Thread by @CCRobertsARK

1/8

On this week's @ARKInvest newsletter (📰), I plan to discuss some exciting research on molecular (prog)nostics. I'm also going to touch on why long-read #sequencing is so important within oncology.

Here's some context:

2/8

We believe multi-cancer liquid biopsy screening will allow dozens of cancers to be detected earlier. However, not all tumors are lethal. Some may never cause symptoms.

By reflexing to surgery, we risk overtreating patients whose tumors may never need it.

3/8

Companies like Veracyte (VCYT) and Castle (CSTL) offer prognostic tests, using a combination of machine learning and gene expression, to risk-stratify suspicious nodules beyond standard cytopathology.

biospace.com/article/releas…

4/8

We believe molecular prognostic testing could give physicians a better understanding of a tumor's eventual fate, perhaps reducing the risk of overtreatment. This approach is not a panacea, and there's still much learning to do, but the future looks bright.

5/8

Molecular prognostic tests result from the collision of #MachineLearning and next-gen sequencing, both of which rapidly are becoming cheaper. We think that these tests, even if given annually for decades, should be considerably cheaper than the cost of cancer treatment.

6/8

I still need to go in depth, but I was really excited by this paper (nature.com/articles/s4159…) published in Nature by @RFitzgerald_lab.

She summarizes much better than I could, so I'll direct readers to her thread:

7/8

While her work utilized low pass short-reads, I'll note that copy number variants (CNVs) are particularly amenable to long-read platforms (so long as tissue isn't degraded).

Here's another great paper by @mike_schatz and team that discusses this:

genome.cshlp.org/content/early/…

8/8

In summary, we'll dive into long-reads, prognostics, and other ways to tackle overtreatment throughout the week. Please add any and all questions and I'll try to address them in my weekly newsletter.