ICG combined with medical adhesive in preoperative localization of complex pulmonary nodules: a retrospective study

Abstract

Background

Precise pulmonary nodule localization, especially for deep-seated or pleural-adherent lesions, remains challenging. Conventional methods face accuracy limitations and high complication rates. This study validates a CT-guided indocyanine green (ICG)-medical adhesive composite, leveraging fluorescence stability and adhesive retention to optimize video-assisted thoracoscopic surgery (VATS) precision in complex anatomical scenarios.

Methods

This retrospective single-center study evaluated a novel CT-guided preoperative localization technique combining ICG with medical adhesive in 262 patients (288 nodules). Preoperative localization was performed under CT guidance using a composite mixture of ICG and medical adhesive. The mixture was injected within 15 mm of the target nodule, and VATS was conducted with intraoperative fluorescence imaging guidance.

Results

Technical success was achieved in all cases (100%), with a mean procedural duration of 12.3 ± 4.2 min. Complications included mild pneumothorax (7.3%), self-limited hemorrhage (5.3%), and mild cough (8.0%), resolving spontaneously. The method successfully localized 57 nodules ≥ 2 cm from the pleura and 28 cases with pleural adhesions (including 7 total pleural adhesions). All nodules were resected without thoracotomy conversion. The fluorescence signal remained detectable for a maximum duration of 8 days post-localization, enabling flexible surgical scheduling.

Conclusions

ICG-medical adhesive localization offers a safe, effective, and adaptable strategy for preoperative pulmonary nodule localization, particularly in anatomically challenging cases. Its low complication rates, sustained fluorescence, and reliability in complex scenarios support broader clinical adoption to enhance VATS precision.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12880-026-02183-9.

Background

Pulmonary nodules present persistent challenges for diagnosis and treatment. Conventional modalities, including percutaneous biopsy, bronchoscopy, and positron emission tomography-computed tomography (PET-CT), often lack sufficient diagnostic precision [1, 2]. While video-assisted thoracoscopic surgery (VATS) has become indispensable for both diagnostic and therapeutic interventions [3], thoracic surgeons need to address critical intraoperative challenges such as rapid localization of non-palpable nodules, tumor resection with maximal oncologic precision, and navigation of anatomically complex scenarios—particularly those involving subpleural nodules exceeding 2 cm in depth or cases complicated by pleural adhesions. Nodules located deeper than 2 cm are often non-palpable, making precise localization dependent on reliable imaging or tactile markers. Additionally, pleural adhesions complicate anatomical dissection and increase conversion risk, and intraoperative dissection of adhesions further complicates nodule localization. Both of which significantly increase the difficulty of successful localization and minimally invasive resection.

While numerous standalone localization techniques exist, each carries inherent limitations [4–10]. In recent years, combining two positioning methods has emerged as a new strategy. For instance, the combination of medical adhesive with methylene blue has been shown to enhance localization accuracy and safety, with a lower complication rate compared to traditional hookwire localization [11]. Similarly, the use of a mixture of indigo carmine and lipiodol via bronchoscopic navigation demonstrated a 96.7% success rate with no procedure-related discomfort or complications [12]. Additionally, the development of a fluorescent iodized emulsion combining indocyanine green (ICG) and lipiodol has enabled precise localization and resection of pulmonary nodules, particularly those that are deep-seated [13]. These combined localization methods allow surgeons to synergize complementary benefits.

Medical adhesive is made from synthetic α-cyanoacrylate, which is a class of tissue sealants that polymerize rapidly upon contact with blood [14]. Crucially, the polymerized adhesive forms a hard nodule that provides surgeons with tactile feedback [15]. Previous studies have shown that preoperative CT-guided localization with medical adhesive is a simple, safe, and effective technique [16, 17]. However, research on the combined application of ICG and medical adhesive remains limited. In this retrospective analysis, we evaluated the utility of this novel combined technique for preoperative pulmonary nodule localization, with a particular focus on complex scenarios such as nodules deeper than 2 cm from the pleural surface and those presented with pleural adhesions.

Methods

Ethical approval and study design

This study was approved by the Institutional Review Board of Huashan Hospital, Fudan University (Approval No. KY-2025082), and patient informed consent was waived due to the retrospective nature of the study. All procedures involving human participants were performed in accordance with the ethical standards of the institutional research committee and according to the 1964 Helsinki Declaration and its later amendments, or equivalent ethical standards. From June 2023 to February 2025, 262 consecutive patients undergoing CT-guided ICG combined with medical adhesive localization prior to VATS were enrolled based on stringent criteria: (1) Pulmonary nodules measuring 6–20 mm in maximal diameter; (2) Radiological suspicion of malignant potential (≥ 3 months follow-up demonstrating size progression or solid component increase warranting surgical intervention); (3) The cardiopulmonary function was suitable for VATS procedures; (4) Explicit consent for preoperative localization. Detailed workflow of patient selection is illustrated in Fig. 1.

Fig. 1.

The workflow of patient selection

Localization procedure

All patients undergoing localization receive detailed preoperative instructions from nurses to ensure they fully understand the procedure and potential post-localization discomforts such as pain and cough. This approach significantly alleviates patient anxiety. During the preparation phase, 25 mg ICG (Dandong Yichuang Pharma) was mixed with 10 mL sterile aqueous solvent. A precisely measured aliquot (0.1 mL containing 0.25 mg ICG) was combined with 0.2 mL medical adhesive (Guangzhou Baiyun Medical Adhesive) prior to injection. Next, patients were positioned in supine, prone, or lateral decubitus positions based on the planned puncture trajectory. A CT scan (Siemens, 2 mSv per scan, total 5 mSv) is then performed to confirm the needle’s position, angle, and insertion depth based on the size, location of the pulmonary nodule and its relationship with surrounding tissues. Following routine disinfection, draping, and local anesthesia, the localization is performed using puncture needle (Bard Peripheral Vascular). Another CT scan is immediately conducted to verify the needle tip is within 15 mm of the nodule. Upon confirmation, 0.3 mL of ICG-medical adhesive mixture is rapidly injected, and the needle is promptly withdrawn to prevent adhesion between the needle tip and lung tissue. Post-procedure CT is repeated to confirm the adhesive and nodule positions, as well as to exclude complications such as pneumothorax and hemothorax (Fig. 2).

Fig. 2.

A patient underwent preoperative combined ICG and medical adhesive localization under CT guidance. A: CT (lung window) demonstrating a 0.6 cm nodule (white arrow) 0.7 cm from visceral pleura in RS6. B: Prepare ICG, medical adhesive, sterile aqueous solvent, and the puncture needle. C: Position the patient in a prone position and mark the puncture site (white arrow). D: Perform the puncture according to the preplanned angle and depth using puncture needle (white arrow). E: Use CT imaging to verify the position of the puncture needle tip and the target nodule (white arrow). F: After injection, post-procedural CT confirmation showing the ICG and medical adhesive mixture (green arrow) 0.5 cm from the target nodule (white arrow). CT = computed tomography, ICG = indocyanine green

VATS procedure

The patient was positioned sideways with the affected lung facing upward and ventilated with a double lumen endotracheal tube while under general anesthesia. After standard sterilization, two small incisions were made: a 3 cm incision at the 4th rib space near the anterior axillary to place the thoracoscopic operative instruments, and a 1 cm incision at the 7th rib space along the mid-axillary line as the observation port. During the VATS, surgeons switched to a fluorescent imaging mode to clearly pinpoint the pulmonary nodule visually. When necessary, they could also palpate the pre-marked medical adhesive spot with fingers or oval forceps to double-check the location. Based on the location of the nodule, a wedge resection or segmentectomy was performed using a disposable linear stapler. Segmentectomy was indicated for patients with deep-seated nodules where wedge resection could not guarantee a safe surgical margin, or nodules located at the anatomical center of a segment or crossing intersegmental planes. In segmentectomy cases, the ICG-medical adhesive marker served to visually verify that the target nodule was centrally located within the resected segment and distantly positioned from the intersegmental plane, thereby ensuring adequate oncologic margins. After the specimen was removed, the lesion was located and intraoperative frozen section examination was performed. For patients with benign lesions, the surgery was concluded. For patients with non-invasive lung cancer, routine lymph node sampling was carried out. Patients with invasive lung cancer indicated by frozen section and without contraindications underwent lobectomy and lymph node dissection (Fig. 3).

Fig. 3.

A patient underwent preoperative combined ICG and medical adhesive localization under CT guidance. A: Preoperative CT (lung window) demonstrating 0.7 cm pGGN (white arrow) 0.3 cm from visceral pleura in LS6. B: Right lateral position CT-guided deployment: Needle tip positioned 0.3 cm from the target nodule (white arrow). C: Post-procedural CT confirmation showing the ICG and medical adhesive mixture (green arrow) 0.3 cm from target nodule (white arrow). D: A 3 cm incision is made as the operation port, and a 1 cm incision as the observation port. E: Fluorescence-guided visualization (2 cm staining field) and a linear cutting stapler was used for lung wedge resection. F: Resected specimen demonstrating a 0.7 cm gray-white tumor (white dashed line) with a 0.3 cm margin from the medical adhesive site (green dashed line) and the postoperative pathology indicated MIA. CT = computed tomography, pGGN = pure ground-glass nodule, ICG = indocyanine green, MIA = minimally invasive adenocarcinoma

Observation indicators

This study documented the baseline characteristics of each patient and pulmonary nodule. Nodule size refers to the maximum diameter measured radiologically. Distance to pleural surface was defined as the shortest measurement between the nodule and either the nearest visceral pleura or interlobar fissure. When incomplete development of interlobar fissures was identified during VATS exploration, the distance from the nodule to the parietal pleura was recorded. Localization duration was calculated as the interval between the initial CT scan and the completion of the final CT scan during the localization procedure. The localization-to-surgery interval was defined as the time from completion of the final localization CT scan to surgical incision. Successful localization was confirmed when the final CT demonstrated the ICG-medical adhesive mixture within 1.5 cm of the target nodule. Localization-related complications included mild pneumothorax (small crescent-shaped gas occupying < 30% thorax on CT), self-limited hemorrhage (CT showing > 2 cm peri-needle track ground-glass opacity) and mild cough (induced by the irritative odor of medical adhesive).

Surgical parameters were additionally recorded. Local pleural adhesion was defined as limited fibrous bands confined to a small thoracic region, whereas extensive adhesion referred to diffuse fibrotic involvement across multiple areas or the entire pleural cavity. Operation time encompassed the period from initial skin incision to chest closure completion.

Statistical analysis

All statistical analyzes were performed using SPSS software (version 22.0; IBM Corporation, Armonk, NY, USA). Continuous variables are expressed as mean ± standard deviation, and categorical variables are presented as number (percentage).

Results

A total of 262 consecutive patients with 288 pulmonary nodules underwent preoperative localization, with technical success achieved in all cases (100%). The general characteristics of patients and nodules are detailed in Table 1. The study population had a mean age of 50.33 ± 10.52 years, comprising 93 males (35.5%) and 169 females (64.5%). The mean diameter of nodules measured 8.8 ± 2.6 mm with a mean pleural distance of 17.6 ± 11.4 mm. The nodules < 2 cm from pleural surface accounted for 80.2% (231/288) of cases, while 19.8% (57/288) demonstrated deeper locations (≥ 2 cm) (Fig. 4A-D). Notably, 75.7% (218/288) of the nodules were pure ground-glass nodules (pGGNs). The mean procedural duration was 12.3 ± 4.2 min. Localization-related complications included mild pneumothorax (7.3%, 19/262), self-limited hemorrhage (5.3%, 14/262), and mild cough (8.0%, 21/262), all of which resolved spontaneously without intervention. VATS was conducted within 24 h post-localization in 43 patients (16.4%), on post-localization day 2 in 124 cases (47.3%), and day 3 in 79 cases (30.2%). 16 patients (6.1%) underwent delayed VATS beyond 72 h. Notably, one patient who developed influenza-related fever underwent VATS 8 days after localization yet still exhibited strong fluorescence that enabled precise nodule resection (Supplementary Fig. 1).

Table 1.

The general characteristics of patients and nodules

Number of patients	262
Age, year	50.33 ± 10.52
Gender
Male	93 (35.5)
Female	169 (64.5)
Nodule size, mm	8.8 ± 2.6
Number of patients with nodules
1 nodule	239 (91.2)
2 nodules	20 (7.6)
3 nodules	3 (1.2)
Number of nodules	288
Distance to pleural surface, mm	17.6 ± 11.4
< 2 cm	231 (80.2)
≥ 2 cm	57 (19.8)
Type of nodules
Pure GGN	218 (75.7)
Solid nodule	26 (9.1)
Mixed GGN	18 (6.2)
Pulmonary nodule location
RUL	95 (36.3)
RML	17 (6.5)
RLL	65 (24.8)
LUL	52 (19.8)
LLL	33 (12.6)
Localization duration, min	12.3 ± 4.2
Localization-to-surgery interval, d	1.8 ± 3.4
Number of successful localization	288 (100)
Localization-related complications
Mild pneumothorax	19 (7.3)
Self-limited pulmonary hemorrhage	14 (5.3)
Mild Cough	21 (8.0)

Data are presented as mean ± standard deviation or number of patients with the percentage in parentheses

CT = computed tomography, mm = millimeter, GGN = Ground-glass nodule, RUL = right upper lung, RML = right middle lung, RLL = right lower lung, LUL = left upper lung, LLL = left lower lung, min = minute, d = day

Fig. 4.

A–D: A patient with deep-located nodule underwent localization and surgery. A: The RS8 nodule was located approximately 4 cm from the pleura (white arrow). B: The patient was placed in the supine position for localization. The ICG-medical adhesive composite (green arrow) was positioned 0.2 cm from the nodule (white arrow). C: Fluorescence imaging clearly demonstrated the localization site. D: The nodule (white arrow) was identified in the resected segmentectomy specimen, with pathology confirming AIS. E–H: A patient with total pleural adhesions underwent localization and surgery. E: The RS2 nodule was located approximately 1.5 cm from the pleura (white arrow). F: The patient was placed in the left lateral position. The ICG-medical adhesive composite (green arrow) was positioned 0.2 cm from the nodule (white arrow). G: Intraoperative exploration revealed total pleural adhesions, which were fully dissected. H: Fluorescence imaging clearly delineated the nodule’s location (green arrow). ICG = indocyanine green; AIS = adenocarcinoma in situ

Surgical outcomes and histopathological findings are summarized in Table 2. All 288 nodules were successfully identified in the initial surgical specimen, eliminating the need for secondary resection. No cases required conversion to thoracotomy. Resection modalities included wedge resection (68.3%, 179/262), segmentectomy (29.8%, 78/262), and lobectomy (1.9%, 5/262). Pleural adhesions were identified in 28 patients (10.7%), with 21 demonstrating partial pleural adhesions and 7 exhibiting total pleural adhesions (Fig. 4E-H). The mean operative time was 46.8 ± 5.2 min.

Table 2.

Surgical outcomes and histopathological findings

Surgical type
Wedge	179 (68.3)
Segmentectomy	78 (29.8)
Lobectomy	5 (1.9)
Pleural adhesion	28 (10.7)
Partial pleural adhesion	21 (8.0)
Total pleural adhesion	7 (2.7)
Operation time, min	46.8 ± 5.2
Nodule identified in first excision	262 (100)
Histopathologic results
Chronic inflammation	14 (4.9)
Fibrosis	8 (2.8)
Cryptococcus infection	10 (3.5)
Granuloma	15 (5.2)
Atypical adenomatous hyperplasia	41 (14.2)
Adenocarcinoma in situ	67 (23.3)
Minimally invasive adenocarcinoma	108 (37.4)
Invasive adenocarcinoma	19 (6.6)
Metastatic tumor	6 (2.1)

Data are presented as mean ± standard deviation or number of patients with the percentage in parentheses

min = minute

Discussion

The widespread adoption of low-dose computed tomographic (LDCT) screening has significantly increased the detection of early-stage lung cancers manifesting as small pulmonary nodules, contributing to reduced lung cancer mortality [18, 19]. Notably, this has led to an increased detection of pGGNs [20]. In our study, pGGNs constituted 75.7% of all nodules. These lesions pose a distinct surgical challenge as they are non-palpable and often impossible to visualize during standard VATS, which adds a further complexity to localization. However, precise intraoperative localization remains challenging. Failed localization may necessitate conversion to thoracotomy or result in missed nodules, exposing patients to heightened risks [21]. Consequently, developing reliable methods to accurately locate these nodules, minimize complications, and boost positioning accuracy has become a priority for thoracic surgeons. Notably, ICG combined with medical adhesive represents an understudied yet promising strategy. Our study innovatively applied this combined ICG-medical adhesive method for preoperative pulmonary nodule localization. We found this combined method to be safe and effective while enabling flexible surgical scheduling, with enhanced performance in anatomically challenging cases, specifically those requiring localization of nodules ≥ 2 cm beneath the pleural surface and pleural adhesions. These findings support broader clinical adoption.

Complication rates vary significantly among pulmonary nodule localization methods. Hook-wire localization demonstrates higher complication rates with pneumothorax and hemorrhage rates reaching 39.2% and 21.6%, respectively, in a study by Park et al. [4], while iodized oil localization reduced hemorrhage incidence to 5.8% but was associated with an elevated pneumothorax rate of 45.0%. Huang et al. [22] reported pneumothorax (18.9%) and hemothorax (27.3%) rates for the microcoil localization technique. Emerging materials exhibit improved safety profiles. Wang et al. [23] observed lower pneumothorax (17.7%) and hemothorax (6.5%) rates using ICG localization. Cen et al. [7] demonstrated the lower complication rates with medical adhesive, achieving pneumothorax (8.5%) and hemothorax (7.9%) rates. In our cohort, pneumothorax and hemorrhage rates were 7.3% and 5.3%, respectively, lower than previously reported values. This may be attributed to: (1) We planned the puncture path preoperatively to avoid blood vessels and emphysematous tissues; (2) Medical adhesive easily forms a sealed cavity in the needle tract, effectively reducing pneumothorax, and can also serve as an embolic agent if hemorrhage occurs around the needle tract; (3) Preoperative education alleviated patient anxiety, which minimized needle displacement risks by stabilizing both involuntary patient movement during localization and anxiety-induced muscle hypertonia. While cough (7.7% -14.4%) has been reported during medical adhesive application in prior studies [11, 17], we observed an 8% incidence of self-limiting cough, all resolving spontaneously without clinical intervention. Based on our clinical experience, we have identified the following key insights: a controlled injection duration of 2–3 s, infusion rate of 0.05 mL/s, concurrent breath-holding before the injection, and total injection volume ≤ 0.3 mL.

In addition to complication rates, the clinical utility of localization techniques should be evaluated by outcomes that directly align with clinical goals. While numerous studies have described technical success rates focused on intraoperative identification of the localization marker itself, reports of histological confirmation success rate, a clinically meaningful metric that may better reflect the true value of a localization technique, remain limited. In contrast, our study achieved a 100% histological confirmation success rate with all 288 nodules dentified in the initial surgical specimen and confirmed pathologically without secondary resection or diagnostic failure. This demonstrates the reliability of our ICG-medical adhesive technique as it ensures technical success with definitive pathological diagnosis.

Studies show that using dyes like methylene blue alone for localization can cause diffusion [8, 9, 24]. While ICG fluorescence imaging offers low cost, minimal toxicity and allergic reactions [25], can also diffuse when used alone for pulmonary nodule localization [23], for its binding to plasma proteins and hepatic metabolism predispose it to tissue diffusion [26]. However, combining medical adhesive with ICG effectively mitigates dispersion through the following mechanisms: The primary component α-cyanoacrylate in medical adhesive forms a stable composite with ICG at the injection site [7], creating a physical barrier which restricts ICG migration via blood flow or interstitial fluid. This enhances local ICG retention, prolonging fluorescence intensity and optimizing imaging contrast (Fig. 5A-C). This allows flexible surgical scheduling: procedures from Tuesday to Friday can utilize same-day or pre-operative localization, while Monday surgeries may employ markers placed the preceding Friday. In our study, in patients undergoing localization within a week prior to surgery, nodules were successfully resected with sustained fluorescence guidance. A previous study showed the longest interval between ICG localization for lung nodules and surgery is 6 days [10]. In our study, the longest visualization time was 8 days. A similar study has shown that fluorescence could still be detected one week after ICG localization [23]. The persistence of this fluorescence improves patient comfort by eliminating activity or positional restrictions during the waiting period. Our study did not systematically evaluate the duration of fluorescence persistence; however, even if fluorescence fades, surgeons can still locate the nodule by feeling the medical adhesive with their fingers, ensuring safe nodule localization.

Fig. 5.

Schematic diagram illustrating the ICG-medical adhesive localization procedure and its clinical efficacy. A: A mixture of 0.1 mL ICG and 0.2 mL medical adhesive is prepared for injection. B: Under CT guidance, the ICG-adhesive composite is injected via the positioning needle into the perinodular region of the target pulmonary nodule. C: Intraoperative fluorescence imaging confirms the localized ICG-medical adhesive composite during VATS and a stable composite is formed at the injection site which establishes a physical barrier to restrict ICG migration. ICG = indocyanine green, CT = computed tomography, VATS = video-assisted thoracoscopic surgery

The correlation between nodule-to-pleura distance and localization efficacy remains underexplored. Previous studies demonstrated electromagnetic navigation bronchoscopic localization (ENBL) as a safe and accurate method for intraoperative localization of small, deep, or subsolid nodules, achieving a success rate of 98.1% with a mean nodule-to-pleura distance of 22 mm [27]. However, ENBL requires prolonged operative time, specialized equipment, and long learning curve [28]. Moreover, studies utilizing microcoil or hookwire localization reported elevated failure risks for nodules < 1 cm or > 2.5 cm from the pleura, primarily due to pneumothorax [29–31]. In our cohort, 57 nodules (19.8%) located ≥ 2 cm from the pleura achieved 100% successful localization and resection, potentially attributable to reduced pneumothorax rates (7.3%). It is well-established that deeper nodules increase intraoperative localization and margin control challenges. Our experience highlights two critical strategies: (1) Segmentectomy is prioritized over wedge resection for deep nodules requiring safe margins; (2) Meticulous trajectory design to avoid vascular structures and emphysematous regions. This requires the surgeon’s solid anatomical skills and rich experience, both crucial for improving the safety and effectiveness of sublobar resection.

Pleural adhesions constitute an independent risk factor for conversion to thoracotomy during VATS and postoperative complications, while extensive adhesions significantly impede nodule localization [32, 33]. Current preoperative imaging modalities including X-ray [34], CT [35], and ultrasound [36] exhibit limited accuracy in adhesion prediction. Our study first validates the reliability of ICG combined with medical adhesive localization in 28 pleural adhesion cases (100% success), including 7 total pleural adhesions. Our experience is as follows: (1) Fingers are used through incision for fan-shaped blunt dissection of the adhesion, expanding the area as much as possible. The camera is reinserted, and electrocoagulation hooks are used through the main operating port for total pleural adhesions; (2) Fluorescence-guided segmentectomy for deep nodules—dissect hilar structures after adhesion release, prioritize vascular management per intraoperative anatomy until the nodule is completely resected at a safe distance under fluorescence guidance. Our study shows that this localization technique is effective in guiding surgery even in complete pleural adhesion, providing a reliable preoperative localization method for surgeons.

This single-center retrospective study has inherent limitations: modest sample size and lack of comparative analysis with conventional techniques. Additionally, although some experience has been accumulated in localization technology and surgical procedures, further optimization and improvement are still required to enhance their effectiveness in broader clinical applications.

In conclusion, this study demonstrates the safety and efficacy of ICG combined with medical adhesive for preoperative pulmonary nodule localization, particularly revealing its unique advantages in challenging scenarios such as deep-seated nodules and pleural adhesions. While acknowledging current limitations, our findings empower thoracic surgeons to navigate complex pulmonary nodules where the ICG-adhesive composite serves as a fluorescent and tactile marker guiding precise resection, providing new insights into the preoperative localization of pulmonary nodules.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Wenhao Wang: Study conception and design, data acquisition and interpretation, statistical analysis, drafting of the manuscript; Yulong Tan: Study conception and design, data acquisition and interpretation, data collection, images analysis, drafting of the manuscript; Dong Xu: Data collection, images analysis, drafting of the manuscript; Haoxin Liu: Statistical analysis, drafting of the manuscript; Yifeng Qian: Statistical analysis, supervision, critical revision of the manuscript; Huijun Zhang: Study conception and design, data acquisition and interpretation, supervision, critical revision of the manuscript; Meng Shi: Study conception and design, data acquisition and interpretation, images analysis, critical revision of the manuscript; Xiaofeng Chen: Study conception and design, data acquisition and interpretation, images analysis, supervision, critical revision of the manuscript. All authors read and approved the final manuscript.

Funding

None.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Institutional Review Board of Huashan Hospital, Fudan University (Approval No. KY-2025082), and patient informed consent was waived due to the retrospective nature of the study. All procedures involving human participants were performed in accordance with the ethical standards of the institutional research committee and according to the 1964 Helsinki Declaration and its later amendments, or equivalent ethical standards.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.