Abstract
Objective:
Pretreatment predicted postoperative diffusing capacity of the lung for carbon monoxide (DLCO) has been associated with operative mortality in patients receiving induction therapy for resectable non-small cell lung cancer (NSCLC). It is unknown whether a reduction in pulmonary function after induction therapy and before surgery affects the risk of morbidity or mortality. We sought to determine the relationship between induction therapy and perioperative outcomes as a function of postinduction pulmonary status in patients undergoing surgical resection for NSCLC.
Methods:
We retrospectively reviewed 1001 patients with pathologic stage I-III NSCLC who received induction therapy before lung resection. Pulmonary function was defined according to American College of Surgeons Oncology Group major criteria: DLCO ≥50% = normal; DLCO <50% = impaired. Patients were categorized into five subgroups according to combined pre- and postinduction DLCO status: normal-normal, normal-impaired, impaired-normal, impaired-impaired, and preinduction only (without postinduction pulmonary function test measurements). Multivariable logistic regression was used to quantify the relationship between DLCO categories and dichotomous endpoints.
Results:
On multivariable analysis, normal-impaired DLCO status was associated with an increased risk of respiratory complications (odds ratio [OR], 2.29 [95% confidence interval {CI}, 1.12-4.49]; p=0.019) and in-hospital complications (OR, 2.83 [95% CI, 1.55-5.26]; p<0.001). Type of neoadjuvant therapy was not associated with an increased risk of complications, compared with conventional chemotherapy.
Conclusions:
Reduced postinduction DLCO may predict perioperative outcomes. The use of repeat pulmonary function testing may identify patients at higher risk of morbidity or mortality.
Keywords: Diffusing capacity of the lung for carbon monoxide, DLCO, Non-small cell lung cancer, pulmonary function testing
INTRODUCTION
Induction chemotherapy with or without radiotherapy has become a mainstay in the treatment of locally advanced non-small cell lung cancer (NSCLC), with improved 5-year survival over surgical resection alone.1-3 Trials aimed at demonstrating the safety and feasibility of induction immunotherapy, which is generally well tolerated, are ongoing.4 However, early results on operative risk and short-term outcomes are promising.5, 6 All induction agents, regardless of their mechanism of action, can potentially cause undue damage to the healthy lung parenchyma surrounding the tumor.7, 8 Studies have shown that select surgical patients treated with induction therapy—in particular, those undergoing right pneumonectomy—have a higher risk of perioperative mortality.9, 10
Preoperative performance tests, including pulmonary function tests (PFTs), can play an important role in predicting perioperative morbidity and mortality.11-13 The American College of Chest Physicians (ACCP) recommends a careful preoperative physiologic assessment using spirometry and diffusion-capacity testing to evaluate the operability of the patient before anatomical resection.14 In fact, profound reductions in forced vital capacity, forced expiratory volume in 1 second (FEV1), and diffusing capacity of the lung for carbon monoxide (DLCO) are major determinants of morbidity and mortality among patients undergoing thoracic surgery.15, 16 In this setting, the literature consistently indicates that standard chemotherapy with or without radiotherapy can have a negative effect on these parameters—in particular, DLCO.17, 18
Previous work from our institution has shown that pretreatment predicted postoperative (ppo) DLCO is associated with operative mortality in patients who receive induction therapy for resectable NSCLC.19 However, it is unknown whether a reduction in pulmonary function after induction therapy and before surgery affects a patient’s risk of morbidity or mortality. To assess the relationship between induction therapy and postoperative outcomes as a function of preoperative pulmonary status, we analyzed PFT results before and after induction therapy in patients undergoing surgery for NSCLC. The primary endpoint of this study was postoperative respiratory complications. Secondary endpoints were in-hospital complications, major complications, length of stay, and 30-day mortality.
METHODS
Patients
After approval from our institutional review board, we queried our prospectively maintained database to identify all eligible patients treated from 2000 to 2019. All patients gave informed consent for participation in the approved study protocol and for publication of the study data. Eligible patients had pathologic stage I-III NSCLC or no viable tumor at the time of surgery as a result of induction therapy (p0), received induction therapy, had complete results of pre- and postinduction PFTs, and underwent an oncologic resection. Patients who received only postinduction PFTs were treated as a separate treatment group (Supplementary Table 1 and 2). Within this final patient cohort, two groups of patients were identified on the basis of the type of induction treatment they received: conventional systemic chemotherapy and immunotherapy (immune checkpoint inhibitor [ICI] therapy). The conventional systemic chemotherapy group was further categorized into three groups on the basis of the number of treatment cycles received: abbreviated (<4 cycles), full (≥4 cycles), and unknown (receipt of systemic chemotherapy but no reported cycle duration). Although the number of cycles needed to reach a clinically complete regimen course varies by patient, the 4-cycle cutoff was based on the 2015 American Society of Clinical Oncology recommendations for treatment of advanced NSCLC.20 Patients in the ICI group received two therapeutic doses of induction ICI therapy, with 39 of 46 patients receiving therapy through a clinical trial without standard chemotherapy (Figure 1).
Figure 1.
CONSORT diagram. ICI, immune checkpoint inhibitor; NSCLC, non-small cell lung cancer; PFT, pulmonary function test.
PFTs
Five hundred nineteen patients had pre- and postinduction PFTs performed (forced vital capacity, FEV1, and DLCO). The American College of Surgeons Oncology Group (ACOSOG) major criteria were used to define clinical cutoffs for pulmonary function, with preoperative FEV1 and DLCO <50% indicating impaired status21, 22 and ≥50% indicating normal status. We categorized the changes from pre- to postinduction FEV1 and DLCO using these definitions. Using this categorization model, we arranged patients into four distinct subgroups according to their pulmonary function status before and after induction therapy: normal-normal, normal-impaired, impaired-normal, and impaired-impaired. Patients with only preinduction PFTs were included as a reference group. A sensitivity analysis was performed to verify our findings, using ACOSOG minor criteria, where DLCO <60% indicates impaired pulmonary status, again categorized into the four aforementioned groups.
Endpoints
The primary endpoint of this study was postoperative respiratory complications. A postoperative respiratory complication was defined as an in-hospital or postdischarge adverse event related to the pulmonary system within 30 days after surgery. Secondary endpoints were in-hospital complications, major complications, and 30-day mortality. Major complications were defined as grade ≥3 complications according to the Clavien-Dindo system.23 A complete list of adverse events reported in this cohort can be found in Supplementary Table 3.
Statistical analysis
Summary statistics were used to assess differences in clinicopathologic characteristics and clinical outcomes by the four induction therapy groups (full, abbreviated, unknown, and ICI). Analysis of covariance (ANCOVA) was initially performed to determine the association between induction group and postinduction PFT measures after controlling for preinduction PFT measures, which included post hoc analysis with a false discovery rate adjustment. The ANCOVA post hoc analysis used the estimated marginal means, which reduced the bias of imbalances in the data. Postinduction DLCO score was adjusted by preinduction DLCO score for each induction group; all observations were weighted, with less weight given to outliers.
Postoperative outcomes were compared between categories of pre- and postinduction PFTs using the Wilcoxon rank sum test for continuous outcomes and the Chi-squared test for dichotomous outcomes. To quantify the relationships between factors and outcomes, univariable logistic regression models were generated for the outcomes of interest (respiratory complications, in-hospital complications, and major complications), with consideration of clinical covariates and the categories of pre- and postinduction pulmonary function status. Multivariable models were developed on the basis of a backward selection process starting with factors with p<0.1 from univariable analyses. PFT status group, induction treatment type, and procedure type were prespecified to be included in the multivariable models regardless of statistical significance. Thirty-day mortality did not have enough events for the construction of a multivariable model. Length of stay (LOS) was assessed using multivariable linear regression with a log transformation of the outcome to account for skewness.
Statistical significance was defined using a 2-sided test with an alpha of 0.05. Analyses were conducted using R version 3.6.2. We used the cor.test, aov, glm, and lm functions from the stats package, the Anova function from the car package, the emmeans_test function from the emmeans package.
RESULTS
Patient characteristics
In total, 1001 patients were included in the study; the majority were women (n=543 [55%]), and the median age for all patients was 65 years (range, 30-88 years). Half of all patients (n=497 [50%]) had a preexisting cardiac comorbidity, including valvular disease, hypertension, coronary artery disease, and atrial fibrillation. Ninety percent of patients (n=875) were ever-smokers, and one-third (n=304 [30%]) had preexisting chronic obstructive pulmonary disease. Most patients had clinical stage II or III disease (n=908 [91%]) (Table 1).
Table 1.
Summary of clinical and pathologic characteristics by treatment group
| Characteristic | Full (N=453) | Abbreviated (N=364) | ICI (N=46) | Unknown (N=138) | P valuea |
|---|---|---|---|---|---|
| R1/R2 resection margin | 44 (9.7) | 22 (6.0) | 4 (8.7) | 7 (5.1) | 0.14 |
| Age at time of surgery, years | 66 (58-71) | 65 (58-72) | 68 (59-73) | 65 (58-70) | 0.6 |
| BMI | 26.9 (23.9-30.4) | 26.9 (24.4-30.8) | 28.2 (24.7-32.7) | 25.3 (22.9-29.1) | 0.008 |
| Procedure type | 0.7 | ||||
| Lobectomy/bilobectomy | 360 (79) | 300 (82) | 38 (83) | 110 (80) | |
| Pneumonectomy | 54 (12) | 43 (12) | 5 (11) | 14 (10) | |
| Sublobar resection | 39 (8.6) | 21 (5.8) | 3 (6.5) | 14 (10) | |
| Pulmonary comorbidity | 139 (31) | 104 (29) | 16 (35) | 40 (29) | 0.8 |
| Smoking history | 0.2 | ||||
| Ever | 402 (89) | 314 (86) | 43 (93) | 115 (83) | |
| Never | 51 (11) | 50 (14) | 3 (6.5) | 23 (17) | |
| Cardiac comorbidity | 193 (43) | 163 (45) | 28 (61) | 64 (46) | 0.12 |
| FEV1 preinduction | 85 (74-98) | 85 (73-100) | 88 (74-99) | 83 (70-100) | 0.7 |
| Unknown | 1 | 3 | 1 | 0 | |
| DLCO preinduction | 77 (64-92) | 77 (62-93) | 82 (68-92) | 75 (63-89) | 0.6 |
| Unknown | 28 | 30 | 1 | 13 | |
| Induction | |||||
| Yes | 453 (100) | 364 (100) | 46 (100) | 138 (100) | |
| FEV1 postinduction | 82 (71-95) | 82 (69-93) | 91 (81-98) | 84 (70-95) | 0.037 |
| Unknown | 208 | 170 | 6 | 42 | |
| DLCO postinduction | 62 (54-78) | 66 (56-76) | 80 (69-90) | 67 (56-79) | <0.001 |
| Unknown | 214 | 170 | 6 | 42 | |
| Approach type | <0.001 | ||||
| Open | 382 (84) | 321 (88) | 22 (48) | 125 (91) | |
| VATS | 71 (16) | 43 (12) | 24 (52) | 13 (9.4) | |
| Pathologic stage | <0.001 | ||||
| 0 | 24 (5.3) | 17 (4.7) | 4 (8.7) | 18 (13) | |
| I | 105 (23) | 100 (27) | 7 (15) | 30 (22) | |
| II | 102 (23) | 93 (26) | 9 (20) | 46 (33) | |
| III | 222 (49) | 154 (42) | 26 (57) | 44 (32) | |
| Surgery year | <0.001 | ||||
| 2000-2004 | 47 (10) | 113 (31) | 0 (0) | 35 (25) | |
| 2005-2009 | 113 (25) | 141 (39) | 0 (0) | 38 (28) | |
| 2010-2014 | 154 (34) | 69 (19) | 1 (2.2) | 41 (30) | |
| 2015-2019 | 139 (31) | 41 (11) | 45 (98) | 24 (17) |
Data are median (interquartile range) or no. (%). DLCO, diffusing capacity of the lung for carbon monoxide; FEV1, forced expiratory volume in 1 second; ICI, immune checkpoint inhibitor; R1, microscopic disease; R2, macroscopic disease; VATS, video-assisted thoracoscopic surgery.
Statistical tests performed: Pearson’s Chi-squared test; Kruskal-Wallis rank sum test.
Patient treatment regimen features
Overall, 753 patients (75%) underwent lobectomy, 71 (7%) underwent sublobar resection, 56 (6%) underwent bilobectomy, and 116 (12%) underwent pneumonectomy. An open approach (thoracotomy) was used in the majority of patients (n=850 [85%]) (Table 1). Most patients (n=952) received a conventional induction treatment regimen—either full, abbreviated, or unknown. Almost all conventional chemotherapy regimens comprised platinum-based chemotherapy (n=855 [90%]) plus a second chemotherapeutic agent (n=853 [90%]), such as etoposide, gemcitabine, pemetrexed, or taxane. Sixty-one patients (6%) received a third agent, including tyrosine kinase inhibitors, vascular endothelial growth factor inhibitors, vinca alkaloids, topoisomerase inhibitors, mitomycin C, and 5-fluorouracil. There were a few exceptions to the conventional regimens: 22 patients (2%) received only 1 of the secondary agents, and 52 patients (5%) received only a tyrosine kinase inhibitor. Forty-six patients (4%) received induction ICI therapy, including atezolizumab (n=33), nivolumab (n=9), and the combination of ipilimumab and nivolumab (n=4) (Figure 1). Furthermore, 114 patients (11%) received induction chemoradiation (median dose, 5040 cGy [interquartile range {IQR}, 4500-5790 cGy]). Of these 114 patients, 106 (93%) received concurrent therapy, 5 (6%) received sequential therapy, and 2 (1%) had no reported timing of treatment administration.
Characteristics and outcomes by induction therapy group
Several baseline characteristics differed between the ICI group and the conventional systemic chemotherapy groups. Patients in the ICI group had an overall higher pathologic stage (p<0.001) and were more likely to undergo a minimally invasive procedure performed between 2015 and 2019 (p<0.001) (Table 1). Length of stay was shorter in the ICI group (median, 4 days [IQR, 3-6 days]; p<0.001) than in the conventional chemotherapy groups. Patients in the abbreviated chemotherapy group had a 38% longer LOS (effect estimate [Exp{β}], 1.38 [95% confidence interval {CI}, 1.20-1.60]; p<0.001), and those in the full group had a 26% longer LOS (Exp[β], 1.26 [95% CI, 1.09-1.46]; p=0.002), compared with patients in the ICI group (Table 2, Supplementary Table 4). All other outcomes did not significantly differ by induction therapy group (Table 2).
Table 2.
Comparison of clinical outcomes by treatment group
| Characteristic | Full (N=453) | Abbreviated (N=364) |
ICI (N=46) | Unknown (N=138) |
P valuea |
|---|---|---|---|---|---|
| LOS | 5.0 (4.0-7.0) | 6.0 (4.0-8.0) | 4.0 (3.0-5.8) | 5.5 (4.0-8.0) | <0.001 |
| Respiratory complication | 77 (17) | 66 (18) | 11 (24) | 26 (19) | 0.7 |
| In-hospital complication | 160 (35) | 131 (36) | 19 (41) | 49 (36) | 0.9 |
| Major complicationb | 37 (8.2) | 31 (8.5) | 6 (13) | 7 (5.1) | 0.3 |
| 30-day mortality | 8 (1.8) | 8 (2.2) | 1 (2.2) | 1 (0.7) | 0.7 |
Morbidity and mortality in the overall cohort
Total perioperative morbidity was 32%, with the majority of cases related to a respiratory event. Overall, 17 patients (2%) died within 30 days of surgery (13 deaths occurred during the index procedure in-hospital course; 12 were attributable to respiratory complications). Of the 116 patients who underwent a pneumonectomy, 2 (2%) died within 30 days of surgery.
PFTs
Median preinduction FEV1 was 80% (IQR, 73%-98%), and median preinduction DLCO was 77% (IQR, 63%-92%). After induction therapy, median FEV1 increased to 83% (IQR, 71%-95%), whereas median DLCO decreased to 66% (IQR, 55%-79%). Postinduction FEV1 and DLCO were both statistically significantly higher in the ICI group than in the conventional systemic chemotherapy groups (p=0.037 and p<0.001, respectively) (Table 1). ANCOVA showed that a statistically significant difference in postinduction DLCO between the induction groups (p<0.0001) remained after adjustment for preinduction DLCO. Post hoc analysis found that mean postinduction DLCO was statistically significantly higher in the ICI group (76.4% [95% CI, 72.4%-80.5%]) than in the abbreviated (67.9% [95% CI, 66.0%-69.9%]), full (64.9% [95% CI, 63.2%-66.6%]), and unknown (66.3% [95% CI, 63.6%-69.1%]) chemotherapy groups (Supplementary Figure 1).
Outcomes by change from preinduction to postinduction DLCO
Patients were categorized according to the previously described ACOSOG subgroups for pre- and postinduction preoperative PFT measurements: normal-normal, normal-impaired, impaired-normal, impaired-impaired, and preinduction only (Table 3 and Supplementary Table 5). Distribution of FEV1 and DLCO PFT subgroups was similar among the four induction therapy groups (p=0.7 and p=0.2, respectively; Supplementary Table 6). Of note, relatively few patients had normal-impaired (n=9) or impaired-impaired (n=3) FEV1 status (Supplementary Table 7). For this reason, although we analyzed the association of change in FEV1 status and patient outcomes, we focused our discussion on DLCO status (Supplementary Table 8).
Table 3.
Pre- and postinduction ACOSOG groupings for DLCO status (N=519)
| Characteristic | Postinduction DLCO | Total (N=519) |
|
|---|---|---|---|
| Normal (N=483) | Impaired (N=86) | ||
| Preinduction DLCO | |||
| Normal | 418 (96) | 57 (70) | 475 (92) |
| Impaired | 19 (4.3) | 25 (30) | 44 (8.5) |
Data are no. (%) of patients. “Impaired” defined as DLCO <50%; “normal” defined as DLCO ≥50%. ACOSOG, American College of Surgeons Oncology Group; DLCO, diffusing capacity of the lung for carbon monoxide.
Overall, pre- to postinduction DLCO status was significantly associated with respiratory complications, in-hospital complications, and 30-day mortality (Table 4). Compared with normal-normal DLCO status in multivariable analyses, normal-impaired DLCO status had a significantly higher odds of respiratory complications (odds ratio [OR], 2.23; 95% CI, 1.12-4.29; p=0.019) and in-hospital complications (OR, 2.83; 95% CI, 1.55-5.26; p<0.001). Multivariable analysis demonstrated that there were no statistically significant differences in outcomes according to type of neoadjuvant therapy. However, whereas sublobar resection was associated with decreased in-hospital complications, pneumonectomy was associated with worse major complications. Furthermore, an open surgical approach was associated with an increased risk of in-hospital complications, male sex was associated with an increased risk of respiratory, in-hospital, and major complications, and more recent year of surgery was associated with worse in-hospital complications (Table 5).
Table 4.
Postoperative outcomes by change from pre- to postinduction DLCO status
| Characteristic | Impaired- normal (N=19) |
Both impaired (N=25) |
Both normal (N=418) |
Normal- impaired (N=57) |
Pre- only (N=426) |
P valuea |
|---|---|---|---|---|---|---|
| LOS | 7.0 (6.0-8.5) | 4.0 (4.0-7.0) | 5.0 (4.0-7.0) | 6.0 (4.0-7.0) | 5.0 (4.0-7.0) | 0.074 |
| Respiratory complication | 6 (32) | 4 (16) | 56 (13) | 15 (26) | 81 (19) | 0.023 |
| In-hospital complication | 9 (47) | 7 (28) | 138 (33) | 35 (61) | 150 (35) | <0.001 |
| Major complicationb | 2 (11) | 3 (12) | 33 (7.9) | 7 (12) | 32 (7.5) | 0.7 |
| 30-day mortality | 0 (0) | 2 (8.0) | 5 (1.2) | 4 (7.0) | 5 (1.2) | 0.002 |
Data are median (interquartile range) or no. (%). “Impaired” defined as DLCO <50%; “normal” defined as DLCO ≥50%. DLCO, diffusing capacity of the lung for carbon monoxide; IQR, interquartile range; LOS, length of stay.
Statistical tests performed: Kruskal-Wallis test; chi-squared test of independence.
Major complications: Clavien-Dindo grade ≥3.23
Table 5.
Multivariable analysis of primary and secondary outcomesa
| Respiratory complication | In-hospital complication | Major complication | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Characteristic | OR | 95% CI | P value | OR | 95% CI | P value | OR | 95% CI | P value |
| DLCO change | |||||||||
| Both normal | — | — | — | — | — | — | |||
| Pre- only | 1.53 | 1.05-2.26 | 0.030 | 1.12 | 0.83-1.52 | 0.5 | 1.00 | 0.59-1.71 | >0.9 |
| Impaired-normal | 3.19 | 1.07-8.57 | 0.027 | 2.09 | 0.79-5.44 | 0.13 | 1.93 | 0.29-7.45 | 0.4 |
| Both impaired | 1.39 | 0.39-3.90 | 0.6 | 0.90 | 0.33-2.20 | 0.8 | 1.86 | 0.41-6.16 | 0.4 |
| Normal-impaired | 2.23 | 1.12-4.29 | 0.019 | 2.83 | 1.55-5.26 | <0.001 | 1.42 | 0.54-3.33 | 0.4 |
| Treatment group | |||||||||
| Full | — | — | — | — | — | — | |||
| Abbreviated | 1.08 | 0.73-1.59 | 0.7 | 1.23 | 0.88-1.70 | 0.2 | 1.09 | 0.64-1.83 | 0.7 |
| ICI | 1.81 | 0.80-3.83 | 0.13 | 1.22 | 0.60-2.44 | 0.6 | 1.40 | 0.45-3.67 | 0.5 |
| Unknown | 1.15 | 0.66-1.94 | 0.6 | 1.09 | 0.70-1.69 | 0.7 | 0.65 | 0.25-1.44 | 0.3 |
| Procedure type | |||||||||
| Lobectomy/bilobectomy | — | — | — | — | — | — | |||
| Pneumonectomy | 0.63 | 0.33-1.12 | 0.14 | 1.30 | 0.84-1.99 | 0.2 | 1.91 | 1.01-3.42 | 0.036 |
| Sublobar | 0.56 | 0.24-1.14 | 0.14 | 0.35 | 0.17-0.65 | 0.002 | 0.15 | 0.01-0.73 | 0.066 |
| Sex | |||||||||
| Female | — | — | — | — | — | — | |||
| Male | 1.68 | 1.19-2.38 | 0.003 | 1.43 | 1.08-1.89 | 0.012 | 1.75 | 1.08-2.87 | 0.024 |
| Approach type | |||||||||
| VATS | — | — | — | — | — | — | |||
| Open | 1.85 | 1.20-2.89 | 0.006 | ||||||
| Surgery year | 1.05 | 1.01-1.08 | 0.004 | ||||||
CI, confidence interval; DLCO, diffusing capacity of the lung for carbon monoxide; ICI, immune checkpoint inhibitor; N/A, not applicable; OR, odds ratio; VATS, video-assisted thoracoscopic surgery.
A full univariable analysis is presented in Supplementary Table 10.
To verify the associations discovered between change in DLCO and the four outcomes of interest, we performed additional analyses with a DLCO <60% threshold—the minor criterion for ACOSOG high-risk surgery. ICI therapy was associated with an increased risk of respiratory complications. Normal-impaired DLCO status remained significantly associated with increased risk of in-hospital complications, although the association with respiratory complications was no longer statistically significant (Supplementary Table 9).
DISCUSSION
The relative perioperative risks associated with induction therapy remain a matter of controversy. Brunelli et al. found that induction chemotherapeutic agents were associated with a 10% to 20% reduction in DLCO.25 Furthermore, this decrease in DLCO after induction therapy has been associated with perioperative respiratory complications.26 However, limited research has been performed to identify functional markers that predict perioperative morbidity and mortality among patients undergoing surgical resection after induction therapy.12, 18 The present study represents one of the largest series in which perioperative morbidity and mortality were evaluated as a function of quantifiable changes in FEV1 and DLCO after induction chemotherapy.19, 27 On the basis of expert opinion, the ACCP developed a preoperative lung function algorithm to assess patient operability for thoracotomy and major anatomic resection.14 The algorithm relies on ppoDLCO and ppoFEV1—in combination with patient functional status, stair climbing, or shuttle walking—to risk-stratify patients before surgery.14 Research from our institution showed that baseline DLCO and ppoDLCO are reliable predictors of mortality in patients undergoing resection for NSCLC.19 The present study expands on these findings to evaluate whether changes in PFT measures, before and after induction therapy, can be used to predict perioperative morbidity and mortality.
Multiple phase III trials of induction ICI therapy, either alone or in combination with chemotherapy, are currently underway to definitively identify the perioperative risks related to ICI therapy. The present study represents one of the few early studies to directly assess the potential adverse effects of ICI therapy on the pulmonary function status of surgical candidates. In this setting, our ICI cohort had statistically significantly higher postinduction DLCO, compared with the conventional chemotherapy groups, but they had equivalent rates of pulmonary and major complications. Suzuki et al., who performed a prospective study on the use of ICI therapy in the treatment of patients with locoregionally advanced and metastatic NSCLC, reported increased rates of respiratory complications in their cohort. They observed that patients with higher DLCO or alveolar volume were more likely to develop ICI-related interstitial lung disease. Nevertheless, DLCO was not a predictor of pulmonary disease on univariable or multivariable analysis.28 Despite differences in disease stage between our study and the study from Suzuki et al., the incongruous relationship of patient DLCO and perioperative pulmonary adverse events suggests that more research is necessary to definitively determine what role, if any, pre- or postinduction DLCO can play in predicting perioperative pulmonary complications in patients treated with induction ICI therapy.
The current ACCP guidelines give a weak, Grade 2C, recommendation for physicians to perform additional PFTs, including diffusion capacity, after induction therapy.14,29 Although most patients had normal DLCO status after induction, 11% of patients went from normal to impaired DLCO status, with an attendant increase in respiratory and in-hospital complications (Tables 3 and 5). Our findings reaffirm that DLCO testing should be performed after induction therapy, as this may provide important data to surgeons assessing patient operability. Furthermore, to mitigate perioperative risk, patients with a reduction in DLCO after induction therapy should be considered for further screening, preoperative pulmonary physical therapy, or parenchymal-sparing operations, when appropriate.30,31
It may be asked why the 57 patients whose DLCO status went from normal to impaired underwent surgery, as these patients had potentially inadequate postinduction pulmonary function status (Table 3). Careful chart review indicated that these patients were ultimately deemed to be surgical candidates by other measures, including confirmatory ventilation-perfusion scans and satisfactory clinical evaluations. The full clinical picture is necessary for appropriate patient selection, and no single variable renders a patient inoperable. The differences in postoperative short-term outcomes depend on many preoperative clinical variables, rather than on DLCO alone. Nevertheless, our findings suggest that more attention should be paid to these patients, considering their potential increased risk of perioperative morbidity and mortality. The current guidelines recommend a multidisciplinary-team approach to assess operability that is dependent on a thoracic surgical oncologist with the aid of a cardiologist and pulmonologist to assess patient candidacy before curative surgical resection.14 We recommend complementary and alternative diagnostic tests at the time of the postinduction impaired PFTs as well as pulmonary optimization with 3 weeks of preoperative high-intensity rehabilitation followed by surgery.32
Our study has several limitations. Although the prospective collection of data from our institution acts to mitigate potential confounders, selection bias is unavoidable in a retrospective review. All patients in our study received two sets of PFTs during their preoperative course. Repeat preoperative testing may represent a unique characteristic of this patient population but comparative analysis between this cohort and the preinduction-only and postinduction-only PFT groups was not different across perioperative risk measures (Supplementary Table 2). Similarly, patients who did not undergo surgery because of poor postinduction PFT results were not included in this study. We were therefore unable to evaluate differences in pulmonary function status between these cohorts. Most patients received both sets of PFTs from our hospital’s pulmonary diagnostics laboratory, but a minority of patients received one or both sets of PFTs from an outside institution, which may have introduced intrainstitutional observer bias. Our findings are limited by the use of DLCO unadjusted for patient anemia. DLCO corrected for hemoglobin concentration is the preferred measurement tool for diffusion capacity, owing to the well-described incidence of post–induction therapy anemia, which can potentially falsely lower DLCO values. We performed sensitivity analyses of our unadjusted pre- and postinduction DLCO values against DLCO adjusted for hemoglobin in those patients with reported values and found that they were highly concordant (preinduction DLCO Spearman coefficient=0.98; postinduction DLCO Spearman coefficient=0.95). Although the pre- and postinduction median DLCO and FEV1 values were within the limits of normal, the range for these metrics was broad; therefore, patients at the extremes may have affected overall outcomes. There were several instances in which a patient shifted from normal to impaired DLCO status despite only a minor change in DLCO. Nevertheless, a reduction was considered a true change in pulmonary status when it met the categorical standard based on the predetermined ACOSOG cutoffs. These findings were then verified through auxiliary analysis using the ACOSOG minor criteria.
This study adds to the accumulating evidence that preoperative pulmonary function status can be used to identify associations between functional parameters and postoperative outcomes (Figure 2). We have shown that the change from pre- to postinduction DLCO—in particular, from normal to impaired—after induction therapy may be an important predictor of perioperative morbidity and mortality. These findings support and augment the current ACCP guidelines. Accordingly, we recommend repeat testing after induction therapy to aid in appropriate patient selection before major pulmonary resection as a means of mitigating operative risk.
Figure 2.
Graphical abstract. Treatment decision tree according to pre- and post-induction PFTs. 575 patients with pathologic stage I-III NSCLC who received induction therapy before lung resection and pre- and post-induction pulmonary function testing were included. Patients were categorized into four subgroups according to ACOSOG major criteria: normal-normal, normal-impaired, impaired-normal, and impaired-impaired. Changes from pre- to postinduction pulmonary function status were associated with differences in perioperative outcomes; therefore, repeat pulmonary function testing should be performed after induction therapy. If the patient’s status changed from normal to impaired the authors recommend re-evaluation, pre-operative pulmonary optimization and even sublobar resection, when appropriate, to reduce the perioperative risk.
Implications: Changes in DLco from pre- to postinduction pulmonary function status may predict perioperative outcomes; therefore, repeat pulmonary function testing should be performed after induction therapy to assess patient operability. If the patient’s status changes from normal to impaired, based on ACOSOG major criteria, re-evaluation, pulmonary optimization and even sublobar resection, in the appropriate context, should be assessed.
PFT, Pulmonary function testing; DLco, Diffusing capacity of the lung for carbon monoxide; NSCLC, Non-small cell lung cancer; SBRT, Stereotactic body radiation therapy
Supplementary Material
Supplementary Figure 1. Analysis of covariance post hoc diffusing capacity of the lung for carbon monoxide (DLCO) analysis. Data are estimated marginal means of the postinduction DLCO for each induction group after adjustment of preinduction DLCO. ICI, immune checkpoint inhibitor. **p<0.01; ***p<0.001; ****p<0.0001.
Central Picture:
Does pulmonary function status mediate induction therapy’s association to postop outcomes?
Central Message:
Changes from pre- to postinduction pulmonary function status may be associated with perioperative outcomes; therefore, repeat pulmonary function testing should be performed after induction therapy.
Perspective Statement:
It is unknown whether a reduction in pulmonary function after induction therapy and before surgery affects a patient’s risk of morbidity or mortality. To assess the relationship between induction therapy and postoperative outcomes, we analyzed pulmonary function test results before and after induction therapy in patients undergoing surgery for non-small cell lung cancer.
Acknowledgments
David B. Sewell of the Department of Surgery, Memorial Sloan Kettering Cancer Center, provided editorial assistance.
Funding:
This work was supported, in part, by the National Institutes of Health/National Cancer Institute (P30 CA008748 and T32 CA009501).
Glossary of Abbreviations
- ACCP
American College of Chest Physicians
- ANCOVA
analysis of covariance
- ASOCOG
American College of Surgeons Oncology Group
- DLCO
diffusing capacity of the lung for carbon monoxide
- FEV1
forced expiratory volume in 1 second
- ICI
immune checkpoint inhibitor
- IQR
interquartile range
- LOS
length of stay
- NSCLC
non-small cell lung cancer
- PFT
pulmonary function test
- ppo
predicted postoperative
- VATS
video-assisted thoracoscopic surgery
Footnotes
COI statement: Daniela Molena serves as a consultant for Johnson & Johnson, Urogen, and Boston Scientific. James M. Isbell has equity in LumaCyte LLC, serves as an uncompensated consultant to Roche-Genentech, and has received institutional research support from Grail and Guardant Health. Bernard J. Park has served as a proctor for Intuitive Surgical and consultant for COTA. Matthew J. Bott serves as a consultant for AstraZeneca. David R. Jones serves as a consultant for Merck and AstraZeneca. Gaetano Rocco has a financial relationship with Scanlan and serves as a consultant for AstraZeneca. All other authors have no potential conflicts to disclose.
IRB approval: IRB 18-391; Approved 9/7/18. All patients gave informed consent for participation in the approved study protocol and for publication of the study data.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES
- 1.Van Meerbeeck JP, Kramer GW, Van Schil PE, Legrand C, Smit EF, Schramel F, et al. Randomized controlled trial of resection versus radiotherapy after induction chemotherapy in stage IIIA-N2 non–small-cell lung cancer. J Natl Cancer I. 2007;99(6):442–50. [DOI] [PubMed] [Google Scholar]
- 2.International Adjuvant Lung Cancer Trial Collaborative Group. Cisplatin-based adjuvant chemotherapy in patients with completely resected non–small-cell lung cancer. N Engl J Med. 2004;350(4):351–60. [DOI] [PubMed] [Google Scholar]
- 3.Waller D, Peake MD, Stephens RJ, Gower NH, Milroy R, Parmar MK, et al. Chemotherapy for patients with non-small cell lung cancer: the surgical setting of the Big Lung Trial. Eur J Cardiothorac. 2004;26(1):173–82. [DOI] [PubMed] [Google Scholar]
- 4.Pisters KM, Ginsberg RJ, Giroux DJ, Putnam JB Jr, Kris MG, Johnson DH, et al. Induction chemotherapy before surgery for early-stage lung cancer: a novel approach. J Thorac Cardiovasc Surg. 2000;119(3):429–39. [DOI] [PubMed] [Google Scholar]
- 5.Forde PM, Chaft JE, Smith KN, Anagnostou V, Cottrell TR, Hellmann MD, et al. Neoadjuvant PD-1 blockade in resectable lung cancer. N Engl J Med. 2018;378(21):1976–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Yang CF, McSherry F, Mayne NR, Wang X, Berry MF, Tong B, et al. Surgical outcomes after neoadjuvant chemotherapy and ipilimumab for non-small cell lung cancer. Ann Thorac Surg. 2018;105(3):924–9. [DOI] [PubMed] [Google Scholar]
- 7.McDonald S, Rubin P, Phillips TL, Marks LB. Injury to the lung from cancer therapy: clinical syndromes, measurable endpoints, and potential scoring systems. Int J Rad Oncol. 1995;31(5):1187–203. [DOI] [PubMed] [Google Scholar]
- 8.Nishino M, Giobbie-Hurder A, Hatabu H, Ramaiya NH, Hodi FS. Incidence of programmed cell death 1 inhibitor–related pneumonitis in patients with advanced cancer: a systematic review and meta-analysis. JAMA Oncol. 2016;2(12):1607–16. [DOI] [PubMed] [Google Scholar]
- 9.Martin J, Ginsberg RJ, Abolhoda A, Bains MS, Downey RJ, Korst RJ, et al. Morbidity and mortality after neoadjuvant therapy for lung cancer: the risks of right pneumonectomy. Ann Thoracic Surg. 2001;72(4):1149–54. [DOI] [PubMed] [Google Scholar]
- 10.Albain KS, Swann RS, Rusch VW, Turrisi III AT, Shepherd FA, Smith C, et al. Radiotherapy plus chemotherapy with or without surgical resection for stage III non-small-cell lung cancer: a phase III randomised controlled trial. Lancet. 2009;374(9687):379–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Cerfolio RJ, Talati A, Bryant AS. Changes in pulmonary function tests after neoadjuvant therapy predict postoperative complications. Ann Thorac Surg. 2009;88(3):930–6. [DOI] [PubMed] [Google Scholar]
- 12.Margaritora S, Cesario A, Cusumano G, Cafarotti S, Corbo GM, Ferri L, et al. Is pulmonary function damaged by neoadjuvant lung cancer therapy? A comprehensive serial time-trend analysis of pulmonary function after induction radiochemotherapy plus surgery. J Thorac Cardiovasc Surg. 2010;139(6):1457–63. [DOI] [PubMed] [Google Scholar]
- 13.Nomori H, Shiraishi A, Cong Y, Shoji K, Misawa M, Sugimura H, et al. Impact of induction chemoradiotherapy on pulmonary function after lobectomy for lung cancer. J Thorac Cardiovasc Surg. 2018;155(5):2129–37. [DOI] [PubMed] [Google Scholar]
- 14.Brunelli A, Kim AW, Berger KI, Addrizzo-Harris DJ. Physiologic evaluation of the patient with lung cancer being considered for resectional surgery: diagnosis and management of lung cancer: American College of Chest Physicians evidence-based clinical practice guidelines. Chest. 2013;143(5):e166S–90S. [DOI] [PubMed] [Google Scholar]
- 15.Boushy SF, Billig DM, North LB, Helgason AH. Clinical course related to preoperative and postoperative pulmonary function in patients with bronchogenic carcinoma. Chest. 1971;59:383–91 [DOI] [PubMed] [Google Scholar]
- 16.Gass GD, Olsen GN. Preoperative pulmonary function testing to predict postoperative morbidity and mortality. Chest. 1986;89(1):127–35. [DOI] [PubMed] [Google Scholar]
- 17.Gopal R, Starkschall G, Tucker SL, Cox JD, Liao Z, Hanus M, et al. Effects of radiotherapy and chemotherapy on lung function in patients with non–small-cell lung cancer. Int J Rad Oncol. 2003;56(1):114–20. [DOI] [PubMed] [Google Scholar]
- 18.Rivera MP, Detterbeck FC, Socinski MA, Moore DT, Edelman MJ, Jahan TM, et al. Impact of preoperative chemotherapy on pulmonary function tests in resectable early-stage non-small cell lung cancer. Chest. 2009;135(6):1588–95. [DOI] [PubMed] [Google Scholar]
- 19.Barnett SA, Rusch VW, Zheng J, Park BJ, Rizk NP, Plourde G, et al. Contemporary results of surgical resection of non-small cell lung cancer after induction therapy: a review of 549 consecutive cases. J Thorac Oncol. 2011;6(9):1530–6. [DOI] [PubMed] [Google Scholar]
- 20.Hanna N, Johnson D, Temin S, Baker S Jr, Brahmer J, Ellis PM, et al. Systemic therapy for stage IV non-small-cell lung cancer: American Society of Clinical Oncology clinical practice guideline update. J Clin Oncol. 2017;35(30):3484–515. [DOI] [PubMed] [Google Scholar]
- 21.Fernando HC, Landreneau RJ, Mandrekar SJ, Hillman SL, Nichols FC, Meyers B, et al. Thirty- and ninety-day outcomes after sublobar resection with and without brachytherapy for non–small cell lung cancer: results from a multicenter phase III study. J Thorac Cardiovasc Surg. 2011;142(5):1143–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Fernando HC, Timmerman R. American College of Surgeons Oncology Group Z4099/Radiation Therapy Oncology Group 1021: A randomized study of sublobar resection compared with stereotactic body radiotherapy for high-risk stage I non–small cell lung cancer. J Thorac Cardiovasc Surg. 2012;144(3):S35–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg. 2004;240(2):205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51(6):1173–82. [DOI] [PubMed] [Google Scholar]
- 25.Brunelli A Preoperative functional workup for patients with advanced lung cancer. J Thorac Dis. 2016;8(Suppl 11):S840. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Takeda SI, Funakoshi Y, Kadota Y, Koma M, Maeda H, Kawamura S, et al. Fall in diffusing capacity associated with induction therapy for lung cancer: a predictor of postoperative complication? Ann Thorac Surg. 2006;82(1):232–6. [DOI] [PubMed] [Google Scholar]
- 27.Shin S, Choi YS, Jung JJ, Im Y, Shin SH, Kang D, et al. Impact of diffusing lung capacity before and after neoadjuvant concurrent chemoradiation on postoperative pulmonary complications among patients with stage IIIA/N2 non-small-cell lung cancer. Respir Res. 2020;21(1):1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Suzuki Y, Karayama M, Uto T, Fujii M, Matsui T, Asada K, et al. Assessment of immune-related interstitial lung disease in patients with NSCLC treated with immune checkpoint inhibitors: a multicenter prospective study. J Thorac Oncol. 2020;15(8):1317–27. [DOI] [PubMed] [Google Scholar]
- 29.Guyatt G, Gutterman D, Baumann MH, Addrizzo-Harris D, Hylek EM, Phillips B, Raskob G, Lewis SZ, Schünemann H. Grading strength of recommendations and quality of evidence in clinical guidelines: report from an American College of Chest Physicians task force. Chest. 2006. Jan 1;129(1):174–81. [DOI] [PubMed] [Google Scholar]
- 30.Rosero ID, Ramírez-Vélez R, Lucia A, Martínez-Velilla N, Santos-Lozano A, Valenzuela PL, et al. Systematic review and meta-analysis of randomized, controlled trials on preoperative physical exercise interventions in patients with non-small-cell lung cancer. Cancers. 2019;11(7):944. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Donington J, Ferguson M, Mazzone P, Handy J Jr, Schuchert M, Fernando H, et al. American College of Chest Physicians and Society of Thoracic Surgeons consensus statement for evaluation and management for high-risk patients with stage I non-small cell lung cancer. Chest. 2012;142(6):1620–35. [DOI] [PubMed] [Google Scholar]
- 32.Stefanelli F, Meoli I, Cobuccio R, Curcio C, Amore D, Casazza D, et al. High-intensity training and cardiopulmonary exercise testing in patients with chronic obstructive pulmonary disease and non-small-cell lung cancer undergoing lobectomy. Eur J Cardiothorac Surg. 2013;44(4):e260–5. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure 1. Analysis of covariance post hoc diffusing capacity of the lung for carbon monoxide (DLCO) analysis. Data are estimated marginal means of the postinduction DLCO for each induction group after adjustment of preinduction DLCO. ICI, immune checkpoint inhibitor. **p<0.01; ***p<0.001; ****p<0.0001.