Large-scale transcriptome profiling of 1,523 diagnostic prostate tumors from randomized STAMPEDE phase 3 trials linked expression signatures and immunohistochemistry to 14-year survival. The Decipher RNA classifier was both prognostic and predicted survival benefit from docetaxel in metastatic disease, and a transcriptome-based PTEN inactivity classifier identified docetaxel-sensitive, metabolically perturbed tumors. High tumor androgen receptor signaling associated with longer survival while increased proliferation predicted shorter survival. These results support clinical implementation of transcriptome classifiers to guide treatment selection in advanced prostate cancer.
0:19Welcome to Base by Base, the paper cast that brings genomics to you, wherever you are. Thanks for listening, and don't forget to follow and rate us in your podcast app. So, to start us off today, I want you to imagine a high stakes real world clinical scenario.
0:33A patient is diagnosed with advanced prostate cancer. Right, which is unfortunately a very common reality. Yeah, exactly. And the doctor has these highly effective life-saving therapies available, specifically things like powerful hormone blockers or, you know, aggressive chemotherapy, but there is a massive catch.
0:54They don't actually know which one to use. Exactly. The doctor does not know which combination will work best for this specific tumor. So it is essentially an educated guessing game. What really happens when we treat cancer blindly without knowing its unique molecular vulnerabilities?
1:08Scary thought It really is. And how could this change if the tumor's own RNA could just hand us an instruction manual for exactly which drugs to use? Well, today we celebrate the work of the stampede trial team, and the researchers at UCL who have advanced our understanding of predictive biomarkers in advanced prostate cancer.
1:27We really have to acknowledge the monumental effort here, specifically highlighting authors Emily Grist, Peter Doody Magni, Gerhard Atard, and the MRC Clinical Trials Unit. Okay, let's unpack this because to really understand the magnitude of what this research team has done.
1:43We need the lay of the land, right? Like what is the baseline for treating advanced prostate cancer today? So the foundational treatment is androgen deprivation therapy or ADT. Basically, you're targeting the disease's primary driver by cutting off testosterone.
1:59But, you know, we've known for a while that ADT alone just isn't enough for long-term control and advanced disease. Right. The cancer eventually figures that a way around it. Exactly. So the standard of care is moved to treatment intensification.
2:09Doctors combine standard ADT with either a more powerful hormone blocker, an androgen resector pathway inhibitor, or ARPI, like Aberatorone, or they combine the ADT with chemotherapy, specifically dose and taxel.
2:22Got it. So we have a fantastic toolkit. But returning to that analogy I used earlier, we have no instruction manual telling us when to use the hammer, which is the chemotherapy, versus the wrench, you know, the hormone therapy.
2:33That is the perfect way to put it. And if we connect this to the bigger picture, the why here is so critical. Dawson Taxle, the chemotherapy, it's highly effective, but it comes with severe physical and financial toxicity.
2:45Like neuropathy, massive fatigue, immune suppression, all of that. Yeah, exactly. So overtreating with chemotherapy when a patient would have done just fine on hormone pills causes immense unnecessary harm.
2:58But on the flip side, under treating like withholding chemo from someone with a rapidly growing tumor, that invariably costs lives. Until now, a predictive biomarker to make that exact choice just hasn't existed.
3:10Which is why the methodology of this deep dive is just jaw dropping. This wasn't some tiny pilot study in a lab. They looked at 1523 patients from the stampede platform trial. With up to 14 years of survival follow-up data, that is incredibly rare in oncology.
3:27Massive. And to get the molecular data, they extracted RNA from diagnostic tumor biopsies, specifically, formal and fixed, paraffin embedded or FFPE tissue, then they ran these clinical grade pan transcripto micro arrays to test, I think it was 59 different biological expression signatures.
3:47Yep, 59 different signatures. It's a really comprehensive look at the tumors biology at the exact moment of diagnosis. Okay, wait, I have to push back here. Isn't RNA famously fragile? I mean, how on earth do researchers get reliable transcript on wide data from preserved, chemically treated biopsies that are, in some cases, over a decade old?
4:08It's a great question, and it's the main reason this hasn't been done routinely before. Formal and fixation cross-links the tissue, and basically degrades the RNA into tiny little fragments. Right. So standard R&E sequencing would just fail, right?
4:20Because it needs longer intact strands. Exactly. If you try to run modern RNA sequencing on a 10-year-old wax block. Your failure rate is going to be way too high for a clinic to rely on. But that's where the clinical grade micro array technology comes in.
4:32Oh, interesting. How is that different? Well, microarrays don't try to sequence the entire transcript. They use thousands of highly specific synthesized probes attached to a chip. They're just looking for short, specific target sequences.
4:44Oh, I see. So it's kind of like if you have a highly classified shredded document. Standard sequencing is trying to tape whole sentences back together, which is impossible if it's shredded too finely. That's a great analogy.
4:55But the micro array is just acting like a word search. It just looks for specific unique words that survived the Shredder. If that tiny fragment is there, it binds and you get a signal. Precisely. It has a slightly narrower dynamic range than modern RNA sequencing, but it is vastly more robust for old, degraded FFPE tissue.
5:15And that's what makes it perfect for rapid clinical implementation, because it uses the exact type of tissue samples, already sitting in pathology labs around the world. Wow, so they basically bypass the biggest technical hurdle.
5:27Okay, so they have this robust RNA data from over a 1000 patients, and they have 14 years of survival data. Let's get into the results. Sure. And to start, we really need to clarify a crucial scientific distinction for our listeners.
5:38The difference between prognostic and predictive biomarkers. Right, because those terms get thrown around interchangeably, but they are very different. Very different. Prognostic means predicting how aggressive the disease is overall, regardless of what treatment you get.
5:52Predictive means predicting if a specific treatment will actually work. Got it. So let's start with the prognostic discoveries. What did the baseline RNA tell them about survival? Well, tumors with higher androgen receptor or AR signaling at diagnosis.
6:09Those were strongly associated with longer survival. That makes sense. It means the tumor is still relying on the hormone pathways, so hormone blockers will still have a strong effect. Exactly. Conversely, they found that tumors with high cellular proliferation, so cells that are dividing rapidly, which they measured by RNA cell cycle signatures and the Kai 67 protein.
6:28Those were associated with much shorter survival. Fast growing equals more aggressive, makes total biological sense. Okay, here's where it gets really interesting. How did they move from just predicting overall survival to predicting who actually needs the chemotherapy?
6:44This is the breakthrough. They evaluated several signatures, but the standout was the decipher score. Wait, decipher? Isn't that already used in early stage localized prostate cancer? Yes, exactly. It's an RNA classifier normally used to predict metastasis after surgery.
7:00But in this study, they repurposed it for metastatic disease, and they prove that a high decipher score strongly predicts a survival benefit from dose taxel. Wait, really? So a test design for early stage disease works as a chemotherapy predictor in late stage disease.
7:15It does. The statistical interaction was highly significant, with a P value of .039. If you had a high decipher score, adding dosataxol extended your life. If you had a lower decipher score, the chemo gave you absolutely no overall survival benefit.
7:30That is incredible. I mean, we're talking about sparing men months of brutal chemotherapy side effects because their RNA says they won't even benefit from it. Exactly. But they didn't stop there. They also found a 2nd major predictive biomarker.
7:43PTN in activation. PTN, right. That's a classic tumor suppressor gene, isn't it? acts like a brake pedal for cell divisions. Spot on. When P10 is inactivated, a growth pathway called PI3K gets permanently switched on.
7:56The cells just divide out of control. So what did the stampede data show about these P10 inactive tumors? Well, 1st they showed that patients with P10 and activation had very short survival if they were only treated with hormone therapies.
8:09They were essentially resistant to standard ADT. But, and I'm guessing here, because they are dividing so fast, they are extremely sensitive to chemotherapy. Yes, highly sensitive to dose of taxel. The statistical interaction there was incredibly strong with a P value of .0000.
8:27Wow. So what does this all mean? How does this change the day-to-day reality for a doctor and a patient? It permanently shifts the paradigm. A doctor can now take a standard single biopsy at diagnosis, run this clinical grade test, and identify if the tumor is high decipher or P10 inactive.
8:44And if it is, they bring out the heavy artillery. They prescribe the dose of taxyl, knowing it will actually work. Exactly. And conversely, if the tumor is PTN active and has a lower decipher score. The doctor can confidently say, we don't need to put you through chemotherapy.
9:00The hormone blockers are enough. That is the ultimate goal of precision medicine, giving the right drug to the right patient at the right time. It really is. But this raises an important question about clinical guidelines.
9:11To change practice, you need level one B evidence. And because they pre-specified their statistical plan before looking at the 14 year survival data, they actually achieved that level one B standard here.
9:22Which is incredibly hard to do with archival tissue. It prevents people from just fishing for good statistics. But of course, we always have to look at the limitations of any study. What's the boundary here?
9:33The main limitation is how fast the field of oncology moves. The stampede trial data they used didn't include a modern triple therapy arm. Ah, right. triple therapy. That's where patients get standard ADT, plus an advanced hormone blocker, plus the chemotherapy, all at the same time, right?
9:48Correct. That's becoming standard for very aggressive cases today. Because the historical trial didn't test all 3 together. We don't know for apps so certain if the decipher score, or PTNN status, perfectly predicts benefit in that exact triple therapy scenario.
10:05So extrapolating these predictions to the 3 drug combo requires further validation in newer trials. Exactly. We have the biological rationale, but we still need that final clinical proof for triplet therapy.
10:16Still, the leap forward here is just monumental. So to wrap this all up, How would you summarize the core takeaway from this deep dive? I'd say it comes down to this. By leveraging clinical grade RNA micro arrays on standard archival biopsies, we can finally peer into the transcriptome of advanced prostate cancer.
10:34We can use that data to predict exactly which patients need chemotherapy, transforming what used to be a clinical guessing game into true precision medicine. What does this mean for the future of personalized oncology?
10:45If an RNA signature can unlock the exact rate treatment for advanced prostate cancer, How many other cancers are out there, just waiting for their own molecular instruction manuals to be decoded. This episode was based on an open access article under the CCBY4 license.
11:02You can find a direct link to the paper and the license in our episode descriptions. If you enjoyed this, follow or subscribe in your podcast app and leave a five-star rating. If you'd like to support our work, use the donation link in the description.
11:14Now stay with us for an original track created especially for this episode and inspired by the article you've just heard about. Thanks for listening and join us next time as we explore more science, base by base.
11:43In the quiet glow of late night screams A million messages in between From frozen slides And stubborn stains We learn what stays and what remains. Some tumors race, some hold their line. Some hide the map entangled sign, but the noise.
12:15A pattern grows. A truth, a transcript finally shows. Read the signal. Don't guess the fight Let the data turn on the light If it's built to break We hit it hard Pick the Play the right card Read the signal Steady and clear.
12:47Bring the best chance closer here. When A.R. burns like a steady flame The days stretch out The odds can change. But high speed cells with restless strife. Can steal the time we try to buy the tense silence of different cost metabolic streets where rules get lost.
13:32Not every answer comes out clean. Not every promise stays routine. Do we consort what's sharp and true? From long ears, yeah, from number two. So when the moment calls our name, we choose with care. Not hope alone, not playing.
14:03Breathe the same no heart. Don't guess the fight Let the data turn on the life It's built to break We hit it hard Pick the ride ride and play the right Breathe the signal Steady and clear Stronger decisions here by here. Read the signal.
14:40right here
