Medical thoracoscopy (MT), also called local anesthetic thoracoscopy or pleuroscopy, is a minimally invasive technique used to examine the pleural space and obtain targeted biopsies of the parietal pleura1–3. It is considered the technique of choice in the diagnosis of pleural effusion (PE) with criteria of exudate of uncertain etiology4,5. In a review published in 2010 that included several case series, MT with pleural biopsy using conventional forceps had a sensitivity of 92.6% for the diagnosis of malignant PE3. MT can be performed with a rigid thoracoscope or a semi-rigid pleuroscope1–3. The results of the major studies comparing diagnostic yield between rigid and semi-rigid MT showed no significant differences between the two techniques6–8. However, when thoracoscopy is performed with a semi-rigid pleuroscope, 6-8 biopsies should be obtained due to its smaller size1–3. The use of this technique is now expanding and it is considered safe, with a rate of serious complications of less than 5%1,3,9,10. However, the reported diagnostic efficacy and safety have not been uniform over all studies, especially in the case of malignant pleural involvement3,9,10. This factor, in addition to a growing need for larger histological samples for the molecular characterization of tumors11, has promoted the development of new pleural biopsy techniques such as cryobiopsy (CB). CB has been described as a safe endoscopic technique with a high diagnostic yield, initially in the diagnosis of endobronchial tumors and more recently in the study of interstitial lung disease12,13. Its validity and safety in the study of PE have been confirmed by several groups, although these studies have reduced case numbers and different designs14, which limits, at least individually, the value of their scientific evidence. Indeed, these new pleural biopsy techniques are not included in the clinical practice guidelines4,5.

In this setting, then, it seems an appropriate time to review the available evidence on this topic by performing a systematic pooled analysis of the published literature on the diagnostic yield and safety of parietal pleural CB in the diagnosis of PE versus conventional flexible biopsy forceps (FBF), and to evaluate some aspects such as the implementation of this novel diagnostic technique in routine clinical practice. This study was carried out with these objectives.

A systematic review and subsequent meta-analysis of existing evidence was performed from January 2010 to May 2019 following the recommendations of the PRISMA statement for reviews and meta-analyses. A search of the literature was conducted in the electronic databases PubMed (MEDLINE), Embase, The Cochrane Library, Web of Science, and TheClinicalTrials.gov for pleural CB during MT. The search terms were: «semi-rigid pleuroscopy», «rigid thoracoscopy», «medical thoracoscopy», «cryobiopsy», «pleural effusion», «cryoprobe biopsy». A manual search was also performed to capture studies or reviews that were not electronically selected. We included meta-analyses and systematic reviews, randomized, cohort, case-control, and case-series studies that described the yield of pleural CB in the study of PE during TM. Studies with fewer than 10 patients and isolated clinical cases were excluded. The search was not limited to any language.

The systematic review was complemented by analyzing the list of references of the studies included and the main clinical practice guidelines.

Two authors (MBR and CMR) independently selected abstracts to determine their eligibility and excluded studies that did not meet the inclusion criteria. Discrepancies were resolved through discussion with a third author (AFV). Eligible studies were retrieved in complete format and reviewed independently by 2 authors (MBR and CMR).

Relevant information from each of the studies was collected from evidence tables.

The main variables analyzed were the diagnostic yield of parietal pleura samples obtained by cryoprobe versus those obtained with FBF. Yield was calculated by dividing the number of subjects in whom biopsies showed definitive histological findings (including nonspecific pleuritis) for the final diagnosis of the patient. Complications, such as degree of bleeding from CB or FBF and procedure-related pain were also recorded as a primary variable. Bleeding was considered mild when it was self-limiting, moderate when the use of topical vasoactive drugs (adrenaline) was required, and severe when intervention with electrocautery or argon plasma was necessary1,3.

Secondary variables were the mean size of the CB (in mm), the depth of the sample, the artifact level, defined as crush affecting more than 25% of the sample area or interpretability, and degree of integrity14.

With respect to the procedure, the cryoprobe freeze time (in seconds) and duration were analyzed.

CB molecular diagnosis yield in malignant PE due to lung cancer was also analyzed as a secondary variable.

The clinical and epidemiological characteristics of the patients (age, sex) and the specific etiology of benign and malignant PE were also described in general terms.

The quality of the studies was assessed according to the levels of evidence and grades of recommendation described in the scientific guidelines15 and 27 items from the PRISMA statement for systematic reviews and meta-analysis were verified for the drafting of this manuscript16. The results of the QUADAS-2 evaluation of all the studies included in this systematic review and meta-analysis were also summarized.

The proportions and 95% confidence intervals (95% CI) of diagnostic yield and complications, particularly bleeding caused by the cryoprobe and FBF, were estimated in each study and jointly. This information was displayed using forest plots drawn up with the R 3.4.2. "meta" library statistical software package. The heterogeneity between the results of the studies was evaluated using the Cochran Q statistical test and the I2 statistic. Significant heterogeneity was considered if I2 was >50% and/or p < 0.1.

This systematic review and meta-analysis evaluates the efficacy and safety of pleural CB versus FBF biopsy techniques performed during MT in the study of PE. Although variability is high, the level of quality is acceptable. CB is shown to be a safe technique, with an overall yield of 95% for establishing the etiology of PE. Pleural samples obtained with cryoprobe are larger, deeper, better quality, and more interpretable, and have fewer artifacts than those obtained with FBF. However, the data are insufficient to draw firm conclusions about a potential advantage in the molecular characterization of samples obtained in lung cancer patients with PE.

In one of the studies, special difficulties were described for the performance of CB in patients with a very dense pleural surface14,24, so it is highly likely that the macroscopic appearance and consistency of metastatic implants is important in the choice of this biopsy technique compared to conventional methods.

A limitation of semi-rigid MT is the small sample size, so many experts recommend 6 to 10 biopsies1–3. As described in this pooled analysis, the samples obtained by CB from the parietal pleura are significantly larger than those obtained with FBF14,19,20,22–24. This may be of particular importance in the diagnosis of malignant PE, as a large sample is needed to perform not only morphological, immunohistochemical and molecular studies, but also to differentiate metastatic PE from different organs or hematological effusion from pleural mesothelioma. In a very recent study, not included in this review since it only included 5 cases with pleural mesothelioma, the size of the samples obtained with cryoprobe was 3 times greater than with FBF. The depth needed to establish the diagnosis of mesothelioma was achieved in 4 cases using CB compared to 1 using FBF, and the percentage of artifacts was significantly lower for CB compared to FBF (9% vs 35%, p < 0.001)18. With CB, the diagnosis of mesothelioma could be established successfully in all cases, while with FBF, diagnosis was only achieved in 1 case18. Despite these results, further studies seem necessary to determine the optimal biopsy technique for the diagnosis of pleural mesothelioma.

With the development of new targeted therapies, including immunotherapy for the treatment of non-small cell lung cancer, the demand for better quality tumor samples has increased11. In a small descriptive study conducted by our group on the diagnostic yield of semi-rigid MT for the molecular characterization of malignant PE in non-small cell lung cancer, molecular testing was performed on 14 adenocarcinomas. Sufficient useful material was obtained to determine the EGFR mutation status in 100% and ALK translocation in 90%25. Only 2 studies in this meta-analysis analyzed cryoprobe samples for the detection of EGFR gene mutation, but this was achieved in all tumor samples analyzed. Given the growing importance and development of targeted therapies, whenever new molecular alterations or conduction mutations are identified, new studies are needed to establish the indication of parietal pleural cryoprobe as a new biopsy technique for optimizing the sample and identifying the greatest number of molecular alterations possible26,27.

The methodology of this review has been rigorous, with independent double screening of abstracts, review of articles, data extraction, and evaluation of the quality of studies. The rigorous and evidence-based process used in the construction of the QUADAS-2 tool allowed us to assess and qualify the risk of bias and concerns regarding applicability. As such, the authors consider that this review provides a reliable summary of the existing bibliography. However, it has limitations. The studies included were mainly observational and there is a remarkable rate of heterogeneity in the main objectives studied.

A systematic review describing similar results has recently been published28. However, one of its limitations is that the only randomized controlled trial published so far was not included among the selected studies24.

The results of this meta-analysis show that CB of the parietal pleura during MT is a safe procedure with a high diagnostic yield for PE. Obtaining larger, better quality samples with fewer artifacts would shorten the procedure time, because fewer biopsies are needed than with FBF, and obtaining larger tissue specimens may help reduce the number of tests and avoid repeat procedures. If we take into account that tissue is now the standard material for establishing the diagnosis of non-small cell lung cancer and, given the continuous search for new molecular markers, the most significant limiting factor is the amount of tumor available, pleural CB may be considered the best option for establishing morphological, immunohistochemical, and molecular diagnosis in this group of patients, although a greater body of scientific evidence that is more homogeneous and higher quality is required.

Another limitation of the technique is that it is not standardized. Probe type and freeze time vary between hospitals, and all studies in this meta-analysis were performed in single centers. Multicenter studies with a rigorous methodology are needed to evaluate the routine use of this new technique as a tool during MT and its application in routine clinical practice.

The authors state that they have no conflict of interests.