Headache is one of the most common neurological complaints among children and adolescents, accounting for approximately 40%–50% of pediatric neurology visits [1,2]. Because pediatric headache encompasses a broad spectrum of etiologies ranging from benign to potentially serious, neuroimaging is frequently performed to exclude secondary causes. Magnetic resonance imaging (MRI) is the preferred modality for evaluating pediatric headache. Incidental findings (IFs), defined as unexpected abnormalities unrelated to the presenting symptom, are detected in approximately one-fifth of cases [3,4]. Among these IFs, the reported prevalence of nonspecific white matter lesions (WMLs) varies widely (0.7% to 47%), reflecting heterogeneity in MRI resolution, lesion-size thresholds, and inclusion criteria across pediatric cohorts [1,5-9]. According to established radiologic classification systems [10-12], IFs on brain MRI are categorized into four grades based on clinical urgency, ranging from grade 1 (no referral required) to grade 4 (immediate referral required). WMLs are typically considered grade-2 findings and are generally regarded as benign but warrant routine clinical assessment. Nevertheless, when neurological abnormalities or atypical clinical features are present, such lesions must be distinguished from other causes, including ischemic, demyelinating, and metabolic disorders [5,13]. Numerous studies have investigated the prevalence, risk factors, and temporal evolution of WMLs in adults, particularly among patients with migraine [14-19]. In contrast, pediatric data remain limited, and most existing studies are cross-sectional [1,5,6,20-22], with only a few reporting longitudinal follow-up or progression patterns [7,13,23].

Given this knowledge gap, the present study aimed to determine the prevalence of WMLs in pediatric patients with headache and to describe their radiologic changes over time and associated clinical characteristics. By elucidating the natural history of these lesions, we sought to provide clinically relevant evidence to support the interpretation and management of incidental WMLs in pediatric headache practice.

This retrospective observational study enrolled children and adolescents (<18 years) who presented with headache at a tertiary-care teaching hospital between September 2006 and July 2023. In total, 2,315 pediatric patients presenting with headache were screened. After excluding patients whose primary complaint was not headache (n=95) and those with neurological or systemic disorders known to cause WMLs (n=18), 515 patients who underwent brain MRI within 1 year of presentation were included in the final analysis. Headache diagnoses were established by a pediatric neurologist according to the International Classification of Headache Disorders, 3rd edition (ICHD-3).

Brain MRI was performed at the discretion of the treating physician in accordance with pediatric headache imaging guidelines [2] and, in some cases, at the caregiver’s request despite the absence of guideline-based indications. Abnormal neurological signs were defined as objective focal deficits identified on examination (e.g., weakness, ataxia, and unilateral mydriasis) or other abnormal physical findings (e.g., macrocephaly). Abnormal neurological symptoms were defined as subjective focal complaints, including transient weakness, numbness, or visual disturbances. Standard MRI sequences included sagittal and axial T2-weighted imaging, T2 fluid-attenuated inversion recovery (T2-FLAIR), T1-weighted imaging, diffusion-weighted imaging, apparent diffusion coefficient mapping, and susceptibility-weighted imaging. Gadolinium contrast agent was administered only when clinically indicated. All scans were interpreted by board-certified radiologists as part of routine clinical practice.

MRI findings were categorized as normal or abnormal. Abnormal findings were further classified as incidental or serious according to established radiologic criteria [3-5,12]. Rhinosinus disease (including sinusitis or mucosal thickening) was regarded as incidental unless it was considered a potential cause of headache or showed clinical improvement after treatment, in which case it was classified as secondary.

WMLs were defined as nonenhancing T2- or FLAIR-hyperintense foci ≥3 mm in diameter [14]. Lesions were described as discrete (single or multiple, well-demarcated) or confluent (merged without clear margins). For prevalence estimation, only WMLs documented as abnormal findings in the official radiology reports were included.

Follow-up MRI was not performed according to a predefined study protocol but was obtained at the discretion of the treating physician, primarily when clinically indicated (e.g., worsening headache symptoms) or at parental request. Radiologic progression was defined as any increase in lesion number or size on follow-up MRI; in the Results, such findings are described as ‘interval radiologic changes.’ All WML images were independently reviewed by a pediatric neurologist to confirm lesion characteristics.

Clinical data included age, sex, headache duration and frequency, headache type, medical history, family history of migraine, indications for neuroimaging, and MRI findings.

The study was approved by the Institutional Review Board (IRB) of Daegu Catholic University Medical Center (IRB No. DCUMC 2025-08-004). Informed consent was waived owing to the retrospective design, and the study was conducted in accordance with the Declaration of Helsinki.

Continuous variables are presented as the mean±standard deviation and range, whereas categorical variables are presented as counts and percentages. Because of the small number of cases, no a priori sample size calculation or hypothesis testing was performed for the WML subgroup. Accordingly, the analysis was descriptive in nature.

A total of 515 pediatric patients who underwent brain MRI for headache were included. The mean age was 10.7±3.6 years (range, 2 to 17), and 259 (50.3%) were boys. Of these, 47 (9.1%), 240 (46.6%), and 228 (44.3%) were aged <6, 6–11, and >11 years, respectively. Headache duration was less than 3 months in 197 patients (38.3%), and headaches occurred daily in 252 (48.9%); data for these variables were unavailable for a small subset of patients. A total of 145 patients (28.2%) had a medical history, and 101 (19.6%) had a family history of migraine. MRI was performed for clinical indications in 267 patients (51.8%). The main indications were progressive headache (18.1%), young age (<6 years; 9.1%), morning or awakening headache (8.0%), and abnormal neurological symptoms or signs (7.8% and 3.3%, respectively). Less common indications included posterior headache, headache aggravated by the Valsalva maneuver, absence of a family history of migraine, and worst-ever headache. Because several patients met multiple criteria, these categories were not mutually exclusive.

According to the ICHD-3, 164 patients (31.8%) were diagnosed with migraine, including 113 (21.9%) without aura, 49 (9.5%) with aura, and two (0.4%) with migraine variants. Tension-type headache was diagnosed in 37 patients (7.2%), and secondary headache was diagnosed in 14 (2.7%), including five patients with symptomatic sinusitis and nine with serious intracranial lesions identified on MRI. The remaining 300 patients (58.3%) were classified as having unspecified primary headache.

Abnormal MRI findings were observed in 126 patients (24.5%). Serious abnormalities occurred in nine patients (1.7%), including four tumors, two cases of moyamoya disease, and one case each of acute disseminated encephalomyelitis, cerebellitis, and subdural hematoma. In contrast, IFs were observed in 117 patients (22.7%), the most common of which was rhinosinus disease (n=73, 14.2%) (Table 1). Because some patients had multiple abnormalities, the totals exceeded the number of patients. Only a small subset of patients with abnormal neurological signs or symptoms had clinically significant lesions on MRI.

Nonspecific WMLs were detected in eight of the 515 patients (1.6%). The clinical and radiologic characteristics of these patients are summarized in Table 2. Among the eight patients, seven were diagnosed with unspecified primary headache and one with migraine with aura. No patient had focal neurological deficits at presentation, and no consistent clinical or demographic risk factors were identified. Most lesions were supratentorial (n=7, 87.5%), with a mean maximum diameter of 10.0 mm (range, 4.3 to 24.9). Follow-up MRI was available for three patients (37.5%); two demonstrated interval increases in lesion extent on repeat imaging. Headache frequency or severity also increased during the same period. Importantly, no new focal neurological deficits were documented in any patient during follow-up. To illustrate the clinical spectrum and radiologic course, three representative cases (patients 6–8) are presented below.

Patient 6: A 3-year-old girl developed intermittent headaches after recovery from Kawasaki disease. MRI performed at parental request revealed a solitary cerebellar T2-hyperintense lesion that remained radiologically stable over 7 years of follow-up (not shown).

Patient 7: A 5-year-old boy presented with progressively more frequent headaches without focal neurological deficits. Baseline MRI revealed confluent hyperintense lesions in the periventricular trigones (Fig. 1A). Follow-up MRI performed 2 years later demonstrated mild enlargement of the lesions (Fig. 1B). Headache frequency also increased during the same period. No new neurological deficits were observed.

Patient 8: A 13-year-old girl had a long-standing history of frontal headaches that progressed from twice monthly to nearly daily. MRI performed for endocrine evaluation 4 years earlier had incidentally revealed multiple small subcortical T2-hyperintense lesions (Fig. 1C and E). Follow-up MRI, obtained because of worsening headaches, demonstrated an increased extent of the lesions (Fig. 1D and F). Another MRI scan obtained 3 years later demonstrated radiologic stability despite fluctuating headache severity.

The interval radiologic changes in representative patients 7 and 8 are illustrated in Fig. 1.

This study demonstrated that nonspecific WMLs are uncommon in pediatric patients with headache, occurring in only 1.6% of our cohort. Although interval radiologic changes were observed in a small subset, none of the patients experienced neurological deterioration. Reported prevalence rates of WMLs in pediatric patients with headache vary widely, ranging from 0.7% to 47% [1,5-9] and up to 60% in selected migraine cohorts [20-22,24,25]. This heterogeneity likely reflects differences in MRI resolution, diagnostic criteria, and inclusion thresholds across studies. For example, earlier studies reported higher rates of WMLs in pediatric patients with migraine [21,22,25]. However, Ackley et al. [6] recently observed similar frequencies among patients with migraine, other primary headaches, and controls, suggesting that most pediatric WMLs are incidental rather than migraine-specific. Their occurrence in the healthy pediatric population (1% to 2%) further supports this interpretation [3,12].

Longitudinal data from adults indicate that WMLs may enlarge, remain stable, or regress over time [14,15,18], suggesting that they represent a dynamic rather than a static process. Although pediatric data are limited, available studies suggest similar patterns: most lesions remain unchanged, whereas some progress or resolve without neurological consequences [7,13,23]. In our series, interval radiologic changes were observed in two of the three patients who underwent follow-up imaging, without accompanying neurological deterioration. However, because repeat MRI was available only in a small, clinically selected subset, these observations should not be considered representative of pediatric patients with headache and WMLs more broadly. Notably, incidental WMLs have also been reported in otherwise healthy children. Furthermore, apparent interval changes on repeat imaging may partly reflect detection variability, differences in imaging technique, or other technical factors rather than true pathological progression. Taken together, these considerations support a cautious interpretation of radiologic changes in pediatric patients with headache.

The pathophysiological basis of WMLs remains unclear. No consistent association has been demonstrated between WMLs and headache type, duration, or frequency, or with migraine aura or the presence of a patent foramen ovale, in either adults or children [16,18,20,22,25]. Histopathologically, increased white matter signal intensity may correspond to gliosis, demyelination, or axonal loss secondary to microvascular or inflammatory injury [5,16,18]. In adults, WMLs are commonly considered markers of small-vessel disease and aging [18]. However, in migraine, WMLs typically occur in the deep and subcortical white matter, predominantly within the frontal and parietal lobes, rather than in the periventricular regions more commonly affected in older adults, suggesting a different underlying pathological mechanism [5,16,19,21,22]. Proposed mechanisms, including transient cerebral hypoperfusion, inflammatory injury with blood–brain barrier disruption, and cortical spreading depolarization, have been derived largely from adult migraine studies [16]. Reports of lesion regression further support the possibility that some MRI changes may reflect transient, potentially reversible processes, such as fluid shifts or resorption of interstitial fluid, rather than permanent tissue damage [14,18,23]. Because these mechanisms have largely been inferred from adult migraine studies, their applicability to pediatric WMLs remains uncertain, as children generally lack vascular risk factors or chronic microangiopathy [5,20]. Accordingly, pediatric-specific longitudinal studies are needed to better define the biological basis and clinical significance of headache-associated WMLs.

When these lesions are detected incidentally, clinicians often face uncertainty regarding their clinical significance and the need for further evaluation or follow-up. This concern may increase when interval radiologic changes are observed on repeat imaging. To date, no consensus guidelines have been established for the management of incidental WMLs in pediatric patients. Our findings, consistent with previous reports [7,13,23], suggest that pediatric WMLs are often clinically silent and may fluctuate without adverse outcomes. Given the largely benign nature of nonspecific WMLs in pediatric patients with headache, a conservative strategy of clinical observation may be reasonable in selected patients with normal neurological examinations, rather than routine short-interval repeat MRI.

This study has several limitations. First, headache classification may have been imprecise because younger children often have difficulty describing their symptoms, and limited follow-up hindered strict application of the ICHD-3 criteria, resulting in a relatively high proportion of unspecified primary headache. Second, as this was a single-center retrospective study, imaging protocols were not standardized, and MRI acquisition was largely determined at the discretion of individual physicians. Third, quantitative volumetric analysis was not performed, and both MRI protocols and scanner hardware varied over the long study period. These factors may have affected lesion detectability and assessment of interval changes, and we could not determine whether the observed increases represented true enlargement of preexisting lesions or detection of newly visible lesions. Finally, follow-up MRI was available in only three patients and was performed on the basis of clinical indication rather than a predefined protocol, limiting conclusions regarding the natural evolution of pediatric WMLs and potentially introducing selection bias in the assessment of interval changes. The small number of WML cases further precluded statistical comparisons.

In conclusion, nonspecific WMLs were rare in pediatric patients with headache. Among the small subset of patients who underwent follow-up MRI, interval radiologic changes were observed, but no neurological deterioration occurred during the observation period. These findings may help clinicians contextualize incidental WMLs and support counseling for patients and their families. Larger prospective studies with systematic follow-up are needed to further clarify the natural history and clinical significance of pediatric WMLs.