The Global Epidemic of Early-Onset Cancer: Nature, Nurture, or Both?

Related Link: Annals of Oncology

Dr. Ogino: Hello. I am Shuji Ogino. I am Professor of Pathology and Epidemiology at Harvard University and Chief of Molecular Pathological Epidemiology at the Brigham and Women's Hospital. Also I'm a member of the Broad Institute of MIT and Harvard.

Today I will talk about early-onset cancer, on which I wrote an editorial because this is a growing problem, and we actually now see a tip of the iceberg. But I can predict that, when the young people get older in the future, there are many other chronic diseases and later-onset cancers as well. And we think prevention is so important. So, I will, in the editorial, put a picture similar to this, a little bit modified, but this illustrates the early-onset cancer etiological research framework. For the first few minutes, I will focus on the top, which is an assessment of exposures in early life of young, cancer-free people.

The incidence of early-onset cancer increases every year

That's very important. And then we need to use existing child cohorts in working with pediatric researchers. I can explain why. So, we see a cancer rise in older adults. You go back, to even 1950, and we start to see the rise of cancer incidence increasing in elderly people because obviously we experienced the lifestyle change, which started gradually around that time around the world so that this star represents colorectal cancer. This is an example of a colorectal cancer, by the way. But people born at that time, obviously were very young, didn't get the cancer. However, decades later, adult cancer increased even more throughout the 20th century. And it's now plateaued because we have a good screening method. But, for young people, when they get older, then we start to see early-onset cancer around the 1990s. And then we are continuing seeing an increase every year in the younger generations.

Each birth-cohort has its unique risk-factor exposures

Birth cohort means people born in a similar time like this. So, I can explain birth cohort, as I said, as people born in similar time. Each birth cohort has unique risk-factor exposures, distributions, starting from conception and then that's different from every other birth cohort. And then we now know that successive birth cohorts, the younger generations, have a higher cancer risk at a given age. So, especially when they get old, they get more cancer from this phenomenon. We can't predict.

One risk factor may influence another may influence another

So, there are many possible risk factors, and then we don't really know which are risk factors exactly, and how they impact each other. Not yet. We have to do a lot of studies, and it's a big challenge because we need to know [risk factors from] early life to adulthood. All of these factors can be contributing risk factors, in addition to the first items. And then, second, are chronic diseases. Obviously, there are risk factors known to be associated with cancer, like obesity. And the challenge is that no study can control for confounding without a proper dataset. The confounding means that those risk factors may influence each other so that other factors can look like a risk factor, which is a big challenge in epidemiology.

Biospecimen research

So, in the second part, we like to focus more on biospecimen research, which is an exciting part and which can give us great insight. We published a few years ago, a review in Nature, and this [slide] is modified from the figure published. Here we presented an ideal study for early-onset cancer, starting from early life — really early life, zygote, conception. And then the fetus, infancy, childhood, adolescence. Those are collectively what we call early life, before adulthood. So, the birth cohort effect really implicates cancer risk-factor exposure that started to happen in early life to adulthood, So, starting from early life and continuing to adulthood, we have to think of the time sequence, which is an enormous challenge. And, then, what's good about biospecimens is that we can get biospecimens from all these young people in early life and then freeze them. We can keep early-life information, biological information, for future studies, which is great. Also, obviously, we are now accumulating early-life electronic health records that can be utilized, especially the use of AI can be powerful.

I'm also a pathologist interested in tumor tissue analysis, which can provide a great clue to pathogenesis. So, for example, [considering] colorectal carcinoma, colorectal cancer, we and others have shown a unique feature for early-onset cancer, such as LINE-1 hypomethylation and immune suppression. And, also, colibactin, a mutational signature. Colibactin is produced by a bacterial strain called pKs+ E coli. This can give us great etiological insights too.

Linking epidemiological factors and pathological features

Then we can utilize a molecular pathological epidemiology approach, which I have been working on for decades. And we can link epidemiological factors and the pathological features. We can really link utilizing big cohort studies. Actually, we have already conducted studies to link long-term exposure with some of the tumor features, which are early-onset cancer features. Although these studies are not looking at early-onset cancer because we don't have enough case numbers, we actually show that those lifestyle dietary factors — especially dietary factors — link to LINE-1 hypomethylated CRC incidence; alcohol and low folate and an inflammatory diet or aspirin non-use can be linked to immunosuppressed CRC incidence, and a Western diet is linked to pKs + E coli–high CRC incidence.

Cohort incident tumor biobank

Our prospective cohort incident tumor biobank is very unique. This PCITB is a new biobank encompassing all of these cohort follow-up elements, and information and biospecimens and integrated analysis. So, that is not easily done by others, yet. But we need to make more of these biobanks to study.

By the way, about this biobank, we put it on the website. You can look at it, free of charge. So, please take a look.

The effect of long-term risk-factor exposures

So, to summarize my speech, short speech, here, rising early-onset cancers are a big problem, [but] we see a tip of the iceberg. Starting in early life we see the effect of long-term risk-factor exposure. What we see now are results of decades-long exposures, and then actually many chronic diseases will increase. And this is a kind of warning sign. What we must do is raise awareness, and we need to utilize existing child cohorts to enable life-course studies and then fund research with long-term impact.