FLEX February 2024

398

R. Locke et al. / International Journal of Pediatric Otorhinolaryngology 78 (2014) 393–401

Table 4 The results of fine needle aspiration cytology in various published paediatric studies.

Sites

Unsatisfactory

Benign

Malignant

False Pos

False Neg

Sensitivity (%)

Specificity (%)

Study

N

Liu [50]

50 29 92

H&N H&N

9 1 6 5 2 7 9 1 0 5 1 0

38 26 53 88 38 86 68 19 93 70 71

3 3

0 2 2 1 1 0 4 2 0 0 2 0

1 1 1 0 1 1 3 1 0 4 1 1

67 67 97 92 95 92 95 89 98 93

100

Ramadan [49] Cohen [53] Ponder [48] Eisenhut [54] Taylor [51] *Kardos [44] *Mobley [46] Schaller [52] Silverman [56] *Wakely [45] Rapkiewicz [55]

92 96 99 99

All LN All All LN All All All

33 13 13 19 37 20 13 36 41 14

106 288

100

273

64

100

153

95 97

89 32

H&N

100

100 100

135 112

97

85

H&N

100

The studies marked with an asterisk almost certainly report on the same cohort of children, but the degree of overlap is difficult to establish. (H&N = all masses in the head and neck region, LN = lymphadenopathy in neck, groin and axillae, All = any mass at any site).

distinguish between tuberculosis nodes and lymphomatous nodes but not between lymphomatous and metastatic nodes. The study however only makes brief comment on the large number of ‘indeterminate’ biopsies and is unclear as to whether open biopsy or multiple FNAC attempts were used [38]. Nemati et al. used grey-scale ultrasonography, colour Doppler and power Doppler ultrasonography to examine 50 children with cervical lymphadenopathy. They do not comment on how the patients were selected and the standard to which ultrasound is compared to in each case is unclear. Overall there remains a high range in sensitivity and specificity, suggesting these tests should not be relied upon on their own for diagnosis [39]. In a review of imaging modalities in children with cervical lymphadenopathy a significant overlap in radiological findings is described. Ultrasound is confirmed as the first imaging modality of choice in the majority of cases with use of computed tomography, magnetic resonance imaging and positron emission tomography in selected cases [40]. On the question of inter-reporter variability a recent study has shown that the introduction of a reporting protocol has improved standardisation of paediatric cervical ultrasound reports [41]. However the benefit in patient management has yet to be shown. Simple haematological tests are often requested in children with cervical lymphadenopathy. However, as discussed above the diagnostic utility of any investigation can only be determined if the whole series of children presenting with lymphadenopathy is reported. Haemoglobin values of less than 10 g/dl were found in three children in one series, two of whom had lymphoma [22]. The finding of malignant cells on peripheral blood smears was very unusual in any series. White cell count was not found to be helpful in one series [19], but counts of greater than 15,000 were found in 6 children in a further series, 3 of whom had lymphoma [22]. The three children with platelet counts below 100,000 in this same series had serious pathology. We found only one study in which a cohort of children with lymphadenopathy who underwent routine serological testing (for toxoplasma, cytomegalovirus, Epstein–Barr virus (EBV), and Brucella) [20]. Bartonella (the agent causing cat-scratch disease) was not tested for. Of 382 children tested, EBV was found in 5%, cytomegalovirus in 3%, toxoplasmosis in 1% and Brucella in none. Abdel-aziz et al. identified 15% positive serology for EBV in 160 children presenting with lymphadenopathy in Egypt however this paper would appear to predominantly include patients with acute lymphadenitis and fever [42]. Lactate dehydrogenase level has been shown to be elevated in children with malignancy, as a marker of cell turnover, and can also be useful in monitoring response to treatment [8]. 3.6. Blood tests

3.7. Tissue sampling

Bain et al. from Cambridge reported a series of 15 children with neck lumps (12 of which were lymph nodes) who underwent cutting needle biopsy under ultrasound guidance [43]. The children ranged in age from 3 months to 16 years with only 3 under 6 years. Sixteen or 18 gauge needles were used. All had EMLA cream applied 1 h before the procedure, followed by local anaesthetic infiltration. Sedation was used in 2 cases but a degree of parental restraint was the norm. The histological diagnosis was correct in every case. This technique supplies a useful amount of tissue for histopathology, however cutting needle biopsy requires consideration of the risks of haemorrhage and nerve damage (although neither was encountered in this small series) and of acceptability, particularly with regard to the adequacy of local anaesthesia without sedation in young children. While clearly successful in this series, it remains to be seen whether this technique will find more widespread acceptance. FNAC has been the subject of many studies and some controversy. The relevant studies are presented in Table 4. However, we were unable to identify any studies specific to FNAC of cervical lymph nodes in children. The paediatric series we found are all retrospective. Three of the papers seem to report on the same cohort of children [44–46]. It should be noted that some figures in the table differ from those quoted in the papers as some re-calculations were required to make sure that data presented in the table are comparable. Values are calculated according to the number of aspirations, not the number of children (where there were children who underwent the procedure more than once) because data presented per child was not always available. Aspirates judged suspicious but not diagnostic of malignancy were included with the definite malignancies for the purpose of calculations, but inadequate samples were excluded. In one study the usefulness of FNAC and the incidence of tuberculosis lymphadenopathy is discussed however the inadequate aspirate rate is not mentioned [47]. Some of the studies report that FNAC was performed without any anaesthetic or sedation [46,50,51], others used local anaes thesia alone [51] or in combination with sedation [52,53] and some used sedation in selected cases only [45,55]. No study reported any complications from FNAC. It can be seen from the table that FNAC has a high specificity, between 92% and 100%. However, the real clinical issue is how sensitive is FNAC. In other words, can serious pathology be reliably excluded with an FNAC result. Sensitivity varies in the series identified varied between 67% and 100%. The heterogeneity of these studies makes it difficult to draw firm conclusions, except to say that it is far from clear that FNAC can be relied upon to exclude serious pathology.