Analysis of perioperative risk factors and development of predictive models for complications following CT guided microwave ablation of pulmonary nodules

Abstract

Lung cancer has the highest incidence and mortality rates globally. Currently, curative surgical resection remains the primary treatment for early-stage lung cancer. However, 60% of lung cancer patients are ineligible for surgery due to various reasons (such as poor cardiopulmonary function or refusal of surgery). Microwave ablation of pulmonary nodules is an effective local treatment method that preserves lung parenchyma. However, being an invasive procedure, postoperative complications are inevitable in some patients. There is currently a scarcity of research that examines perioperative factors comprehensively, with few predictive models available in this context. Through analyzing the clinical data of 117 patients undergoing microwave ablation, our study identified BMI (Body mass index), emphysema, reduced lung diffusion capacity, and number of pleural punctures as independent risk factors for pneumothorax in patients with pulmonary nodules undergoing microwave ablation. Additionally, postoperative electrolyte imbalance was recognized as an early independent predictor for pneumothorax in these patients. Emphysema, tumor-pleura distance ≤ 10 mm, and maximum ablation power represent independent risk factors for postoperative pleural effusion after lung microwave ablation, while postoperative neutrophils count serves as an early independent predictive factor. The nomogram prediction models based on the above variables demonstrate strong predictive value, offering valuable guidance for clinical decision-making.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-98636-2.

Introduction

The “Global Cancer Statistics 2022,” released in 2024, provides a systematic analysis of global cancer incidence and mortality in 2022 based on the latest Global Cancer Observatory data. In 2022, there were 2.48 million new cases of lung cancer globally, accounting for approximately one-eighth of all cancer diagnoses, with 1.82 million deaths attributed to lung cancer, representing nearly one-fifth of all cancer-related deaths¹. Lung cancer has emerged as the most prevalent and deadliest malignancy worldwide, with both incidence and mortality rates on the rise. Additionally, the lungs rank as the second most common site for metastasis among all cancers, with approximately one-third of cancer-related deaths involving lung metastases². The increasing utilization of high-resolution, low-dose computed tomography (CT) scans³, particularly among individuals participating in lung cancer screening programs and routine health check-ups, has led to the detection of pulmonary nodules with increasing frequency. For early-stage non-small cell lung cancer (NSCLC), surgical resection with curative intent remains the primary treatment approach^4,5; however, for various reasons (such as poor cardiopulmonary function or advanced age, which renders anesthesia or surgery intolerable), about 60% of lung cancers cannot be surgically resected^6,7. The main alternative treatments for patients with early-stage NSCLC who are ineligible for surgery include stereotactic body radiation therapy (SBRT) and image-guided thermal ablation (IGTA). IGTA, as a standalone treatment modality, has emerged as the third major local tumor treatment following surgery and radiotherapy^8,9. At present, thermal ablation techniques primarily include radiofrequency ablation (RFA), microwave ablation (MWA), cryoablation, laser ablation, irreversible electroporation and high-intensity focused ultrasound ablation (HIFU).

Due to the relatively special tissue structure of the lungs, MWA demonstrates higher convection and lower heat settlement effects in inflated lung tissue than other thermal ablation techniques, as well as strong penetrating power. Therefore, MWA has a high thermal efficiency, rapid temperature increase, short ablation time, a relatively uniform high-temperature thermal field, and a wide range of necrosis in the coagulation zone^10,11. Moreover, as a lung parenchyma-preserving local treatment, MWA has a relatively small impact on lung function, allowing lung function to generally return to preoperative levels within one month postoperatively⁹. MWA mainly generates high temperatures of 60–150 °C in the surrounding tissue through an electromagnetic field, causing denaturation, coagulation, and necrosis of tissue proteins, thereby achieving the goal of treating lung lesions^6,12. Furthermore, the efficacy and safety of MWA for the treatment of early-stage primary lung cancer and limited pulmonary metastases have been confirmed^10,13. Several retrospective studies have indicated that for high-risk patients with stage I lung cancer who are unsuitable for surgery, MWA can achieve a prognosis similar to lobectomy^13–16. Compared with surgery, percutaneous image-guided MWA is preferable for patients with underlying lung disease and other comorbidities^13,17. With the increasing promotion of the concept of enhanced recovery in surgery, clinicians are placing more emphasis on postoperative rehabilitation and quality of life for patients. Therefore, the timely identification of risk factors for postoperative complications and early intervention is crucial. MWA has also garnered attention regarding postoperative complications and related risk factors in the treatment of pulmonary nodules. Previous studies have mainly focused on predicting the impact of various factors preoperatively and intraoperatively on the occurrence of complications in patients, while this study systematically analyzed the entire perioperative data of patients undergoing CT-guided MWA of pulmonary nodules, encompassing preoperative, intraoperative, and postoperative phases. For the first time, specific preoperative lung function indicators, easily accessible postoperative blood test results with minimal patient harm, and clinical data were included in this study. By identifying risk factors for postoperative complications in patients undergoing pulmonary nodule MWA, this study seeks to offer more precise clinical decision-making guidance for healthcare professionals.

Materials and methods

Inclusion and exclusion criteria

Inclusion criteria: (1) Pulmonary nodules with confirmed pathology or a high suspicion of malignancy on chest CT; (2) Inoperable due to comorbidities such as advanced age, previous history of lung resection, poor cardiopulmonary function, or refusal of surgery due to significant anxiety or fear; (3) Unilateral or bilateral lung nodules ≤ 3 or ≤ 5 oligometastases, respectively; (4) Eastern Cooperative Oncology Group (ECOG) performance status of ≤ 2.

Exclusion criteria: (1) Severe bleeding tendency or coagulation dysfunction uncorrectable within a short timeframe (prothrombin time > 18 s, prothrombin activity < 40%, platelet count < 50 × 10⁹/L); (2) Poor infection control; (3) Severe dysfunction of the liver, kidney, heart, or brain; (4) Occurrence of acute myocardial infarction or acute cerebral infarction within the past 30 days; (5) Lesions adjacent to major blood vessels, pulmonary hilum, main bronchi, or mediastinum; (6) Presence of poorly controlled extrapulmonary malignancies.

Research subject

A retrospective analysis was conducted on 112 patients (undergoing a total of 117 MWA sessions) with pulmonary nodules treated in the Department of Thoracic Surgery at our hospital from November 2020 to May 2024, during which a total of 125 pulmonary nodules were successfully ablated. The data were sourced from patient medical records, nursing documentation, and various auxiliary examinations, all of which were collected by one individual and subsequently confirmed and validated by another. The analysis of complication data centered around the aggregate number of lung MWA procedures (117 sessions), focusing on complications with an incidence rate exceeding 10%. Out of the total cohort, pneumothorax was observed in 30 patients, accounting for 25.6%, while postoperative pleural effusion occurred in 23 patients, representing 19.7% of cases. All patients included in this study were informed of the relevant surgical risks preoperatively, and their informed consent was obtained. This retrospective study adhered to the principles of the Helsinki Declaration and was approved by the Ethics Committee of the Third Affiliated Hospital of Soochow University (Approval No. 2021-Technology-95).

Ablation methods

All patients received an enhanced chest CT before MWA to evaluate the location, number and size of the tumors, as well as their relative relationships to important organs, blood vessels and bronchi. All laboratory tests were completed 1–3 days before MWA. In cases of preoperative electrolyte imbalances, manual intervention was necessary for correction before reevaluation. Relevant staging tests, such as bone scans or head CT scans, and PET/CT as needed, were also conducted to evaluate lymph node involvement and distant metastases. Under the guidance of multi-slice spiral CT (scan parameters: voltage 120 kV, current 250 mA, pitch 0.984, rotation speed 0.6 s, thin-layer scanning with a slice thickness interval of 1 mm), MWA was performed using the ECO MWA system (Nanjing ECO Medical Equipment Co., Ltd.), with a frequency of 2450 MHz and adjustable continuous wave output power ranging from 0 to 60 W. The effective length of the microwave antenna was 10–15 cm, with an outer diameter of 1.6 mm. The power and time parameters of the MWA device were set according to the size and location of the nodules. Prior to ablation, 5–10 mL of 1% lidocaine was administered for local anesthesia. A water-cooling system was simultaneously used during ablation to prevent heat damage to adjacent tissues or organs caused by the rapid expansion of the ablation zone. And complete ablation was achieved when the ground-glass shadow beyond the edge of the tumor measured about 5 to 10 mm. Following removal of the ablation needle, a CT scan was performed immediately to assess the ablation status and check for complications such as pneumothorax. The entire surgical procedure took place in a sterile environment, with continuous monitoring of the patient’s blood pressure, heart rate, and oxygen saturation to ensure stable vital signs. After surgery, routine chest X-rays should be performed 6 h postoperatively to detect delayed complications. In subsequent treatment or follow-up examinations, chest tube drainage may be necessary for patients with pneumothorax causing > 30% lung collapse, pleural effusion causing > 10% lung collapse, or for those with significant symptoms¹⁸. The treatment decisions for all patients were made collaboratively by thoracic surgeons, respiratory and critical care physicians, radiologists, and oncologists. And all surgical procedures were performed by the same thoracic surgeon and radiologist, both with over ten years of work experience.

Observational indicators

Preoperative information

Gender, age, BMI (Body Mass Index), COPD (Chronic Obstructive Pulmonary Disease), emphysema, history of ipsilateral lung surgery, maximum nodule diameter, tumor-pleura shortest distance, FVC (Forced Vital Capacity), FVC as a percentage of predicted, FEV1 (Forced Expiratory Volume in 1 s), FEV1 as a percentage of predicted, PEF (Peak Expiratory Flow), FEV1/FVC, DLCO (Carbon Monoxide Diffusing Capacity), pulmonary ventilation function and pulmonary diffusion function.

Intraoperative information

Surgical duration, maximum ablation power, cumulative ablation time, cumulative length traversing lung tissue, involvement of pleura in ablation, number of ablated lesions, and frequency of pleural puncture.

Postoperative information

Hospitalization costs, duration of hospital stay, length of postoperative stay, postoperative albumin levels, WBC (White Blood Cell) and NEUT (Neutrophil) levels within 24 h postoperatively, occurrence of electrolyte disturbances postoperatively (sodium, potassium, chloride), and postoperative fever status.

Statistical methods

The statistical analysis was conducted using SPSS 25.0 and R version 4.1.1. Measurement data were presented as the median and interquartile range[M()], and group comparisons were conducted using the Mann-Whitney U test. Categorical data were expressed as frequencies and proportions, with the χ² test or Fisher’s exact test used for comparisons. Univariate and multivariate logistic regression analyses were employed to identify risk factors for pneumothorax and pleural effusion following MWA of pulmonary nodules, with statistical significance set at P < 0.05. Based on the results of the multivariate analysis, a nomogram model was developed, and the accuracy of the model was assessed by plotting a ROC (Receiver Operating Characteristic) curve. Additionally, the Bootstrap method was used to resample the data 1,000 times to create “virtual samples.” This allowed us to simulate the sampling distribution of the initial data and to plot a calibration curve to evaluate the consistency of the model.

Results

Comparative analysis of perioperative clinical data in patients with pneumothorax and pleural effusion

In this study, pneumothorax occurred in 30 out of 117 patients, representing 25.6%. Postoperatively, pleural effusion was observed in 23 cases, accounting for 19.7% of the total. In comparison to patients without pneumothorax, those with pneumothorax showed a significant increase in hospitalization costs (23245.95 vs. 20432.01 yuan, P = 0.001), length of stay (4.48 vs. 2.93 days, P = 0.001), and length of postoperative stay (47.72 vs. 39.50 h, P = 0.012). Similarly, compared to patients without pleural effusion, those with pleural effusion exhibited a significant increase in hospitalization costs (23718.01 vs. 20991.28 yuan, P = 0.047), length of stay (5.90 vs. 2.94 days, P = 0.004), and length of postoperative stay (50.67 vs. 40.13 h, P = 0.007), as shown in Table S1.

Results from logistic regression analysis

Results of the univariate regression analysis

Compared to patients without pneumothorax, male gender (OR(Odds Ratio) 3.520, P = 0.009), BMI (OR 0.763, P = 0.002), COPD (OR 4.100, P = 0.030), emphysema (OR 6.045, P < 0.001), maximum tumor diameter (OR 2.074, P = 0.006), DLCO (OR 0.853, P = 0.007), reduced lung diffusion function (OR 2.969, P = 0.013), maximum ablation power (OR 1.113, P = 0.004), cumulative ablation time (OR 1.414, P = 0.026), cumulative needle tract length (OR 1.022, P = 0.045), pleural involvement during ablation (OR 3.000, P = 0.012), number of pleural punctures (OR 4.779, P = 0.001), postoperative NEUT (OR 1.265, P = 0.011), and postoperative electrolyte disturbances (OR 3.200, P = 0.013) showed statistically significant differences. An OR value greater than 1 indicates a risk factor, while an OR value less than 1 indicates a protective factor, as shown in Table 1.

Table 1.

Results of univariate and multivariate analysis in patients with pneumothorax.

Categorization	Univariable analyses				Multivariable analyses
	B	OR	95%CI	P	B	OR	95%CI	P
Preoperative information
Gender
Male	1.259	3.520	1.369–9.055	0.009
Female		Ref
Age(years)	0.028	1.028	0.988–1.070	0.168
BMI(kg/m2)	−0.270	0.763	0.642–0.907	0.002	−0.303	0.739	0.588–0.928	0.009
COPD
Yes	1.411	4.1	1.150–14.614	0.030
No		Ref
Emphysema
Yes	1.799	6.045	2.324–15.728	< 0.001	1.774	5.894	1.644–21.133	0.006
No		Ref				Ref
History of ipsilateral lung surgery
Yes	0.462	1.587	0.572–4.406	0.375
No		Ref
Maximum nodule diameter(cm)	0.730	2.074	1.232–3.493	0.006
Tumor-pleura shortest distance
≤ 10 mm	0.310	1.364	0.591–3.149	0.467
> 10 mm		Ref
FVC(L)	−0.162	0.851	0.430–1.684	0.643
FVC%	−0.024	0.976	0.951–1.002	0.066
FEV1(L)	−0.387	0.679	0.323–1.427	0.307
FEV1%	−0.019	0.981	0.961–1.002	0.071
PEF(L/min)	−0.295	0.745	0.554–1.001	0.051
FEV1/FVC(%)	−0.033	0.968	0.933–1.004	0.081
DLCO(ml/min/mmHg)	−0.159	0.853	0.760–0.957	0.007
Pulmonary ventilation function
Normal		Ref
Decline	0.660	1.934	0.831–4.500	0.126
Pulmonary diffusion function
Normal		Ref				Ref
Decline	1.088	2.969	1.255–7.023	0.013	1.335	3.798	1.038–13.899	0.044
Intraoperative information
Surgical duration(min)	0.029	1.030	0.994–1.066	0.100
Maximum ablation power (W)	0.107	1.113	1.034–1.198	0.004
Cumulative ablation time (min)	0.347	1.414	1.043–1.919	0.026
Cumulative length traversing lung tissue(mm)	0.022	1.022	1.000–1.043	0.045
Ablation involving the pleura
Yes	1.099	3.000	1.275–7.059	0.012
No		Ref
Number of ablated lesions
1		Ref
2	0.600	1.822	0.408–8.134	0.432
Number of pleural punctures
1		Ref				Ref
≥ 2	1.564	4.779	1.858–12.295	0.001	2.155	8.631	2.357–31.608	0.001
Postoperative information
Postoperative WBC(*109/L)	0.144	1.155	0.976–1.367	0.093
Postoperative NEUT(*109/L)	0.235	1.265	1.055–1.517	0.011
Postoperative electrolyte disturbance
Yes	1.163	3.200	1.278–8.015	0.013	1.634	5.124	1.514–17.341	0.009
No		Ref				Ref
Fever
Yes	0.568	1.765	0.687–4.534	0.238
No		Ref
Constant					3.623

Note: B: Regression Coefficient; OR: Odds Ratio; CI: Confidence Interval; Ref: Reference. BMI: Body Mass Index; COPD: Chronic Obstructive Pulmonary Disease; FVC: Forced Vital Capacity; FEV1: Forced Expiratory Volume in 1 second; PEF: Peak Expiratory Flow; DLCO: Carbon Monoxide Diffusing Capacity; WBC: White Blood Cell; NEUT: Neutrophil. A P-value of less than 0.05 indicates statistical significance. An OR greater than 1 signifies a risk factor, while an OR less than 1 indicates a protective factor.

Note: B: Regression Coefficient; OR: Odds Ratio; CI: Confidence Interval; Ref: Reference. BMI: Body Mass Index; COPD: Chronic Obstructive Pulmonary Disease; FVC: Forced Vital Capacity; FEV1: Forced Expiratory Volume in 1 s; PEF: Peak Expiratory Flow; DLCO: Carbon Monoxide Diffusing Capacity; WBC: White Blood Cell; NEUT: Neutrophil. A P-value of less than 0.05 indicates statistical significance. An OR greater than 1 signifies a risk factor, while an OR less than 1 indicates a protective factor.

Compared with patients without postoperative pleural effusion, the following factors showed statistically significant differences: COPD (OR 6.282, P = 0.005), emphysema (OR 8.883, P < 0.001), maximum tumor diameter (OR 1.945, P = 0.018), tumor-pleura shortest distance ≤ 10 mm (OR 7.680, P < 0.001), FEV1/FVC (OR 0.950, P = 0.010), DLCO (OR 0.854, P = 0.012), reduced pulmonary diffusion function (OR 3.021, P = 0.023), maximum ablation power (OR 1.163, P = 0.001), postoperative albumin (OR 0.827, P = 0.041), postoperative NEUT (OR 1.465, P < 0.001), and postoperative fever (OR 2.912, P = 0.034). An OR value greater than 1 indicates a risk factor, while an OR value less than 1 indicates a protective factor, as detailed in Table 2.

Table 2.

Results of univariate and multivariate analysis in patients with postoperative pleural effusion.

Categorization	Univariable analyses				Multivariable analyses
	B	OR	95%CI	P	B	OR	95%CI	P
Preoperative information
Gender
Male	0.999	2.715	0.984–7.491	0.054
Female		Ref
Age(years)	0.010	1.010	0.969–1.053	0.630
BMI(kg/m2)	−0.061	0.940	0.798–1.108	0.463
COPD
Yes	1.838	6.282	1.720–22.943	0.005
No		Ref
Emphysema
Yes	2.184	8.883	3.194–24.710	< 0.001	2.110	8.252	2.362–28.825	0.001
No		Ref				Ref
History of ipsilateral lung surgery
Yes	−0.457	0.633	0.170–2.366	0.497
No		Ref
Maximum nodule diameter(cm)	0.665	1.945	1.123–3.369	0.018
Tumor-pleura shortest distance
≤ 10 mm	2.039	7.680	2.604–22.651	< 0.001	1.618	5.044	1.430–17.784	0.012
> 10 mm		Ref				Ref
FVC(L)	0.201	1.222	0.588–2.541	0.591
FVC%	−0.009	0.991	0.964–1.018	0.494
FEV1(L)	−0.260	0.771	0.344–1.728	0.528
FEV1%	−0.018	0.982	0.960–1.004	0.108
PEF(L/min)	−0.041	0.960	0.714–1.290	0.785
FEV1/FVC(%)	−0.051	0.950	0.914–0.988	0.010
DLCO(ml/min/mmHg)	−0.158	0.854	0.754–0.966	0.012
Pulmonary ventilation function
Normal		Ref
Decline	0.655	1.926	0.759–4.886	0.168
Pulmonary diffusion function
Normal		Ref
Decline	1.106	3.021	1.164–7.839	0.023
Intraoperative information
Surgical duration(min)	0.014	1.014	0.977–1.053	0.463
Maximum ablation power (W)	0.151	1.163	1.065–1.271	0.001	0.136	1.145	1.028–1.276	0.014
Cumulative ablation time (min)	0.259	1.296	0.951–1.767	0.101
Cumulative length traversing lung tissue(mm)	0.007	1.007	0.985–1.030	0.529
Ablation involving the pleura
Yes	0.830	2.294	0.909–5.789	0.079
No		Ref
Number of ablated lesions
1		Ref
2	0.982	2.670	0.589–12.102	0.203
Number of pleural punctures
1		Ref
≥ 2	0.882	2.416	0.883–6.606	0.086
Postoperative information
Postoperative albumin (g/L)	−0.189	0.827	0.690–0.992	0.041
Postoperative WBC(*109/L)	0.176	1.192	0.993–1.431	0.059
Postoperative NEUT(*109/L)	0.382	1.465	1.187–1.809	< 0.001	0.286	1.332	1.029–1.724	0.030
Postoperative electrolyte disturbance
Yes	0.204	1.227	0.429–3.505	0.703
No		Ref
Fever
Yes	1.069	2.912	1.084–7.824	0.034
No		Ref
Constant					−10.632

Note: B: Regression Coefficient; OR: Odds Ratio; CI: Confidence Interval; Ref: Reference. BMI: Body Mass Index; COPD: Chronic Obstructive Pulmonary Disease; FVC: Forced Vital Capacity; FEV1: Forced Expiratory Volume in 1 second; PEF: Peak Expiratory Flow; DLCO: Carbon Monoxide Diffusing Capacity; WBC: White Blood Cell; NEUT: Neutrophil. A P-value of less than 0.05 indicates statistical significance. An OR greater than 1 signifies a risk factor, while an OR less than 1 indicates a protective factor.

Note: B: Regression Coefficient; OR: Odds Ratio; CI: Confidence Interval; Ref: Reference. BMI: Body Mass Index; COPD: Chronic Obstructive Pulmonary Disease; FVC: Forced Vital Capacity; FEV1: Forced Expiratory Volume in 1 s; PEF: Peak Expiratory Flow; DLCO: Carbon Monoxide Diffusing Capacity; WBC: White Blood Cell; NEUT: Neutrophil. A P-value of less than 0.05 indicates statistical significance. An OR greater than 1 signifies a risk factor, while an OR less than 1 indicates a protective factor.

Results of multiple regression analysis

Pneumothorax: Factors with P < 0.05 in univariate analysis were subjected to multivariable logistic stepwise regression. The analysis identified five independent risk factors: BMI (OR 0.739, P = 0.009), emphysema (OR 5.894, P = 0.006), reduced lung diffusion capacity (OR 3.798, P = 0.044), number of pleural punctures (OR 8.631, P = 0.001), and postoperative electrolyte imbalance (OR 5.124, P = 0.009), all showing statistically significant differences, as shown in Table 1.

Pleural effusion: Multivariable logistic regression analysis was conducted on factors with P < 0.05 in univariate analysis. The results identified four independent risk factors: emphysema (OR 8.252, P = 0.001), tumor-pleura shortest distance ≤ 10 mm (OR 5.044, P = 0.012), ablation power (OR 1.145, P = 0.014), and postoperative NEUT (OR 1.332, P = 0.03), all showing statistically significant differences, as presented in Table 2.

Development and verification of the predictive models for pneumothorax and pleural effusion

Based on the results of the logistic stepwise regression, nomogram plots were created using R version 4.1.1. Each risk factor corresponds to a point score above the line, with the total points indicating the probability of the corresponding outcome below. For pneumothorax, five independent predictive factors were considered: BMI, emphysema, lung diffusion capacity, number of pleural punctures, and postoperative electrolyte disturbance (Fig. 1). The regression equation is: logit(P) = 3.623 − 0.303 × BMI + 1.774 × emphysema + 1.335 × (pulmonary diffusion function decline) + 2.155 × (number of pleural punctures ≥ 2) + 1.634 × (postoperative electrolyte disturbance), as shown in Table 1. For postoperative pleural effusion, four independent predictive factors were included: emphysema, tumor-pleura shortest distance, maximum ablation power, and postoperative NEUT levels (Fig. 2). The regression equation is: logit(P) = −10.632 + 2.110 × emphysema + 1.618 × (tumor-pleura shortest distance ≤ 10 mm) + 0.136 × (maximum ablation power) + 0.286 × (postoperative NEUT), as shown in Table 2. These Nomogram plots can assist in identifying patients at risk of pneumothorax or pleural effusion following MWA, facilitating effective risk stratification and informing decisions regarding further monitoring or intervention.

Fig. 1.

Nomogram depicting the risk of pneumothorax in patients undergoing microwave ablation for pulmonary nodules: A line graph presentation.

Fig. 2.

Nomogram for the risk of pleural effusion after microwave ablation in patients with pulmonary nodules: A line graph presentation.

The ROC evaluation curve for predicting pneumothorax occurrence (Fig. 3) in patients undergoing MWA of pulmonary nodules yielded an AUC (Area Under the Curve) of 0.863 (95% CI: 0.787–0.920), surpassing the AUCs of individual risk factor curves. The model’s cutoff values for specificity and sensitivity are 87.4% and 73.3%, respectively. Additionally, the ROC evaluation curve for predicting pleural effusion occurrence (Fig. 4) yielded an AUC of 0.888 (95% CI: 0.808–0.982), also exceeding the AUCs of individual risk factor curves, with a specificity of 75.5% and sensitivity of 91.3%. These findings indicate that the nomogram exhibits strong discriminative ability and accuracy in identifying individuals at high risk of pneumothorax or pleural effusion following MWA.

Fig. 3.

ROC curves of the predictive model for the risk of pneumothorax in patients with microwave ablation of lung nodules.

Fig. 4.

ROC curves of the predictive model for the risk of postoperative pleural effusion in patients with microwave ablation of lung nodules.

Simultaneously, the consistency of the pneumothorax (Fig. 5) and pleural effusion (Fig. 6) predictive models was assessed by conducting 1000 bootstrap resamplings and plotting calibration curves. The results indicated a strong correlation between the predicted and actual values, with the proximity between the reference and observed data lines suggesting a close alignment between the predicted and actual event risks.

Fig. 5.

Calibration curves for a pneumothorax risk model developed through the bootstrap method.

Fig. 6.

Calibration curves for a postoperative pleural effusion risk model developed through the bootstrap method.

Discussion

Despite the advancements in robotic assistance systems¹⁹and medical imaging navigation technology²⁰in interventional radiology, which have improved the precision, safety, and efficiency of MWA technology and reduced complications, complications following MWA for pulmonary nodules still present a certain rate of incidence, which not only increases the financial and psychological burden on patients but also adversely affects postoperative recovery. Therefore, enhancing clinicians’ ability to predict the risks of post-ablation complications, enabling early diagnosis and providing effective treatment are crucial for improving cure rates and promoting patients’ postoperative recovery. Potential post-MWA complications include pneumothorax, pleural effusion, post-ablation syndrome, pleural reactions, bronchopleural fistula, bleeding, pulmonary infections, rib fractures, pseudoaneurysm of the pulmonary artery, pulmonary embolism, tumor seeding or dissemination along the needle tract, chest wall burns, nerve injuries, nonspecific respiratory failure, acute heart failure, and even death, all of which have been documented in relevant case reports^6,18,21–28. Although some complications may be unavoidable, the safety and predictability of the surgery are key prerequisites for selecting the surgical approach. Identification of perioperative risk factors for complications such as pneumothorax and pleural effusion in patients undergoing MWA for lung nodules, followed by timely intervention and rigorous preoperative preparation, can significantly reduce the incidence of postoperative complications, accelerate recovery, lower hospitalization costs, and shorten hospital stays.

Pneumothorax is the most common complication observed in patients undergoing lung MWA, with an incidence rate of 25.6% in this retrospective study, consistent with rates reported in previous studies ranging from 8.5 to 63%^{6,7,14,26,29–37}. The occurrence of pneumothorax can be attributed to thermal and mechanical lung tissue damage caused by ablation, which may lead to abnormal connections between the alveoli or airways and the pleural cavity³⁸. Most post-ablation pneumothorax cases are self-limiting and resolve spontaneously without requiring specific interventions such as thoracentesis, local skin incision for air drainage, or closed chest tube drainage. In our study, 30% (9/30) of patients with pneumothorax required chest tube drainage, similar to rates of 3.5–40% reported in previous studies^{23,32,33,35,39–41}. We found that low preoperative BMI, concomitant emphysema, impaired lung diffusion function, increased number of pleural punctures during the procedure, and postoperative electrolyte imbalances were independent predictive factors for pneumothorax following lung nodule MWA.

Emphysema is characterized by permanent enlargement of the distal air spaces in the lung, which results in the destruction of alveolar walls⁴². This condition manifests as increased radiolucency in chest X-rays and as areas of decreased lung density during inspiration on CT scans. It is also recognized as a critical predictive factor for pneumothorax following thermal ablation procedures^{6,21–23,32,43–45}. The anatomical distribution of emphysema may influence the development of pneumothorax, as the presence of emphysematous bullae can facilitate the occurrence of pneumothorax upon needle puncture. Previous studies have identified the number of pleural punctures as an independent risk factor for pneumothorax post-MWA^6,32, which aligns with our findings. During MWA procedures, needle punctures through the pleura create channels communicating with the pleural cavity, which can be further enlarged by respiratory-induced lung movements. An increased frequency of pleural punctures can lead to a greater quantity and size of pleural fissures, thereby elevating the risk of pneumothorax. While there is no conclusive evidence yet on the role of BMI as a risk factor for post-ablation pneumothorax, Ogawa et al. identified an association between low BMI and emphysema in male smokers with COPD⁴⁶. Additionally, research by Wang et al. identified the thickness of the chest wall traversed by the puncture needle as an independent risk factor influencing pneumothorax incidence during lung biopsy procedures⁴⁷. In patients with low BMI, the relatively thin soft tissue of the chest wall may reduce the stabilizing effect on the puncture needle, increasing the difficulty of puncture and consequently the risk of pneumothorax. Conversely, individuals with a high BMI may exhibit a more compact body shape with a smaller thoracic cavity, which limits lung respiratory movements and enhances needle stability, thereby reducing the risk of pneumothorax. These findings suggest that low BMI may act as a potential risk factor for pneumothorax following lung nodule MWA. This study is the first to incorporate preoperative lung function indicators and identify impaired lung diffusion capacity as a factor impacting postoperative pneumothorax occurrence. In the study by Ostridge et al.⁴⁸, a strong correlation was observed between CT quantification of pulmonary emphysema and pulmonary diffusion capacity. The severity of emphysema directly impacts lung diffusion capacity, with more severe emphysema leading to poorer diffusion function. This is attributed to structural damage in the alveoli of emphysema patients, where the supportive tissue of alveolar walls is compromised, resulting in collapse of small airways and permanent airflow obstruction. Additionally, the reduced number of capillaries in alveolar walls also further limits the exchange of oxygen and carbon dioxide. Also, in this study, a bivariate χ² test was conducted on emphysema and impaired lung diffusion function confirmed this association (P< 0.001). This study reveals that postoperative electrolyte disturbances may indicate the occurrence of pneumothorax in patients. When a patient experiences pneumothorax, inadequate oxygen supply due to lung compression can lead to hypoxia, increasing anaerobic metabolism and lactate production. This may result in intracellular acidosis, affecting cell membrane potential and potassium-hydrogen exchange (K⁺/H⁺ exchange). Concurrently, hypoxia can disrupt cellular energy metabolism, impairing the function of the sodium pump (Na⁺/K⁺-ATPase) on the cell membrane, thereby affecting the transmembrane transport of sodium and potassium ions^49,50.

Furthermore, other relevant studies have reported that a history of lung surgery^36,43,51, male gender^6,36,43, age^6,18,36,51, tumor diameter^6,18,35, tumor located in the middle or lower lobes of the lung^6,18,52, length of microwave needle traversal through lung tissue^18,51, pleural involvement during ablation^27,32,38,53, number of tumors ablated^36,45,51, history of lung ablation³², passage of ablation needle through fissures^18,32,44,54, prolonged ablation time³⁵, higher ablation power⁵⁵, use of multiple ablation electrodes⁶, preoperative radiotherapy⁴⁵, peribronchial tumors⁴⁵, and lack of burning in the puncture track³⁷ are also important influencing factors for the occurrence of pneumothorax following MWA of lung nodules.

Pleural effusion is a common complication following MWA, second only to pneumothorax, with an incidence rate of 19.7% in this retrospective study, consistent with previous findings ranging from 1 to 60%^{6,29,31,33,56}. The occurrence of post-ablation pleural effusion typically arises from pleuritis induced by thermal conduction to the pleura⁴³, as higher pleural temperatures have been significantly correlated with effusion formation⁵⁷. More severe complications often stem from pleural injury²⁵, leading to loss of elasticity and impaired ability to repair pleural openings, potentially triggering pleuritis with accompanying pleural effusion. In general, most patients with pleural effusion post-ablation are asymptomatic, usually discovered incidentally during post-ablation chest CT scans. Typically, conservative management involving oxygen therapy, nutritional support, and analgesics leads to the absorption and resolution of effusion without the need for further interventions such as thoracentesis or closed thoracic drainage. Drainage tube placement for pleural effusion was required in 8.7% (2/23) of cases in our study, aligning with previous reports of 0–10%^{6,23,25,29,33,43,56}. Our study identified pre-existing pulmonary emphysema, tumor-pleura shortest distance ≤ 10 mm, higher ablation power, and elevated postoperative neutrophil count as independent predictive factors for post-ablation pleural effusion among patients receiving MWA for pulmonary nodules.

In 2008, Tajiri et al.⁵⁷compared two groups of lung tumor patients undergoing RFA surgery differentiated by the maximum pleural temperature reached during the procedure: <40 °C and ≥ 40 °C. They found that the rate of pleural effusion was significantly higher in the group with pleural temperatures ≥ 40 °C. The conclusion drawn was that higher pleural temperatures were associated with the occurrence of pleural effusion, which is a widely accepted notion in current literature^6,18,58. It is important to note that when the target lesion lies close to the pleura, the heat generated during ablation can easily transfer to the pleura, resulting in thermal injury and elevating the risk of pleural effusion. In the research conducted by Zheng et al.²³, it was observed that patients undergoing microwave ablation of pulmonary nodules were more likely to develop pleural effusion when the distance between the tumor and the pleura was less than 1 cm. This aligns with our own study results, which demonstrate a notable rise in the incidence of pleural effusion when the distance from the chest wall to the target tumor is less than 10 mm^{6,23,25,57,58}. Xu et al.‘s previous research suggested a potential correlation between pulmonary emphysema and pleural effusion, although emphysema may be more closely associated with pneumothorax⁵⁸. The possible mechanism is that reduced pulmonary perfusion and ventilation in emphysema make lung parenchyma more susceptible to thermal injury²¹, which is consistent with our findings. Additionally, Xu et al. reported that maximum power is a significant predictive factor for the occurrence of pleural effusion⁵⁸, consistent with our findings. The potential mechanism is that the higher microwave energy transmitted to the lung parenchyma may elevate pleural temperatures and increase the risk of pleuritis. This study is the first to incorporate postoperative blood test results within 24 h and found that elevated neutrophil counts may serve as an indicator for the occurrence of postoperative pleural effusion. Neutrophils, as the first responders during the body’s acute inflammatory response, play a crucial role in defending against inflammation and various pathogens as the initial host defense system. Additionally, during inflammation, the lifespan of neutrophils significantly increases several-fold upon activation to ensure their sustained presence at the site of inflammation⁵⁹. Furthermore, related research has shown a significant association between increased neutrophils in lung cancer patients after MWA and poor prognosis⁵⁵.

Other relevant studies have reported that factors such as the use of cluster electrodes⁶, frequency of pleural puncture⁵⁸, multiple lesions ablated in one session^6,18, prolonged ablation time^6,18, a decrease in the length of the aerated lung traversed by the applicator⁶, larger tumor diameter³⁷, and the lack of burning along the puncture trajectory³⁷ are also important influencing factors for the occurrence of postoperative pleural effusion following MWA of pulmonary nodules.

Based on the independent risk factors identified in this study, a visual nomogram predictive model was developed for pneumothorax and pleural effusion in patients undergoing MWA for pulmonary nodules. By integrating relevant clinical data, preoperative risk stratification of patients undergoing MWA can be conducted to assess the postoperative risks of related complications. The predictive model’s discriminative ability was evaluated using ROC curves, yielding an AUC of 0.863 for pneumothorax prediction and 0.888 for pleural effusion, demonstrating great discriminative performance. Additionally, utilizing the bootstrap method with 1000 resamples, a calibration curve for the model was generated, showing strong agreement between predicted and actual outcomes. In clinical practice, incorporating this predictive model can facilitate early identification of patients at high risk for pneumothorax or pleural effusion post-MWA, enabling more rigorous preoperative preparation, timely diagnosis, intervention, and ultimately, preventing adverse outcomes.

This study has several limitations, including its single-center design, limited sample size, and lack of external validation of the dataset. And overfitting might be present when conducting consistency validation of predictive models. Additionally, the study did not differentiate between immediate-onset and delayed-onset pneumothorax in patients, nor did it address cases where pneumothorax coexisted with pleural effusion. In the future, further validation of this predictive model is expected to be supplemented through large-sample multicenter studies.

In summary, a low preoperative BMI, concomitant emphysema, impaired lung diffusion function, and an increased number of pleural punctures are independent risk factors for pneumothorax in patients undergoing MWA for pulmonary nodules. Additionally, postoperative electrolyte imbalance was recognized as an early independent predictor for pneumothorax in these patients. Preexisting emphysema, the shortest tumor-pleura distance of ≤ 10 mm, and higher ablation power are independent risk factors for postoperative pleural effusion in these patients. while elevated NEUT levels within 24 h postoperatively serves as an early independent predictive factor. The predictive model developed from these factors demonstrates high accuracy and clinical applicability. Moving forward, MWA, as an independent technique akin to surgery and SBRT, is expected to play an increasingly significant role in the comprehensive treatment of pulmonary tumors.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

LC: Writing – original draft, Investigation, Methodology. XX: Data curation, Formal analysis. KL: Validation. CY: Visualization, Project administration. HL: Supervision, Resources. LZ: Writing – review & editing. All authors reviewed the manuscript. Lei Chen and Xiangyu Xie contributed equally to this work.

Data availability

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.