Initial results of pulmonary resection following neoadjuvant nivolumab in patients with resectable non-small cell lung cancer

Abstract

Objective

We conducted a phase I trial of neoadjuvant nivolumab, a monoclonal antibody to the PD-1 checkpoint receptor, in patients with resectable non-small cell lung cancer (NSCLC). We analyzed perioperative outcomes to assess the safety of this strategy.

Methods

Patients with untreated stage I-IIIA NSCLC underwent neoadjuvant therapy with two cycles of nivolumab (3 mg/kg), 4 and 2 weeks before resection. Patients underwent invasive mediastinal staging as indicated and posttreatment computed tomography. Primary study endpoints were safety and feasibility of neoadjuvant nivolumab followed by pulmonary resection. Additional surgical details were collected through chart review.

Results

Of 22 patients enrolled, 20 underwent resection. One was unresectable; another had small cell histologic subtype. There were no delays to surgical resection. Median (range) time from first treatment to surgery was 33 (17–43) days. There were 15 lobectomies, 2 pneumonectomies, 1 bilobectomy, 1 sleeve lobectomy, and 1 wedge resection. Of 13 procedures attempted via a video-assisted thoracoscopic surgery (VATS) or robotic approach, 7 (54%) required thoracotomy. Median (range) operative time was 228 (132–312) minutes; estimated blood loss was 100 (25–1000) mL; length of hospital stay was 4 (2–17) days. There was no operative mortality. Morbidity occurred in 10 of 20 patients (50%). The most common postoperative complication was atrial arrhythmia (6/20 [30%]). Major pathologic response was identified in 9 of 20 patients (45%).

Conclusion

Neoadjuvant therapy with nivolumab was not associated with unexpected perioperative morbidity or mortality. More than half of VATS/robotic cases were converted to thoracotomy, often due to hilar inflammation and fibrosis.

INTRODUCTION

Immune checkpoint inhibitors have rapidly become a pivotal treatment option for patients with advanced non-small cell lung cancer (NSCLC). These agents, which block immune inhibitory signals on the cancer cell surface, allow for recognition of tumor antigens by the patient’s T lymphocytes and ultimately result in destruction of tumor cells by the immune system. Initial studies of these drugs in patients with metastatic NSCLC and melanoma suggested that they are capable of producing significant and often durable treatment responses in subsets of patients with metastatic cancer.^1,2 Long-term outcomes from clinical trials of nivolumab in previously treated patients with advanced NSCLC demonstrated a 2-year overall survival of 23%–29% and 5-year overall survival of 16%.^1,3 Subsequent studies using these agents in combination with chemotherapy and radiation also suggested promising efficacy in patients with locally advanced NSCLC.⁴ Recent results of the PACIFIC trial showed a 12-month progression-free survival of 16.8 months in patients with unresectable stage III NSCLC treated with definitive chemoradiotherapy followed by durvalumab (anti–programmed death-ligand 1 [PD-L1]) versus 5.6 months in patients given placebo.⁴ The success of these agents in the metastatic setting has generated considerable enthusiasm among clinicians, and high-visibility marketing campaigns have resulted in rapidly expanding interest among patients.

On the basis of the success of these trials, numerous studies have been initiated to evaluate the utility of immunotherapeutic agents as neoadjuvant or adjuvant therapy in patients with surgically resectable NSCLC. The administration of these agents in the preoperative setting offers several theoretical advantages. Drug treatment while the tumor is in situ may provide immunologic benefits in terms of increased antigen load and may facilitate tumor recognition and drug efficacy.⁵ Similarly, correlative studies on posttreatment tumor specimens offer a unique opportunity to study underlying biologic and immunologic mechanisms and could yield important insights regarding treatment sensitivity and resistance. Despite these potential benefits, data regarding the safety and feasibility of surgical resection following treatment in this setting remain scarce. Initial retrospective case series in patients with previously unresectable or metastatic disease have suggested overall feasibility but have also raised concerns about technical challenges and perioperative issues unique to this patient population.^6,7

We recently completed a clinical trial investigating the safety and feasibility of the anti– programmed cell death protein 1 (PD-1) antibody nivolumab as a neoadjuvant treatment in patients with surgically resectable NSCLC. Data regarding tolerability and efficacy of this treatment approach have been recently reported.⁸ However, the unique mechanism of action of these agents and the theoretical challenges this may pose for subsequent lung resection warrants a more granular analysis of operative details and postoperative outcomes in this patient cohort. The aim of this study was to further explore the implications of preoperative nivolumab on subsequent lung resection in these patients, with particular attention given to technical details and postoperative outcomes.

METHODS

The study was approved by the institutional review boards at the two participating study sites, Johns Hopkins Hospital and Memorial Sloan Kettering Cancer Center, and all patients gave written informed consent. The clinical trial was registered at www.clinicaltrials.gov (NCT02259621). The trial details were previously published⁸ but are summarized here in brief. The trial was a single-arm study that involved two doses of nivolumab (3 mg/kg) administered intravenously 4 weeks and 2 weeks before surgery. Entry criteria included age ≥18 years with stage I to IIIA biopsy-proven NSCLC. Eligible patients were evaluated by a thoracic surgeon before enrollment, to confirm resectability. Nivolumab was the sole neoadjuvant therapy; patients were eligible for adjuvant therapy with chemotherapy or radiation as indicated. Patients were specifically excluded if they had a history of immunodeficiency, had active autoimmune or infectious disease, were receiving immunosuppressive therapy, or had a concurrent malignancy. All patients underwent a standard preoperative staging workup that included pretreatment tumor biopsy, contrast-enhanced computed tomography (CT) scan of the chest, positron emission tomography/CT scan, and invasive mediastinal nodal staging with endobronchial ultrasound or mediastinoscopy as indicated. Patients also underwent brain imaging with CT or MRI when clinically indicated and posttreatment CT before resection. Resection of the primary tumor and lymphadenectomy were performed according to institutional standards.

Patient demographic characteristics, pathologic data, and clinical outcomes were prospectively collected in accordance with the study protocol. To collect additional operative details, we performed a retrospective chart review on the study patients. Retrospectively abstracted data include operative approach, extent of resection, operative time, hospital length of stay, postoperative morbidity, rationale for conversion of MIS procedures to thoracotomy if necessary, and other details related to the operative experience.

Tumors were staged according the 7^th edition of the American Joint Committee on Cancer’s TNM staging criteria. Major pathologic response was defined as ≤10% of viable tumor cells remaining on postoperative pathologic review.⁹ Operative time was extracted from the operating room documentation and was defined as the duration from incision to wound closure. Surgical complications were defined according to the Society of Thoracic Surgeons database criteria.

RESULTS

Between August 2015 and October 2016, 22 patients were enrolled in the study and received neoadjuvant treatment with nivolumab. One patient was subsequently removed when histologic analysis of a preoperative biopsy demonstrated small cell carcinoma. Histologic subtypes for the remaining patients were adenocarcinoma (14/21 [67%]), squamous cell carcinoma (5/21 [24%]), pleomorphic carcinoma (1/21 [5%]), and adenosquamous carcinoma (1/21 [5%]). Preoperative clinical staging was as follows: 4 of 21 (19%) stage I, 10 of 21(48%) stage II, and 7 of 21 (33%) stage IIIA. Median age was 67 years (range, 55–84 years). Eleven patients (52%) were female; 18 (86%) were current or former smokers (Table 1). Twenty patients underwent resection. A single patient with a T2N2 squamous cell tumor developed a postobstructive pneumonia after a single dose of nivolumab and underwent uncomplicated resection at that time; this was the only grade 3 or higher adverse effect of nivolumab treatment. One patient with stage IIIA disease was deemed unresectable due to tracheal invasion noted on preoperative bronchoscopy, and surgery was aborted. The majority of patients (18/21 [86%]) had stable disease on posttreatment CT imaging. Two patients (10%) had a partial response, and 1 (5%) had progression of disease radiographically. There were no treatment-related delays to scheduled resections. Median time from the second dose of nivolumab to resection was 18 days (range, 11–29 days).

Table 1.

Characteristics of enrolled patients (N=21)

Characteristics
Age, years, median (range)	67 (55–84)
Gender
Female	11 (52)
Male	10 (48)
Clinical stage
IA T1N0	2 (10)
IB T2N0	2 (10)
IIA	5 (24)
T1bN1	1
T2aN1	4
IIB	5 (24)
T3N0	3
T2bN1	2
IIIA	7 (33)
T1N2	2
T2N2	2
T3N1	2
T4N1	1
Histologic subtype
Adenocarcinoma	14 (67)
Squamous cell	5 (24)
Adenosquamous	1 (5)
Pleomorphic	1 (5)
Smoking history
Never	3 (14)
Former/current	18 (86)
Extent of resection (n=20)
Lobectomy	15 (75)
Pneumonectomy	2 (10)
Wedge	1 (5)
Sleeve lobectomy	1 (5)
Bilobectomy	1 (5)
Approach (n=20
Thoracotomy	14 (70)
Thoracoscopy	3
Robotic	3

Unless otherwise indicated, data are no. (%).

Of the 20 patients undergoing resection, 15 (75%) underwent lobectomy, 2 (10%) underwent pneumonectomy, 1 (5%) each underwent sleeve lobectomy, bilobectomy, and wedge resection. Two of the lobectomies included en bloc wedge resection from an adjoining lobe. Fourteen of the 20 resections (70%) were performed via an open approach, of which 7 cases (35%) utilized thoracotomy as the initial strategy. Of the 13 cases that were initiated using a minimally invasive (video-assisted thoracoscopic surgery or robotic) approach, 7 (54%) were converted to thoracotomy.

Conversion to thoracotomy was performed in 1 of 4 (25%) patients with clinical stage I tumors. Operative reports note that this was due to dense, vascularized chest wall adhesions. Similarly, 1 of 2 clinical stage IIA cases were converted to thoracotomy. This patient had a postobstructive pneumonia preoperatively, and conversion was required due to dense adhesions in the fissure and between the left lower lobe and the aorta. The remainder of conversions occurred in patients with more-advanced stages of disease (5 of 7 [71%] stage IIB and IIIA cases). Also, 5 of 8 cases (62%) with either clinical or pathologic evidence of lymph node metastases that were attempted thoracoscopically required conversion. Operative reports from these cases note inflammation and/or dense adhesions in either the fissure or surrounding hilar and mediastinal nodal stations, which contributed to conversion. The presence of a major pathologic response (MPR) did not appear to be a factor in conversion to an open approach, as 4 of 6 cases with MPR (67%) that were attempted thoracoscopically were completed in this manner.

Perioperative outcome data are summarized in Table 2. Median surgical time was 228 minutes (range, 132–312 minutes). Median estimated blood loss was 100 mL (range, 25–1000 mL). No patient required blood transfusion. Median hospital length of stay was 4 days (range, 2–17 days). There was no operative mortality. A single patient died of traumatic injury on postoperative day 61. Ten of 20 patients (50%) experienced any postoperative morbidity. Atrial arrhythmia was the most common morbidity, occurring in 6 patients (30%). One of these patients suffered a concomitant myocardial infarction. A single patient developed pneumonia and a pulmonary embolism. Prolonged air leak and urinary retention each occurred in a single patient. A single patient who underwent lobectomy developed an empyema 4 months postoperatively requiring flap drainage. A single patient undergoing thoracoscopic lobectomy experienced an injury to the bronchus intermedius during dissection of the subcarinal space requiring conversion to thoracotomy for repair.

Table 2.

Perioperative outcomes of 20 patients undergoing pulmonary resection following neoadjuvant nivolumab

Perioperative Outcomes
Surgical time, minutes, median (range)	228 (132–312)
Estimated blood loss, mL, median (range)	100 (25–1000)
Length of stay, days, median (range)	4 (2–17)
Operative mortality	0a
Any morbidity
Atrial arrhythmia	6 (30)
Atrial arrhythmia	6 (30)
Pneumonia	1 (5)
Pulmonary embolism	1 (5)
Myocardial infarction	1 (5)
Prolonged air leak	1 (5)
Urinary retention	1 (5)
Empyema	1 (5)

^aOne patient died of traumatic injury on postoperative day 61.

On pathologic analysis, 9 of 20 patients (45%) had a major pathologic response (defined as ≤10% residual viable tumor)^8,9; 8 (40%) had pathologic downstaging. Two patients (pretreatment T1bN1 pleomorphic carcinoma and T3N1 squamous) had no residual viable tumor on final pathologic assessment. Currently, median follow-up is 20 months, with 18 of 20 patients (90%) alive at last follow-up. Three of 20 patients (15%) developed recurrent disease: 1 patient with solitary brain metastasis 2 months postresection treated with stereotactic radiotherapy, 1 patient with mediastinal nodal recurrence treated with definitive chemoradiation, and 1 patient with distant metastatic disease. One resected patient has died of NSCLC.

COMMENT

Several prospective randomized trials have demonstrated improved survival and tolerability with PD-1 and PD-L1 blockade in select patients with metastatic NSCLC, both in combination with platinum-doublet chemotherapy and as a single agent, compared with chemotherapy.^10–13 The improved disease-free and overall survival at 24 months with immunotherapy, compared with chemotherapy, in the metastatic setting has led to multiple clinical trials of immunotherapeutic agents in the neoadjuvant setting (Table 3). Evaluation of perioperative outcomes from these trials is necessary to address the feasibility and safety of pulmonary resection following neoadjuvant immunotherapy. These surgical outcomes are relevant to the planning and execution of future trials and surgical practice.

Table 3.

Select ongoing trials of neoadjuvant immunotherapeutic agents in patients with resectable NSCLC

Trial Identifier	Phase	Stage	Neoadjuvant Intervention	Primary Endpoint	Target Accrual
NCT02998528	3	IB-IIIA	Nivolumab + Ipilimumab vs Nivolumab + Platinum Doublet vs Platinum Doublet	MPR	624
NCT03158129	2	I-IIIA	Nivolumab +/− Ipilimumab	MPR	66
NCT02259621	2	IB-IIIA	Nivolumab +/− Ipilimumab	MPR	30
NCT02927301	2	IB-IIIA	Atezolizumab	MPR	180
NCT02572843	2	IIIA(N2)	Cisplatin/Docetaxel + Durvalumab	Event-free survival	68
NCT02818920	2	IB-IIIA	Pembrolizumab (neoadjuvant + adjuvant)	Surgical feasibility	32
NCT03237377	2	IIIA	Durvalumab + radiation	Safety	32

MPR, major pathologic response.

Chaft et al. first reported a series of 5 patients with advanced NSCLC who underwent pulmonary resection for persistent local disease after treatment with T-cell checkpoint inhibitors.⁷ This report suggested that pulmonary resection was feasible but cautioned that mediastinal and hilar fibrosis may develop as a result of response to treatment. More recently, Bott et al. reported a series of 19 patients who underwent pulmonary resection for metastatic disease or previously unresectable NSCLC and concluded that these resections, although challenging, were feasible, without undue morbidity.⁶ Finally, Yang et al. reported surgical outcomes from TOP1201, a phase II trial that evaluated the feasibility and safety of neoadjuvant chemotherapy plus ipilimumab (monoclonal antibody to CTLA-4) in patients with stage II-III NSCLC.¹⁴ Similarly, this study demonstrated that resection in these patients was safe and feasible, with perioperative outcomes similar to those in a cohort of patients who received neoadjuvant platinum-based chemotherapy.

The current study represents the largest experience to date of planned pulmonary resection following the neoadjuvant administration of immunotherapeutic agents in resectable NSCLC. Pulmonary resection in 20 patients with stage I-IIIA NSCLC following neoadjuvant treatment with 2 doses of nivolumab did not result in unexpected morbidity or mortality. There were no operative mortalities—a single patient died of a traumatic injury on postoperative day 61. This compares favorably to the 0% to 7% mortality in prior trials of neoadjuvant chemotherapy followed by pulmonary resection.^15–20 In addition, rates of perioperative morbidities were similar to those reported in previous multicenter trials.^15–20 Specifically, in comparison to the Southwest Oncology Group (SWOG) S9900 trial comparing chemotherapy followed by surgery with surgery alone in patients with stage IB-IIIA NSCLC, patients in the current study demonstrated favorable rates of pneumonia (5% vs 7%), reintubation (0% vs 7%), prolonged air leak (5% vs 9%), and respiratory failure (0% vs 7%). The rate of atrial arrhythmia was higher in this study than in SWOG S9900 (30% vs 16%).¹⁵ It is unclear whether this is due to the treatment strategy or instead is related to the relatively small size of the study.

The majority of resections (14/20) in this study were performed through an open approach. In 7 of 20 cases, this was the initial approach selected by the operating surgeon. The remainder were cases that were converted from an initial minimally invasive approach. As several of these cases involved sizeable tumors with hilar and/or mediastinal nodal disease, the necessity of conversion is not surprising. However, in most cases, this appeared to be related to inflammatory response and fibrosis at the primary tumor and involved nodal stations, presumably related to treatment response. Conversion to thoracotomy appeared to be more frequent in more-advanced stages of disease (25% of stage I cases, compared with 71% of stage IIB/IIIA cases). Furthermore, 62% of patients with node-positive tumors (either clinical or pathologic) who underwent an attempt at minimally invasive resection were converted to thoracotomy. Interestingly, the degree of pathologic response to treatment did not seem to be a significant factor, as the majority of MPR cases (67%) that were attempted thoracoscopically were completed without conversion.

Whereas some studies have suggested that minimally invasive lobectomy following neoadjuvant chemotherapy is feasible in patients with stage IIIA (N2) disease,^21–23 a recent analysis of the National Cancer Database by Krantz et al. demonstrated that lobectomy was accomplished by a minimally invasive approach in only 25.7% of such patients.²⁴ In the study by Bott et al., 4 of 11 anatomic resections (36%) were accomplished via a minimally invasive approach.⁶ Surgical approach was not reported in the operative results of the TOP1201 trial.¹⁴ Although the current study contains few patients, the necessity of thoracotomy does not seem to have affected morbidity and early mortality rates.

It is evident from this study that radiologic assessment of treatment response with CT may not be accurate following neoadjuvant immunotherapy. This is an important consideration as trials continue to accrue and for clinical practice. Major pathologic response was identified in 9 of 20 of patients (45%), whereas posttreatment CT most commonly demonstrated stable disease, with partial response in only 2 patients and progression in 1 patient. The 2 patients in the current study with complete pathologic response demonstrated stable disease radiographically. This phenomenon, oftentimes referred to as “pseudo-progression,”^25,26 has been previously documented and is likely related to T-cell infiltration and peri-tumoral inflammation during the early stages of treatment.

In addition to this study, two currently underway trials are evaluating combination nivolumab and ipilimumab in the neoadjuvant setting. CheckMate 816 (NCT02998528) is a multicenter randomized trial comparing neoadjuvant nivolumab + ipilimumab with nivolumab + platinum doublet or platinum doublet alone in patients with stage IB (>4 cm) to IIIA resectable NSCLC. The NEOSTAR trial (NCT03158129) compares the efficacy of the nivolumab/ipilimumab combination to nivolumab monotherapy in the neoadjuvant setting in patients with stage I-IIIA NSCLC. The rationale for the addition of ipilimumab is based on results from the CheckMate 012 study, which demonstrated improved progression-free survival, compared with nivolumab monotherapy, in patients with stage III-IV NSCLC.^27,28 Additional neoadjuvant trials are underway evaluating other immunotherapeutic agents, including the PD-L1 antibody atezolizumab in a multi-institutional trial through the Lung Cancer Mutation Consortium (NCT02927301), durvalumab (NCT02572843), and pembrolizumab alone or in combination with chemotherapy (Table 3).

As the use of T-cell checkpoint inhibitors in the treatment of NSCLC becomes more widespread in clinical practice and as trials of these drugs administered in the neoadjuvant setting accrue, it is likely that surgeons will be asked to perform pulmonary resections on increasing numbers of patients who have received immunotherapy. It is incumbent on surgeons to report surgical outcomes under this novel treatment strategy. This analysis reports the perioperative outcomes in the largest series to date of patients treated with neoadjuvant nivolumab for resectable NSCLC. Although pulmonary resection in this setting may be challenging, our experience suggests that these procedures can be accomplished safely with outcomes similar to historical outcomes after neoadjuvant chemotherapy. The relative effectiveness of this strategy in comparison to neoadjuvant platinum-based chemotherapy will hopefully be clarified by currently accruing clinical trials.

Supplementary Material

Video

Stephen R. Broderick discussing the implications of the study.

Central Picture:

Pre- (top) and post- (bottom) nivolumab. Imaging often underestimates pathologic response.

Central Message:

Pulmonary resection after neoadjuvant therapy with nivolumab did not result in undue morbidity or mortality. Major pathologic responses were identified despite stable disease radiographically.

Perspective Statement.

Immune checkpoint inhibitors have durable efficacy in advanced NSCLC, and their role in the neoadjuvant setting is under investigation. Pulmonary resection after neoadjuvant PD-1 blockade can be done without excessive treatment-related toxicity, operative morbidity, or mortality. Thoracic surgeons should play an active role in the planning and execution of clinical trials in this developing field.

Glossary of Abbreviations

CT

Computed tomography

MPR

Major pathologic response

NSCLC

Non-small cell lung cancer

PD-1

Programmed cell death protein 1

PD-L1

Programmed death-ligand 1VATS – Video-assisted thoracoscopic surgery