The ESTRO 38 meeting, held earlier this week in Milan, saw over 100 companies exhibiting at the trade show. Physics World spoke to some of the newest vendors, in the dedicated “Start-Up Corner”.
Beads of glass track delivered dose
TRUEinvivo is developing a high-performance thermoluminescent detector array based on small, biocompatible silica beads. The company’s DOSEmapper system is designed to track the dose delivered during radiation treatments. “We are bringing an in vivo dosimetry system to the market that can be placed inside the body and measure radiation dose at high resolution,” explains Shakar Jafari, TRUEinvivo’s founder and CTO.
The microsilica beads can be arranged into various configurations. For prostate cancer treatments, for example, the beads can be strung onto a thread that fits into a standard catheter. The catheter can be placed into the patient’s urinary tract where the beads measure dose to the prostate and bladder during treatment — at up to 1 mm resolution. Alternatively, the beads can be assembled on a mesh around a balloon to measure dose in cavities. Jafari notes that DOSEmapper is suitable for use with about 70% of cancer types.
After treatment, the beads are placed into an automated thermo-luminescent reader that heats them one at a time and measures the emitted light signal from each individual bead. Read-out of 100 beads typically takes about 15 minutes, and the beads exhibit a linear response from mGy to kGy levels. The measurements can then be compared with the patient’s treatment plan, enabling plan adaptation if necessary.
Artificial intelligence addresses segmentation woes
Mahmudul Hasan, CEO of Helsinki-based MVision, founded the company to solve a problem — namely, the complex, time-consuming nature of contouring images for radiotherapy planning. “Image segmentation is highly challenging and can lead to delays in treating patients,” he explains. “It is time consuming because the process is mostly manual or semi-automatic with atlas-based systems, and both need extensive manual fixing.”
Hasan describes how in a previous role, he was attending treatment planning training course and was challenged to contour a prostate on a scan. “I was completely lost, so I asked an oncologist for help, and found that I could do it by following instructions,” he says. This led him to surmise that artificial intelligence (AI) could be used to automate such tasks, saving time and adding consistency to the radiotherapy planning process.
So in 2017, Hasan teamed up with another software engineer, a medical physicist and an AI research scientist to create MVision. “We developed our first product in 11 months,” he says. The company’s AI-driven contouring tool takes a CT scan and produces a 3D contoured model in just two or three minutes. “It is now CE marked and deployed in two Finnish hospitals,” says Hasan, noting that in future developments, he aims to reduce the tool’s processing time to just a few seconds. “Deep learning-based algorithms trained with real clinicians’ experience perform as well as a clinical expert, which saves time and brings consistency into the workflow.”
Radiotherapy system targets breast cancer
EhmetDx‘s MammoKnife is the world’s first self-shielded radiotherapy system dedicated to treating breast cancer. “It is entirely self-shielded, so you don’t need a radiotherapy bunker, you can install it in a health clinic or even a mobile trailer,” explains CTO and co-founder Neal Clinthorne. “We aim to give radiation oncologists a tool that does not limit their creativity and that can provide better cancer care around the world.”
The MammoKnife, which incorporates a 6 MV linac and a multileaf collimator, treats patients in the prone position, thus minimizing heart and lung toxicity. Offering full 360° access to the breast, it can irradiate from more angles than possible with conventional systems, improving cosmetic outcome.
Targeted at treatment of early-stage breast cancer, MammoKnife can deliver accelerated whole and partial breast irradiation, intensity-modulated treatments, multi-target techniques and radiosurgery. Clinthorne points out that the path to CE and FDA regulatory clearance is simplified as the system is “substantially equivalent” to existing devices. He predicts that MammoKnife will be released for sale in 24–30 months.
User-independent QA device offers two-in-one testing
MD Arge has created a next-generation linac quality assurance system that performs both geometric and dosimetric tests using a single device. The UFC (user free control) device measures the light field and the radiation field together, with an adjustable resolution of 1 cm to 1.25 µm, enabling comparison of the radiation and mechanical isocentres. “One device can do most of required tests together, and it is user independent,” says co-founder Deniz Çelik.
Çelik explains that while other quality assurance systems require careful positioning on the treatment couch, UFC uses an inclinometer on the measurement array to align its angle and detect the measurement point itself. “You simply put the device on the couch and it finds the isocentre and the correct angle automatically,” he says.
Dose measurement is performed using a high-resolution CsI(Tl) scintillator photodiode array, while a silicon photodiode array measures the light field. The idea arose from Çelik’s PhD thesis (at Istanbul University), and he established MD Arge in 2017 to commercialize the device.