Many advances have been made in molecularly targeted therapies for NSCLC, particularly since 2013.^1-3 It started with the FDA approval of an agent for patients with EGFR alterations, followed by approvals for NSCLC with driver mutations in ALK, ROS1, BRAF^V600E, NTRK, RET, METex14, KRAS^G12C, EGFRex20, and HER2^mut. In patients with no targetable molecular alterations, immunotherapy is used alone or in combination for advanced NSCLC. Whether treatment with a targeted agent occurs in the first, second, or third line of therapy depends on the specific alteration.

Approximately 69% of patients with advanced NSCLC have a known targetable driver mutation.^1,4-15 The most commonly altered genes are KRAS in 25% (KRAS^G12C in 14%), EGFR sensitizing in 17%, and ALK in 7%. In addition, 3% of patients have more than 1 targetable mutation. In 31% of patients, no known oncogenic drivers are detected.

For patients with advanced NSCLC who do not have a targetable molecular alteration, tumor histology and PD‑L1 expression status determine first-line treatment.^1,15,16 Ultimately, patients receive checkpoint inhibitor (CPI) monotherapy or dual therapy with or without chemotherapy if their Eastern Cooperative Oncology Group (ECOG) performance status (PS) is 0-2.

To know how to treat patients, we must test for molecular alterations and PD‑L1 expression status up front. PD-L1 expression is measured using IHC but molecular alterations can be detected in various ways, including single-gene testing (polymerase chain reaction, fluorescence in situ hybridization, or IHC), tissue-based NGS, and ctDNA-based NGS.^16-22 An important factor in choosing assays is that we need PD-L1 and molecular alteration results quickly and in parallel. Below I will discuss the benefits and drawbacks to each technology.

The tests assessing biopsy tissue are single-gene testing and tissue-based NGS. Single-gene testing uses a small amount of tissue, has short turnaround time, and is low cost, but it often requires sequential testing that may require additional biopsy material and ultimately takes longer and costs more than the alternatives. Therefore, current recommendations are to use large NGS panels evaluating DNA or RNA. NGS can detect novel alterations and test various genes concurrently and may detect comutations. However, it requires sufficient tissue so might require a new tissue biopsy. It also has a longer turnaround time and can miss tumor heterogeneity depending on the sampling. The main difference between DNA and RNA-based NGS is that RNA-NGS is more sensitive than DNA-NGS for gene fusion detection.

If a tissue biopsy is not feasible or insufficient tissue is available for tissue-based NGS, ctDNA-based NGS is particularly useful. ctDNA-NGS is minimally invasive and can be used to monitor patients who are receiving targeted treatment. The drawbacks are lower sensitivity compared with tissue biopsy and that negative results require confirmation using a tissue biopsy because of the possibility of a false negative.

Soon, we probably will measure the expression of additional proteins by IHC to determine eligibility for novel antibody‒drug conjugates (ADCs). So, we increasingly need huge biopsies for all this testing.

Patients with NSCLC who progress on or after frontline CPI have limited treatment options.^16,23 The standard option is generally docetaxel with or without a VEGFR2 inhibitor or another single-agent chemotherapy. Unfortunately, the long-term magnitude of benefit in response rate, progression-free survival (PFS), and overall survival (OS) is low. The challenge today is improving patient outcomes in second and third lines of therapy compared with docetaxel, particularly for patients with advanced disease.

ADCs are a new class of biopharmaceutical important for the treatment of NSCLC.^24,25 ADCs are composed of 3 parts: a chimeric/humanized monoclonal IgG antibody, a cytotoxic payload, and a linker attaching the two.

Ideally, the antibody targets an antigen found only on tumor cells that is not downregulated by binding of the antibody and has minimal nonspecific binding. The antibody can be IgG class 1, 2, 3, or 4 depending on the desired serum half-life, C1q binding, and Fcγ avidity.

The payload varies but is often an antimicrotubule or TOP1 inhibitor. It should be nonimmunogenic and nontoxic while the ADC circulates but highly potent in small quantities.

The linker ensures that the payload stays attached to the antibody in circulation but is efficiently released in tumor cells upon binding of the ADC to its target. The linker can be cleavable or noncleavable.

The properties of ADCs vary depending on the antibody, linker, and payload. The right combination of these components is necessary for optimal efficiency and limited toxicity.

Several ADCs are being evaluated in patients with advanced NSCLC. Trastuzumab deruxtecan was recently approved by the FDA for adult patients with unresectable or metastatic NSCLC. Additional investigational ADCs include datopotamab deruxtecan and sacituzumab govitecan, which target TROP-2; patritumab deruxtecan, which targets HER3; tusamitamab ravtansine, which targets CEACAM5; and telisotuzumab vedotin, which targets C-MET.^26-28 These ADCs differ in their properties, including linker type, payload, and the ratio of drug to antibody.

In general, hematologic toxicities are common with ADCs and include thrombocytopenia, neutropenia, and anemia.²⁹ Digestive toxicity also occurs and mainly includes grade 1/2 vomiting, diarrhea, and nausea. General toxicities such as fever and fatigue also are mainly grade 1/2. Of note, the differences in toxicity profiles among ADCs reflect the different cytotoxic payloads and sites of antibody binding. Some ADCs have liver toxicity and ocular toxicity. Tusamitamab ravtansine, which targets CEACAM5, is associated with keratopathy. Pneumonitis can occur—particularly with ADCs targeting HER2, such as trastuzumab deruxtecan—as can alopecia and peripheral neuropathy. Most of these toxicities are easy to manage.

CEACAM5 is overexpressed in multiple tumor types, including nonsquamous NSCLC.^30-32 CEA interactions promote tumor progression and metastasis.³² Membrane‑bound CEA on a cancer cell can directly engage a cell-surface receptor on a target cell. Secreted CEA can act in a paracrine manner, stimulating the secretion of protumorigenic and prometastatic cytokines. The binding of CEA to DC-SIGN induces tolerance in dendritic cells. CEA interacting with CEACAM1 on natural killer cells inhibits their cytotoxicity. CEA interactions are important targets in cancer, particularly lung cancer.

CEACAM5 expression is heterogenous but shows higher expression in nonsquamous vs squamous NSCLC.³³ Different areas of a tumor vary in membrane expression of CEACAM5. Some zones have polarized expression on the apical side of the membrane, and other zones have whole-membrane expression.

Pooled data from 214 NSCLC biopsies found high CEACAM5 expression in 24.3% of samples, moderate expression in 30.4%, and negative expression in 45.3%.³³ Among patients with high expression of CEACAM5, the most common expression pattern was mixed (12.1%), followed by whole-membrane expression (9.3%). Polarized expression was present in only 2.8% of the samples.

CEACAM5 expression by IHC correlates well with CEACAM5 protein expression as measured by ELISA and mRNA expression as measured by RNA sequencing.³³ The best and probably simplest way to measure CEACAM5 expression is with IHC.

Most cases of lung adenocarcinoma and lung squamous cell carcinoma had weak (0.5-1+) or moderate (1-2+) CEACAM5 expression.³⁴ By contrast, normal lung tissue did not express CEACAM5. This makes CEACAM5 a good target for an ADC.

In a phase I/II study of tusamitamab ravtansine, 888 patients with NSCLC were prescreened for CEACAM5 expression.³⁵ CEACAM5 was highly expressed in 172 patients (19.4%) and moderately expressed in 210 patients (23.6%).

Based on the expression profile of CEACAM5, tusamitamab ravtansine—an ADC-targeting CEACAM5 with a cytotoxic payload of maytansinoid DM4—has been tested in patients with advanced solid tumors, including NSCLC.³⁶ In the first-in-human, dose‑escalation phase I study excluding NSCLC, patients started at 5 mg/m² and went up to 210 mg/m² administered at 2-week intervals.

The most frequent primary tumor locations were colorectal (58.1%), stomach (22.6%), and gastroesophageal junction (9.7%).³⁶ Approximately 60% of patients overexpressed CEACAM5 on tissue biopsy by IHC, and the circulating CEA level was high in 66.7% of patients included in this phase I trial.

Tusamitamab ravtansine had good tolerability.³⁶ The only dose-limiting toxicity (DLT) in the first 2 cycles was grade 3 keratopathy, which affected patients receiving doses of 120 mg/m² and 150 mg/m². The maximum tolerated dose (MTD) was determined to be 100 mg/m².

We must be cautious in interpreting these results because of the low number of patients.³⁶ In total, 11 patients (35.5%) had a best overall response of stable disease, 3 patients (9.7%) had confirmed partial response (PR), and the rest had progressive disease. Among the 3 patients with a PR, 2 experienced grade 3 keratopathy. Two of the 3 patients with objective responses had membrane CEACAM5 expression graded as ≥2+ in 100% of their tumor cells; both patients had colorectal cancer.

Subsequently, part 2 of this phase I trial escalated the loading dose of tusamitamab ravtansine on Day 1, cycle 1, followed by the MTD (100 mg/m²) every 2 weeks. Part 3 escalated the dose on an every-3-week schedule.³⁷

Part 2 included 28 patients, with 1 patient in the lung cancer group.³⁷ Approximately two thirds of patients had high expression of CEACAM5 by IHC, and 75% had high circulating CEACAM5. The most common toxicities with tusamitamab ravtansine when investigating different loading doses in cycle 1 were keratopathy, keratitis, dry eye, peripheral sensory neuropathy, asthenia, and nausea.³⁷ Two of the 9 DLT-evaluable patients experienced a DLT at the 170 mg/m² loading dose level. One patient developed grade 2 keratitis and withdrew from therapy. The other developed grade 2 keratopathy, underwent a treatment delay, and then resumed therapy at a reduced dose.

In part 3, 15 patients evaluated every-3-week dosing; most had digestive tumors.³⁷ In total, 75% of patients had high expression of CEACAM5 by IHC, and the circulating CEA level was high in 100% of patients.

Two of the 3 DLT-evaluable patients experienced a DLT at 190 mg/m².³⁷ One patient experienced grade 3 liver toxicity during cycle 1, which recovered after drug withdrawal, and the other patient had grade 2 keratopathy during cycle 1, which recovered after treatment delay and dose reduction.

Keratopathy and keratitis were the most frequent DLTs in the escalation and loading dose parts. Increased transaminases and keratopathy DLTs were reported in the every-3-week dosing part. It is important to remember that corneal events are reversible and manageable by dose modification (dose delay and/or dose reduction).

In the phase II expansion part of this trial, tusamitamab ravtansine was evaluated in patients with advanced solid tumors at a dose of 100 mg/m² every 2 weeks.³¹ This phase of the trial included a cohort of 64 patients with nonsquamous NSCLC with high expression of CEACAM5 (≥2+ in ≥50%) and another cohort of 28 patients with nonsquamous NSCLC and moderate CEACAM5 expression (≥2+ in ≥1% and <50%) by IHC. Tumor assessment was every 4 cycles (8 weeks).

The primary endpoint was the overall response rate (ORR) in the expansion phase of the trial. Secondary endpoints included safety, recommended phase II dose identification, and duration of response (DoR).

The NSCLC cohorts were balanced, with a median age of 62.5 years and an ECOG PS of 1 in 70.7% of patients.³¹ Most patients with high CEACAM5 expression were male (57.8%), whereas male patients represented only 35.7% of the moderate CEACAM5 expression cohort. Patients were heavily pretreated, with a median of 3 prior regimens; 60.9% of patients previously received an antitubulin drug, and 75.0% previously received anti‒PD‑1/PD‑L1 therapy.

ORR was promising in patients with high CEACAM5 expression and was achieved by 13 patients (20.3%).³¹ Only 2 patients (7.1%) with moderate CEACAM5 expression achieved an ORR. All responses were confirmed PR. So, clearly the best responses appear to occur in patients with the highest expression of CEACAM5.

Among patients with high CEACAM5 expression, prior treatment with an antitubulin agent was more common among nonresponding patients than responding patients, at 64.7% vs 46.2%.³¹ Exposure to PD-1/PD-L1 inhibitors was quite similar, at 72.5% in nonresponders and 61.5% in responders.

At 5.6 months, the median DoR was promising in the CEACAM5 high‑expression population.³¹ It was not calculated for patients with moderate CEACAM5 expression.

The tolerability profile was generally good, with toxicity mainly grade 1/2.³¹ The most significant treatment-emergent adverse events (TEAEs) were corneal, with all-grade keratopathy/keratitis affecting 38.0% of patients and grade ≥3 affecting 10.9%. All-grade asthenia affected 37.0% of patients, peripheral neuropathy affected 27.2%, and diarrhea affected 22.8%. Hematologic toxicity was primarily grade 1/2 and included neutropenia (4.4%), anemia (75.8%), and thrombocytopenia (12.2%).

Grade 1/2 corneal AEs affected 27.2% of patients, and grade 3 affected 10.9%.³¹ In total, 25 patients (27.2%) developed a corneal TEAE leading to dose modification. All 25 patients had ≥1 dose delay, 10 patients (10.9%) had ≥1 dose reduction, and 1 patient (1.1%) permanently discontinued treatment.

Ocular events related to the DM4 payload of tusamitamab ravtansine are reversible noninflammatory deposits starting at the periphery of the cornea. The first occurrence was within the first 4 cycles of treatment for 28 patients (80%). It was manageable with dose delay and/or dose reduction, with a median time to recovery of 18.5 days. Primary prophylaxis is not effective; treatment with topical ophthalmologic corticosteroids when it occurs is recommended.

You can continue patients’ therapy if they develop grade 1 toxicity, but if it is grade ≥3, you must discontinue tusamitamab ravtansine. It can take time to recover, so if we see a patient develop grade 1 keratopathy, we quickly dose reduce. In my experience, dose reducing really improves tolerability, and keratopathy goes away.

Because tusamitamab ravtansine showed promising activity in heavily pretreated patients with advanced nonsquamous NSCLC and high CEACAM5 expression, the results are to be confirmed in an ongoing phase III trial. CARMEN‑LC03 is a confirmatory phase III trial evaluating tusamitamab ravtansine vs docetaxel in patients with metastatic nonsquamous NSCLC.³⁰ Patients enrolled on this trial are required to have high tumor CEACAM5 expression, an ECOG PS of 0/1, and metastatic disease progression after platinum‑based chemotherapy and CPI therapy. Patients were excluded if they had untreated brain metastases, a history of leptomeningeal disease, a history of unresolved corneal disorder, or previously received docetaxel. Randomization is 1:1 to receive tusamitamab ravtansine 100 mg/m² every 2 weeks vs docetaxel 75 mg/m² every 3 weeks. Treatment will continue until progression or unacceptable toxicity. The primary endpoints are PFS by blinded review committee and OS.

Some preclinical data suggest synergistic activity between tusamitamab ravtansine and ramucirumab.³⁸ The CARMEN‑LC04 study is a single-arm phase II trial evaluating the combination of these 2 agents in adults with metastatic nonsquamous NSCLC and an ECOG PS of 0/1. All patients have high CEACAM5 expression and had disease progression during or after platinum-based chemotherapy or CPI therapy.

Part 1 of the trial is a safety run-in with 6-12 patients receiving both drugs every 2 weeks. Treatment continues until disease progression, unacceptable toxicity, or patient withdrawal. The primary endpoint of part 1 is DLT at cycle 1 and cycle 2 and determination of the recommended phase II dose. Part 2 of the trial is an assessment of antitumor activity using the recommended phase II dose in 30 patients. The primary endpoint for this portion of the trial is ORR.

CARMEN‑LC05 is a phase II trial evaluating tusamitamab ravtansine in combination with pembrolizumab with or without platinum‑based chemotherapy, with or without pemetrexed (NCT04524689).³⁹ This trial is enrolling adults with advanced nonsquamous NSCLC with no prior CPI treatment for metastatic disease with moderate/high CEACAM5 expression, any PD-L1 expression, and an ECOG PS of 0/1. Patients cannot have EGFR, BRAF, ALK, or ROS1 alterations. The primary endpoint is DLT, and secondary endpoints are TEAEs, serious AEs, and ORR.

The first efficacy data with this combination are early but promising.³⁹ In total, 13 patients had confirmed PRs, and PR occurred in patients regardless of PD‑L1 expression status and high or intermediate CEACAM5 expression.

The response rates also are promising when evaluating by CEACAM5 expression.³⁹ The response rate was 36% in patients with high CEACAM5 expression (≥50%) compared with 20% in patients with intermediate CEACAM5 expression (1%-49%). These are early data, so we must be cautious until DoR and median PFS data confirm the magnitude of benefit and the role of this combination in first-line treatment of nonsquamous NSCLC.

The most frequent TEAEs were nausea (44%), diarrhea (36%) and asthenia (32%).³⁹ Grade ≥3 TEAEs were observed in 17 patients (68.0%) and grade 5 events in 4 patients (16.0%). All grade 5 events occurred in the arm of patients receiving tusamitamab ravtansine in combination with pembrolizumab, platinum-based chemotherapy, and pemetrexed, but they were deemed unrelated to tusamitamab ravtansine.

Corneal TEAEs of any grade were reported in 6 patients (24.0%) and were manageable with dose modification. Pneumonitis and peripheral neuropathy affected 16% and 28% of patients, respectively.

High levels of circulating CEA (≥100 ng/mL) have been found in the blood of some patients with NSCLC, but the relationship between tumor CEACAM5 expression and circulating CEA is unknown (NCT05245071).⁴⁰ The impact, if any, of circulating CEA on tusamitamab ravtansine efficacy also is not known. Some researchers have hypothesized that patients with negative or moderate tumor CEACAM5 expression but high circulating CEA still may benefit from this therapy. The CARMEN-LC06 trial will look at this question in 38 adults with metastatic nonsquamous NSCLC. The primary endpoint is ORR, and key secondary endpoints include TEAEs, PFS, and DoR.

Moving now to novel CEACAM5-targeted agents, NEO‑201 is a monoclonal antibody generated against tumor‑associated antigens from colorectal cancer.⁴¹ NEO-201 specifically binds to CEACAM5 and CEACAM6 variants and shows both antibody‑dependent cellular cytotoxicity and complement‑dependent cytotoxicity against cancer cells. A phase I trial evaluated NEO-201 in 17 patients with solid tumors, including adenocarcinoma of the pancreas, colorectal cancer, and breast cancer.

Because of the low number of patients, we must be cautious in interpreting the results, but a best response of stable disease was observed in 4 of the 9 evaluable patients with colorectal cancer.

DLTs with NEO-201 included grade 4 febrile neutropenia and prolonged neutropenia.⁴² The most frequent grade 3/4 toxicities were neutropenia (94%), white blood cell decrease (59%), lymphocyte decrease (29%), and febrile neutropenia (24%). Based on safety and pharmacokinetic data, the recommended phase II dose was established as 1.5 mg/kg.

NEO-201 also will be evaluated in combination with pembrolizumab in a currently enrolling phase I/II trial for patients with NSCLC, head and neck squamous cell carcinoma, cervical cancer, and uterine cancer progressing on frontline chemotherapy (NCT03476681). Patients will receive NEO-201 at the recommended phase II dose of 1.5 mg/kg every 2 weeks, with pembrolizumab given 1 day after NEO-201 at 400 mg IV every 6 weeks.

The last CEACAM5-targeted therapy I will discuss is the ADC M9140, which uses a TOP1 inhibitor payload (NCT05464030). A first-in-human clinical trial is evaluating the safety, tolerability, pharmacokinetics, and preliminary clinical activity of M9140 in patients with advanced solid tumors.

We continue to move forward in the treatment of lung cancer with novel targeted therapies.^43,44 Initially, we targeted the cell cycle using chemotherapy and radiotherapy. These are still important treatments. Signal transduction targeted treatments came next for alterations in EGFR, ALK, ROS1, BRAF, NTRK, MET, RET, KRAS, and NRG1. Agents targeting new molecular alterations are coming. Now we are targeting surface proteins with ADCs, an important new class of treatment for NSCLC. The cell-surface proteins being targeted with ADCs are ERBB2, ERBB3, TROP2, and CEACAM5, with other protein targets under investigation.

We currently test patients for molecular alterations and PD-L1 expression, but we will soon need to do more IHC testing for surface protein expression.^1,45 We ultimately may need to test for expression of CEACAM5, HER3, HER2, MET, and TROP-2 to better determine which therapy or combination of therapies should be used for our patients with advanced NSCLC.

Several ongoing phase III trials are evaluating ADCs targeting TROP-2, HER3, HER2, and CEACAM5 in NSCLC (NCT05089734, NCT04656652, NCT05215340, NCT04619004, NCT05048797, NCT04154956). Many of these trials are evaluating these ADCs vs docetaxel in the second or third line of therapy after platinum-based chemotherapy plus anti‒PD-1/PD-L1.

Finally, there are several mechanisms of resistance to ADCs. The first is through downregulation of target antigen expression by tumor cells.²⁴ If a cell does not express the target antigen, ADCs do not bind, and payload molecules are not released. Another resistance mechanism is through altered intracellular trafficking pathways or drug breakdown in lysosomes. In this case, the ADC will be ejected from the cell via endosome recycling prior to payload release. The third mechanism of ADC resistance is through the upregulation of drug efflux pumps in the cell membrane. Although the payload gets released into the cytoplasm, the cell is protected from cytotoxic damage if there is active payload efflux moving it to the exterior of the cell. Although we have identified several mechanisms of resistance to ADCs, we still need to learn more about which mechanisms are the most important and how to treat patients who develop resistance.