenic or not. Similarly, there are other colleagues who are working on several prediction tasks related to ICU, for example, and they have a direct impact on the patient’s health.