This transcript was automatically generated from the audio recording and may contain inaccuracies, including errors or typographical mistakes.

Dr. Strosberg: I am Jonathan Strosberg. I am a Professor of Gastrointestinal Oncology at the Moffitt Cancer Center in Florida, and I am joined today by Dr. Daniel Morgenstern. We will be talking about poorly differentiated extrapulmonary neuroendocrine carcinomas in general, and specifically about the potential role of DLL3-targeting drugs.

Just as a way of introduction, when we think of neuroendocrine neoplasms, we are talking about a very heterogeneous category of tumors ranging from extremely slow-growing to extremely aggressive. On the one hand, we have well-differentiated neuroendocrine tumors that tend to be relatively slow-growing. We distinguish between three different grades, grade 1 with the Ki-67 less than 3%, grade 2 with Ki-67 of 3 to 20%, grade 3 with Ki-67 greater than 20%. Now, those well-differentiated grade 3 tend to be fairly aggressive, but then on the other end of the spectrum is poorly differentiated neuroendocrine carcinomas. You will notice we use the word tumor to describe well-differentiated even if it is metastatic, carcinoma to describe poorly differentiated. Those are extremely aggressive cancers. We can distinguish between small-cell histology, which is similar to small-cell lung cancer, and large-cell histology, although hard to say whether that is really clinically meaningful or prognostic.

Poorly differentiated neuroendocrine carcinomas are always high grade. The Ki-67 is always above 20%, and in most cases it is much higher than that 70, 80, 90%. By poorly differentiated referring to the morphology, which, unlike well-differentiated NETs, does not resemble the normal endocrine cells of origin. With well-differentiated NETs, the cells tend to be arranged in islets or trabecular, they are well organized with typical salt and pepper chromatin. With poorly differentiated neuroendocrine carcinomas, they are often pleomorphic, a high nucleotide-to-hyperplastic ratio, and often large areas of necrosis.

Now, when we talk about neuroendocrine carcinomas in general, the large majority are lung neuroendocrine carcinomas and specifically small-cell lung cancer, in fact, more than 90% are small-cell lung cancer. Extrapulmonary is roughly 10% of the pie. It is very heterogeneous in terms of site of origin. Maybe about a third of those extrapulmonary originate in the gastrointestinal tract, another third originate either in the urinary tract or the female genital tract or the male genital tract. Then maybe another third of those are unknown primary.

Within the gastrointestinal tract, the most common primary sites are colorectal and pancreas, esophagus is also seen occasionally, gallbladder is another potential primary site, stomach. Within the female general urinary tract, it is mostly cervix, but occasionally ovary, and then for males, prostate NEC is something that we see relatively frequently, but it typically evolves from adenocarcinoma as it differentiates and sometimes bladder. These are some of the sites of extrapulmonary neuroendocrine cancers.

Usually pathologists can fairly easily distinguish a well-differentiated neuroendocrine tumor from a poorly differentiate neuroendocrine carcinoma. But it could be ambiguous in some cases. In those cases, looking for different mutational patterns on next-generation sequencing can be helpful. Just like small-cell lung cancer, we commonly see mutations in p53 or RB1. With NECs, sometimes SMAD4. Sometimes we see mutations that are common to adenocarcinomas from this corresponding organ, so things like KRAS or BRAF for colon or pancreas because in many cases, it is thought that these cancers arise from the same cell of origin and simply acquire a neuroendocrine differentiation as they develop. On the other hand, neuroendocrine tumors tend to have characteristic mutations like MEN1, DAXX, ATRX, those kinds of mutations indicate it is probably well-differentiated, especially pancreatic NETs.

Sometimes PET imaging can help. Somatostatin receptor imaging is commonly done for neuroendocrine tumors, and those commonly lights up avidly for well-differentiated neuroendocrine tumors, much less commonly for poorly differentiated neuroendocrine cancers. FDG PET was probably not that helpful because you would expect it to be avid in or so hyper metabolic uptake in well-differentiated G3 tumors. FDG PET will almost always be positive for poorly differentiated neuroendocrine carcinoma, but also sometimes positive for an aggressive well-differentiated NET.

There is also a mixed histology cancer. It is called MINEN, mixed neuroendocrine, non‑neuroendocrine carcinoma. Used to be called MANEC, mixed adenoneuroendocrine carcinoma. The strict definition is that at least 30% of either the neuroendocrine or non-neuroendocrine carcinoma can be identified by the pathologist. Although, whether it really needs to be 30% to meet this diagnostic criteria is not completely clear. It again, reflects the fact that many neuroendocrine carcinomas originate from similar precancerous cells to adenocarcinomas, and there can be two branches of differentiation.

As far as prognosis, just like small-cell lung cancer, extrapulmonary neuroendocrine cancers are highly aggressive cancers. They are usually identified in the metastatic stage. In many cases, the performance status is too poor at even a diagnosis to initiate treatment, and patients who do not start treatment generally have a life expectancy of less than two months. The median survival for all-comers is about 7-8 months, but obviously it varies by stage. For local stage, according to the Seer database, it is about 20 months for local, about 14 months for regional, and about six months for patients with distant metastatic disease.

The treatment for extrapulmonary neuroendocrine carcinomas has been borrowed from small-cell lung cancer traditionally because they are histologically similar, although there are differences. For example, we all know that small-cell lung cancer is almost 100% associated with smoking. That is not the case for extrapulmonary neuroendocrine cancer. We are not talking about exactly the same malignancy. But in any case, the standard first-line therapy has been platinum etoposide for several decades now. The outcomes tend to be quite similar with cisplatin or carboplatin. There is data that switching etoposide with irinotecan results in more or less the same outcomes.

The response rates are typically 30-50%, maybe a little bit lower than small-cell lung cancer. The optimal number of cycles is unclear. For small-cell lung cancer, it tends to be four cycles. We are not sure whether that is true for extrapulmonary, and in some cases, will continue beyond four cycles, although most patients cannot tolerate much more than six cycles.

The benefit of adding a checkpoint inhibitor like atezolizumab has been demonstrated only in small-cell lung cancer, where we are conducting a randomized trial in extrapulmonary neuroendocrine cancers to see whether there is similar survival benefit in this population. It has not been determined yet. There have been single arm studies of adding checkpoint inhibitors to platinum etoposide. But it is hard to know whether it truly benefits patients based on single-arm studies.

As I said, the platinum etoposide combination was borrowed from small-cell lung cancer. But of course, we use our own platinum-based regimens in gastrointestinal cancers, it just tends to be oxaliplatin more often than cisplatin. There is some retrospective data suggesting that FOLFIRINOX, a combination of 5-FU, oxaliplatin and irinotecan might have encouraging outcomes in gastrointestinal NECs. We are awaiting data from a randomized Phase II study that is comparing FOLFIRINOX to platinum etoposide.

Beyond first-line, there is really not much that has been promising. If there has been more than a 3-6-month interval of benefit after first-line platinum etoposide, we tend to resume the same chemotherapy. But other options include FOLFOX, FOLFIRI, sometimes we use capecitabine temozolomide. There is data suggesting that the outcomes are fairly poor with all these second-line regimens with median survivals of approximately six months.

There was a small prospective study looking at nal-IRI and 5-FU that showed an overall survival of six months. Another study, a negative study, that showed that adding bevacizumab to FOLFIRI did not result in any overall survival benefit. There are also anecdotal-based reports that in patients with targetable mutations like BRAF, which are fairly uncommon. Targeted treatments might be somewhat beneficial.

As far as checkpoint inhibitor therapy in the second-line, there have been studies suggesting that single-agent checkpoint inhibitors are relatively ineffective. There has been mixed data with combinations such as ipilimumab-nivolumab, one very small prospective study on EP NEC showed a response rate of 26%. Other similar studies have shown lower response rates. In our experience it is really a very small percentage, more or less 10% that experienced long-term benefit from combination immunotherapy. Guidelines like NCCN indicate that these are an option. But it is mostly because the chemotherapy is so ineffective in this setting.

For early-stage tumors, which are relatively uncommon, data is even more limited. The guidelines generally recommend multimodal therapy. Surgery alone is clearly inadequate. We generally recommend at least four cycles of either neoadjuvant or adjuvant platinum etoposide and in some cases the local regional treatment should be radiation rather than surgery. That seems to be the case for esophageal neuroendocrine carcinomas or other locally advanced primary sites. In some cases we do trimodality treatment: surgery, radiation and adjuvant therapy. But this is all based on anecdotal data and theory rather than prospective studies.

To conclude, thus far, the progress in management of extrapulmonary neuroendocrine cancers has been extremely limited, and the standard of care for stage IV disease remains platinum etoposide and the role of checkpoint inhibitors in unselected patients is debatable. There is an urgent need for new therapies, and one promising new direction, our DLL3 drugs targeting using bispecific engagers. Dr. Morgenstern is going to talk more about that. Thanks.

Dr. Morgenstern: Hi. I am Dr. Daniel Morgenstern, Director of Thoracic Oncology at Washington University School of Medicine. My task here is to talk about DLL3, which has provided some hope for patients with small-cell lung cancer, also for extrapulmonary neuroendocrine tumors.

DLL3 is a delta-like ligand 3, is a member of a Notch signaling pathway which regulates the cell fate decisions through four receptors, Notch1-4, and five canonic ligands. JAG1, JAG2, DLL1, DLL3, DLL4. In this pathway, Notch suppresses ASCL1, which is a key transcriptional driver of neuroendocrine differentiation. In this way, Notch activation is associated with suppressed neuroendocrine differentiation. Unlike other ligands which are located in the cell surface, DLL3 is normally confined to intracellular compartments, mostly golgi apparatus.

Upregulation of ASCL1 neuroendocrine tumors leads to DLL3 overexpression and aberrant expression in the cell surface, while inhibits Notch-signaling in cis. Cis means the same cell, unlike the other Notch ligands, which are in turn from another cell leading to ASCL1-high phenotype.

In summary, ASCL1 is the driver of neuroendocrine differentiation. It is inhibited by Notch, which in turn is inhibited by DLL3. DLL3 is expressed in the majority of patients' small-cell lung cancer and other neuroendocrine tumors. In a study recently published, DLL3 was positive in more than 90% of patients' small-cell lung cancer, approximately 80% of other tumors. DLL3 is rarely expressed in normal cells, making it an attractive target for these neuroendocrine tumors.

There are several ways to target DLL3: through antibody conjugates, T-cell engagers, CAR T-cells, and monoconjugates. A brief of ADCs, antibody-drug conjugates, there are many promising ADCs currently testing small-cell lung cancer, with some emerging data in extrapulmonary neuroendocrine carcinomas, which are called EP-NECs.

The main targets for a disease in endocrine tumors are DLL3, Trop2, B7-H3, and CR6. DLL3 was one of the first targets for disease, with two drug tested. Rovalpituzumab tesirine (Rova-T) and SC-002. But despite initial promising results, both drugs were discontinued development because of a lack of improving outcomes compared to topotecan/Rova-T and marked toxicity from both, which has been attributed to the PBD payload.

Despite these initial failures, those studies show that the DLL3 is a valuable target and there is a new generation of DLL3 disease of promising results. Of those, ZL-1310 has the most data available. We have a promising response rate above 50% in small-cell lung cancer. Unfortunately for EP-NECs, there is no data for DLL3 disease. The only ADC with some data is ABBV-706, which is an antibody-drug conjugate targeting SEZ6, and we are using a topoisomerase-1 payload out of nine patients with GEPs, gastro enteropancreatic tumors, two responded, leading to a 22% response rate.

In contrast, T-cell engagers have a completely different mechanism of action. The T-cell engagers bind to CD3 complex on T-cells and target antigen on tumors. This low binding brings the tumors in close proximity of the T-cells of the formation of immunological synapse, leading to MCH1 independent T-cell activation. The T-cell engagers in the most advanced stage of development tarlatamab and obrixtamig. Both are bispecific T-cell engagers targeting CD3 on T-cells and DLL3 on tumor cells.

The first one is tarlatamab, which is composed of two single-chain variable fragments connected by a linker and a non-functional FC just to increase the serum half-life. The initial studies have been done exclusively in small-cell lung cancer, and the drug is already approved for this indication. Perhaps the most important study was at DeLLphi-304, where patients were randomized to second-line, tarlatamab or standard-of-care, which could include lurbinectedin, topotecan or amrubicin. Tarlatamab was associated with improvement in OS, median PFS and objective response. There are several ongoing studies with tarlatamab, and a special one is called RAGNAR, which is evaluating tarlatamab alone or in combination FOLFIRI in patients with DLL3-positive GEP-NECs previously treated with a platinum-based regime, and I assume those will be patients that receive platinum etoposide.

The second bispecific antibody is obrixtamig, which was until recently known as BI764532, which, unlike tarlatamab, has an IgG-like scaffold, so the whole molecule looks like an IgG. The Phase I study involved patients with small-cell lung cancer, large-cell lung cancer, which I will call LCNEC from now on in EP-NECs. This study had four different regimens, and the overall response rate for patients treated in two of those regimens was 21%. More interestingly presented to ASCO this year, among the 60 patients of extrapulmonary NECs, the objective response rate was 40% to those with DLL3-high, which was defined more than 50% of cells positive and only 3.3% for those of DLL3-low.

Obrixtamig received fast-track designation for DLL3 extrapulmonary neuroendocrine cancers. And the diagram five is a study looking at obrixtamig in patients with small-cell lung cancer or NECs. They should have received at least two prior lines and they will compare two doses followed by an expansion cohort and most promising dose.

There is also the diagram seven, which is a Phase I study evaluating carboplatin etoposide + obrixtamig as a first-line therapy for patients with DLL3-positive EP-NECs, LCNECs, or NECs of unknown primary site. This is an important study, which may actually define the role of obrixtamig or bispecific T-cell engagers in the first-line setting.

The other interesting DLL3 T-cell engagers will have other BiTEs. We have three specific T-cell=activating constructs, which we call them TriTACs. The one in the most advanced stage of development is called MK-6070, which was formerly known as HPN328, and in the Phase I study, there were 13 patients with neuroendocrine neoplasms, of which six achieved a response, which is 46%. Although the results are promising, certainly better than results from a standard of care in those patients, there are some potential complications such as cytokine release syndrome, which usually is low-grade and treated with acetaminophen, dexamethasone or fluids, but some cases may be more severe, requiring tocilizumab.

Another fear of complication from bispecific T-cell engagers is the immune effector cell-associated neurotoxicity ICANS syndrome or ICANS. This toxicity usually occurs slightly after the CRS, and it may present with confusion, impaired attention, more to weakness, dysgraphia. Patients should be treated with dexamethasone or methylprednisolone.

Another way to target DLL3s with chimeric antigen receptor CAR cells. These are synthetic receptors that allow T-cells to recognize the tumor-associated antigens independent of MHC. There are two drugs, two CAR Ts.

AMG 199 was the first one to be tested. Five patients were evaluated, there was one partial response to stable disease. The treatment was relatively well-tolerated, but the study is suspended.

The second one, which is still ongoing, is LB2102, another DLL3 CAR T which is evaluated in small-cell lung cancer in LCNEC. The Phase I study was initially presented in 2025. Nine patients treated, eight with small-cell one with LCNEC. The treatment was well-tolerated and there was one partial response.

Another interesting way of CAR therapy is there are three CAR NK cells. Those studies use NK92, which is NK cell line, natural killer cell, established in 1992 from a patient with Hodgkin's lymphoma. These cells have increased cytotoxicity compared to primary activated NK cells. They can be engineered as CAR and not for donor-independent or off-the-shelf options. The manufacturing will be ready. Patients will not have to wait that long. It is going to be tested only in small-cell lung cancer.

The fourth way to target DLL3 is through a variety of monotherapy. One of the drugs is Lu177-DTPA-C16. We just break down this drug, C16 is the antibody. C stands for Stemcentrx. That is the original antibody that was tested in the Robo-T SC002 studies. The DTPAs decay later to bind to the activated isotope. In this case, lutetium-177. This is the same radioactive compound that was used for lutetium dotatate and in lutetium vipivotide tetraxetan. There is no clinical data available at this time.

A second radioactive isotope is the AC227 dotatate octreotide. Here comes the name 'tate' is a somatostatin analogue that will bind to somatostatin receptor two, SSR2. Dota is the chelator and actinium AC225 is an alpha-emitting radioactive isotope. This is currently being tested in small-cell lung cancer and other neuroendocrine tumors.

In summary, DLL3 is a very important target for neuroendocrine tumors. There are many ways to target. We are not so sure how those drugs could be used in sequence. At this point, most of the studies will not allow prior use of another mechanism of targeting DLL3, so we have to choose which one would be the most promising. Responses in second-line and beyond seemed to be better across trials comparing with the existing treatments for LCNECs, and I have little question that at least one of the DLL3-targeting drugs will soon become one of the standard therapies in this setting.

Dr. Morgenstern: How do you see ADCs or bispecific antibodies? Perhaps, think a place maybe in previously treated patients, maybe it is a little bit too early to claim the first-line.

Dr. Strosberg: Yes. The natural progression is to start in second-line and beyond, and then try adding it to standard therapy in the first-line. I think that is what is going to happen with these bispecifics. As far as ADCs, as you pointed out, there has not been any significant success so far with ADCs targeting DLL3, but the bispecific and trispecific T-cell engagers look quite promising. There is so many cases where patients are really not in good enough shape to receive any second-line therapy, and so if we can add these drugs to first-line in addition to chemotherapy or in some cases, maybe instead of chemotherapy, it may be more beneficial to a larger population of patients.

Dr. Morgenstern: I also think that it is very important that we have at least four reasonable targets: DLL3, SEZ6, Trop2, B7-H3. If someone receives DLL3 ADC with a Trop1 payload, that is over. But if they receive a DLL3 bispecific for example, there is nothing to prevent the patient from receiving sequential ADC of Trop1 payload and one of the other three targets. I can see at least possibly two lines of therapy that will be extrapolated from small-cell to EP-NECs. The number of patients are small, so it is hard to have a 200-patient randomized study for EP-NECs, but having 20 or 30 of some reasonable response is probably going to be enough. The obrixtamig for example, had patients with EP-NECs, and the fasttrack includes neuroendocrine tumors that are DLL3-positive.

Dr. Strosberg: Yes. We definitely need to identify more targets. We are actually planning to do a study on identification of targets in EP-NETs and NECs. That would be interesting to see if we can duplicate some of the things that are being done in small-cell.

Dr. Morgenstern: To me some of those targets, they should be expressed overexpressed in neuroendocrine tumors. For example, DLL3 is a target of ASCL1, which is the master of neuroendocrine differentiation. The same is true for SEZ6. They have excellent correlation. I would suspect there would be no difference.

Dr. Strosberg: Well, DLL3 definitely is expressed highly in poorly differentiated neuroendocrine carcinomas. The literature suggests it is not well expressed in well-differentiated neuroendocrine tumors.

Dr. Morgenstern: True. Yes. NECs not NETs. Yes. Because NETs, you can do other things. Everolimus + octreotide, this is the next step where you borrow the small-cell treatments for a specific reason, because, sure, the smoking and demographics are not the same, and the percentage of patients of early-stage, NECs is a little bit higher than small-cell. But in terms of differentiation and in genomic expression they are quite similar.

Dr. Strosberg: I think extrapulmonary NECs in many ways are a merger between small-cell in the primary site. For example, I just saw gallbladder NEC recently, and it is highly expressing HER2, which gallbladder cancers do, or cervical NECs can be associated with HPV. They are a hybrid between the corresponding adenocarcinomas or squamous cell carcinomas and small-cell lung cancer or large-cell lung cancer.

Dr. Morgenstern: Or perhaps early differentiation before the symptoms appear and before treatment is given. I also noticed that even the pattern of metastasis, very little brain metastases, and they behave just like the carcinoma counterpart, the gallbladder NECs go to the liver. They do not go to the brain like the small-cell lung cancers. It is fascinating.

Dr. Strosberg: Yes. Very interesting.