Prognostic factors following complete resection of non-superior sulcus lung cancer invading the chest wall

Abstract

OBJECTIVES

Locally advanced non-small-cell lung cancer (NSCLC) with chest wall invasion carries a high risk of recurrence and portends poor survival (30–40% and 20–50%, respectively). No studies have identified prognostic factors in patients who underwent R0 resection for non-superior sulcus NSCLC.

METHODS

A retrospective review was conducted for all chest wall resections for NSCLC from 2004 to 2018. Patients with superior sulcus tumours, partial (<1 rib) or incomplete (R1/R2) resection or distant metastasis were excluded. Disease-free survival (DFS) and overall survival (OS) were estimated using the Kaplan–Meier method. Cox proportional hazards modelling was used to determine factors associated with DFS and OS.

RESULTS

A total of 100 patients met inclusion criteria. Seventy-three (73%) patients underwent induction therapy, and all but 12 (16%) patients experienced a partial radiological response. A median of 3 ribs was resected (range 1–7), and 67 (67%) patients underwent chest wall reconstruction. The 5-year DFS and OS were 36% and 45%, respectively. Pathological N2 status [hazard ratio (HR) 3.12, confidence interval (CI) 1.56–6.25; P = 0.001], intraoperative blood transfusion (HR 2.24, CI 1.28–3.92; P = 0.005) and preoperative forced vital capacity (per % forced vital capacity, HR 0.97, CI 0.96–0.99; P = 0.013) were associated with DFS. Increasing pathological stage, lack of radiological response to induction therapy (HR 7.35, CI 2.35–22.99; P = 0.001) and cardiovascular comorbidity (HR 2.43, CI 1.36–4.36; P = 0.003) were associated with OS.

CONCLUSIONS

We demonstrate that blood transfusion and forced vital capacity are associated with DFS after R0 resection for non-superior sulcus NSCLC, while radiological response to induction therapy greatly influences OS. We confirm that pathological nodal status and pathological stage are reproducible determinants of DFS and OS, respectively.

INTRODUCTION

En bloc resection is recommended as the primary treatment modality for T3N0-1 or T4N0-1 non-small-cell lung cancer (NSCLC) invading the chest wall [1], which accounts for only 5% of newly diagnosed NSCLC in published surgical series [2]. Determinants of survival after resection are extent of nodal involvement and completeness of resection. In patients with T3 tumours, the 5-year overall survival (OS) of 61% has been reported for node-negative disease, compared with 18–21% for N2 disease [3]. A drastic survival advantage has also been noted for patients who underwent an R0 resection compared with patients who underwent an R1 or R2 resection. Our institution has previously reported 5-year OS as low as 4% in cases of incomplete resection, with minimal survival benefit over those who did not undergo resection [4].

Complete (R0) resection of locally advanced NSCLC has traditionally been defined by 2-cm margins [5, 6]. In patients with completely resected tumours, only nodal status has been consistently shown to be a predictor of survival [3, 7], though previous reports have postulated that sex, number of ribs resected, tumour size and degree of histological differentiation may also play a role [8, 9]. In addition, several studies have demonstrated an OS benefit in patients with tumours invading the superior sulcus/thoracic inlet who underwent induction chemoradiotherapy followed by surgery [7, 10, 11]. However, aside from a small series of 34 patients [12], no data supporting the use of induction therapy for non-superior sulcus tumours involving the chest wall have been reported and few studies have attempted to elucidate predictors of survival in such patients after a complete resection.

To address this knowledge gap, the primary objective of our study was to identify predictors of 5-year disease-free survival (DFS) and OS in patients with non-superior sulcus tumours involving the chest wall who underwent a complete resection. A secondary objective was to determine whether induction therapy, which is used almost routinely for superior sulcus tumours [7, 10, 11], is associated with survival after the chest wall resection of non-superior sulcus tumours.

METHODS

Patients and outcomes

Following institutional review board approval, we retrospectively reviewed a prospectively maintained database for all chest wall resections performed for NSCLC at a single institution between 2004 and 2018. Patients with tumours designated T3 or T4 by American Joint Committee on Cancer (AJCC) Cancer Staging Manual, 8th edition [13], and nodal status N0–N2 was included. Patients with superior sulcus tumours, those who underwent partial (<1 full rib) or incomplete (R1 or R2) resection and those with distant metastasis at diagnosis were excluded (Supplementary Material, Fig. S1).

Preoperative comorbidities were classified using the American Society of Anesthesiologists (ASA) physical status classification system. Cardiac comorbidities included hypertension, arrhythmia, coronary artery disease/myocardial infarction and congestive heart failure. Pulmonary comorbidities included asthma, chronic obstructive pulmonary disease, pulmonary hypertension, interstitial lung disease and tuberculosis. Radiological response to induction therapy was characterized according to Response Evaluation Criteria in Solid Tumors (RECIST) [14]. Lack of radiological response to induction therapy was defined as progressive or stable disease by RECIST criteria. Mortality during initial hospitalization and at 30 and 90 days post-surgery was reported, as were complications at these time points using the Clavien–Dindo classification system [15]. Readmissions within 30 and 90 days after discharge from initial hospitalization were also reported. Readmissions at other healthcare facilities, when noted in the medical record, were also captured.

Statistical analysis

The distributions of patient characteristics were summarized as number (proportions) or median (interquartile range) and compared between groups using Fisher’s exact test or Wilcoxon rank-sum test, respectively. Incidence of overall complications and major complications was compared between those with and without chest wall reconstruction using Fisher’s exact test. DFS was defined from the time of surgery to recurrence or death, and OS was defined from the time of surgery to death; patients were censored at the time of last follow-up otherwise. DFS and OS were estimated using the Kaplan–Meier method and compared between groups using the log-rank test. Median follow-up was calculated using the reverse Kaplan–Meier method [16]. Cox regression analysis was used to identify factors associated with DFS and OS using hazard ratios (HRs) and 95% confidence intervals (CIs). Multivariable models were constructed using backwards elimination approach starting with all factors with P < 0.1 on univariable analysis. All variables were completely observed except primary tumour standardized uptake value (N = 7, 7% missing data), FEV1 (N = 2, 2%), forced vital capacity (FVC) (N = 4, 4%) and diffusion capacity of the lungs for carbon monoxide (DLCO) (N = 5, 5%). There was no statistically significant evidence that the missing data violated the missing completely at random assumption under the significance level of 0.05 (P = 0.2), based on Little’s missing completely at random test [17]. We did not perform any imputation procedure for the missing data. Collinearity between factors such as FVC and cardiac comorbidity were assessed. The proportionality assumption was assessed using Schoenfeld residuals, and none of the factors were found to have violated the proportionality assumption. Analyses were conducted with Stata 15.0 (StataCorp, College Station, TX, USA) and R 3.5.1 (Vienna, Austria). Statistical tests were 2-sided, and a P-value <0.05 was considered statistically significant.

RESULTS

Patients and clinicopathological characteristics

A total of 100 R0 chest wall resections were performed for non-superior sulcus NSCLC from 2004 to 2018 (Table 1). The median follow-up was 5.5 years [interquartile range (IQR) 4.2–6.9 years]. Nearly all patients (n = 97; 97%) had a history of tobacco use, with a median of 40.0 pack-years. Only 27% (n = 27) were ASA class I or II. Preoperative pulmonary function (median, IQR) was as follows: FEV1, 81% (68–93%); FVC, 91% (79–103%); and DLCO, 65% (59–81%).

Table 1:

Patient, disease and treatment characteristics

Characteristics	Patients (N = 100), n (%) or median (IQR)
Age at surgery (years)	63 (56–70)
Sex
Female	45 (45)
Male	55 (55)
BMI	24.4 (21.3–28.0)
Smoking
Never	3 (3)
Ever	97 (97)
Pack-years (n = 97)	40.0 (25.0–54.5)
Tumour size (cm)a	5.8 (4.1–7.3)
Primary tumour SUV (n = 93)	15.0 (11.0–21.0)
Preoperative histology
Adenocarcinoma	38 (38)
Squamous cell carcinoma	31 (31)
Unspecified NSCLC	31 (31)
FEV1 (%) (n = 98)	81.0 (68.0–93.0)
FVC (%) (n = 96)	90.5 (79.0–102.5)
DLCO (%) (n = 95)	65.0 (59.0–81.0)
Induction therapy
None	27 (27)
Chemotherapy only	24 (24)
Chemoradiotherapy	49 (49)
Radiological response to induction therapy (n = 73)
Yes	61 (84)
No	12 (16)
ASA class
I or II	27 (27)
III or IV	73 (73)
Estimated blood loss (l)	0.4 (0.3–0.8)
Intraoperative transfusion
No	70 (70)
Yes	30 (30)
Surgery time (h)	5.3 (4.5–6.8)
Tumour location
Left	34 (34)
Right	66 (66)
Number of ribs resected
1–2	28 (28)
3–4	65 (65)
≥5	7 (7)
Location of ribs resected
Anterior/anterolateral	22 (22)
Posterior/posterolateral	67 (67)
Both	11 (11)
Concurrent vertebral body resection
No	78 (78)
Yes	22 (22)
Chest wall reconstruction
No	33 (33)
Yes	67 (67)
Final pathology
Adenocarcinoma	50 (50)
Squamous cell carcinoma	30 (30)
Unspecified NSCLC	6 (6)
No viable tumour	14 (14)
Pathological nodal status
N0	77 (77)
N1	9 (9)
N2	14 (14)
Pathological disease stageb
0	14 (14)
I	9 (9)
II	48 (48)
III	27 (27)
IV	2 (2)
Adjuvant therapy
None	71 (71)
Chemotherapy only	26 (26)
Radiotherapy only	3 (3)

^aComputed tomography.

^bAJCC 8th edition [13].

AJCC: American Joint Committee on Cancer; ASA: American Society of Anesthesiologists; BMI: body mass index; DLCO: diffusion capacity of the lungs for carbon monoxide; FEV1: forced expiratory volume per second; FVC: forced vital capacity; IQR: interquartile range; NSCLC: non-small-cell lung cancer; SUV: standardized uptake value.

Seventy-three (73%) patients underwent induction therapy, with the lack of radiological response (stable disease or progression of disease) in 12 (16%) patients. Induction chemotherapy regimens are listed in Supplementary Material, Table S1. At least 3 ribs were resected in 72 (72%) patients (Supplementary Material, Table S2). Vertebral body resection was performed concurrently with rib resections in 22 (22%) patients. Reconstruction of the resulting chest wall defect was performed in 67 (67%) patients: anterior or anterolateral, 17 (25%) patients; posterior or posterolateral, 42 (63%) patients; and both anterior/anterolateral and posterior/posterolateral, 8 (12%) patients. Reconstruction details are listed in Supplementary Material, Table S3. Of the 33 patients who did not undergo reconstruction, 18 patients had fewer than 3 ribs resected and 15 patients with at least 3 ribs resected had high posterior defects not requiring reconstruction. Thirty (30%) patients required blood transfusion during the operation, with a median blood loss of 850 ml in these patients. Twenty-seven of the 30 (90%) patients who received a transfusion underwent induction therapy. In addition, nearly one-third (30%, n = 9/30) of transfusions were performed in patients who underwent vertebral body resection, with a median blood loss of 625 ml. On final pathology, half of tumours were adenocarcinoma (n = 50; 50%), while 14 (14%) patients had a complete pathological response to induction therapy. Of these 14 patients, only 5 (36%) patients died (with only one cancer-related death) and 4 (80%) of those 5 deaths occurred >5 years after surgery. Interestingly, all 14 patients who experienced a complete pathological response received induction chemoradiotherapy, though this did not translate to improved DFS or OS (Supplementary Material, Fig. S2). Finally, 29% of patients (n = 29) underwent adjuvant therapy.

Postoperative outcomes

The median length of hospital stay was 7 days (IQR 5–11 days), with 44 (44%) patients experiencing an in-hospital complication; most (n = 28; 64%) were classified as minor (grade I or II) (Table 2). In comparing those who underwent chest wall reconstruction to those who did not, rates of overall complications (48% vs 36%, P = 0.29) and major complications (grade III–V) (41% vs 25%, P = 0.49) did not significantly differ. Thirteen (13%) patients required intensive care unit (ICU) admission during their postoperative course. Preoperative FEV1 and FVC values in patients who experienced ICU admission were significantly lower than in those who did not (Table 3). There were 2 (2%) patients who died during initial hospitalization, with no intraoperative deaths. At 30 days post-surgery, an additional 1 (1%) patient died, and from 31 to 90 days post-surgery, an additional 3 (3%) patients died, yielding a 90-day mortality of 6.0% (n = 6). Causes of death for these 6 patients are listed in Supplementary Material, Table S4. Twelve (13%) patients experienced readmission within 30 days of initial hospital discharge, and an additional 9 (9%) patients were readmitted between 31 and 90 days. Nearly half (4/9, 44%) of readmissions between 31 and 90 days were for the management of neurological symptoms secondary to brain metastases. The median length of stay after readmission was 6 days (range 3–113 days). Causes of readmission for these 21 patients are listed in Supplementary Material, Table S5. Thirty-two (32%) patients experienced a recurrence (12 local), with 23 (44%) disease-related deaths. Fifty-two percent of the cohort in total (n = 52) died by the end of the study period. The 5-year DFS and OS were 36% (95% CI 34–56%) and 45% (95% CI 25–47%), respectively, for all pathological stages.

Table 2:

Outcomes

Outcomes	Patients (N = 100), n (%) or median (IQR)
Length of stay (days)	7.0 (5.0–11.0)
ICU admission
No	87 (87)
Yes	13 (13)
Invasive ventilation
No	90 (90)
Yes	10 (10)
Complication
In hospital	44 (44)
Grades I and II	28 (28)
Grades III–V	16 (16)
Discharge 30 days (n = 93)	9 (9.7)
Grades I and II	1 (1.1)
Grades III–V	8 (8.6)
31–90 days (n = 97)	8 (8.2)
Grades I and II	2 (2.1)
Grades III–V	6 (6.2)
Readmission
Discharge 30 days (n = 93)	12 (13)
31–90 days (n = 97)	9 (9.3)
Recurrence	32
Locoregional	12 (12)
Distant	18 (18)
Locoregional and distant	2 (2)
Death	52
In hospital	2 (2.0)
Discharge 30 days (n = 98)	1 (1.0)
31–90 days (n = 97)	3 (3.1)
>90 days (n = 94)	46 (49)

ICU: intensive care unit; IQR: interquartile range.

Table 3:

Preoperative pulmonary function in relation to postoperative ICU admission

Test	ICU admission (n = 13)	No ICU admission (n = 87)	P-value
FEV1 (%)	68 (60–83)	82 (70–95)	0.013
FVC (%)	83 (65–92)	92 (81–104)	0.014
DLCO (%)	63 (55–79)	68 (60–81)	0.231

Values are presented as median (IQR) and assessed between ICU admission status using Wilcoxon rank-sum test.

DLCO: diffusion capacity of the lungs for carbon monoxide; FEV1: forced expiratory volume per second; FVC: forced vital capacity; ICU: intensive care unit; IQR: interquartile range.

Factors associated with survival

The results of univariable analyses for DFS and OS are listed in Supplementary Material, Tables S6 and S7. A total of 96 and 100 patients were included in the multivariable models for DFS and OS, respectively. On multivariable analysis for DFS, increasing FVC (per %FVC, HR 0.97, 95% CI 0.96–0.99; P = 0.013), intraoperative blood transfusion (HR 2.24, 95% CI 1.28–3.92; P = 0.005) and pathological N2 disease (HR 3.12, 95% CI 1.56–6.25; P = 0.001) were significantly associated with DFS (Table 4). In addition to analysing FVC as a continuous variable in our Cox regression, we investigated its association with DFS and OS after dividing patients into 2 groups: one with FVC above the median of 90.5% and one with FVC below the median. Patients with FVC above the median had a statistically significant survival advantage, with the 5-year DFS and OS of 47% and 57%, respectively, compared with 24% and 32% for those with FVC below the median (Fig. 1). Furthermore, not only was receipt of blood transfusion associated with worse survival (Fig. 2), receipt of even 1 U of blood was significantly associated with worse DFS (HR 2.80, 95% CI 1.40–5.61; P = 0.004).

Table 4:

Multivariable Cox model for DFS (n = 96) and OS (N = 100)

Factor	Univariable analysis		Multivariable analysis
	HR (95% CI)	P-value	HR (95% CI)	P-value
DFS (n = 96)
Increasing FVC (per %)	0.97 (0.96–0.99)	0.008	0.97 (0.96–0.99)	0.013
Intraoperative transfusion	2.48 (1.44–4.27)	<0.001	2.24 (1.28–3.92)	0.005
Pathological nodal status
N0	Reference
N1	1.94 (0.86–4.36)	0.11	1.97 (0.87–4.49)	0.10
N2	3.31 (1.68–6.53)	0.001	3.12 (1.56–6.25)	0.001
OS (N = 100)
Cardiac comorbidity	1.99 (1.15–3.46)	0.014	2.43 (1.36–4.36)	0.003
Lack of radiological response to induction therapya	2.26 (0.89–5.72)	0.085	7.35 (2.35–22.99)	0.001
Pathological stageb
0	Reference
I	1.64 (0.43–6.19)	0.5	2.66 (0.67–10.63)	0.2
II	1.65 (0.63–4.32)	0.3	3.58 (1.16–11.07)	0.027
III/IV	2.61 (0.96–7.07)	0.059	5.17 (1.65–16.20)	0.005

^a n = 12 patients experienced a lack of response by RECIST criteria (n = 7 with stable disease and n = 5 with progression of disease), while n = 61 patients experienced a partial response.

^bAJCC 8th edition [13].

AJCC: American Joint Committee on Cancer; CI: confidence interval; DFS: disease-free survival; FVC: forced vital capacity; HR: hazard ratio; OS: overall survival; RECIST: Response Evaluation Criteria in Solid Tumors.

Figure 1:

Five-year (A) disease-free survival and (B) overall survival in patients with preoperative FVC above or below the median of 90.5%. FVC: forced vital capacity.

Figure 2:

Five-year (A) disease-free survival and (B) overall survival in patients who received blood transfusion and patients who did not.

Multivariable analysis for OS revealed that lack of radiological response to induction therapy (defined as stable or progressive disease by RECIST criteria) was strongly associated with worse OS (HR 7.35, 95% CI 2.35–22.99; P = 0.001). Patients who had progression of disease despite induction therapy had significantly worse OS than those with stable disease or a partial response, with the 2-year OS of 20%, compared with 68% and 74%, respectively (log-rank P = 0.015, Supplementary Material, Fig. S3). Cardiac comorbidity and increasing pathological stage were also associated with worse OS, as shown in Table 4.

DISCUSSION

Our study identified several important predictors of survival in patients who underwent an R0 resection for non-superior sulcus NSCLC invading the chest wall. Higher preoperative FVC was associated with improved DFS, whereas receipt of intraoperative blood transfusion was associated with worse DFS. We reaffirmed that pathological N2 disease was the factor most strongly associated with DFS. In addition to the presence of cardiac comorbidity and increasing pathological stage, we discovered that lack of radiological response to induction therapy was associated with a substantially higher risk of poor OS.

Perioperative blood transfusion has been linked to poor outcomes after lung resection in numerous studies [18, 19]. A recent report from our institution of over 4800 patients who underwent anatomic pulmonary resection for NSCLC showed that receipt of >1 U of blood perioperatively was associated with greater risk of recurrence and worse DFS and OS [20]. In our study, receipt of even 1 U of blood was associated with worse DFS, which underscores the importance of judicious use of blood transfusion in patients undergoing pulmonary resection, especially for locally advanced tumours requiring chest wall resection. Interestingly, the transfusion rate in the first half of the study period (31%, 2004–2011) was higher than in the second half (28%, 2011–2018), despite a median blood loss of 600 vs 1200 ml in these cases, respectively. In addition, 27 of the 30 (90%) patients who received a transfusion underwent induction therapy and it is likely that transfusions were also given more liberally in these patients. As such, we caution against reflex transfusion in patients with low haemoglobin/haematocrit values given haemodynamic stability during the operation. Although a standardized transfusion protocol was not employed over the study period, in general, if the intraoperative haemoglobin level is noted to drop below a certain threshold (usually <7–8 mg/dl in patients without pre-existing cardiac disease) or there is bleeding at such a rate that causes haemodynamic instability, transfusions may be initiated.

Preoperative pulmonary function testing is a prerequisite for operative planning in cases of chest wall resection. The impact of predicted postoperative DLCO on outcomes after lung resection, especially in patients who received induction therapy, has previously been established [21]. However, DLCO is the representative of pulmonary vascular and alveolar properties, and though FEV1 and FVC may also be affected in parenchymal lung disease, the latter 2 are also representative of chest wall mechanics and are of greater relevance in cases of chest wall resection. A study by Liu et al. [22] found that postoperative FVC was significantly lower in propensity-matched patients who underwent concomitant parenchymal and chest wall resection for NSCLC than in patients who underwent only parenchymal resection, while FEV1 and DLCO did not differ between the 2 groups. Impaired chest wall stability, loss of intercostal muscle function, and chest wall fibrosis after chest wall resection may simulate a restrictive chest wall process, thereby altering FEV1 and FVC. Despite this, to our knowledge, no prior studies have shown higher FVC to be an independent predictor of improved survival after chest wall resection. Though FVC was treated as a continuous variable for regression analysis, FVC below the median value of 90.5% associated with both worse DFS and OS on Kaplan–Meier analysis. Furthermore, patients who experienced ICU admission postoperatively had significantly lower preoperative FVC than those who did not (median 83, IQR 65–92 vs median 92, IQR 81–104, P = 0.014). However, increasing preoperative FVC was only an independent predictor of DFS, not OS, possibly due to the 26 non-recurrence-related DFS events (total 58 DFS events). Collectively, these findings indicate that FVC, rather than FEV1 or DLCO, may be of greatest relevance in chest wall surgery and potentially identifies a subset of patients at higher risk of postoperative morbidity and mortality.

Perhaps the most striking finding of our study is the degree to which radiological response to induction therapy was associated with OS. Several prior studies have demonstrated an OS benefit in patients with tumours invading the superior sulcus/thoracic inlet who underwent induction chemoradiotherapy followed by surgery, achieving the 5-year OS rates of 44–56% [10, 11] compared with 25–38% in patients who did not receive induction chemoradiotherapy [7]. Aside from a recent small series of 34 patients [12], to our knowledge, no prior studies have demonstrated a survival benefit considering only patients with non-superior sulcus NSCLC. Prior studies have also routinely excluded patients who had radiological progression of disease despite induction therapy. Of the 73 patients in our study who received induction therapy, 5 (7%) patients experienced radiological progression of disease and 7 (10%) patients had stable disease. Of the 5 patients with the progression of disease, 4 (80%) patients died within 7 months of surgery. The survival advantage in patients receiving induction therapy is historically most pronounced in cases of complete pathological response, with the 5-year OS of 70–90% compared with 40–50% in patients with residual disease [10, 23]. In our study, only 5 (36%) of the 14 patients who experienced a complete pathological response to induction therapy died (with only one cancer-related death) and 4 (80%) of those 5 deaths occurred >5 years after surgery. Taken together, this suggests that response to induction therapy, rather than simply receipt of induction therapy (as is the case for patients with NSCLC invading the superior sulcus), greatly influences survival in patients who undergo chest wall resection for non-superior sulcus NSCLC. In patients without a radiological response to induction therapy, additional adjuvant therapy should be considered, as surgery alone may not provide a long-term benefit. It is also worth noting that although adjuvant therapy was not shown to be associated with DFS or OS in our study, the low proportion of patients who received adjuvant therapy (29%) likely prohibits detection of this association. A subgroup analysis of only node-positive (N1 or N2) patients was attempted but excluded from final analysis, as there were only 10 patients who received adjuvant therapy for node-positive disease.

Limitations

The strengths of our study include a long follow-up period (median 5.5 years) and a focus strictly on patients with non-superior sulcus tumours. A limitation of the study is that specific induction regimens were not standardized, thus complicating the assessment of response to induction therapy. In addition, while the current study was limited to those who underwent complete surgical resection, in future studies, the authors intend to compare the effectiveness of induction therapy in all cases of non-superior sulcus NSCLC invading the chest wall, including cases where resection was not performed and those who underwent incomplete resection. Finally, though the majority of our cohort received induction therapy, the proportion of patients who did not experience a radiological response to therapy was quite low (16%), limiting the generalizability of our results.

CONCLUSIONS

In patients who undergo complete resection of non-superior sulcus NSCLC invading the chest wall, we demonstrate that preoperative FVC and intraoperative blood transfusion are major contributors to DFS, while the presence of cardiac comorbidity was significantly associated with OS. Characterization of radiological response to induction therapy may also be useful in identifying patients who are at increased risk of mortality and who may benefit from adjuvant therapy.

Funding

This work was supported in part by the National Cancer Institute [P30 CA008748].

Conflict of interest: Valerie Rusch has received commercial research grants from Genentech and other remuneration from Intuitive Surgical. James M. Isbell has stock ownership in LumaCyte, LLC. Bernard J. Park has served as a proctor for Intuitive Surgical and consultant for COTA. David R. Jones serves as a senior medical advisor for Diffusion Pharmaceuticals and a consultant for Merck and AstraZeneca. Gaetano Rocco has financial relationships with Baxter, Scanlan, and Medtronic.

Author contributions

Gregory D. Jones: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Writing—original draft; Writing—review & editing. Raul Caso: Data curation; Formal analysis; Investigation; Writing—review & editing. Jae Seong No: Data curation; Formal analysis; Investigation; Methodology; Writing—original draft; Writing—review & editing. Kay See Tan: Formal analysis; Project administration; Validation; Writing—review & editing. Joseph Dycoco: Formal analysis; Investigation; Validation; Writing—review & editing. Manjit S. Bains: Formal analysis; Investigation; Methodology; Writing—review & editing. Valerie W. Rusch: Formal analysis; Investigation; Methodology; Writing—review & editing. James Huang: Investigation; Writing—review & editing. James M. Isbell: Investigation; Writing—review & editing. Daniela Molena: Investigation; Writing—review & editing. Bernard J. Park: Investigation; Writing—review & editing. David R. Jones: Investigation; Writing—review & editing. Gaetano Rocco: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Writing—review & editing.

ABBREVIATIONS

AJCC

American Joint Committee on Cancer

ASA

American Society of Anesthesiologists

CI

Confidence interval

DFS

Disease-free survival

FVC

Forced vital capacity

HR

Hazard ratio

ICU

Intensive care unit

IQR

Interquartile range

NSCLC

Non-small-cell lung cancer

OS

Overall survival

RECIST

Response Evaluation Criteria in Solid Tumors