A systematic review of ultrasonography-guided transcutaneous fine needle aspiration cytology in the diagnosis of laryngeal malignancy

Clinical studies of patients undergoing ultrasonography-guided transcutaneous FNAC for diagnosis of suspected laryngopharyngeal cancer were included. Studies of any experimental or observational design in humans were included. Animal studies and human studies not reporting adequacy or accuracy rate for FNAC, or where the abstract or full text were unavailable in the English language were excluded.

Two reviewers (ME and DY) independently performed the literature searches and screened abstracts. The following databases were searched: MEDLINE^®, PubMed^®, Embase^®, Web of Science™, Cochrane Library and ClinicalTrials.gov (via Cochrane). The search terms used are listed in Figure 1. No limit was placed on language or year of publication.

Two reviewers (DY and ME) independently screened the title and abstract of all records identified from the database searches. Studies describing ultrasonography-guided transcutaneous FNAC in patients with suspected laryngopharyngeal cancer were assessed against the inclusion and exclusion criteria, with any disagreement resolved by discussion with a third reviewer (TS). Studies without an accessible abstract or full text after the title/abstract screening were followed up by attempting to contact the study authors. If they remained unavailable, the study was excluded. Studies using other methods of imaging guidance or tissue sampling were also excluded, as were any not reporting sample adequacy. Studies in which transcutaneous FNAC outcomes were grouped with those of other methods were followed up by contacting the study authors. If data remained undistinguishable, the study was excluded. Studies presenting overlapping or duplicate populations were limited to the largest study sharing data where it was not possible to disambiguate them.

Potentially relevant studies identified from the initial searches and abstract screening then underwent full-text screening by the two independent reviewers prior to data extraction. Conflicts on selection were resolved by discussion between the reviewers.

Data were extracted by one reviewer (DY) and subsequently checked by a second reviewer (ME). Extracted data were collected in an Excel^® spreadsheet (Microsoft, Redmond, WA, US).

Two reviewers independently assessed the risk of bias using the Brazzelli risk of bias checklist for non-randomised studies.⁴ Studies were also evaluated according to the Oxford Centre for Evidence Based Medicine (OCEBM) grading system.⁵ Discrepancies between the reviewers were resolved by discussion.

A total of 568 studies were identified from the literature search, of which 3 met the inclusion criteria.^6–8 The process of identifying and screening records is displayed in Figure 2.

The three studies included in this review involved a total of 215 patients. Two of the studies were cohort analyses from India: a prospective study of 24 patients and a retrospective study of 189 patients. Also included was a retrospective case series of two patients in the US. The studies were published between 2013 and 2020. The study characteristics including FNAC technique and OCEBM level of evidence⁵ are summarised in Table 1.

The results of each individual study are summarised in Table 2. Parasuraman et al demonstrated a diagnostic accuracy rate of FNAC gaining a cytological diagnosis of 87.5%.⁶ Of the 24 patients included, 21 samples showed squamous cell carcinoma (SCC), 2 samples demonstrated atypical cells and 1 sample contained inadequate material for analysis. The three latter patients proceeded to have direct laryngoscopy and biopsy, and were then confirmed to also have SCC. FNAC was reported to be tolerated well by all patients with no complications.

Mukundapai et al investigated the largest cohort, with 189 patients; 146 patients provided satisfactory aspirate material while the remaining 43 patients were unable to provide adequate samples for cytology.⁷ Of the satisfactory samples, 132 were labelled neoplastic, 6 samples were benign and 8 cases were given descriptive labels such as necrosis or mucinous material. Among the samples identified as neoplastic, eight patients were labelled as having poorly differentiated carcinoma, possibly high-grade SCC. The descriptive reports and poorly differentiated samples tended to be investigated further to confirm diagnosis. However, only four cases were biopsied and confirmed as SCC, three patients died prior to biopsy and the remaining patients were lost to follow-up. Mukundapai et al were able to ascertain a cytomorphological diagnosis in 138 patients, giving a diagnostic accuracy rate of 73.0%. They reported a small number of cases of mild haemoptysis but otherwise no complications from the procedure.

The case series reported by Lopchinsky et al included two patients who underwent ultrasonography-guided FNAC and who were both successfully diagnosed with SCC, giving a 100% adequacy rate for the case series.⁸ No complications were mentioned by the authors.

Pooling of the results of the 215 patients from the included studies shows that 171 patients produced adequate samples for cytological analysis. Of these, 161 patient samples were able to lead to a diagnosis. This gives an overall adequacy rate of 74.9%.

All included studies were graded at level 3 or 4 of the OCEBM levels of evidence.⁵ The risk of bias assessment is presented in Table 3. Certain studies had a high risk of bias due to retrospective data collection and lack of sufficient follow-up.

The studies included in our review all report positive results for the use of FNAC as a method of reaching a histological diagnosis for laryngeal malignancy, citing the safety and speed of the procedure. This suggests that FNAC could offer an alternative to direct laryngoscopy and biopsy. Although two studies were retrospective and only one was prospective, it was deemed reasonable to pool the results across all three studies as the study population and intervention in each study appeared to be similar.

In all studies, FNAC was performed successfully in most patients, with minimal complications. Parasuraman et al found it to be successful in almost all patients, demonstrating FNAC to be a viable option for individuals who are not fit for direct laryngoscopy under general anaesthesia.⁶

Mukundapai et al had a greater rate of sample inadequacy, which was attributed to several factors, such as size and site of lesion, and uncooperative patients.⁷ The reason for this difference in sample adequacy may be due to the individual study designs and patient selection criteria. Mukundapai et al used a retrospective design looking at all patients who were investigated by FNAC in a larger tertiary centre. These patients were not chosen based on selection criteria and were instead investigated by FNAC as a primary investigation of their tumour. Lopchinsky et al reported on a case series demonstrating successful use of FNAC in patients who were not candidates for operative laryngoscopy and biopsy.⁸

In the context of other evidence, FNAC also carries proven use in the diagnosis of other head and neck conditions such as thyroid pathology. One retrospective study from 2023 conducted by Osseis et al looked at the diagnostic accuracy of FNAC in thyroid tumours, using histopathology as a reference standard.⁹ It was found that FNAC had an overall accuracy rate of 75.89%, demonstrating moderate diagnostic accuracy.

The limitations of our systematic review should be considered when interpreting its results. The study by Mukundapai et al is limited by its retrospective design⁷ and that by Lopchinsky et al is limited as it is a case series.⁸ Mukundapai et al included all patients investigated with FNAC and did not include any specific selection criteria to isolate certain target populations. Conversely, as the study by Parasuraman et al had a prospective design, this meant specific selection criteria were set for patients who were likely to produce better results (i.e. mainly patients with advanced laryngeal and hypopharyngeal mass lesions [T3 and T4 lesions]).⁶ They also excluded patients with calcified thyroid cartilage, vascular lesions and lesions not detected by ultrasonography. This could have skewed their results in favour of higher diagnostic accuracy.

Neither Parasuraman et al nor Lopchinsky et al provided information on patient follow-up or long-term outcomes,^6,7 Moreover, FNAC was not compared directly with laryngoscopy and biopsy in any of the three studies in our review.^6–8

Despite these limitations, all three studies included representative samples of the target population, gave sufficient information on how patients were selected for inclusion and had clearly defined outcome measures (adequacy and diagnostic accuracy of FNAC).^6–8 Patients were selected consecutively within similar age ranges and all FNAC was performed by adequately trained staff in appropriate environments.

Across all three studies, if only including the patients for whom adequate samples for analysis were obtained, the average diagnostic accuracy for FNAC increases from 74.9% to 94.2%, meaning that when an adequate sample is produced, FNAC offers a high diagnostic yield. Parasuraman et al suggest immediately assessing samples for adequacy prior to analysis as a method to ensure adequate samples.⁶ In fact, guidelines from the National Institute for Health and Care Excellence recommend the presence of a cytopathologist at the time of aspiration for this specific purpose, and many hospitals have implemented ‘one stop’ head and neck clinics to facilitate this in other head and neck malignancies.¹⁰ Consequently, a ‘one stop’ clinic in which diagnostic FNAC is used for laryngeal malignancy could be an excellent way to achieve quick and safe diagnoses while avoiding general anaesthesia.