QuantaCell, IBM & CHU Jean Perrin team wins the JFR Challenge 2019 of the French Radiology Society

nodule detection cancer quantacell JRF SFR

QuantaCell is proud to be part of the lung cancer nodule detection challenge with Dr Julien Brehant from University Hospital Jean Perrin (Clermont-Ferrand) and IBM Montpellier (Cognitive Systems),  .

The objective of the challenge was to “distinguish 3D scanner exams with lung cancer nodules greater than 100mm3”

The risk of malignancy of a pulmonary nodule rises with the increase in its diameter. Screening studies using semi-automated volume measurements have shown higher accuracy and reproducibility compared to diameter measurements, and it has been shown that small nodules (those with a volume <100 mm3 or diameter <5 mm) are not predictive for lung cancer. For this question, teams had different tasks: to recognize pulmonary nodule on 3D CT chest scanners, to segment them, to estimate their volume and to classify them into probable benign (volume < 100mm3) or probably malignant nodules (volume ≥ 100 mm3).   
3D retina unet quantacell ibm
3D Retina U-NET

This challenge was a difficult detection problem because of the respiratory and circulatory systems network, leading to potential confusion due to high contrast images.

Our approach was to implement a pre-processing function that handles original images and annotations, isolating the lungs from the rest of the body using Hounsfield Units (HU) using a 3D U-NET segmentation algorithm and .postprocessing This detection allowed us to automatically reject artifacts detected outside the lungs.

This step was followed by a nodule detection model, based on the « Retina U-Net 3D » architecture  https://www.nature.com/articles/s41591-019-0447-x , that accepted the preprocessed image input and predict the detected nodules by generating the coordinates of a rectangular parallelepipedal box, but also intensity attributes and segmentation results.

Finally, we used a classification model (SVM support vector machine after feature extraction), to classify the patient as pathological having at least one nodule greater than 100 mm3 or normal.

Thanks to IBM team we used 3 IBM Power AC922 servers with Nvidia V100 32 Go  graphic cards (GPU – Graphics Processing Unit with respectively 4, 4 et 6, for a total of 14 GPUs)

https://www.ibm.com/us-en/marketplace/power-systems-ac922 .

That processing power together with the capacity to handle large models allowed to process successfully the test images in the required time with the best result of the ten teams in the challenge.

data challenge Deep learning cancer detection Quantacll ibm
The trophy. Thanks deep learning !

That processing power together with the capacity to handle large models allowed to process successfully the test images in the required time with the best result of the ten teams in the challenge.

The method was publish into the journal “Diagnostic and Interventional Imaging” in  2020 nov (DOI: 10.1016/j.diii.2020.10.004)

https://www.sciencedirect.com/science/article/pii/S2211568420302497

https://www.researchgate.net/publication/345922043_Artificial_intelligence_solution_to_classify_pulmonary_nodules_on_CT

Winning team Julien Brehant Thibault Besson Victor Racine
The winning team !

QUANTACELL