Detecting cancer faster with optical innovation

You never know when an opportunity to innovate might present itself and, if seized, how much of a difference you can make in people’s lives.

Such an opportunity presented itself to us at Broadcom in the area of cancer detection. We have been innovating in optical technology for more than 50 years. Those innovations have largely focused on improving technologies like communications and sensors. This time, our innovation ultimately helped doctors and their patients — and led to other applications we hadn’t even considered.

It began with a little-known technology called “silicon photomultiplier,” or SiPM. SiPM is a next-generation optical sensor that allows people to “see” things that they could not detect without it. SiPM can sense, measure, and quantify radiation levels as low as even a single photon of light. Even more, its electrical response to a single photon is extremely fast. Those capabilities are fundamental to PET scanners, which help diagnose multiple forms of cancer. It can also be used for laser imaging technology, or LIDAR, for autonomous vehicles and 3D scanning and mapping applications, astrophysics, and related fields.

An Opportunity for Technology Disruption

We looked at the market to see where we could design and advance optical technology and possibly disrupt a market. We discovered that the market for silicon sensors used in PET scanners and other nuclear imaging purposes was mostly dominated by a single vendor. That was an opportunity we felt we could seize, and a market in which our expertise in developing custom chips would allow us to do something different.

We learned that the predominant optical technology used for nuclear medicine, the photomultiplier tube, was both old and expensive. At the same time, potential partners and customers were looking into a newer technology: SiPM. But there was a catch: they were finding it difficult and expensive to use it widely in new generation systems.

SiPM is a disruptive technology. Photomultiplier tubes are akin to the valve amplifiers used in old radios and TVs. SiPM, on the other hand, can be compared with semiconductor transistors, which replaced valve amplifiers, boosting the miniaturization and performance of all electronic systems and devices.

SiPM has many advantages over photomultiplier tubes: smaller size, lower power, lower cost, and longer lifetimes. With recent developments, their performance has become much better than the tubes in terms of sensitivity and speed of response. For all these reasons, SiPM is the only choice for modern PET systems.

That gave us an opportunity to be the disruptive technology that remakes this market.

Taking Advantage

Our first step was to create a team and manufacturing facility. We enlisted an expert with the best overview on SiPM state-of-the-art technology. We were adamant that we wanted the best SiPM possible; our consultant pointed us to a small research institute in Italy.

The institute was developing SiPMs on silicon wafers, which were adequate for research but not for volume production. We moved the manufacturing to our own foundry partners, changed the size of the chips, and cut costs while improving performance. After developing the packaging, we launched our first product with a global partner.

In the following years, we improved the most critical parameter of our SiPMs to provide an optimized sensor for PET machines. As a result, in 2023 our lead customer achieved the world record for the fastest “coincidence time resolution,” an important parameter for the resolution and quality of PET images.

Innovating New Market Opportunities

In creating this first application in PET scanners, we recognized that SiPM technology could be used in other applications. Today we are selling our SiPM-based products to several other global manufacturers while developing new SiPM applications, such as:

• Automotive LIDAR technology
• Mining operations where companies are using it to “see” into mountains for elements such as copper
• Flow cytometry or devices used to analyze characteristics of cells or particles—which can help diagnose certain blood cancers

Not Just Detecting but Processing Data

For state-of-the-art PET scanners, recording the radiation events as fast as possible is crucial. Timing and precision are paramount. The machines need to see and read the light emitted by the scintillator in tens-hundreds of picoseconds — that’s a trillionth of a second.

This feature is extremely challenging to improve because it depends on all the components of the detection module, including the processing of the electrical signals. Today, PET manufacturers use external discrete components to capture the information, but that’s both cumbersome and expensive. To overcome this limitation, we created a new kind of chip: a “read-out” ASIC. We’ve optimized both the ASIC and our SiPM technology for applications that need to “see” into low-light situations.

Now, customers can generate the electrical signal, amplify it, convert it to digital information through our SiPM modules and read-out ASICs. No data loss and very fast processing, which is crucial for applications such as cancer detection with state-of-the-art PET scanners.

It took us eight years to develop and optimize our SiPM technology, disrupting the market. Now we are ready to continue innovating further to make a difference in fields like nuclear medicine.

To learn more about our SiPM technology, go here.

About Martin Weigert

Martin Weigert is the Vice President and General Manager of the Industrial Fiber Products Division. He is responsible for the design and development of fiber optics products targeting industrial, automotive and emerging markets.