Diagnostic Accuracy and Complication Rates of Fusion Images Created Using Real-Time Ultrasound with CT for Identification of Peripheral Lung Lesions in Patients Undergoing Biopsy

Background: Fusion image improves lesion detectability and can be an effective tool for percutaneous ultrasound (US)-guide procedure. We describe the clinical benefit of US-guided lung biopsy using fusion image. Purpose: To retrospectively compare the diagnostic accuracy and complication rates of US-guided lung biopsy with B-mode alone and those of a fusion image created using real-time US and computed tomography (CT). Materials and Methods: Between September, 2013 and September, 2016, 50 peripheral lung lesions in 50 patients (40 males, 10 females; median, 74 years old) were performed by US-guided percutaneous cutting needle biopsy using the B-mode alone or fusion image. Final diagnoses were based on surgical outcomes or clinical follow-up results for at least 12 months after biopsy. To assess prebiopsy characteristics, all lesions were divided into two groups: group 1 (identification on B-mode) and group 2 (identification on fusion image). Results: Of 50 peripheral lesions, 40 lesions (80%) were detected by means of B-mode alone (group 1), and 10 lesions (20%) were identified by fusion image (group 2). The diagnostic accuracy of group 1 was 90% (36/40 lesions), and the diagnostic accuracy of group 2 was 100% (10/10 lesions). Nodule type and the size of the lesions showed significant group wise differences (p < 0.001 and p = 0.02, respectively). Pneumothorax occurred immediately after the first How to cite this paper: Imamine, R., Kobayashi, H., Akuta, K., Matsuki, M., Isoda, H. and Togashi, K. (2019) Diagnostic Accuracy and Complication Rates of Fusion Images Created Using Real-Time Ultrasound with CT for Identification of Peripheral Lung Lesions in Patients Undergoing Biopsy. Open Journal of Radiology, 9, 36-47. https://doi.org/10.4236/ojrad.2019.91004 Received: January 8, 2019 Accepted: January 22, 2019 Published: January 25, 2019 Copyright © 2019 by author(s) and Scientific Research Publishing Inc. This work is licensed under the Creative Commons Attribution International License (CC BY 4.0). http://creativecommons.org/licenses/by/4.0/ Open Access


Introduction
In the era of personalized medicine in oncology, the requirement for a least invasive biopsy technique that can provide accurate diagnosis is becoming increasingly important.Several investigators have described the utility and safety of ultrasound (US)-guided transthoracic cutting needle biopsy and fine needle aspiration in chest lesions [1]- [10].However, these studies were conducted only in the conventional B mode, and thus occurrence of mis-targeting is possible.Although careful US examination is required to detect the contact area between the peripheral lung lesion and the pleura, conditions such as pleural adhesions and bullous emphysema may interfere with US-visualization of the pleural surface [11] [12].
Recently, fusion image with real-time US and computed tomography (CT) has been introduced in interventional radiology [13] [14].In the chest, fusion image can be used to assist in detecting the pleural-lesion contact area in extremely small solid nodules.Moreover, the fusion image of the pleural membrane associated with solid component of part-solid ground-glass nodule (GGN), results in localization of the lesion.The effectiveness of fusion image for US-guided lung biopsy has not been reported.The purpose of this study was to compare the diagnostic accuracy and complication rates associated with US-guided lung biopsy with B-mode alone and those of fusion images created using real-time US and CT.

Patients
The institutional review board approved this retrospective study and waived the requirement for informed consent.This retrospective study was conducted at a single institution.Between September, 2013 and September, 2016, US-guided percutaneous cutting needle biopsy was conducted for 63 thoracic lesions.Biopsy was performed for all patients within 4 weeks after a chest lesion was detected with CT.Of these, 12 mediastinal lesions and one lung lesion were excluded because of the use of contrast-enhanced US for detection of the lesion [15]

US Examination and US-Guided Biopsy Procedure
All patients were evaluated using the GE LOGIQ Q9 US imaging system (GE Healthcare Worldwide, Chicago, IL, USA).US examinations were performed by an interventional radiologist (with experience of >30 years).Patients were placed in the supine, prone, or right/left lateral decubitus positions.The indication of fusion image depended on the same interventional radiologist.Initially, US using the conventional B-mode was conducted to identify CT-detected lesions.
When a solid nodule was not identified through conventional B-mode, fusion image with real-time US and CT was performed using a commercially available US unit (Figure 1).Moreover, fusion image was successful in localizing the part-solid GGN with pleural involvement, despite non-detection of the ground-glass opacity within the sonographic window (Figure 2).For the fusion image, the manubrium of the sternum served as a reference point to synchronize the spatial relationship between the CT and US images in supine position.In the prone or lateral position, the first or 12 th spinous process was used as a reference point.Once because no pathologist was on site.If the patients complained of symptoms associated with pneumothorax or hemoptysis after the procedure, a chest CT was obtained.In addition, posteroanterior chest radiographs of all patients were obtained on the following day.

Data Collection and Study Definition
One interventional fellowship-trained radiologist (with experience of >10 years) evaluated medical records for consecutive patients presenting for US-guided biopsies and assessed the lesions' characteristics and procedure parameters.One radiologist (with experience of >30 years) and the same interventional fellowship-trained radiologist independently measured the longest diameters of all lesions on CT.The longest diameter was measured along the long axis of the lesion on the lung window setting of CT.The lesion-pleura contact arc length (LPCAL) of all lesions was measured on CT twice on different days by the same interventional fellowship-trained radiologist.The pleural lesion contact area was defined as a cross-sectional view of the peripheral lung lesion in contact with the overlying pleural surface in the CT mediastinum setting.The LPCAL was measured as a straight-line approximation, in order to obtain the total value of each value (<10 mm) separately measured along the maximum curvature length of the lesion on the mediastinal window setting of CT.The mean values were accepted for all quantitative data.The lesions were divided into two groups: group 1 (identification on B-mode) and group 2 (identification on fusion image).
The pathological results of the specimens were classified as malignant, benign, and non-diagnostic.Insufficient specimens with a paucity of suspicious cellularity were considered non-diagnostic.For statistical analysis, the pathologically malignant results were regarded as positive in biopsy, while the pathologically benign results were considered as negative in biopsy.The diagnostic accuracy of US-guided biopsy was calculated using the follow formula: diagnostic accuracy (%) = number of lesions accurately diagnosed (true positive + true negative)/total number of lesions.A true-positive case included a malignant biopsy result confirmed by the surgical specimen or the post-procedural malignant clinical course.A true-negative case included a benign biopsy result together with surgical confirmation or the subsequent clinical course showing no increase in size or the disappearance of lesion with or without administration of antibiotics on follow-up CT for at least 12 months.A case was considered as a false negative when a benign biopsy result was contradicted by the finding of malignancy at surgical resection, biopsy of other organs, or tumor growth observed at follow-up CT examination (performed 3 months after biopsy).Tumor growth was defined according to the Response Evaluation Criteria in Solid Tumors (RECIST).Final diagnosis was defined as a clinical diagnosis at the end of the follow-up period or by September, 2017.

Statistical Analyses
tween the two groups, the Mann-Whitney's U test was used to compare quantitative variables, and the chi-square test was used to compare qualitative variables.
A p value < 0.05 was considered to indicate a statistically significant difference.
Data processing and analysis were performed with commercially available software (MedCalc version 16.8.4;MedCalc Software, Ostend, Belgium).

Results
Of 50 peripheral lesions, 40 lesions (80%) were detected by means of B-mode alone (group 1), and 10 lesions (20%) were identified by means of fusion image (group 2).In this study, no patient was noted as a registration error in the image fusion approach.The patients' and lesions' characteristics, and procedure parameters between the two groups are shown in Table 1.Nodule type (p < 0.001) and size (p = 0.02) of the lesions showed significant differences between the two groups, whereas the other values showed no significant differences.Although pneumothorax occurred immediately after the first puncture in five of 50 (10%) lesions, no symptoms developed in all patients; and the sizes of the five lesions ranged from 4 mm to 20 mm.These five patients had sufficient tissues available for pathological diagnosis of malignancy.The final diagnosis is summarized in Table 2.As malignant lesions, 36 lesions of primary lung cancer (non-small cell lung cancer/small cell lung cancer = 31/5), 5 lesions of lung metastasis from other organs, 1 lesions of malignant mesothelioma were confirmed.And the remaining 8 lesions were benign lesions (epithelioid cell granuloma/nonspecific = 2/6).The definitive diagnoses in group 1 were confirmed by surgery (n = 11) or clinical course (n = 29), and 10 lesions in group 2 were confirmed by surgery (n = 4) or clinical course (n = 6).Diagnostic accuracy between the two groups is shown in Table 3.The diagnostic accuracy of group1 and group 2 were 90% and 100%, respectively.Four lesions with diagnostic failure (three pathological benign lesions and one non-diagnostic lesion) are summarized in Table 4.Of these, malignancy was confirmed in two lesions by lobectomy (No. 1 and No. 3), in one lesion by repeat US-guided lung biopsy (No. 2), and in one lesion by US-guided subclavicular lymph node biopsy (No. 4).With regard to the relationship between lesion-size and LPCAL, the diagnostic accuracy for malignant and benign lesions is summarized in Table 5.All lesions' characteristics using fusion image are shown in Table 6.In this study, life-threatening major complications and hemoptysis were not observed.

Figure 1 .Figure 2 .
Figure 1.Pulmonary nodule in a 61-year-old woman with a history of minor salivary gland tumor (adenoid cystic carcinoma) and breast cancer.(a) Chest CT scan showed a 4-mm peripheral solid nodule (arrow) in the right middle lobe; (b) Although B-mode alone failed to reveal the solid nodule, localization of the solid nodule was possible using fusion image with real-time US (white arrow) and CT (black arrow); (c) The inserted line was visualized as a dotted line; subsequently, the biopsy needle (short arrows) was introduced into the lesion (long arrows) along the dotted line.Only one puncture was performed due to pneumothorax.The biopsy revealed metastatic lesion of minor salivary gland tumor (adenoid cystic carcinoma).

Table 1 .
Patients' and lesions' characteristics and procedure parameters between the two groups.

Table 2 .
Final diagnosis between the two groups.

Table 3 .
Diagnostic accuracy between the two groups.

Table 4 .
Results of lesions with diagnostic failure.

Table 5 .
Results of biopsies according to size and LPCAL between the two groups.: true positive, FN: false negative, ND: non-diagnostic, SB: specific benign, NB: non-specific benign. TP

Table 6 .
Results of biopsy lesions using fusion image.