Clinical updates of approaches for biopsy of pulmonary lesions based on systematic review

From the eligible articles, investigators CJD and FQD reviewed the following information, i.e., PMID, year of publication, study design, number of patients, average age of patients, nodules size and location, operation time, biomarkers for detection, diagnostic sensitivity, relative complication morbidity, treatment of complications, outcome and follow-up period.

In last 20 years, the overall sensitivity, specificity, and accuracy of CT-guided PTNB were 92.52 ± 3.14%, 97.98 ± 3.28%, and 92.28% ± 5.40%, respectively. The top two complications of CT-guided PTNB were pneumothorax (1111/4822:23.04%) and hemorrhage (287/3503:8.19%), respectively. Two cases with severe complications were reported [2, 3]. Bronchial artery embolization was performed in one patient due to massive hemoptysis [3]. The other one suffered from cardiopulmonary arrest leading to death [2].

Diagnostic accuracy and incidence of complications seemed to be decreased [3‐5] and increased [2‐9], respectively, by smaller lesion size or longer needle path length (P < 0.05).

The detection rate of lung cancer by ENB increased gradually (Fig. 1a) and was recently reported to be 96.8% [10]. There seemed to be no significant correlation between detection rate and number of cases, average age of patients, sex, nodule size, lobar location of nodule, mean distance from pleura to nodule and operation time. As shown in Fig. 1b, pneumothorax was the top one complication following ENB (86/1648:5.2%). In 86 pneumothorax cases, 34 cases (34/86) were administrated with closed drainage [10‐21], and one case (1/86) was managed with manual aspiration and observation [19]. The other 51 cases with mild pneumothorax were discharged for rehabilitation. Intriguingly, the incidence of pneumothorax was significantly negatively correlated with nodule size (R = − 0.512, P = 0.018, Fig. 1c). The three hemorrhage cases were observed carefully without further intervention and were discharged for rehabilitation [16, 22]. Three cases of respiratory failure were reported without detailed depiction [16]. There were no ENB related death [10‐30]. ENB plus EBUS imaging seemed to yield a higher detection rate as compared with sole use of ENB (59% vs. 88% [20] and 71.42% vs. 73.07% [11]). Surprisingly, studies combining fluoroscopy with ENB to confirm navigation success reported lower diagnostic yields (56.3 vs. 69.2% without fluoroscopy, p = 0.006) [31].

The detection rate of lung node by EBUS-TBNA remained to be 86.06 ± 9.70%. The diagnostic sensitivity, specificity, accuracy, positive predictive value and negative predictive value of EBUS-TBNA for the mediastinal staging of lung cancer were 85.48% ± 12.89%, 99.09% ± 3.15%, 92.88% ± 4.99%, 98.70% ± 3.03%, 83.03% ± 15.46%, respectively. As shown in Fig. 2a, the top three complications following EBUS-TBNA were hemorrhage (53/8662:0.61%), pneumothorax (46/12432:0.37%) and infection (34/11250:0.30%), respectively. Four hemorrhage cases were administrated with further intervention with one perioperative death. The other 49 cases with mild hemorrhage were discharged for rehabilitation [32, 33]. In 46 pneumothorax cases, nine cases (9/46) and 37 cases (37/46) were administrated with closed drainage and conservative treatment, respectively [32‐35]. Perioperative mortality was relatively low (4/11189:0.04%). Besides the above mentioned one case died of severe hemorrhage, there was one case died of cerebral infarction and two unexplained deaths [32, 33, 36].

The detection rate of lung cancer by mediastinoscopy increased slightly (Fig. 2b) which was reported to be 96% in recent years [37]. The diagnostic sensitivity, specificity, accuracy, positive predictive value and negative predictive value of mediastinoscopy for the mediastinal staging of lung cancer were 82.83% ± 10.63%, 100%, 93.98% ± 4.68%, 100%, 87.64% ± 13.00%, respectively. Intriguingly, the positive rates of 4^thR (91.5% ± 9.35%) and 7th (80.56% ± 19.47%) lymph node were significantly higher than others (P = 0.03) (Fig. 2c). As shown in Fig. 2d, hoarseness (67/4387:1.53%) was the top one complication following mediastinoscopy. Among the abovementioned 67 cases with hoarseness, nine cases (9/67) suffered from permanent hoarseness, two cases (2/67) recovered partially by vocal cord medialization and six cases (6/67) recovered within a few months [37‐45]. Perioperative mortality was relatively low (4/2137: 0.19%). The death causes among three cases were aortic laceration, stroke, and cardiac arrest, respectively, and one case die of unexplained cause [46].

The mean sensitivities of a variety of methods to detect CTC remained to be 63.05%. As shown in Fig. 3a, sensitivity of PCR seemed to be highest (78.81 ± 14.72%). Sensitivity of Density-gradient, ISET and Magnetic bead seemed to be higher than 60% (71.32% ± 2.8%, 67.75% ± 21.22% and 67.85% ± 25.24%, respectively). Specificity of ISET, PCR and Cell search was relatively high (100%, 90.88 ± 0.53% and 94.33% ± 9.82%, respectively). There was no published data regarding specificity of Magnetic bead and density-gradient.

Intriguingly, there are a variety of biomarker combination for CTCs identification by using PCR yielding different sensitivities. As shown in Fig. 3b, the sensitivity of Multimarker assay seemed to be highest (90%). Besides, the sensitivity of the combination of TSA-9, KRT-19, Pre-proGRP was satisfactory (84.3%).