跳到正文
原文
MedPage Today·· 5 天前AI 评分45

《Lancet》揭示成人1型糖尿病三种亚型

Early-Onset Colorectal Cancer; AI-Enabled Prior Auth; Type 1 Diabetes Subtypes

AI 导读

《Lancet》报道,德国糖尿病研究中发现成人1型糖尿病存在三种亚型:经典型和混合型,后者比前者早发病约6年。研究通过抗体、血脂等指标区分亚型,并发现不同亚型在低血糖、视网膜病变和心血管自主神经病变风险上存在差异。研究结果提示早期识别这些亚型有助于免疫调节治疗,可能延缓或阻止疾病进展。

正文

— Also in TTHealthWatch: T cells and ALS

November 14, 2026 • 9 min read

TTHealthWatch is a weekly podcast from Texas Tech. In it, Elizabeth Tracey, director of electronic media for Johns Hopkins Medicine in Baltimore, and Rick Lange, MD, president of Texas Tech Health El Paso, look at the top medical stories of the week.

This week's topics include early-onset colorectal cancer, artificial intelligence (AI)-enabled prior authorization, T cells and amyotrophic lateral sclerosis (ALS), and different metabolic subtypes of adult type 1 diabetes.

Program notes:

0:41 Subtypes of type 1 diabetes in adults

1:41 Three distinct subtypes

2:41 Not investigated before?

3:41 Reduced islet cell autoimmunity

4:14 Early-onset colorectal cancer

5:15 Sidedness in body

6:16 Increased stromal-to-epithelial ratio

6:50 AI-enabled prior authorization

7:50 Was a change in frequency of procedure

9:10 T cells in ALS

10:10 Cerebrospinal (CSF) versus circulation

11:20 Cytomegalovirus interacts with existing conditions

12:52 End

Transcript:

Elizabeth: What are the differences between early-onset colorectal cancer and that that occurs in older people?

Rick: Does AI-enabled prior authorization change clinical practice?

Elizabeth: What's the role of T cells in ALS?

Rick: And different metabolic subtypes of adult type 1 diabetes.

Elizabeth: That's what we're talking about this week on TTHealthWatch, your weekly look at the medical headlines from Texas Tech University Health Sciences Center in El Paso. I'm Elizabeth Tracey, a Baltimore-based medical journalist.

Rick: And I'm Rick Lange, president of Texas Tech Health El Paso.

Elizabeth: Rick, why don't we turn to The Lancet? I find this notion of, gosh, subtypes of type 1 diabetes to be pretty intriguing.

Rick: Most often when we talk about type 1 diabetes, we talk about it in kids, and it's an autoimmune disease. There are antibodies that form against the beta cells that cause their death and decrease insulin, and therefore result in diabetes. But the same thing can happen in adults. It looks like there's a lot of different heterogeneity, different clinical presentations. Are these distinct disease entities, or are they just different stages of a shared autoimmune process?

They did a prospective cohort study. They analyzed data from adults aged 18 to 69 that were in the German Diabetes Study that had clinically diagnosed type 1 within the preceding 12 months. And then they asked whether the results were reproducible in a prospective follow-up of the same patients. They assessed different antibodies, their triglycerides, their HDL, and tried to characterize clinical and metabolic subtypes. And then they looked at insulin sensitivity, residual B-cell function, and looked at the trajectory.

And what they discovered is that there were three distinct subtypes. Two had antibodies to glutamic acid decarboxylase -- I'm going to call it GADA -- and one was GADA-negative. I want to focus on the GADA-positive. They characterized these into two different groups. One was what they called classic type 1 diabetes. These were adults that had a lower BMI [body mass index], high insulin sensitivity, and severe B-cell dysfunction. And the second group that were GADA-positive, they had a higher BMI, lower insulin sensitivity, and they had better B-cell function. They call those the mixed type 1 diabetes.

It appeared that the mixed type 1 diabetes was actually just an earlier form of the classic, about 6 years earlier. If we can identify these individuals, and if we know it's an immune response, perhaps doing immunomodulatory therapy can change things.

Elizabeth: A couple questions about this. I'm always wondering about, gosh, why didn't anyone ever look at this before to see about this heterogeneity, which certainly we see every time we look at something that presents clinically in a certain way. Gosh, there's an awful lot under there; that's that iceberg notion. The other thing is you didn't mention the negative, the other group.

Rick: Yep. It takes a large number of patients, a large amount of data, and you have to do a cluster analysis. And if we take these similarities, then what are the differences between this similar group and that similar group?

Not only are there differences with regard to GADA and the timing of when the diabetes is likely to occur, there was also differences with the risk of hypoglycemia, retinopathy, and cardiovascular autonomic neuropathy. And in fact, they were higher in the mixed group than they were in the classic group.

With regard to the GADA-negative group, had other different antibodies that were responsible for their diabetes. The B-cell function was intermediate between the other two groups. There was a more gradual beta-cell decline, a lower diabetes genetic risk, and reduced islet cell autoimmunity, suggesting it was pathogenetically different and also heterogeneous.

Elizabeth: Clearly, this has important clinical implications and needs to be distilled into something that can be just tested pretty routinely and then employed.

Rick: Yep. When you identify an adult with type 1 diabetes, if we can catch them very early in that disease, when they're in the mixed-diabetes group, we can change their immune system and either delay or prevent progression of the disease.

Elizabeth: Let's turn from here then, speaking of things that happen early in life, to Nature Medicine. And this is a study looking at insights into early-onset colorectal cancer from a single-nucleus RNA-sequencing atlas of colorectal cancer across the age continuum.

What these folks did was they looked at 189 primary colorectal cancer tumors spanning eight decades of life with at least 20 tumors in most age decades. They did a number of analyses on this, stratifying tumors by their microsatellite status, the tumor sidedness in the person, the stage, the patient's sex, neoadjuvant treatment history.

What they found was that tumors from the younger patients were characterized by increased stromal and decreased epithelial proportions. They confirmed that in an external cohort of 712 colorectal cancer tumors. Epithelial cells from the younger patients demonstrated increased expression of what are called long non-coding RNAs, a trend toward increased heterogeneity and copy number alterations. Left-sided colorectal cancer tumors had more cell-to-cell interactions.

What they basically conclude is yes, indeed, there are things that we can look at. This stromal versus epithelial proportion and the sidedness seem to be distinguishing characteristics, and maybe help further understand why this is happening more often in younger people.

Rick: The incidence of colorectal cancer before the age of 50 has been increasing about 1 to 3% per year. It is now the leading cause of cancer death in adults 18 to 49 years of age. Unfortunately, the survival of individuals with early colorectal cancer is poor. Part of it's poor because 58% of these individuals are in late stages of the disease, but also, as you mentioned, we're trying to distinguish are there characteristics from the tumor that make it more resistant to therapy?

This increased stromal-to-epithelial ratio -- there's more connective tissue and less cells -- makes it more resistant to standard chemotherapy. So identifying these individuals early and identifying unique markers that may be more responsive to particular types of therapy are important.

Elizabeth: Clearly points to the need to develop new treatments that are targeted toward these particular characteristics versus the ones that have been classically identified.

Rick: Are these targets of treatment, and are they actually driving the tumor, or are they just bystanders?

Elizabeth: Let's turn to JAMA.

Rick: AI-enabled prior authorization and particularly in Medicare.

This is really controversial because prior authorization requires the clinician to submit a justification for a specific medical service. The reason it's controversial is commercial insurers use this as a way of denying insurance and denying payment. If the procedure is necessary, that's really frustrating to the patient and to the physician, obviously. And in the past, Medicare has not required any prior authorizations with rare exceptions.

However, in January of this year, CMS initiated what's called the Wasteful and Inappropriate Service Reduction, called the WISeR model, talking about prior authorization, to make sure that the procedure that's being done is not wasteful and it's not inappropriate because these are services that are considered vulnerable to fraud, waste, and abuse.

So what Medicare said is, OK, we're going to use AI-enabled prior authorization for some particular procedures, one being epidural steroid injection. Before we're going to reimburse hospitals or physicians for this particular procedure, we're going to submit these individuals to prior authorization using AI to do that. Did that really change the use of epidural steroid injection? And by the way, we're going to compare the six states in whom AI authorization was enacted with all the other states.

When they looked at it, the use of AI authorization led to a 15% reduction relative to the baseline use of epidural steroid injection. Did the other procedures that didn't require authorization change? Not at all. Just the ones in whom prior authorization was used. It looked like there was a dip and then there was a recovery. The question is, over the next 6 months, will that maintain or not? Is it just because there's a new process and it slowed things down, and now it's going to recur? Or does it actually do what it's supposed to do? Make sure the test is both appropriate and not wasteful.

Elizabeth: As we know, AI models train. And so one question I have is, as this thing trains and it gathers more data, and perhaps there are more people who push back and say, oh, wait a second, this is really indicated here, there might be a training impact as well.

Rick: Yeah. You're right. So this is a 6-year project. So we're going to want to follow it for a long period of time. This is just the initial results from the first two quarters, more data to come.

Elizabeth: OK. Turning back to Nature Medicine then, let's take a look at this notion of T cell involvement in amyotrophic lateral sclerosis, or ALS.

This study is looking at T-cell responses in folks who have it. Previous research has implicated T cells here. And what they're using is single-cell RNA and T-cell receptor sequencing, longitudinal immune profiling, and imaging flow cytometry to identify long-lived intrathecal clonotypes and antigen-presenting T-cell complexes that have features of immunological synapse formation.

What they find is that, sure enough, in the CSF, in the cerebrospinal fluid, they do find these higher frequencies of different subtypes of T cells there that were associated with slower disease progression. Whereas when they noted that these same subtypes of CD4 cells, the T cells, were found in the circulation, they were associated with faster progression of ALS.

They did find that there were also these differences in T-cell clonotypes that were shared between the blood and the CSF exhibited distinct transcriptional states. They also noted that cytomegalovirus-specific T cells in the cerebrospinal fluid and serum antibodies were associated with faster progression and shorter survival in the sporadic form of ALS, which is the majority of the cases, I would note, versus that that's genetically associated. I, at least, am happy to see further characterization of this disease and potentially ways that might point to an intervention.

Rick: Yeah. Neuroinflammation is now recognized as being an important part of the central feature of ALS. About three-fourths of the cases are sporadic, about a fourth are from mutations, and ALS from a complex interplay between both genetic susceptibility and environmental influences that causes neuroinflammation. And, by the way, cytomegalovirus probably isn't the thing that initiates it, it just interacts with all this together. The fact that we can identify different types of T cells, those that regulate or those that help promote inflammation, those different populations in the blood and in the brain hopefully will give us some insight in how best to modulate the neuroinflammation.

Elizabeth: Or even to slow it down. I was gratified to see that the authors of this study did look at other neurotropic viruses and assess their role. So they looked specifically at herpes viruses and they found that this was specific to cytomegalovirus.

One of the questions I have, of course, is that cytomegalovirus is just so very common. What is that interaction? And if it was really a pronounced interaction, I would think that ALS would be a good deal more common than it is.

Rick: Yeah. It's not a uniform driver of pathology, but it interacts with the underlying disease mechanisms. It's a very complicated interplay in terms of the neuroinflammatory response.

Elizabeth: More coming, no doubt.

On that note then, that's a look at this week's medical headlines from Texas Tech. I'm Elizabeth Tracey.

Rick: And I'm Rick Lange. Y'all listen up and make healthy choices.

来源:MedPage Today · medpagetoday.com