Immunotherapy Before Surgery for Esophageal and Gastroesophageal Junction Cancer: Results from a Phase Ib Trial

Table of Contents

• Key Points
• Background: Why This Research Matters
• How the Study Was Designed
• Who Participated in the Study
• Safety Results and Side Effects
• Feasibility: Could Patients Get to Surgery?
• How Well Did the Treatments Work?
• Biomarker Findings: PD-L1 and LAG-3
• Circulating Tumor DNA (ctDNA) Analysis
• What These Findings Mean for Patients
• Limitations of the Study
• Recommendations and Next Steps
• Frequently Asked Questions
• Source Information

Key Points

• In a 32-patient phase Ib trial, nivolumab plus chemoradiotherapy met its safety goal for resectable stage II or III esophageal and gastroesophageal junction cancer.
• Adding relatlimab during chemoradiation caused unacceptable toxicity, so the trial was amended to give it only during induction, which was well tolerated.
• Pathological complete response occurred in 40% of the nivolumab group and 21.4% of the dual immunotherapy group among 29 patients who had surgery.
• Two-year recurrence-free survival was 72.5% and two-year overall survival was 82.6% overall; these early results need confirmation in larger randomized trials.
• Undetectable ctDNA at multiple time points was linked to longer survival, but ctDNA testing remains exploratory and not yet standard of care.

Background: Why This Research Matters

Over the past few decades, there have been limited treatment advances for patients with resectable gastroesophageal cancers. The majority of patients experience disease progression and death within 5 years from diagnosis. This makes new treatment approaches urgently needed.

Recently, a phase III study called CheckMate 577 led to Food and Drug Administration (FDA) approval of adjuvant nivolumab. The approval is for patients with completely resected esophageal or gastroesophageal junction (E/GEJ) cancer who had residual pathologic disease after neoadjuvant chemoradiotherapy (CRT). This approval represented a paradigm shift in how doctors manage operable stage II and stage III disease.

Early stage gastroesophageal tumors may express high levels of programmed cell death ligand 1 (PD-L1). PD-L1 is a protein that helps cancer cells hide from the immune system. When tumors have high PD-L1 levels, this indicates an adaptive immune resistance mechanism that may be reversed by anti-PD-1 antibodies. Anti-PD-1 antibodies are drugs that block this hiding mechanism and help the immune system attack cancer cells.

Preclinical and human studies have indicated that neoadjuvant CRT (chemotherapy and radiation given before surgery) may have a PD-L1 priming effect in operable E/GEJ cancer. This supports combining CRT with immune checkpoint inhibitors (ICI). Immune checkpoint inhibitors are drugs that release the brakes on the immune system.

Lymphocyte-activation gene 3 (LAG-3) is a co-inhibitory receptor that is highly expressed in gastroesophageal cancers. A co-inhibitory receptor is a protein on immune cells that dampens their activity. Combined anti-LAG-3 and anti-PD-1 therapy has the potential to modulate immune checkpoint pathways, re-invigorate exhausted T cells, and thus enhance antitumor immune responses. T cells are a type of immune cell that can kill cancer cells.

Limited data exist for the use of neoadjuvant PD-1 pathway inhibitors in combination with CRT in operable E/GEJ cancer, with conflicting results to date. While the integration of immunotherapy into the treatment paradigm of gastroesophageal cancer is of paramount importance, the broad efficacy of immunotherapy in this disease remains elusive.

There has been an expanding application of neoadjuvant immunotherapy for operable cancer. This is based on the notion and evidence that immunotherapy capitalizes on the primary tumor as a source of tumor antigens that could "prime" tumor-specific T cells. These primed T cells could then seek out micrometastatic disease that ultimately drives recurrence after curative-intent surgery. Micrometastatic disease refers to tiny clusters of cancer cells that have spread but are too small to detect with standard imaging.

Assessments of systemic tumor burden by analyses of circulating tumor DNA (ctDNA) may accurately and rapidly determine tumor regression with neoadjuvant ICI. ctDNA refers to small fragments of tumor DNA that circulate in the bloodstream. Preoperative ctDNA assessments may identify individuals more likely to attain pathologic complete responses with neoadjuvant ICI. Postoperative ctDNA detection may identify patients with minimal residual disease (MRD) and an increased risk of disease relapse that may benefit from sequential therapy. MRD refers to cancer cells that remain in the body after treatment but are not detectable by standard tests.

Nevertheless, the predictive versus prognostic role of ctDNA remains unclear. No studies to date have examined ctDNA in patients with gastroesophageal cancer treated in the neoadjuvant immuno-CRT setting. This study aimed to fill that gap.

How the Study Was Designed

This was a phase Ib clinical trial conducted at multiple centers. A phase Ib trial is an early-stage study designed primarily to evaluate safety and find the right dose of a treatment.

Patients aged 18 years and older with clinical stage II or stage III distal E/GEJ adenocarcinoma or squamous cell carcinoma (SCC) were eligible to enroll. Adenocarcinoma is a type of cancer that starts in glandular cells. Squamous cell carcinoma is a type of cancer that starts in flat, thin cells lining the esophagus. All patients had to have surgically resectable disease and be candidates for standard of care CRT followed by surgery.

The study enrolled patients into two treatment groups consecutively:

• Arm A: Nivolumab every 2 weeks for two induction cycles, then three additional doses given concurrently with CRT
• Arm B: Nivolumab and relatlimab every 2 weeks according to the same schedule

Patients were enrolled in Arm B only after safety and feasibility objectives were met in Arm A. The primary endpoint of the trial was safety. The secondary endpoint was feasibility. Exploratory endpoints included overall survival (OS), recurrence-free survival (RFS), major pathological response (MPR), and pathological complete response (pCR) rates.

Baseline CT and PET/CT scans were obtained before the first dose of neoadjuvant treatment. PET/CT was obtained again after completion of neoadjuvant treatment before surgery. Tumor samples were collected at baseline, after two cycles of induction immunotherapy, and at the time of surgery.

Serial blood samples were collected at baseline, start of cycle 2, start of cycle 3, before surgery, and within 3 to 12 weeks after surgery. Exploratory analyses of serial ctDNA and assessment of neoantigen-specific T cells in peripheral blood, together with gene expression analyses, were also performed on serial biospecimens in each arm. Neoantigen-specific T cells are immune cells that recognize specific abnormal proteins produced by cancer cells.

Arm B was amended to nivolumab plus relatlimab only during induction due to toxicities. This means that after the first 9 patients in Arm B experienced significant side effects, the remaining 7 patients received only two cycles of immunotherapy before chemoradiation, without continuing immunotherapy during chemoradiation.

Who Participated in the Study

From August 2017 to July 2021, 42 patients were screened and 32 patients were enrolled. Of the 42 patients screened, 8 did not meet inclusion criteria and 2 withdrew consent.

The clinical and pathological characteristics of the 32 enrolled patients are summarized below:

• Age: Mean age was 63 years (standard deviation 7.95 years). Median age was 65 years (range 39 to 73 years). In Arm A, mean age was 60.25 years. In Arm B, mean age was 65.75 years.
• Sex: 26 patients (81.25%) were male and 6 patients (18.75%) were female. Both arms had the same distribution: 13 males and 3 females each.
• Smoking history: 17 patients (53.12%) had never smoked. 15 patients (46.88%) were former smokers.
• Clinical stage: 20 patients (62.5%) had stage II disease. 12 patients (37.5%) had stage III disease.
• Nodal involvement: 24 patients (75.0%) had cancer that had spread to lymph nodes.
• Anatomic location: 26 patients (81.25%) had primary esophageal tumors. 6 patients (18.75%) had gastroesophageal junction tumors.
• Histology: 28 patients (87.5%) had adenocarcinoma. 4 patients (12.5%) had squamous cell carcinoma.

Patients predominantly had adenocarcinoma histology (87.5%), primary esophageal tumors (81.3%), and nodal involvement (75.0%).

Sixteen patients received nivolumab as induction for two cycles, then in combination with CRT for a total of five doses (Arm A). In Arm B, the first 9 of 16 patients received nivolumab and relatlimab following the same schedule as in Arm A. The remaining 7 patients received only nivolumab and relatlimab as induction due to a protocol amendment for toxicity.

Overall, 28 patients completed the full course of neoadjuvant therapy. Four patients (one patient in Arm A and three patients in Arm B) required ICI discontinuation due to immunotherapy-related adverse events (irAEs).

Nivolumab and relatlimab combined with CRT demonstrated unacceptable toxicity, requiring a protocol amendment per predefined early stopping rules. Six of the first nine patients (66%) treated with dual ICI plus CRT developed grade 3 or higher irAEs. These included pericarditis (inflammation of the sac around the heart) in 2 of 9 patients (22%) and adrenal insufficiency (when the adrenal glands do not produce enough hormones) in 2 of 9 patients (22%).

The nivolumab/relatlimab arm was thus amended to include two cycles of nivolumab and relatlimab only as induction before chemoradiation. This amended schedule was subsequently well tolerated.

All patients received both cycles of induction ICI. 92.1% of planned systemic therapy cycles were administered.

After completion of neoadjuvant therapy, two patients were ineligible for surgery. One patient had disease progression, and one patient had CRT-related decreased performance status. One patient declined surgery. Twenty-nine surgical candidates underwent Ivor Lewis esophagectomy within a median of 8 weeks from completion of CRT (range 3.6 to 11.4 weeks). Ivor Lewis esophagectomy is a surgical procedure to remove part of the esophagus and the upper part of the stomach.

Eight patients (28%) received adjuvant therapy as per standard of care. Adjuvant therapy is treatment given after surgery to reduce the risk of cancer returning. Three patients in Arm A received adjuvant FOLFOX (folinic acid, 5-fluorouracil, and oxaliplatin) with a median duration of 12 weeks. Four patients in Arm B received adjuvant nivolumab, all ongoing at the time of data lock.

Safety Results and Side Effects

The clinical trial met its primary endpoint of safety for Arm A, which evaluated nivolumab plus chemoradiation. However, Arm B required an amendment to mitigate toxicity.

Treatment-related adverse events (TRAEs) of any grade occurred in all patients. TRAEs are side effects caused by treatment. These were most often related to chemotherapy or radiation.

Grade 3 or higher TRAEs occurred in 31.3% of patients (95% confidence interval [CI]: 16.1–50.0%). Grade 3 side effects are serious but not life-threatening. The rate was 18.8% in Arm A (95% CI: 4.0–45.6%) and 43.8% in Arm B (95% CI: 19.8–70.1%).

Overall, 17 patients (53.1%, 95% CI: 34.7–70.9%) experienced immune-related adverse events (irAEs). The rate was 31.3% in Arm A (95% CI: 11.0–58.7%) and 75.0% in Arm B (95% CI: 47.6–92.7%). These were generally low grade.

The most common irAEs were:

• Dermatitis (skin inflammation): 31.3% of patients
• Elevated AST (aspartate aminotransferase) or ALT (alanine transaminase) liver enzymes: 12.5% of patients
• Hypothyroidism (underactive thyroid): 12.5% of patients

There was no grade 3 or higher treatment-related pneumonitis (lung inflammation) or acute respiratory failure. One patient in Arm B experienced grade 2 pneumonitis, most consistent with a radiation-induced process. This resolved with a short course of prednisone, which was already prescribed for concomitant irAE dermatitis.

Grade 3 or higher irAEs occurred in 8 patients (25.0%, 95% CI: 11.5–43.4%) of all treated patients. These were more prevalent with the dual ICI regimen: 2 patients (12.5%, 95% CI: 0.016–38.3%) in Arm A and 6 patients (37.5%, 95% CI: 15.2–64.6%) in Arm B. The most common grade 3 or higher irAE was dermatitis (4 patients, 12.5%, 95% CI: 3.5–29.0%). Pericarditis (2 patients, 6.3%) and adrenal insufficiency (2 patients, 6.3%) were also notable.

Both cases of pericarditis were from Arm B and required hospitalization. One patient presented after the first week of concurrent ICI and CRT with acute chest pain and EKG changes suggestive of pericarditis. This resolved with NSAIDs (nonsteroidal anti-inflammatory drugs) and colchicine. The second patient had already completed their neoadjuvant course and presented with acute pericardial effusion (fluid around the heart) in the setting of cardiogenic shock on postoperative day 11. This resolved with a pericardial window procedure, NSAIDs, and colchicine.

Neoadjuvant immunotherapy was discontinued in four patients due to grade 3 irAEs. Two patients had dermatitis, one had elevated liver enzymes, and one had pericarditis.

With respect to primary adrenal insufficiency, both patients received significant radiation to the adrenal glands. However, adrenal insufficiency was likely related to dual ICI rather than radiation.

Overall, irAEs necessitated immunotherapy hold in 18.8% of patients (95% CI: 7.2–36.4%). This was 2 patients (12.5%, 95% CI: 1.6–38.3%) in Arm A and 4 patients (25.0%, 95% CI: 7.3–52.4%) in Arm B. IrAEs necessitated discontinuation in 12.5% of patients (95% CI: 3.5–29.0%). This was 1 patient (6.3%, 95% CI: 0–30.2%) in Arm A and 3 patients (18.8%, 95% CI: 4.0–45.6%) in Arm B.

None of the seven patients in the amended Arm B who received immunotherapy only as induction experienced an irAE-related treatment hold or discontinuation.

There was one death in the immediate postoperative period due to septic shock unrelated to systemic therapy. Septic shock is a severe whole-body response to infection.

Feasibility: Could Patients Get to Surgery?

Feasibility was assessed through the proportion of eligible patients who proceeded to surgery without substantial treatment-related delay. Substantial delay was defined as more than 11 weeks from completion of chemoradiation.

A Bayesian continuous monitoring plan was used to monitor if the proportion was evidently greater than 75%. This means at most 25% of cases could have substantial delays.

Of the patients eligible for surgery after completion of neoadjuvant therapy, one patient declined surgery (Arm B). Another patient underwent resection at 11.4 weeks after chemoradiation due to travel logistics (Arm B). All other operative candidates proceeded to planned surgery without substantial delay. This represented 100% feasibility.

How Well Did the Treatments Work?

In the cohort of 29 patients that underwent surgery, the pCR rate was 40.0% in Arm A (95% CI: 16.3–67.7%) and 21.4% in Arm B (95% CI: 4.7–50.8%). A pathological complete response (pCR) means no cancer cells were found in the removed tissue.

Nine patients (31.03%) had a College of American Pathologists (CAP) tumor regression score of grade 0 (G0). Seven patients (24.14%) had G1, ten patients (34.48%) had G2, and three patients (10.34%) had G3. These grades correspond to zero, 1–10%, 11–50%, and 51–100% residual viable tumor, respectively.

Major pathological response (MPR), defined as 10% or less residual tumor cells after neoadjuvant therapy, occurred in 53.5% of patients in Arm A (95% CI: 26.6–78.7%) and 57.1% in Arm B (95% CI: 28.9–82.3%).

By histology, pCR and MPR rates were:

• Adenocarcinoma: pCR 30.8% (95% CI: 14.3–51.8%) and MPR 50.0% (95% CI: 30.0–70.0%)
• Squamous cell carcinoma: pCR 33.33% (95% CI: 0–90.6%) and MPR 100.0% (95% CI: 29.2–100%)

The R0 resection rate was 100%. R0 resection means all visible cancer was removed during surgery. The median follow-up time was 36.4 months in the overall cohort (95% CI: 29.9–43.9). It was 43.9 months in Arm A (95% CI: 43.2–not reached) and 27.5 months in Arm B (95% CI: 18.7–33.9).

Median recurrence-free survival (RFS) was 41.1 months overall (95% CI: 29.4–not reached). RFS is the time after treatment during which no cancer returns. It was 34.1 months in Arm A (95% CI: 21.6–not reached) and not reached in Arm B.

The overall 2-year RFS rate was 72.5% (95% CI: 57.8–90.1%). It was 62.5% in Arm A (95% CI: 42.8–91.4%) and 87.1% in Arm B (95% CI: 71.8–100%).

The overall 2-year OS rate was 82.6% (95% CI: 69.7–97.8%). It was 75.0% in Arm A (95% CI: 56.5–99.5%) and 93.8% in Arm B (95% CI: 82.6–100%). Median overall survival (OS) was not reached in either arm.

Biomarker Findings: PD-L1 and LAG-3

In evaluating baseline tumor specimens with adequate tissue for immunohistochemistry, PD-L1 combined positive score (CPS) was evaluable for 29 of 32 patients (90.6%). Immunohistochemistry is a laboratory method that uses antibodies to check for certain proteins in tissue. CPS is a scoring system that measures how much PD-L1 is present on tumor and immune cells.

Eleven tumors (11 of 29, 37.9%) were negative for PD-L1 (CPS less than 1). Eighteen tumors (18 of 29, 62.1%) were positive for PD-L1. There was no significant difference in baseline PD-L1 CPS between patients in Arm A and Arm B (Fisher's exact test, P = 0.95).

In resected specimens, PD-L1 expression was evaluable for 15 of 20 patients (75%) with residual tumor. Three tumors (3 of 15, 20%) were negative for PD-L1 (CPS less than 1). Twelve tumors (12 of 15, 80%) were positive for PD-L1.

Human epidermal growth factor receptor 2 (HER2) expression was evaluable in 15 of 20 resected specimens (75%). The majority were HER2 negative (13 patients, 86.67%). HER2 is a protein that can promote cancer cell growth. Thirteen of thirteen evaluable resected tumors (100%) were mismatch repair (MMR) proficient by immunohistochemistry. MMR proficiency means the cells have normal ability to repair DNA errors.

Overall, baseline PD-L1 expression was not associated with pCR or MPR. There was a trend toward enrichment for baseline PD-L1 CPS of 5 or higher in tumors with pCR (Fisher's exact test, P = 0.089).

Patients with tumors with baseline PD-L1 CPS of 5 or higher had a longer RFS. Median RFS was not reached for CPS of 5 or higher versus 29.34 months for CPS less than 5 (log-rank test, P = 0.013). There was a trend toward longer OS (median OS not reached for both groups; log-rank test, P = 0.13).

Similarly, patients with adenocarcinomas harboring a baseline PD-L1 CPS score of 5 or higher had a longer RFS. Median RFS was not reached versus 29.34 months (log-rank test, P = 0.026).

Researchers then evaluated changes in PD-L1 expression in resected tumors after immuno-CRT compared to baseline. They did not observe significant differences (n = 14; Fisher's exact test, P = 0.26 for all patients and n = 13; Fisher's exact test, P = 0.24 for the adenocarcinoma subset).

There were no significant differences in OS or RFS for patients with PD-L1 increase after immuno-CRT (n = 5). This was compared to those without PD-L1 expression changes (n = 9; log-rank test, P = 0.82 and P = 0.42 for OS and RFS, respectively).

As part of the study's exploratory analyses, researchers evaluated the association between baseline LAG-3 expression and pathologic response. Patients who attained an MPR (n = 11) had a higher LAG-3 expression at baseline compared to patients with a non-MPR (n = 11). Median LAG-3 expression was 6.68 versus 6.01, respectively (two-sided Wilcoxon rank-sum test, P = 0.016).

These findings were driven by Arm B. Patients who attained an MPR (n = 4) had a higher baseline LAG-3 expression (median LAG-3 expression 6.77 versus 5.95, respectively; two-sided Wilcoxon rank-sum test, P = 0.016).

Circulating Tumor DNA (ctDNA) Analysis

Circulating tumor DNA (ctDNA) analysis was a key exploratory component of this study. ctDNA refers to small fragments of DNA released by tumor cells into the bloodstream. Measuring ctDNA can provide information about tumor burden and response to treatment.

Researchers analyzed serial blood samples to detect tumor-derived variants. Tumor-derived variants are specific genetic changes found in the cancer cells. Most tumor-derived variants were patient-specific. There were no recurrent mutations associated with therapeutic efficacy.

Twelve patients without any tumor-derived variants detected at any time point were classified in the ctDNA undetectable group (ctDNA UD). These patients were considered nonevaluable for subsequent circulating tumor burden assessments.

Researchers defined ctDNA as detected (ctDNA+) if at least one tumor-derived variant was found. 37.5% of patients (12 of 32) had only clonal hematopoiesis (CH)-derived or germline variants detected. Clonal hematopoiesis refers to age-related changes in blood cells that can produce DNA variants not related to cancer. Germline variants are inherited genetic changes.

The study found that patients with undetectable ctDNA after ICI induction, before surgery, and after surgery had a significantly longer RFS and OS. ctDNA clearance was reflective of neoantigen-specific T cell responses. This means that when ctDNA disappeared from the blood, it corresponded with the immune system mounting a specific attack against the tumor.

These findings highlight the potential of ctDNA analysis to dynamically assess systemic tumor burden during neoadjuvant ICI. This may open a therapeutic window for future intervention.

What These Findings Mean for Patients

This study provides important insights into the safety profile of combined PD-1 and LAG-3 blockade in gastroesophageal cancer. The dual immunotherapy approach (nivolumab plus relatlimab) combined with chemoradiotherapy caused unacceptable toxicity in the initial dosing schedule. This required a protocol amendment to give the dual immunotherapy only during the induction phase before chemoradiation.

For patients considering immunotherapy before surgery for esophageal or gastroesophageal junction cancer, this study offers several key takeaways:

• Nivolumab plus chemoradiotherapy was safe and feasible in this patient population.
• Adding relatlimab to nivolumab during chemoradiation caused too many serious immune-related side effects, including pericarditis and adrenal insufficiency.
• When relatlimab was given only during the induction phase (before chemoradiation), the treatment was well tolerated.
• Pathological complete response rates were 40% in the nivolumab group and 21.4% in the dual therapy group.
• Two-year recurrence-free survival was 72.5% overall, and two-year overall survival was 82.6% overall.
• Patients with higher baseline PD-L1 expression (CPS of 5 or higher) had longer recurrence-free survival.
• Patients with higher baseline LAG-3 expression were more likely to achieve major pathological response.
• Undetectable ctDNA at multiple time points was associated with significantly longer recurrence-free and overall survival.

The ctDNA findings are particularly promising. They suggest that a simple blood test may help doctors monitor how well immunotherapy is working. If ctDNA becomes undetectable after treatment, patients may have a better prognosis. If ctDNA remains detectable, doctors may consider additional treatments.

The biomarker findings also have implications for treatment selection. Patients with high PD-L1 expression may benefit more from immunotherapy. Patients with high LAG-3 expression may benefit specifically from LAG-3 inhibitors like relatlimab. However, these findings need confirmation in larger studies.

Limitations of the Study

This was a phase Ib trial with a small number of patients. Only 32 patients were enrolled, with 16 in each treatment arm. Small studies can show trends but cannot definitively prove that one treatment is better than another.

The study was not randomized. Patients were enrolled consecutively into Arm A first, then Arm B. This means that differences between the arms could be due to factors other than the treatment itself.

Arm B required a protocol amendment during the study. The first 9 patients received dual immunotherapy throughout chemoradiation, while the remaining 7 patients received dual immunotherapy only during induction. This change makes it harder to interpret the results for Arm B as a whole.

The follow-up time differed between the arms. Arm A had a median follow-up of 43.9 months, while Arm B had a median follow-up of 27.5 months. This difference could affect the survival comparisons.

The ctDNA analysis was exploratory. While the findings are promising, they need to be validated in larger, prospective studies before ctDNA testing can be routinely used to guide treatment decisions.

The biomarker analyses were also exploratory and involved small numbers of patients. The associations between PD-L1, LAG-3, and treatment response need confirmation in larger studies.

The study population was predominantly male (81.25%) and had adenocarcinoma histology (87.5%). The findings may not apply equally to female patients or to patients with squamous cell carcinoma.

Recommendations and Next Steps

Based on this study, several recommendations emerge for patients and clinicians:

1. For patients considering neoadjuvant immunotherapy: Nivolumab plus chemoradiotherapy is a safe and feasible option for patients with resectable stage II or stage III esophageal or gastroesophageal junction cancer.
2. Regarding dual immunotherapy: Adding relatlimab to nivolumab during chemoradiation caused unacceptable toxicity, but because of promising long-term efficacy the researchers said future studies should further evaluate the optimal sequencing of dual immunotherapy given together with chemoradiation. However, giving relatlimab only during the induction phase before chemoradiation appears safe and warrants further study.
3. Biomarker testing: Testing for PD-L1 and LAG-3 expression at baseline may help identify patients who are more likely to benefit from immunotherapy. Patients with high PD-L1 expression may have longer recurrence-free survival. Patients with high LAG-3 expression may be more likely to achieve major pathological response.
4. ctDNA monitoring: Blood tests for ctDNA may help monitor treatment response. Undetectable ctDNA after treatment is associated with better outcomes. However, ctDNA testing is not yet standard of care and should be done in the context of clinical trials or research settings.
5. Future research: Larger randomized trials are needed to confirm these findings. Future studies should also investigate whether adjusting treatment based on ctDNA results can improve outcomes.

The researchers concluded that their findings provide insights into the safety profile of combined PD-1 and LAG-3 blockade in gastroesophageal cancer. They also highlighted the potential of ctDNA analysis to dynamically assess systemic tumor burden during neoadjuvant ICI. This may open a therapeutic window for future intervention.

For patients with esophageal or gastroesophageal junction cancer, this study represents another step forward in understanding how immunotherapy can be best used before surgery. While the dual immunotherapy approach requires further refinement, the ctDNA findings offer hope for more personalized treatment monitoring in the future.

Frequently Asked Questions

Who was eligible to join this trial?

Patients aged 18 and older with clinical stage II or stage III distal esophageal or gastroesophageal junction adenocarcinoma or squamous cell carcinoma could enroll. All needed surgically resectable disease and had to be candidates for standard chemoradiotherapy followed by surgery. Of 42 people screened, 32 enrolled; 8 did not meet inclusion criteria and 2 withdrew consent.

What were the two treatment groups in this trial?

One group received nivolumab every two weeks for two induction cycles, then three more doses with chemoradiotherapy. The other received nivolumab plus relatlimab on the same schedule. Patients joined the second group only after safety and feasibility goals were met in the first. The second group was later amended because of toxicity.

What side effects did patients experience?

Treatment-related side effects occurred in all patients, most often from chemotherapy or radiation. Grade 3 or higher side effects occurred in 31.3% overall: 18.8% in the nivolumab group and 43.8% in the dual immunotherapy group. Immune-related side effects occurred in 53.1% overall, most commonly skin inflammation, elevated liver enzymes, and underactive thyroid.

Why was the dual immunotherapy group changed during the trial?

Six of the first nine patients (66%) receiving nivolumab plus relatlimab with chemoradiotherapy developed grade 3 or higher immune-related side effects, including pericarditis and adrenal insufficiency. This met predefined early stopping rules. The group was amended so the remaining seven patients received the dual immunotherapy only as induction before chemoradiation, which was well tolerated.

How many patients were able to have surgery after treatment?

Of patients eligible for surgery after finishing neoadjuvant therapy, one declined and one had resection at 11.4 weeks due to travel logistics. All other operative candidates proceeded without substantial delay, representing 100% feasibility. Twenty-nine patients underwent Ivor Lewis esophagectomy within a median of 8 weeks after chemoradiation, and the R0 resection rate was 100%.

What did the trial find about pathological complete response?

Among the 29 patients who had surgery, no cancer cells were found in the removed tissue in 40.0% of the nivolumab group and 21.4% of the dual immunotherapy group. Major pathological response, meaning 10% or less residual tumor, occurred in 53.5% and 57.1% respectively. These are early-phase results from a small trial.

What did the ctDNA blood test results suggest?

Patients with undetectable circulating tumor DNA after immunotherapy induction, before surgery, and after surgery had significantly longer recurrence-free and overall survival. ctDNA clearance corresponded with immune responses against the tumor. These analyses were exploratory, so ctDNA testing is not yet standard of care and should be done in research settings.

When should a patient with resectable stage II or stage III esophageal or gastroesophageal junction cancer seek a second opinion before immunotherapy and surgery?

A second opinion is reasonable when weighing neoadjuvant options for resectable stage II or stage III disease. Nivolumab plus chemoradiotherapy met its safety goal, while adding relatlimab during chemoradiation caused unacceptable toxicity, including pericarditis and adrenal insufficiency, and required a protocol change. Pathological complete response was 40% with nivolumab and 21.4% with dual therapy; two-year recurrence-free survival was 72.5% and overall survival 82.6%. Baseline PD-L1, LAG-3 and ctDNA results may inform selection. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: Neoadjuvant nivolumab or nivolumab plus LAG-3 inhibitor relatlimab in resectable esophageal/gastroesophageal junction cancer: a phase Ib trial and ctDNA analyses

Authors: Ronan Joseph Kelly, Blair V. Landon, Ali Hussainy Zaidi, Dipika Singh, Jenna VanLiere Canzoniero, Archana Balan, et al.

Publication: Nature Medicine, Volume 30, April 2024, pages 1023–1034

DOI: https://doi.org/10.1038/s41591-024-02877-z

ClinicalTrials.gov registration: NCT03044613

Received: 5 June 2023

Accepted: 16 February 2024

Published online: 19 March 2024

This patient-friendly article is based on peer-reviewed research. It is intended for educational purposes only and should not replace consultation with a qualified healthcare provider. Patients should discuss their individual treatment options with their medical team.