Prophylactic thyroidectomy in children with multiple endocrine neoplasia type 2

Abstract

Background

In patients with multiple endocrine neoplasia type 2 (MEN2) syndrome, genetic testing offers early diagnosis, stratifies the risk of developing medullary thyroid cancer (MTC) and informs the timing of thyroidectomy. The efficacy of treatment, which depends on timely and safe surgery, is not well established.

Methods

This was a retrospective review of diagnostic and clinicopathological outcomes of prophylactic thyroidectomy in children with MEN2 between 1995 and 2013 in the UK. American Thyroid Association (ATA) 2009 guidelines were used as a benchmark for adequate treatment.

Results

Seventy-nine children from 16 centres underwent total thyroidectomy. Thirty-eight patients (48 per cent) underwent genetic testing and 36 (46 per cent) had an operation performed above the age recommended by the ATA 2009 guidelines; pathology showed MTC in 30 patients (38 per cent). Late surgery, above-normal preoperative calcitonin level and MTC on pathology correlated with late genetic testing. Twenty-five children had lymphadenectomy; these patients had more parathyroid glands excised (mean difference 0·61, 95 per cent c.i. 0·24 to 0·98; P = 0·001), and were more likely to have hypocalcaemia requiring medication (relative risk (RR) 3·12, 95 per cent c.i. 1·54 to 6·32; P = 0·002) and permanent hypoparathyroidism (RR 3·24, 1·29 to 8·11; P = 0·010) compared with those who underwent total thyroidectomy alone. Age did not influence the development of complications.

Conclusion

Late genetic testing may preclude age-appropriate surgery, increasing the risk of operating when MTC has already developed. Early genetic testing and age-appropriate surgery may help avoid unnecessary lymphadenectomy and improve outcomes.

Introduction

Precision medicine is the watchword for preventative and personalized treatments based on phenotypic, biomarker and genetic characteristics¹, whereas surgical precision has become an idiom for doing something very well². Both play an important part in managing patients with multiple endocrine neoplasia type 2 (MEN2), an autosomal dominant hereditary cancer syndrome associated with medullary thyroid cancer (MTC). MTC remains the principal MEN2-related cause of death³. The estimated prevalence of MEN2 (1 in 30 000) implies that there may be about 2000 patients at risk of developing MTC in the UK⁴.

Early identification of affected children followed by prophylactic thyroidectomy is currently the only effective strategy to prevent the development and spread of MTC. In the past, clinical observation with regular measurement of basal and stimulated calcitonin levels was the only available strategy for early detection⁴. The discovery of the causative role for RET mutations in MEN2 allowed the early distinction between affected and non-affected children^5,6. Correlation between specific RET mutations and the onset and course of MTC enables risk stratification based on genetic signatures, and led to recommendations for prophylactic thyroidectomy at the earliest stages of C-cell disease⁷. The American Thyroid Association (ATA) guidelines⁸ of 2009 marked a significant shift from a pure DNA-based model of planning thyroidectomy to an integrated genetic, clinical, radiological and biochemical concept^9,10. The modern strategy for prophylactic management of MTC in MEN2 syndrome, based on tailoring surgery to subcategories of patients with a distinct risk of disease, is complemented by precisely timed, safe and meticulously performed surgery. Thyroidectomy in young children is known to have a higher rate of complications than that in adults, and the consequences of these complications can profoundly affect well-being and quality of life⁹. The present study aimed to assess how precision medicine has been implemented in the management of MEN2 in the UK.

Methods

A retrospective nationwide medical record review was designed to investigate the timeliness, quality and efficacy of thyroidectomy in children with MEN2. All university hospitals in the UK managing patients with MEN2 were invited to identify children who had genetic testing and DNA-directed thyroidectomy for MEN2 between 1995 and 2013. Only children who were diagnosed and treated before they turned 16 years were included, as this age is a limit to being considered a child in the UK.

Patients diagnosed after 1995 were included, as discovery of a causative RET mutation in 1994 was a pivotal event in developing best clinical strategies¹⁰ and the UK health system was an early adopter of genetic testing. The diagnosis of MEN2 syndrome was based on detection of RET proto-oncogene mutations at multiple exons (10, 11, 13, 14, 15 and 16) or non-RET mutations that were proven to cause the syndrome¹¹ and confirmed in duplicate DNA samples. The closing year was chosen to allow postoperative follow-up of at least 1 year, as it included basal calcitonin sampling at long-term follow-up.

Permission for data collection from medical records of children was sought from the participating institutions. All information identifying the children was removed from the medical records before data collection. Because deidentified data were used, and it was not possible to trace any of the data to an individual, individual patient consent was not required. Only information required for answering the research question was extracted and coded. Each centre was visited by a research team, which entered anonymized data from hospital records into a research database. Information on diagnostic, clinical and pathological outcomes was recorded systematically according to a standard protocol to ensure accuracy of collected data.

The ATA 2009 guidelines⁸ were used as a benchmark for adequate management of children with MEN2, even though some of children had surgery before the guidelines were published, as the integrated genetic, clinical, radiological and biochemical management concept was prevalent long before 2009. Children with a clinical diagnosis of MEN2A, MEN2B and familial medullary thyroid cancer were stratified into four ATA 2009 risk groups based on the DNA profile.

Measures of outcome

Preoperative and postoperative basal calcitonin levels were determined by either immunoradiometric or chemiluminescence assays. Preoperative basal calcitonin levels were recorded as within or above the normal range, as each laboratory referred to a locally approved interval of normal values based on test sensitivity. Within abnormal results, a calcitonin level over 40 pg/ml was identified, as such a level suggests the presence of nodal metastases¹². Postoperative basal calcitonin levels were recorded as undetectable, detectable within or above normal limits, and exceeding 40 pg/ml.

The age of the children at the time of DNA testing and at the time of surgery, and type of operation (total thyroidectomy (TT), with or without lymph node dissection (LND)) were recorded and compared with recommendations for each of the risk classes defined by the ATA 2009 guidelines. Postoperative complications, including but not limited to laryngeal nerve and parathyroid function, were recorded and graded according to the Clavien–Dindo classicification¹³.

When LND was performed, it was noted whether central or lateral neck compartments were involved, as described in the operating notes and pathology report. Single lymph nodes occasionally found at pathology did not count as neck compartment dissection unless lymphadenectomy was stated in the operative note; however, such single nodes were included in the lymph node count and positivity analysis.

The number of parathyroid glands identified, preserved in situ, excised or autografted was also noted. Postoperative hypocalcaemia was defined as a serum calcium level below 2·2 mmol/l, without or with clinical symptoms. Calcium levels were recorded as sampled on the first day after surgery. The number of patients whose clinical status or blood samples indicated hypocalcaemia during the hospital stay was noted. Symptomatic hypocalcaemia was further classified as requiring treatment with intravenous or oral calcium and α-calcidol during the hospital admission, for less than 12 months (transient hypoparathyroidism), or for 12 months and over (permanent hypoparathyroidism).

Histopathological results were classified as normal thyroid, C-cell hyperplasia (CCH) or MTC. CCH was defined by evidence of more than 50 C-cells observed in at least three low-power fields at 100 × magnification or equivalent. MTC was diagnosed in the event of atypia of C-cells and their nuclei, in combination with spreading of C-cells through the basement membrane as a sign of infiltrative C-cell proliferation¹⁴. The risk of finding MTC at pathology was computed against timely genetic testing and timely operation, by ATA risk class and by preoperative calcitonin status. Tumour size larger than 10 mm, lymph node involvement, multifocality, lymphovascular invasion and extrathyroidal extension were recorded. If multiple foci of MTC were present, the largest tumour dimension was used for analysis. Lymph node status was defined as N0 (no lymph node metastases), N1 (lymph node metastases present) or Nx (no LND).

Statistical analysis

ANOVA was used to analyse differences in mean values, and one-way ANOVA for ranks (Kruskal–Wallis) test for differences in median values, between multiple groups. Univariable associations between categorical variables were evaluated by χ² test and logistic regression. Where appropriate, odds ratios or relative risks with 95 per cent confidence intervals were estimated for dichotomous outcomes, whereas weighted mean differences with 95 per cent confidence intervals were calculated for continuous outcomes.

Differences were explored in univariable analysis between characteristics of patients (age, sex), disease (risk-level genotype, preoperative basal calcitonin level, presence of MTC), type of surgery (with or without LND) and complications. Candidate variables for a multivariable model were chosen based on clinical relevance and significance in univariable analysis. Statistical significance was set at 0·050. Data were analysed using SPSS® version 22.0.0 software (IBM, Armonk, New York, USA).

Results

Between January 1995 and April 2013, 79 children (36 boys) diagnosed with MEN2 on the basis of genetic testing by the age of 16 years were identified from 16 UK centres participating in the study. The majority of children were found to have MEN2A and the most common RET mutation was in codon 634 (Table 1). Preoperative basal calcitonin levels were above normal in 49 patients (62 per cent) and exceeded 40 pg/ml in 16 (20 per cent). Most children with a basal calcitonin level over 40 pg/ml were in the high-risk groups based on ATA 2009 guidelines (Table 2).

Table 1

Germline mutations in relation to clinical syndromes and American Thyroid Association 2009 risk classes

		By clinical syndrome			By ATA 2009 risk group
Mutation	No. of patients	FMTC	MEN2A	MEN2B	A	B	C	D
RET codon
609	2	0	2	0	0	2	0	0
620	9	0	9	0	0	9	0	0
634	40	0	40	0	0	0	40	0
649	2	2	0	0	2	0	0	0
768	3	2	1*	0	3	0	0	0
790	2	2	0	0	2	0	0	0
791	1	1	0	0	1	0	0	0
891	5	5	0	0	5	0	0	0
918	4	0	0	4	0	0	0	4
Not known†	6	1	5	0	1	5	0	0
Non-RET FMTC‡	5	5	0	0	5	0	0	0
Total	79	18	57	4	19	16	40	4

*

Part of a verified multiple endocrine neoplasia (MEN) type 2A kindred (phaeochromocytoma in family history).

†

Patient underwent DNA test, but type of mutation could not be found on notes; syndrome type was verified on medical notes.

‡

Verified familial medullary thyroid cancer (FMTC) kindred with non-RET mutation. ATA, American Thyroid Association.

Table 2

Preoperative and operative management of children who had a genetic test and underwent surgery

		ATA 2009 risk group
	All patients (n = 79)	A (n = 19)	B (n = 16)	C (n = 40)	D (n = 4)	P§
Sex ratio (M : F)	36 : 43	10 : 9	11 : 5	14 : 26	1 : 3
Timing of DNA test
Age at DNA test (years)*	4·3 (0·3–15·2)	3·1 (0·3–15·0)	4·5 (0·3–14·0)	5·2 (0·3–15·2)	2·8 (0·3–9·0)	0·723¶
Age at genetic test as recommended by ATA 2009 guidelines	41 (52)	10 (53)	10 (63)	20 (50)	1 (25)	0·585
Late genetic test	38 (48)	9 (47)	6 (37)	20 (50)	3 (75)
Above-normal calcitonin	29 of 38 (76)	4 of 9 (44)	4 of 6 (67)	18 of 20 (90)	3 of 3 (100)	0·039
MTC on pathology	21 of 38 (55)	1 of 9 (11)	2 of 6 (33)	15 of 20 (75)	3 of 3 (100)	0·003
Biochemical presentation
Interval between genetic test and surgery (months)*	9·8 (1·1–159)	20·4 (3·4–159)	17·2 (5·7–54·7)	7·7 (1·1–69·7)	4·9 (3·9–77·4)	0·046¶
Above-normal preoperative calcitonin	49 (62)	10 (53)	7 (44)	28 (70)	4 (100)	0·089
Preoperative calcitonin > 40 pg/ml	16 (20)	1 (5)	2 (13)	10 (25)	3 (75)	0·011
Timing of surgery
Age at surgery (years)*	6·2 (0·8 –15·6)	7·6 (4·4–15·3)	5·7 (3·3–15·6)	6·2 (0·8–15·5)	3·1 (0·8–15·4)	0·195¶
Timing of surgery as recommended by ATA 2009 guidelines	43 (54)	19 (100)	7 (44)	16 (40)	1 (25)	< 0·001
Late surgery	36 (46)	–‡	9 (56)	24 (60)	3 (75)
Above-normal calcitonin	28 of 36 (78)	–	4 of 9 (44)	21 of 24 (88)	3 of 3 (100)	0·019
MTC on pathology	20 of 36 (56)	–	2 of 9 (22)	15 of 24 (63)	3 of 3 (100)	0·031
Operative management
TT alone	54 (68)	16 (84)	13 (81)	24 (60)	1 (25)	0·042
Age at TT alone (years)†	7·1 (0·8–15·6)	8·5 (4·2–15·3)	7·4 (3·7–8·5)	6·2 (1·7–15·5)	2·8 (0·8–15·4)	0·223#
TT alone as recommended by ATA 2009 guidelines	48 of 54 (89)	16 of 16 (100)	11 of 13 (85)	21 of 24 (88)	0 of 1 (0)	0·016
TT and above-normal preoperative calcitonin	29 of 54 (54)	8 of 16 (50)	6 of 13 (46)	14 of 24 (58·3)	1 of 1 (100)	0·693
TT + LND	25 (32)	3 (16)	3 (19)	16 (40)	3 (75)	0·042
Age at TT + LND (years)†	8·1 (0·8–15·5)	12·6 (11·8–13·5)	5·3 (3·7–8·5)	8·1 (1·7–15·5)	6·5 (0·8–15·4)	0·215#
TT + LND as recommended by ATA 2009 guidelines	11 of 25 (44)	1 of 3 (33)	0 of 3 (0)	7 of 16 (43·8)	3 of 3 (100)	0·097
TT + LND and above-normal preoperative calcitonin	20 of 25 (80)	2 of 3 (67)	1 of 3 (33)	14 of 16 (88)	3 of 3 (100)	0·126

Values in parentheses are percentages unless indicated otherwise; values are

*

median (range) and †mean (range).

‡

No age limit specified for this class in the American Thyroid Association (ATA) 2009 guidelines. MTC, medullary thyroid carcinoma; TT, total thyroidectomy; LND, lymph node dissection.

§

χ² test, except

¶

one-way ANOVA for ranks (Kruskal–Wallis) test and

#

ANOVA.

Timing of genetic testing

Median age at genetic testing was 4·3 years. Forty-one children (52 per cent) had a genetic test by the age recommended in the ATA 2009 guidelines. All children were asymptomatic for C-cell disease at the time of diagnosis. The majority of children who underwent genetic testing beyond the recommended age had above-normal preoperative calcitonin levels and had already developed MTC by the time of surgery; both observations were significantly more frequent in high-risk groups (Table 2).

Late surgery, above-normal preoperative calcitonin level and MTC on pathology correlated with late genetic testing (Table 3).

Table 3

Univariable correlation studies

	Spearman ρ	P
Age-appropriate DNA test
Detectable basal calcitonin	–0·208	0·066
Above-normal basal calcitonin	–0·283	0·011
Calcitonin > 40 ng/ml	–0·397	< 0·001
Appropriately timed surgery*	0·594	< 0·001
MTC	–0·343	0·002
MTC
Appropriately timed surgery*	–0·331	0·003
Above-normal basal calcitonin	0·344	0·002

MTC, medullary thyroid carcinoma.

*

According to American Thyroid Association 2009 guidelines.

In multivariable analysis, the odds of finding MTC at pathology were sevenfold higher in children with a preoperative calcitonin level above 40 pg/ml (Table 4). The odds of having a preoperative calcitonin exceeding 40 pg/ml were decreased by 17 per cent by having a timely DNA test, and by 84 per cent if a child was in ATA risk class A.

Table 4

Multivariable binary logistic regression analysis of risk factors for medullary thyroid carcinoma

Dependent variable	Explanatory variable	Odds ratio	P
MTC found at pathology	Calcitonin > 40 pg/ml	7·23 (1·13, 46·38)	0·037
Calcitonin > 40 pg/ml	Appropriately timed DNA test*	0·83 (0·02, 0·44)	0·003
	ATA risk class A	0·16 (0·01, 0·46)	0·016
Appropriately timed surgery*	Appropriately timed DNA test*	12·70 (3·84, 41·97)	< 0·001

Values in parentheses are 95 per cent confidence intervals.

*

According to American Thyroid Association 2009 guidelines. MTC, medullary thyroid carcinoma.

Timing of surgery

The median age of children undergoing surgery was 6·2 years. The median interval between genetic testing and surgery was 9·8 months, and was significantly longer in lower-risk groups. Forty-three children (54 per cent) had surgery at the appropriate age according to the ATA 2009 guidelines, with the proportion significantly higher in low-risk groups. The majority of children who had late surgery presented with above-normal preoperative calcitonin levels and MTC on histopathology (Table 2).

In, multivariable analysis, surgery at the recommended age was significantly dependent on a genetic test performed by the recommended time (Table 4).

Type of surgery

Paediatric, ear, nose and throat or general surgeons with expertise in endocrine surgery performed all operations. Fifty-four children (68 per cent) underwent TT alone. TT was performed more frequently in lower- than in higher-risk groups. More than three-quarters of patients had TT alone in keeping with ATA 2009 guidelines, and children in lower-risk groups were more likely to have TT as recommended. Twenty-nine children (54 per cent) who had TT also had above-normal preoperative calcitonin levels.

Twenty-five children (32 per cent) had TT and central compartment (level 6) LND; three also had lateral neck compartment dissection (level 3,4). LND was performed more frequently in higher- than lower-risk groups, and predominantly in patients with raised preoperative calcitonin levels. The decision to perform LND was in keeping with ATA 2009 recommendations in 11 of 25 children. Recurrent laryngeal nerves were seen and preserved in all patients.

Postoperative complications

No patient developed permanent significant dysphonia, postoperative bleeding or needed reoperation. Five children had complications not related to postoperative hypocalcaemia (Table 5). Only procedures including LND were associated with complications of Clavien–Dindo grade II or above, including a patient who needed ICU care and intubation. Age at surgery was not associated with postoperative complications.

Table 5

Complications, management of parathyroid glands and duration of hospital stay by extent of surgery

					Estimated effect of TT + LND versus TT alone
	All patients (n = 79)	TT alone (n = 54)	TT + LND (n = 25)	P‡‡	Effect size*§	P¶¶
Complications related to calcium control
Hypocalcaemia on POD 1	39 (49)	20 (37)	19 (76)	0·001	2·05 (1·36, 3·10)	< 0·001
Hypocalcaemic during hospital stay	23 (29)	12 (22)	11 (44)	0·048	1·98 (1·02, 3·86)	0·040
Hypocalcaemia requiring treatment during admission#	22 (28)	9 (17)	13 (52)	0·001	3·12 (1·54, 6·32)	0·002
Hypocalcaemia requiring treatment for < 12 months (transient hypoparathyroidism)#	3 (4)	2 (4)	1(4)	0·686	1·08 (0·09, 12·54)	0·950
Hypocalcaemia requiring treatment for ≥12 months (permanent hypoparathyroidism)#	15 (19)	6 (11)	9 (36)	0·009	3·24 (1·29, 8·11)	0·010
Non-calcium-related complications
Complications other than hypocalcaemia	5 (6)	2 (4)**	3 (12)††	0·319	3·55 (0·25, 22·70)	0·180
Clavien–Dindo ≥ I	3 (4)	0 (0)	3 (12)††	0·029	16·96 (0·84, 341·79)	0·069
Duration of hospital stay (days)†	3·2(2·7)	2·7(1·7)	4·2(4·0)	0·018§§	1·53 (–0·10, 3·16)¶	0·070##
Management of parathyroids‡
No. of parathyroids seen during surgery		3·15 (3)	3·08 (3)	0·763§§	–0·07 (–0·55, 0·41)¶	0·770##
No. of parathyroids preserved during surgery		3·48 (4)	2·96 (3)	0·006§§	–0·52 (–0·91, –0·13)¶	0·010##
No. of parathyroids autografted during surgery		0·09 (0)	0·32 (0)	0·030§§	0·23 (–0·03, 0·49)¶	0·080##
No. of parathyroids in histology specimen		0·39 (0)	1·0 (1)	< 0·001§§	0·61 (0·24, 0·98)¶	0·001##

Values in parentheses are percentages unless indicated otherwise;

*

values in parentheses are 95 per cent confidence intervals;

†

values are mean(s.d.);

‡

values are mean (median).

§

Relative risk, except

¶

mean difference.

#

Treatment consisting of calcium and α-calcidol.

**

Mild cough, transient minor hoarseness, resolved before discharge;

††

stridor and decreased saturation in recovery, wound infection, ICU admission and intubation for phrenic nerve palsy, respiratory distress and aspiration pneumonia. TT, total thyroidectomy; LND, lymph node dissection; POD, postoperative day.

‡‡

χ² test, except

§§

ANOVA.

¶¶

Mantel–Haenszel test, except

##

inverse-variance method.

The majority of postoperative complications were related to calcium control, and were more common among in patients who underwent LND rather than TT alone. Patients who underwent LND had significantly fewer parathyroid glands preserved in situ and were at higher risk of parathyroid excision than those who had TT alone.

Thirty-nine children (49 per cent) had a postoperative calcium measurement below normal on the first postoperative day, and this was more frequent when LND was performed (Table 5). Twenty-two of these children had symptomatic hypocalcaemia, which required treatment with calcium and α-calcidol in the immediate postoperative period. Eighteen of the 79 patients (23 per cent) were still being treated for a low calcium level up to 3 months after surgery, and 15 (19 per cent) beyond 12 months.

Mean duration of postoperative hospital stay was 3·2 days and was significantly longer among children who had LND.

Histopathology

The majority of children showed either CCH or MTC. Such findings were significantly more frequent in higher-risk groups (Table 6). The youngest age at which CCH alone and MTC was found among patients in risk class C was 1·5 and 0·8 years respectively, whereas the oldest child with a normal thyroid was aged 2·4 years. In class B, the youngest patient with MTC was 4·3 years old; this child had a codon 620 mutation.

Table 6

Histopathology results and follow-up

		ATA 2009 risk group
	All patients (n = 79)	A (n = 19)	B (n = 16)	C (n = 40)	D (n = 4)	P§
Thyroid pathology						< 0·001
Normal thyroid	10 of 79 (13)	7 of 10 (70)	2 of 10 (20)	1 of 10 (10)	0
CCH	39 of 79 (49)	10 of 39 (26)	11 of 39 (28)	18 of 39 (46)	0
MTC	30 of 79 (38)	2 of 30 (7)	3 of 30 (10)	21 of 30 (70)	4 of 30 (13)
Multifocal	11	0	3	7	1
≥ 10 mm	2	0	0	1	1
LN data
Children who had LN dissection	25	3	3	16	3
No. of LNs dissected*	5·1 (0–15)	9·5 (4–15)	3·7 (0–10)	4·7 (2–15)	6·5 (6–7)	0·453¶
Children with positive LNs	1 of 25 (4)	0 of 3 (0)	0 of 3 (0)	0 (0)	1 of 3 (33)‡	0·054
No. of positive LNs	2	0	0	0	2	0·054¶
Age of patients with specific pathological features (years)
Youngest child with MTC	0·8	11·8	4·3	0·8	0·8
Youngest child with CCH	1·5	4·4	3·7	1·5	–
Oldest child with normal thyroid	0	15·3	5·0	2·4	0
Follow-up†
Postoperative basal calcitonin
Undetectable	27 of 71 (38)	4 of 16 (25)	8 of 14 (57)	15 of 37 (41)	0 of 4 (0)	0·118
Detectable, within normal range	20 of 71 (28)	3 of 16 (19)	4 of 14 (29)	12 of 37 (32)	1 of 4 (25)	0·788
Above normal range (on at least 1 occasion)	24 of 71 (34)	9 of 16 (56)	2 of 14 (14)	10 of 37 (27)	3 of 4 (75)	0·021
Calcitonin ≥ 40 pg/ml	1 of 71 (1)	0 of 16 (0)	0 of 14 (0)	0 of 37 (0)	1 of 4 (25)‡	0·001
Clinical recurrence	1 of 79 (1)	0 of 19 (0)	0 of 16 (0)	0 of 40 (0)	1 of 4 (25)‡	< 0·001

Values in parentheses are percentages unless indicated otherwise;

*

values are mean (range).

†

Median 104·5 (range 18·2–240·6) months.

‡

The same patient. ATA, American Thyroid Association; CCH, C-cell hyperplasia; MTC, medullary thyroid carcinoma; LN, lymph node.

§

χ² test, except

¶

ANOVA.

Of the 25 patients who had LND, only one had positive lymph nodes. This child had MEN2B diagnosed by a genetic test at 9 years of age and underwent surgery aged 15 years, with a preoperative calcitonin level exceeding 40 pg/ml.

Follow-up

Median follow-up was 104·5 (range 18·2–240·6) months. Postoperative calcitonin results were available from the medical records of 71 children (90 per cent). Calcitonin was undetectable in 27 of these patients (38 per cent) and detected on at least one occasion during follow-up in 44 (62 per cent). Twenty children (28 per cent) had a postoperative calcitonin level within normal limits, and in 24 (34 per cent) it was above normal on at least one occasion. The only child with positive lymph nodes had a postoperative calcitonin level exceeding 40 pg/ml and developed clinical recurrence. No significant association was found between postoperative calcitonin status at follow-up and type of surgery (TT alone or with LND).

Discussion

Hereditary MTC, the first manifestation of MEN2, is highly penetrable. It arises from normal C-cells and progresses early in an age-dependent process to CCH and then to carcinoma. MTC is highly malignant and not curable unless diagnosed and treated before extrathyroidal spread develops¹⁵. Prophylactic thyroidectomy enabled by individual DNA profiling, and directed by guidelines built on phenotypic, biomarker and genetic characteristics, is an example of precision medicine. Precision medicine predicates that patients get the right treatment at the right time, with minimum ill consequences and maximum efficacy¹⁶. Timeliness of management of children with MEN2 has two main components: timing of genetic testing and timing of thyroidectomy, because children can benefit from prophylactic thyroidectomy only when mutations predisposing to MEN2 are identified early enough.

In the present study, timing of the genetic test was the only independent factor that predicted whether thyroidectomy was performed by the recommended age. Half of the children had a late genetic test, two-thirds showed abnormal preoperative calcitonin levels, half underwent thyroidectomy beyond the age recommended by the ATA 2009 guidelines, and half of those who had late surgery presented with MTC on histopathology. Overall, children in high-risk groups were less likely to have a genetic test or surgery at the recommended age, probably as a consequence of the shorter interval from birth to a truly prophylactic procedure.

Univariable analysis showed that late genetic testing, operation beyond the recommended age and above-normal preoperative basal calcitonin levels were factors associated with a higher rate of MTC in the specimen, whereas the only significant risk factor for MTC in multivariable analysis was a calcitonin level exceeding 40 pg/ml, as shown previously^8,17. A calcitonin level over 40 pg/ml was also significantly proportional to a delayed genetic test. Although it is desirable for all children to be tested early, the age at which they undergo genetic testing may be affected by the clinical and social context of diagnosis.

A window of opportunity is often lost when children, particularly those with MEN2B, are themselves index cases. Fewer than 20 per cent of children with MEN2B develop mucosal neuromas, intestinal ganglioneuromatosis-related constipation or tearless crying during the first year of life¹⁸, and manifestations of the syndrome can be subtle and difficult to recognize¹⁹. All children with MEN2B in this study were index cases and had a late genetic test as a consequence of a vague clinical picture. The opportunity for early diagnosis may also be lost when parents are index cases, but the children undergo genetic testing in order to distinguish sporadic from familial disease beyond the recommended age for testing.

Because genetic screening involves the dissemination of complex information, socioethical barriers, including inadequate decision-making capacity, limited access to healthcare, and cultural, religious or economic factors, may interfere with disclosure of test results to family members²⁰ and with decisions affecting children of families at risk. Up to 15 per cent of affected families do not accept screening, and gene carriers may delay or refuse prophylactic thyroidectomy for their children²¹. There appears to be no clear guidance from professional societies on informing family members at risk when an affected relative is reluctant to do so²². Genetic counsellors may not be available, and genetic issues may be poorly appreciated²³.

Adequate preventative surgery for hereditary MTC has to amalgamate technical skills with choice of procedure appropriate to the patients most likely to benefit from it, and should be performed safely and with minimal surgical invasiveness; all these objectives have been seen as premises of best surgical outcomes²⁴. All children in this study underwent TT, with no deaths or postoperative bleeding requiring reoperation. The recurrent laryngeal nerve was seen and preserved in all patients.

Twenty-five children also had lymph node dissection, which was undertaken significantly more often in high-risk classes, where the majority of patients presented with abnormal calcitonin levels. Lymphadenectomy was, however, performed in only 11 children in keeping with ATA 2009 guidelines; more than half of the children in ATA class C had LND performed outside the recommendations. On the other hand, the majority of children in lower-risk classes (A and B) had TT alone in line with ATA recommendations. In these patients, lymphadenectomy was chosen less frequently than in higher-risk classes, even when the patients presented with above-normal preoperative calcitonin levels.

This suggests that in the presence of factors such as high-risk mutations and older age at presentation (mean age 8·1 years), surgeons may be inclined to judge that a window of opportunity to cure MTC by performing thyroidectomy alone has already closed²⁵. However, it has recently been shown that, when MTC has developed, lower-risk classes exhibit metastatic potential similar to that in higher-risk classes²⁶. With studies showing no lymph node metastases when the preoperative basal calcitonin level is below 40 pg/ml, the latest guidelines²² now consider TT without central compartment dissection adequate for patients with a basal calcitonin level below 40 pg/ml. In the present study, the only child found to have positive lymph nodes at pathology had MEN2B and a preoperative calcitonin level exceeding 40 pg/ml. The present series thus shows that, when information on disease biology is limited, personalization of treatment based on the experience of the operating surgeon may be preferred over current recommendations.

Children who had lymphadenectomy stayed in hospital significantly longer and had a significantly higher risk of postoperative hypocalcaemia, with one-third developing permanent hypoparathyroidism. Age did not have any impact on the frequency and severity of complications, which depended solely on whether lymphadenectomy was performed. Limited experience of paediatric thyroidectomy for MEN2 may have contributed to the rate of postoperative hypocalcaemia. Surgical experience is a well recognized contributor to both positive perioperative and oncological outcomes²⁷, including in paediatric surgery²⁸, so discussion on whether treatment of these rare cases should take place in fewer centres of excellence may have a positive impact on outcomes of prophylactic thyroidectomy for MEN2.

At a median follow-up of 104·5 months, the sole clinical recurrence had developed in the only child with MEN2B who presented with a preoperative calcitonin level above 40 pg/ml. Follow-up calcitonin status was not associated with type of operation (TT alone or with lymphadenectomy). Detectable but fluctuating calcitonin levels are difficult to interpret²⁹. In the present cohort, these were recorded in children with MTC but also in those with a normal thyroid or CCH; this finding raises the issue not only of cancer persistence or recurrence, but also incomplete or ‘less than total’ thyroidectomy in patients with benign disease. The majority of these patients were not treated in line with ATA guidelines, yet the 8-year follow-up data did not show any convincing untoward oncological outcomes. However, the case mix and follow-up interval in this study do not allow sufficiently informed speculation over the prognostic significance of detectable levels of postoperative calcitonin. Accepting the concept of an undetectable calcitonin level as evidence of cure would suggest that nearly two-thirds of children in the present study were not cured, posing a risk of them becoming patients again, and raising the important issue of continuing follow-up. Central administration of information collected in a systematic and comprehensive way in a registry has been seen as a way to solve the practical problems of follow-up³⁰.

Although this study has limitations related to the retrospective collection of data over a long period, the results are believed to be a true reflection of the management of children with MEN2 in the UK over the past 23 years. The findings suggest that undertaking genetic testing as early as possible could increase the rate of truly prophylactic thyroidectomy in children with MEN2. Research is needed to understand patients' preferences, opinions of healthcare professionals and parents' acceptance of genetic testing.

Collaborators

Other members of the MEN2 in Children UK Collaborative Group are: P. Hindmarsh, M. Dattani, H. Spoudeas, R. Amin (University College Hospital and Great Ormond Street Hospital, London); J. Watkinson, N. Shaw (Birmingham Children's Hospital, Birmingham); S. Alvi, R. Squire (Leeds General Infirmary, Leeds); B. Harrison, J. Wales, N. Wright (Sheffield Children's Hospital, Sheffield); P. Clayton (Royal Manchester Children's Hospital, Manchester); F. Eatock (Royal Belfast Hospital for Sick Children, Belfast); C. Buchanan (King's College Hospital, London); A. Albanese, G. Bano (St George's Hospital, London); H. Storr, M. Korbonits, E. Gevers (St Bartholomew's Hospital, London); T. Randell, L. Denvir (Nottingham Children's Hospital, Nottingham); J. Davies (Southampton General Hospital, Southampton); T. Cheetham, T. Lennard (Newcastle Royal Victoria Hospital, Newcastle upon Tyne); J. Warner (University Hospital of Wales, Cardiff); D. Smith (Ninewells Hospital, University of Dundee, Dundee); L. Crowne (Bristol Royal Infirmary, Bristol); T. Hulse, L. Izatt (St Thomas' Hospital, London); J. Blair (Alder Hey Children's Hospital, Liverpool).

[Corrections added on 8 May 2018, after first online publication: T. Randal has been corrected to T. Randell and L. Crown corrected to L. Crowne]

Acknowledgements

The study was funded by the Association for Multiple Endocrine Neoplasia Disorders (AMEND).

Disclosure: The authors declare no conflict of interest.