SCLC has historically been considered a one-size-fits-all diagnosis associated with a poor prognosis.¹ Until recently, when chemoimmunotherapy became the frontline standard of care (SoC),² platinum plus etoposide was the standard treatment for patients with SCLC for many decades, with more than 40 phase III trials failing to improve upon platinum‑based therapy during that time.^1,3

In 2019, the FDA approved combination therapy with the PD-L1 checkpoint inhibitor atezolizumab and carboplatin/etoposide, and shortly thereafter in 2020, the PD-L1 checkpoint inhibitor durvalumab and platinum/etoposide, for first-line treatment of patients with ES-SCLC.^4,5 This advance transformed outcomes for a subset of patients with ES-SCLC, approximately 15% to 20%,^6-9 but there remains a clear need for new treatment strategies across the disease continuum. For a review of the most recent clinical data on SoC therapies in the treatment of advanced ES-SCLC, please visit the companion module here.

As this typical H&E micrograph of an SCLC tumor biopsy shows, these tumor cells are almost all nucleus, with very little cytoplasm. Further, they are extremely mitotically active; it is not unusual to see a Ki67 >90%.¹⁰  SCLC tumor cells are also very transcriptionally active. Historically, SCLC diagnoses were confirmed by expression of the lung-specific marker TTF1 along with the neuroendocrine markers synaptophysin, chromogranin, and CD56 (ie, NCAM1). However, we are increasingly appreciating that there is more nuance and heterogeneity to the biology of SCLC, which we will discuss further.

One of the challenges impeding personalization of therapy for advanced SCLC, like we have for advanced non-small-cell lung cancer (NSCLC),¹¹ has been the failure to identify genomic drivers amenable to conventional targeted therapy. This is not due to a lack of effort by the research community, but rather is a product of the biology of SCLC, which does not possess any of the genomic oncogenic driver mutations we see in NSCLC (eg, EGFR mutations, ALK or ROS1 rearrangements, etc). Instead, SCLC is typically associated with the loss of tumor suppressor genes. Specifically, there is almost universal dual loss of TP53 and RB1 across SCLC.¹²

This diagram on the right represents a groundbreaking effort from George and colleagues¹³ to perform genomic analysis on upwards of 80 resected SCLC tumors. In addition to the universal loss of TP53 and RB1, they observed frequent MYC amplifications and alterations in epigenetic modifiers, but again a complete lack of our standard oncogenic driver mutations. Although this work advanced our understanding of SCLC biology, unfortunately, it did not move us toward personalized treatment strategies for patients with SCLC.

Although the genomic landscape in SCLC is relatively homogeneous—at least as it pertains to TP53 and RB1 and an apparent lack of oncogenic driver mutations—considerable heterogeneity at the transcriptional level has been identified. Efforts by our group¹⁴ and others¹² have defined 4 primary transcriptional subtypes of SCLC that are defined by the presence or absence of 3 transcription factors: ASCL1, NEUROD1, and POU2F3.¹⁴ The subtypes SCLC-A, SCLC-N, and SCLC-P correspond to increased transcription of ASCL1, NEUROD1, and POU2F3, respectively. There is a fourth SCLC subtype, SCLC-I (ie, SCLC inflamed), that is associated with lower expression (not necessarily no expression) of these 3 transcription factors, but has a much higher expression of immune genes in general, a sort of inflammatory gene expression profile. The diagram on the left illustrates the continuum across these transcriptional subtypes of SCLC, going from more neuroendocrine-like to less neuroendocrine-like, and from less inflamed to more inflamed.

These 4 SCLC subtypes bear out in retrospective analyses of patient outcomes with SoC chemoimmunotherapy regimens in the frontline setting.¹⁴ As predicted, the SCLC-I subset does better with chemoimmunotherapy, with a higher likelihood of response and long-term survival.

Of note, in contrast to genomic driver mutations, transcriptional subtypes are not necessarily fixed; they can evolve across the course of treatment and may be a potential source of therapeutic resistance. As we attempt to integrate these subtypes into clinical trials as predictive biomarkers, we are cognizant of changes in transcription profiles as a potential mechanism of resistance.

Application of a biomarker test that relies on tissue biopsy acquisition and a complex downstream molecular analysis, such as used in the transcriptional profiling studies described above, is not necessarily compatible with assigning a patient who has SCLC to an appropriate treatment regimen on a trial or in the clinic. To navigate this challenge, we need to think about more rapid but equally reliable ways to assign biomarkers to these patients.

To address this challenge, Heeke and colleagues¹⁵ recently demonstrated that the variable subtypes of SCLC have unique DNA methylation profiles that not only can be detected in the tumor, but most importantly, can be detected in circulating tumor DNA (ctDNA) in plasma. The latter is achieved by testing for DNA methylation of ctDNA in a “liquid biopsy”, which is a simple blood sample. In addition to being able to assign patients to a subtype this way, liquid biopsy has the potential for longitudinal monitoring of response to therapy and potential subtype evolution.

I am also hopeful that in the near future we will be able to integrate blood-based detection of biomarkers for the targeted therapies under development I will talk about next.

Among emerging targets for biomarker-driven therapies under development in the treatment of SCLC, I will focus on DLL3,¹⁶ B7-H3,¹⁷ and TROP-2.¹⁸

These are representative of a number of cell surface proteins (which are either associated specifically with SCLC and other neuroendocrine tumors, or across tumors but which also happen to be found in SCLC) that theoretically could be leveraged as beacons to deliver cytotoxic therapies to tumor cells via strategies like ADCs, or as ways to engage the immune system via strategies such as CAR T-cell therapy or BiTE therapy.

We will now look at emerging clinical data for these very strategies.

Of these cell surface-targeting strategies, the agent that is the furthest along in development is tarlatamab. Tarlatamab is a DLL3-targeting BiTE, that is, a dual antibody with one component that targets DLL3, which is expressed on the tumor cell surface, and a complementary component that binds to CD3, which is expressed on the surface of T-cells.^16,19 Tarlatamab forces a physical interaction between the tumor cell and T-cell that is otherwise frequently not happening in the immune deserts that we associate with SCLC. Aside from the 15% to 20% of patients with SCLC who have the inflamed subtype and experience long-term survival with chemoimmunotherapy,^6-9 SCLC tumor microenvironments are devoid of T-cells. Tarlatamab provides a physical way to overcome this limitation and introduce the immune system within proximity of the tumor.^16,19

The most compelling demonstration of tarlatamab efficacy in ES-SCLC thus far was the open-label phase II DeLLphi301 study.²⁰ In DeLLphi301, patients with ES-SCLC who had received at least 2 prior lines of therapy were assigned to tarlatamab 10 mg or 100 mg, with the 10 mg dose ultimately being selected for dose expansion. Of note, step-up dosing of tarlatamab was done in cycle 1, with 1 mg given on Day 1 then each full dose given on Day 8 and Day 15; full dose tarlatamab was given every 2 weeks thereafter. DLL3 expression positivity was not required for enrollment on the trial.

The primary endpoint of the study was objective response rate (ORR), with secondary endpoints including duration of response (DoR), disease control rate (DCR), progression-free survival (PFS), and overall survival (OS).

The cohort of patients receiving tarlatamab 10 mg achieved an ORR of 40%, which in the context of third-line ES-SCLC is impressive.²⁰ The ORR was a little lower at 32% in the cohort of patients receiving tarlatamab 100 mg. As the waterfall plot shows, quite a few of those responses are deep, near-complete responses.

Furthermore, responses were observed in patients who had confirmed DLL3 expression and those who either did not have DLL3 expression or where DLL3 expression was undefined because of a dearth of adequate tumor tissue. Thus, it is unclear whether there is a distinct threshold for DLL3 expression that is necessary for efficacy of this agent. 

Although a 40% ORR with tarlatamab 10 mg in the third-line setting of ES-SCLC is impressive in its own right, what is really impressive is the duration of some of these responses.²⁰ In the figure, you can see that several patients have ongoing responses for a year or more, which is almost unheard of in the heavily pretreated setting for relapsed SCLC.

To me, this is indicative of an immune response to the tumor, which although was the design intent with BiTEs, was not a foregone conclusion when these drugs were first introduced.

Here we see the survival data for both the tarlatamab 10 mg and tarlatamab 100 mg cohorts.²⁰ Of PFS and OS, the OS data is the most compelling to me.

The median OS with tarlatamab 10 mg is 14.3 months—in the third-line setting and beyond, in heavily-pretreated patients—which is longer than the median OS observed in the phase III trials leading to FDA approval of first-line chemoimmunotherapy in ES-SCLC. In IMpower133, median OS with first-line atezolizumab plus carboplatin/etoposide was 12.3 months.⁶ In CASPIAN, median OS with first-line durvalumab plus platinum/etoposide was 13.0 months.⁹

These data set an impressive benchmark in the treatment of ES-SCLC,²⁰ and are enthusiastically being followed up in randomized phase III trials, including moving this agent into earlier lines of therapy for ES-SCLC, and even earlier into the treatment of limited-stage SCLC (LS-SCLC).

• DeLLphi-304: second-line tarlatamab vs SoC chemotherapy after first-line platinum-based regimen in patients with ES-SCLC (NCT05740566)
• DeLLphi-305: first-line tarlatamab vs durvalumab after durvalumab plus platinum/etoposide in treatment of ES-SCLC (NCT06211036)
• DeLLphi-306: tarlatamab vs placebo after concurrent chemoradiation therapy in LS-SCLC (NCT06117774)

T-cell–mediated therapy, including BiTEs, like tarlatamab, and CAR T-cell therapy, is associated with a risk of CRS. CRS is an exaggerated inflammatory response characterized by a rapid elevation of cytokine levels that can be associated with mild flu-like symptoms or severe, life-threatening symptoms, including respiratory distress or shock leading to organ failure, as well as neurologic side effects like immune effector cell-associated neurotoxicity syndrome (ICANS).^21,22 Most of what we know about these T-cell–related adverse events (AEs) are from clinical trials and clinical experience with CAR T-cell therapy, which has a slightly different mechanism of action than BiTEs. It is emerging in other diseases, particularly hematologic malignancies where these drugs are approved and in clinical use, that these AEs may be less severe with BiTEs than with CAR T-cell therapy.

This figure depicts CRS and ICANS during treatment with tarlatamab 10 mg in DeLLphi-301.²⁰ ICANS was not much of an issue, with only a few documented cases, none of which were severe.

CRS was observed primarily during step-up dosing in cycle 1, most predominantly with the 1 mg dose on Day 1 and the first full 10 mg dose on Day 8. These were almost exclusively mild events, with very few incidences of grade 3/4 CRS. However, after the first 2 weeks of step-up dosing, CRS became virtually nonexistent in patients with ES-SCLC receiving tarlatamab 10 mg.

Most cases of CRS that did occur were managed with some combination of acetaminophen, intravenous hydration, and glucocorticoids. Tocilizumab, an immunosuppressive biologic,²³ was only needed to treat CRS in 5% of the tarlatamab 10 mg cohort; nearly all cases (98%) resolved.²⁰

Although we will need to be prepared to address CRS in patients who are receiving tarlatamab, because of its mild nature, I am hopeful we will be able to primarily manage it with steroids, anti-inflammatories, and fluids, and outside of the ICU setting, with an eventual goal of outpatient administration. Patient education will be important, including telling them about the typical timing for CRS onset with tarlatamab, during which they would need to watch closely for characteristic symptoms. Some trials of tarlatamab plan to test a strategy where there is a 1 hour outpatient observation period after which patients are required to be within a certain distance of a hospital for a defined period of time.

BI 764532 is another DLL3-targeted BiTE under investigation. This drug has a comparable design to tarlatamab, where you have both antibodies that bind to DLL3 and CD3 in the same molecule attempting to deliver the immune system to the tumor microenvironment.²⁴

This slide describes the first-in-human phase I dose escalation trial of BI 764532, with a primary goal of describing the drug’s safety and maximum tolerated dose as well as its pharmacokinetics (PK) and antitumor activity.²⁴ BI 764532 was initially given every 3 weeks, then based on emerging PK data, was switched to weekly with inclusion of step-up dosing during the first 3 cycles to reduce risk of CRS, which is common with BiTE therapy across disease types.^21,22

Of note, and in contrast to the DeLLphi-301 trial, enrollment on this study required demonstration of DLL3 expression on tissue, although the tissue could be archival.²⁴ Furthermore, the population in this study was a bit more heterogeneous. In addition to patients with advanced SCLC, this study also included patients with extrapulmonary high-grade neuroendocrine carcinoma (epNEC) and with pulmonary large cell neuroendocrine carcinoma (LCNEC). Patients were required to be ineligible for or to have relapsed on standard therapy, which for patients with SCLC needed to include 1 or more prior lines of platinum-based chemotherapy. 

Here we see the overall response data, including patients with advanced SCLC, epNEC, and LCNEC, for BI 764532.²⁴ In the waterfall plot, there appears to be differential outcomes at a dose cutoff of 90 µg/kg, with a trend toward responses in the patients who received higher doses.

Across all dose levels, partial responses were seen in 18% of all comers, 19% of patients with SCLC, and 38% of patients with LCNEC, the latter of which was a very small cohort (n = 8).

Here we see the response data for the higher doses of BI 764532, ≥90 µg/kg, which are thought to be the meaningful therapeutic doses of this agent.²⁴ The rate of partial responses in patients with SCLC is a bit higher, at 26%, with 3 of 5 patients with LCNEC achieving a partial response as well.

With the caveat that this is a dose-finding study and ORR is not the primary endpoint, these data seem to suggest that limiting your population to patients who are DLL3-positive does not necessarily have a significant impact on outcomes.  

Similar to what we saw with tarlatamab in DeLLphi-301,²⁰ BI 764532 elicited meaningful, durable responses in patients with advanced SCLC and LCNEC, again likely indicative of a real antitumor immunity occurring due to these BiTE therapies.²⁴

Together, the data for these DLL3-targeted BiTEs engender optimism in a disease where immune responses are not the norm—as mentioned, patients that are deriving long-term benefit from the addition of immunotherapy to chemotherapy is on the order of 15% to 20%.^6-9

We will now switch gears to discuss ADCs, another one of the strategies I mentioned earlier in terms of delivering therapy directly to the tumor.

The mechanism by which ADCs work is often analogized to a Trojan horse,²⁵ which I think is apt. The antibody component of an ADC binds to a cell surface molecule on the tumor cell, which is then internalized. Hiding on the antibody, connected via a linker, is a cytotoxic payload. Release of the payload into the tumor cell upon internalization of the ADC causes death not only in the cell that internalized it but also potentially in bystander cells.

Ifinatamab deruxtecan is an ADC being evaluated in ES-SCLC that targets B7-H3, a molecule expressed frequently on the surface of tumors, especially SCLC.²⁶ The payload of ifinatamab deruxtecan is the topoisomerase 1 inhibitor deruxtecan, which is derived from exatecan.

This slide describes the first-in-human, multicohort phase I/II dose escalation/expansion trial of ifinatamab deruxtecan (I-DXd), with a primary goal of identifying a dose of I-DXd that is both tolerable and efficacious.²⁶ This trial did not require demonstration of B7-H3 expression, so was theoretically conducted in all comers. However, B7-H3 expression was evaluated retrospectively.

The given analysis included 21 patients with SCLC dosed with I-DXd ≥6.4 mg/kg every 3 weeks. These patients had received median of 2 prior systemic therapies (range: 1-7), with 81.8% having received chemoimmunotherapy.

In this small cohort of patients with refractory SCLC, I-DXd achieved an ORR of 52.4%, with a CR rate of 4.8%.²⁶ On the left, we see another waterfall plot with the majority of patients showing some level of tumor reduction, something we are not used to seeing in the setting of relapsed SCLC. The DoR with I-DXd was 5.9 months, but in the figure on the right, you can see that several responses are ongoing well beyond that.

Simliar to the DLL3-targeted BiTE therapies, I-DXd appears to be a therapeutically meaningful agent in the setting of relapsed SCLC, and these data are also being followed up in a randomized phase III trial.

• IDeate-2: I-DXd vs physician’s choice of SoC therapy (topotecan, lurbinectedin, amrubicin) in relapsed ES-SCLC pretreated with 1 or more platinum-based regimen (NCT06203210)

Regarding the safety of I-DXd, grade ≥3 treatment-emergent AEs were seen in 36.4% of patients and were associated with death in 1 patient (4.5%).²⁶ As the table shows, the most common treatment-emergent AEs with I-DXd were nausea, fatigue, anemia, and vomiting. The rate of treatment-emergent AEs that led to treatment discontinuation was relatively high at 22.7%, so we will need to take a closer look at toxicities associated with I-DXd as more patients are treated.

This was a small number of patients with ES-SCLC, but 3 (13.6%) did develop grade 1/2 interstitial lung disease (ILD),²⁶ an AE we need to watch for when using ADCs that have deruxtecan as a payload in our patients with lung cancers.^27-29

With regard to diagnosis and management of ILD, should I-DXd become available in the clinic for SCLC, being aware of the risk for this pulmonary toxicity is the first step. If a patient experiences symptoms of a dry cough and occasional chest discomfort, imaging should be conducted promptly. If radiographic changes are seen (ie, potential evidence of grade 1 pneumonitis), treatment should be held while evaluations are done to determine other potential causes of the patient’s symptoms. For low-grade pneumonitis, oral systemic corticosteroids can be considered. For higher grades of pulmonary toxicity, empiric high-dose corticosteroids such as methylprednisolone 2 mg/kg every 6 to 8 hours or equivalent should be implemented, in addition to permanent drug discontinuation.³⁷

As previously mentioned, B7-H3 expression was not required for patients to be enrolled on this trial, but it was evaluated retrospectively.²⁶ As this figure shows, while the patient with the highest B7-H3 expression level was also the 1 patient who had a complete response, activity of I-DXd was otherwise observed across B7-H3 expression levels, which was quite robust across the board.

However, the verdict is still out on whether expression observed in a single biopsy, which may be heterogenous, is critical for the response here and whether it is a reliable threshold. It will be interesting to continue to explore the correlation between B7-H3 expression and response to I-DXd in larger, more definitive studies to see whether an expression threshold should be set in prioritizing this therapy for a patient. It might be that only a very low level of target expression is needed for these agents to be active, in particular since ADCs are known to exhibit a bystander effect.

Lastly, we will discuss another ADC, sacituzumab govitecan,^30,31 which targets TROP-2, a cell surface protein highly expressed across tumor types, including NSCLC and SCLC.³² This drug is probably familiar to many medical oncologists for its FDA-approved use in some other solid tumors, such as breast cancer.³³ It is also being investigated in NSCLC (NCT05089734; NCT05609968) and has shown promising preliminary activity with a tolerable safety profile both as monotherapy and in combination with immune checkpoint inhibition.^34,35

The payload of sacituzumab govitecan is topoisomerase 1 inhibitor SN38, which is derived from irinotecan.^30,31 Although its payload is different than I-DXd, sactizumab govitecan has a comparable mechanism of action.

Sacituzumab govitecan is being evaluated in relapsed ES-SCLC as part of the open-label phase II TROPiCS-03 basket trial.³⁶ In the ES-SCLC cohort, which required patients to have received 1 prior line of platinum-based chemotherapy and anti—PD-1/PD-L1 therapy, sacituzumab govitecan was delivered at 10 mg/kg on Day 1 and Day 8 of every 21-day cycle. The primary endpoint was ORR.

Here we see the response data for sacituzumab govitecan in patients with relapsed ES-SCLC.³⁶ The ORR was 37%, all confirmed partial responses. In the waterfall plot, you can see that several other patients were approaching a partial response, with 77% experiencing some amount of tumor reduction.

This is another agent to be excited about in the relapsed setting for ES-SCLC. In general, there is much enthusiasm and optimism for ADCs, many of which are under development with data anxiously awaited, for the treatment of SCLC.