NEWS
Young Colon Tumors Carry Extra Scaffolding That Resists Chemo
A 189-tumor atlas shows early-onset colorectal cancer carries more stromal scaffolding, a mismatch for extra chemo and screening that still starts at 45.
A 189-tumor cell atlas published on October 29, 2026, found that early-onset colorectal cancer in adults under 50 carries more support tissue and fewer cancer cells than tumors in older patients. The extra scaffolding is already linked, in other solid cancers, to chemotherapy bouncing off the tumor, a problem for a disease the American Cancer Society ranks as the leading cause of cancer death under 50.
Kimmie Ng, director of Dana-Farber Cancer Institute’s Young-Onset Colorectal Cancer Center, and colleagues at the Broad Institute and Mass General Brigham mapped gene activity in 850,222 nuclei. The map does not give clinics a new drug. It does show why dose-heavy regimens built for older, cancer-cell-heavy tumors keep falling short in the youngest patients.
Younger Tumors Carry More Scaffolding Than Cancer Cells
The paper in Nature Medicine is a single-nucleus atlas of 189 tumors taken at surgery, 54 from people diagnosed before age 50 and 135 from people diagnosed at 50 or older. Patients ranged from age 29 to over 90, with 20 or more samples in most decades. First author Caroline Porter, of the Broad’s Klarman Cell Observatory, used frozen tissue from Brigham and Women’s Hospital operations stored in Dana-Farber’s gastrointestinal biobank.
Among microsatellite-stable tumors, the common subtype, younger patients had more stromal cells and fewer epithelial cancer cells. Nuclei clustered by cell type, DNA-repair status, and which side of the colon the tumor sat on, not by a clean young-versus-old split. Age moved the mix of cells, not the identity of the cancer cells themselves.
Ng announced the atlas the day it posted.
Our atlas of young-onset CRC is finally out in @NatureMedicine! We found a unique biology in younger patients that may explain the aggressive nature of this disease. This was a monumental effort with colleagues at @MassGenBrigham and @broadinstitute. @DFarberYoungCRC https://t.co/JVLxChISWM
— Kimmie Ng, MD, MPH (@KimmieNgMD) October 29, 2026
“This is the largest single-cell atlas of colorectal cancer published to date, with the largest number of young patients under the age of 50,” Ng said. “In addition to having a large number of tumor samples, we also had samples from patients spanning the whole spectrum of age, including people in their 20s and people over 90.”
The team then asked whether the same mix showed up outside the 189-sample map.
THREE LAYERS OF THE SAME MIX
| Cohort | Tumors | What held with younger age |
|---|---|---|
| Single-nucleus atlas (MSS) | 134 (45 under 50) | More stroma, fewer epithelial cells |
| Bulk gene-activity set | 564 MSS (126 under 50) | Same epithelial drop and stromal rise |
| Stained tissue images | 712 MSS (18 under 50) | Odds 2.70 for a high stroma-to-cancer-cell ratio under 50 versus 70+ |
The image set came from the Nurses’ Health Study and the Health Professionals Follow-Up Study. Compared with patients 70 and older, the adjusted odds of a high support-to-cancer-cell ratio were 2.70 under age 50, 1.83 at ages 50 to 54, and 1.51 at ages 55 to 69. The trend ran down the age scale rather than snapping off at 50.
The Extra Treatment Young Patients Already Get
Ng’s clinic already treats these patients harder, and they are healthier at baseline. Survival still fails to pull ahead, and the gap is widest under 35.
Their survival, especially for the very youngest patients under the age of 35, may actually not be as good as somebody in their 40s or someone in their 60s, despite receiving more aggressive treatment and being at baseline healthier.
Kimmie Ng, director, Young-Onset Colorectal Cancer Center, Dana-Farber Cancer Institute
The paper points to a mechanical reason. A higher share of support tissue has been tied to chemotherapy resistance in several solid cancers. Dana-Farber’s write-up of the work says stromal cells act as scaffolding that can shield a tumor and keep drugs and immune cells from getting in. The authors write that this mix, plus high activity of programs that help tumor cells loosen and spread, offers a basis for the poor survival seen in early-onset disease, particularly the youngest patients.
That is a hypothesis, not a prescription to cut chemo. The 189 patients were not scored here for how their tumors responded to drugs. What the atlas does is recast the old paradox. Extra oxaliplatin in a healthier body still has to cross a thicker support layer than the older tumors those doses were tuned on.
Young patients in this cohort were also more often left-sided (odds 3.61) and more often stage IV at resection (odds 2.55). Prior studies cited in the paper found young patients 58% more likely than older patients to be diagnosed with advanced-stage disease. The stromal excess remained after the team accounted for stage, so later diagnosis does not fully explain the cell mix.
Half of Under-50 Cases Sit Below the Screening Age
The American Cancer Society and the U.S. Preventive Services Task Force lowered average-risk screening from 50 to 45 in 2018 and 2021. The atlas’s age slope does not match that line. Support-cell burden was highest under 50, still elevated at 50 to 54, and only then eased toward the pattern seen at 70.
ACS figures for 2026 make the gap concrete. Colorectal cancer is the third most common cancer in U.S. men and women and the second most common cause of cancer death overall. In people under 50 it is first, after mortality in that age group rose 1.1% per year from 2005 and moved from fifth in the early 1990s to first in 2023. Overall cancer deaths under 50 fell 44% from 1990 to 2023, from 25.5 to 14.2 per 100,000. Colorectal cancer was the outlier that climbed.
THE SCREENING GAP IN ACS FIGURES
- New cases in 2026: An estimated 158,850 colorectal cancers, including 24,640 in people from birth through 49, which is 16% of the total, and 55,230 deaths, including 3,890 under 50.
- Incidence trend: Rates rose 3 percent a year in ages 20 to 49 from 2013 through 2022, and 0.4% a year at ages 50 to 64, while death rates for all ages fell about 2% a year.
- Stage at diagnosis: Three in four cancers in people younger than 50 are already regional or distant, and 27% are distant.
- Who is invited: Half of under-50 diagnoses fall in ages 45 to 49, the new screening window, yet only 37% of people 45 to 49 are up to date with screening.
The other half of under-50 diagnoses sit below 45, where average-risk programs do not invite anyone. Those patients are also the ones whose tumors sit furthest along the stromal gradient. Nearly half of all new colorectal cancers now occur under age 65, up from 27% in 1995. The same mismatch shows up in cancers rising outside the screening age more broadly: incidence is climbing in people the calendar still treats as too young to test.
How the Atlas Controlled for Confounding
Older comparisons often split patients at 50, mixed DNA-repair subtypes, and ignored which side of the colon was involved. This team modeled age as a continuous variable, stratified by microsatellite status, and adjusted for sidedness, stage, sex, and whether the patient had chemo or radiation before surgery. That design is why small, gradual shifts survived.
Ana C. Anderson of the Gene Lay Institute, a co-senior author, said the need to understand the rise in younger adults is urgent, and that it starts with biology across the whole lifespan.
By looking beyond conventional age cutoffs, we’ve uncovered biological patterns that earlier studies may have missed and created a foundation for the field to investigate them further.
Ana C. Anderson, PhD, Gene Lay Institute, Mass General Brigham
Broad Institute public notes on the paper stress differences in cell composition and expression rather than a new mutation unique to the young. In this cohort, age was not linked to sex, microsatellite status, or prior neoadjuvant treatment. Left-sided and stage IV tumors were.
THE ROAD TO THE ATLAS
- Mid-1990s: Early-onset colorectal cancer incidence in the United States begins rising at 1% to 3% per year even as overall rates fall.
- 2018: The American Cancer Society lowers the average-risk screening age from 50 to 45.
- 2021: The U.S. Preventive Services Task Force follows, extending a grade B screening recommendation to ages 45 to 49.
- 2023: Colorectal cancer becomes the leading cause of cancer death among Americans under 50.
- October 29, 2026: Porter, Ng, Anderson, and Ramnik J. Xavier publish the 189-tumor atlas in Nature Medicine.
Ng’s center feeds the work through its Beyond CRC Project, which lets patients donate tissue and clinical data. Single-nucleus sequencing was chosen because it works on frozen archives, so the team could use tumors already tied to long records instead of waiting for fresh samples.
Tumor Cells in the Young Express More FN1
Epithelial cells from younger patients showed higher activity of long noncoding RNAs and of FN1, a gene tied to cell movement and scar-like tissue. Left-sided microsatellite-stable tumors in the young also had more cell-to-cell contacts, with FN1, GDF, and ACTIVIN signals, pathways that help lining cells loosen and spread, running hotter in specific cell types. Copy-number changes looked more mixed in the young, though that trend was weaker.
Dana-Farber counted more than 200 genes whose activity differed across several cell types. Many sit on the path from a stuck lining cell to one that can move. Ng called the RNA finding the one that most needed follow-up, because those molecules had not been tied to early-onset colorectal cancer and many have never been described.
“Many of the long non-coding RNAs that showed gene expression differences between younger and older patients have not been previously characterized, so we don’t know what they are doing,” Ng said. She is working with colleague Marios Giannakis to find out. Some of the same RNAs showed up across different cancer-cell types. A large share are still “dark,” with no known job.
Ramnik Xavier, a co-senior author at the Broad Institute and Mass General Brigham, said the next tests should ask how these cell programs meet chemicals from diet and from gut microbes. That is the causal step the atlas does not take. Composition and signaling can sit downstream of food, microbes, or something else that tracks with birth year. The map names the neighborhood. It does not name the landlord.
The Map Does Not Yet Change Care
Mass General Brigham’s release on the paper is blunt about the limit: the findings do not prove these differences cause more aggressive disease. The method also recovered relatively few immune cells, so whether the extra scaffolding keeps T cells out is still open. In the 43 microsatellite-unstable tumors, no lineage-level cell type shifted with age the way stroma did in the stable majority.
WHAT WE KNOW
- Cell mix: Microsatellite-stable tumors from younger patients hold more stromal support cells and fewer epithelial cancer cells, confirmed in bulk gene data and in 712 stained tumors.
- Signals: FN1 and long noncoding RNAs run higher in young epithelial cells, and left-sided young tumors show more EMT-related contacts.
- Clinic pattern: Young patients in the atlas were more often left-sided and stage IV, matching the wider pattern of later diagnosis.
WHAT IS UNCONFIRMED
- Cause: The atlas does not show that extra stroma, dark RNAs, or EMT signals make the cancer spread or kill.
- Drugs: No one measured whether these 189 tumors actually resisted the extra chemo young patients already receive.
- Immunity: Too few immune cells were captured to say whether the scaffolding blocks T cells.
Ng said the group had been frustrated at not being able to pin down biological differences and now has a place to look: stromal cells and their contacts with immune cells. Functional work still has to show those contacts drive the disease. Until it does, average-risk screening still starts at 45, and standard chemo still goes in at young-adult doses. Xavier’s diet and microbe tests, and Giannakis’s RNA work, are the experiments that could move either line. Half of the under-50 cases will keep arriving in people those screening programs do not invite.
Frequently Asked Questions
What Is Early-Onset Colorectal Cancer?
It is colorectal cancer diagnosed before age 50. In this atlas, age was not linked to sex, microsatellite status, or whether the patient had treatment before surgery, even though young patients had more left-sided and stage IV tumors. Most cases are not inherited syndromes, though the samples were annotated for Lynch syndrome so those tumors could be flagged.
Does the Atlas Mean People in Their 30s Should Get Colonoscopies?
No. Average-risk screening in the United States still starts at 45, and this paper did not test a new screening age. People with Lynch syndrome, inflammatory bowel disease, or a close relative with colorectal cancer already start earlier under existing rules. The atlas’s value for screening is that support-cell burden declined gradually with age rather than dropping at 50, which is an argument for studying risk tools below 45, not a new invitation to the procedure.
What Is Microsatellite-Stable Colorectal Cancer?
It is the majority subtype, without the DNA-repair defect that defines microsatellite-unstable (MSI-high) tumors. The stromal excess was measured in 134 microsatellite-stable tumors. The 43 MSI-high tumors, only nine of them early-onset, did not show the same lineage-level age shift, so the scaffolding story should not be copied onto that smaller group.
Where Is the Atlas Data Deposited?
Processed count matrices and an AnnData object are in the Gene Expression Omnibus under accession GSE312983. Raw sequences sit in dbGaP under phs004493.v1.p1, a controlled-access repository. Clinical metadata for the 189 samples are in the paper’s Supplementary Table 1.
Should Young Patients Get Different Chemotherapy Because of This Study?
Not on this evidence. The authors cite stromal-rich tumors as harder to treat in several solid cancers, but they did not measure drug response in these patients. Ng has said the hope is that a clearer molecular picture will lead to new detection and treatment approaches. Until trials test stroma-directed or RNA-directed ideas, standard regimens stay the default.
Disclaimer: This article is news reporting and analysis of a published research paper and public cancer statistics. It is for information only and is not medical advice, a diagnosis, or a treatment recommendation for any person. It does not tell readers whether to start, skip, or change colorectal cancer screening, chemotherapy, or any other care. Anyone with bowel symptoms, a family history of colorectal cancer, or questions about screening age should talk with a qualified physician, such as a gastroenterologist or oncologist, before acting. Case counts, death counts, screening rates, and study findings reflect the American Cancer Society figures and the Nature Medicine paper cited here and may change as new data and trials appear.
-
NEWS3 months agoPennsylvania Measles Deaths Collide With Falling MMR Coverage
-
NEWS3 months agoFDA Approves XFG COVID Vaccines While the Advice Panel Is Frozen
-
NEWS3 months agoProgranulin and JAK2 Split the Macrophages That Heal
-
FITNESS3 months agoMuscle-Boosting Drugs Near Approval for Children With SMA
-
NEWS3 months agoAlnylam Defends Amvuttra After Wainua’s ATTR-CM Trial Miss
-
MENTAL HEALTH3 months agoAfter the ADHD Invoice, GPs and Waiters Carry the Cost
-
NEWS2 months agoRasonque’s List Price Shifts the Pancreatic Cancer Bill
-
NEWS3 months agoFDA Flagged This Sprout Plant 16 Months Before the Outbreak
