Yeah. The unfortunate truth about science news reporting is that usually it's not sensational, so they need to play things up for clicks and ad revenue. A lot of the time it ends up in somewhat misleading semi-truths like this.

Well, the article refers to both :)

I think you'd be right about the "number of diagnoses" statement in the title, but I think the discussion is about the deaths due to cancer, which have also increased and would not have as strong of a correlation for the reasons others mentioned

Even in the US, autopsies are not always performed. Ima quote WebMD because I'm bone idle:

Although laws vary, nearly all states call for an autopsy when someone dies in a suspicious, unusual, or unnatural way.

Many states have one done when a person dies without a doctor present. Twenty-seven states require it if the cause of death is suspected to be from a public health threat, such as a fast-spreading disease or tainted food.

According to a 2012 DOJ report, only 8.5% of US deaths result in autopsy.

I mean sure. But the data is likely comparative and can be looked at just within countries that have been getting autopsies since the 90s.