Acoustic Surgery: When tumors become audible

Bayern Innovativ: Dr. Esmaeli, could you briefly explain to us what Surag Medical does and who your products are intended for?

Nazila Esmaeili: SURAG Medical stands for Surgical Audio Guidance. We focus on the development of vibroacoustic sensor technology. We started working on this topic as a research project, and based on this, we founded the company in 2021. The initial focus was on medical applications, on needle-based interventions, to be more precise. Our goal was to provide complementary information and feedback during these procedures. Today, we also focus on non-medical applications. With the current R-Kit, we can explore applications in different areas.

Bayern Innovativ: Which medical application or procedure is the focus of your technology, and at what point in the clinical process does it create the greatest added value?

Nazila Esmaeili: We started with needle-based procedures. The biggest advantage of our technology is that interactions occur in every surgical procedure, so we have access to a rich source of information with a high dynamic range.
For needle-based interventions, we have already demonstrated the clinical benefit of real-time feedback during blind punctures, for example spinal needle and Veress needle insertion. Recently, in a collaborative project, we have started to insert needles into tumor lesions. In this way, we can "listen" to tumors and generate information that enables us to differentiate between different types of tumors. This represents a major advance in diagnostics because we can provide a real-time diagnosis through this tumour characterization. This means there are no waiting times for a biopsy and everything is faster and safer for patients.
We have also investigated other medical applications, including laparoscopic and robotic surgery, and our current focus is on advancing the digitalization of surgery.

Bayern Innovativ: Can you explain to me in simple terms what exactly the sensor picks up from the vibrations?

Nazila Esmaeili: If you have a medical instrument such as a needle, for example, we attach the sensor to the proximal end, i.e. the part that is closer to the surgeon. The sensor is therefore completely minimally invasive. The sensor itself has no contact with the patient. With this surgical instrument, you begin to interact with the tissue. For example, the doctor takes the needle and begins to insert it into the body of the person to be treated in order to take a blood sample. Or it could be inserted into a tumor to take samples, or into the spine to inject medication or take samples such as cerebrospinal fluid. As the needle begins to interact with the tissue and penetrate layers to reach the desired site, there are micro-vibrations at the point of contact between the instrument and the tissue. These micro-vibrations propagate along the shaft of the instrument and do not require any active component to be embedded in the instrument. This means that it is natural information, we do not generate it. We simply attach our sensor to the proximal end and record these micro-vibrations. The sensor transmits the vibrations wirelessly to a processing unit in real time. In the current version of the R-Kit, an iPad with special software installed acts as the processing unit. There, the data is received and processed in real time to provide acoustic feedback, i.e. to bring the vibration into an audible range, and to store it for further analysis. This is our product, our R-Kit. Researchers, universities and research centers, can buy it to conduct research in relevant areas. That's why I said earlier that we don't just limit ourselves to the medical field. Depending on the procedure, auditory feedback can be presented in different forms. With interventions like the tumour example, it goes beyond just listening: we can characterize tissue-specific behaviours, classify them and correlate them with diagnoses.

Bayern Innovativ: What role does AI play in your products and how does it work?

Nazila Esmaeili: In the current R-Kit, AI is partially integrated into the software. It supports users in annotating their data and enables automatic recognition and classification of certain tasks - depending on the respective application in which the system is used. We are continuously developing our product in this area, particularly because digitalization is becoming increasingly important in surgery.
In medical applications, the role of AI lies mainly in how the collected data is processed and the feedback that can be provided to clinical users.

Bayern Innovativ: What kind of output does the user receive when using the tools in practice? Are there acoustic signals, warnings, text displays or other notifications?

Nazila Esmaeili: In the current version of the R-Kit, the feedback is audified. You can hear the sounds of micro-vibrations by amplifying them and bringing them into the audible range. But the type of feedback can vary depending on the procedure. With audification feedback, for example, when a needle is inserted through different layers, you can hear and thus recognize the transitions between the individual layers. Another form of feedback can be indicators. For example, instead of giving acoustic feedback during a needle puncture, we can use AI to convert the vibroacoustic data into a visual or acoustic indicator to confirm that the needle is moving from one tissue to the next tissue type. Or back to the application example in tumor diagnostics: in combination with AI, the vibroacoustic data allows conclusions to be drawn about the tissue properties of tumors and can provide diagnostic information about the tumor type, for example whether a tumor is benign or malignant.

So we have three levels of feedback:
1. Simple audification - you listen and interpret it.
2. Advanced indicators that help to perform and guide an intervention.
3. decision-based feedback that helps to determine a treatment strategy.

Bayern Innovativ: Can you give a simple example of the benefits?

Nazila Esmaeili: I'll give you an example from our work on tumor characterization. Here we have started to collect data from removed salivary gland tumors. The normal clinical procedure is as follows: Affected individuals go to the doctor with appropriate symptoms and the first step is an ultrasound examination. If a tumor is found, a percutaneous ultrasound-guided needle biopsy is performed. Even with ultrasound, it can be a challenge to see where the needle is going. We can improve the guidance so that a suitable position is achieved to take samples. If samples are taken, they must first be sent to pathology for diagnosis. This can take 7 to 10 days, in some cases even up to two weeks. The patient has to wait this long before any necessary surgery can be planned. Our technology can shorten this waiting time.
What we can improve as a result is

1. better guide the needle to reach a suitable site for sampling by combining vibroacoustics with ultrasound.
2. we can make a diagnosis by "listening" to and characterizing the tumor. And we can tell in real time and with a certain probability whether the tumor is benign or malignant.
3. we offer a cost-effective technology that improves the procedure described in several ways, firstly by saving time and secondly by enabling more precise navigation.

Bayern Innovativ: How do you test your tools and make sure they work?

Nazila Esmaeili: When we develop a new version of our systems, we first start with laboratory tests. We usually use ex-vivo animal tissue or synthetic tissue. It depends on the type of procedure we want to focus on. Later we expand to tests on animal cadavers, body donations or real patients. However, we always start with laboratory tests, as these provide us with criteria for improving the system specifically for the desired application. Our laboratory tests can be divided into two categories: synthetic data and ex-vivo animal data. Only in the next step do we carry out experiments on animal cadavers. Tests on real patients are always the last stage. This data is particularly important before a system is launched on the market.

Bayern Innovativ: Why are you focusing on vibroacoustics instead of other solutions such as camera-based imaging or other sensors?

Nazila Esmaeili: Vibroacoustic data is a rich source of information with high dynamic bandwidth and is easier to process and store compared to image or video data. In addition, the good thing about vibroacoustics is that it can complement existing technologies. Our aim is not to replace existing technologies. In laparoscopic procedures, for example, surgeons rely on video data from the camera placed in the patient's body to perform the operation. This means that neither hearing nor touch can be used due to the indirect access to the surgical field. Vibroacoustic sensor technology can remedy this situation. By attaching the sensor to the laparoscopic instruments, we can make the interactions between the instrument and tissue audible as acoustic feedback. Surgeons can hear what is happening during the procedure - for example when cutting or palpating tissue. With this feedback, they can also recognize the underlying structures, such as blood vessels, by hearing the pulsation of the blood vessels. This means that damage to critical structures can be avoided and operations can be performed more safely.

Bayern Innovativ: What challenges are you currently facing when using AI?

Nazila Esmaeili: In general, one of the biggest challenges is gaining access to high-quality and well-annotated data, especially in medical applications. With the R-Kit, we enable users to access vibroacoustic data for their specific applications. Access to well-annotated data is one of our main goals. The R-Kit software helps users to annotate their data with simple labeling tools and also provides automatic annotation and labeling of the data. The current version already works efficiently, but we are continuously improving it by providing even more precise labeling information and adding more user-friendly features.
Apart from that, AI systems in medical applications require strict validation and must meet regulatory requirements before they can be used in clinical practice. This is often a major hurdle for companies developing AI-based solutions. Since we provide three levels of feedback in our case, we can initially introduce solutions that are not based on AI. This allows us to collect high-quality and well-annotated clinical data. This helps us to develop more robust AI-based solutions for future clinical applications.

Bayern Innovativ: Where are you today and what is your next big milestone in the next 12 months?

Nazila Esmaeili: Our R-Kit is already available on the market and we are in contact with several universities and hospitals to make this system available to them. We have released the latest version 2025. The product structures data and stores it based on projects and experiments. Videos can also be synchronized. Researchers from various fields can buy it and use it depending on their area of application.
After a lot of development work over the last few years, we are pleased to have now started our study on vibroacoustic tumour monitoring to characterize tumour tissue. We collect data, listen to it and analyze it. We have a cooperation with the University Hospital Giessen and Marburg for this project. Our aim is to generate more evidence in both clinical and non-clinical settings. We are also planning projects with companies that are interested in vibroacoustic sensor technology and would like to use it for their own applications.

Bayern Innovativ: Thank you very much, Dr. Esmaeili, for the interview.

Surag Medical GmbH is a partner in Bayern Innovativ's Health Partner Network. Bayern Innovativ GmbH brings together players from science, industry and practice, thus creating spaces for exchange and knowledge transfer. On behalf of the Free State of Bavaria, it supports companies and start-ups in gaining orientation in the complex innovation ecosystem. In this way, promising approaches are sustainably transformed into marketable solutions and Bavaria is strengthened as an innovation region.