David Molina

Influence of gray level and space discretization on brain tumor heterogeneity measures obtained from magnetic resonance images.

PURPOSE Tumor heterogeneity in medical imaging is a current research trend due to its potential relationship with tumor malignancy. The aim of this study is to analyze the effect of dynamic range and matrix size changes on the results of different heterogeneity measures. MATERIALS AND METHODS Four patients harboring three glioblastomas and one metastasis were considered. Sixteen textural heterogeneity measures were computed for each patient, with a configuration including co-occurrence matrices (CM) features (local heterogeneity) and run-length matrices (RLM) features (regional heterogeneity). The coefficient of variation measured agreement between the textural measures in two types of experiments: (i) fixing the matrix size and changing the dynamic range and (ii) fixing the dynamic range and changing the matrix size. RESULTS None of the measures considered were robust under dynamic range changes. The CM Entropy and the RLM high gray-level run emphasis (HGRE) were the outstanding textural features due to their robustness under matrix size changes. Also, the RLM low gray-level run emphasis (LGRE) provided robust results when the dynamic range considered was sufficiently high (more than 8 levels). All of the remaining textural features were not robust. CONCLUSION Tumor texture studies based on images with different characteristics (e.g. multi-center studies) should first fix the dynamic range to be considered. For studies involving images of different resolutions either (i) only robust measures should be used (in our study CM entropy, RLM HGRE and/or RLM LGRE) or (ii) images should be resampled to match those of the lowest resolution before computing the textural features.

Tumour heterogeneity in glioblastoma assessed by MRI texture analysis: a potential marker of survival

OBJECTIVE: The main objective of this retrospective work was the study of three-dimensional (3D) heterogeneity measures of post-contrast pre-operative MR images acquired with T1 weighted sequences of patients with glioblastoma (GBM) as predictors of clinical outcome. METHODS: 79 patients from 3 hospitals were included in the study. 16 3D textural heterogeneity measures were computed including run-length matrix (RLM) features (regional heterogeneity) and co-occurrence matrix (CM) features (local heterogeneity). The significance of the results was studied using Kaplan-Meier curves and Cox proportional hazards analysis. Correlation between the variables of the study was assessed using the Spearmans correlation coefficient. RESULTS: Kaplan-Meyer survival analysis showed that 4 of the 11 RLM features and 4 of the 5 CM features considered were robust predictors of survival. The median survival differences in the most significant cases were of over 6 months. CONCLUSION: Heterogeneity measures computed on the post-contrast pre-operative T1 weighted MR images of patients with GBM are predictors of survival. ADVANCES IN KNOWLEDGE: Texture analysis to assess tumour heterogeneity has been widely studied. However, most works develop a two-dimensional analysis, focusing only on one MRI slice to state tumour heterogeneity. The study of fully 3D heterogeneity textural features as predictors of clinical outcome is more robust and is not dependent on the selected slice of the tumour.

Using metaheuristic algorithms for parameter estimation in generalized Mallows models

HighlightsWe deal with the problem of parameter estimation in Generalized Mallows model (GMM).We deal with 22 real datasets, all of them but one created by the authors.We have designed two experiments varying the maximum evaluations allowed.Obtained results significantly improve the previous competing approaches. This paper deals with the problem of parameter estimation in the generalized Mallows model (GMM) by using both local and global search metaheuristic (MH) algorithms. The task we undertake is to learn parameters for defining the GMM from a dataset of complete rankings/permutations. Several approaches can be found in the literature, some of which are based on greedy search and branch and bound search. The greedy approach has the disadvantage of usually becoming trapped in local optima, while the branch and bound approach, basically A* search, usually comes down to approximate search because of memory requirements, losing in this way its guaranteed optimality. Here, we carry out a comparative study of several MH algorithms (iterated local search (ILS) methods, variable neighborhood search (VNS) methods, genetic algorithms (GAs) and estimation of distribution algorithms (EDAs)) and a tailored algorithm A* to address parameter estimation in GMMs. We use 22 real datasets of different complexity, all but one of which were created by the authors by preprocessing real raw data. We provide a complete analysis of the experiments in terms of accuracy, number of iterations and CPU time requirements.

Using extension sets to aggregate partial rankings in a flexible setting

This paper deals with the rank aggregation problem in a general setting; in particular, we approach the problem for any kind of ranking: complete or incomplete and with or without ties. The underlying idea behind our approach is to take into account the so-called extension set of a ranking, that is, the set of permutations that are compatible with the given ranking. In this way we aim to manage the uncertainty inherent to this problem, when not all the items are ranked and/or when some items are equally preferred. We develop two approaches: a general one based on this idea, and a constrained version in which the extension set is limited by not allowing non-ranked items to be placed in an existent bucket of tied items. We test our proposal by coupling it with two different algorithms. We formalize our approaches mathematically and also carry out an extensive experimental evaluation by using 22 datasets. The results show that the use of our extension sets-based approaches to compute the precedence matrix, clearly outperforms the standard way of computing the preference matrix by only using the information explicitly provided by the rankings in the sample.

Lack of robustness of textural measures obtained from 3D brain tumor MRIs impose a need for standardization

Purpose Textural measures have been widely explored as imaging biomarkers in cancer. However, their robustness under dynamic range and spatial resolution changes in brain 3D magnetic resonance images (MRI) has not been assessed. The aim of this work was to study potential variations of textural measures due to changes in MRI protocols. Materials and methods Twenty patients harboring glioblastoma with pretreatment 3D T1-weighted MRIs were included in the study. Four different spatial resolution combinations and three dynamic ranges were studied for each patient. Sixteen three-dimensional textural heterogeneity measures were computed for each patient and configuration including co-occurrence matrices (CM) features and run-length matrices (RLM) features. The coefficient of variation was used to assess the robustness of the measures in two series of experiments corresponding to (i) changing the dynamic range and (ii) changing the matrix size. Results No textural measures were robust under dynamic range changes. Entropy was the only textural feature robust under spatial resolution changes (coefficient of variation under 10% in all cases). Conclusion Textural measures of three-dimensional brain tumor images are not robust neither under dynamic range nor under matrix size changes. Standards should be harmonized to use textural features as imaging biomarkers in radiomic-based studies. The implications of this work go beyond the specific tumor type studied here and pose the need for standardization in textural feature calculation of oncological images.

Geometrical Measures Obtained from Pretreatment Postcontrast T1 Weighted MRIs Predict Survival Benefits from Bevacizumab in Glioblastoma Patients

Background Antiangiogenic therapies for glioblastoma (GBM) such as bevacizumab (BVZ), have been unable to extend survival in large patient cohorts. However, a subset of patients having angiogenesis-dependent tumors might benefit from these therapies. Currently, there are no biomarkers allowing to discriminate responders from non-responders before the start of the therapy. Methods 40 patients from the randomized GENOM009 study complied the inclusion criteria (quality of images, clinical data available). Of those, 23 patients received first line temozolomide (TMZ) for eight weeks and then concomitant radiotherapy and TMZ. 17 patients received BVZ+TMZ for seven weeks and then added radiotherapy to the treatment. Clinical variables were collected, tumors segmented and several geometrical measures computed including: Contrast enhancing (CE), necrotic, and total volumes; equivalent spherical CE width; several geometric measures of the CE ‘rim’ geometry and a set of image texture measures. The significance of the results was studied using Kaplan-Meier and Cox proportional hazards analysis. Correlations were assessed using Spearman correlation coefficients. Results Kaplan-Meier and Cox proportional hazards analysis showed that total, CE and inner volume (p = 0.019, HR = 4.258) and geometric heterogeneity of the CE areas (p = 0.011, HR = 3.931) were significant parameters identifying response to BVZ. The group of patients with either regular CE areas (small geometric heterogeneity, median difference survival 15.88 months, p = 0.011) or those with small necrotic volume (median survival difference 14.50 months, p = 0.047) benefited substantially from BVZ. Conclusion Imaging biomarkers related to the irregularity of contrast enhancing areas and the necrotic volume were able to discriminate GBM patients with a substantial survival benefit from BVZ. A prospective study is needed to validate our results.

Consensus-based journal rankings: A complementary tool for bibliometric evaluation

Annual journal rankings are usually considered a tool for the evaluation of research and researchers. Although they are an objective resource for such evaluation, they also present drawbacks: (a) the uncertainty about the definite position of a target journal in the corresponding annual ranking when selecting a journal, and (b) in spite of the nonsignificant difference in score (for instance, impact factor) between consecutive journals in the ranking, the journals are strictly ranked and eventually placed in different terciles/quartiles, which may have a significant influence in the subsequent evaluation. In this article we present several proposals to obtain an aggregated consensus ranking as an alternative/complementary tool to standardize annual rankings. To illustrate the proposed methodology we use as a case study the Journal Citation Reports, and in particular the category of Computer Science: Artificial Intelligence (CS:AI). In the context of the consensus rankings obtained by the different methods, we discuss the convenience of using one or the other procedure according to the corresponding framework. In particular, our proposals allow us to obtain consensus rankings that avoid crisp frontiers between similarly ranked journals and consider the longitudinal/temporal evolution of the journals.

Textural features and SUV-based variables assessed by dual time point 18F-FDG PET/CT in locally advanced breast cancer

AimTo study the influence of dual time point 18F-FDG PET/CT in textural features and SUV-based variables and their relation among them.MethodsFifty-six patients with locally advanced breast cancer (LABC) were prospectively included. All of them underwent a standard 18F-FDG PET/CT (PET-1) and a delayed acquisition (PET-2). After segmentation, SUV variables (SUVmax, SUVmean, and SUVpeak), metabolic tumor volume (MTV), and total lesion glycolysis (TLG) were obtained. Eighteen three-dimensional (3D) textural measures were computed including: run-length matrices (RLM) features, co-occurrence matrices (CM) features, and energies. Differences between all PET-derived variables obtained in PET-1 and PET-2 were studied.ResultsSignificant differences were found between the SUV-based parameters and MTV obtained in the dual time point PET/CT, with higher values of SUV-based variables and lower MTV in the PET-2 with respect to the PET-1. In relation with the textural parameters obtained in dual time point acquisition, significant differences were found for the short run emphasis, low gray-level run emphasis, short run high gray-level emphasis, run percentage, long run emphasis, gray-level non-uniformity, homogeneity, and dissimilarity. Textural variables showed relations with MTV and TLG.ConclusionSignificant differences of textural features were found in dual time point 18F-FDG PET/CT. Thus, a dynamic behavior of metabolic characteristics should be expected, with higher heterogeneity in delayed PET acquisition compared with the standard PET. A greater heterogeneity was found in bigger tumors.

Computing the consensus permutation in Mallows distribution by using genetic algorithms

We propose the use of a genetic algorithm in order to solve the rank aggregation problem, which consists in, given a dataset of rankings (or permutations) of n objects, finding the ranking which best represents such dataset. Though different probabilistic models have been proposed to tackle this problem (see e.g. [12]), the so called Mallows model is the one that has more attentions [1]. Exact computation of the parameters of this model is an NP-hard problem [19], justifies the use of metaheuristic algorithms for its resolution. In particular, we propose a genetic algorithm for solving this problem and show that, in most cases (specially in the most complex ones) we get statistically significant better results than the ones obtained by the state of the art algorithms.

Papel predictivo y pronóstico de las variables volumétricas metabólicas obtenidas en la 18F-FDG PET/TC en el cáncer de mama con indicación de quimioterapia neoadyuvante

espanolObjetivo Investigar la utilidad de las variables metabolicas obtenidas en la 18F-FDG PET/TC en la prediccion de la respuesta a quimioterapia neoadyuvante (QNA) y el pronostico en el cancer de mama locamente avanzado (CMLA). Material y metodos Estudio prospectivo que incluye a 67 pacientes con CMLA, indicacion de QNA y 18F-FDG PET/TC basal. Se obtuvieron las variables SUV (SUVmax, SUVmedio y SUVpico) y volumetricas, tales como el volumen tumoral metabolico (VTM) y la glucolisis total lesional (GTL). Los tumores se agruparon en fenotipos moleculares y fueron clasificadas como respondedores y no respondedores tras la finalizacion de la QNA. Se obtuvo el estado libre de enfermedad (eLE), supervivencia libre de enfermedad (SLE) y supervivencia global (SG). Se realizo analisis univariante y multivariante para estudiar el potencial de todas las variables en la prediccion de la eLE, SLE y SG. Resultados Catorce pacientes se clasificaron como respondedoras. La media ± DE de la SLE y de la SG fue de 43 ± 15 y 46 ± 13 meses, respectivamente. El SUV y la GTL mostraron una relacion significativa (p Conclusion Las variables metabolicas obtenidas con la 18F-FDG PET/TC, de forma distinta a las variables de SUV, fueron buenos predictores tanto de la respuesta al tratamiento quimioterapico neoadyuvante y el pronostico. EnglishAim To investigate the usefulness of metabolic variables using 18F-FDG PET/CT in the prediction of neoadjuvant chemotherapy (NC) response and the prognosis in locally advanced breast cancer (LABC). Material and methods Prospective study including 67 patients with LABC, NC indication and a baseline 18F-FDG PET/CT. After breast tumor segmentation, SUV variables (SUVmax, SUVmean and SUVpeak) and volume-based variables, such as metabolic tumor volume (MTV) and total lesion glycolysis (TLG), were obtained. Tumors were grouped into molecular phenotypes, and classified as responders or non-responders after completion of NC. Disease-free status (DFs), disease-free survival (DFS), and overall survival (OS) were assessed. A univariate and multivariate analysis was performed to study the potential of all variables to predict DFs, DFS, and OS. Results Fourteen patients were classified as responders. Median ± SD of DFS and OS was 43 ± 15 and 46 ± 13 months, respectively. SUV and TLG showed a significant correlation (p Conclusion Volume-based metabolic variables obtained with 18F-FDG PET/CT, unlike SUV based variables, were good predictors of both neoadjuvant chemotherapy response and prognosis.AIM To investigate the usefulness of metabolic variables using 18F-FDG PET/CT in the prediction of neoadjuvant chemotherapy (NC) response and the prognosis in locally advanced breast cancer (LABC). MATERIAL AND METHODS Prospective study including 67 patients with LABC, NC indication and a baseline 18F-FDG PET/CT. After breast tumor segmentation, SUV variables (SUVmax, SUVmean and SUVpeak) and volume-based variables, such as metabolic tumor volume (MTV) and total lesion glycolysis (TLG), were obtained. Tumors were grouped into molecular phenotypes, and classified as responders or non-responders after completion of NC. Disease-free status (DFs), disease-free survival (DFS), and overall survival (OS) were assessed. A univariate and multivariate analysis was performed to study the potential of all variables to predict DFs, DFS, and OS. RESULTS Fourteen patients were classified as responders. Median±SD of DFS and OS was 43±15 and 46±13 months, respectively. SUV and TLG showed a significant correlation (p<0.005) with the histological response, with higher values in responders compared to non-responders. MTV and TLG showed a significant association with DFs (p=0.015 and p=0.038 respectively). Median, mean and SD of MTV and TLG for patients with DFs were: 8.90, 13.73, 15.10 and 33.78, and 90.54 and 144.64, respectively. Median, mean and SD of MTV and TLG for patients with non-DFs were: 16.72, 29.70 and 31.09 and 90.89, 210.98 and 382.80, respectively. No significant relationships were observed with SUV variables and DFs. Volume-based variables were significantly associated with OS and DFS, although in multivariate analysis only MTV was related to OS. No SUV variables showed an association with the prognosis. CONCLUSION Volume-based metabolic variables obtained with 18F-FDG PET/CT, unlike SUV based variables, were good predictors of both neoadjuvant chemotherapy response and prognosis.