Posts Tagged :

Oncology

Bianca Muresan

INTERVIEW | Quantification of body composition to detect malnutrition in oncological patients

Bianca Tabita Muresan, master in personalized and community nutrition by the University of Valencia and currently PhD student in nutrition and endocrinology, took time to answer some questions about the relation between malnutrition in oncology and quantitative imaging after her one week training period at QUIBIM.

Tell us about yourself and how did you come to know QUIBIM before your training period with us?

I am a PhD student at the Faculty of Medicine (University of Valencia) and my doctorate thesis is about the diagnosis of sarcopenia and measurement of body composition in cancer patients, using the CT scans previously done to the patient for the disease diagnose or the radiotherapy treatment planning. Specifically, I use the powerful information hidden in CT scans to detect malnutrition in those cancer patients with high risk of developing it, as for example happens in patients suffering from head and neck or lung cancer, or cancers that affects the digestive system.  After detecting malnourished cancer patients, I work together with the Endocrinology and Nutrition Department to translate nutrition goals for improving cancer treatment collateral effects (as for example fatigue, vomiting, diarrhea, dysphagia, among others), or to assist with nutritional supplements of feeding tubes to help keep up a patient’s strength during treatment.

Due to my Master practicum and some research projects I collaborate since four years ago with many investigators of La Fe Health Research Institute in Valencia. Moreover, my thesis director (Dr. Alegría Montoro Pastor), recommended me to contact QUIBIM and GIBI for a short internship during my university studies and I knew that could improve my performance at work, as well as help me to reach new skills about imaging biomarkers.

 

Your field of expertise is malnutrition in Oncology. The Quantitative Medical Imaging field has significantly increased in the last years, but in hospitals, we still use anthropometric mediations as for example body mass index (kg/m2) for evaluating body composition, so what do you think is the main reason for using medical imaging?

Anthropometric variables, as for example body mass index, are the most commonly employed measures for detecting nutritional status in epidemiology, because of their simplicity and easy data collection. The problem is that in clinical practice, these parameters have a significant inter and intra observer variability. Moreover, they don’t allow the detection of major body compartments (such as lean body mass or adipose tissue), the estimation of body composition or to description of body fat distribution, which have been proved to be highly correlated to different degrees of malnutrition and sarcopenic obesity. In those cases, medical imaging, as for example computed tomography (CT) and magnetic resonance imaging (MRI) would facilitate the quantification of body fat and muscle mass distribution. For this purpose, 3D image segmentation techniques are applied, which allow the differentiation of subcutaneous fat, visceral fat, muscles and fatty infiltration within the muscular tissue. As of today, most of the 3D image segmentation techniques require manual correction but new artificial intelligence (AI) algorithms based on Convolutional Neural Networks (CNN) are providing promising results in the field.

What means sarcopenia and sarcopenic obesity and which are the effects of these terms to oncological outcomes? How could you measure these terms with medical imaging?

Sarcopenia is defined as a loss of skeletal muscle mass and decrease muscle function with or without loss of body weight and body fat. This condition could also occur in patients who present overweight, coexisting both pathologies: sarcopenia and obesity. The prevalence of sarcopenia in cancer patients was found to be a bad prognostic factor for disease progression and survival, as well as a negative predictor of toxicity levels and treatment complications.

In clinical practice we use bio–impedance analysis (BIA) for the evaluation of lean body mass and total adipose tissue, and this could also be analyzed by dual x-ray absorptiometry (DEXA). In addition, CT and MRI have shown to be excellent tools in assessing muscle mass tissue and different fat areas inside the body (subcutaneous, visceral and intramuscular). In the last years, several studies have suggested different cut-off values for detecting low muscle mass and low muscle density.  With this information, we try to detect pathological processes as for example pre-sarcopenia (which means loss of muscle mass without loss of muscle strength), myosteatosis (which means fat within and around skeletal muscle) and visceral obesity. The evaluation of muscle function for the diagnosis of sarcopenia is completed by measuring handgrip strength.

Which indicators are you working with at present and which ones do you think are the best by incorporating quantitative imaging in the future to analyze body composition in hospitals?

At the moment, we measure skeletal muscle mass, intramuscular adipose tissue, visceral adipose tissue and subcutaneous adipose tissue in cancer patients before starting cancer treatment. On the other hand, we try to study full body composition before starting radio – chemotherapy. As being a nutritionist, this technique helps to improve my work using the best technology. Related to the quantification of these imaging biomarkers, we have already sent different communications to Spanish congresses and written scientific articles for my PhD thesis, which are now in review.  Moreover, our department also works trying to find out different anatomical locations with major loss of muscle mass before cancer treatment, as well as correlating the prevalence of sarcopenia with toxicity levels and quality of life after antineoplastic treatment.

 For the future, I believe it would be helpful to include bone measurements to identify patients with osteopenia and in those patients with MRI, determine fat concentration in the liver (steatosis) would also be important for detecting important nutritional problems.

Finally, tell us shortly how did you find the experience at QUIBIM and GIBI these days and how could this training period improve your work?

 First of all, I would like to thank all the teamwork of QUIBIM and GIBI for offering me the opportunity to learn about the most advanced technology. I have definitely improved my skills about discovering interesting quantitative imaging biomarkers as for example fat liver and iron concentration, which are very important for future nutrition researches. It was absolutely an enriching experience. Thank you! 

Thank you Bianca for your time!

QUIBIM se posiciona en el mercado americano analizando hábitats tumorales y enfermedades difusas hepáticas

Ha cerrado acuerdos con la compañía EnvoyAI y ha abierto oficina en Silicon Valley

QUIBIM sigue creciendo en el desarrollo de algoritmos de análisis de imágenes médicas basados en modelos biológicos y en inteligencia artificial. La compañía, que cerró el año 2017 con un crecimiento del 500% en sus métricas de negocio (número de análisis y facturación), ha avanzado con la expansión al mercado americano de sus algoritmos para cáncer y enfermedades difusas hepáticas. Entre los hitos más relevantes, además de la reciente apertura de oficina en Palo Alto (Silicon Valley, California), QUIBIM ha establecido una alianza con la empresa EnvoyAI, dedicada a la integración y distribución de algoritmos de inteligencia artificial para imágenes médicas. EnvoyAI fue adquirida el año 2016 por TeraRecon, empresa con un software de visualización radiológica que se encuentra actualmente instalado en el 85% de los hospitales de Estados Unidos.

Estos algoritmos de QUIBIM para el mercado americano, ahora mismo en proceso de certificación por la Food and Drug Administration (FDA), caracterizan tanto las diferentes subregiones o habitats internos de un tumor como analizan por biopsia virtual hepática las enfermedades difusas tipo esteatosis, sobrecarga de hierro, inflamación y fibrosis.

En el ámbito de la oncología, QUIBIM proporciona una metodología que extrae datos de textura, celularidad y proliferación vascular en los tumores, siendo aplicable en lesiones cerebrales, de mama, próstata, hígado y recto, entre los principales tumores sólidos. Esta metodología puede aplicarse a las imágenes de Resonancia Magnética y Tomografía Computarizada, proporcionando una información pronóstica sobre la evolución de los pacientes y su respuesta a los diferentes tratamientos.

En el ámbito de las enfermedades difusas hepáticas, QUIBIM proporciona un algoritmo para realizar una biopsia virtual hepática, sin dañar al paciente, a partir de las imágenes de Resonancia Magnética y extrayendo la proporción de grasa, concentración de hierro y proporciones de inflamación y fibrosis hepática. Este análisis es de especial relevancia para el estudio y seguimiento de la esteatohepatitis y la hemocromatosis, y en la evaluación de la historia natural y su modificación por el tratamiento en las hepatopatías crónicas y la cirrosis. Actualmente esta información se obtiene a partir de biopsias convencionales, lo que implica un riesgo para el paciente y un sesgo inherente al procedimiento dado que sólo se obtiene información del lugar de la punción, mientras que el algoritmo de QUIBIM permite obtener información de todo el hígado de manera segura.

En palabras de su CEO, el Dr. Ángel Alberich-Bayarri, resumiendo estos hitos “estamos en una fase muy intensa de la compañía y con un crecimiento significativo, y gracias al esfuerzo del equipo hemos podido acelerar varios hitos que teníamos previstos en fases más tardías, como por ejemplo la alianza con EnvoyAI, la solicitud a FDA y la apertura de oficina en EEUU. Los algoritmos que hemos seleccionado para este mercado aportan un valor diferencial que ya hemos podido verificar, recibiendo solicitudes desde algunos hospitales antes de iniciar acciones comerciales.”

QUIBIM es una empresa biotecnológica nacida en Valencia y está especializada en la extracción de información cuantitativa de las imágenes médicas radiológicas y de medicina nuclear, mediante técnicas originales y avanzadas de procesamiento computacional. Estos parámetros extraídos reciben el nombre de Biomarcadores de Imagen y aportan rasgos extraídos de las imágenes médicas, relacionadas con procesos biológicos normales, enfermedades o respuestas terapéuticas.

Además, el equipo ha desarrollado la plataforma QUIBIM Precision® de análisis de imágenes médicas en la nube, que puede instalarse en versiones privadas para hospitales y para compañías farmacéuticas que desarrollen ensayos clínicos. A partir de imágenes de rayos X, Ecografía, TAC, Resonancia Magnética o PET, QUIBIM es capaz de aplicar algoritmos avanzados de análisis con metodologías basadas en procesamiento por GPU (unidades de procesamiento gráfico), Machine Learning o Big Data. El software de QUIBIM permite aportar una mayor información en los diagnósticos y poder evaluar de forma temprana la respuesta a los tratamientos farmacológicos. La compañía fue seleccionada en 2017 para el programa SME Instrument Fase II de la Comisión Europea.

Login QUIBIM

Quibim Precision V2.0

It is a real pleasure for Quibim to announce the release of Precision 2.0. On this new version, we have focused on improving its performance, making it more powerful and user friendly. Keep reading to know how we did it…
read more

Visit of the group of Dr. Regina Beets-Tan and Dr. Erik Ranschaert to QUIBIM headquarters

The satisfaction of bringing together the best experts in the world at home

Dr. Luis Martí-Bonmatí (QUIBIM co-founder and director of advisory board) and myself had been locally organising the joint congress of the European Society of Oncological Imaging (ESOI) and the European Society of Medical Imaging Informatics (EuSoMII) for approximately 1 year, and the event finally took place past October from 6th to 8th at our home. The Venue was our hospital: La Fe Polytechnics and University Hospital, where we develop our professional activity and where QUIBIM headquarters are placed, since QUIBIM is a spin-off Company of La Fe Health Research Institute, the institutional arm of the hospital to perform research.

The topic of the congress was “Imaging Informatics in Oncology” and it was the result of joining both Societies fields like Oncology and Medical Imaging Informatics from ESOI and EUSOMII, respectively. The congress program combined workshops with plenary sessions. The workshops covered different concrete areas like Response to treatment, Tumor Boards, 3D printing, and Imaging Informatics for clinicians and computer scientists. The complete program can be found here.

I had the opportunity to present on our stepwise development of Imaging Biomarkers process and the associated bottlenecks for validation. Here you can find the slides of the presentation.

Our group had a really active participation, either from the Biomedical Imaging Research Group (GIBI230) from our Hospital and from QUIBIM.

The main scientific presentations were:

  1. ProstateChecker, a tool for the multi-variate analysis of Prostate Cancer from T2, diffusion and perfusion MR sequences, by David García-Juan, post-processing biomedical engineer at QUIBIM
  2. Cloud architecture for Imaging Biomarkers analysis, by Rafael Hernández, CTO at QUIBIM
  3. Spatial registration on PET-CT scans and quantitative structured report for treatment response evaluation on lymphoma patients, by Fabio García-Castro, post-processing biomedical engineer at QUIBIM
  4. Web application for PI-RADS 2.0 Structured Reporting, by Enrique Ruiz-Martínez, CTO at GIBI230 group
  5. Business analytic, metric, key performance indicator. Automating the monitorization of performance indicators in Radiology departments, by Enrique Ruiz-Martinez, CTO at GIBI230 group
  6. Platform for the integration of Imaging Biomarkers in Radiology Departments, by Enrique Ruiz-Martinez, CTO at GIBI230 group
  7. Integration of Redmine as a tool to manage Clinical Trials in Radiology, by Amadeo Ten-Esteve, Clinical Trial Manager at GIBI230 group.
  8. Artificial intelligence techniques applied to medical imaging. Deep learning applied to automated chest X-ray screening, by Belén Fos-Guarinos, internship biomedical engineering student at GIBI230 group.
  9. Automatic vertebrae localization in pathological spine CT using Decision Forests, by Ana Jiménez Pastor, internship telecommunications engineering student at QUIBIM.
  10. Automated segmentation of muscle using Neural Networks, by Sara Rocher, internship biomedical engineering student at GIBI230 group.
  11. Automatic classification of intensity-vs-time curves in breast DCE-MRI by K-means clustering and Dynamic Time Warping curve matching, by Fabio García-Castro, post-processing biomedical engineer at QUIBIM

In the meantime, the QUIBIM stand was boiling! With several experts interested…

QUIBIM CEO, Angel Alberich-Bayarri explaining QUIBIM advantages to interested attendees to EUSOMII - ESOI congress

QUIBIM CEO, Angel Alberich-Bayarri explaining QUIBIM advantages to interested attendees to EUSOMII – ESOI congress

Besides the opportunity to present the work of our group, we had some remarkable visits to our headquarters, like the visit of Dr. Regina Beets-Tan group together with Dr. Erik Ranschaert (see photo).

Visit of the group of Dr. Regina Beets-Tan and Dr. Erik Ranschaert to QUIBIM headquarters

Visit of the group of Dr. Regina Beets-Tan and Dr. Erik Ranschaert to QUIBIM headquarters

Industry also visited us, in this case it was Agfa Healthcare. The Global Senior Solution Manager, Mr. Chris Townend and the National Sales Manager, José Vicente Puig visited our headquarters and attended the demo of our solution.

Visit of José Vicente Puig (National Sales Manager at Agfa Healthcare) and Chris Townend (Global Senior Solution Manager at Agfa Healthcare)

Visit of José Vicente Puig (National Sales Manager at Agfa Healthcare) and Chris Townend (Global Senior Solution Manager at Agfa Healthcare)

 

The last day we were lucky to have Regina Beets-Tan and Sergey Mozorov, new presidents of ESOI and EUSOMII, respectively, at our headquarters.

 

Visit of Sergey Morozov and Regina Beets-Tan. Photo with QUIBIM founders

Visit of Sergey Morozov and Regina Beets-Tan. Photo with QUIBIM founders

 

The experience speaks by itself.

Summer School on Oncology Imaging Biomarkers

Oncology Imaging Biomarkers

Imaging biomarkers define objective quantitative characteristics extracted from medical images that are related to normal biological processes, clinical endpoints in diseases, or the response to treatment.

In order to develop an imaging biomarker, it is necessary to carry out a series of steps to validate its relation with the reality studied and to check its clinical and technical validity. As we published in 2012 (https://www.ncbi.nlm.nih.gov/pubmed/21733539) and later adopted by the European Society of Radiology (ESR) (https://www.ncbi.nlm.nih.gov/pubmed/23397519), these steps allow for an efficient way of adoption of imaging biomarkers either in clinical routine or clinical trials. This process includes defining tests for the concepts and mechanisms; obtaining standardized and optimized anatomic, functional, and molecular images; analyzing the data with computer models; displaying data appropriately; obtaining the appropriate statistic measures; and conducting tests on the principle, efficacy, and effectiveness. Although important advances are being developed nowadays in the field of image processing algorithms for the extraction of quantitative information from images, the integration of this information in clinical routine requires the existence of validated biomarkers, according to specific protocols for their analysis, in terms of the processing algorithms needed, the models applied and the way the results are generated.

All these topics will be addressed in the upcoming Summer School that we are organising from the 25th to 29th of next July in Valencia under the umbrella of the European Institute for Biomedical Imaging Research (EIBIR). In case you are interested please register to the course and you will enjoy an unforgettable week in Valencia.

Registrations at: http://www.eibir.org/scientific-activities/joint-initiatives/biomedical-image-analysis-platform/eibir-summer-school/2016-2/

 

EIBIR Summer School on Oncology Imaging Biomarkers Poster

EIBIR Summer School on Oncology Imaging Biomarkers