EUVIMEDDescoperă EUVIMED
Open access · CC BY 4.0

Development of a lung nodule volumetry tool for quality assurance assessment of low‐dose lung cancer screening CT protocols

Yasmina Siddiki, Lucia Gallego Manzano, Chiara Pozzessere, David C Rotzinger, Damien Racine, Veronika Vitzthum

Medical Physics · 2026-12 · 53 (12)

DOI: 10.1002/mp.70776 ↗

At a glance

Journal article in Medical Physics, published online on 2026-11-19. Written by Yasmina Siddiki and 5 co-authors. 22 references. The full text is freely available (open access, CC BY 4.0).

RezumatDate bibliograficeCiteRelated studies

Abstract (original language)

Background

Low‐dose CT (LDCT) is the preferred modality for lung nodule detection and follow‐up, yet standardized quality‐control criteria are lacking. Volumetric accuracy is clinically important but strongly dependent on segmentation and reconstruction settings, particularly for small, low‐contrast subsolid nodules, and frameworks for optimizing LDCT protocols based on volumetric performance remain limited.

Purpose

To develop a lung nodule segmentation and volumetry tool that evaluates volumetric accuracy across CT reconstruction settings and to demonstrate how accuracy and repeatability metrics can support protocol comparison as a quality‐assurance assessment.

Methods

A custom semi‐automatic tool was developed in Python and evaluated using the QRM Thorax phantom on a GE Revolution Apex CT system. Fifteen artificial nodules were analyzed, combining five diameters (3–12 mm) and three attenuation levels (100, ‐630, ‐800 HU). Reconstruction parameters were systematically varied, including kernel (STANDARD, LUNG), slice thickness (0.625–2.5 mm), spacing (0.625, 1.25 mm), algorithm (DLIR‐H and ASiR‐V at 80%), and field of view (18, 36 cm). Performance was assessed using mean relative volume error (target within 20%), repeatability coefficient (RC), and size‐specific RMSE; the best‐performing configuration per nodule type served as reference. Reconstruction effects were evaluated using factorial ANOVA, Levene's test, Welch's t‐test, and bootstrap analysis with false‐discovery‐rate correction.

Results

Overall mean relative volume errors ranged from −7.1% to 6.7% for solid nodules and −14.7% to 5.8% for type 1 subsolid nodules mm; type 2 subsolid nodules showed larger deviations (−27.1% to −12.1%). Nodule size and reconstruction kernel were the dominant contributors to mean volume error, while slice thickness and spacing mainly influenced variability. Significant differences from the reference were found for four solid‐nodule settings, none for type 1, and one for type 2 subsolid settings, with degradation occurring mainly at increased slice thickness or reduced field of view. The 3 mm subsolid nodules showed consistently high variability across all settings.

Conclusions

The tool enabled quantitative comparison of CT reconstruction settings using volumetric accuracy and repeatability metrics. For clinically relevant sizes ( mm), volumetric performance remained within the predefined technical‐suitability range across most settings, supporting use of the tool for protocol evaluation and future dose‐optimization studies.

Cite

APA

Siddiki, Y., Manzano, L. G., Pozzessere, C., Rotzinger, D. C., Racine, D., & Vitzthum, V. (2026). Development of a lung nodule volumetry tool for quality assurance assessment of low‐dose lung cancer screening CT protocols. Medical Physics, 53(12). https://doi.org/10.1002/mp.70776

Vancouver

Siddiki Y, Manzano LG, Pozzessere C, Rotzinger DC, Racine D, Vitzthum V. Development of a lung nodule volumetry tool for quality assurance assessment of low‐dose lung cancer screening CT protocols. Medical Physics. 2026;53(12). doi:10.1002/mp.70776

BibTeX
@article{Siddiki2026,
  author = {Siddiki, Yasmina and Manzano, Lucia Gallego and Pozzessere, Chiara and Rotzinger, David C and Racine, Damien and Vitzthum, Veronika},
  title = {Development of a lung nodule volumetry tool for quality assurance assessment of low‐dose lung cancer screening CT protocols},
  journal = {Medical Physics},
  year = {2026},
  volume = {53},
  number = {12},
  doi = {10.1002/mp.70776}
}
Start for free →