San Jose, CA. Today NE Scientific CEO Andrea Borsic has pitched at the Early Stage Challenge organized by NVIDIA, competing for a $100,000 prize. The prize was won by Artomatix, a company operating in the field of video games. Our presentation was well received, and it has been very exciting to compete with other companies at the cutting edge of GPU computing.

The Early Stage Challenge is a business pitch/competition that takes place during the NVIDIA GPU Technology Conference (GTC), where a group of 12 selected companies, which have received less than $ 1 million in funding, compete for a $100,000 prize offered by NVIDIA. Each company has 4 minutes for presenting their business idea, based on the use of GPU technology, and 4 minutes to respond to questions from a panel of judges.

This year the competition was moderated by Scott Budman Business & Technology Reporter of NBC and the panel of judges was formed by:

  • Jeff Herbst Vice President of Business Development, NVIDIA
  • Savitha Srinivasan Partner, IBM’s Venture Capital Group, IBM
  • Saeed Amidi Founder & CEO Plug and Play Tech Center
  • Rob Enderle President and Principal Enderle Group

NE Scientific Presentation Summary

NE Scientific presented a computer platform for guidance in Radio Frequency Ablation of cancer which is being developed by the company, and for which the company has received funding from the National Cancer Institute.

Radio Frequency Ablation (RFA) is a surgical procedure where Radio Frequency (RF) energy is applied to tissues for killing tumoral masses. RF is applied by inserting a needle percutaneously into the volume of the tumor. This needle is electrically connected to an RF power generator which provides energy that will heat up tissues to more than 100 degrees Celsius, killing a volume of tissues around the electrode.

Needles available on the market can consist simply in a straight metallic shaft, or they might be able to deploy several metallic filaments into the tissues.

The computer simulation below shows the deployment of an RFA needle for treatment of a liver tumor. The simulated electrode deploys an umbrella of filaments into the tissues for encompassing a larger volume.


The electrode and the filaments are used to apply radio frequency energy to the tissues. Typically 100 to 150 Watts of power are applied for 10 minutes, raising the temperatures of the tissues to more than 100 degrees Celsius. The purposes of this intervention is to kill a ball of tissues that completely encompasses the tumoral mass (and a 1cm margin all around it, for safety) so that the malignant tissues are destroyed. The body will later eliminate on its own the resulting dead tissues.

The computer simulation below, driven by an RFA Physics Library developed by NE scientific, and running on GPUs, shows a successful ablation. The orange volume represents the volume of tissues being killed, and the green volume represents the tumoral mass. The tumoral mass is completely encompassed by the volume of killed tissues, and therefore the malignancy is completely destroyed.


While RF Ablation (RFA) has been successfully utilized in the past 15 to 20 years, there are limitations that NE Scientific is addressing. RFA is generally preferable to chemotherapy, as it does not present the side effects of drug-based treatments. RFA would potentially be preferable to open surgery, as less invasive, however open surgery however is still the preferred treatment option as by having direct view of the tissues it is easier to guarantee the full resection of the malignancy. Generally the 5-years survival rate with open resection surgery is higher than with RFA, as there are smaller risks of relapse due to non-complete destruction of the tumoral mass.

NE Scientific is addressing this issue by developing a computerized platform for guidance in RFA which will display on screen to surgeons exactly which tissues have been killed and which not. Physicians will be able therefore to immediately asses which further tissues need to be treated, and to consistently achieve full narcotization of the malignant tissues and of a safety margin around them.

We are addressing also a second problem, which has been limiting physicians so far. In RFA vessels in the proximity of the tumor remove heat from the ablation site (as the temperature of blood is 37 deg C versus the 100 deg C temperature of the heated tissues). This will generally result in a smaller ablation volume and in a altered geometry for the ablation. This results often in incomplete ablations, where untreated malignant tissues are left behind.

The computer simulation below, driven by the RFA  Physics Library developed by NE Scientific, shows how a vessel (a portion of a vessel is considered here, represented in violet color), is able to alter and shrink the ablation volume (orange) in such a way that part of the tumor (green) is not treated (a small green bulge is seen sticking out from the orange surface in proximity of the vessel at the end of the ablation).


NE Scientific is developing a software / computer system that will identify vessels from pre-operative CT images, and use this information to predict exactly the boundary of the killed tissues. Physicians will therefore be able to see on screen, as in the simulations above, the effect of a vessel, and to take actions to guarantee that the tumoral mass is fully treated.

The physics behind the simulations above are particularly complex, and they require many small time-steps to follow the temporal progression of the boundary of killed tissues. Previous works using CPUs for computing report simulations times for a 10 minutes ablation to be in excess of three hours. NE Scientific has developed a fast RFA Physics Library, accelerated on GPUs, which is able to run these simulations in 5 minutes, a speed up of 36 times, and two times faster than real-time speed. We have also a slight simplified model that runs in 15 seconds, and which can be used for super rapid assessment. Use of GPU technology, and the in-house developments of NE Scientific, enable therefore, for the first time, real-time operation. This will allow adoption of these tools in the clinical environment, and use in the operating room, for real-time evaluation of effectiveness of the treatment,  a step which has not been possible before.