Quick links
Quick links
J Chest Surg 2024; 57(1): 44-52
Published online January 5, 2024 https://doi.org/10.5090/jcs.23.094
Copyright © Journal of Chest Surgery.
Hye Rim Na , M.D.1, Seok Whan Moon
, M.D., Ph.D.1, Kyung Soo Kim
, M.D., Ph.D.1, Mi Hyoung Moon
, M.D., Ph.D.1, Kwanyong Hyun
, M.D.2, Seung Keun Yoon
, M.D.1
1Department of Thoracic and Cardiovascular Surgery, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul; 2Department of Thoracic and Cardiovascular Surgery, St. Vincent’s Hospital, College of Medicine, The Catholic University of Korea, Suwon, Korea
Correspondence to:Seung Keun Yoon
Tel 82-2-2258-2858
Fax 82-2-594-8644
E-mail skycs@catholic.ac.kr
ORCID
https://orcid.org/0000-0002-2609-2148
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background: Visceral pleural invasion (VPI) is a poor prognostic factor that contributes to the upstaging of early lung cancers. However, the preoperative assessment of VPI presents challenges. This study was conducted to examine intraoperative pleural carcinoembryonic antigen (pCEA) level and maximum standardized uptake value (SUVmax) as predictive markers of VPI in patients with clinical T1N0M0 lung adenocarcinoma.
Methods: A retrospective review was conducted of the medical records of 613 patients who underwent intraoperative pCEA sampling and lung resection for non-small cell lung cancer. Of these, 390 individuals with clinical stage I adenocarcinoma and tumors ≤30 mm were included. Based on computed tomography findings, these patients were divided into pleural contact (n=186) and non-pleural contact (n=204) groups. A receiver operating characteristic (ROC) curve was constructed to analyze the association between pCEA and SUVmax in relation to VPI. Additionally, logistic regression analysis was performed to evaluate risk factors for VPI in each group.
Results: ROC curve analysis revealed that pCEA level greater than 2.565 ng/mL (area under the curve [AUC]=0.751) and SUVmax above 4.25 (AUC=0.801) were highly predictive of VPI in patients exhibiting pleural contact. Based on multivariable analysis, pCEA (odds ratio [OR], 3.00; 95% confidence interval [CI], 1.14–7.87; p=0.026) and SUVmax (OR, 5.25; 95% CI, 1.90–14.50; p=0.001) were significant risk factors for VPI in the pleural contact group.
Conclusion: In patients with clinical stage I lung adenocarcinoma exhibiting pleural contact, pCEA and SUVmax are potential predictive indicators of VPI. These markers may be helpful in planning for lung cancer surgery.
Keywords: Adenocarcinoma of lung, Visceral pleural invasion, Pleural carcinoembryonic antigen, Maximum standardized uptake value, Thoracic surgery
Lung cancer is a leading cause of cancer-related deaths, with a 5-year relative survival rate of 22% [1]. Advances in screening examinations have facilitated the detection of smaller and less invasive cancers, thereby increasing the likelihood of surgical treatment. However, for patients eligible for surgery, debate remains as to whether to perform sublobar or lobar resection for early-stage non-small cell lung cancer (NSCLC). A recent randomized controlled trial demonstrated that segmentectomy is not inferior to lobectomy in terms of overall survival in clinical stage IA peripheral NSCLC. However, that study also reported a twofold increase in locoregional relapse following segmentectomy [2]. Consequently, careful selection of the most suitable surgical approach for patients is crucial to minimize the risk of recurrence in stage IA lung cancers.
Extensive research has been conducted into factors that contribute to the invasiveness and recurrence of small-size lung cancers. A retrospective study involving 327 patients with stage IA NSCLC who underwent sublobar resection identified lymphatic permeation, microscopic positive surgical margin, and visceral pleural invasion (VPI) as significant predictors of locoregional recurrence [3]. In the eighth edition of the tumor-node-metastasis (TNM) classification system, VPI is considered a T2 descriptor even for tumors smaller than 3 cm and results in the upstaging of IA tumors to IB [4,5]. Pathological VPI is characterized by the extension of the tumor beyond the elastic layer of the pleura. According to the modified Hammar classification, pleural invasion can be categorized into 4 stages, from PL0 to PL3 [6]. PL0 signifies the absence of VPI, with the tumor located in the subpleural lung tissue, just beneath the elastic layer. PL1 represents an extension beyond the elastic layer, while PL2 indicates further progression that reaches the pleural surface. Tumors invading any part of the parietal pleura are classified as PL3. Of these categories, PL1 and PL2 denote VPI, suggesting potential invasion into the pleural space.
Unlike pathologic evaluation, the clinical diagnosis of VPI has not been clearly established. Some researchers have proposed combinations of features of the tumor-pleura relationship on computed tomography (CT). These features include direct pleural contact of the tumor, pleural retraction, and pleural tags [7-10]. However, a study by Kim et al. [7], which examined CT-defined VPI and its correlation with disease-free survival in a retrospective review of 695 patients, revealed that approximately one-half of the findings were false positives. Consequently, the accuracy of these findings ranged from 62.7% to 72.3%, suggesting a low level of reliability as an independent predictor [7].
Multiple separate studies have proposed alternative methods for the preoperative assessment of VPI, such as the measurement of circulating tumor cells from serum and pre-resectional pleural lavage tumor cytology. However, these approaches require further research for validation [8,11]. The maximum standardized uptake value (SUVmax) and serum carcinoembryonic antigen (CEA) level have also been identified as significant risk factors for VPI, but no definite cutoff values have been established for diagnosis [12-14]. In the present study, given the close association of VPI with invasion into the pleural cavity, we hypothesized that tumor markers in the pleural fluid could serve as predictive markers for VPI. To our knowledge, no investigations have yet been published regarding the association between pleural CEA (pCEA) and VPI in early-stage adenocarcinomas.
This study involved a retrospective review of patients with clinical stage I adenocarcinoma, with tumors measuring 30 mm or less, who underwent intraoperative pCEA sampling prior to lung resection. The primary objectives were to evaluate the predictive value of pCEA and SUVmax in the clinical diagnosis of VPI and to ascertain whether these parameters could aid in decision-making regarding lung surgery.
A retrospective review was conducted of the electronic medical records of patients who underwent intraoperative pCEA collection followed by surgical resection for NSCLC at Seoul St. Mary’s Hospital in Seoul, Korea. The review covered a study period from January 2017 to December 2022. The exclusion criteria included patients with clinical stage II to IV adenocarcinoma as confirmed by radiological imaging, those with pathologically confirmed non-adenocarcinoma, and those who had previously undergone lung resection. Among the patients with clinical stage I lung adenocarcinoma, those with a tumor larger than 30 mm were additionally excluded (Fig. 1). This study received approval from the institutional review board at Seoul St. Mary’s Hospital (approval no., KC23RASI0381), and informed consent was waived due to the retrospective nature of the study.
Preoperative CT images were captured using a Siemens device (Somatom, Erlangen, Germany). These images were then reconstructed in axial, coronal, and sagittal planes, with a section thickness of either 1 or 3 mm. All measurements were conducted using a lung window setting, which had a window level of 600 Hounsfield units (HU) and a window width of 1,600 HU. Within this lung window setting, pulmonary tumors that directly contacted the pleural surface or interlobar fissure were classified as pleural contact tumors. Tumors that did not meet this criterion were categorized as non-pleural contact tumors (Fig. 2). Furthermore, tumors exhibiting signs of a pleural tag, defined as linear soft tissue strands linking the tumor and the pleural surface, were included in the non-pleural contact category. Preoperative 18F-fluorodeoxyglucose (18F-FDG) positive emission tomography/computed tomography (PET/CT) images were acquired using a Discovery 710 device (GE Healthcare, Milwaukee, WI, USA). All patients were required to fast for at least 6 hours prior to examination. Images were then captured 60 minutes after intravenous injection of 18F-FDG (0.12 mCi/kg) and at intervals of 1.5 to 2 minutes per bed position.
Patients diagnosed with clinical stage IA lung cancer were considered eligible for surgical treatment. The standard surgical approach involved performing a lobectomy accompanied by systemic lymph node dissection. However, for patients at low risk, with peripheral tumors either measuring less than 2 cm or exhibiting more than 50% ground-glass opacity, sublobar resection was considered in line with the National Comprehensive Cancer Network guidelines [15]. For these patients, lobe-specific lymph node dissection or sampling were also potential options. Surgical procedures were performed using videoscopic-assisted thoracoscopic surgery (VATS), with the number of incisions ranging from 2 to 4, based on the surgeon’s preference. Following skin incision and visualization of the thoracic cavity, a minimum of 2 mL of pleural effusion fluid was collected for CEA analysis before any other lung manipulation. Pleural effusion samples were most frequently obtained from the costodiaphragmatic recess beneath the inferior pulmonary ligament (Fig. 3). However, pCEA sampling was not conducted in patients with diffuse lung adhesion or in cases in which the pleural fluid was contaminated by massive bleeding. The measurement of pCEA was performed using an electrochemiluminescence immunoassay with the Elecsys CEA kit (Roche Diagnostics, Penzberg, Germany), and results were typically available approximately 30 minutes after sampling.
The resected lung specimens were fixed in formalin, deparaffinized, and stained with hematoxylin-eosin. Elastic staining was employed for the pathological confirmation of VPI. The final report incorporated details including tumor size, histological type, differentiation, pleural invasion, lymphovascular invasion, resection margin, nodal status, pathologic stage, elastic stain, and immunohistochemistry. The presence of VPI was determined using the modified Hammar classification system, as follows: PL0 represented a tumor within the subpleural lung parenchyma or superficially invading the pleural connective tissue below the elastic layer, PL1 denoted tumor invasion beyond the elastic layer, PL2 represented invasion reaching the pleural surface, and PL3 indicated invasion of the parietal pleura. The PL1 and PL2 categories describe VPI and are incorporated within T2 staging [6].
Data representing baseline characteristics are presented as frequencies with percentages for categorical variables and as mean±standard deviation for continuous variables. The Mann-Whitney U test was employed for comparisons of continuous variables, while the chi-square or Fisher exact tests were used for categorical variables. The association between variables and VPI was assessed using receiver operating curve (ROC) analysis, with the area under the curve (AUC) calculated for each index. Logistic regression was applied to analyze risk factors for VPI, and those with p-values of less than 0.2 were included in the multivariable analysis. All statistical computations were performed using R ver. 4.0.4 (R Foundation for Statistical Computing, Vienna, Austria), utilizing the Epi, ggplot2, and moonBook packages. A p-value of less than 0.05 was considered to indicate statistical significance.
The study population comprised 390 patients with clinical stage I adenocarcinoma, with tumors measuring 30 mm or less, who underwent pulmonary resection. Of these, 50 patients exhibited pathologic VPI, while the remaining 340 did not. A significant difference was observed in the proportion of patients with elevated pCEA levels between the VPI group (58.0%) and the non-VPI group (18.8%) (p<0.001). Elevated SUVmax was also noted more frequently in the VPI group (81.2%) compared to the non-VPI group (30.2%) (p<0.001). Moreover, patients with pathologic VPI showed significant differences in the extent of lung resection, tumor size, nodule type, pleural contact on CT, nodal status, pathological TNM stage, and predominant histological subtype. Table 1 details the characteristics of the study population.
Table 1. Clinicopathologic characteristics of patients with clinical stage I adenocarcinoma with tumor ≤30 mm (n=390)
Characteristic | VPI absent (n=340) | VPI present (n=50) | p-value |
---|---|---|---|
Age (yr) | 64.3±10.2 | 66.7±10.0 | 0.025 |
Sex | 0.480 | ||
Female | 212 (62.4) | 28 (56.0) | |
Male | 128 (37.6) | 22 (44.0) | |
Smoking status | 0.330 | ||
Never | 225 (66.2) | 29 (58.0) | |
Ever | 115 (33.8) | 21 (42.0) | |
Pleural CEAa) | <0.001 | ||
Normal | 276 (81.2) | 21 (42.0) | |
Elevated | 64 (18.8) | 29 (58.0) | |
Serum CEAb) | 0.119 | ||
Normal | 285 (84.1) | 37 (74.0) | |
Elevated | 54 (15.9) | 13 (26.0) | |
SUVmaxc) | <0.001 | ||
Normal | 162 (69.8) | 9 (18.8) | |
Elevated | 70 (30.2) | 39 (81.2) | |
Extent of resection | 0.003 | ||
Sublobar | 139 (40.9) | 9 (18.0) | |
Lobectomy | 201 (59.1) | 41 (82.0) | |
Tumor size (mm) | <0.001 | ||
≤10 | 159 (46.8) | 0 | |
>10, ≤20 | 120 (35.3) | 22 (44.0) | |
>20, ≤30 | 61 (17.9) | 28 (56.0) | |
Nodule type | <0.001 | ||
Pure GGN | 84 (24.7) | 0 | |
Part solid | 143 (42.1) | 12 (24.0) | |
Pure solid | 113 (33.2) | 38 (76.0) | |
Pleural contact on CT | <0.001 | ||
Pleural contact | 148 (43.5) | 38 (76.0) | |
Non-pleural contact | 192 (56.5) | 12 (24.0) | |
Nodal status | <0.001 | ||
N0 | 324 (95.3) | 39 (78.0) | |
N1–N2 | 16 (4.7) | 11 (22.0) | |
Pathological TNM stage | <0.001 | ||
Stage I | 320 (95.0) | 35 (71.4) | |
Stage II | 10 (3.0) | 3 (6.1) | |
Stage III | 7 (2.1) | 8 (16.3) | |
Stage IV | 0 | 3 (6.1) | |
Predominant subtype | <0.001 | ||
Lepidic | 152 (44.7) | 4 (8.0) | |
Acinar | 154 (45.3) | 33 (66.0) | |
Papillary | 16 (4.7) | 2 (4.0) | |
Solid | 9 (2.6) | 5 (10.0) | |
Micropapillary | 2 (0.6) | 4 (8.0) | |
Cribriform | 1 (0.3) | 1 (2.0) | |
Mucinous | 5 (1.5) | 0 | |
Mixed | 1 (0.3) | 1 (2.0) |
Values are presented as mean±standard deviation or number (%).
VPI, visceral pleural invasion; CEA, carcinoembryonic antigen; SUVmax, maximum standardized uptake value; GGN, ground-glass nodule; CT, computed tomography; TNM, tumor-node-metastasis.
a)The cutoff values for pleural CEA were 2.56 ng/mL for the pleural contact group and 2.105 ng/mL for the non-pleural contact group. b)The cutoff values for serum CEA were 3.8 ng/mL for never-smokers and 5.5 ng/mL for ever-smokers. Preoperative serum CEA levels were available for 389 patients at our hospital. c)The cutoff values of SUVmax were 4.25 for the pleural contact group and 4.65 for the non-pleural contact group. SUVmax values were available for 280 patients at our hospital.
Of a total of 390 patients, 186 individuals who exhibited direct pleural contact on CT were evaluated regarding the diagnostic performance of pCEA and SUVmax in predicting VPI. The ROC curve for our model is illustrated in Fig. 4. pCEA was associated with an AUC of 0.751, with a threshold value of 2.565 ng/mL (sensitivity, 63.2%; specificity, 77.7%; positive predictive value [PPV], 10.9%; negative predictive value [NPV], 57.9%). SUVmax demonstrated an AUC of 0.801, with a threshold level of 4.25 (sensitivity, 77.8%; specificity, 78.1%; PPV, 8.9%; NPV, 45.1%).
Table 2 presents the characteristics of the 186 patients who exhibited pleural contact. Of these, 38 patients (20.4%) exhibited pathologic VPI, while 148 patients (79.6%) did not. Univariate analysis revealed that elevated pCEA level, SUVmax, tumor size, and nodule type were significant risk factors for VPI (p<0.001). Similarly, in the multivariate analysis, pCEA (odds ratio [OR], 3.00; 95% confidence interval [CI], 1.14–7.87; p=0.0256), SUVmax (OR, 5.25; 95% CI, 1.90–14.50; p=0.0014), and nodule type (OR, 3.89; 95% CI, 1.42–10.68; p=0.0084) were identified as independent risk factors for VPI.
Table 2. Factors associated with VPI in patients with CT findings of pleural contact (n=186)
Variable | VPI | Univariate analysis | Multivariate analysis | ||||||
---|---|---|---|---|---|---|---|---|---|
Absent | Present | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | |||
Total | 148 (79.6) | 38 (20.4) | |||||||
Pleural CEAa) | <0.001 | ||||||||
Normal | 115 (77.7) | 14 (36.8) | Ref | Ref | |||||
Elevated | 33 (22.3) | 24 (63.2) | 3.64 (1.25–10.64) | 0.018 | 3.00 (1.14–7.87) | 0.026 | |||
Age (yr) | 64.6±10.0 | 67.2±10.9 | 0.176 | 1.01 (0.97–1.06) | 0.579 | ||||
Sex | 0.239 | ||||||||
Female | 103 (69.6) | 22 (57.9) | Ref | Ref | |||||
Male | 45 (30.4) | 16 (42.1) | 4.45 (0.79–25.2) | 0.091 | 2.17 (0.79–5.99) | 0.135 | |||
Smoking status | 0.282 | ||||||||
Never | 109 (73.6) | 24 (63.2) | Ref | ||||||
Ever | 39 (26.4) | 14 (36.8) | 0.37 (0.06–2.26) | 0.283 | |||||
Serum CEAb) | 0.024 | ||||||||
Normal | 127 (85.8) | 26 (68.4) | Ref | ||||||
Elevated | 21 (14.2) | 12 (31.6) | 0.42 (0.12–1.53) | 0.188 | |||||
SUVmaxc) | <0.001 | ||||||||
Normal | 82 (78.1) | 8 (22.2) | Ref | Ref | |||||
Elevated | 23 (21.9) | 28 (77.8) | 6.32 (2.08–19.20) | 0.001 | 5.25 (1.90–14.50) | 0.001 | |||
Tumor size (mm) | <0.001 | 1.03 (0.45–2.35) | 0.946 | ||||||
≤10 | 69 (46.6) | 0 | |||||||
>10, ≤20 | 48 (32.4) | 19 (50.0) | |||||||
>20, ≤30 | 31 (20.9) | 19 (50.0) | |||||||
Nodule type | <0.001 | 4.37 (1.46–13.11) | 0.008 | 3.89 (1.42–10.68) | 0.008 | ||||
Pure GGN | 29 (19.6) | 0 | |||||||
Part solid | 70 (47.3) | 8 (21.1) | |||||||
Pure solid | 49 (33.1) | 30 (78.9) |
Values are presented as number (%) or mean±standard deviation, unless otherwise stated.
VPI, visceral pleural invasion; CT, computed tomography; OR, odds ratio; CI, confidence interval; CEA, carcinoembryonic antigen; Ref, reference; SUVmax, maximum standardized uptake value; GGN, ground-glass nodule.
a)The cutoff value for pleural CEA was 2.565 ng/mL. b)The cutoff values for serum CEA were 3.8 ng/mL for never-smokers and 5.5 ng/mL for ever-smokers. c)The cutoff value for SUVmax was 4.25 SUVmax values were available for 141 patients at our hospital.
Table 3 presents an analysis of the association between risk factors and VPI within the non-pleural contact group. Of 204 patients, 12 (5.9%) exhibited VPI, while 192 (94.1%) did not. Based on multivariate logistic regression, tumor size (OR, 4.66; 95% CI, 1.16–18.69; p=0.0298) emerged as the sole statistically significant predictor of VPI.
Table 3. Factors associated with VPI in patients with CT findings of non-pleural contact (n=204)
Variable | VPI | Univariate analysis | Multivariate analysis | ||||||
---|---|---|---|---|---|---|---|---|---|
Absent | Present | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | |||
Total | 192 (94.1) | 12 (5.89) | |||||||
Pleural CEAa) | 0.030 | ||||||||
Normal | 155 (80.7) | 6 (50.0) | Ref | ||||||
Elevated | 37 (19.3) | 6 (50.0) | 2.03 (0.44–9.34) | 0.365 | |||||
Age (yr) | 64.0±10.3 | 69.7±9.1 | 0.064 | 1.04 (0.97–1.12) | 0.280 | 1.05 (0.99–1.11) | 0.078 | ||
Sex | 0.874 | ||||||||
Female | 109 (56.8) | 6 (50.0) | Ref | ||||||
Male | 83 (43.2) | 6 (50.0) | 0.58 (0.07–4.77) | 0.616 | |||||
Smoking status | 0.327 | ||||||||
Never | 116 (60.4) | 5 (41.7) | Ref | ||||||
Ever | 76 (39.6) | 7 (58.3) | 2.08 (0.28–15.19) | 0.470 | |||||
Serum CEAb) | 0.684 | ||||||||
Normal | 158 (82.7) | 11 (91.7) | Ref | ||||||
Elevated | 33 (17.3) | 1 (8.3) | 0.21 (0.02–2.15) | 0.187 | |||||
SUVmaxc) | <0.001 | ||||||||
Normal | 98 (77.2) | 3 (25.0) | Ref | Ref | |||||
Elevated | 29 (22.8) | 9 (75.0) | 3.37 (0.52–21.9) | 0.202 | 3.47 (0.74–16.42) | 0.116 | |||
Tumor size (mm) | <0.001 | 4.31 (0.84–22.1) | 0.079 | 4.66 (1.16–18.69) | 0.030 | ||||
≤10 | 90 (46.9) | 0 | |||||||
>10, ≤20 | 72 (37.5) | 3 (25.0) | |||||||
>20, ≤30 | 30 (15.6) | 9 (75.0) | |||||||
Nodule type | 0.029 | 0.81 (0.17–3.77) | 0.790 | ||||||
Pure GGN | 55 (28.6) | 0 | |||||||
Part solid | 73 (38.0) | 4 (33.3) | |||||||
Pure solid | 64 (33.3) | 8 (66.7) |
Values are presented as number (%) or mean±standard deviation, unless otherwise stated.
VPI, visceral pleural invasion; CT, computed tomography; OR, odds ratio; CI, confidence interval; CEA, carcinoembryonic antigen; Ref, reference; SUVmax, maximum standardized uptake value; GGN, ground-glass nodule.
a)The cutoff value for pleural CEA was 2.105 ng/mL. b)The cutoff values for serum CEA were 3.8 ng/mL for never-smokers and 5.5 ng/mL for ever-smokers. c)The cutoff value for SUVmax was 4.65 SUVmax values were available for 139 patients at our hospital.
VPI is a potent prognostic factor in small NSCLC and contributes to the upstaging of tumors. However, the preoperative assessment of VPI is challenging and heavily dependent on radiological examination. In the present study, intraoperative pCEA and SUVmax demonstrated high diagnostic value for predicting VPI in early-stage lung adenocarcinomas, particularly in patients exhibiting direct pleural contact.
Several studies have noted the association between SUVmax and VPI in NSCLC. Elevated uptake of 18F-FDG is correlated with increased tumor invasiveness, which is also linked to VPI [14,16]. Zhang et al. [17] observed that PET-CT, which provides both structural and functional information, is relatively intuitive and easy to interpret, making it a more suitable approach than CT imaging. Our study also revealed a significant correlation between SUVmax and VPI through multivariate analysis. High diagnostic value was demonstrated in the prediction of VPI, with a cutoff value of 4.25. We were able to review a larger number of patients than in previous studies because our institution routinely recommends PET-CT for patients suspected of having primary lung cancer. However, not all institutions perform PET exams preoperatively. Therefore, assessing pCEA may be a useful method for predicting VPI in these patients.
CEA is widely used as a marker to predict prognosis, recurrence, and treatment efficacy in patients with NSCLC [18]. In addition to serum CEA, the potential role of pCEA has been investigated, particularly in the diagnosis of malignant effusion [19-22]. While pCEA has demonstrated high specificity in predicting malignant effusion, a recent meta-analysis has indicated that this marker alone may be insufficient due to its low sensitivity. This review further emphasized that overall accuracy could be improved by combining various tumor markers [20]. In the present study, we measured the pCEA level through VATS, a method that is generally more accurate than sampling from thoracentesis. To our knowledge, this is the first study to examine the role of pCEA in predicting VPI among patients with early-stage lung adenocarcinoma.
The relationship between elevated pCEA levels and pathologic VPI is currently not well understood. CEA, a glycoprotein involved in cell adhesion, is typically produced during fetal development and ceases production before birth. However, it has been found to be overexpressed in many malignancies, particularly colorectal cancers and breast and lung adenocarcinomas [18]. Therefore, the increased release of these cell-surface proteins from tumors located at the pleural surface, or their leakage through the lymphatic system, could potentially explain the elevated pCEA levels in NSCLC patients with VPI [23]. This micrometastasis can contribute to poor prognosis in small lung adenocarcinoma. Consequently, the evaluation of VPI using intraoperative pCEA could be beneficial in determining the extent of lung resection required during surgery. It also serves as an efficient diagnostic tool in terms of cost, time, and performance.
This study had several limitations. First, the retrospective design inherently introduced the potential for selection bias, as the patient inclusion process may have been influenced by specific factors. Second, the differentiation of the groups based on the presence of pleural contact was not performed by radiologists. Instead, to minimize the bias associated with diverse and detailed radiographic interpretations, multiple cardiothoracic surgeons re-evaluated the CT images and employed a simple categorization method to determine the presence or absence of direct pleural contact. Third, the examination of a wider range of tumor markers could have yielded a more thorough analysis. However, we chose CEA due to its relatively high sensitivity and specificity compared to other markers, a finding supported by prior studies. Fourth, we did not investigate the prognosis or long-term survival outcomes of the patients, primarily due to insufficient follow-up duration. Therefore, future data collection will allow us to assess survival outcomes and prognosis. Finally, to validate these results, further prospective studies with larger sample sizes or involving multiple institutions are required.
The findings of this study underscore the diagnostic value of pCEA and SUVmax for predicting VPI in clinical T1N0M0 lung adenocarcinoma, especially in instances involving direct pleural contact. This could aid clinicians in making informed decisions about the extent of lung resection required for patients with stage I lung adenocarcinoma in which VPI is suspected.
Author contributions
Conceptualization: HRN, SHM, SKY. Data curation: SHM, KSK, MHM, KYH, SKY. Formal analysis: HRN, SKY. Funding acquisition: not applicable. Investigation: HRN, SHM, KSK, MHM, KYH, SKY. Methodology: SHM, KYH, SKY. Project administration: SHM, SKY. Resources: SHM, KSK, MHM, KYH, SKY. Software: SKY. Supervision: SHM, SKY. Validation: SHM, SKY. Visualization: HRN, SKY. Writing–original draft: HRN. Writing–review & editing: HRN, SKY. Final approval of the manuscript: all authors.
Conflict of interest
No potential conflict of interest relevant to this article was reported.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
J Chest Surg 2024; 57(1): 44-52
Published online January 5, 2024 https://doi.org/10.5090/jcs.23.094
Copyright © Journal of Chest Surgery.
Hye Rim Na , M.D.1, Seok Whan Moon
, M.D., Ph.D.1, Kyung Soo Kim
, M.D., Ph.D.1, Mi Hyoung Moon
, M.D., Ph.D.1, Kwanyong Hyun
, M.D.2, Seung Keun Yoon
, M.D.1
1Department of Thoracic and Cardiovascular Surgery, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul; 2Department of Thoracic and Cardiovascular Surgery, St. Vincent’s Hospital, College of Medicine, The Catholic University of Korea, Suwon, Korea
Correspondence to:Seung Keun Yoon
Tel 82-2-2258-2858
Fax 82-2-594-8644
E-mail skycs@catholic.ac.kr
ORCID
https://orcid.org/0000-0002-2609-2148
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background: Visceral pleural invasion (VPI) is a poor prognostic factor that contributes to the upstaging of early lung cancers. However, the preoperative assessment of VPI presents challenges. This study was conducted to examine intraoperative pleural carcinoembryonic antigen (pCEA) level and maximum standardized uptake value (SUVmax) as predictive markers of VPI in patients with clinical T1N0M0 lung adenocarcinoma.
Methods: A retrospective review was conducted of the medical records of 613 patients who underwent intraoperative pCEA sampling and lung resection for non-small cell lung cancer. Of these, 390 individuals with clinical stage I adenocarcinoma and tumors ≤30 mm were included. Based on computed tomography findings, these patients were divided into pleural contact (n=186) and non-pleural contact (n=204) groups. A receiver operating characteristic (ROC) curve was constructed to analyze the association between pCEA and SUVmax in relation to VPI. Additionally, logistic regression analysis was performed to evaluate risk factors for VPI in each group.
Results: ROC curve analysis revealed that pCEA level greater than 2.565 ng/mL (area under the curve [AUC]=0.751) and SUVmax above 4.25 (AUC=0.801) were highly predictive of VPI in patients exhibiting pleural contact. Based on multivariable analysis, pCEA (odds ratio [OR], 3.00; 95% confidence interval [CI], 1.14–7.87; p=0.026) and SUVmax (OR, 5.25; 95% CI, 1.90–14.50; p=0.001) were significant risk factors for VPI in the pleural contact group.
Conclusion: In patients with clinical stage I lung adenocarcinoma exhibiting pleural contact, pCEA and SUVmax are potential predictive indicators of VPI. These markers may be helpful in planning for lung cancer surgery.
Keywords: Adenocarcinoma of lung, Visceral pleural invasion, Pleural carcinoembryonic antigen, Maximum standardized uptake value, Thoracic surgery
Lung cancer is a leading cause of cancer-related deaths, with a 5-year relative survival rate of 22% [1]. Advances in screening examinations have facilitated the detection of smaller and less invasive cancers, thereby increasing the likelihood of surgical treatment. However, for patients eligible for surgery, debate remains as to whether to perform sublobar or lobar resection for early-stage non-small cell lung cancer (NSCLC). A recent randomized controlled trial demonstrated that segmentectomy is not inferior to lobectomy in terms of overall survival in clinical stage IA peripheral NSCLC. However, that study also reported a twofold increase in locoregional relapse following segmentectomy [2]. Consequently, careful selection of the most suitable surgical approach for patients is crucial to minimize the risk of recurrence in stage IA lung cancers.
Extensive research has been conducted into factors that contribute to the invasiveness and recurrence of small-size lung cancers. A retrospective study involving 327 patients with stage IA NSCLC who underwent sublobar resection identified lymphatic permeation, microscopic positive surgical margin, and visceral pleural invasion (VPI) as significant predictors of locoregional recurrence [3]. In the eighth edition of the tumor-node-metastasis (TNM) classification system, VPI is considered a T2 descriptor even for tumors smaller than 3 cm and results in the upstaging of IA tumors to IB [4,5]. Pathological VPI is characterized by the extension of the tumor beyond the elastic layer of the pleura. According to the modified Hammar classification, pleural invasion can be categorized into 4 stages, from PL0 to PL3 [6]. PL0 signifies the absence of VPI, with the tumor located in the subpleural lung tissue, just beneath the elastic layer. PL1 represents an extension beyond the elastic layer, while PL2 indicates further progression that reaches the pleural surface. Tumors invading any part of the parietal pleura are classified as PL3. Of these categories, PL1 and PL2 denote VPI, suggesting potential invasion into the pleural space.
Unlike pathologic evaluation, the clinical diagnosis of VPI has not been clearly established. Some researchers have proposed combinations of features of the tumor-pleura relationship on computed tomography (CT). These features include direct pleural contact of the tumor, pleural retraction, and pleural tags [7-10]. However, a study by Kim et al. [7], which examined CT-defined VPI and its correlation with disease-free survival in a retrospective review of 695 patients, revealed that approximately one-half of the findings were false positives. Consequently, the accuracy of these findings ranged from 62.7% to 72.3%, suggesting a low level of reliability as an independent predictor [7].
Multiple separate studies have proposed alternative methods for the preoperative assessment of VPI, such as the measurement of circulating tumor cells from serum and pre-resectional pleural lavage tumor cytology. However, these approaches require further research for validation [8,11]. The maximum standardized uptake value (SUVmax) and serum carcinoembryonic antigen (CEA) level have also been identified as significant risk factors for VPI, but no definite cutoff values have been established for diagnosis [12-14]. In the present study, given the close association of VPI with invasion into the pleural cavity, we hypothesized that tumor markers in the pleural fluid could serve as predictive markers for VPI. To our knowledge, no investigations have yet been published regarding the association between pleural CEA (pCEA) and VPI in early-stage adenocarcinomas.
This study involved a retrospective review of patients with clinical stage I adenocarcinoma, with tumors measuring 30 mm or less, who underwent intraoperative pCEA sampling prior to lung resection. The primary objectives were to evaluate the predictive value of pCEA and SUVmax in the clinical diagnosis of VPI and to ascertain whether these parameters could aid in decision-making regarding lung surgery.
A retrospective review was conducted of the electronic medical records of patients who underwent intraoperative pCEA collection followed by surgical resection for NSCLC at Seoul St. Mary’s Hospital in Seoul, Korea. The review covered a study period from January 2017 to December 2022. The exclusion criteria included patients with clinical stage II to IV adenocarcinoma as confirmed by radiological imaging, those with pathologically confirmed non-adenocarcinoma, and those who had previously undergone lung resection. Among the patients with clinical stage I lung adenocarcinoma, those with a tumor larger than 30 mm were additionally excluded (Fig. 1). This study received approval from the institutional review board at Seoul St. Mary’s Hospital (approval no., KC23RASI0381), and informed consent was waived due to the retrospective nature of the study.
Preoperative CT images were captured using a Siemens device (Somatom, Erlangen, Germany). These images were then reconstructed in axial, coronal, and sagittal planes, with a section thickness of either 1 or 3 mm. All measurements were conducted using a lung window setting, which had a window level of 600 Hounsfield units (HU) and a window width of 1,600 HU. Within this lung window setting, pulmonary tumors that directly contacted the pleural surface or interlobar fissure were classified as pleural contact tumors. Tumors that did not meet this criterion were categorized as non-pleural contact tumors (Fig. 2). Furthermore, tumors exhibiting signs of a pleural tag, defined as linear soft tissue strands linking the tumor and the pleural surface, were included in the non-pleural contact category. Preoperative 18F-fluorodeoxyglucose (18F-FDG) positive emission tomography/computed tomography (PET/CT) images were acquired using a Discovery 710 device (GE Healthcare, Milwaukee, WI, USA). All patients were required to fast for at least 6 hours prior to examination. Images were then captured 60 minutes after intravenous injection of 18F-FDG (0.12 mCi/kg) and at intervals of 1.5 to 2 minutes per bed position.
Patients diagnosed with clinical stage IA lung cancer were considered eligible for surgical treatment. The standard surgical approach involved performing a lobectomy accompanied by systemic lymph node dissection. However, for patients at low risk, with peripheral tumors either measuring less than 2 cm or exhibiting more than 50% ground-glass opacity, sublobar resection was considered in line with the National Comprehensive Cancer Network guidelines [15]. For these patients, lobe-specific lymph node dissection or sampling were also potential options. Surgical procedures were performed using videoscopic-assisted thoracoscopic surgery (VATS), with the number of incisions ranging from 2 to 4, based on the surgeon’s preference. Following skin incision and visualization of the thoracic cavity, a minimum of 2 mL of pleural effusion fluid was collected for CEA analysis before any other lung manipulation. Pleural effusion samples were most frequently obtained from the costodiaphragmatic recess beneath the inferior pulmonary ligament (Fig. 3). However, pCEA sampling was not conducted in patients with diffuse lung adhesion or in cases in which the pleural fluid was contaminated by massive bleeding. The measurement of pCEA was performed using an electrochemiluminescence immunoassay with the Elecsys CEA kit (Roche Diagnostics, Penzberg, Germany), and results were typically available approximately 30 minutes after sampling.
The resected lung specimens were fixed in formalin, deparaffinized, and stained with hematoxylin-eosin. Elastic staining was employed for the pathological confirmation of VPI. The final report incorporated details including tumor size, histological type, differentiation, pleural invasion, lymphovascular invasion, resection margin, nodal status, pathologic stage, elastic stain, and immunohistochemistry. The presence of VPI was determined using the modified Hammar classification system, as follows: PL0 represented a tumor within the subpleural lung parenchyma or superficially invading the pleural connective tissue below the elastic layer, PL1 denoted tumor invasion beyond the elastic layer, PL2 represented invasion reaching the pleural surface, and PL3 indicated invasion of the parietal pleura. The PL1 and PL2 categories describe VPI and are incorporated within T2 staging [6].
Data representing baseline characteristics are presented as frequencies with percentages for categorical variables and as mean±standard deviation for continuous variables. The Mann-Whitney U test was employed for comparisons of continuous variables, while the chi-square or Fisher exact tests were used for categorical variables. The association between variables and VPI was assessed using receiver operating curve (ROC) analysis, with the area under the curve (AUC) calculated for each index. Logistic regression was applied to analyze risk factors for VPI, and those with p-values of less than 0.2 were included in the multivariable analysis. All statistical computations were performed using R ver. 4.0.4 (R Foundation for Statistical Computing, Vienna, Austria), utilizing the Epi, ggplot2, and moonBook packages. A p-value of less than 0.05 was considered to indicate statistical significance.
The study population comprised 390 patients with clinical stage I adenocarcinoma, with tumors measuring 30 mm or less, who underwent pulmonary resection. Of these, 50 patients exhibited pathologic VPI, while the remaining 340 did not. A significant difference was observed in the proportion of patients with elevated pCEA levels between the VPI group (58.0%) and the non-VPI group (18.8%) (p<0.001). Elevated SUVmax was also noted more frequently in the VPI group (81.2%) compared to the non-VPI group (30.2%) (p<0.001). Moreover, patients with pathologic VPI showed significant differences in the extent of lung resection, tumor size, nodule type, pleural contact on CT, nodal status, pathological TNM stage, and predominant histological subtype. Table 1 details the characteristics of the study population.
Table 1 . Clinicopathologic characteristics of patients with clinical stage I adenocarcinoma with tumor ≤30 mm (n=390).
Characteristic | VPI absent (n=340) | VPI present (n=50) | p-value |
---|---|---|---|
Age (yr) | 64.3±10.2 | 66.7±10.0 | 0.025 |
Sex | 0.480 | ||
Female | 212 (62.4) | 28 (56.0) | |
Male | 128 (37.6) | 22 (44.0) | |
Smoking status | 0.330 | ||
Never | 225 (66.2) | 29 (58.0) | |
Ever | 115 (33.8) | 21 (42.0) | |
Pleural CEAa) | <0.001 | ||
Normal | 276 (81.2) | 21 (42.0) | |
Elevated | 64 (18.8) | 29 (58.0) | |
Serum CEAb) | 0.119 | ||
Normal | 285 (84.1) | 37 (74.0) | |
Elevated | 54 (15.9) | 13 (26.0) | |
SUVmaxc) | <0.001 | ||
Normal | 162 (69.8) | 9 (18.8) | |
Elevated | 70 (30.2) | 39 (81.2) | |
Extent of resection | 0.003 | ||
Sublobar | 139 (40.9) | 9 (18.0) | |
Lobectomy | 201 (59.1) | 41 (82.0) | |
Tumor size (mm) | <0.001 | ||
≤10 | 159 (46.8) | 0 | |
>10, ≤20 | 120 (35.3) | 22 (44.0) | |
>20, ≤30 | 61 (17.9) | 28 (56.0) | |
Nodule type | <0.001 | ||
Pure GGN | 84 (24.7) | 0 | |
Part solid | 143 (42.1) | 12 (24.0) | |
Pure solid | 113 (33.2) | 38 (76.0) | |
Pleural contact on CT | <0.001 | ||
Pleural contact | 148 (43.5) | 38 (76.0) | |
Non-pleural contact | 192 (56.5) | 12 (24.0) | |
Nodal status | <0.001 | ||
N0 | 324 (95.3) | 39 (78.0) | |
N1–N2 | 16 (4.7) | 11 (22.0) | |
Pathological TNM stage | <0.001 | ||
Stage I | 320 (95.0) | 35 (71.4) | |
Stage II | 10 (3.0) | 3 (6.1) | |
Stage III | 7 (2.1) | 8 (16.3) | |
Stage IV | 0 | 3 (6.1) | |
Predominant subtype | <0.001 | ||
Lepidic | 152 (44.7) | 4 (8.0) | |
Acinar | 154 (45.3) | 33 (66.0) | |
Papillary | 16 (4.7) | 2 (4.0) | |
Solid | 9 (2.6) | 5 (10.0) | |
Micropapillary | 2 (0.6) | 4 (8.0) | |
Cribriform | 1 (0.3) | 1 (2.0) | |
Mucinous | 5 (1.5) | 0 | |
Mixed | 1 (0.3) | 1 (2.0) |
Values are presented as mean±standard deviation or number (%)..
VPI, visceral pleural invasion; CEA, carcinoembryonic antigen; SUVmax, maximum standardized uptake value; GGN, ground-glass nodule; CT, computed tomography; TNM, tumor-node-metastasis..
a)The cutoff values for pleural CEA were 2.56 ng/mL for the pleural contact group and 2.105 ng/mL for the non-pleural contact group. b)The cutoff values for serum CEA were 3.8 ng/mL for never-smokers and 5.5 ng/mL for ever-smokers. Preoperative serum CEA levels were available for 389 patients at our hospital. c)The cutoff values of SUVmax were 4.25 for the pleural contact group and 4.65 for the non-pleural contact group. SUVmax values were available for 280 patients at our hospital..
Of a total of 390 patients, 186 individuals who exhibited direct pleural contact on CT were evaluated regarding the diagnostic performance of pCEA and SUVmax in predicting VPI. The ROC curve for our model is illustrated in Fig. 4. pCEA was associated with an AUC of 0.751, with a threshold value of 2.565 ng/mL (sensitivity, 63.2%; specificity, 77.7%; positive predictive value [PPV], 10.9%; negative predictive value [NPV], 57.9%). SUVmax demonstrated an AUC of 0.801, with a threshold level of 4.25 (sensitivity, 77.8%; specificity, 78.1%; PPV, 8.9%; NPV, 45.1%).
Table 2 presents the characteristics of the 186 patients who exhibited pleural contact. Of these, 38 patients (20.4%) exhibited pathologic VPI, while 148 patients (79.6%) did not. Univariate analysis revealed that elevated pCEA level, SUVmax, tumor size, and nodule type were significant risk factors for VPI (p<0.001). Similarly, in the multivariate analysis, pCEA (odds ratio [OR], 3.00; 95% confidence interval [CI], 1.14–7.87; p=0.0256), SUVmax (OR, 5.25; 95% CI, 1.90–14.50; p=0.0014), and nodule type (OR, 3.89; 95% CI, 1.42–10.68; p=0.0084) were identified as independent risk factors for VPI.
Table 2 . Factors associated with VPI in patients with CT findings of pleural contact (n=186).
Variable | VPI | Univariate analysis | Multivariate analysis | ||||||
---|---|---|---|---|---|---|---|---|---|
Absent | Present | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | |||
Total | 148 (79.6) | 38 (20.4) | |||||||
Pleural CEAa) | <0.001 | ||||||||
Normal | 115 (77.7) | 14 (36.8) | Ref | Ref | |||||
Elevated | 33 (22.3) | 24 (63.2) | 3.64 (1.25–10.64) | 0.018 | 3.00 (1.14–7.87) | 0.026 | |||
Age (yr) | 64.6±10.0 | 67.2±10.9 | 0.176 | 1.01 (0.97–1.06) | 0.579 | ||||
Sex | 0.239 | ||||||||
Female | 103 (69.6) | 22 (57.9) | Ref | Ref | |||||
Male | 45 (30.4) | 16 (42.1) | 4.45 (0.79–25.2) | 0.091 | 2.17 (0.79–5.99) | 0.135 | |||
Smoking status | 0.282 | ||||||||
Never | 109 (73.6) | 24 (63.2) | Ref | ||||||
Ever | 39 (26.4) | 14 (36.8) | 0.37 (0.06–2.26) | 0.283 | |||||
Serum CEAb) | 0.024 | ||||||||
Normal | 127 (85.8) | 26 (68.4) | Ref | ||||||
Elevated | 21 (14.2) | 12 (31.6) | 0.42 (0.12–1.53) | 0.188 | |||||
SUVmaxc) | <0.001 | ||||||||
Normal | 82 (78.1) | 8 (22.2) | Ref | Ref | |||||
Elevated | 23 (21.9) | 28 (77.8) | 6.32 (2.08–19.20) | 0.001 | 5.25 (1.90–14.50) | 0.001 | |||
Tumor size (mm) | <0.001 | 1.03 (0.45–2.35) | 0.946 | ||||||
≤10 | 69 (46.6) | 0 | |||||||
>10, ≤20 | 48 (32.4) | 19 (50.0) | |||||||
>20, ≤30 | 31 (20.9) | 19 (50.0) | |||||||
Nodule type | <0.001 | 4.37 (1.46–13.11) | 0.008 | 3.89 (1.42–10.68) | 0.008 | ||||
Pure GGN | 29 (19.6) | 0 | |||||||
Part solid | 70 (47.3) | 8 (21.1) | |||||||
Pure solid | 49 (33.1) | 30 (78.9) |
Values are presented as number (%) or mean±standard deviation, unless otherwise stated..
VPI, visceral pleural invasion; CT, computed tomography; OR, odds ratio; CI, confidence interval; CEA, carcinoembryonic antigen; Ref, reference; SUVmax, maximum standardized uptake value; GGN, ground-glass nodule..
a)The cutoff value for pleural CEA was 2.565 ng/mL. b)The cutoff values for serum CEA were 3.8 ng/mL for never-smokers and 5.5 ng/mL for ever-smokers. c)The cutoff value for SUVmax was 4.25 SUVmax values were available for 141 patients at our hospital..
Table 3 presents an analysis of the association between risk factors and VPI within the non-pleural contact group. Of 204 patients, 12 (5.9%) exhibited VPI, while 192 (94.1%) did not. Based on multivariate logistic regression, tumor size (OR, 4.66; 95% CI, 1.16–18.69; p=0.0298) emerged as the sole statistically significant predictor of VPI.
Table 3 . Factors associated with VPI in patients with CT findings of non-pleural contact (n=204).
Variable | VPI | Univariate analysis | Multivariate analysis | ||||||
---|---|---|---|---|---|---|---|---|---|
Absent | Present | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | |||
Total | 192 (94.1) | 12 (5.89) | |||||||
Pleural CEAa) | 0.030 | ||||||||
Normal | 155 (80.7) | 6 (50.0) | Ref | ||||||
Elevated | 37 (19.3) | 6 (50.0) | 2.03 (0.44–9.34) | 0.365 | |||||
Age (yr) | 64.0±10.3 | 69.7±9.1 | 0.064 | 1.04 (0.97–1.12) | 0.280 | 1.05 (0.99–1.11) | 0.078 | ||
Sex | 0.874 | ||||||||
Female | 109 (56.8) | 6 (50.0) | Ref | ||||||
Male | 83 (43.2) | 6 (50.0) | 0.58 (0.07–4.77) | 0.616 | |||||
Smoking status | 0.327 | ||||||||
Never | 116 (60.4) | 5 (41.7) | Ref | ||||||
Ever | 76 (39.6) | 7 (58.3) | 2.08 (0.28–15.19) | 0.470 | |||||
Serum CEAb) | 0.684 | ||||||||
Normal | 158 (82.7) | 11 (91.7) | Ref | ||||||
Elevated | 33 (17.3) | 1 (8.3) | 0.21 (0.02–2.15) | 0.187 | |||||
SUVmaxc) | <0.001 | ||||||||
Normal | 98 (77.2) | 3 (25.0) | Ref | Ref | |||||
Elevated | 29 (22.8) | 9 (75.0) | 3.37 (0.52–21.9) | 0.202 | 3.47 (0.74–16.42) | 0.116 | |||
Tumor size (mm) | <0.001 | 4.31 (0.84–22.1) | 0.079 | 4.66 (1.16–18.69) | 0.030 | ||||
≤10 | 90 (46.9) | 0 | |||||||
>10, ≤20 | 72 (37.5) | 3 (25.0) | |||||||
>20, ≤30 | 30 (15.6) | 9 (75.0) | |||||||
Nodule type | 0.029 | 0.81 (0.17–3.77) | 0.790 | ||||||
Pure GGN | 55 (28.6) | 0 | |||||||
Part solid | 73 (38.0) | 4 (33.3) | |||||||
Pure solid | 64 (33.3) | 8 (66.7) |
Values are presented as number (%) or mean±standard deviation, unless otherwise stated..
VPI, visceral pleural invasion; CT, computed tomography; OR, odds ratio; CI, confidence interval; CEA, carcinoembryonic antigen; Ref, reference; SUVmax, maximum standardized uptake value; GGN, ground-glass nodule..
a)The cutoff value for pleural CEA was 2.105 ng/mL. b)The cutoff values for serum CEA were 3.8 ng/mL for never-smokers and 5.5 ng/mL for ever-smokers. c)The cutoff value for SUVmax was 4.65 SUVmax values were available for 139 patients at our hospital..
VPI is a potent prognostic factor in small NSCLC and contributes to the upstaging of tumors. However, the preoperative assessment of VPI is challenging and heavily dependent on radiological examination. In the present study, intraoperative pCEA and SUVmax demonstrated high diagnostic value for predicting VPI in early-stage lung adenocarcinomas, particularly in patients exhibiting direct pleural contact.
Several studies have noted the association between SUVmax and VPI in NSCLC. Elevated uptake of 18F-FDG is correlated with increased tumor invasiveness, which is also linked to VPI [14,16]. Zhang et al. [17] observed that PET-CT, which provides both structural and functional information, is relatively intuitive and easy to interpret, making it a more suitable approach than CT imaging. Our study also revealed a significant correlation between SUVmax and VPI through multivariate analysis. High diagnostic value was demonstrated in the prediction of VPI, with a cutoff value of 4.25. We were able to review a larger number of patients than in previous studies because our institution routinely recommends PET-CT for patients suspected of having primary lung cancer. However, not all institutions perform PET exams preoperatively. Therefore, assessing pCEA may be a useful method for predicting VPI in these patients.
CEA is widely used as a marker to predict prognosis, recurrence, and treatment efficacy in patients with NSCLC [18]. In addition to serum CEA, the potential role of pCEA has been investigated, particularly in the diagnosis of malignant effusion [19-22]. While pCEA has demonstrated high specificity in predicting malignant effusion, a recent meta-analysis has indicated that this marker alone may be insufficient due to its low sensitivity. This review further emphasized that overall accuracy could be improved by combining various tumor markers [20]. In the present study, we measured the pCEA level through VATS, a method that is generally more accurate than sampling from thoracentesis. To our knowledge, this is the first study to examine the role of pCEA in predicting VPI among patients with early-stage lung adenocarcinoma.
The relationship between elevated pCEA levels and pathologic VPI is currently not well understood. CEA, a glycoprotein involved in cell adhesion, is typically produced during fetal development and ceases production before birth. However, it has been found to be overexpressed in many malignancies, particularly colorectal cancers and breast and lung adenocarcinomas [18]. Therefore, the increased release of these cell-surface proteins from tumors located at the pleural surface, or their leakage through the lymphatic system, could potentially explain the elevated pCEA levels in NSCLC patients with VPI [23]. This micrometastasis can contribute to poor prognosis in small lung adenocarcinoma. Consequently, the evaluation of VPI using intraoperative pCEA could be beneficial in determining the extent of lung resection required during surgery. It also serves as an efficient diagnostic tool in terms of cost, time, and performance.
This study had several limitations. First, the retrospective design inherently introduced the potential for selection bias, as the patient inclusion process may have been influenced by specific factors. Second, the differentiation of the groups based on the presence of pleural contact was not performed by radiologists. Instead, to minimize the bias associated with diverse and detailed radiographic interpretations, multiple cardiothoracic surgeons re-evaluated the CT images and employed a simple categorization method to determine the presence or absence of direct pleural contact. Third, the examination of a wider range of tumor markers could have yielded a more thorough analysis. However, we chose CEA due to its relatively high sensitivity and specificity compared to other markers, a finding supported by prior studies. Fourth, we did not investigate the prognosis or long-term survival outcomes of the patients, primarily due to insufficient follow-up duration. Therefore, future data collection will allow us to assess survival outcomes and prognosis. Finally, to validate these results, further prospective studies with larger sample sizes or involving multiple institutions are required.
The findings of this study underscore the diagnostic value of pCEA and SUVmax for predicting VPI in clinical T1N0M0 lung adenocarcinoma, especially in instances involving direct pleural contact. This could aid clinicians in making informed decisions about the extent of lung resection required for patients with stage I lung adenocarcinoma in which VPI is suspected.
Author contributions
Conceptualization: HRN, SHM, SKY. Data curation: SHM, KSK, MHM, KYH, SKY. Formal analysis: HRN, SKY. Funding acquisition: not applicable. Investigation: HRN, SHM, KSK, MHM, KYH, SKY. Methodology: SHM, KYH, SKY. Project administration: SHM, SKY. Resources: SHM, KSK, MHM, KYH, SKY. Software: SKY. Supervision: SHM, SKY. Validation: SHM, SKY. Visualization: HRN, SKY. Writing–original draft: HRN. Writing–review & editing: HRN, SKY. Final approval of the manuscript: all authors.
Conflict of interest
No potential conflict of interest relevant to this article was reported.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Table 1 . Clinicopathologic characteristics of patients with clinical stage I adenocarcinoma with tumor ≤30 mm (n=390).
Characteristic | VPI absent (n=340) | VPI present (n=50) | p-value |
---|---|---|---|
Age (yr) | 64.3±10.2 | 66.7±10.0 | 0.025 |
Sex | 0.480 | ||
Female | 212 (62.4) | 28 (56.0) | |
Male | 128 (37.6) | 22 (44.0) | |
Smoking status | 0.330 | ||
Never | 225 (66.2) | 29 (58.0) | |
Ever | 115 (33.8) | 21 (42.0) | |
Pleural CEAa) | <0.001 | ||
Normal | 276 (81.2) | 21 (42.0) | |
Elevated | 64 (18.8) | 29 (58.0) | |
Serum CEAb) | 0.119 | ||
Normal | 285 (84.1) | 37 (74.0) | |
Elevated | 54 (15.9) | 13 (26.0) | |
SUVmaxc) | <0.001 | ||
Normal | 162 (69.8) | 9 (18.8) | |
Elevated | 70 (30.2) | 39 (81.2) | |
Extent of resection | 0.003 | ||
Sublobar | 139 (40.9) | 9 (18.0) | |
Lobectomy | 201 (59.1) | 41 (82.0) | |
Tumor size (mm) | <0.001 | ||
≤10 | 159 (46.8) | 0 | |
>10, ≤20 | 120 (35.3) | 22 (44.0) | |
>20, ≤30 | 61 (17.9) | 28 (56.0) | |
Nodule type | <0.001 | ||
Pure GGN | 84 (24.7) | 0 | |
Part solid | 143 (42.1) | 12 (24.0) | |
Pure solid | 113 (33.2) | 38 (76.0) | |
Pleural contact on CT | <0.001 | ||
Pleural contact | 148 (43.5) | 38 (76.0) | |
Non-pleural contact | 192 (56.5) | 12 (24.0) | |
Nodal status | <0.001 | ||
N0 | 324 (95.3) | 39 (78.0) | |
N1–N2 | 16 (4.7) | 11 (22.0) | |
Pathological TNM stage | <0.001 | ||
Stage I | 320 (95.0) | 35 (71.4) | |
Stage II | 10 (3.0) | 3 (6.1) | |
Stage III | 7 (2.1) | 8 (16.3) | |
Stage IV | 0 | 3 (6.1) | |
Predominant subtype | <0.001 | ||
Lepidic | 152 (44.7) | 4 (8.0) | |
Acinar | 154 (45.3) | 33 (66.0) | |
Papillary | 16 (4.7) | 2 (4.0) | |
Solid | 9 (2.6) | 5 (10.0) | |
Micropapillary | 2 (0.6) | 4 (8.0) | |
Cribriform | 1 (0.3) | 1 (2.0) | |
Mucinous | 5 (1.5) | 0 | |
Mixed | 1 (0.3) | 1 (2.0) |
Values are presented as mean±standard deviation or number (%)..
VPI, visceral pleural invasion; CEA, carcinoembryonic antigen; SUVmax, maximum standardized uptake value; GGN, ground-glass nodule; CT, computed tomography; TNM, tumor-node-metastasis..
a)The cutoff values for pleural CEA were 2.56 ng/mL for the pleural contact group and 2.105 ng/mL for the non-pleural contact group. b)The cutoff values for serum CEA were 3.8 ng/mL for never-smokers and 5.5 ng/mL for ever-smokers. Preoperative serum CEA levels were available for 389 patients at our hospital. c)The cutoff values of SUVmax were 4.25 for the pleural contact group and 4.65 for the non-pleural contact group. SUVmax values were available for 280 patients at our hospital..
Table 2 . Factors associated with VPI in patients with CT findings of pleural contact (n=186).
Variable | VPI | Univariate analysis | Multivariate analysis | ||||||
---|---|---|---|---|---|---|---|---|---|
Absent | Present | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | |||
Total | 148 (79.6) | 38 (20.4) | |||||||
Pleural CEAa) | <0.001 | ||||||||
Normal | 115 (77.7) | 14 (36.8) | Ref | Ref | |||||
Elevated | 33 (22.3) | 24 (63.2) | 3.64 (1.25–10.64) | 0.018 | 3.00 (1.14–7.87) | 0.026 | |||
Age (yr) | 64.6±10.0 | 67.2±10.9 | 0.176 | 1.01 (0.97–1.06) | 0.579 | ||||
Sex | 0.239 | ||||||||
Female | 103 (69.6) | 22 (57.9) | Ref | Ref | |||||
Male | 45 (30.4) | 16 (42.1) | 4.45 (0.79–25.2) | 0.091 | 2.17 (0.79–5.99) | 0.135 | |||
Smoking status | 0.282 | ||||||||
Never | 109 (73.6) | 24 (63.2) | Ref | ||||||
Ever | 39 (26.4) | 14 (36.8) | 0.37 (0.06–2.26) | 0.283 | |||||
Serum CEAb) | 0.024 | ||||||||
Normal | 127 (85.8) | 26 (68.4) | Ref | ||||||
Elevated | 21 (14.2) | 12 (31.6) | 0.42 (0.12–1.53) | 0.188 | |||||
SUVmaxc) | <0.001 | ||||||||
Normal | 82 (78.1) | 8 (22.2) | Ref | Ref | |||||
Elevated | 23 (21.9) | 28 (77.8) | 6.32 (2.08–19.20) | 0.001 | 5.25 (1.90–14.50) | 0.001 | |||
Tumor size (mm) | <0.001 | 1.03 (0.45–2.35) | 0.946 | ||||||
≤10 | 69 (46.6) | 0 | |||||||
>10, ≤20 | 48 (32.4) | 19 (50.0) | |||||||
>20, ≤30 | 31 (20.9) | 19 (50.0) | |||||||
Nodule type | <0.001 | 4.37 (1.46–13.11) | 0.008 | 3.89 (1.42–10.68) | 0.008 | ||||
Pure GGN | 29 (19.6) | 0 | |||||||
Part solid | 70 (47.3) | 8 (21.1) | |||||||
Pure solid | 49 (33.1) | 30 (78.9) |
Values are presented as number (%) or mean±standard deviation, unless otherwise stated..
VPI, visceral pleural invasion; CT, computed tomography; OR, odds ratio; CI, confidence interval; CEA, carcinoembryonic antigen; Ref, reference; SUVmax, maximum standardized uptake value; GGN, ground-glass nodule..
a)The cutoff value for pleural CEA was 2.565 ng/mL. b)The cutoff values for serum CEA were 3.8 ng/mL for never-smokers and 5.5 ng/mL for ever-smokers. c)The cutoff value for SUVmax was 4.25 SUVmax values were available for 141 patients at our hospital..
Table 3 . Factors associated with VPI in patients with CT findings of non-pleural contact (n=204).
Variable | VPI | Univariate analysis | Multivariate analysis | ||||||
---|---|---|---|---|---|---|---|---|---|
Absent | Present | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | |||
Total | 192 (94.1) | 12 (5.89) | |||||||
Pleural CEAa) | 0.030 | ||||||||
Normal | 155 (80.7) | 6 (50.0) | Ref | ||||||
Elevated | 37 (19.3) | 6 (50.0) | 2.03 (0.44–9.34) | 0.365 | |||||
Age (yr) | 64.0±10.3 | 69.7±9.1 | 0.064 | 1.04 (0.97–1.12) | 0.280 | 1.05 (0.99–1.11) | 0.078 | ||
Sex | 0.874 | ||||||||
Female | 109 (56.8) | 6 (50.0) | Ref | ||||||
Male | 83 (43.2) | 6 (50.0) | 0.58 (0.07–4.77) | 0.616 | |||||
Smoking status | 0.327 | ||||||||
Never | 116 (60.4) | 5 (41.7) | Ref | ||||||
Ever | 76 (39.6) | 7 (58.3) | 2.08 (0.28–15.19) | 0.470 | |||||
Serum CEAb) | 0.684 | ||||||||
Normal | 158 (82.7) | 11 (91.7) | Ref | ||||||
Elevated | 33 (17.3) | 1 (8.3) | 0.21 (0.02–2.15) | 0.187 | |||||
SUVmaxc) | <0.001 | ||||||||
Normal | 98 (77.2) | 3 (25.0) | Ref | Ref | |||||
Elevated | 29 (22.8) | 9 (75.0) | 3.37 (0.52–21.9) | 0.202 | 3.47 (0.74–16.42) | 0.116 | |||
Tumor size (mm) | <0.001 | 4.31 (0.84–22.1) | 0.079 | 4.66 (1.16–18.69) | 0.030 | ||||
≤10 | 90 (46.9) | 0 | |||||||
>10, ≤20 | 72 (37.5) | 3 (25.0) | |||||||
>20, ≤30 | 30 (15.6) | 9 (75.0) | |||||||
Nodule type | 0.029 | 0.81 (0.17–3.77) | 0.790 | ||||||
Pure GGN | 55 (28.6) | 0 | |||||||
Part solid | 73 (38.0) | 4 (33.3) | |||||||
Pure solid | 64 (33.3) | 8 (66.7) |
Values are presented as number (%) or mean±standard deviation, unless otherwise stated..
VPI, visceral pleural invasion; CT, computed tomography; OR, odds ratio; CI, confidence interval; CEA, carcinoembryonic antigen; Ref, reference; SUVmax, maximum standardized uptake value; GGN, ground-glass nodule..
a)The cutoff value for pleural CEA was 2.105 ng/mL. b)The cutoff values for serum CEA were 3.8 ng/mL for never-smokers and 5.5 ng/mL for ever-smokers. c)The cutoff value for SUVmax was 4.65 SUVmax values were available for 139 patients at our hospital..
2024; 57(3): 255-262