A serial image analysis architecture with positron emission tomography using machine learning combined for the detection of lung cancer.

INTRODUCTION AND OBJECTIVES: Lung cancer is the second type of cancer with the second highest incidence rate and the first with the highest mortality rate in the world. Machine learning through the analysis of imaging tests such as positron emission tomography/computed tomography (PET/CT) has become a fundamental tool for the early and accurate detection of cancer. The objective of this study was to propose an image analysis architecture (PET/CT) ordered in phases through the application of ensemble or combined machine learning methods for the early detection of lung cancer by analyzing PET/CT images. MATERIAL AND METHODS: A retrospective observational study was conducted utilizing a public dataset entitled "A large-scale CT and PET/CT dataset for lung cancer diagnosis." Various imaging modalities, including CT, PET, and fused PET/CT images, were employed. The architecture or framework of this study comprised the following phases: 1. Image loading or collection, 2. Image selection, 3. Image transformation, and 4. Balancing the frequency distribution of image classes. Predictive models for lung cancer detection using PET/CT images included: a) the Stacking model, which used Random Forest and Support Vector Machine (SVM) as base models and complemented them with a logistic regression model, and b) the Boosting model, which employed the Adaptive Boosting (AdaBoost) model for comparison with the Stacking model. Quality metrics used for evaluation included accuracy, precision, recall, and F1-score. RESULTS: This study showed a general performance of 94% with the Stacking method and a general performance of 77% with the Boosting method. CONCLUSIONS: The Stacking method proved to be a model with high performance and quality for lung cancer detection when analyzing PET/CT images.

Evaluación del valor pronóstico de los parámetros volumétricos metabólicos calculados con la 18F-FDG PET/TC y su valor añadido a las características moleculares en pacientes con linfoma B difuso de células grandes / Evaluation of the prognostic value of the metabolic volumetric parameters calculated with 18F-FDG PET/CT and its value added to the molecular characteristics in patients with diffuse large b-cell lymphoma

Introducción y objetivos: El volumen metabólico tumoral (VMT) y la glicólisis total de la lesión (TLG) son predictores pronósticos en los pacientes con linfoma B difuso de células grandes (LBDCG). El objetivo del presente estudio es evaluar el impacto pronóstico de los parámetros volumétricos basales calculados con la tomografía por emisión de positrones/tomografía computarizada con 18F-fluorodesoxiglucosa (18F-FDG PET/TC) y su valor agregado a las características moleculares en pacientes con LBDCG tipo no especificado (NOS). Metodología: Se trata de un estudio retrospectivo observacional, en el que se incluyeron 35 pacientes sometidos a un 18F-FDG PET/TC basal previo al tratamiento. Se realizó un análisis univariable de los parámetros volumétricos (VMT y TLG), estudio inmunohistoquímico y traslocaciones cromosómicas. El método para el cálculo de los parámetros volumétricos fue el umbral SUV 2,5. La comparación entre los modelos predictivos se seleccionó en función del valor de criterio de información de Akaike (AIC), bayesiano (BIC) y C de Harrell, después de realizar un modelo de regresión de riesgos proporcionales de Cox. Además, se realizó un análisis univariable de los parámetros volumétricos, según los datos del estudio inmunohistoquímico utilizando la prueba de Wilcoxon-Mann-Whitney. Resultados: Al realizar un análisis univariable se evidenció que el VMT y la TLG son predictores de la supervivencia libre de progresión (SLP) y de la supervivencia global (SG), con una alta capacidad de discriminación. El añadir el VMT y la TLG al estudio inmunohistoquímico y a la traslocación cromosómica proporcionó un mejor valor pronóstico a la SLP y SG en los pacientes diagnosticados con LBDCG tipo NOS. Así mismo, se evidenció que los valores de los parámetros volumétricos eran menores en los pacientes que presentaron un fenotipo células B centro germinal (GCB) frente a los pacientes con fenotipo células B activadas (ABC) que presentaron valores mayores (AU)

Introduction and objectives: Metabolic tumor volume (MTV) and total lesion glycolysis (TLG) are prognostic predictors in patients with diffuse large B-cell lymphoma (DLBCL). The objective of this study is to evaluate the prognostic impact of the baseline volumetric parameters calculated with positron emission tomography/computed tomography with 18F-fluorodeoxyglucose (18F-FDG PET/CT) and its added value to the molecular characteristics in patients with DLBCL not otherwise specified (NOS). Methodology: This is a retrospective observational study, which included 35 patients who underwent a baseline 18F-FDG PET/CT prior to treatment. A univariate analysis of the volumetric parameters (MTV and TLG), immunohistochemical study and chromosomal translocations were performed. The method for calculating the volumetric parameters was the SUV 2.5 threshold. The comparison between the predictive models was selected based on the information criterion value of Akaike (AIC), bayesian (BIC) and Harrell's C, after performing a Cox proportional hazards regression model. In addition, a univariate analysis of the volumetric parameters was performed according to the data of the immunohistochemical study using the Wilcoxon-Mann-Whitney test. Results: A univariate analysis revealed that VMT and TLG are predictors of progression-free survival (PFS) and overall survival (OS), with a high discrimination capacity. Adding VMT and TLG to the immunohistochemical study and chromosomal translocation provided a better prognostic value for PFS and OS in patients diagnosed with DLBCL-NOS. Likewise, it was evidenced that the values of the volumetric parameters were lower in patients who presented a germinal center B cell phenotype (GCB) compared to patients with an activated B cell phenotype (ABC) who presented higher values. Conclusion: MTV and TLG added to the immunohistochemical study and chromosomal translocation provided a better prognostic value for PFS and OS (AU)

Evaluation of the prognostic value of the metabolic volumetric parameters calculated with 18F-FDG PET/CT and its value added to the molecular characteristics in patients with diffuse large B-cell lymphoma.

INTRODUCTION AND OBJECTIVES: Metabolic tumor volume (MTV) and total lesion glycolysis (TLG) are prognostic predictors in patients with diffuse large B-cell lymphoma (DLBCL). The objective of this study is to evaluate the prognostic impact of the baseline volumetric parameters calculated with positron emission tomography/computed tomography with 18F-fluorodeoxyglucose (18F-FDG PET/CT) and its added value to the molecular characteristics in patients with DLBCL not otherwise specified (NOS). METHODOLOGY: This is a retrospective observational study, which included 35 patients who underwent a baseline 18F-FDG PET/CT prior to treatment. A univariate analysis of the volumetric parameters (MTV and TLG), immunohistochemical study and chromosomal translocations were performed. The method for calculating the volumetric parameters was the SUV 2.5 threshold. The comparison between the predictive models was selected based on the information criterion value of Akaike (AIC), bayesian (BIC) and Harrell's C, after performing a Cox proportional hazards regression model. In addition, a univariate analysis of the volumetric parameters was performed according to the data of the immunohistochemical study using the Wilcoxon-Mann-Whitney test. RESULTS: A univariate analysis revealed that VMT and TLG are predictors of progression-free survival (PFS) and overall survival (OS), with a high discrimination capacity. Adding VMT and TLG to the immunohistochemical study and chromosomal translocation provided a better prognostic value for PFS and OS in patients diagnosed with DLBCL-NOS. Likewise, it was evidenced that the values of the volumetric parameters were lower in patients who presented a germinal center B cell phenotype (GCB) compared to patients with an activated B cell phenotype (ABC) who presented higher values. CONCLUSION: MTV and TLG added to the immunohistochemical study and chromosomal translocation provided a better prognostic value for PFS and OS in patients diagnosed with DLBCL-NOS.

Evaluación del valor pronóstico de los diferentes métodos de cálculo del volumen metabólico tumoral con 18F-FDG PET/TC, en pacientes con linfoma B difuso de células grandes / Evaluation of the prognostic value of different methods of calculating the tumour metabolic volume with 18F-FDG PET/CT, in patients with diffuse large cell B-cell lymphoma

INTRODUCCIÓN Y OBJETIVOS: El volumen metabólico tumoral (VMT) es un indicador de pronóstico prometedor en el linfoma B difuso de células grandes (LBDCG). El objetivo del presente estudio es evaluar los diferentes métodos para el cálculo del VMT basal con la tomografía por emisión de positrones/tomografía computarizada con 18F-fluorodesoxiglucosa (18F-FDG PET/TC) en pacientes con LBDCG, relacionando cada uno de los volúmenes medidos con la supervivencia libre de progresión (SLP) y la supervivencia global (SG). METODOLOGÍA: Se trata de un estudio de cohortes retrospectivo analítico, en el que se incluyeron 34 pacientes sometidos a un 18F-FDG PET/TC basal previo al tratamiento. Comparamos tres umbrales SUV 2,5, SUV 40% del SUV máximo y SUV medio hepático (PERCIST), para el cálculo de los biomarcadores VMT y glucólisis total de la lesión (TLG) relacionándolos con la SLP y SG. El mejor modelo predictivo se seleccionó en función del valor de criterio de información de Akaike (AIC) después de realizar una regresión de riesgos proporcionales de Cox. RESULTADOS: Con relación a la SLP, muestran diferencias estadísticamente significativas: VMT 2,5, TLG 2,5, VMT 40%, TLG 40%, VMT y TLG calculados con el umbral PERCIST. Entre estos, el que tiene un AIC menor es VMT 2,5, por lo que se considera el mejor parámetro para predecir la SLP. Con respecto a la SG, muestra diferencias estadísticamente significativas: VMT 2,5, VMT y TLG calculados con el umbral PERCIST. Entre estos tres, el que tiene un AIC menor es VMT 2,5, por lo que se considera el mejor parámetro para predecir la SG. Además, un mayor valor de VMT y TLG, se asocia a peor SLP y SG. CONCLUSIÓN: El VMT calculado con el umbral SUV 2,5 parece ser el mejor parámetro para predecir la SLP y SG en los pacientes diagnosticados con LBDCG con el 18F-FDG PET/TC

INTRODUCTION AND OBJECTIVES: Metabolic tumor volume (MTV) is a promising indicator of prognosis in diffuse large B-cell lymphoma (DLBCL). The aim of the present study is to evaluate the different methods for the calculation of the basal metabolic tumor volume with 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) in the patients with DLBCL, relating each one of the volumes measured with progression-free survival (PFS) and overall survival (OS). METHODOLOGY: This is a retrospective analytical cohort study, in which 34 patients underwent to 18F-FDG PET/CT baseline prior to treatment. We compared three SUV thresholds 2.5, SUV 40% of the maximum SUV and SUV mean hepatic uptake (PERCIST) for the calculation of MTV and total lesion glycolysis (TLG) biomarkers, relating them to the PFS and OS. The best predictive model was selected based on the Akaike's information criterion (AIC) after performing a Cox proportional hazards regression. RESULTS: In relation to the PFS, they show statistically significant differences: MTV 2.5, TLG 2.5, MTV 40, TLG 40, MTV and TLG calculated with the PERCIST threshold. Among these, the one that has a lower AIC is MTV 2.5, so it is considered the best parameter to predict the PFS. With respect to OS, it shows statistically significant differences: MTV 2.5, VMT and TLG calculated with the PERCIST threshold. Among these three, the one with the lowest AIC is MTV 2.5, which is why it is considered the best parameter to predict OS. In addition, a higher value of MTV and total tumor glycolysis (TLG), is associated with worse PFS and OS. CONCLUSION: The MTV calculated with the threshold SUV 2.5 seems to be the best parameter to predict PFS and OS in patients diagnosed with DLBCL with 18F-FDG PET/CT

Evaluation of the prognostic value of different methods of calculating the tumour metabolic volume with 18F-FDG PET/CT, in patients with diffuse large cell B-cell lymphoma. / Evaluación del valor pronóstico de los diferentes métodos de cálculo del volumen metabólico tumoral con 18F-FDG PET/TC, en pacientes con linfoma B difuso de células grandes.

INTRODUCTION AND OBJECTIVES: Metabolic tumor volume (MTV) is a promising indicator of prognosis in diffuse large B-cell lymphoma (DLBCL). The aim of the present study is to evaluate the different methods for the calculation of the basal metabolic tumor volume with 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) in the patients with DLBCL, relating each one of the volumes measured with progression-free survival (PFS) and overall survival (OS). METHODOLOGY: This is a retrospective analytical cohort study, in which 34 patients underwent to 18F-FDG PET/CT baseline prior to treatment. We compared three SUV thresholds 2.5, SUV 40% of the maximum SUV and SUV mean hepatic uptake (PERCIST) for the calculation of MTV and total lesion glycolysis (TLG) biomarkers, relating them to the PFS and OS. The best predictive model was selected based on the Akaike's information criterion (AIC) after performing a Cox proportional hazards regression. RESULTS: In relation to the PFS, they show statistically significant differences: MTV 2.5, TLG 2.5, MTV 40, TLG 40, MTV and TLG calculated with the PERCIST threshold. Among these, the one that has a lower AIC is MTV 2.5, so it is considered the best parameter to predict the PFS. With respect to OS, it shows statistically significant differences: MTV 2.5, VMT and TLG calculated with the PERCIST threshold. Among these three, the one with the lowest AIC is MTV 2.5, which is why it is considered the best parameter to predict OS. In addition, a higher value of MTV and total tumor glycolysis (TLG), is associated with worse PFS and OS CONCLUSION: The MTV calculated with the threshold SUV 2.5 seems to be the best parameter to predict PFS and OS in patients diagnosed with DLBCL with 18F-FDG PET/CT.