Targeting Pulmonary Nodules Prior to Video-Assisted Thoracic Surgery: Comparison of Two Different Techniques

Abstract

Objectives

The increasing use of CT imaging and the implementation of lung cancer screening have led to a rising detection of pulmonary nodules. Accurate management requires reliable localization techniques enabling pathological diagnosis and potential curative resection. This study compared hook-wire localization and cone-beam computed tomography (CBCT)-guided detection in terms of diagnostic yield.

Methods

A retrospective analysis was performed in 2 French academic centres, including 260 patients with 266 nodules (January 2012-May 2019). The primary end-point was diagnostic yield, defined as the proportion of successfully localized nodules. Secondary end-points included localization and operative times, peri- and postoperative complications, chest tube duration, and hospital stay.

Results

Diagnostic yield was 88.0% with hook-wire localization and 96.9% with CBCT, corresponding to an absolute difference of 8.9% (95% CI, –0.2 to 17.9; P = .053). Complication rates (P = .63), drainage duration (P = .13), and hospital stay (P = .16) did not differ between groups. Operative time (P = .042) and localization time (P = .001) were significantly longer with CBCT and hook-wire, respectively.

Conclusions

Hook-wire and CBCT localization demonstrated comparable diagnostic yields and acceptable safety profiles. CBCT required longer operative room utilization, whereas hook-wire placement was performed in the radiology department, saving resources for other surgeries.

The widespread use of CT imaging in clinical practice, together with advances in image resolution and the implementation of lung cancer screening programs in smokers, has led to a steady increase in the detection of pulmonary nodules.

Graphical abstract

INTRODUCTION

The widespread use of CT imaging in clinical practice, together with advances in image resolution and the implementation of lung cancer screening programs in smokers, has led to a steady increase in the detection of pulmonary nodules.^1,2 A proportion of these nodules represent early-stage non-small cell lung cancer (NSCLC), for which timely diagnosis and treatment can improve survival. Management strategies vary worldwide depending on nodule type and size, ranging from radiological surveillance to diagnostic or curative surgical resection. However, small, pure ground-glass, or deep-seated mixed or solid nodules remain difficult to localize, particularly with the increasing use of video-assisted thoracoscopic surgery (VATS), which eliminates tactile palpation.

In the absence of localization, deep or small nodules carry a high risk of surgical conversion. Galetta et al. reported a 63% conversion rate for nodules <1 cm and located >5 mm from the pleural surface.³ Definitions of “non-palpable” nodules vary across centers, generally including nodules <1 cm or those located at least 1-2 cm from the pleura, especially when of ground-glass or mixed density.^4,5 As a result, there is no consensus regarding indications for localization. Current recommendations define surgical indications at multidisciplinary tumour boards, typically including any mixed nodule with a solid component >6 mm, any ground-glass nodule >6 mm with interval growth, and any solid nodule ≥8 mm or with progression.^6–9

Multiple localization techniques are available, including hook-wire, coils, methylene blue, lipiodol, indocyanine green, technetium-99m, barium, intraoperative ultrasound, electromagnetic navigation, and robot-assisted bronchoscopy. Reported complication rates vary widely. Hook-wire placement is associated with pneumothorax (7%-49%), pulmonary Haemorrhage (15%-30%), wire dislodgement (1%-7%), and surgical conversion rates up to 16%.^10,11 Technique selection largely depends on local expertise and resources, such as radiologist availability, CT scanner access, and interventional endoscopy. Comparative studies have produced inconsistent results: for instance, Park et al. found no significant difference between hook-wire and lipiodol localization in 250 patients.¹²

Continuous refinements are underway, such as modified hook-wire systems,¹⁰ methylene blue combined with collagen,¹³ and radiotracers, although their efficacy may be limited in emphysematous patients.¹⁴ Other approaches include electromagnetic navigation bronchoscopy^15,16 or virtual 3D reconstruction software for anatomical segmentectomy. Most available studies are limited by small sample sizes and significant inter-centre variability, limiting the generalizability of results.

The aim of this study was to compare the diagnostic yield of 2 localization techniques, hook-wire and cone-beam computed tomography (CBCT), before VATS resection of pulmonary nodules.

PATIENTS AND METHODS

Study design and population

This multicentre retrospective descriptive study was conducted between January 2012 and May 2019 in 2 French academic centres: Jean Perrin Comprehensive Cancer Center (Clermont-Ferrand, France–centre 1) and Rennes University Hospital (Rennes, France–centre 2). The study was classified as MR004, sponsored by Jean Perrin Comprehensive Cancer Center, and registered with the institutional data protection officer and the French Health Data Hub. Rennes University Hospital participated as an investigating centre. The study period was determined to exclude the learning curves of the different operators.

Ethic Committee number was 2022/CE61 and was obtained on December 7, 2022.

Patients were identified through operating room listings cross-referenced with departmental databases. All patients were notified by mail about the study and were given the opportunity to decline participation. Written informed consent was obtained from all patients.

Inclusion and exclusion criteria

Eligible patients were those who underwent preoperative localization of pulmonary nodules followed by minimally invasive VATS in the absence of a preoperative pathological diagnosis.

Exclusion criteria included: nodules >3 cm in diameter, nodules >3 cm from the pleural surface, palpable or visible nodules at surgery that would not require a targeting technique, nodules with a known pathological diagnosis before resection, resections performed by thoracotomy, and patients who did not consent to data use.

Localization techniques

The choice of localization technique was centre-dependent and based on equipment availability and institutional workflow rather than on nodule characteristics or surgeon preference. One centre routinely used CT-guided hook-wire localization, whereas the second centre had access to a hybrid operating room allowing CBCT-guided localization.

Hook-wire localization was performed in the radiology department prior to surgery. After a low-dose non-contrast CT scan, the radiologist inserted the hook-wire (Ariane’s thread, Biosphere Medical, Louvres, France) under local anaesthesia, adjusting its trajectory and depth with sequential CT slices. A final CT scan verified correct positioning and excluded immediate complications before transfer to the operating room (Figure 1).

Figure 1.

(A) Hook-Wire Placement Under CT-Scan; (B) Video-Thoracoscopic View

CBCT localization was performed intraoperatively in a hybrid operating room equipped with a fixed angiographic C-arm system. After induction of general anaesthesia, double-lumen endotracheal intubation, and lateral decubitus positioning, an initial cone-beam CT acquisition was performed to confirm patient positioning and lung inflation status. Following trocar placement and creation of a pneumothorax, a second CBCT acquisition was obtained. The target pulmonary nodule was manually segmented on 3D reconstructed images. This segmentation was then automatically projected onto real-time fluoroscopic images, creating an augmented reality overlay. Using this overlay in combination with standard thoracoscopic visualization, surgical instruments were guided towards the lung surface adjacent to the targeted lesion, allowing accurate identification of the resection area (Figure 2). Localization and resection were therefore performed during a single operative session, without patient transfer outside the operating room.

Figure 2.

Intraoperative Localization with CBCT. (A) straightforward Fluoroscopy; (B) video-Thoracoscopic View; (C) Augmented Fluoroscopy; (D) View of the Hybrid Operating Room

Surgical procedure

All resections were performed by VATS according to institutional practice. Patients were excluded if the nodule was visible or palpable intraoperatively. Wedge resection was performed as the initial procedure. If frozen section confirmed malignancy, completion lobectomy or anatomical segmentectomy with systematic lymphadenectomy was undertaken. The extend of the parenchymal resection was based on nodule’s characteristics according to current guidelines and patients’ respiratory reserve. Segmentectomy was extended to lobectomy if frozen section analysis of the segmental bronchus lymph node was positive.

Outcomes

The primary end point was diagnostic yield, defined as the proportion of successfully localized nodules among all nodules targeted. Success was based on nodule palpation within the resected specimen and/or pathological confirmation on frozen or final analysis. Resection margin status was not included in the definition of localization success, as it reflects oncological adequacy rather than localization performance.

Secondary end points included procedure-specific morbidity (pneumothorax, alveolar Haemorrhage, radiation exposure, chest drainage), postoperative complications according to the Clavien-Dindo classification,¹⁷ operative time, localization time, chest tube duration, hospital stay, pathological characteristics of the nodules, and need for surgical conversion.

Data collection

Collected data included:

• Patient characteristics: centre, sex, age, smoking history, FEV1, DLCO, and tumour location (side and lobe).

• Preoperative characteristics: number of nodules, targeted nodules, nodule size, distance from pleura, localization technique, and immediate complications.

• Perioperative characteristics: type of resection (wedge, segmentectomy, lobectomy), lymphadenectomy, and surgical conversion.

• Time metrics: Targeting time was defined differently according to the technique. For CT-guided hook-wire localization, targeting time corresponded to the interval between the first and last CT acquisition in the radiology department. For CBCT localization, targeting time was defined as the interval between skin incision and the last CBCT acquisition performed intraoperatively in the hybrid operating room. Operative time was defined as skin incision to skin closure and therefore included both localization and resection steps for CBCT.

• Pathology: histological diagnosis of the resected nodules.

• Postoperative characteristics: chest tube duration, hospital stay, and postoperative complications graded by Clavien-Dindo.¹⁷

Statistical analysis

Descriptive statistics were used to summarize the data. Quantitative variables were expressed as mean ± SD or median (interquartile range), and categorical variables as frequencies and percentages. Comparisons were performed using Fisher’s exact test or χ² test for categorical variables and the Mann-Whitney U or Kruskal-Wallis test for continuous variables, depending on group numbers. Missing data were infrequent (<2% for key variables) and were handled using complete-case analysis. An α risk of 5% was used. All statistical tests were two-tailed, and a P-value <.05 was considered statistically significant. Given the exploratory nature of secondary analyses, P values were not adjusted for multiplicity. To address potential selection bias related to the retrospective design and centre-dependent allocation, an exploratory propensity score-matched analysis was performed as a sensitivity analysis. The propensity score was estimated using a logistic regression model including age, sex, nodule size, distance from the pleura, nodule density and lobe location. Patients were matched 1:1 using nearest-neighbour matching without replacement and a caliper of 0.2 SDs of the logit of the propensity score.

RESULTS

A total of 261 patients with 267 pulmonary nodules were initially identified. One patient declined data use and was excluded, leaving 260 patients and 266 nodules for analysis. Most patients (n = 255) had a single nodule targeted and resected; 4 patients had 2 nodules and 1 had 3.

At centre 1, 205 patients with 208 nodules underwent hook-wire localization. At centre 2, 55 patients with 57 nodules were included, of which 24 were localized by hook wire and 33 by CBCT. Patients were classified into 2 groups according to the localization technique.

Primary endpoint

Diagnostic yield was 88.0% with hook wire and 96.9% with CBCT, corresponding to an absolute risk difference of +8.9% (95% CI, −0.2 to +17.9; P = .053), without a statistically significant difference (Table 1). In the CBCT group, one localization failure corresponded to inadvertent targeting of a non-planned nodule and was counted as a failure of the primary end point. Given the centre-dependent allocation of localization techniques, a centre-adjusted analysis was performed. After adjustment for centre, the direction and magnitude of the effect on diagnostic yield remained similar, with no statistically significant difference between techniques. Furthermore, in a sensitivity analysis restricted to centre 2, where both localization techniques were used, diagnostic yield was numerically higher with CBCT, with no statistically significant difference between techniques.

Table 1.

Patients’ Characteristics

		Hook-wire group (n, %)	CBCT group (n, %)	P
History of cancer		118 (51%)	18 (55%)	.67
Sex	F	118 (51%)	15 (45%)	.58
	M	115 (49%)	18 (55%)
Smoking history		156 (68%)	17 (52%)	.055
DLCO, mean (SD)		75.3 (17.9)	77.6 (15.7)	.46
FEV1, mean (SD)		92.6 (19.4)	101 (22.3)	.058
Nodule size (mm), median [Q25-75]		10.0 [7.00; 15.0]	12.0 [9.00; 15.0]	.075
Distance from the pleura (mm), median [Q25-75]		9.71 [6.63; 15.6]	8.00 [5.00; 15.0]	.39
Nodule localization	RUL	69 (30%)	10 (30%)	.92
	LLL	51 (22%)	6 (18%)
	LUL	49 (21%)	8 (24%)
	RLL	48 (21%)	8 (24%)
	ML	16 (6.9%)	1 (3%)
Number of targeted nodules per patient	1	228 (98%)	33 (100%)	1
	2	4 (1.6%)	0 (0%)
	3	1 (0.4%)	0 (0%)

Abbreviations: F, female; LLL, left lower lobe; LUL, left upper lobe; M, male; ML, middle lobe; RLL, right lower lobe; RUL, right upper lobe.

After propensity score matching, 32 patients were included in each group (Supplementary material 1). Baseline characteristics were well balanced between groups. Diagnostic yield remained numerically higher in the CBCT group, with no statistically significant difference between techniques, consistent with the primary analysis.

Preoperative characteristics

Baseline demographics were comparable between groups (Table 1). Patients in the hook-wire group had a higher number of nodules on preoperative CT than those in the CBCT group (P < .001). Mean localization time was significantly longer with hook wire (32 min) compared with CBCT (13 min; P < .001, Table 3). Mean nodule size was 10 mm in the hook-wire group and 12 mm in the CBCT group (P = .075). Mean nodule-to-pleura distance was 11.3 mm, with a small but significant difference between groups (P < .045).

Table 3.

Perioperative Times

	Hook-wire group (n = 233)	CBCT group (n = 33)	P
Duration of the surgery (min), median [Q25-75]	65.0 [34.2; 138]	90.0 [60.0; 125]	.042
Duration of the targeting procedure (min), median [Q25-75]	32.0 [25.0; 42.0]	13.0 [11.0; 16.0]	<.001
Duration of chest tube drainage (hours), median [Q25-75]	48.0 [24.0; 72.0]	48.0 [24.0; 48.0]	.13
Hospital length of stay (hours), median [Q25-75]	96.0 [72.0; 120]	96.0 [72.0; 120]	.16

Procedural complications were more frequent with hook-wire localization: 36% developed a pneumothorax (only one requiring chest tube drainage) and 8.6% an alveolar Haemorrhage. No haemothorax or other hemorrhagic events were observed (Table 2). No complication related to CBCT-guided localization were observed.

Table 2.

Perioperative Results

		Hook-wire group (n, %)	CBCT group (n, %)	P
Diagnostic yield		207 (89%)	32 (96.9%)	.053
Alveolar haemorrhage		20 (8.6%)	0 (0%)	.15
Conversion to thoracotomy		14 (6%)	2 (6.1%)	1
Conversion for targeting failure		4 (1.7%)	0 (0%)	1
Lymph node dissection		88 (38%)	23 (70%)	<.001
Wedge		222 (95%)	26 (79%)	<.01
Segmentectomy		17 (7.3%)	0 (0%)	.14
Lobectomy		62 (27%)	13 (39%)	.13
Postoperative complicationsa	0	198 (85%)	28 (85%)	.63
	1	10 (4.3%)	1 (3%)	–
	2	13 (5.6%)	1 (3%)	–
	3	10 (4.3%)	2 (6.1%)	–
	4	2 (0.86%)	1 (3%)	–

^aComplications according to the Clavien-Dindo classification.¹⁷ Bold results are statistically significant.

Perioperative outcomes

Mean operative time (Figure 3) was significantly shorter in the hook-wire group (65 min) compared with CBCT (90 min; P = .042, Table 3). The extent of pulmonary resection differed significantly between groups (P = .01, Table 1).

Figure 3.

Box Plots of Operative Time According to Targeting Technique. Box Plots Show Median, Interquartile Range, and Range (Excluding Outliers). Comparisons Are Descriptive and Reflect Differences in Workflow Organization Rather than Procedural Efficiency

The overall thoracotomy conversion rate was 6.0% with hook wire and 6.1% with CBCT (P = 1.0). Conversion due to localization failure did not differ between groups (P = 1.0). Among the 16 conversions (14 hook wire, 2 CBCT), the main causes were pleural adhesions (n = 5), arterial injury (n = 3), and dissection difficulties (n = 3). Four conversions were attributed to poor hook-wire positioning.

Lymphadenectomy was more frequently performed in the CBCT group (P = .001, Table 1), including in 2 patients with benign nodules.

Nodule characteristics

Malignancy was confirmed in 78% of resected nodules in the hook-wire group and 82% in the CBCT group (P = .59, Table 4). No significant differences in histological subtypes were observed (Table 4).

Table 4.

Nodules’ Characteristics

	Hook wire group (n, %)	CBCT group (n, %)	P
Benign	52 (22%)	6 (19%)	.65
Malignant	181 (78%)	27 (82%)	.59
Adenocarcinoma	122 (52%)	20 (61%)	.37
Neuro endocrine tumour	6 (2.6%)	0 (0%)	1
Squamous cell carcinoma	13 (5.6%)	1 (3%)	1
Metastasis	34 (15%)	7 (21%)	.32
Solid nodule	117 (51%)	15 (45%)	.58
Part solid nodule	38 (16%)	8 (24%)	.26
Ground-glass opacity	78 (33%)	10 (30%)	.72

In the hook-wire group, 52 benign nodules were resected (50 by wedge resection, 2 by segmentectomy). No lobectomy for benign disease was performed, although one mediastinal lymphadenectomy was undertaken pending frozen section. In the CBCT group, 6 benign nodules were resected, all by wedge resection; 2 of these cases also had lymphadenectomy.

Postoperative outcomes

No significant differences were observed in postoperative complications between the groups (P = .63, Table 2). The absolute risk difference for overall postoperative complications between CBCT and hook-wire localization was −0.2% (95% CI, −12.1 to +11.8), indicating no clinically meaningful difference between techniques. Similarly, chest tube duration (P = .13) and hospital stay (P = .16) were comparable (Table 3).

COMMENT

In this multicentre retrospective study, hook-wire and CBCT localization both achieved high diagnostic yields prior to VATS resection of pulmonary nodules (88% and 96.9%, respectively), with no significant difference between the 2 techniques. These findings are consistent with previous reports, which describe success rates of 93%-100% for hook-wire^18,19 and up to 100% for CBCT.^4,20,21 The slightly lower hook-wire yield observed in our cohort may be explained by our stricter definition of success, which required both palpation of the nodule in the resected specimen and histological confirmation, whereas other studies often accept the presence of the wire on CT or indirect signs such as Haemorrhage despite potential wire displacement.

The 96.9% technical success rate reported with CBCT in our series must be interpreted with caution, as one targeting error occurred on a wrong nodule. Nevertheless, intraoperative 3D imaging clearly enhances accuracy, which may explain the trend towards higher yield. The borderline P value observed for diagnostic yield should be interpreted in light of the marked imbalance in group sizes and the limited number of patients in the CBCT group. The study was not powered to detect small differences in diagnostic yield between techniques, and the wide CI around the absolute risk difference reflects this limited precision. Importantly, the point estimate consistently favoured CBCT across analyses, suggesting a potential advantage that warrants confirmation in larger prospective studies.

Several groups have combined CBCT with other methods—such as hook-wire, coils, methylene blue, or electromagnetic navigation—to further increase precision and reduce complication risks.^4,20,21 Reported detection rates in these hybrid approaches range from 94% to 100%, though often at the cost of increased operative time and radiation exposure.

Complication rates in our series were within the ranges reported in the literature. In the hook-wire group, pneumothorax occurred in 36% of patients, which is lower than previous reports (7%-53%),^18,22,23,24 and only one case required chest tube placement (0.4%), consistent with published data (0%-4.6%).²² Alveolar Haemorrhage occurred in 8.6%, also lower than the 13%-36% range described in Asian series.¹⁸ No haemothorax or air embolism was observed.²⁵ With CBCT, pneumothorax is inherent to VATS and not procedure-related, and no additional complications were noted. These results confirm the overall safety of both techniques.

Hook-wire localization was associated with significantly shorter operative times, as the procedure was performed beforehand in the radiology department, allowing the surgical team to optimize operating room use. By contrast, CBCT required intraoperative imaging, lengthening surgical time despite a shorter localization step. Similar increases in operating room occupancy have been reported when CBCT is combined with hook-wire, coils, or electromagnetic navigation (+30—45 min on average).^4,19 Thus, the choice of technique should account for institutional workflow and resource allocation. Interpretation of operative time comparisons requires caution, as hook-wire localization was performed in the radiology suite prior to surgery, whereas CBCT localization and resection occurred during a single session in the hybrid operating room. These differences reflect distinct organizational workflows rather than procedural inefficiency.

Malignancy rates were high in both groups (78%-82%), reflecting appropriate selection criteria based on radiological features and tumour board decisions. The proportion of benign nodules resected was higher in the hook-wire group (22%) compared with reported rates of 10%-16%,^11,18 possibly related to the higher frequency of multiple nodules in this group. Most benign lesions were treated by wedge resection, with only rare segmentectomies for deep lesions. Importantly, all patients in our cohort had a complete pathological diagnosis, highlighting the value of surgical excision compared with percutaneous or endoscopic biopsy, which carries lower diagnostic yields.²⁶ Localization techniques also assist the pathologist in identifying the region of interest within the specimen.

Several perioperative differences between groups were observed, including a higher rate of lymphadenectomy and a greater proportion of anatomical resections in the CBCT group. These differences were not directly related to the localization technique itself but rather reflected variations in surgical strategy and institutional practice. As such, they should be considered potential confounders when interpreting secondary outcomes such as operative time or postoperative morbidity. Importantly, the primary end-point—diagnostic yield—was defined independently of resection extent or lymphadenectomy, limiting the impact of these factors on the main study objective.

This study has several limitations. Its retrospective design introduces potential bias, and the limited number of patients in the CBCT group reduces statistical power. Allocation of technique was centre-dependent, precluding randomization. However, sensitivity analyses adjusting for centre and restricted to centre where both techniques were available showed similar trends, suggesting that the observed differences were not solely driven by centre-specific workflows. Radiation exposure could not be directly compared due to different measurement units, although mean doses were 452 mGy·cm in the hook-wire group and 3380 µGy·m² in the CBCT group. For CT-guided hook-wire localization, radiation exposure was expressed as dose-length product (DLP, mGy·cm), which reflects the total radiation dose integrated over the scanned length during conventional computed tomography acquisition. For CBCT-guided localization performed in the hybrid operating room, radiation exposure was reported as dose-area product (DAP, µGy·m²). DAP represents the integral of air kerma across the irradiated field. DAP reflects both radiation intensity and exposed surface area but does not directly correspond to scan length or volumetric dose. Because DLP and DAP quantify radiation exposure using fundamentally different physical principles and geometrical assumptions, direct conversion between these metrics is not reliable without additional assumptions regarding beam geometry, patient size, and tissue absorption. Consequently, radiation exposure associated with CT-guided hook-wire and CBCT-guided localization could not be directly compared in this study. Radiation metrics were therefore reported descriptively for each technique, in accordance with current reporting standards.

Finally, long-term oncological outcomes were not assessed.

Neither hook-wire nor CBCT demonstrated superiority in diagnostic yield, but each technique offers distinct advantages. Hook-wire is simple, widely available, and performed outside the operating room, whereas CBCT provides real-time intraoperative guidance but requires a hybrid suite and greater resource mobilization. Ultimately, the choice should depend on local expertise, logistics, and cost considerations, while maintaining the priority of achieving a definitive pathological diagnosis with minimal patient risk. As technological advances continue, localization techniques are likely to become an integral part of both the diagnostic and therapeutic management of pulmonary nodules, in close collaboration between radiologists and thoracic surgeons.

SUPPLEMENTARY MATERIAL

Supplementary material is available at ICVTS online.

FUNDING

No funding supported this manuscript.

CONFLICTS OF INTEREST

The authors have no disclosure regarding this manuscript. The current manuscript has been read and approved by all co-authors.

No generative artificial intelligence tools were used in the preparation of this manuscript.

DATA AVAILABILITY

The data underlying this article will be shared on reasonable request to the corresponding author.

INFORMED CONSENT STATEMENT

Patients gave their informed consents for using their medical data for the purpose of the presented study. Written informed consent was obtained from all patients.

ETHICS COMMITTEE NUMBER AND DATE

2022/CE61 approved on December 7, 2022. The committee approved the use of the database according to the WMA Declaration of Taipei requirements.