metricas
covid
Revista Española de Medicina Nuclear e Imagen Molecular (English Edition) Evaluation of the diagnostic efficacy of [18F]FDG PET radiomic analysis for tumo...
Journal Information
Share
Download PDF
More article options
Original Article
Uncorrected Proof. Available online 3 December 2025
Evaluation of the diagnostic efficacy of [18F]FDG PET radiomic analysis for tumor differentiation in patients with non-small cell lung cancer
Evaluación de la eficacia diagnóstica del análisis radiómico de la PET con [18F]FDG para la diferenciación tumoral en pacientes con cáncer no microcítico de pulmón
Xavier Boulvard Cholleta,b,
Corresponding author
xleboulvard@riojasalud.es

Corresponding author.
, María P. Garrastachu Zumarána, Leonardo G. Romero Roblesa, Albert Santapau Traveriaa, M. Clara Albornoz Almadaa, Rafael Ramírez Lasantaa, José G. Pichelc, Roberto C. Delgado Boltona,b,d
a Servicio de Medicina Nuclear, Hospital Universitario San Pedro, Centro de Investigación Biomédica de La Rioja (CIBIR), 26006 Logroño, Spain
b Unidad de Investigación Biomédica, Centro de Investigación Biomédica de La Rioja (CIBIR), Fundación Rioja Salud, 26006 Logroño, Spain
c Unidad de Cáncer de Pulmón y Enfermedades Respiratorias (CIBIR), Fundación Rioja Salud, 26006 Logroño, Spain
d Servicio Cántabro de Salud, Spain
Article information
Abstract
Full Text
Bibliography
Download PDF
Statistics
Figures (2)
Tables (3)
Table 1. Distribution of patients by grade of differentiation.
Tables
Table 2. Statistical analysis of the variables comparing well-differentiated tumors with moderately and poorly differentiated tumors.
Tables
Table 3. Indicators of efficacy.
Tables
Show moreShow less
Abstract
Purpose

The aim of this study was to evaluate the utility of radiomics based on [18F]FDG PET/CT imaging for the prediction of histological differentiation grade in patients with non-small cell lung cancer (NSCLC) of either adenocarcinoma or squamous cell carcinoma subtype.

Materials and methods

A single-center retrospective observational study was conducted, including 150 patients with histologically confirmed NSCLC who underwent [18F]FDG PET/CT imaging prior to complete surgical resection of the tumor. Patients were excluded if they showed no significant FDG uptake, lacked differentiation data, or had insufficient tumor volume. Image segmentation and feature extraction were performed using LIFEx software, obtaining both textural and morphological features. Predictive variables were selected using classical statistical techniques and LASSO regression, and model performance was evaluated using ROC curve analysis.

Results

Among the 150 patients included, 25 had well-differentiated tumors, 88 moderately differentiated, and 37 had poorly differentiated. Comparison between well-differentiated tumors and those with moderate or poor differentiation revealed statistically significant differences in several radiomic features. The LASSO model identified seven variables with high discriminative power. The model achieved a sensitivity of 76%, specificity of 99.2%, positive predictive value of 95%, negative predictive value of 95.4%, and an overall accuracy of 95.3%.

Conclusions

Radiomic features extracted from [18F]FDG PET/CT images can predict tumor differentiation grade in NSCLC with high specificity. This non-invasive approach may serve as a valuable adjunct to histopathological evaluation, potentially aiding clinical decision-making.

Keywords:
Non-small cell lung cancer
[18F]FDG
PET-CT
Radiomic
Differentiation
Resumen
Objetivo

Evaluar la utilidad de la radiómica usando imágenes PET-TC con [18F]FDG para predecir de forma no invasiva el grado de diferenciación histológica en pacientes con cáncer de pulmón no microcítico (CPNM) de tipo adenocarcinoma o carcinoma escamoso.

Material y métodos

Se realizó un estudio observacional retrospectivo en un único centro, incluyendo 150 pacientes con CPNM confirmados histológicamente, que se sometieron a una PET-TC con [18F]FDG previa a la resección quirúrgica completa del tumor. Se excluyeron pacientes sin captación significativa de [18F]FDG, sin datos de diferenciación o con volumen tumoral insuficiente. Las imágenes fueron segmentadas y analizadas con el software LIFEx, extrayendo características radiómicas texturales y morfológicas. Se emplearon métodos estadísticos clásicos y regresión LASSO para la selección de variables predictoras, evaluando el rendimiento diagnóstico mediante curvas ROC.

Resultados

De los 150 pacientes analizados, 25 presentaban tumores bien diferenciados, 88 moderadamente diferenciados y 37 pobremente diferenciados. La comparación entre tumores bien diferenciados frente a los moderada o pobremente diferenciados mostró diferencias estadísticamente significativas en múltiples variables radiómicas. El modelo LASSO identificó siete variables predictivas con alto valor discriminativo. La curva ROC del modelo alcanzó una sensibilidad del 76%, especificidad del 99,2%, valor predictivo positivo del 95%, valor predictivo negativo del 95,4% y una exactitud global del 95,3%.

Conclusiones

Las características radiómicas obtenidas de imágenes [18F]FDG PET-TC permiten predecir con alta especificidad el grado de diferenciación tumoral en el CPNM, ofreciendo una alternativa no invasiva y potencialmente útil para complementar la evaluación histopatológica en la toma de decisiones clínicas.

Palabras clave:
Cáncer de pulmón no microcítico
[18F]FDG
PET-CT
Radiómica
Diferenciación

Article

These are the options to access the full texts of the publication Revista Española de Medicina Nuclear e Imagen Molecular (English Edition)
Member
If you are a member of the Spanish Society of Nuclear Medicine and Molecular Imaging (SEMNIM), you can access the full text of the contents of the Revista Española de Medicina Nuclear e Imagen Molecular through the journal links published on the SEMNIM website (link: http://socios.semnim.es/?ir-a=remnim), after logging in as a member.

If you experience access problems, you can contact the SEMNIM Technical Secretariat by email at secretaria.tecnica@semnim.es or by phone at +34 619 594 780.
Subscriber
Subscriber

If you already have your login data, please click here .

If you have forgotten your password you can you can recover it by clicking here and selecting the option “I have forgotten my password”
Subscribe
Subscribe to

Revista Española de Medicina Nuclear e Imagen Molecular (English Edition)

Purchase
Purchase article

Purchasing article the PDF version will be downloaded

Purchase now
Contact
Phone for subscriptions and reporting of errors
From Monday to Friday from 9 a.m. to 6 p.m. (GMT + 1) except for the months of July and August which will be from 9 a.m. to 3 p.m.
Calls from Spain
932 415 960
Calls from outside Spain
+34 932 415 960
E-mail
Article options
Tools