FLEX February 2024

397

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

Table 2 The association between lymph node size and the proportion with malignant histology.

to be reactive in nature whereas 20% of unilateral nodes were discovered to have histology that required treatment [33]. Some studies have suggested that the prevalence of malignancy rises as the number of involved lymph node groups increases. The prevalence for children with only one lymph node group involved was 1.4%, rising to 20% for those with 4 or more groups involved [20]. Another study showed 25% of children with 1 or 2 lymph node groups involved had a malignancy compared with 71% of those with 3 or more groups involved [19]. Supraclavicular nodes have a particularly high prevalence of malignancy compared to other sites in the neck, as shown in Table 2. Bamji et al. [11] found that 35% of supraclavicular nodes were lymphomas, which emphasises the importance of considering such a mass as lymphoma. It is suggested that supraclavicular nodes of any size should be examined with a high index of suspicion [19,22]. In 3 studies, the proportion of children who had a serious cause for their lymphadenopathy and an abnormal chest X-ray (78% [15], 83% [22], 69% [19]) was found to be significantly higher than the proportion of those with serious diagnosis and a normal chest X ray (18% [15], 25% [22], 15% [19]). What cannot be ascertained is the diagnostic utility of chest X-ray in the presence of lymphade nopathy as only the children with a serious cause for the lymphadenopathy were compared. In the algorithm described by Umapathy for the management of cervical lymphadenopathy, chest X-ray is suggested only if the patient has had contact with tuberculosis [35]. Ultrasound scanning is widely available, non-invasive and quick to obtain. Gianfelice et al. reported that ultrasound is particularly useful in thyroid and congenital masses but is not useful in distinguishing reactive lymphadenopathy from malig nancy [36]. However, they made no attempt to describe nodal architecture. Many of the comments are based on extrapolation from studies of ultrasound in adults and this may be unsound due to the greater mass of active lymphoid tissue in children and the different range of pathology encountered. Papakonstantinou et al. specifically studied ultrasonographic features including nodal architecture, hilar shape and vascularity in cervical lymphadenop athy in children [37]. The authors report on 103 children, aged 2 months to 14 years, presenting with cervical lymphadenopathy over a 3-year period. The series includes acute infections as well as noninflamed neck masses. Of the 11 children with heterogeneous echo-texture nodes, 5 had infectious mononucleosis and 6 had lymphoma. Blurred nodal margins and formation of a nodal mass were found in lymphoma and infected cases. Round shape (ratio of long to short axis less than 2) was found in 9% of reactive nodes but 78% of lymphomas. A narrow or absent hilum was found in 6% of reactive nodes but all lymphomas. Pandey et al. in a prospective study of patients presenting with cervical lymphadenopathy and confirmed by clinical examination used ultrasonography with FNAC to distinguish tuberculosis, lymphomatous and metastatic nodes. Identification of strong internal echoes, hypoechoic centre, echogenic thin layer, matting, presence of hilus and sharp margins were statistically significant to 3.5. Radiological investigations

Size (cm)

Karadeniz et al. [20] (%)

Soldes et al. [19] (%)

< 1

23 18

18

1–2 2–3 2–4

9

–

38

22

–

> 3 > 4

–

58

52

–

to confirm more serious pathology [15,20,22,30,31]. This may be due to sampling error relating to the choice of lymph node excised or it may relate to limitations in histopathological processing and reporting. All but one of these studies are over 20 years old.

3.4. Can serious pathology be predicted on clinical grounds?

We identified 11 studies which attempt to establish the predictive value of various clinical features for malignancy and other serious pathology. Ten are retrospective studies of a series of excised lymph nodes [1,4,14,19,21–23,32–34] and only one is a prospective study of children presenting with lymphadenopathy [20]. In a child with a known history of previous malignancy, the development of an enlarged lymph node in the neck is clearly suspicious. Two out of eight such children (25%) had a malignant cause for the lymphadenopathy in one series [19]. Lymph node size was found to be associated with the likelihood of malignancy in 2 studies [18,19]. Table 2 compares node size and rate of malignancy. Srouji et al. [33] reported that 75% of those with lymph nodes larger than 3 cm in diameter had histology requiring treatment. All those nodes with fluctuating size were found to be reactive. The reported duration of lymphadenopathy in the majority of studies was not found to be associated with the presence of serious pathology [15,19,20,22,33,34]. In 2 studies where adequate data was presented the prevalence of malignancy was lower in children where the lymphadenopathy had been present for a longer period of time [19,22]. The consistency of the enlarged lymph nodes on palpation was reported as being unhelpful for diagnosis in 2 studies although neither gave detailed figures [16,22]. In another study, 71% of lymph nodes which were judged to be fixed had a malignant diagnosis [19]. One study considered that tenderness is a specific feature warranting concern [35], however this was contradicted by another that showed tender nodes were reactive [33]. A malignant diagnosis or tuberculosis was found in 57% of children with persistent fever or weight loss in 1 study [15], but fever was not found to be predictive in another [19]. Malignancy or other serious pathology was found in 80% of children with hepatosplenomegaly in one series [22], and in 40% with splenomegaly and 50% with hepatomegaly in another [15]. Splenomegaly was not associated with serious diagnosis in a subsequent series [19]. Lymph node location is another potential predictor of serious pathology (Table 3). In a recent study all bilateral nodes were found

Table 3 The proportion of children with a malignant or other serious diagnosis, according to lymph node group involved.

Karadeniz et al. [20] (%)

Soldes et al. [19] (%)

Knight et al. [15] (%)

Lake and Oski [22] (%)

Site

89

75

61

80

Supraclavicular Posterior triangle Parotid/preauricular Deep cervical chain

8 0

0

32 14

12

Submandibular