Retrospective Cohort Analysis of the Impact of Puberty on Plexiform Neurofibroma Growth in Patients with Neurofibromatosis Type 1

Abstract

Objective

To assess the hypothesis that plexiform neurofibroma (PN) growth rates increase during puberty.

Study design

PN growth rates before and during puberty were compared in a retrospective cohort of children with neurofibromatosis type 1 with puberty defined by Tanner staging. Of 33 potentially eligible patients, 25 had adequate quality magnetic resonance imaging for volumetric analysis and were included in ≥1 anchor cohort. Volumetric analysis was performed for all available imaging studies within the 4 years before and after puberty, and before and after 9- and 11-year-old anchor scans. Linear regression was performed to estimate the slope of change (PN growth rate); growth rates were compared with paired t test or Wilcoxon matched-pairs signed rank test.

Results

There were no significant difference in rates of PN growth in milliliters per month or milliliters per kilogram per month in the prepubertal vs pubertal periods (mean, 1.33 ± 1.67 vs 1.15 ± 1.38 [P = .139] and −0.003 ± 0.015 vs 0.002 ± 0.02 [P = .568]). Percent increases of PN volumes from baseline per month were significantly higher pre-pubertally (1.8% vs 0.84%; P = .041) and seemed to be related inversely to advancing age.

Conclusions

Puberty and its associated hormonal changes do not seem to influence PN growth rate. These findings support those previously reported, but from a typical population of children with neurofibromatosis type 1 with puberty confirmed by Tanner staging.

Plexiform neurofibromas (PN) are benign peripheral nerve sheath tumors that develop in nearly one-half of patients with neurofibromatosis type 1 (NF1). PN can cause significant morbidity, including pain, disfigurement, and functional deficits, and may undergo malignant transformation.^1,2 There is concern that growth rates of PN accelerate during periods of hormonal excess, such as puberty and pregnancy.³ Although prior studies have evaluated the impact of pregnancy on tumor growth in NF1 with inconsistent findings,^4,5 only 1 prior study has evaluated the natural history of PN growth during puberty. That study did not identify a significant difference in prepubertal and pubertal growth rates in the 16 evaluated patients. However, that study had several limitations, including a sample of patients referred to the National Cancer Institute (NCI) for treatment of their PN and estimations of puberty based on age.⁶ The estimation of puberty based on age poses a threat to the validity of these findings, because there is an increased prevalence of abnormal patterns of pubertal development in children with NF1.⁷ Thus, given the potential limitations in interpretation of prior data, the true impact of puberty on PN in the general population of children with NF1 remains unknown.

Treatment for PN is often initiated with the onset of progressive growth and/or functional deficits; however, despite the successes of mitogen-activated kinase (MEK) inhibitors for these tumors, many morbidities once developed, such as nerve damage, cannot be reversed fully.^8,9 Thus, it is crucial to identify whether periods of accelerated PN growth exist in children to support the timing for surveillance as well as the initiation of treatment. Based on this knowledge gap, the objective of this study was to compare PN growth rates before and during puberty in a representative, clinic-based population of children with NF1 where puberty was rigorously defined by specific Tanner staging criteria.

Methods

This single-center, retrospective study received approval from the institutional review board. Informed consent was waived. The study sample consisted of children with a PN who received longitudinal care at the Children’s Hospital of Philadelphia Neurofibromatosis Program and had a clinical diagnosis of NF1 using the National Institutes of Health Consensus criteria¹⁰ or a documented pathogenic constitutional NF1 gene mutation. Three time periods (anchor cohorts) of interest were established (puberty, 9 years of age, and 11 years of age). The 9- and 11-year-old cohorts were selected as estimations of puberty for comprehensive comparison with the prior study of NCI-based patients.⁶ Patients were screened for eligibility for inclusion in each of the cohorts and were included in an anchor cohort if they had at least 2 magnetic resonance imaging (MRI) studies of adequate quality for volumetric analysis in the 2 years before and in the 2 years after the anchor of interest between January 1, 2002, and December 31, 2020 (a minimum of 4 MRIs total). Anchor images for analysis were obtained within 3 months before or after the first notation of onset of puberty, at 9 years of age, and at 11 years of age. All imaging studies performed within the 4 years before and after the anchor images were included in the analysis. For the pubertal cohort, patients were excluded if pubertal status was not adequately documented or if the patient was greater than Tanner stage 3 at time of first MRI, to capture patients in the early stages of pubertal development. For all cohorts, patients were excluded if they received treatment with a MEK inhibitor, cabozantinib (multiple tyrosine kinase inhibitor), or pegylated interferon at any time between the onset of puberty or 8 years of age (whichever occurred first) and 16 years of age, because these are the only therapies to date that have shown a significant prolongation of tumor growth or response rate.^8,9 Patients were also excluded if they underwent debulking surgery or received a MEK inhibitor for a non-PN tumor. Patients who received conventional cytotoxic chemotherapy for a non-PN tumor, such as glioma, were not excluded.

Puberty was defined by Tanner staging as noted in the medical chart by either a pediatric endocrinologist or neuro-fibromatosis specialty clinician. The onset of puberty in boys was defined as the first notation in the medical chart of Tanner 2, 2–3, or 3 pubic hair or testes. Onset of puberty in girls was defined as first notation of Tanner 2, 2–3, or 3 breast development. If the latter information was not available, then onset of puberty in girls was defined as first notation in the clinical chart of Tanner 3 pubic hair (however, if menarche occurred >2 years after this date, then this was not considered onset of puberty and patients were not eligible).^11–13 Charts were abstracted for demographic and clinical characteristics, including age at onset of puberty, sex, body weight in kilograms at the time of the MRI, target tumor location, tumor-directed therapy, date of birth, and dates of all MRI scans. For the analysis of tumor volume in milliliters, volumetric analysis was performed at the NCI as previously described.¹⁴ Patients were assessed on MRI for the presence of distinct nodular lesions (DNLs) within the target PN; DNLs were defined as encapsulated-appearing, well-demarcated lesions lacking a central dot sign.¹⁵

Linear regression analyses were performed to estimate the slope of volume change (PN growth rate) for each subject in target tumor volume per patient weight (mL/kg/month), absolute volume change (mL/month), and percent change from baseline volume (%/month). Although the Response Evaluation in Neurofibromatosis and Schwannomatosis international collaboration recommends change in absolute volume as a response criterion for PN clinical trials,⁸ change in target tumor volume per patient weight was also calculated to evaluate the relationship of tumor volume to changes in body mass with patient growth and development. Baseline tumor volume for the prepubertal, pre-9- and 11-year-old rate analyses was the volume calculated for the first image obtained within the 2-year period before the anchor image; the baseline tumor volume for the pubertal, 9- and 11-year old rates was the anchor image volume. For descriptive statistics, the mean was used for consistency with the prior study, unless otherwise indicated.⁶ To test if there is a difference in the rate of tumor growth before vs after puberty, the slopes constructed for each patient for the intervals of interest were compared (prepubertal/pubertal, pre/post age 9 and pre/post age 11) using a paired t test or Wilcoxon matchedpairs signed rank test. The data were analyzed by patient sex and in a cohort restricted to DNL as well, to assess for potential differences.

To assess further if the rate of tumor growth changes with age, the slopes were constructed for pre-9 years, 9–11 years, post-11 years for patients with sufficient data for inclusion in all 3 cohorts and compared using paired t test or Wilcoxon signed-rank test. All analyses were performed using Stata 15 (StataCorp) and a 2-sided P value of <.05 was considered statistically significant.

Results

Of the 33 potentially eligible patients identified, 8 were not evaluable owing to inadequate quality of imaging for volumetric analysis; 25 patients had adequate image quality for volumetric analysis and were included in ≥1 anchor cohort. Five patients screened for inclusion were determined to be ineligible owing to treatment with a MEK inhibitor; initiation of therapy occurred at a mean of 2.8 years before the onset of puberty (range, 0–5 years).

Pubertal Cohort

For 15 patients, slopes of PN growth could be calculated before and during puberty. Six patients (40%) were female and 9 (60%) were male (comparison of demographics of CHOP cohort with the NCI puberty cohort provided in Table I). Longitudinal change in tumor volume was also analyzed for all 15 patients, with a median duration of imaging follow-up of 70 months (Figure 1). The mean age of pubertal onset was 11.5 years (range, 7.2–13.6 years). The mean target tumor volume at pubertal onset was 136.8 mL with a mean initial target tumor volume per patient body weight of 3.7 mL/kg. One patient had a target tumor with a distinct nodular appearance (DNL) on MRI. There was no significant difference in tumor growth over time in mL/month and mL/kg/month before vs after the onset of puberty (Table II, Figure 2, A and B). However, there was a significant difference in percent change from baseline target tumor volume per month before vs after the onset of puberty (P = .041), with tumors increasing 1.8% from baseline per month prepubertally and 0.84% during the pubertal period (Table II, Figure 2, C). In a comparison of prepubertal and postpubertal growth rates with the overall longitudinal growth rate of a patient’s target PN, 71% of patients had a higher absolute rate of growth prepubertally, whereas only 36% had a higher rate postpubertally (Table III). Of note, 2 patients were in a PN-directed clinical trial during the period of interest (tipifarnib¹⁶ and methotrexate/vinblastine, with the latter limited to only 3 months duration of exposure to chemotherapy before the onset of puberty).

Table I.

Characteristics of the Children’s Hospital of Philadelphia puberty cohort in comparison with the NCI puberty cohort⁶

Patient characteristics	CHOP cohort (n = 15)	NCI cohort (n = 16)
Sex
Females	6(40)	5(31)
Males	9(60)	11 (69)
Target tumor location
Head/neck	10(67)	NR
Trunk	4(27)
Extremity	1 (6)	9 (56)†
PN-directed therapy during period of interest	2(13)*
Age pubertal onset, years		12.5 ± 2.7‡
Females	10.5 ± 2.1
Males	12.3 ± 0.9	10.7 ± 3.0‡
Tumor volume on MRI at pubertal onset, mL	136.8 ± 141.9	516
Duration of MRI observations, years	8.1 ± 2.5	7.6
Tumor burden/patient body weight (mL/kg)
Initial	3.68 ± 4.7	27.6 ± 29.8
Final	2.92 ± 3.6	26.4 ± 29.8

NR, not reported.

Values are mean ± SD or number (%).

^*Tipifarnib and methotrexate/vinblastine.

^†No patients included received treatment with a MEK inhibitor.

^‡Estimation based on age and change in Tanner staging.

Figure 1.

Percent change of each target PN volume over time for the pubertal cohort (n = 15). Asterisks indicate the onset of puberty for each patient. The median duration of imaging follow-up for patients in the pubertal cohort was 70 months (range, 37–99 months).

Table II.

Rates of PN growth by pubertal status and age

Measurements of PN growth	Pubertal (n = 15)	P value	9-Year-old (n = 16)	P value	11-Year-old (n = 17)	P value
Rate of absolute PN growth, mL/month
Pre anchor	1.33 ± 1.67	.139	1.09 ± 1.29		6.02 ± 20.06	.905
Post anchor	1.16 ± 1.38		1.31 ± 1.42	.133	10.16 ± 37.8
Rate of PN growth per body weight, mL/kg/month
Pre anchor	−0.003 ± 0.015	.568	0.025 ± 0.05		0.15 ± 0.57	.079
Post anchor	−0.002 ± 0.022		−0.001 ± 0.017	.03	0.343 ± 1.42
Percent change of tumor volume above baseline per month
Pre anchor	1.8 ± 2.73	.041	1.95 ± 2.67		2.0 ± 2.8	.035
Post anchor	0.84 ± 0.74		1.03 ± 0.75	.029	1.03 ± 0.95

Values are mean ± SD

Figure 2.

Paired comparison of growth rates of PN before and after puberty. Growth rates in milliliters per month (A), milliliters per kilogram month (B), and percent change from baseline tumor volume (C). No significant difference was found in absolute growth or growth by body weight however percent increase of tumor volume above baseline per month was significantly different, with higher rates observed in the prepubertal period of evaluation.

Table III.

Per-patient comparison of prepubertal and pubertal growth rates with longitudinal growth rate of target PN in puberty cohort in milliliters per month and percent change from baseline per month (n = 14^*)

Per-patient longitudinal absolute rates of growth (mL/month)	Absolute rate prepuberty (mL/month)	Absolute rate pubertal (mL/month)	Longitudinal percent change from baseline (%/month)	Percent change prepuberty (%/month)	Percent change pubertal (%/month)
3.68	3.83	3.5	1.45	1.23	0.78
0.015	−0.0042	−0.03	0.282	−0.09	0.69
1.59	1.84	2.76	1.49	1.4	1.53
2.5	3.28	1.8	1.06	1.15	0.624
0.204	0.13	0.23	1.21	0.654	0.9
0.44	1.13	0.39	0.31	0.825	0.198
1.85	1.49	2.7	2.62	1.16	1.47
1.63	1.78	1.73	4.4	1.44	0.91
0.002	−0.006	−0.031	0.0011	−0.06	−0.12
0.298	0.38	0.35	8.75	11.1	2.9
0.069	0.13	0.06	1.21	2.3	0.58
0.37	0.29	0.24	0.367	0.223	0.154
0.0225	0.023	0.02	0.568	0.51	0.147
0.0765	0.12	0.025	2.37	3.48	0.61
Proportion of patients with growth rate above the longitudinal rate	71%	36%	Proportion of patients with percent change above the longitudinal rate	36%	14%

^*One patient was excluded from longitudinal analysis owing to tumor debulking.

A repeat analysis was performed restricted to typical appearing PN in the pubertal period to evaluate the potential impact of DNL growth pattern on the study findings (n = 14). A lack of significant differences remained for both absolute rates and tumor growth by body weight prepubertally vs pubertal and compared with the entire cohort inclusive of the DNL (1.4 vs 1.2 mL/month [P = .178]; −0.003 vs −0.001 mL/kg/month [P = .55]). An additional post hoc analysis was performed and restricted to patients in the pubertal period who did not receive treatment with tipifarnib or methotrexate/vinblastine (n = 13). There was no significant difference in tumor growth over time in mL/month and mL/kg/month before compared with the onset of puberty (0.97 vs 0.84 mL/month [P = .468]; 0.003 vs 0.008 mL/kg/month [P = .258]); however, a significant difference was observed in percent change from baseline target tumor volume per month before vs after the onset of puberty (P = .046), with tumors increasing 1.8% from baseline per month prepubertally but only 0.76% during the pubertal period.

There was no significant difference in absolute and per body weight tumor growth rates in a subanalysis by sex (Table IV). Both sexes demonstrated higher rates of percent change in tumor volume per month prepubertally; however, a statistically significant difference was only observed in males.

Table IV.

Analysis of PN growth rates during puberty by sex

Characteristic	Females	P value	Males	P value
Rate of absolute tumor growth per month, mL/month
Prepubertal	1.32 ± 2.1	.68	1.33 ± 1.43	.485
Pubertal	1.16 ± 1.59		1.15 ± 1.32
Rate of tumor growth per body weight, mL/kg/month
Prepubertal	0.002 ± 0.01	.917	−0.006 ± 0.02	.594
Pubertal	0.009 ± 0.03		−0.01 ± 0.01
Percent change of tumor volume above baseline per month
Prepubertal	2.8 ± 4.1	.463	1.1 ± 1.0	.021
Pubertal	1.3 ± 0.86		0.53 ± 0.49

Values are mean ± SD.

Age-based Cohorts

Age-based cohorts were evaluated as estimations of puberty for comprehensive comparison with the prior study of NCI-based patients. Sixteen of the 25 patients met criteria for inclusion in the 9-year-old cohort analysis (7 females, 9 males). The mean tumor volume at 9 years of age was 108.4 mL. There was a significant difference in PN growth rates in mL/kg/month and percent change from baseline before and after 9 years of age (Table II), with higher growth rates occurring before 9 years of age. There was no significant difference in absolute rates of tumor growth in mL/month before vs after 9 years of age.

Seventeen patients had adequate data for inclusion in the 11-year-old cohort analysis (6 females, 11 males). The mean tumor volume at 11 years of age was 407.4 mL (range, 5.9–4853.0 mL). There was no significant difference in the tumor growth over time in mL/kg/month before vs after 11 years of age, nor was there a significant difference in mL/month before or after 11 years of age (Table II). There was a significant difference in percent change from baseline, with tumors increasing by a mean of 2% per month before the age of 11 and 1.03% after (P = .035).

Ten patients (5 females, 5 males) had sufficient longitudinal clinical and imaging data for a continuous age-based comparison of tumor growth rates across the periods of interest (before 9, 9–11 years of age, ≥11 years of age) (Table V). The median number of images included in the analysis was 10 per subject (range, 7–12 per subject). The mean age of onset of puberty for this restricted cohort was 11.4 years. Tumor growth by body weight was significantly different between the 9-year-old and 11-year-old periods of observation, with more rapid tumor growth observed before 9 years of age compared with 11 years of age and older (0.034 vs 0.005 mL/kg/month; P = .012). This difference was not significant when comparing tumor growth rates before age 9 vs 9–11 years of age. As was observed in the larger cohort, the percent change per month from baseline seemed to be inversely correlated with age with higher percent change in growth in patients <9 years of age, yet this difference did not reach statistical significance.

Table V.

Restricted analysis of patients with sufficient observations for inclusion in all cohort analyses (age-based analysis)

Characteristics	n = 10	P value
Sex
Female	5 ± 50	NA
Male	5 ± 50
Age of pubertal onset, years	11.4 ± 2.0	NA
Target tumor volume, mL
9 Years	112	NA
11 Years	145
Puberty	142.5
Percent change from baseline per month
Pre 9 years	2.3 ± 3.37	Ref
9–11 years	0.98 ± 0.87	.075
Post 11 years	0.86 ± 0.90	.114
Rate of tumor growth per month, mL/kg/month
Pre 9 years	0.034 ± 0.07	Ref
9–11 years	−0.003 ± 0.03	.141
Post 11 years	0.005 ± 0.03	.012
Rate of absolute tumor growth per month, mL/month
Pre 9 years	1.15 ± 1.62	Ref
9–11 years	1.32 ± 1.7	.721
Post 11 years	1.4 ± 1.47	.646

NA, not applicable; Ref, reference.

Values are mean ± SD unless otherwise noted.

Longitudinal change in tumor volume was analyzed for all 25 patients, with a median duration of imaging follow-up of 90 months (range, 35.0–134.4 months). The median number of distinct MRI time points included per subject in the longitudinal analysis was 8 (range, 4–15) (Figure 3). Of note, in the assessment of the 25 patients in the entire longitudinal cohort, 3 tumors were outliers in growth (increase of tumor volume of >400% from baseline). In a post hoc analysis comparing these 3 patients (2 males, 1 female) with the rest of the cohort, there were no differences in in age at first MRI (P = .490), baseline tumor volume (P = .247), or tumor location (distributed across head, neck, and chest). Importantly, all 3 patients remain in active follow-up; to date, none of the tumors are concerning for transformation to malignant peripheral nerve sheath tumor, although 1 tumor underwent surgical intervention 2 years after the study period with pathology and molecular analysis consistent with atypical neurofibromatous neoplasm of uncertain biological potential.

Figure 3.

Percent change of each target PN volume over time (n = 25). The mean percent increase in tumor volume per month was 2.6%, with a median of 1.2%. The median duration of imaging follow-up for patients was 90.0 months (range, 35.0–134.4 months).

Discussion

Puberty and its associated hormonal changes do not seem to influence growth of PN in children with NF1. This study did not identify a relationship between the onset of puberty and an acceleration in tumor growth in the present cohort. In addition, there was no difference in growth rates before vs at pubertal onset when stratified by patient’s sex assigned at birth. Furthermore, the growth rate of PN appeared to correlate inversely with age; younger patients experienced a higher increase in tumor volume per month as compared with older children.

These findings support the previously reported results from Dagalakis et al, despite substantial clinical differences between the 2 study cohorts.⁶ In the former study, males were over-represented, baseline tumor volumes were much greater, and there were more patients receiving tumordirected therapy during the period of interest compared with the present study. The greater mean tumor volume of the NCI cohort is the likely cause for the apparent differences in growth by body weight, with the NCI cohort appearing to have a greater change in monthly tumor volumes (−0.16 mL/kg/month compared with −0.002 mL/kg/month). Importantly, despite these differences in clinical features, neither study identified a period of increased growth rates during puberty. In addition, the present study’s findings add to the growing literature describing increased rates of PN growth in younger children.^17–19 This increased rate of growth observed with younger age was sustained in a comparison of prepubertal with longitudinal volumes and in the restricted longitudinal analysis of patients included in all anchor cohorts, further supporting the association of young age and increased rates of tumor growth.

The present study has several limitations, including a small sample size, which may have limited the detection of a true difference in growth rate during puberty. However, the present study’s results are consistent with those of the previously published NCI study; the 2 studies together suggest that acceleration of growth rates with the onset of puberty is unlikely. The present study also did not capture serologic correlates of puberty; however, the rigorous clinical definition of puberty likely lessened the impact of misclassification of pubertal onset. In addition, analysis of the relationship of tumor location to growth rate was limited by sample size. No patients included in this analysis had tumors of the external genitalia or reproductive organs. Thus, this study was not able to examine the impact of elevated sex hormones on the growth of tumors within reproductive anatomy, which have been previously reported to increase in size during periods of hormone excess.²⁰ Finally, this study excluded patients who received treatment with a MEK inhibitor and it is possible that the rate of tumor growth would have accelerated during puberty in these 5 patients. However, the exclusion of these patients is not likely to have impacted the results significantly because these patients were initiated on MEK inhibitors before the onset of puberty, suggesting clinically significant PN growth starting before pubertal development. Although the behavior of PN during puberty in patients receiving MEK inhibitors is of interest, this will likely be examined in the upcoming phase II clinical trial of selumetinib for the prevention of PN growth and morbidity in NF1 conducted by the NF Clinical Trials Consortium.

In summary, identifying key periods of accelerated PN growth in patients with NF1 is crucial for anticipatory guidance for patients and families and to define the optimal time for initiation of therapy in patients with inoperable and/or progressive tumors. Puberty does not seem to impact or accelerate tumor growth. In addition, age seems to correlate inversely with rate of growth, with an increased rate of growth at <9 years of age compared with the pubertal and 11-year-old cohorts. Additional prospective studies of the natural history of PN during puberty for both treated and untreated patients are necessary to validate further these findings and ultimately to improve patient care.

Glossary

DNL

Distinct nodular lesion

MEK

Mitogen-activated kinase

MRI

Magnetic resonance imaging

NCI

National Cancer Institute

NF1

Neurofibromatosis type 1

PN

Plexiform neurofibroma