Preoperative versus intraoperative image-guided localization of multiple ipsilateral lung nodules

Abstract

Open in new tabDownload slide

OBJECTIVES

Computed tomography (CT)-guided localization of multiple ipsilateral pulmonary nodules remains challenging. Hybrid operating rooms equipped with cone-beam CT and laser navigation systems have the potential for improving clinical workflows and patient outcomes.

METHODS

Patients with multiple ipsilateral pulmonary nodules requiring localization were divided according to the localization method [preoperative CT-guided (POCT group) localization versus intraoperative CT-guided (IOCT group) localization]. The 2 groups were compared in terms of procedural efficacy, safety and radiation exposure.

RESULTS

Patients in the IOCT (n = 12) and POCT (n = 42) groups did not differ in terms of demographic and tumour characteristics. Moreover, the success and complication rates were similar. Notably, the IOCT approach allowed multiple nodules to be almost simultaneously localized—resulting in a shorter procedural time [mean difference (MD) −15.83 min, 95% confidence interval (CI) −7.97 to −23.69 min] and lower radiation exposure (MD −15.59 mSv, 95% CI −7.76 to −23.42 mSv) compared with the POCT approach. However, the total time under general anaesthesia was significantly longer in the IOCT group (MD 34.96 min, 95% CI 1.48–68.42 min), despite a similar operating time. The excess time under anaesthesia in the IOCT group can be attributed not only to the procedure per se but also to a longer surgical preparation time (MD 21.63 min, 95% CI 10.07–33.19 min).

CONCLUSIONS

Compared with the POCT approach, IOCT-guided localization performed in a hybrid operating room is associated with a shorter procedural time and less radiation exposure, albeit at the expense of an increased time under general anaesthesia.

INTRODUCTION

The advent of low-dose computed tomography (CT) screening for lung cancer has markedly increased the number of small pulmonary nodules detected, which require either biopsy or surgical removal [1]. An accurate lesion localization is an essential prerequisite for their successful surgical removal through video-assisted thoracoscopic surgery (VATS) [2, 3]. Before transferring the patient to an operating room, most pulmonary nodules are currently localized using either the hookwire or dye injection techniques performed in an interventional radiology suite [4–6]. Although this approach resulted in a high procedural success rate with a low incidence of complications [4, 6], its implementation in patients presenting with multiple nodules is technically demanding [7–9]. Under these circumstances, the main limitation lies in the need for multiple pleural punctures—ultimately increasing the risk of pneumothorax. Other caveats of this technique include the prolonged procedural time (which may cause anxiety and discomfort) and the increased radiation exposure.

Theoretically, all of these limitations may be overcome by implementing a single-stage lesion localization and removal performed in a hybrid operating room (HOR) equipped with a suitable localization system [10]. The objective of this study is to present our initial experience with intraoperative CT (IOCT)-guided localization of multiple ipsilateral pulmonary nodules implemented within a HOR environment. We also sought to compare this new technique with the traditional preoperative CT (POCT)-guided approach in terms of procedural efficacy, safety and radiation exposure.

METHODS

Study patients

This is a retrospective study of prospectively collected data aimed at comparing the IOCT group approach with the traditional 2-stage POCT group technique for pulmonary lesion localization and removal. The study sample consisted of consecutive patients presenting with multiple ipsilateral pulmonary nodules requiring localization before surgical removal between 1 February 2012 and 1 April 2019. The study protocol was reviewed and approved by the local Institutional Review Board (CGMH-IRB 201600671A3).

Indications for localization and technique selection

Patients presenting with solid lesions underwent localization in the presence of small (diameter < 10 mm) and/or deeply located (distance from the visceral pleura > 10 mm) nodules. We also localized all of subpleural cavitary lesions and subsolid nodules—regardless of their depth and/or size. Since its introduction in 2007, POCT-guided localization has been the standard approach in our hospital. However, a HOR has been opened in our hospital and shared with the cardiovascular department as of 2016. We consequently implemented the IOCT approach whenever the HOR was available. If this was not the case, the traditional POCT technique was utilized.

Preoperative computed tomography-guided localization

CT-guided localization was performed by an experienced interventional radiologist [11]. Patients were positioned in the CT scanner (GE HiSpeed, Milwaukee, WI, USA) in order to achieve the shortest direct distance from the skin to the lesion of interest. CT images were obtained at a slice thickness of 2.5 mm. A direct vertical needle trajectory was used to reach the target lesion whenever possible. The puncture site was cleansed carefully and a small incision was made with a scalpel through the skin under local anaesthesia. Thereafter, a 10.7-cm long, 20-gauge cannula needle housing a 20-cm long double-thorn hookwire (DuaLok^®, Bard Peripheral Vascular, Inc., Tempe, AZ, USA) was gradually inserted into the chest wall under sequential CT guidance. If feasible, the pulmonary lesion was pierced through the cannula needle. Subsequently, the hookwire was advanced along the cannula until the needle tip was correctly positioned within the nodule of interest or in its close proximity. Superficial lesions were localized through the injection of patent blue V dye (0.5 ml, patent blue V 2.5%; Guerbet, Aulnay-sous-Bois, France) via a 22-gauge spinal needle (length: 8.9 cm). Upon completion of the first localization, the patient was repositioned to localize all of the remaining lesions of interest with the same technique. In case of occurrence of a pneumothorax precluding additional lung punctures, the procedure was interrupted. A final CT scan was acquired to assess complications (e.g. pneumothorax or pulmonary haemorrhage). Patients were transferred to a general ward prior to surgery.

Intraoperative computed tomography-guided localization

IOCT-guided localization was carried out in a HOR equipped with a C-arm cone-beam CT apparatus (ARTIS zeego; Siemens Healthcare GmbH, Erlangen, Germany) and a Magnus surgical table (Maquet Medical Systems, Wayne, NJ, USA). Both localization and surgery were performed by a single team of thoracic surgeons. The procedural workflow has been previously reported [10, 12]. In brief, we placed the patient in the lateral decubitus position upon induction of general anaesthesia. An initial scan for surgical planning was obtained during end-inspiration breath-holding using a 6-s acquisition protocol (6 s DynaCT Body). We modelled the entering trajectory in the isotropic data set under the syngo Needle Guidance of a syngo X-Workplace (Siemens Healthcare GmbH). After measuring the distances between the skin and the pleura and from the pleura to the lesion (Fig. 1A), we marked the skin needle entry by projecting a laser-target cross onto the patient’s surface. After puncture, the needle was properly fixed in the subcutaneous tissue (Fig. 1B) and the targeting procedure was repeated for the second nodule (Fig. 2). Upon completion of all subcutaneous insertions, the needles were advanced simultaneously into the lesions of interest. A second CT scan was acquired to confirm their appropriate positioning before the simultaneous release of the markers. Video 1 illustrates the procedural details.

Surgical treatment

Upon completion of VATS wedge resection (under hookwire or dye guidance), the surgical specimen was cut along the lesion maximum diameter to allow macroscopic inspection of the surgical margins. Lobectomy was performed in patients with adequate pulmonary function who had a confirmed diagnosis of primary lung cancer with either a tumour size ≥2 cm or a tumour size <2 cm in the presence of resection margins lower than the tumour diameter.

Outcome assessment

Accuracy, efficacy and safety indices were the main outcomes of interest and were defined as follows. The accuracy index included (i) the rate of successful targeting (calculated as the number of successful targeting procedures divided by the number of all localization procedures) and (ii) successful targeting rates in the operating field (defined as the number of successful targeting procedures minus the number of wire dislodgements or dye fading/spillage occurring in the operation field divided by the number of all localization procedures). The efficacy index was defined as the time required for tumour localization. Specifically, it was calculated from the start of the preprocedural CT to the completion of the postprocedural CT scan in the POCT group. In patients who underwent IOCT, it was defined as the time elapsed from the docking of the C-arm to the end of the localization procedure (i.e. when the C-arm was retracted from the table to the park position). Other efficacy measures included (i) the time required for surgical preparation—which was defined as the time between completion of the general anaesthesia and skin incision in the POCT group and as the time between completion of general anaesthesia and C-arm docking plus C-arm parking to skin incision in the ICOT group; (ii) the time at risk, which was defined as the time between the completion of localization and skin incision; (iii) the operating time; and (iv) the time under general anaesthesia. The occurrence of complications (i.e. pneumothorax and lung haemorrhage) served as the safety index and was recorded upon completion of the follow-up CT scan immediately following localization. In accordance with the 2010 British Thoracic Society guidelines, the size of a pneumothorax was divided into ‘large’ and ‘small’ depending on the presence of a rim <2 cm or ≥2 cm between the lung margin and the chest wall, respectively [13]. The radiation dose delivered to patients was quantified by calculating the effective dose (ED). The radiation dose of multi-detector computed tomography implemented during the POCT procedure was expressed as the dose length product (mGy-cm). The radiation dose was subsequently converted to an ED using an appropriate conversion factor (0.014 mSvGy⁻¹cm⁻¹) [14]. As far as IOCT procedures are concerned, cone beam computed tomography and fluoroscopy doses were expressed as dose area product (mGy × cm²). Conversion factors of 0.146 and 0.12 mSvGy⁻¹cm⁻² were used for ED calculation, respectively [15, 16].

Statistical analysis

Continuous variables were given as means ± standard deviations and compared using 2-sample Student’s t-tests. Intergroup comparisons of continuous data were performed by calculating mean differences (MDs) and their 95% confidence interval (CI) taking the POCT arm as reference. Categorical data were reported as counts and percentages and compared with the χ² test or the Fisher’s exact test (when a cell value was lower than 5). Intergroup comparisons of categorical variables were performed by calculating the percentage differences and their 95% CI taking the POCT arm as reference. All analyses were conducted with the statistical software R (R Foundation for Statistical Computing, Vienna, Austria; http://www.R-project.org/).

RESULTS

Characteristics of patients and pulmonary nodules

During the study period, we identified 57 patients with multiple ipsilateral pulmonary nodules requiring localization. After the exclusion of 3 patients with missing clinical data, the IOCT and POCT groups consisted of 12 and 42 patients who harboured 24 and 84 pulmonary lesions, respectively. Table 1 summarizes the general characteristics of the study patients. According to preoperative CT findings, 39 (36.1%) nodules were classified as solid, whereas the remaining 69 (63.9%) were subsolid. The mean lesion size on preoperative CT images was 6.79 mm, whereas the mean tumour depth-to-size ratio was 1.39.

Safety and efficacy of intraoperative computed tomography-guided versus preoperative computed tomography-guided localization

Table 2 depicts the accuracy, safety and radiation exposures associated with each localization procedure. All of the 12 (100%) patients in the IOCT group successfully underwent localization in a single position (i.e. lateral decubitus). In contrast, 4 (9.5%) patients in the POCT group required position changes during the first and second tumour localization (from supine to prone or from prone to supine). Overall, we successfully localized 106 nodules—with 1 failure per group. In the POCT group, the failure occurred during localization of the second lung nodule because of a large pneumothorax. Although all nodules underwent complete localization in the IOCT group, postprocedural cone-beam CT imaging revealed a case of needle displacement (which was at a 3-cm distance from the target). The successful targeting rates during localization in the POCT and IOCT groups were 98.8% and 95.8%, respectively (P = 0.397).

According to postprocedural cone-beam CT results (or CT findings), a total of 15 (27.8%) patients had pneumothorax—without significant intergroup differences [IOCT group (25%) vs POCT group (28.6%)]. Most cases of pneumothorax were small, the only exception being 2 cases of large pneumothorax in the POCT group (of whom one was treated with pigtail drainage).

Radiation exposure—expressed by ED—was also significantly lower in the IOCT group compared with the POCT group (MD −15.59 mSv, 95% CI −7.76 to −23.42 mSv). Table 3 summarizes the main efficacy indices. The IOCT group was characterized by a shorter localization procedural time (MD −15.83 min, 95% CI −7.97 to −23.59 min). Because all procedures were performed in a HOR, the patient time at risk was also significantly lower in the IOCT group than in the POCT group (MD −189.52 min, 95% CI −160.91 to −218.14 min). However, the total time under general anaesthesia was significantly longer in the IOCT group (MD 34.96 min, 95% CI 1.48–68.42 min), which can be attributed not only to the procedure per se but also to a longer surgical preparation time (MD 21.63 min, 95% CI 10.07–33.19 min).

Table 4 depicts the operative and perioperative outcomes. Among the 106 nodules for which the localization procedure was successful, we failed to complete marker-guided tumour VATS resection for 3 nodules, either because of wire dislodgement (n = 2) in POCT group or dye spillage (n = 1) in IOCT group. Therefore, the rate of successful marker-guided resection was 95.4%, without significant intergroup differences. However, the scheduled thoracoscopic resection was successfully performed under guidance of the puncture hole positioned on the lung surface. There were no cases of open conversion. All patients initially underwent wedge resection. The findings from intraoperative frozen sections revealed the presence of 64 malignancies (45 primary tumours and 19 metastases) and 44 benign lesions. In the presence of 2 distinct lesions, the distribution was as follows: 2 malignancies (n = 28), 1 malignancy and 1 benign lesion (n = 8) and 2 benign lesions (n = 18). Patients who had an intraoperative frozen section diagnosis of either malignancies or precancerous lesions underwent systematic lymph node dissection. Two patients in the POCT group who were intraoperatively diagnosed with primary lung tumours had an inadequate resection margin upon gross lesion inspection. One of these patients underwent lobectomy, which was unfeasible in the other case because of a poor pulmonary reserve. There were no in-hospital deaths. The mean length of hospital stay after surgery was 4.69 days, without significant intergroup differences.

DISCUSSION

To our knowledge, this study is the first to compare the efficacy and safety of IOCT versus POCT for localizing multiple ipsilateral pulmonary nodules. Besides the benefit of decreasing the time at risk by taking advantage of performing localization in a HOR, IOCT-guided localization for multiple ipsilateral lung tumour localization also reduces the procedural time and decreases the radiation exposure compared with the POCT approach. Our pilot data suggest that IOCT-guided localization may be a valuable approach in the presence of multiple ipsilateral pulmonary nodules.

In our experience, 2 major factors may have contributed to the superior performance of IOCT-guided localization. First, the multiaxial robot C-arm mounted on the cone-beam CT apparatus in the HOR allows to complete localization with the patients of the IOCT group lying in the lateral decubitus position—ultimately allowing the widest possible access for the transthoracic puncture and potentially eliminating the need for repositioning between the first and second puncture. Although a similar approach could theoretically be implemented during POCT-guided localization as well, several reasons may practically hamper its use. As the examination table is not specifically designed for the lateral decubitus position, it may be difficult for the patient to remain motion-free when the examination table slides into and out of the CT scanner. Furthermore, even though localization can be accomplished in the lateral decubitus position, its maintenance during transfer to the operating room may be troublesome. Second, the accurate needle path outlined by the syngo Needle Guidance software allowed to carry out 2 consecutive pleural punctures almost simultaneously (Video 1). This approach not only minimized the time lag between 2 punctures (ultimately decreasing the procedural time), but also reduced the total number of CT scans required throughout the entire procedure (thereby lowering the radiation exposure).

Previous reports focusing on the POCT approach identified intraprocedural position changes as the main risk factor for pneumothorax development [8]. Although the IOCT approach does actually abrogate the need for patient repositioning, we were unable to reduce significantly the pneumothorax rate. One potential explanation for this phenomenon is the use of positive pressure ventilation during the IOCT procedure. It is indeed well-known that inflation of the lung under positive pressure may predispose to pneumothorax after pleural puncture. However, all cases of pneumothorax in the IOCT group were classified as small and none of them required drainage.

Despite a shorter procedural time and a lower radiation exposure observed in the IOCT group, localization in the HOR significantly increased the time under general anaesthesia. The excess time under anaesthesia in the IOCT group can be attributed not only to the procedure per se but also to a longer surgical preparation time. Specifically, the use of IOCT required the preparation of sufficiently long cables and tubes to enable a collision-free C-arm rotation during the CBCT scan—ultimately increasing the preparation complexity. Another potential caveat of localization performed in a HOR is the increased risk of air embolism [17]. When a needle simultaneously crosses an air-containing space and the adjacent pulmonary vein, a fistula may form and air can enter the vein when alveolar air pressure exceeds the pulmonary venous pressure (an event that, for example, may be triggered by positive ventilation) [17]. To prevent air embolism, the introducer needle must be occluded by the inner stylet, saline drops or a finger. Moreover, the patient should suspend respiration when the needle is manipulated [18].

Limitations

Several limitations of our study merit comment. First, the selection of markers was affected by the physician’s preferences and was not uniform across groups. Although more patients in the POCT group received hookwire localization, there were no significant intergroup differences in terms of D/S ratio. We believe that this imbalance may at least in part reflect the radiologist’s preference. Unfortunately, we were unable to address this issue through a propensity-matched analysis because of the small sample size, which would ultimately have been underpowered. The marker choice is unlikely to have had a major impact on our results, at least in term of success rates and procedural time [19]. However, we are aware that markers may have an impact on the occurrence of pneumothorax—although this was not the case in our study [19]. Second, IOCT has become feasible in our institute as of 2016 only and its use thoracic surgery has remained limited. Consequently, allocation to either IOCT or POCT for lesion localization was not randomized. Third, the IOCT-guided approach significantly increased the time under anaesthesia—potentially resulting in higher medical costs. Unfortunately, we were unable to provide a more in-depth economic analysis due to the lack of cost data. Fourth, this study was aimed at investigating 2 different percutaneous CT-guided approaches. Consequently, we did not compare IOCT or POCT with other non-percutaneous techniques (e.g. electromagnetic navigation bronchoscopy or virtual bronchoscopy-guided marker injection) [20–22]. Future ad hoc studies will be required to specifically analyse the efficacy and cost-effectiveness of percutaneous versus non-percutaneous localization approaches.

CONCLUSIONS

Compared with the POCT approach for localizing multiple ipsilateral pulmonary nodules, IOCT-guided localization performed in a HOR is associated with a shorter procedural time and less radiation exposure, albeit at the expense of an increased time under general anaesthesia.

Funding

This study was financially supported by a grant [CMRPG3F1813] from the Chang Gung Memorial Hospital, Taiwan.

Conflict of interest: none declared.

Author contributions

Yin-Kai Chao: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Resources; Supervision; Validation; Writing—Original draft; Writing—Review & Editing. Hsin-Yueh Fang: Data curation; Formal analysis; Methodology; Writing—Original draft. Kuang-Tse Pan: Conceptualization; Data curation; Methodology; Writing—Review & Editing. Chih-Tsung Wen: Conceptualization; Data curation; Methodology; Writing—Review & Editing. Ming-Ju Hsieh: Supervision; Validation.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• CIs

  Confidence interval

 
• CT

  Computed tomography

 
• ED

  Effective dose

 
• HORs

  Hybrid operating rooms

 
• IOCT

  Intraoperative CT-guided

 
• MD

  Mean difference

 
• POCT

  Preoperative CT-guided

 
• VATS

  Video-assisted thoracoscopic surgery