Circulating Tumor-Macrophage Fusion Cells and Circulating Tumor Cells Complement Non–Small-Cell Lung Cancer Screening in Patients With Suspicious Lung-RADS 4 Nodules

Abstract

PURPOSE

Low-dose computed tomography (LDCT) screening of high-risk patients decreases lung cancer–related mortality. However, high false-positive rates associated with LDCT result in unnecessary interventions. To distinguish non–small-cell lung cancer (NSCLC) from benign nodules, in the present study, we integrated cellular liquid biomarkers in patients with suspicious lung nodules (lung cancer screening reporting and data system [Lung-RADS] 4).

METHODS

Prospectively, 7.5 mL of blood was collected from 221 individuals (training set: 90 nonscreened NSCLC patients, 74 high-risk screening patients with no/benign nodules [Lung-RADS 1-3], and 20 never smokers; validation set: 37 patients with suspicious nodules [Lung-RADS 4]). Circulating tumor cells (CTCs), CTC clusters, and tumor-macrophage fusion (TMF) cells were identified by blinded analyses. Screening patients underwent a median of two LDCTs (range, 1-4) with a median surveillance time of 30 (range, 11-50) months.

RESULTS

In the validation set of 37 Lung-RADS 4 patients, all circulating cellular biomarker counts (P < .005; Wilcoxon test) and positivity rates were significantly higher in 23 biopsy-proven NSCLC patients (CTCs: 23 of 23 [100%], CTC clusters: 6 of 23 [26.1%], and TMF cells: 15 of 23 [65.2%]) than in 14 patients with biopsy-proven benign nodules (6 of 14 [42.9%], 0 of 14 [0%], and 2 of 14 [14.3%]). On the basis of cutoff values from the training set, logistic regression with receiver operating characteristic and area under the curve analyses demonstrated that CTCs (sensitivity: 0.870, specificity: 1.0, and area under the curve: 0.989) and TMF cells (0.652; 0.880; 0.790) complement LDCT in diagnosing NSCLC in Lung-RADS 4 patients.

CONCLUSION

Cellular liquid biomarkers have a potential to complement LDCT interpretation of suspicious Lung-RADS 4 nodules to distinguish NSCLC from benign lung nodules. A future prospective, large-scale, multicenter clinical trial should validate the role of cellular liquid biomarkers in improving diagnostic accuracy in high-risk patients with Lung-RADS 4 nodules.

INTRODUCTION

Non–small-cell lung cancer (NSCLC) accounts for approximately 85% of lung malignancies. Annual screening of high-risk patients with chest low-dose computed tomography (LDCT) reduces lung cancer–related mortality by 20%.^1,2 However, LDCT screening has the major limitation of high false-positive rates, subsequently leading to unnecessary interventions, morbidities, and costs.^3,4 The American College of Radiology developed a lung cancer screening reporting and data system (Lung-RADS) to improve LDCT interpretation.⁵ Suspicious nodules are detected in approximately 4% of high-risk patients and categorized as Lung-RADS 4 with a cancer risk of > 5%, whereas Lung-RADS 1-3 are defined as absent or radiographically benign-appearing nodules.⁵ Nevertheless, even suspicious Lung-RADS 4 findings have a high false-positive rate, a clinical gap that blood-based liquid biomarkers could fill to increase the accuracy of LDCT screening.

CONTEXT

• Key Objective

• Chest low-dose computed tomography (LDCT) screening of high-risk patients decreases lung cancer–related mortality. However, false-positive rates associated with nodules detected by screening result in unnecessary interventions. This study aims to integrate cellular liquid biomarkers with LDCT to enhance non–small-cell lung cancer (NSCLC) diagnosis by distinguishing suspicious lung cancer screening reporting and data system (Lung-RADS) 4 nodules from benign nodules.

• Knowledge Generated

• In a prospective study of 221 patients, training (n = 184) and validation (n = 37) cohorts demonstrated that circulating tumor cells and circulating large tumor-macrophage fusion cells improve the sensitivity and specificity of NSCLC diagnosis in patients with suspicious Lung-RADS 4 nodules as determined by LDCT screening.

• Relevance

• Cellular liquid biomarkers can complement LDCT interpretation of suspicious Lung-RADS 4 nodules to differentiate NSCLC from benign nodules. A future multicenter trial should validate the role of cellular liquid biomarkers in improving diagnostic accuracy of NSCLC screening in high-risk patients with suspicious Lung-RADS 4 nodules.

Circulating tumor cells (CTCs) are considered as micrometastatic seeds of the primary tumor and detached cancer cells in the blood that can serve as easily accessible liquid biopsies.⁶ Tumor cell migration into the blood is understood to be an early event in carcinogenesis, yet detection of CTCs in early-stage patients is challenging because of low cell counts.⁷ CTCs are identified by US Food and Drug Administration criteria on the basis of epithelial cell adhesion molecule expression, followed by immunohistochemical identification as cytokeratin 8/18 and/or 19+ and leukocyte marker CD45–, with a 4′,6-diamidino-2-phenylindole+ nucleus.⁸ CTCs and CTC clusters (≥ 2 CTCs in aggregate) are prognostic for patients with NSCLC.^9,10 Recently, our group has described a unique CTC subtype in the blood that is large (≥ 30 μm) and exhibits epithelial and myeloid phenotypes.¹¹ These circulating tumor-macrophage fusion (TMF) cells in the blood of patients with nonmetastatic NSCLC are highly prognostic for poor survival after curative resection.¹¹ Since TMF cells were specific to NSCLC diagnosis and rarely detected in long-term smokers without lung cancer, it was already hypothesized that TMF cells might be clinically applicable liquid biomarkers for early detection of lung cancer.¹² As a proof-of-concept study on early detection of lung cancer, CTCs were identified in some patients with chronic obstructive pulmonary disease years before lung cancer was diagnosed by computed tomography scans, indicating the clinical utility of CTCs as an early diagnostic marker.¹³ A recent study demonstrated an improved accuracy of LDCT by integrating CTCs.¹⁴ By contrast, a multicenter trial described that a size-based CTC isolation technology was not suitable for identifying NSCLC in LDCT patients.¹⁵ However, this all-comer screening trial did not incorporate Lung-RADS interpretation that stratifies LDCT patients on the basis of detailed radiographic findings.

In the present prospective study, novel CTC subtypes, such as TMF cells, are integrated into standardized Lung-RADS criteria with a specific focus on suspicious Lung-RADS 4 nodules. This pilot study suggests a potential role for CTCs and TMF cells in complementing Lung-RADS findings by LDCT to identify NSCLC in patients with suspicious Lung-RADS 4 nodules.

METHODS

Patients were enrolled between July 2016 and December 2018, with a written informed consent. Analytic laboratory personnel were separated from the recruiting staff and blinded to any patient-specific findings. Patients were recruited on the basis of Lung-RADS criteria, as outlined below. This selection approach allowed us to investigate a focused, preselected primary study group of patients with suspicious Lung-RADS 4 nodules (low incidence of approximately 4%) at highest risk of cancer.⁵ The training set (total n = 184) included treatment-naïve nonscreened NSCLC patients (n = 90), high-risk chronic smokers without/with benign-appearing lung nodules (Lung-RADS 1-3; n = 74), and healthy never smokers (n = 20). The validation set consisted of Lung-RADS 4 patients with suspicious lung nodules (total n = 37). Upon tissue biopsy, patients with suspicious nodules were categorized into benign (n = 14) and NSCLC (n = 23). Inclusion criteria for lung cancer LDCT screening were defined as recommended by the US Preventive Services Task Force (grade B): 55-80 years, ≥ 30 pack-year smoking history, and current smokers/quit within the past 15 years.² Patients with a concurrent or history of cancer were excluded. Exclusion criteria for nonscreened NSCLC patients was the presence of a pre-existing cancer of any origin. Institutional Review Board approval was obtained at the University of Missouri and Truman VA Hospital (institutional review board No.: 2010166-MU/2004401-VA). All experiments complied with the standards of the Declaration of Helsinki and were performed in accordance with relevant guidelines and regulations for human subject research. Trials were registered at ClinicalTrials.gov identifiers (NCT02838836/NCT03551951). Additional methodologic details on LDCT imaging, clinicopathological data, liquid biomarker detection, and statistical analysis are provided in the Data Supplement.

RESULTS

Diagnostic Challenge in Lung-RADS 4 Patients and Study Patients' Characteristics

To illustrate the diagnostic challenge within the study group chosen for this analysis, two similar-appearing Lung-RADS 4 nodules are shown in Figure 1A. Upon tissue biopsy of these suspicious lung nodules, one turned out to be a benign granuloma (Fig 1A [a]) and the other a stage IA NSCLC (Fig 1A [b]). To address this diagnostic challenge and to accurately diagnose NSCLC, we incorporated circulating cellular biomarkers (CTCs, CTC clusters, and TMF cells). A total of 221 study patients were divided into training (n = 184) and validation sets (n = 37; Table 1). Patient clinical characteristics of training and validation sets (total N = 221) are presented in Table 1. The training set consisting of nonscreened NSCLC patients (n = 90), high-risk smoking controls with Lung-RADS (1-3; n = 74), and healthy controls (n = 20) was recruited to develop a model that incorporated cellular liquid biomarkers for NSCLC detection. The validation set consisted of 37 prospective high-risk screening patients with Lung-RADS 4 nodules. All these 37 Lung-RADS 4 patients underwent tissue biopsy. We determined that 23 (62%) patients had NSCLC, whereas 14 (38%) patients had benign lung nodules (eg, granulomata). Because of the benign LDCT findings, the 74 high-risk screening patients with Lung-RADS 1-3 did not undergo further tissue biopsy. Lung-RADS 1-3 patients were further observed over a median time of 30 (range, 11-50) months, undergoing a median of 2 (range, 1-4) LDCT scans over this period. During this surveillance period, only one (1.4%) of the Lung-RADS 1-3 patients with benign findings developed a lung cancer.

FIG 1.

Diagnostic challenge to distinguish suspicious lung nodules and characteristics of cellular liquid biomarkers. (A) Representative low-dose computed tomography scans of two different high-risk screening patients with similar-appearing suspicious lung nodules (arrows) categorized as Lung-RADS 4. (A.a) Biopsy revealed a benign granuloma; (A.b) biopsy revealed NSCLC. (B) After blood draws (7.5 mL), cells were enriched by size with a microfilter and immunofluorescence staining performed. CTCs (B.a-B.c) were defined as CK+/EpCAM+/CD14/CD45– with a DAPI+ nucleus, (B.d-B.f) CTC clusters (aggregates of ≥ 2 CTCs), (B.g-B.l) circulating TMF cells, and (B.m-B.r) giant TMF cells. Arrow in (B.p): CTC cluster. TMF cells are defined as CK/EpCAM+ and CD14/45+ with ≥ 1 DAPI+ nucleus and ≥ 30 μm size with polymorphic shapes, and some of them were giant with cell sizes ≥ 50 μm. Representative images of merged immunostainings are shown. CK, cytokeratin; CTC, circulating tumor cell; DAPI, 4′,6-diamidino-2-phenylindole; EpCAM, epithelial cell adhesion molecule; Lung-RADS, lung cancer screening reporting and data system; NSCLC, non–small-cell lung cancer; TMF, tumor-macrophage fusion cell.

TABLE 1.

Study Patients (training and validation sets)

CTCs and CTC Clusters As Potential Biomarkers to Complement LDCT Screening to Diagnose NSCLC in Patients With Suspicious Lung-RADS 4 Nodules

To determine whether cellular liquid biomarkers can be integrated into LDCT to differentiate Lung-RADS 4 nodules, CTCs (Fig 1B [a-c]), CTC clusters (Fig 1B [d-f]), TMF cells (Fig 1B [g-l]), and giant TMF cells (Fig 1B [m-r]) were enriched and detected by immunostaining. Within the training set, nonscreened NSCLC patients (n = 90) were compared with high-risk screening patients (with a similar risk profile) with or without benign-appearing lung nodules on LDCT (Lung-RADS 1-3; n = 74), whereas healthy nonsmokers (n = 20) were used as baseline. CTCs were detected in all 90 patients with NSCLC; however, CTCs were detected only in 30 of 74 (40.5%) Lung-RADS 1-3 patients (Table 2). Nonscreened NSCLC patients had significantly higher CTC counts (mean [± SEM]; 22.53 [± 1.16]) compared with Lung-RADS 1-3 patients (0.70 [± 0.57]; P < .0001; Table 2; Fig 2A). CTCs were absent in healthy controls. On the basis of the findings of the training set, quantitative CTC analysis was performed in validation set patients with Lung-RADS 4 suspicious nodules (n = 37). CTCs were present in 78.4% (29 of 37) of all Lung-RADS 4 patients (Table 2). Similar to nonscreened NSCLC patients, all 23 Lung-RADS 4 patients with biopsy-proven NSCLC had CTCs detectable, with a significantly higher mean count of 12.87 (± 1.55; Table 2; Fig 2A). By contrast, of 14 Lung-RADS 4 patients with benign lung nodules, only six patients had CTCs with a mean count of 1.0 (± 0.36; P < .0001; Table 2; Fig 2A).

TABLE 2.

Results on CTCs, CTC Clusters, and TMF Cells in the Study Population

FIG 2.

Counts of cellular liquid biomarkers are higher in patients with NSCLC, including high-risk patients with Lung-RADS 4 nodules on screening low-dose computed tomography with subsequently biopsy-proven NSCLC: (A) CTCs, (B) CTC clusters, (C) circulating TMF cells (all sizes ≥ 30 μm), and (D) giant TMF cells ≥ 50 μm were either rare (CTCs and TMF cells) or absent (CTC clusters, giant TMF cells ≥ 50 μm) in unsuspicious Lung-RADS 1-3 patients versus NS NSCLC patients (training set). Similar findings with respect to all groups of cellular liquid biomarkers were observed in Lung-RADS 4 screening patients with NSCLC versus benign nodules proven by tissue biopsy (validation set). All four types of cellular liquid biomarkers were absent in healthy controls tested. CTC, circulating tumor cell; Lung-RADS, lung cancer screening reporting and data system; NS, nonscreened; NSCLC, non–small-cell lung cancer; TMF, tumor-macrophage fusion cell.

CTC clusters were specifically present in nonscreened NSCLC patients and absent in high-risk and healthy controls (Table 2). Ten of 90 (11.1%) nonscreened NSCLC patients had CTC clusters with a mean count of 0.2 (± 0.07; P < .01; Fig 2B). In alignment with the findings from the training set, CTC clusters were detected in 6 of 23 (26.1%) Lung-RADS 4 patients diagnosed with biopsy-proven NSCLC, with a mean count of 0.52 (± 0.11; P < .001; Table 2; Fig 2B).

TMF Cells Were More Prevalent in NSCLC Patients With or Without LDCT Screening

TMF cells were more prevalent in nonscreened NSCLC patients (73 of 90; 81%) compared with their rare occurrence in screened patients with Lung-RADS 1-3 (4 of 74; 5.4%; Table 2). TMF counts in the training set of nonscreened NSCLC patients were significantly higher (3.28 [± 0.34]) compared with patients with Lung-RADS 1-3 (0.05 [± 0.03]; P < .0001; Table 2; Fig 2C). In the validation set, Lung-RADS 4 patients with NSCLC had significantly higher TMF counts (2.44 [± 0.69]) in comparison with Lung-RADS 4 patients with biopsy-proven benign lung nodules (0.14 [± 0.10]; P < .005; Table 2; Fig 2C). A subclass of TMF cells ≥ 50 μm (giant TMF cells) were exclusively detected in patients with NSCLC, whereas giant TMF cells were absent in high-risk smoking controls (Lung-RADS 1-3) and in healthy nonsmokers (Table 2). Forty-three of 90 (48%) patients with NSCLC were positive for giant TMF cells with a mean count of 0.81 (± 0.15; P < .0001; Table 2; Fig 2D). In light of their prevalence only in nonscreened NSCLC patients in the training set, giant TMF analysis was performed in the validation set of Lung-RADS 4 patients. Significantly higher counts of giant TMF cells (0.70 ± 0.24; P < .01; Table 2; Fig 2D) were observed in biopsy-proven Lung-RADS4 NSCLC patients, whereas they were absent in patients with benign Lung-RADS 4 nodules.

CTCs and TMF Cells Independently Complement LDCT Interpretation for Diagnosis of NSCLC in the Validation Set With Suspicious Lung-RADS 4 Nodules

Univariate logistic regression was performed to analyze the relationship of cellular liquid biomarkers with NSCLC diagnosis in patients with Lung-RADS 4 nodules. The cutoff value for each biomarker was determined using receiver operating characteristic analysis in the training set (nonscreened NSCLC patients [n = 90] v patients with Lung-RADS [1-3; n = 74]). On the basis of the cutoff values from the training set, performance of the models was validated in patients with suspicious Lung-RADS 4 nodules (n = 37). CTCs at a cutoff of 0.533 showed a very high sensitivity of 0.989 and a specificity of 1.0 with an area under the curve (AUC) of 0.992 toward NSCLC diagnosis (Fig 3A). In the validation set, CTCs showed high sensitivity (0.870) and specificity (1.0) to complement LDCT in diagnosing NSCLC in Lung-RADS 4 patients (AUC: 0.989; Fig 3B). Similarly, TMF cells had a sensitivity of 0.811, a specificity of 0.945, and an AUC of 0.894 at a cutoff of 0.518 for NSCLC diagnosis in the training set (Data Supplement). In Lung-RADS 4 patients, TMF cells had a sensitivity of 0.652 and a specificity of 0.880 in NSCLC diagnosis (AUC: 0.790; Data Supplement). In contrast to CTCs and TMF cells, CTC clusters showed very low sensitivity (0.111) at a cutoff of 0.761, a specificity of 1.0, and an AUC of 0.556 in the training set (Data Supplement). In the validation set, CTC clusters were associated with a sensitivity of 0.522, a specificity of 1.000, and an AUC of 0.761 (Data Supplement). Bivariate regression analysis was performed to evaluate the role of TMF cells plus CTC clusters in complementing LDCT screening in NSCLC diagnosis. TMF cells plus CTC clusters at a cutoff of 0.512 were associated with the sensitivity, specificity, and AUC of 0.811, 0.945, and 0.895, respectively, in NSCLC diagnosis within the training set (Fig 3C). At the same cutoff, performance of TMF cells plus CTC clusters was evaluated in patients with suspicious Lung-RADS 4 lung nodules in the validation set. Findings revealed a sensitivity, specificity, and AUC of 0.880, 0.783, and 0.869, respectively, for diagnosis of NSCLC (Fig 3D).

FIG 3.

Univariate and bivariate receiver operating characteristic curve analyses to associate CTCs, TMF cells, and CTC clusters with NSCLC diagnosis in patients with Lung-RADS 4 nodules. (A) The univariate logistic regression model was based on the training set (Lung-RADS 1-3 with no or benign-appearing lung nodules v nonscreened NSCLC) for CTCs to determine the cutoff point with highest AUC, sensitivity, and specificity. (B) Cutoff point from the training set was used to evaluate the performance of CTCs in the validation set of patients with Lung-RADS 4 nodules. (C) Bivariate logistic regression models were chosen to determine the cutoff point to simultaneously include TMF cells plus CTC clusters in NSCLC diagnosis. (D) Performance of incorporating TMF cells plus CTC clusters (in combination) for NSCLC diagnosis in the validation set of patients with suspicious Lung-RADS 4 nodules was evaluated on the basis of the cutoff point determined from the training set. AUC, area under the curve; CTC, circulating tumor cell; Lung-RADS, lung cancer screening reporting and data system; NSCLC, non–small-cell lung cancer; TMF, tumor-macrophage fusion cell.

Despite the significant association of CTCs only and TMF cells plus CTC clusters combined with NSCLC diagnosis in patients with suspicious Lung-RADS 4 nodules, the small sample size in the validation set revealed that all three cellular liquid biomarkers cannot be combined together to diagnose NSCLC.

DISCUSSION

Although higher accuracy of LDCT screening of patients at risk for lung cancer is desired, improvements in screening of Lung-RADS 4 patients with suspicious lung nodules will likely improve cancer-specific outcomes and reduce unnecessary interventions and associated morbidities. The present study demonstrated that integrating cellular liquid biomarkers, CTCs, and TMF cells into standardized LDCT screening protocols can improve accuracy of NSCLC detection in a preselected and clinically relevant subgroup of Lung-RADS 4 patients with suspicious nodules.

The aim of the present study was to provide a more comprehensive cellular liquid biomarker analysis beyond CTC criteria by including TMF cells that are highly prognostic in early-stage NSCLC.¹¹ Our study focused on a subgroup (approximately 4%) of lung cancer screening patients with suspicious Lung-RADS 4 nodules, as defined by standardized radiographic interpretation.⁵ This granular LDCT interpretation and risk stratification may benefit from liquid biomarker integrations, rather than focusing on a broad population with uncategorized lung nodules as other groups have done in comparable studies.¹⁵ More than 90% of screening patients have no or unsuspicious lung nodules on their first LDCT (Lung-RADS 1-3) and just require biannual/annual LDCT follow-up imaging. On the other hand, patients with suspicious Lung-RADS 4 nodules already have a > 5% (and probably much higher) cancer risk.⁵ Therefore, within our screening population, we chose to examine cellular liquid biomarkers in this clinically important group of Lung-RADS 4 patients.

Liquid biopsy integration into early detection of lung cancer has been suggested using various biomarkers, such as circulating tumor DNA,^16-19 extracellular vesicles, microRNAs,²⁰ and metabolomics.²¹ In contrast to cell-free liquid biomarkers released from tumor cells (eg, extracellular vesicles), detection of CTCs and TMF cells as whole cells might be less sensitive because of their scarcity in blood, but potentially more relevant owing to possible replicative potential of cells leading to metastasis.^22,23 Our study highlighted the potential of CTCs to complement LDCT screening in NSCLC diagnosis with high sensitivity. The presence of CTCs plays a negative prognostic role in different solid tumors, including lung cancer.^9,24 Since tumor cell migration into the blood is an early event in cancer, there is a likely role of CTCs and other circulating cancer-associated cells in early detection of lung cancer.^7,15,25 In contrast to a previous LDCT and cellular liquid biomarker study, which had broader study criteria without a preselection by detailed radiographic lung nodule interpretation,¹⁵ our cellular biomarker study included a standardized radiographic imaging interpretation system with Lung-RADS criteria to preselect patients. This system was applied to focus on a study group with suspicious Lung-RADS 4 nodules at highest risk of cancer.⁵

In addition to CTCs, we also detected CTC clusters that have significantly enhanced metastatic potential in comparison with individual CTCs.²⁶ Recently, our group demonstrated that CTC clusters are exclusively detected in patients with NSCLC, but not in high-risk patients with no or benign-appearing lung nodules on LDCT (Lung-RADS 1-3).¹⁰ In the present study, we observed that CTC clusters were also absent in Lung-RADS 4 patients with biopsy-proven benign lung nodules. Furthermore, the applied microfilter technology also allowed the detection of recently identified large (≥ 30 µm), circulating TMF cells with a dual epithelial and macrophage phenotype,^11,27,28 which remain undetected with most other CTC detection platforms.^27-29 These distinctively large TMF cells have been identified in the blood of patients with organ metastasis; importantly, the presence of TMF cells is associated with shorter survival across different solid tumors, including lung cancer.^18,19,30,31 In a recent study, our group showed that among patients with NSCLC who underwent curative resection, TMF cells correlated with shorter survival.¹¹ Importantly, TMF cells were highly specific for patients with NSCLC and found rarely in cancer-free smokers. Investigators suggested that these TMF cells might play a role in early lung cancer detection.¹² A subset of giant TMF cells (≥ 50 µm) are rarely detected in patients with NSCLC, but not in high-risk smokers with biopsy-proven benign lung nodules.^11,29 In the present study, giant TMF cells were also specific for patients with NSCLC, yet their rarity may make them unsuitable biomarkers for early detection. TMF cells may play a critical role in tumor initiation and acquisition of metastatic traits by generating diverse and unique metastatic clones.^19,32-34 Besides being prognostic and a cancer screening biomarker, the role of TMF cells in early NSCLC needs further study, as TMF cells could play important oncoimmunologic and therapeutically targetable roles.

Like in previous studies that reported the presence of CTCs and TMF cells in breast and lung cancer screening patients with benign conditions, the present findings also showed that CTCs and TMF cells were detected in LDCT screening patients with benign nodules.^10,11,28 Patients with a cancer history were excluded since they could carry micrometastatic disease in the form of CTCs. However, false-positive findings that we observed remain of unclear etiology and the causes need to be further explored, as these patients might have a yet undetected early cancer (lung or other origin). Future studies may require a longer-than-30-month follow-up period, which would also include a higher number of LDCT scans with more information on changes in lung nodule characteristics. Finally, a recent follow-up study from the National Lung Screening Trial revealed that on extended follow-up (> 10 years) of high-risk screening patients, the reduction in lung cancer–associated deaths is sustained by annual LDCT screening, which exemplifies the need for longer surveillance.³⁵

The present observational trial interpretation was limited by selection bias as the focus was specifically on Lung-RADS 4 patients. Patients with radiographically similar-appearing suspicious lung nodules are very interesting and a clinically most critical class of screening patients, and as such, because of the aggressiveness of NSCLC, there is an immediate need for further workup. Selection bias was intended to be reduced by prospective and blinded analysis. Then, screening patients are defined by broad inclusion criteria, leading to population heterogeneity, confounding comorbidities, and variable extent of smoking beyond 30 pack-years, which qualify for LDCT screening. Beyond the follow-up provided here, further long-term patient surveillance will need to be performed. Furthermore, the present study was conducted at two independent centers in rural America, such that patients enrolled exhibited similar exposure profiles. To increase scientific rigor and to limit study bias, we separated patients into training (nonscreened NSCLC patients and patients with unsuspicious Lung-RADS 1-3 findings without cancer; healthy controls were used as baseline) and validation sets (patients with suspicious Lung-RADS 4 nodules). Clearly, larger studies containing external validation sets from multiple regional treatment centers will be more impactful.

In summary, the results of this pilot, bi-institutional, prospective trial suggest diagnostic value of integrating CTCs and TMF cells with LDCT screening for detection of NSCLCs in patients with suspicious Lung-RADS 4 lung nodules. Future prospective, multicenter clinical trials with increased Lung-RADS 4 patients with suspicious nodules will validate the clinical applicability of cellular biomarkers in LDCT screening. These future studies could also integrate cellular with cell-free liquid biomarkers for higher accuracy and investigate the combined value in improving early detection of the most lethal cancer worldwide.

AUTHOR CONTRIBUTIONS

Conception and design: Yariswamy Manjunath, Eric T. Kimchi, Kevin F. Staveley-O'Carroll, Klaus Pantel, Guangfu Li, Jussuf T. Kaifi

Financial support: Jussuf T. Kaifi

Administrative support: Aadel A. Chaudhuri, Jussuf T. Kaifi

Provision of study materials or patients: Diego M. Avella Patino, Eric T. Kimchi, Jussuf T. Kaifi

Collection and assembly of data: Yariswamy Manjunath, Kanve Nagaraj Suvilesh, Diego M. Avella Patino, Klaus Pantel, Jussuf T. Kaifi

Data analysis and interpretation: Yariswamy Manjunath, Jonathan B. Mitchem, Diego M. Avella Patino, Eric T. Kimchi, Klaus Pantel, Huang Yi, Guangfu Li, Peter K. Harris, Aadel A. Chaudhuri, Jussuf T. Kaifi

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).