A retrospective study of 701 memory clinic patients tested whether stratifying by age at onset and family history enriches for diagnostically relevant genetic findings. Using an adapted Goldman-score classification with exome sequencing and targeted genotyping in high-risk cases, the authors increased diagnostic yield and evaluated implications for APOE and C9ORF72 testing.
0:00Welcome to Base by Base, the papercast that brings genomics to you wherever you are. Thanks for listening, and don't forget to follow and rate us in your podcast app. Absolutely. Thanks for tuning in. So for today's deep dive, I want to start by posing a real world dilemma, write to you, the listener.
0:16What really happens when a memory clinic decides to completely rethink exactly who gets genetically tested for dementia? Oh man. That question, it honestly completely changes the entire landscape of patient care.
0:30Right. Because, I mean, in the past, genetic testing for dementia was often just actively discouraged. Yeah, highly discouraged. It was viewed by so many in the medical field as this, like, insanely expensive endeavor with super long turnaround times, and, frankly, very low success rate.
0:47Exactly. But now, you know, with these revolutionary and new therapies actually on the horizon, knowing your genetic makeup isn't just about, well, predicting the future anymore. No, it's not. It could determine whether a new treatment might actually help you or, and this is the scary part, actually put you at serious risk.
1:03The stakes have literally never been higher for getting this biological assessment right. Totally. Which brings us to the core mystery we are unraveling today. How could changing a single age cutoff in a clinical checklist prevent doctors from missing life altering genetic clues?
1:21I mean, changing a single line item sounds like such a minor administrative tweak, right? But as we'll see, it is the difference between catching a highly penetrant genetic risk factor and just, you know, letting it slip right past the clinical radar entirely.
1:35is wild. So today we celebrate the work of Teresa Connick, Elizabeth Stugman, and their extensive research team at the Medical University of Vienna, who have advanced our understanding of identifying dementia patients with diagnostically significant genetic findings.
1:49Yeah, the clinical research they've compiled here is just rigorous. It's really impressive For sure. And just for some context, this work is accepted in June 2025. and published in genetics and medicine, which is an official journal of the American College of Medical Genetics and Genomics, where the ACMG.
2:05Right. And to truly grasp the gravity of this research, we really have to look at the sheer scale of the scientific problem that this Vienna team tackled. Yeah, because we throw the word dementia around a lot, right?
2:18But it's actually an umbrella term for several distinct physical changes in the brain. Exactly. If we connect this to the bigger picture, 1000000s of people globally are affected by these various conditions and the vast majority of cases fall under Alzheimer's disease or AD.
2:35Which is what, like more than half? Yeah, about 50 to 60% of the total. In Alzheimer's, we're looking at the accumulation of amyloid plaques and toutangles that essentially disrupt cellular communication.
2:46Right, the blacks and tangles we hear about those a lot. Yeah. And then you have vascular dementia at about 25%, which is driven by impaired blood flow to the brain. Oh, okay. Louis body dementia sits at about 15%.
2:58That one is characterized by abnormal deposits of a protein called alpha synucleene. And finally, frontotemporal dementia or FTD makes up another 15%. That's where we see profound atrophy in the frontal and temporal lobes, which, you know, drastically alters a person's personality and language.
3:16Yeah, FTD is just brutal. And a huge portion of the general public actually has a family history of one of these diseases, which naturally elevates your own risk. Significantly, yeah. So if you have that family history, you might logically walk into a clinic and say, hey, can you test my DNA?
3:32I want to know what I'm facing here. But historically, clinicians hesitated to order those tests. They really did, because the diagnostic yield, meaning, you know, the percentage of times a genetic test actually found a definitive, clear cut cause for the disease, was just frustratingly low.
3:47Like, how low are we talking? Clinicians were finding a true pathogenic variant, only about 16 to 30% of the time. And that's even in patients who develop symptoms unusually early in life. Wow. So testing everyone was basically like looking for a needle in a haystack while blindfolded.
4:03That's a perfect way to put it. Expensive and frustrating. But now, patients desperately need answers because of emerging personalized medicines, right? Exactly. That historical hesitance created a massive gap in our data.
4:15For years, testing everyone just felt economically impossible and clinically futile, but the shift and relevance we're seeing right now is profound. Because it actually changes the treatment plan now. Yes.
4:26Genetic analysis is no longer just this academic exercise to satisfy your curiosity. It is crucial clinical data that dictates medical intervention. And this urgency is particularly acute for early onset dementia, which the medical community generally defines as showing cognitive symptoms before the age of 65.
4:45Okay, and early onset cases are the ones with much higher heritability, right? We are looking for those highly penetrated variants. Yeah, exactly Let's actually make sure we have our definitions locked in for everyone listening.
4:56When we say a gene is highly penetrant. We basically mean that if you possess this specific genetic mutation, you were overwhelmingly likely to develop the disease, right? Regardless of your lifestyle or environment.
5:07Yeah, that captures the biological reality perfectly. It's almost a certainty. In these highly penetrant early onset cases, we're often looking at autosomal dominant inheritance. Okay, so autosomal dominant means, if you get just one copy of the mutated gene from one parent, you are almost guaranteed to get the disease.
5:27Yep, just one copy. and often very early in life. Very early. And because it only takes that one copy, the family history is usually pretty striking. You'll see it cascading through generation after generation.
5:37Okay. For Alzheimer's, we specifically look for autosomal, dominant mutations in genes like APP, PSEN one, and PSEN2. Okay. And for fronto temporal dementia, the major culprits are mutations in the MAPT gene and the programuline gene.
5:53Got it. So we find ourselves in this incredible tension, right? We desperately need to find these specific genetic markers because they increasingly dictate which emerging personalized medicines a patient can even receive.
6:05Exactly. But we still can't afford to just blindly sequence the entire genome of every single person who walks into a clinic complaining of memory loss. No, the healthcare system would collapse. It's too expensive.
6:15So how did the researchers at the Medical University of Vienna actually build a better sorting hat? Like, how do they fix the haystack problem? Well, to solve the systemic inefficiency, they designed a brilliant retrospective study, they went deep into their own memory outpatient clinics database and analyzed 701 well-characterized dementia patients.
6:34Oh, 701. Yeah, these were patient seen between 2017 and 2022. And they were meticulously documented cases. The researchers had access to detailed family histories, verified ages of symptom onset, comprehensive neuropsychological testing results, and even structural MRI imaging for these individuals.
6:53Okay, so 701 patients is a really robust cohort to analyze. But they still needed a filtering mechanism to figure out who among those 701 actually warranted the expensive genetic sequencing, right? Right.
7:05And the innovative step here was exactly how they quantified that familial risk before a single drop of blood ever went to the sequencer. Okay, what did they do? They utilized a framework called a modified Goldman score.
7:17It's basically a numerical scale ranging from one to 4.5. And it essentially grades the objective strength of a patient's family history based on previous generations. I was looking at this scale in the paper, and it is beautifully pragmatic.
7:32Like, a score of one indicates a very strong familial background. That means you have at least 3 affected family members spanning over 2 generations. Yep, very clear lineage. And then a score of 4 means a known, definitively negative family history.
7:48Meaning you have absolute confirmation that no one else in your family have the disease. Right, you've ruled it out. But the addition of the 4.5 score is what really caught my eye here. A 4.5 indicates an unclear or censored family history.
8:00Yeah, that 4.5 score is brilliant. It acknowledges the messy reality of human record keeping. Relatives lose touch, families emigrate and lose their medical records or, you know, ancestors pass away early from other causes before dementia could ever even manifest.
8:14Right, think about your own family tree. To the person listening right now. If you had a great aunt who passed away in a nursing home in the 1980s, her medical records might just say old age or senility.
8:26Exactly. Nobody was running genetic panels on her back then. That missing data creates a massive blind spot in your own genetic risk profile today. And that is why this 4.5 score is so vital. It actively accounts for the ghosts in our medical history, rather than just assuming a lack of data means a lack of risk.
8:44Beautifully said. So the researchers took that nuanced family history score, and combined it with the patient's age at onset, or AAO. Okay, and this brings us to the core mystery we set up earlier. Previous diagnostic models, specifically this highly regarded clinical scheme developed by a researcher named Coryef, they set the strict cutoff for early onset at 60 years of eight.
9:05Yes, 60 was the golden rule. But this Vienna team deliberately rebelled against that standard, they widened the net and changed the early onset cutoff in their filter from 60 years old to 65 years old.
9:17And I just have to push back on this a bit because adjusting an age dial by just 5 years seems like a rounding error, like this adding 60 months to a clinical guideline, really changed the fundamental biology of who we catch.
9:29I know it sounds tiny. But the outcome of that specific five-year shift is honestly the crux of their most surprising finding. We'll analyze exactly why it matters biologically in a second. But first, let's observe how this new grid functioned in practice.
9:43Okay, lay on me. By combining the Goldman score with that controversial 65 year age cutoff, the researchers computationally filtered their massive cohort. They flagged exactly 51 patients out of the initial 701 as being high risk for carrying a clinically relevant genetic variant.
9:59Wow. So they successfully narrowed a haystack of 701 down to just 51 high probability patients. Yes. That alone is a massive win for clinical efficiency. Huge win. And from that high risk pool, 38 patients had complete data profiles and actually provided the necessary consent for advanced analysis.
10:17Okay, 38 patients. Right. And once you narrow the pool down to just 38 individuals, a clinic can finally afford to use the most rigorous, multi-layered genetic tools available. Yeah, you bring out the big guns.
10:29Exactly. They deployed whole XO sequencing to read the protein coding regions of the DNA. They performed targeted APOE genotyping. And critically, they utilize a specialized two-step PCR and southern blotting technique specifically designed to find C90RF 72 repeat expansions.
10:49See, I'm looking at the methodology here, and they didn't just use standard XOM sequencing. They added southern blotting to look for that C9ORF 72 variant. I always thought whole XM sequencing was like the absolute gold standard of modern genetics.
11:02A lot of people think that, yeah. So why go through the extensive trouble of adding this older labor intensive blotting technique? Why wouldn't the XM sequence just catch everything? Oh, this is where it gets fascinating.
11:12It comes down to the physical architectural structure of our DNA. Whole XM sequencing only looks at the Exxons, which is the roughly one to 2% of our genome that actually provides the instructions to build proteins.
11:23Wait, only one to 2%? Yep. It completely ignores the vast stretches of non-coding DNA. And furthermore, exome sequencing operates by reading the DNA in very short fragmented pieces, and then stitching the data together computationally.
11:39So it's highly effective for finding tiny spelling mistakes. like a single incorrect letter in the genetic code. Okay, I think I see where this is going. If exome sequencing is looking for single letter typos.
11:52What exactly is a repeat expansion in that C 9 0 air, a 72 gene? So a repeat expansion is a massive stuttering repetition of a DNA sequence. And it's often located in those exact non-coding regions that exome sequencing just ignores?
12:07Oh, wow. Because the standard sequencer reads in short fragments, it literally trips over these giant structural anomalies. The computational software cannot figure out how to stitch the repeating fragments together, so it just discards the data.
12:19That's crazy. Right. So you need the specialized PCR to amplify that specific region. And then you use southern wadding, which involves running the DNA through an electrically charged gel matrix, and that physically separates the DNA fragments by their actual size and weight.
12:35Okay, so standard X own sequencing is basically like spell checking a book, word by word. It is fantastic for finding a typo on page four. But a repeat expansion is like someone accidentally copy pasting an entire chapter 50 times in the middle of the book.
12:51Yes. The spell checker reads the words, sees that they're all spelled correctly, and just misses the massive structural error entirely. Exactly. So southern blotting is like putting the physical book on a scale and immediately realizing, oh, this is £10 too heavy.
13:05That analogy perfectly illustrates the mechanical limitations of our sequencing technology. Without the scale, you just missed the extra chapters. That's incredible. So they have their 38 highly targeted patients and they deploy this multi-layered testing protocol, the spell checker and the scale.
13:21Did the combination of the Goldman score and that controversial 65 year age cutoff actually reveal the genetic variants the old system was missing. Because the raw numbers from their findings here are incredibly striking.
13:34really are. First, they establish that family history is far more common than many assume, right? It was reported in 42% of their Alzheimer's patients and 48% of their fronto temporal dementia patients.
13:47Yeah, seeing nearly half of the cohort with a familial link highlights exactly why memory clinics are currently just overwhelmed with patients asking about their genetic risk. Absolutely. But the most impactful data point is the leap in diagnostic efficiency.
14:03Before implementing the structured classification scheme, when doctors relied purely on their unstructured subjective clinical judgment to decide who to test... Just going on their gut feeling. Right, their diagnostic yield was sitting at roughly 20%.
14:18Which means 80% of the time. Families were enduring the emotional toll of waiting months for test results, only to be told the doctors still had no idea what was causing the disease. It's devastating. But by switching to this new targeted classification scheme, the Goldman score paired with the age of onset, the clinic's diagnostic yield skyrocketed to 39%.
14:3815 out of the 38 fully analyzed high risk patients had diagnostically relevant actionable variants. They practically doubled their success rate just by organizing the data they already had. Exactly. They uncovered critical disease causing variants in genes like APP, PSE, and one, APT, and per granulin.
14:57But we have to circle back to the age cutoff. Ah, yes, the 5 year difference. Right. Did bumping the dial from 60 to 65 actually matter in the final analysis? It mattered immensely. And it really exposes how rigid clinical guidelines can inadvertently harm patient discovery.
15:14Okay, tell me. If the Vienna clinic had strictly adhered to the 60 year cutoff from that older Koreath model, They would have completely missed 4 patients who carried the highly penetrant APOA 44 genotype.
15:25You're kidding. Nope. And they would have missed one patient carrying the C9 ORF 72 mutation. Wow. So it's like adjusting the sensitivity on a metal detector. By widening the sweep just slightly to age 65, They captured the most critical treasure actionable genetic data that would have otherwise slipped right through the cracks.
15:43That's exactly what happened. But wait, why did those specific APOE 44 patients slip past the 60 year mark in the 1st place? I thought these mutations caused early onset. Right. So this is where we must distinguish between different types of genetic dominance.
15:57We discussed autosomal dominant genes like PSEN one earlier, where one bad copy aggressively drives the disease, usually manifesting very early, often in a patient's 40s or 50s. The APOE 4 gene, however, operates differently.
16:13Being an APOE 44 homozygo, meaning you inherited a copy of the APOE 4 variant from both your mother and your father, is considered semi-dominant. Okay, so a semi-dominant gene means its effects are more cumulative over time, which naturally pushes the onset of symptoms a bit later in life.
16:28Exactly. The biological burden builds more gradually. In this specific Alzheimer's cohort, the median age of symptom onset for those carrying the EOE 44 genotype was 68 years. Oh, wow. The onset for these patients was actively occurring between the ages of 60 and 68.
16:43A rigid cutoff of 60 simply leaves them off the radar because their biology operates on a slightly delayed timeline compared to the autosomal dominant mutations. Right, they were just missing them completely, which brings us to, honestly, the most urgent clinical question of our time.
16:57What does all of this mean for the patient sitting in the clinic today? Why is finding that APOE 44 genotype so incredibly critical right now in this specific moment in medical history? Because we have officially entered the era of anti-ameloid therapies.
17:14Yes, the new drugs. Exactly. Monochormal antibody drugs like Lecanamab and Donamab are making global headlines for their unprecedented ability to actively clear amyloid plaques from the brain in patients with early Alzheimer's.
17:27Which is amazing. It is, but there is a severe biological catch. Patients who carry 2 copies of the APOE4 gene, the exact APOE 44 genotype we just discussed catching, with the 65 year cutoff, they face significantly higher risks of potentially life-threatening side effects from these exact drugs.
17:45Right. You were talking about ARIA, right? amyloid related imaging abnormalities. Yes. ARIA is the clinical term for the adverse events we see on MRI scans after administering these drugs. And it basically manifests in 2 primary forms.
17:58There's ARIAE, which involves severe swelling or edema in the brain tissue, and there's ARAH, which involves microbleeds or even larger catastrophic hemorrhages on the surface of the brain. That is terrifying.
18:10We really have to explain why this happens because this is where the biology gets truly intense. Why does a drug designed to cure the brain, suddenly cause it to swell and bleed specifically in people with the APOE 44 genotype?
18:24It all comes down to the mechanism, which is tied to where the amyloid plaques are physically located. You see, amyloid doesn't just clump up harmlessly in the empty spaces of the brain tissue. Right. It actively builds up inside the walls of the brain's blood vessels, a condition known as cerebral amyloid angiopathy inside the actual vessel walls.
18:45So when these powerful new monoclonal antibodies enter the brain, they bind to the amyloid to dissolve it. But in doing so, they are literally stripping the structural amyloid out of the blood vessel walls.
18:56Oh my god. So the drugs are pulling the mortar right out of the brick wall of the blood vessels. Exactly. And individuals with the APA 44 genotype have a much more aggressive inflammatory response from their brain's immune cells called microglia when the stripping process occurs.
19:11This hyperinflammatory reaction leaves the blood vessels incredibly porous, fragile, and prone to leaking fluid, which is the edema, or rupturing entirely, which is the hemorrhage. Wow. This is exactly why knowing your genetics is no longer just about predicting if you'll get the disease.
19:29It is literally about knowing whether the cure might actually hurt you. Yes. The risk is so pronounced that international regulatory bodies are taking unprecedented action. Really? Like who? Well, when the European Medicines Agency, the EMA, reviewed Laquemby, which is the brand name for LaCanimab, their regulatory approval explicitly excluded APOE 44 patients from receiving the treatment.
19:52Wow, just completely excluded. Yes. The safety profile is simply too hazardous for that specific genetics subgroup. Okay, hearing the mechanism of ARIA. It makes it completely obvious why capturing those APOE 44 patients by shifting the age dial to 65 is an absolute game changer.
20:07But I do have to push back on the studies overall framework just a bit. Because if this goldming score, an age cutoff filter is so effective at saving money by narrowing the testing pool, aren't we still inevitably leaving some people behind?
20:21It's a fair point Like, what about the individuals who have a negative family history simply because they have incomplete medical records? If my grandfather died of a sudden heart attack at age 50, no one on earth would know if he was genetically destined to develop dementia at age 60.
20:37Under this strict tiered system, do I just get filtered out of the sequencing pool entirely? That is a highly valid structural concern, and it's actually one of the primary limitations the researchers themselves acknowledge in their discussion.
20:51Oh, they do? Yeah. Specialized memory clinics inherently operate with a selection bias. They attract highly specific cases. In the Vienna cohort, 37% of the patients had early onset Alzheimer's, but if you look at the general population, early onset cases only make up about 5 to 10% of all Alzheimer's diagnoses.
21:09Okay, so the clinic is already analyzing a heavily skewed population from the moment the door is open. Exactly. And the researchers explicitly warn that a documented negative family history should never definitively rule out genetic testing.
21:22Right, because of all those missing variables. Misdiagnoses in previous generations, early mortality from cardiovascular events, or lost medical records like your grandfather analogy, all of these factors create false negatives in a family history assessment.
21:36So if the filter isn't perfect, how do we balance the astronomical cost of genetic sequencing with the ethical mandate to make sure we don't miss people who desperately need answers? It's tough, but the authors suggest implementing a tiered approach to testing, particularly for fronto temporal dementia.
21:54Okay, how does that work? Well, because FTD is so highly heritable and presents with such wildly variable symptoms, the researchers argue that genetic testing should be considered for all FTD patients, regardless of their family history, whenever economically feasible.
22:08Wow, all of them. Yeah. And to manage the financial burden, clinics could initiate testing with cheaper panel sequencing. Right, panel sequencing is where you only look at a predefined limited set of genes known to cause dementia rather than sequencing the entire 20,000 genes in the human XO, right?
22:24Exactly. Panel testing serves as the wide cost effective net. If the panel comes back negative, but the clinical suspicion remains high, then the clinic can justify bringing up the whole Xome sequencing.
22:36That makes sense. Though it is imperative to remember that even panel testing must be paired with that specialized southern blotting technique to catch the physical weight of the C9 ORF 72 repeat expansions.
22:48scale for the book. Yep. In this study alone, that specific mutation accounted for a staggering 17% of their entire FTD cohort. 17%. is wild. It is genuinely amazing how complex and physically nuanced the genetic testing landscape has become.
23:04But this paper provides clinicians with such a clear, actionable roadmap out of the dark. It really does. It really distills down to a powerful central insight, right? By combining the biological age of symptom onset with a highly structured family history score.
23:18Memory clinics can effectively double their success rate in identifying critical life-altering genetic variants in dementia patients. Crucially, by bravely expanding that early onset definition from 60 to 65 years, clinics ensure they catch those semi-dominant, highly penetrant genotypes like APOE 44.
23:38Which changes everything for treatment. Exactly. In the emerging age of anti-amyloid therapies, discovering that specific genotype dictates the safety and efficacy of the most groundbreaking Alzheimer's treatments humanity has ever developed.
23:51It finally takes the blindfold off while we are looking for that needle in the haystack. Which leaves us with his final thought to mull over. What does this mean for the future of routine clinical checkups as our understanding of personalized genetic medicine continues to evolve?
24:05It's a huge question. It really is. This episode is 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. If you enjoyed this, follow or subscribe in your podcast app and leave a 5 star rating.
24:22If you'd like to support our work, use it a donation link in the description. Now stay with us for an original track created, especially for this episode, and inspired by the article you've just heard about.
24:31Thanks for listening and join us next time as we explore more science, base by base.
