CT-Guided Lung Biopsies Performed with A 16 Gauge Needle and Non-Coaxial Technique: Preliminary Experience with Operator Learning Curve Analysis in Open Access Journal of Medical and Clinical Surgery
Rational and objectives: To assess the diagnostic performance and safety of lung biopsies performed under computed tomography guidance (CTLB) with non-coaxial technique and a 16 Gauge aspiration needle, with a special focus on the operator learning curve.
Materials and Methods: Retrospective observational analysis of 120 CTLB performed with a 16-gauge needle and a non-coaxial technique (study period: January 2017 – April 2020). Patient-related, lesion-related, procedural-related variables and histological results were recorded for each procedure. Study population was divided into two groups based on when the procedures were performed (group A: January 2017 –October 2018; group B: November 2018 – April 2020). Mean procedural time, diagnostic performance and complication rate were compared between the two groups. Appropriate tests were applied for inferential statistical analysis.
Results: Overall sensitivity, specificity, and diagnostic accuracy were 90.2 %, 100.0 %, and 90.8 % respectively. Operator learning curve showed a significant improvement of procedural time, diagnostic accuracy, and sensitivity in group B vs group A (sensitivity: 98.2 % vs 83.5 %; specificity: 100.0 % vs 100.0 %; diagnostic accuracy: 98.3 % vs 83.3%). Complication rate was similar between group A and B. Lesser lesion diameter (p < 0.001) and the presence of emphysema along the needle path (p < 0.001), were associated with increased complication rate, in particular of pneumothorax (p=0.001). Extended contact surface between the lesion and the pleura (p=0.001) resulted a protective factor from procedural complications.
Conclusion: The use of the 16G needles seems a feasible and effective option especially with clinical practice over time.
Keywords: Needle biopsy, CT-guided biopsy, Interventional Oncology, Interventional Radiology, Lung Biopsy Learning Curve.
Lung biopsy performed under computed tomography guidance (CTLB) is an established diagnostic tool to obtain histological diagnosis of pulmonary lesions of unknown origin [1]. The recent development of targeted and immunological therapies for lung cancer increased the need to collect adequate material to achieve the correct diagnosis and to define the molecular profile of the neoplasms in each patient [2]. Coaxial technique is frequently used by interventionalists to increase the amount of tissue obtained during CTLB but the risk of pneumothorax, bleeding and/or fatal air embolism is higher in comparison with non-coaxial technique [3,4]; on the other hand, diagnostic yield of non- coaxial technique has been reported to be lower [4]. The association of larger needles with non-coaxial technique could be an effective compromise between diagnostic performance and risk of complications, but there is lack of evidence about the use of 16 Gauge (G) needles in CTLB practice [5]. The present study reports a preliminary experience on CTLB performed using a 16 G needle in association to non-coaxial, focusing on procedural outcome and the operator learning curve.
All CTLB performed with non-coaxial technique and a 16 G needle, between March 2017 and April 2020, were retrospectively included. All the procedures were performed by an interventional radiologist without experience in CTLB performed with 16 G needle and non-coaxial technique. The study was approved by the Institutional Review Board and the requirement for written informed consent was waived due to the retrospective nature of the study. All patients signed an informed consent to the procedure at the time of examination.
For each CTLB, images from diagnostic CT and/or PET-CT (PET, Positron Emission Tomography) examinations were reviewed in order to decide patient's decubitus for the procedure. Orthogonal scanograms were obtained for lesion localization. A CT scan was performed at the lesion level and skin markers were used to plan the best needle path. After local anesthesia, a 16 G manual aspiration needle (TSK, Surecut, Japan) was advanced into the lesion and a single specimen was obtained under CT-guidance. A final CT scan was performed immediately after the procedure for the identification of intra-procedural complications (Figure 1).
Data collection and variables definition
Patient-related (age and sex), lesion-related (lesion size, CT appearance, PET-CT appearance, lesion location, pleural contact) and procedural-related (patient position, presence of emphysema, parenchymal needle path, procedural time, complications, and final diagnosis) variables were obtained from electronic medical records and radiological images from our database. Lesion related variables were defined by two radiologists in consensus. Lesion maximal diameter was measured on CT axial images with the lung window [6].
Lesion CT appearance was classified as solid or not solid, with or without cavitation, in accordance with the Fleischner society guidelines [7]. Positron Emission Tomography (PET) appearance of the lesions was distinguished in up-taking, up-taking with central photopenic area and not up-taking. Parenchymal needle path was measured from the pleural surface to the entry point into the lesion: when necessary, CT multiplanar reconstructions were used to unravel needle trajectory; when the needle trajectory did not across the lung parenchyma, parenchymal needle path was considered absent. Measurement of the lesion pleural contact was obtained at the level of the largest surface of contact between the lesion and the pleura (Figure 2). Procedural time was retrieved from CT images: first CT image after scanograms was considered the start point of the procedure, the last image of the final CT scan was considered as the end of the procedure.
Regarding complications, pneumothorax and parenchymal bleeding were considered minor complications when clinical irrelevant. Pneumothorax requiring percutaneous drainage and prolonged hospitalization (>2 days), hemorrhage with associated hemoptysis, air systemic embolism and death were considered major complications [8].
Histopathological results of the biopsy samples were compared with definitive histopathological diagnoses obtained from surgical specimens (when lesion surgical excision was performed) or obtained by other minimally invasive procedures, such as bronchoscopy, otherwise considering imaging follow-up. The biopsy results were considered true positive or true negative if CTLB malignant or benign results were in accordance with the final malignant or benign diagnoses respectively; otherwise, results were considered false positive (malignant result at CTLB not confirmed) or false negative (benign result at CTLB with final malignant diagnosis at surgical excision or malignant evolution at imaging follow-up). The study population was divided into two groups of equal size according to the date of execution of the procedure for the operator learning curve analysis: the first 60 patients, whose procedures were performed from March 2017 to October 2018, were enrolled in group A; the remaining 60 patients (November 2018 – April 2020) were assigned to group B. A descriptive and inferential analysis of procedural time, diagnostic performance and complication rate was performed on the overall study population and between the two groups.
Statistical analysis was performed with a dedicated software (SPSS 22, IBM). Continuous variables with normal distribution were reported as mean value ± standard deviation. Student’s t-test or one-way ANOVA were adopted to assess significant differences, as appropriate; correlation between continuous variables normally distributed was investigated with Pearson’s linear correlation. For continuous variables without normal distribution Mann-Whitney U test was employed for inferential statistics. For nominal and ordinal variables, the chi-square or Fisher exact test were applied, as appropriate. A p-value < 0.05 was considered as statistically significant.
A total of 1564 consecutive CTLB were performed in the period March 2017 – April 2020; among these, 120 CTLB performed on 120 patients matched with the inclusion criteria and were finally included (Figure 3). Patient-related and lesion-related variables of the entire study population and of groups A and B are reported in Table 1. Group A and group B did not show significant differences in the variables analyzed. The overall sensitivity, specificity, and diagnostic accuracy were 90.2 %, 100.0 %, and 90.8 % respectively (Table 2). A false negative result occurred in 11/120 (9.2%) CTLB; among these, eight patients showed a high lesion uptake at PET-CT examination with final malignant diagnosis after surgery (Figure 4), three patients showed volumetric increase at imaging follow-up.
The presence of a parenchymal needle path, especially in the presence of emphysema, was strongly associated with increased complication rate (p < 0.001), in particular with pneumothorax (p < 0.001); the same result was observed for lesions with a lesser diameter (p < 0.001). An extended contact surface between the lesion and the pleura resulted a protective factor from procedural complications (p < 0.001). All the remaining results are shown in Table 3. A significant improvement of the procedural time (23.2 ± 6.45 minutes in group A vs 20.1 ± 4.8 minutes in group B; p =0.006), diagnostic accuracy (83.3 % vs 98.3 %; p = 0.004), and sensitivity (82.5 % vs 98.2 %; p = 0.005) was observed in group B in respect to group A; specificity was 100% in both groups and the complication rate was comparable between groups A and B (p = 0.201). Descriptive analysis of diagnostic indicators, procedural time and complication trends is shown in Figure 5.
The present study reports a preliminary experience on CTLB performed with a non-coaxial technique and a 16 G needle, with a particular focus on procedural outcome and operator learning curve. Although CTLB procedural outcome has been extensively studied over years, operator learning curve on CTLB is still an unexplored ground. Su et al [9] previously investigated learning curve for lung biopsies performed on a cone-beam CT system in association with a virtual navigation software, but a comparable analysis of lung biopsies performed with conventional CT systems has never been investigated before.
Learning curve analysis showed procedural streamlining over time as suggested by the reduction of mean procedural time in group B compared to group A. This effect could be explained by the progressive experience-related optimization of the procedural steps, such as a more “fluid” placement of the needle within the lesion, and a fewer number of CT scans required for lesion targeting. A positive shift was evident from group A to group B in terms of sensitivity, specificity and diagnostic accuracy; this improvement was predominantly influenced by the marked reduction of false negative results in group B (1 vs 10 of group A). It is our opinion that these results are the consequence of at least two different factors: (1) an increased amount of adequate material due to an improved use of the needle; (2) a refined analysis of the diagnostic CT / PET-CT images before the procedure, with a more accurate selection of the target portion of the lesions; a more effective needle insertion within the lesion during the procedure; both factors could be traceable back to an increased operator experience over time.
The overall complication rate of study population was 49.57% (59/120 patients), with operative management occurred in 11/120 (9.16%) patients, all treated with chest tube placement for a non-limiting pneumothorax, and with a median hospitalization of 3 days. This data are in agreement with those reported by other authors with smaller needles sizes [10,11], suggesting that the use of non-coaxial technique and 16 G needles is basically safe; however, about 50% of patients had procedural complications (predominantly minor complications), without an improvement of complication rate over time.
Our experience confirms that a smaller diameter of the lesion [12] [13] and the presence of emphysema along the parenchymal needle path [22, 23] are predisposing factors for complications during CTLB. On the contrary, the presence of pleural contact on CT images resulted as a protecting factor from the onset of complications (p = 0.012) [24] with a negative correlation between complication risk and contact surface extent (p < 0.001); a possible explanation of this result could be the presence of inflammatory adhesions or tumor infiltration of the pleura that nailed the lung against the thoracic wall during the needle passage. Comparative studies with post-biopsy surgical findings could better clarify this aspect. This study has some undeniable limitations: the sample size is relatively small with single-arm design, however, the number of procedures included was adequate for a valuable learning curve analysis and we have reason to think that a longer observation period and a greater number of procedures would have flattened the distance between groups A and B in terms of procedural time and diagnostic yield, interfering with the learning curve analysis; furthermore, the inclusion of a second arm would not have added information on the operator learning curve. Secondly, a specific analysis of the quality of pathological specimens was not possible in our population due to the retrospective nature of the study.
In the era of personalized medicine, the goal of CTLB is to provide an adequate amount of material to achieve an accurate diagnosis and, in cases of malignancy, to allow the molecular characterization of the lesion [14]. However, the use of CTLB should take into account the following critical aspects: diagnostic accuracy is extremely variable, ranging from 71% to 97% [6, 15], and procedural complications are relatively common, up to 62% [16] , with pneumothorax being the forerunner [10, 16]. Biopsy technique and needle caliper are essential elements that may affect procedural outcome: the use of a coaxial (needle in needle) technique could improve the diagnostic accuracy if compared to non-coaxial approach [4] but arises the risk of pneumothorax, perilesional bleeding and/or fatal complications [3, 4, 10]. On the other hand, non-coaxial technique is usually performed with small needles reducing chances to obtain both diagnosis and lesion molecular profile [4, 17, 18]. According to our preliminary results, non-coaxial technique in association with 16 G needles seems a feasible approach for CTLB with satisfactory outcomes, comparable with those previously reported for CTLB [19-21]. Practice over time improves procedural length and operator diagnostic performance, maintaining an adequate complication rate.
Regarding our Journal: https://oajclinicalsurgery.com/
Know more about this article
https://oajclinicalsurgery.com/oajcs.ms.id.10036/
https://oajclinicalsurgery.com/pdf/OAJCS.MS.ID.10036.pdf
