This study evaluates an automated, decentralized cfDNA NGS workflow (Oncomine Precision Assay GX with the Genexus system) in 298 patients with advanced solid tumors. The assay achieved 99% sequencing success, detected mut-ctDNA in about half of patients, identified actionable or resistance alterations in 18% of patients, and showed 72% concordance with matched tissue profiling. Detection sensitivity varied by cancer type, tumor burden, and metastatic site, and plasma-only variants were enriched after targeted therapy.
0:00Mm. Yeah, yeah. Late my lights on the quiet floor. A tiny whisper in the bloodstream's roar. Welcome 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.
0:28So I want you to imagine a really critical moment in a clinic. A patient's cancer has returned. And uh, even worse, it's become completely resistant to the standard therapy they were receiving. Right, which is a terrifying scenario.
0:42Yeah, exactly. So the medical team is huddled together and they urgently need to know exactly how this tumor has mutated, you know, so they can find a new targeted treatment. The window to act is incredibly narrow.
0:54But the only available puma tissue they have is years old. It's degraded, sitting in some cold storage archive, essentially just a relic of the past. Yeah, it's basically like looking at a fading, outdated photograph of the tumor from years ago, and then trying to make a life or death decision based on it today.
1:13Wow, yeah, that's a great way to put it. Because the tumor has evolved and adapted, but the doctor's reference point just hasn't. And historically, if you want to process a new tissue bioxy or, you know, even send a blood sample to a massive centralized lab, you're looking at weeks of waiting for those genetic results to come back.
1:31So how could this change if a simple decentralized blood draw one processed right at the local clinic in just 24 hours could provide a real-time high-definition surveillance feed of the cancer's current genetic state?
1:45I mean, what really happens when a community hospital doesn't need to rely on a distant mega lab to understand how a patient's tumor is mutating right this very second? It fundamentally flips the script, you know, on how we handle cancer recurrence.
1:59It takes this process that's usually measured in weeks and shrinks it down to a single day. Which is huge. It is. For a patient with advanced cancer. Time is literally the most valuable currency there is.
2:08Absolutely. So today, we celebrate the work of a team, led by hitting Chan, Shinji Takahashi, and Suki Lo, primarily at the Japanese Foundation for Cancer Research, who have really advanced our understanding of automated and decentralized genomic profiling for advanced solid tumors.
2:26Yeah, we were digging into their 2025 paper, automated and decentralized genomic profiling of plasma cell free DNA for identification of targetable and resistance alterations in advanced solid tumors. Quite a mouthful, but an important one.
2:40Definitely. This was published in the journal Clinical Chemistry, volume 71, issue six. So to really understand why this paper is so impactful. I think we need to look at the current standard of care for precision oncology.
2:51When a patient needs a targeted therapy, doctors rely almost entirely on molecular profiling of tumor tissue, right? They surgically remove a tumor or they take a needle biopsy and that physical tissue is sent off to pathology.
3:03And what happens to it there? Well, in pathology, that chunk of tumor is essentially embedded in wax. It's formal and fixed and paraffin embedded. From there, scientists carve off these microscopic slices, extract the DNA, and sequence it to hunt for those actionable mutations.
3:20Okay, so the underlying assumption is that this block of wax holds the master blueprint of the patient's disease. Exactly. But here's the catch. About 10% of these archive tissue samples completely fail analysis.
3:34Wait, really? 10% But if 10% of these tests completely fail, and we're talking about patients who are, you know, running out of time. Why are we still relying on archive tissue as the gold standard? I mean, that feels like a pretty unacceptable gap in the standard of care.
3:49It's a massive gap, yeah. But you're dealing with profound biological and chemical limitations here. Sometimes the needle biopsy just doesn't yield enough total DNA to even run a sequence. Oh I see. Other times, the tissue slice has a very low content of actual cancer cells relative to the normal healthy tissue around it.
4:08So the tumor signal just gets completely drowned out. That makes sense. But honestly, the biggest enemy is time. The formal and chemicals used to preserve the tissue, actually cross-link and degrade the DNA over time.
4:21It makes the DNA super brittle. So the older the sample, the worse it gets. Exactly. The longer that block sits in an archive, the more the DNA shatters into these tiny, unreadable fragments. Okay, so if the archive tissue is too degraded to read, the obvious solution seems to be just going back to the patient for a fresh sample, right?
4:39Like just take another biopsy to see what the tumor looks like today. Well, in a purely theoretical world, yes. But in clinical reality, going back for a 2nd or 3rd biopsy is highly invasive. Oh, right, because of where the tumors might be.
4:52Exactly. Depending on where the tumor has metastasized. So deep in the lungs or the brain taking a new core sample can be outright dangerous. And it's also just physically and emotionally exhausting for a patient whose body is already ravaged by advanced cancer.
5:06Yeah, I can even imagine. And beyond the physical toll, there's a massive biological flaw in relying on single biopsies anyway. It's driven by a concept known as tumor heterogeneity. Because a tumor isn't just one uniform block of identical clones, right?
5:20Far from it. A tumor is this highly diverse chaotic ecosystem. When you take a needle biopsy, you're really just taking a tiny localized core sample of one specific geographic neighborhood of that tumor.
5:34So the cells on one side could be totally different from the other. Exactly. The cancer cells on the left side of the tumor might be driven by completely different genetic mutations than the cells on the right side.
5:44And if the cancer has spread, the metastatic tumors in the liver, for example, are evolving entirely different resistance mutations than the original primary tumor. Wow. So a single tissue biopsy gives you a snapshot of one neighborhood, but completely misses the drug resistant threats silently evolving on the other side of the city, so to speak.
6:04Precisely, which is the core argument for the liquid biopsy. Instead of stabbing a needle into one specific geographic location, we just draw a vial of blood from the patient's arm. Because tumors shed into the blood.
6:16Right. No matter where they're hiding in the body. Tumors shed microscopic fragments of their genetic material into the bloodstream as their cells die and turn over. This is what we call self-free DNA or CF DNA.
6:29So, a blood sample effectively acts as this, like, systemic whole body surveillance system. Exactly. It sweeps up the genetic debris from every tumor site across the entire body. It captures the primary tumor, the liver metastasis, the bone metastasis, all in one minimally invasive draw.
6:45And that completely bypasses the heterogeneity problem because you're looking at the combined genetic output of the entire disease burden. You nailed it. But wait, liquid biopsies have been around for a few years now.
6:57If drawing blood solves both the invasiveness problem and the heterogeneity problem, why isn't every local clinic already doing this for every single patient? That's the $10000 question. And the answer is infrastructure.
7:09Because until very recently, analyzing that self-free DNA required massive centralized facilities. Oh, so you still couldn't just do it in-house? Nope. You had to draw the blood, package it in these highly specialized temperature controlled tubes, and physically ship it via courier to a massive commercial sequencing facility 100s of miles away.
7:30Which takes time a lot of time. Once it arrived, it sat in a queue. It required complex manual lab prep by specialized technicians, followed by days of running on massive sequencers. And then, a remote bioinformatics team had to clean up and interpret all that data.
7:48So even though drawing the blood is fast, the logistics of actually analyzing it totally negate the speed advantage. The patient and the doctor are still just sitting around for 2 or 3 weeks waiting for a PDF report to be emailed back to them.
8:00Exactly. The centralization is the bottleneck here. The real clinical need is to have a machine sitting right there in the local hospital's pathology lab that can do this automatically. Without the shipping, without the specialized molecular biologist, and definitely without the weeks of waiting.
8:16You got it. And that's exactly what this research team set out to prove whether that local automated approach could actually match the gold standard. Okay, so let's talk about how they did that. They designed this clinical evaluation around a cohort of 298 patients with advanced solid tumors, right?
8:33We're talking severe colorectal pancreatic and breast cancers. And the defining characteristic of these 298 patients is that their cancer had progressed. They had all developed resistance to their standard first line therapies.
8:48Gotcha. So these are the exact patients who urgently need new molecular answers to find a 2nd line targeted drug. Exactly. So the researchers set up a direct head-to-head comparison between two fundamentally different ways of getting those answers.
9:02In one corner, you have the current heavyweight champion, the Foundation one CDX. Which is the centralized FDA approved tissue test. You send the physical tissue block away. It comprehensively screens over 300 genes and the turnaround time is, like we said, measured in weeks.
9:17Right. And in the other corner, you have the decentralized challenger. The team used the on-commine precision asset GX, combined with the Genexis sequencer. So this is an automated system located right at the local institution.
9:29Yep, it's a targeted blood plasmatist. It screens a very focused panel of 50 key pan cancer genes, and it delivers results in a rapid 24 hours. That's incredible. 24 hours. Yeah, you basically just put the blood plasma in and the machine automates the chemistry, the sequencing, the data analysis, and then generates the report locally.
9:49But looking closely at their methodology, there's one step the researchers took that feels completely counterintuitive to me. Yeah, which part? Well, they draw the blood, they spin it down to get the plasma where the tumor DNA is floating and they sequence it.
10:03But then they also take the patient's normal, healthy white blood cells from that same tube. extract the genomic DNA from them and sequence those two. Ah, yes. So my question is why bother? If the entire goal is to hunt for cancer mutations, Why spend sequencing power on the healthy blood?
10:21Because healthy blood lies. Wait, what do you mean healthy blood lies? The researchers are using those white blood cells to build a baseline filter against a biological phenomenon called clonal hematopoysis or CH.
10:32Okay, let's explore the mechanics of clonal hematopoasis, because that sounds really fascinating. It really is. It basically comes down to the evolutionary biology of aging. Deep inside your bone marrow.
10:43You have stem cells, whose entire job is to constantly divide and pump out new white blood cells for your immune system. They do this every minute of every day for your entire life. Now, over decades of constant division, these stem cells naturally make typos.
10:59They accumulate random genetic mutations. So just by virtue of getting older. Exactly. By the time a person is 60 or 70 years old, millions of their circulating white blood cells carry these typos. Now, these lutations are generally benign, you know, they aren't causing leukemia or blood cancer, the immune cells still function normally.
11:17Okay, but I'm guessing there's a catch. A huge catch. The genes that are mutating in these aging stem cells are the exact same genes that drive solid tumors. Genes like TP 53 or KRES. Oh, wow. And that creates a massive diagnostic trap, doesn't it?
11:32It does. When these benign, mutated white blood cells naturally die, they pop open and dump their DNA into the blood plasma right alongside the DNA shedding from the actual lung or breast tumor. Ah, I see where this is going.
11:47Yeah, so if a doctor only sequences the plasma. The machine will detect a massive spike in mutated TP 53, for instance. The doctor will look at the report, assume they just found the driving mutation of the patient's solid tumor, and prescribe a drug based on it.
12:02But it's a ghost. The tumor doesn't have that mutation at all. It was just the benign static noise from the aging white blood cell. Exactly. It's a total phantom. So by sequencing the normal white blood cell separately.
12:12The researchers basically create a subtraction filter. They identify all the benign typos in the healthy blood, and they digitally erase those specific mutations from the plasma results. Right. And whatever is left over is the true verified tumor DNA.
12:27It's a critical layer of quality control. And as we'll see in the data, skipping it has severe consequences. Well, let's look at that data. How did the two methods actually perform in the real world? Because right out of the gate, the decentralized blood test demonstrated a massive logistical win.
12:43It did. The plasma sequencing had a 99% success rate. Almost every single blood sample ran perfectly and generated a readable result. While the archive tissue samples, the supposed gold standard only achieved a 96% success rate.
12:58Yeah. And when you look at the tissue samples that failed, it confirms exactly what we discussed earlier about degradation. The failed tissue samples were significantly older. How much older are we talking?
13:08The median age for a tissue sample that failed to sequence was 1526 days. Wow, that is over 4 years old. Right. Contrast that with the successful tissue samples, which had a median age of roughly 497 days.
13:21If a patient's primary tumor was surgically removed 4 years ago, trying to extract high quality DNA from that ancient block of wax today is just highly prone to failure. And the real-time blood draw completely bypasses that degradation problem.
13:35Okay, so the blood test is faster and it fails less often, but none of that matters if it isn't accurate. When they compare the genetic results of the patients who had both a successful tissue test and a successful blood test, how often do they actually agree?
13:49So the overall concordance rate, meaning the plasma test found the exact same driving mutations as the tissue test, was 72%? Okay, 72% is solid. But, you know, if blood is supposed to be a systemic surveillance system capturing everything in the body, why are 28% of the tests not matching up?
14:07Like, why is the blood missing things that the tissue caught? Because the success of a liquid biopsy is fundamentally dictated by the physical biology of the tumor itself. Not all tumors shed DNA into the bloodstream at the same rate.
14:21Oh. Yeah. The researchers found that the detection rate was extremely high in colorectal and breast cancers. But it plummeted when looking at pancreatic and thyroid cancers. The tumors simply weren't releasing enough material to trigger the machine's sensors.
14:35Okay, let me try an analogy here. Think of the human bloodstream like a city's main river system. Okay, I'm with you. If you build a massive, highly active industrial factory right on the riverbank, it's going to dump a huge amount of detectable waste straight into the water.
14:51So in the body, that factory is a very large tumor, maybe over 20 millimeters in size, or a major metastasis sitting directly in the liver. Right, because the liver is an incredibly vascular organ. An enormous volume of blood flows through it every single minute.
15:08A tumor growing in the liver is constantly being washed over by that current, sweeping those mutated DNA fragments right into the main circulatory river. And the study back that up, right? It did. Patients with liver metastases had a highly successful detection rate of 84.2%.
15:25Okay, so contrast that with a much smaller isolated facility far away from the main river. Say a metastasis that's located only in the lung. It's like a small factory off the grid, shedding very little waste into a much smaller local stream.
15:39Yeah, that's a great way to visualize it. And by the time that tiny amount of DNA reaches the main river where the blood draw happens. It's so diluted, it's practically invisible. Exactly. The studies show that patients with lung only metastases had a surprisingly low detection rate of just 41.5%.
15:54Wow. So liquid biopsy really is constrained by the anatomical reality of the disease. If the tumor isn't shedding into the blood, the machine can't sequence it. Right. But in the cases where shedding was adequate, The rapid blood test proved its true clinical superiority.
16:09It didn't just match the old tissue records. It actually found 63 plasma only alterations across 18% of the patients. Wait, plasma only, meaning these are mutations that simply did not exist in the archive tissue biopsies.
16:24Exactly. The tumor had learned new tricks that the old wax blocks knew absolutely nothing about. So these were dynamically acquired mutations. Yes, they represent the tumor's real time evolutionary response to being attacked by drugs.
16:36The researchers actually noticed a distinct pattern here. Patients who had recently been treated with targeted therapies or immunotherapy were far more likely to harbor these brand new plasma mutations.
16:46How much more likely? Well, 24% of them had new resistance mechanisms. Compared to only 10% of patients who were just receiving broad spectrum standard chemotherapy. That's a huge jump. Why does the type of treatment change how the tumor mutates?
17:00Like, what is it about targeted therapies that forces the cancer to evolve so aggressively? It's the biological principle of selective pressure. Think of standard chemotherapy like carpet bombing. It indiscriminately kills rapidly dividing cells.
17:16Okay. Targeted therapies, however, are like sniper rifles. They are engineered to hunt down and kill cancer cells that display one very specific genetic profile. Right much more precise. Exactly. So when you administer a targeted drug, you effectively wipe out all the susceptible cells in the tumor.
17:34But if even a microscopic fraction of the tumor cells happen to possess a random natural mutation that makes them immune to that specific sniper rifle, they survive. And suddenly, all their competition for oxygen and nutrients is completely dead.
17:47Exactly. The resistant cells are the only ones left, and they have unlimited resources, so they multiply rapidly, taking over the entire tumor mass. And because they're growing so fast, they shed massive amounts of their newly mutated DNA into the bloodstream.
18:02You got it. And relying on the four-year-old tissue biopsy would tell a doctor to keep using the sniper rifle, completely oblivious to the fact that the tumor is now completely immune to it. That is wild.
18:15This decentralized blood draw caught that resistance developing in real time, giving the clinical team the intelligence they needed to pivot to a different drug immediately. It's a staggering clinical victory, but we do have to return to that filter we discussed earlier, the Clonal Hamato Poises, the ghosts and the healthy white blood cells.
18:34Oh, right. What happened when the researchers turn the filter off and just looked at the raw plasma data? The results were alarming, honestly. Without filtering the plasma result against the matched normal white blood cells.
18:4511% of the total genetic variants found by the machine would have been falsely categorized as cancer mutations. One in 10. One in 10 mutations reported by the machine would be a complete phantom. So if a local community hospital buys this sequencer, plugs it in, and just runs plasma samples without taking the extra time and money to sequence the healthy blood control, they are generating a massive error rate.
19:08And the clinical consequences of that error rate are severe. If a doctor sees a phantom mutation on a report, and assumes it belongs to the tumor, they might prescribe a highly toxic, incredibly expensive targeted therapy.
19:23And the patient endures all the severe side effects of the drug, but receives absolutely zero anti cancer benefit. Exactly. Because the tumor never actually had the mutation the drug was designed to attack in the first place.
19:34Man. So when you step back and look at what this all means for the broader medical landscape, the implications are really dual cited. Very much so. On one hand, this study proves that democratizing genomic profiling is entirely feasible.
19:47A decentralized sequencer in a local hospital can successfully and rapidly identify actionable resistance related mutations in almost a 5th of advanced cancer patients, saving weeks of critical waiting time.
19:59Right. It bypasses the severe limitations of degraded, outdated tissue archives and provides a real-time snapshot of how the disease is evolving right now. But I have to play devil's advocate and challenge the ultimate endpoint here.
20:13With a 99% success rate and a 24 hour turnaround, why shouldn't the medical field just declare victory, phase out traditional tissue biopsies entirely, and just move everyone to local blood draws? Because doing so would abandon a significant portion of patients.
20:29Tissue is not obsolete, and it goes right back to the shedding dynamics we explored earlier with the river analogy. Because some tumors just don't shed enough. Exactly. A negative blood test in a patient with a small tumor or an isolated lung atastasis does not mean that cancer isn't mutating.
20:45It simply means the cancer isn't shedding enough DNA into the blood for the machine to actually detect it. Right. The absence of evidence in the blood is not evidence of absence in the tumor. That's exactly right.
20:55If the blood comes back negative, the medical team still has to go digging into the tissue. A tumor 1st or a reflex tissue test remains absolutely mandatory for low shedding cases. Okay, that makes sense.
21:06Furthermore, blood tests really struggle with complex structural changes in the genome, point mutations, you know, single letter typos in the DNA. Those are easy to find in the plasma. But larger changes are harder.
21:19Much harder. Copy number variations, where entire massive paragraphs of a chromosome are duplicated or deleted by the cancer, are incredibly difficult to piece together from the tiny fragmented confetti DNA floating in the blood.
21:33Ah I see. So for complex structural profiling, a solid piece of tumor tissue remains the unquestioned gold standard. So they're really complementary tools. The blood test acts as a rapid, real-time scout, and the tissue biopsy serves as a comprehensive structural map.
21:49They work in tandem, yeah. But the absolute, most critical warning this paper issues to the medical community is about rigor. The filter. The filter. As this decentralized technology spreads to smaller clinics.
22:01Those clinics must adopt comprehensive filtering strategies. They have to run the matched normal blood alongside the plasma to filter out the clonal hematopoasis. The technology is rapid and brilliant, but interpreting the biology requires meticulous discipline to avoid harming the patient with false positive.
22:18Plaster technology demands even stronger safety protocols, right? Exactly. So to summarize the central insight here, decentralize rapid genomic profiling of cell-free DNA from a simple blood draw is a highly successful transformative tool for identifying actionable cancer mutations, especially when traditional tissue biopsies are too old or too degraded to read.
22:38However, its true clinical power relies entirely on understanding a tumor specific biological shedding patterns and meticulously filtering out the background genetic noise of aging healthy cells. That's the perfect takeaway.
22:52It provides the intelligence we need right when we need it, provided we know how to verify the signal. Absolutely. What does this mean for the future of personalized medicine where the timeline from cancer recurrence to a life-saving targeted treatment decision could drop from weeks to a single day?
23:06It's an exciting frontier, that's for sure. This episode was based on an open access article under the CCBY 4.0 license. You can find a direct link to the paper and the license in our episode description.
23:18If you enjoyed this, follow or subscribe in your podcast app and leave a 5 star rating. If you'd like to support our work, use the donation link in the description. Now, stay with us for an original track, created especially for this episode, and inspired by the article you just heard about.
23:32Thanks for listening and join us next time as we explore more science, base by base. Late night lights on the quiet floor. A tiny whisper in the bloodstream's roar, no scalpel map. Just a coded trace. Turning shadows into a target's face.
24:04Running fast to keep the hands off the wheel. Uh, built in checks for what's fake out's real. Oh, if the signal's in, we're in between. And comfort tissue I need to see. Read my blood. Let the answers go.
24:27Find the switch. Find the way to go. Oh, action in the light resistance in the frames can't fight, no rock calling of the name. Don't guess, just know. Some room shout, some lesions hide, some room shout, some roots light up, some slip inside, ah, ah, after the therapy, loose, sparks appear.
24:54Plasma only stories. Suddenly clear. But there's a trick in the older code. Look alike changes the body holds. So we filter the echoes We steady the view, true from the drift and the drift from you. Let the answers blow Find a switch.
25:14Find a way to go. Faster, returns more chances to aim, catch it in time, rewrite the game. Read my Don't guess, just know.
