Method-dependent nomograms for ultrasonographic measurement of the fetal thymus: comparison of the THY-box technique with two- dimensional measurement across trimesters of pregnancy

Technical Summary: Method-Dependent Nomograms for Ultrasonographic Measurement of the Fetal Thymus

Problem Statement
Ultrasonographic assessment of the fetal thymus is utilized as a marker for intrauterine stress, fetal growth restriction, and specific genetic conditions such as 22q11.2 microdeletion. However, clinical practice currently employs different measurement techniques—primarily the standardized "THY-box" acoustic window and direct two-dimensional (2D) measurement in the three-vessel view. These methods are sometimes used interchangeably in clinical settings, despite lacking direct comparative data regarding their agreement and interchangeability. Furthermore, the feasibility of these techniques varies significantly across pregnancy trimesters due to changes in fetal anatomy and acoustic conditions (e.g., thymus isoechogenicity relative to lung tissue in the first trimester). There is a lack of method-specific reference ranges (nomograms) that account for these technical differences and gestational age, creating a risk of misclassification when reference values from one technique are applied to measurements obtained via another.

Methodology
This single-center, cross-sectional study analyzed 147 fetal thymus measurements from 119 singleton pregnancies at the Polish Mother's Memorial Hospital Research Institute and Salve Medica Medical Center.

• Population: Singleton pregnancies with reliably established gestational age. Exclusion criteria included diagnosed congenital heart defects, genetic syndromes affecting thymic development, and high-risk chromosomal abnormalities. Only euploid fetuses with structurally normal hearts were included.
• Measurement Techniques:
  • THY-box: Measured within a standardized acoustic window of the upper mediastinum. Lateral borders were defined by internal thoracic arteries (using Color Doppler), the posterior border by the brachiocephalic artery, and the anterior border by the sternum.
  • 2D Measurement: Measured in the three-vessel view on grayscale imaging, capturing the transverse diameter between lateral borders, distinguished from lung tissue by echogenicity differences and respiratory mobility.
• Study Design: Measurements were categorized by trimester (First: <14 weeks; Second: 14–27 weeks; Third: ≥28 weeks). In a subgroup of 26 fetuses, both methods were applied sequentially during the same examination to assess agreement.
• Statistical Analysis: Agreement was evaluated using Bland-Altman analysis, Deming regression, and the intraclass correlation coefficient (ICC). Nomograms were constructed using second-order polynomial regression on independent observations (one measurement per pregnancy), with standard deviation modeled as a function of gestational age. Additional analyses assessed associations with nuchal translucency (NT), fetal weight, uterine artery pulsatility index, and other variables.

Key Results

• Feasibility:
  • First Trimester: THY-box was feasible in 100% of cases (22/22); 2D measurement was not feasible in any case due to the isoechogenicity of the thymus and lung tissue.
  • Second Trimester: THY-box was feasible in 96% of cases; 2D was feasible in 25% of cases, allowing for paired measurements in 19 fetuses.
  • Third Trimester: 2D measurement was feasible in 100% of cases (32/32); THY-box was feasible in only 19% (6/32) due to ossification and fetal positioning.
• Method Agreement:
  • In the paired subgroup, the 2D method yielded systematically larger values than the THY-box method.
  • Bias: The mean difference (2D minus THY-box) was 3.91 mm (SD 2.76).
  • Limits of Agreement: Ranged from −1.50 mm to 9.31 mm.
  • Correlation: A very strong linear correlation was observed (Pearson r = 0.931; p < 0.001). Deming regression indicated a slope close to unity (2D = 3.78 + 1.008 × THY-box), suggesting a constant additive offset rather than a proportional error.
  • ICC: The intraclass correlation coefficient was 0.82.
• Nomograms:
  • Both methods showed a strong relationship between thymus dimension and gestational age.
  • The THY-box nomogram demonstrated a smaller residual scatter compared to the 2D nomogram, suggesting higher repeatability in the window where both are available.
  • Method-specific centiles (5th, 50th, 95th) were generated for gestational ages ranging from 12 to 38 weeks.
• Associations:
  • Gestational-age-adjusted thymus dimension was positively associated with fetal weight percentile (rho = 0.404; p = 0.001).
  • No significant association was found between thymus dimension and nuchal translucency (rho = 0.113; p = 0.285), uterine artery pulsatility index, fetal sex, mode of conception, or recent infection.

Key Contributions

1. Quantification of Non-Interchangeability: The study establishes that THY-box and 2D techniques, while measuring the same structure, differ by a nearly constant offset of approximately 4 mm. Consequently, reference ranges developed for one method cannot be applied to measurements obtained with the other without risk of misdiagnosis.
2. Trimester-Dependent Feasibility Mapping: The paper defines a practical algorithm for technique selection based on gestational age: THY-box is the primary method for the first and early second trimesters, while 2D measurement becomes the standard in the third trimester.
3. Method-Specific Nomograms: The authors provide the first set of nomograms that are explicitly dependent on the measurement technique and gestational age, allowing for accurate assessment across the entire pregnancy timeline.
4. Clarification of NT Relationship: The study systematically examines and finds no association between nuchal translucency and thymus size in euploid fetuses with structurally normal hearts, suggesting that the developmental link between the heart and thymus manifests primarily in pathological states (e.g., 22q11.2 microdeletion) rather than physiological variability.

Significance and Claims
The authors claim that the primary significance of this work lies in diagnostic safety. The consistent ~4 mm offset between methods means that applying a reference range from one technique to a measurement from another can lead to false classifications of thymic hypoplasia. This has direct implications for the diagnosis of conditions such as fetal growth restriction, intra-amniotic infection, and 22q11.2 microdeletion.

The paper emphasizes that fetal biometric markers are technique-dependent and require method-specific reference values. The provided nomograms are intended to organize thymus assessment across successive pregnancy stages, reducing the risk of misclassification. The authors modestly note that while the study provides ready-to-use reference ranges, their implementation in clinical practice requires further validation of the repeatability of both techniques and confirmation of these offsets in populations with specific pathologies (e.g., growth restriction or infection). The study does not propose new applications but rather seeks to standardize existing practices to prevent diagnostic errors.