A new study shows that the long-term decline in cancer rates may be reversing.

A lot of epidemiology starts with a simple question. But as you think about it, you realize just how complicated the question actually is. Here’s a simple question – is cancer more common today than it was in the past?

Easy, right? Well – let’s take it apart. I mean, the truth is there are more new cases of cancer in the US now than there have ever been before. Last year, 2 million people in the country were diagnosed with cancer. There were about 1 million diagnosed in 1990. Of course, that’s misleading – the population of the US is bigger now than it ever was before. So we don’t want a count really, we want a rate – like a rate per 100,000 individuals.

But even the cancer incidence rate is not entirely clear cut. After all, the population of the US is older now than it was in, say, 1960, and age is the major risk factor for most cancers, so if the rate of certain cancers are increasing, is it really fair to conclude that something worrisome is driving that if it’s really just driven by the overall population getting older? As one of my epi teachers used to say – “the question is what is the question”. So if we want to really understand if something has fundamentally changed in terms of cancer risk over the years, we need to calculate age-adjusted cancer rates.

Even then, there’s a problem. Older people are not just older biologically – they were also born further in the past. I know this sounds like circular reasoning. But it’s a bit deeper than it sounds. Realize that current 70-year olds were exposed to things in the 90s when they were 40. And current 50-year-old had the same exposures, just when they were 20. So birth cohort matters. If there was some horrible cancer-causing chemical released briefly in 1992, for example, you might expect to see a rise in cancer among all those alive in 1992, regardless of what age they were.

In fact, it may be reasonable to think of cancer incidence not in terms of the calendar year, but in terms of the risk by social generation. And if you do that analysis, you’ll find something quite interesting. With each successive generation since the Greatest Generation, cancer incidence has decreased. With one exception: Gen X.

I was born in 1979, the tail end of Gen X. My childhood was spent riding a bike without a helmet, listening to Duran Duran and Wham!, and staying up late to watch Sandler and Farley on SNL. Simpler times. From a public health perspective, my generation was really one of the first to have a clear understanding, from a very young age, of cancer risk factors. We were the ones told to “just say no”, though the effect was not that profound.

We were more health conscious – though the “low fat” craze would prove to be disastrous as our carbohydrate intake soared. Still – a priori I would have thought that our rates of cancer would be lower than our parents.

This paper, appearing in JAMA Network Open, shows us that it is not.

Source: Rosenberg et al. JAMA Network open 2024

Researchers used the well-worn SEER cancer database to capture incident cancer cases among 3.8 million people in the US from 1992-2018. This cohort spanned birth years from 1908-1983, from the Greatest Generation to Gen X.

The key data elements here are age at cancer diagnosis and year of birth. This allowed the researchers to examine how rates of cancer at a given age changed over time.

I’m going to start with a straightforward example. Here are the rates of lung cancer at age 60 among women and men, as a function of birth year. The broad trend is fairly clear – people born in more modern times have a lower risk of getting lung cancer by the time they turn 60.

This makes sense. The campaign against smoking has been one of the more successful public health campaigns over the last 50 years.

But the situation is not so good for some other cancers.

For example, look at the rate of kidney cancers and thyroid cancers. People born later in the century are dramatically more likely to be diagnosed with kidney or thyroid cancer than earlier generations were.

Lumping the major cancers together gives you charts like this. The authors stratify these by race and ethnicity, but pretty much the pattern holds. The Greatest Generation had the highest risk of cancer, which declined and more or less remained flat for the baby boomers and then slowly, inexorably, has begun to rise again – putting my generation, for the first time in about a century – at higher risk of cancer than our parents were.

OK – what’s going on here? There are some benign explanations – pardon the pun. We have made significant strides in reducing the rate of heart disease in the population over the past 100 years. As deaths from heart disease decline, a greater number of people live long enough to be diagnosed with cancer. Of course, age-adjusting the cancer rates should account for this.

Unaccounted for is the fact that this study examines the number of detected new cancers. Detected being a key word here. As technology has advanced, our ability to detect cancers earlier, and even to detect cancers that might never have been detected otherwise, has dramatically increased. I think thyroid cancer is the poster child for this – as thyroid ultrasounds proliferate, incidentally detected thyroid cancers have skyrocketed.

Supporting the idea that this is more detection is the fact that, broadly speaking, cancer mortality has declined over time. Even if I am more likely to be diagnosed with kidney cancer than my parents, I am less likely to die from that. That’s due to earlier detection but also, of course, because we have better therapies now.

There are, of course, malignant explanations for this phenomenon. Environmental exposures have substantially changed from the 1940’s to today – and the stuff my generation was exposed to when we were kids is fundamentally different than the stuff my parents were exposed to. Sure, they had lead and cigarette smoke. But we had industrial chemicals, pesticides, and ultra-processed foods.

What this paper really tells us is that the fight against cancer is not only continuing, it is changing. I’ve seen it anecdotally in my own practice – with younger and younger people presenting with cancers that, by the books I read in med school, are supposed to present in those who are older. We’re making progress, of course – it is clearly better to have cancer now than it was to have cancer 50 years ago. But the goal of eliminating cancer will mean we have to be flexible in our approach. Cancer changes quickly – we have to as well.

A version of this commentary first appeared on Medscape.com.

Methods Man Blog

Why Gen X is the First Generation to Have More Cancer than Their Parents

A new study shows that the long-term decline in cancer rates may be reversing.

The Methods Man