By Anna C. Greene, PhD, NETRF Chief Scientific Officer
Scientific progress rarely comes from a single experiment, or even a single research grant. It happens when discoveries build on one another, researchers bring different expertise to the same problem, and the tools created by one team make the next question possible.
A new study of lung neuroendocrine tumors (NETs) offers a remarkable example.
Published in Molecular Cancer, the study brings together genomic sequencing, gene expression, DNA methylation, spatial analysis, tumor organoids, pathology, and artificial intelligence to create one of the most detailed molecular maps of lung NETs to date. It provides strong evidence that these tumors fall into four biologically distinct molecular groups, including an especially aggressive group enriched for tumors known as “supra-carcinoids.”
The science is exciting. But from NETRF’s perspective, so is how it came together.
The paper acknowledges six NETRF research awards made between 2019 and 2025 that supported five investigators involved in the study: Lynnette Fernandez-Cuesta, PhD; Matthieu Foll, PhD; Talya Dayton, PhD; Jules Derks, MD, PhD; and Nicolas Alcala, PhD. The study was conceived by Drs. Alcala, Foll, and Fernandez-Cuesta, who are also its corresponding authors, with NETRF-supported Drs. Dayton and Derks contributing additional expertise.
This is what it looks like when we invest not only in individual projects, but in a research ecosystem.
Why lung NETs need a better map
Today, lung NETs are primarily classified as typical or atypical carcinoids based on features a pathologist can see under the microscope, including how frequently tumor cells are dividing and whether areas of dead tumor tissue, or necrosis, are present.
Those categories are clinically useful, but they do not capture the full biological diversity of these cancers. Tumors that look similar under the microscope can behave very differently. Conversely, tumors placed in different histologic categories can share important molecular characteristics.
Researchers have been working for years to understand what lies underneath those differences.
In the new study, the team analyzed more than 300 lung NETs using multiple layers of molecular information. Rather than finding two major groups that matched the traditional typical-versus-atypical distinction, the researchers identified four distinct molecular groups: Ca A1, Ca A2, Ca B, and a fourth group enriched for supra-carcinoids. Importantly, both typical and atypical tumors appeared within the molecular groups.
In other words, what a tumor looks like under the microscope is only part of its story.
What are supra-carcinoids?
Supra-carcinoids are especially intriguing because they can look like relatively well-differentiated lung NETs under the microscope while carrying molecular characteristics associated with much more aggressive neuroendocrine cancers.
In this study, the supra-carcinoid-enriched group had a more disrupted genome, was more likely to invade the pleura (the thin tissue lining surrounding the lung), had poorer overall survival, and showed a distinctive immune-rich tumor microenvironment. Interestingly, an immune-rich tumor is not necessarily one in which the immune system is successfully fighting the cancer. In supra-carcinoids, the researchers found evidence of close communication between tumor cells and macrophages, raising the possibility that the surrounding immune environment may help shape the tumor’s aggressive biology.
The researchers also found molecular similarities between these tumors and some forms of small cell lung cancer, suggesting that the biological boundaries between well-differentiated lung NETs and high-grade neuroendocrine carcinomas may be more complicated than previously understood.
The researchers also began to investigate how these different tumor types arise. Their analyses suggest that the major molecular identities of many lung NETs may be established relatively early in tumor development. Supra-carcinoids may be more complex, with evidence suggesting that at least some could evolve from other lung NET molecular groups.
Those findings raise important new questions about why some apparently lower-grade tumors become aggressive and whether researchers could eventually identify that risk earlier.
From molecular discovery to something pathologists can use
Defining molecular groups is scientifically interesting. For the work to ultimately matter to patients, however, researchers need practical ways to recognize those groups.
That is another important aspect of this study.
The investigators used artificial intelligence to examine routine pathology images and identify visual features associated with the molecular subtypes. Remarkably, the deep-learning models were better at predicting the underlying molecular groups than they were at predicting traditional histologic grade.
The team then combined these morphological features with an immunohistochemistry panel measuring three proteins, ASCL1, HNF1A, and OTP. Using the protein markers together with morphology, they correctly assigned 77 of 90 evaluated tumors to their molecular group. The researchers also developed a computational tool called PACMOS that can classify tumors when RNA or DNA methylation data are available.
That is significant because precision medicine requires more than discovering biological differences. We also need feasible ways to identify those differences in actual tumor samples.
Different tumors may have different vulnerabilities
The four groups also showed different patterns of genes and proteins that could someday help guide treatment research.
For example, the Ca A1 group showed the highest expression of DLL3, a potential treatment target. EGFR expression was highest in Ca A1 and supra-carcinoid-enriched tumors, while SSTR2 expression was high in both Ca A2 and Ca B. Two members of the fibroblast growth factor receptor family also showed distinct patterns: FGFR4 expression was highest in Ca A2, and FGFR3 expression was highest in Ca A2 and supra-carcinoid-enriched tumors. These findings suggest potential treatment targets that may be worth investigating in particular molecular groups. However, gene expression alone does not establish whether a tumor will respond to a treatment. Further studies are needed to determine whether these differences can help guide therapy.
That distinction is important. This is not yet a new treatment algorithm. It is a much better map from which researchers can begin asking more precise questions.
The NETRF research ecosystem behind the discovery
For NETRF, one of the most exciting aspects of this paper is seeing years of intentionally connected research converge.
Our investment in this area extends back to a 2017 Petersen Accelerator Award to Hans Clevers, MD, PhD, to develop desperately needed NET laboratory models, including patient-derived organoids. That program helped establish tumor models that researchers could use to study NET biology and therapeutic vulnerabilities. Talya Dayton, PhD, then working in the Clevers laboratory, helped advance those models and later received her own NETRF Investigator Award to study lung NET progression using patient-derived organoids. A supra-carcinoid organoid contributed to the biology explored in this new study.
At the same time, NETRF supported several complementary lines of investigation:
- In 2019, Dr. Fernandez-Cuesta received an Investigator Award to deeply characterize supra-carcinoids and understand why this unusual subset of lung NETs behaves so aggressively.
- In 2022, Dr. Foll received an Investigator Award to reconcile molecular and pathology-based classifications of lung NETs and move those discoveries closer to clinical use.
- Also in 2022, Dr. Derks received a Mentored Research Award to determine whether molecular lung NET subtypes could be identified using practical biomarkers in tumor biopsies, work that helped establish the ASCL1, HNF1A, and OTP marker panel.
- In 2023, Dr. Alcala received a Mentored Research Award to reconstruct how different NET molecular subtypes evolve and how tumor genetics and the immune environment influence their clinical behavior.
- And in 2025, NETRF made its largest award, the $800,000 Petersen Accelerator Award, to Dr. Fernandez-Cuesta to take the next step: understanding the cell states that drive supra-carcinoids and developing better approaches to identify and ultimately intercept aggressive disease.
This ecosystem also helps build the future of the field. NETRF support has helped researchers develop deep expertise in neuroendocrine cancer and, in several cases, progress into independent research leadership. That pipeline is particularly important in a rare cancer field, where sustained progress depends on attracting talented scientists and giving them opportunities to build careers focused on these diseases.
These were not five researchers independently circling the same problem. Their work became interconnected: models informed molecular studies; molecular classifications informed pathology; pathology provided tools for translation; and evolutionary studies asked why the different tumor types arise in the first place.
The new paper brings many of those threads together.
That is exactly the kind of scientific ecosystem NETRF seeks to build.
What needs to happen next?
Important work remains before this molecular classification can influence routine patient care.
The researchers note that diagnostic methods will need additional refinement and validation across institutions. The relationship between molecular subtype and treatment response must be tested prospectively, and the biology of supra-carcinoids, including whether and how some tumors transition toward a more aggressive state, requires further study.
Those are precisely the kinds of questions the next generation of this research is positioned to answer.
What does this mean for patients today?
This study does not change how clinicians diagnose or treat an individual patient with a lung NET today, and molecular classifications such as Ca A1, Ca A2, Ca B, and supra-carcinoid-enriched are not yet part of routine clinical care.
What it does provide is something fundamental: a clearer picture of why lung NETs that appear similar can behave so differently.
Ultimately, better classification could help researchers identify patients at greater risk of aggressive disease, design smarter clinical trials, and match therapies to the biological vulnerabilities of an individual tumor.
For a rare cancer, building that foundation takes time. It also takes continuity.
This paper is a powerful reminder that research funding is not simply a series of individual grants. The greatest return can come when one investment creates a tool, another reveals a biological insight, another translates that insight toward the clinic, and a community of researchers has the resources and relationships to keep building.
That is how a field moves forward.
Dr. Fernandez-Cuesta will continue this conversation at the 2026 Margie & Robert E. Petersen Neuroendocrine Cancer Research Symposium in November, where she will deliver the Challenge Address on what it will take to move neuroendocrine cancer research forward.
Read the study in Molecular Cancer
_____________________________________________________________________________________________________________________
Learn more about NETRF-funded research. Read the The Year in Research 2025.