Tumor on a Chip Explained: The Future of Cancer Drug Testing
- Jörn Menninger
- Dec 11, 2025
- 28 min read
Updated: 6 days ago
What Is This About?
Tumor-on-a-chip technology replaces animal testing with human-relevant cancer models built using microfluidic engineering. This interview explores how organ-on-chip startups are accelerating personalized oncology by testing drug responses on patient-specific tumor environments in the lab.
Introduction
Organ-on-chip technology is replacing animal testing in cancer drug development by creating human-relevant tumor environments on miniaturized platforms. This episode explains how tumor-on-a-chip models accelerate personalized oncology research, enabling pharmaceutical companies to test drug candidates against realistic human tissue before clinical trials. The technology represents one of the most promising intersections of biotech and engineering emerging from European research institutions.
Tumor-on-a-chip technology creates human-relevant cancer microenvironments on miniaturized platforms, enabling drug testing without animal models. The technology accelerates personalized oncology by testing treatments against individual patient tumor profiles before clinical application. Current systems can replicate tumor heterogeneity, immune interactions, and drug metabolism with increasing biological accuracy. The approach addresses both the ethical concerns of animal testing and the scientific limitations of models that poorly predict human responses.
Discover how organ-on-chip cancer models replace animal tests and accelerate personalized oncology with human-relevant tumor environments.
Discover how organ-on-chip cancer models replace animal tests and accelerate personalized oncology with human-relevant tumor environments. Startuprad.io brings you independent coverage of the key developments shaping the startup and venture capital landscape across Germany, Austria, and Switzerland.
This founder interview is part of our ongoing coverage of Scaleup Founder Interviews from Germany, Austria, and Switzerland.
Management Summary
For decades, cancer drug development relied on 2D cell cultures and animal models — systems that fail to behave like human tumors and cost the industry billions in failed trials. In this episode of Startuprad.io, I spoke with Ghazaleh Madani, CEO and Co-Founder of CanChip, whose tumor-on-a-chip microfluidic platform is reshaping how oncology drugs are developed, validated, and personalized.
Her team builds human-relevant tumor environments using microfluidics, perfusion flow, endothelial and immune co-cultures, and real-time biosensing. These chips can model drug response in 48–72 hours, replacing weeks of animal testing and reducing clinical risk.
This is not a future vision. It’s happening now.
If you work in oncology R&D, biotech investing, translational research, or personalized medicine, this model changes your assumptions: faster data, fewer failed trials, reduced costs, and most importantly — better outcomes for patients.
Our Sponsor
Quick break for something every founder should hear. One leak on the dark web can mean account takeovers, impersonation, or a board-level crisis. That’s why we partnered with NordStellar — a business-grade threat-exposure platform from the team behind NordVPN. It gives you early signals before attackers escalate — with data-breach and dark-web monitoring, attack-surface discovery, and cybersquatting detection. You’ll spot exposed credentials, shadow IT, and fake domains fast.
Startuprad listeners get an exclusive 20% Black Friday discount — go to nordstellar.com/startupradio and use code blackfriday20 before December 10, 2025. Don’t wait until your data shows up for sale — visit nordstellar.com/startupradio, code blackfriday20.
Table of Contents
What Is a Tumor on a Chip?
Why Traditional Models Fail in Oncology
Inside the Microfluidic Tumor Environment
Personalized Oncology: Patient-Derived Tumor Models
Real-Time Biosensing and Drug Response Prediction
How Tumor Chips Reduce Animal Trials
Case Study: 72-Hour Drug Response from a Patient Tumor
Startup Execution in Biotech: Lessons from CanChip
Market Outlook: Organ-on-Chip by 2030
FAQs
What Is a Tumor on a Chip?
A tumor on a chip is a microfluidic cancer model that recreates human tumor biology on a transparent, engineered platform. Unlike traditional cell culture plates or animal trials, this system uses:
3D human tumor cells
Immune and endothelial co-cultures
Perfusion to simulate blood flow
Biosensors that measure drug response
As Ghazaleh explains, these chips replicate real tissue, not plastic approximations.
The result: a miniaturized human tumor environment where drugs behave as they would inside a patient — not a mouse.
Why Traditional Models Fail in Oncology
Every oncology researcher knows the story: a drug works beautifully in 2D cultures or mice, only to fail in human trials.
Why?
2D Cell Cultures Are Flat — Humans Aren’t
Cells grow on plastic. No perfusion. No immune interaction. No tumor structure.Ghazaleh:
“2D models behave nothing like human tumors.”
Animal Models Are Not Human Models
Physiology, metabolism, tumor mutation profiles — all different.
Billions are wasted each year predicting human outcomes based on non-human biology.
Drugs Fail Because Models Fail
When your preclinical data is wrong, your clinical phases collapse.
A tumor-on-chip model removes this uncertainty by directly simulating human tumor behavior.
Inside the Microfluidic Tumor Environment
Ghazaleh describes the engineering challenge: the chip must behave like tissue, not like plastic.
Their architecture includes:
3.1 Co-Cultures
Tumor cells
Endothelial cells
Immune cells
Each added for realistic tumor-immune and tumor-vascular interactions.
3.2 Perfusion Flow
A microfluidic system continuously delivers nutrients and drugs, mimicking blood flow.
3.3 A Zero-Absorption Chip Material
Many chips absorb drugs — corrupting data. The CanChip platform eliminates this.
3.4 Transparent, Real-Time Observation
Researchers can visualize how tumors react as drugs circulate through them.
This isn’t a simulation — it’s biology behaving as biology.
Personalized Oncology: Patient-Derived Tumor Chips
This is where the science becomes deeply human.
Using fresh patient tumor samples, the chip can model:
Drug sensitivity
Tumor resistance
Cell death patterns
Immune-tumor interaction
Instead of waiting weeks for results, oncologists can know:
“Which drug will work for this exact patient?”
within 48–72 hours.
Ghazaleh emphasizes:
“Every delay in preclinical development is real suffering in the real world.”
This is personalized oncology without guesswork.
Real-Time Biosensing and Drug Response Prediction
These chips integrate biosensors that detect:
Cell viability
Apoptosis markers
Metabolic shifts
Gene expression changes
Flow and perfusion metrics
This enables real-time feedback, not endpoint assays.
In one case:
“We monitored drug response in real time and had results in 48–72 hours.”
For biotech, that speed is not convenience — it is competitive advantage.
How Tumor Chips Reduce Animal Trials
The industry is shifting.
FDA now accepts non-animal models (2024)
Ghazaleh:
“Animal models do not replicate human response anymore.”
Advantages vs. animals:
More predictive data
Faster testing
Ethical compliance
Lower cost
Better reproducibility
One pharma partner cut animal trials by ~30% using tumor-on-chip models.
This isn’t theoretical reduction — it is operational change.
Case Study: A 72-Hour Drug Response Prediction
A collaborative partner provided patient-derived colorectal tumor cells.
On the chip:
Continuous perfusion
Real-time sensors
Gene expression tracking
And within three days:
“We had the primary drug response results.”
Compared to traditional methods:
Method | Time to Result | Predictiveness |
Animal Trials | Weeks–Months | Low |
2D Cultures | Days | Low |
Tumor-on-Chip | 48–72 Hours | High |
This is the inflection point where oncology moves from trial-and-error to precision.
Startup Execution in Biotech: Lessons from CanChip
Biotech founders often believe their product will “speak for itself.”
It won’t.
Ghazaleh learned:
“A founder must be loud. If no one knows you, the product has no value.”
Her tactical playbook:
1. Be Loud Early
Visibility → partnerships → data → credibility.
2. Launch Fast, Learn Faster
Perfection kills momentum. Pilot studies reveal real issues.
3. Co-Develop with Users
The product you want ≠ the product the market needs.
4. IP Strategy Early
CanChip filed their first patent in 2024.
5. Collaborate Relentlessly
Personalized oncology is not a solo act. Hospitals must engage.
Market Outlook: Organ-on-Chip by 2030
Ghazaleh frames it simply:
“Organ-on-chip will mature from innovation to infrastructure.”
By 2030:
Animal testing becomes obsolete in many oncology contexts
Pharma relies on organ-on-chip pipelines
Personalized therapy decisions are chip-based
AI integrates patient data + chip response for hybrid predictive models
CanChip positions itself as the leader in tumor-on-chip for precise oncology.
When this becomes standard, we’ll look back and say:
“This is where oncology finally became human-first.”
Animal models mislead drug development
Tumor-on-chip creates human-relevant tumor environments
Patient-derived chips deliver 72-hour drug response predictions
FDA acceptance accelerates commercial adoption
Personalized oncology becomes clinically actionable
Quote Box
“We want more collaborations from hospitals. Personalized medicine is not a one-man show.”— Ghazaleh Madani, CEO of CanChip
Market Lens (Value Block)
The organ-on-chip market grows at 30–40% CAGR, driven by:
Regulatory acceptance
AI integration
Pharma demand for predictive models
Ethical pressure to reduce animal tests
This is one of biotech’s few inevitabilities.
Pro Tip (Value Block)
If you're a founder in deep tech:Your breakthrough is only as powerful as your ability to communicate it.
Stat Spotlight
30% reduction in animal trials for some pharma partners using tumor-on-chip.
Relationship Map
Ghazaleh Madani → CEO and Co-Founder → CanChip
Ghazaleh Madani → CEO → CanChip
Jörn "Joe" Menninger → Host of → Startuprad.io
Automated Transcript
1 If you're a biotech founder, investor or pharma innovator 2 trying to accelerate drug screening beyond outdated 3 models, here's the challenge. Traditional tumor models 4 are slow, costly and unreliable. Ghazalay, 5 founder and CEO of Can Chips GmbH has built a 6 microfluid tumor on a chip platform from 7 the Potsdam Science park ecosystem backed by national 8 startup awards and a mission born from a personal oncology 9 journey. Today we'll break down how her tumor on a 10 Chip innovation can help you bring personalized cancer 11 therapy, animal free drug development and 12 predictive tumor microenvironment models into 13 real world impact so you can leap ahead in the competitive 14 biotech space. 15 Welcome to Startupradio, 16 your podcast and YouTube blog covering the German 17 startup scene with news, interviews and 18 live events. 19 Welcome to StartupRadio. Our guest today is Ghazali
20 Madani, co founder and CEO of Can Chips, a 21 Potsdam based biotech startup revolutionizing 22 cancer research with its tumor on a chip 23 microfluid platform. Actually, you'll explain to 24 us what that means soon, right? With Masters of Biochemistry and 25 Molecular Biology from the University of Potsdam and a 26 Bachelor's in Medical Laboratory Science from Isafan University, 27 Ghazali brings rigorous science, entrepreneurial vision and 28 personal motivation. Her mother's cancer journey led 29 directly to her founding mission. Since 2023, Can 30 Chips has developed human cell co cultured 3D 31 microfluid tumor models, achieved major awards 32 including Newcomer of the Year at the German Startup Awards 2025. 33 That's why you are here. Congratulations. And is 34 positioned at the cutting edge intersection of personalized 35 medicine and non animal high predictive 36 clinical testing. Today she joins us to unpack 37
how tumor on a chip microfluids, 38 angiogenesis modeling and biosensor 39 integration combine to create the next frontier in 40 oncology drug development and what that means for founders, 41 investors and former partners alike. Ghazali, welcome 42 to the show and have a to really explain a lot. 43 I do have an audience who listens from time to time to biotech content, 44 but I have to admit it's quite 45 unique. What you are doing? 46 Can you take us along your journey? You started in medical laboratory 47 science in Iran, moved to Masters in 48 Biochemistry and Molecular Biology in Germany and then founded 49 Can Chips in 2023. I bet when you started your bachelor 50 you never imagined that you would go 51 through this once upon a time moment when you realized that tumor 52 on a chip was the entrepreneurial path you felt compelled to
53 take. And can you please explain what a tumor on a chip is? 54 Yes, of Course, first of all, thank you very much for having me. 55 That's an honor to be involved in such wonderful 56 podcasts. And yeah, I mean, then I go back and think 57 about my bachelor. I could never imagine that I am here, here in the 58 position that I am today. And the moment, once 59 upon a time for me was actually during my master degree with 60 the idea of can chip that we were collecting data from like 2D models. 61 And then we were doing animal trials and some of them were working really good. 62 And then I was realizing that they are clinically really not relevant. 63 We were relying on models that they do not behave like 64 real human tumors. And that was disconnecting me from the
65 point that tumor on a chip can stop being just 66 a dream or being in academia and become a 67 mission for me. What is tumor on a chip? And since we are not doctors, 68 can you tell us why this model 69 behavior that differs from the real human being? 70 Why this is a big problem? Yeah, when we are having like the 71 preclinical data for the pharma companies and biotech companies, they 72 normally have 2D models. What is a 2D model is like a 73 flat cell culture where the cells are growing and they are 74 not acting as they can act in human body because we don't have a 75 flat organ. They are all like 76 3D. And the other 77 thing is that they are testing them on animals. And animals cannot behave like 78 humans. And they are the reasons that so many of the drug
79 developments fail in clinical phases. And that is 80 very expensive to develop a drug that is time consuming to develop a 81 drug. And what these tumor on chip models do is that they are 82 3D. They have the microenvironment of the tumor. 83 So that is really a mini tumor on a biological 84 platform that can be used for testing the 85 medications that are in development. In that area, 86 we do not need animal trials anymore. So much so we are reducing 87 the animal trials and they are also having more reliable 88 results. At the end of this story. 89 In the everyday status quo of cancer drug development, 90 which, which is a very, very difficult field 91 because we don't know a lot of what actually causes cancer. So 92 that's always a question, how to treat it. We still rely
93 heavily on animal models and like you said, the 2D 94 cultures. How does your system, the tumor on chip, 95 challenge that status quo? And what is the biggest limitation you saw 96 that led you to build this platform? 97 The way that we challenge the stethoscope per day is by creating the 98 micro environment that actually behaves like human tissue with 99 perfusion with co culture with real time function results. And 100 that means that we have a dynamic platform like the blood is 101 going through the organs, the medium for the cells is going through them. And the 102 results that we have, they are the results that you need to develop a drug. 103 And it's just not a smaller model, it's a human model. And the 104 limitation that we saw to lead us through that is as I said, 105
the results that we get sometimes are a fantasy in 2D model. 106 And we are like wow, we just developed a new drug and when it goes 107 to the clinical trial with human that is not working. And that is 108 the huge limitation because we are talking about millions 109 and billions of euros that yearly is being spent on drugs. And we 110 are talking about 14 years that one drug is taking 111 from the beginning of the drug development to when it's ready to 112 be released to the market. 113 I see, I see. And 114 that's basically where your 115 tumor on a chip really gets 116 started. So. And actually that was all 117 triggered as we already said in the intro, by your personal experience, 118 a personal catalyst. Your mother was diagnosed with 119 breast cancer and that inspired your mission. How did
120 that emotional driver shape your scientific 121 focus and your founding strategy for cancer? 122 I mean the cancer was for me always when I was also in my bachelor 123 degree was for me always a very big question mark. But I 124 realized that when my mom was diagnosed with breast cancer, the 125 science was one of a sudden for me, personal. And 126 I understood that every delay that we have in preclinical 127 phase of the drug development translates to real suffering, 128 real world suffering. That means later drugs, later 129 personalized medicine. And it sharpened my focus that 130 accelerating a realistic cancer model isn't a scientific 131 ambition. That is responsibility that someone like me has when they have an idea 132 and they have passion to bring it into the real world. I mean that's that 133 that's like being so realistic
134 at the same time being also so emotional. Because I 135 can say that each of us had someone in the family that was fighting 136 for cancer or in a different crisis and know what a story 137 is that and when we are one step more forward through personalized medicine, 138 then that is personal for all of us. And 139 due to this catalyst, you chose to integrate two 140 human cell what you call co cultures, perfusion 141 microfluids and multi 142 omics readouts. You gotta explain what that 143 is into chip design. What were the major 144 engineering biological PIV pivots you 145 encountered by developing the tumor microenvironment 146 model. So meaning that setting up a model that 147 the tumor really behaves like a tumor in a human body. 148 Exactly. I mean the point is that when we are talking about these microfluidic
149 systems, you can imagine a resin base that we are working with or 150 some other companies like a microplastic and so on. And we had to think 151 about it, that the chip cannot behave like a plastic or a resin, that 152 should behave like a tissue, that should behave like a human. And making that 153 micro environment to be in that area that was balancing like 154 biology and engineering. And we are really 155 biologists and biochemists by heart and by practice. 156 And that was the irritation that we had to mix these two 157 together. And that was the moment that I was like, 158 okay, that is what can be really good because we found 159 a way to have zero drug absorption in our platform, to have a 160 transplant transparent platform and to have this 161 engineering to come in the real world. Data that we don't talk about just
162 the chip, we talk about like real tissues. We're also 163 going to record a founders world, of course. And you once told in an interview 164 that we need a quick shift towards personalized cancer medicine 165 and we must leave our comfort zone. You'll dare 166 share moments when you personally pushed yourself beyond that comfort zone 167 and what that meant for can chips growth. 168 Until finally after prototyping and validation, can chip 169 achieve proof of concept for what you call tumor on a 170 chip models and secures its first awards and partnerships? What were the 171 breakthrough milestones you hit and how did you validate 172 your platform to pharma and biotech collaborators. 173 After this. Time of the prototyping and so on, we 174 were really in a point that we were ready to go. And I have to 175 say that our, our breakthrough was consistency. There are so many challenges,
176 daily challenges that you have in biotech and specifically when you 177 are managing to have a unique and innovative 178 platform or a service, then we 179 reproduce the same drug response curves in 180 independent runs. We could show the biotech and pharma companies that 181 this platform is ready. You can trust it because the data that we have, 182 they are clinically relevant because the data that they have shows that you can 183 reduce the animal trials and you can have better results faster. And that 184 was the moment that we were coming to talks with biotech and pharma companies 185 to understand their need to also push them a bit out 186 of the comfort zone and tell them let's make it a bit bigger together. 187 And I mean that was a journey, but that was 188 going good. 189 I see from your platform's perspective,
190 what, what are the tactical frameworks you deploy when 191 designing those 3D tumor on a chip models, 192 for example tumor and co cultures, perfusion 193 settings and real time biosensing. How does 194 this differ from the standard approach? 195 The way that we do it is normally with mixing some points 196 together. For example, that is the architecture part. As you said, we have 197 all the micro environment, we just have the tumor cells because that 198 doesn't work as it should. We have the tumor cells, we have the endothelial 199 cells, we have the immune cells that should be included. Then we have the 200 perfusion model that we have the translation of 201 the flow that we have like in human body with the 202 blood. We translate it into the tumor on chip model and 203 we have it on for the drug development. Because the absorbance will be different,
204 the effect and side effects that it has, it will be different. And then 205 was the validation and the readout. The difference that is with the 206 standard assays is that that is more complex so that there 207 are different questions that can be answered with using that platform. 208 For each type of tool for our audience. 209 Because I believe you and me, we had way more touch 210 points to cancer than we would ever like 211 to have. You hopefully mostly in professional manner. 212 For everybody who's not with a medical background in oncology. 213 How many types of tumors out there? I do 214 believe there's so many that even oncologists encounter 215 from time to time a type of tumor they've never seen before. That's 216 true. I mean they are really for each type of tumor there are 217 so many types of sub tumors. When someone has for example
218 colorectal cancer, the other one also has it. The subtypes 219 can be totally different. That is genetic based, that is based on the 220 mutation that the cancer has. And I say it always, if 221 100 companies at the same time work with the platform that we 222 have, still each of us has so much to do because there are really 223 so many areas of cancer that they are non touched, that they are 224 the cancers that you understand at the end of stage and you cannot do something 225 about it. So I really have to look it up exactly 226 how many types of more we have. But I can assure you that there 227 are plenty. And so many of them are still so much unknown to us. 228 It's quite incredible how many different types of team one can have and
229 everyone, every subtype even needs to be treated 230 differently. So it's, it's a huge field 231 and a lot can be gained by, by having the right therapy. 232 So I'm going a little bit back to can chips. What 233 are some of the scaling challenges you faced both 234 biologically and operationally? 235 I mean biologically I have to say we didn't have so many validation 236 problems, you know, the reproducibility of the system and so on. Because my 237 co-founder also had so much experience in the cancer research and 238 microfluidics, we were really going a bit more faster. But the biological 239 challenge that not us, but so many of the companies have is 240 that the variability of the primary human cells that when you are 241 working and saying I am developing a pancreas cancer on chip, you are 242 developing one of the pancreas cancer models or sub models that
243 is even possible to study on and there are many, many different more to 244 go. And that was also the maintaining the barrier 245 integrity across the chips. You know that you want to have so many different 246 cells on the chip, each of them grow differently. So you have to control 247 some of them, you have to let them grow freely and they are the 248 biological aspect of that. And operationally of 249 course that regulatory talks because we know that 250 FDA also accepts these non-animal models. But there is 251 not a golden pathway that you can say if I go this direction, everything's 252 accepted. The good point is that because we are in the 253 preclinical phase, we do not have so many regulatory aspects that we 254 have to go through. But still if you want to convince biotech and pharma company
255 to do less animal trial and come to this non animal 256 models, they need to know which direction they are going. We are 257 active in this area with fda, with EMEA to really 258 try to make it a pathway for all of us. But I have 259 to say operationally these regulations and convincing 260 is a challenge. And 261 will you talk about FDA and so on that are actually the 262 regulatory bodies for Europe? Yes, exactly. 263 Whenever a drug is wanna come into a market, they should 264 definitely have the FDA approval to be allowed. That means that they pass, 265 they pass the preclinical test, clinical test and the drug is safe 266 to be in the market for patients. 267 I was wondering for our audience if you're working on college 268 R&D or biotech investments and you're in, you're hearing
269 this. What would you ask Ghazalie right now 270 about tumor-on-a-chip models that can impact your own project? 271 Guys, we'll be right back after a short break. Dive into 272 strategic decision making and customer partner wins. That 273 really propel Startupradio forward. 274 You mentioned you are working with patient derived 275 tumor cells and integrating biosensors for real time 276 drug response monitoring, which is actually pretty cool that one can do that. 277 Can you walk us through a case study or partner collaboration where 278 this played out? Of course we have a very, very 279 good collaborator that really trusted us from the very 280 beginning, that we use the patient derived colorectal tumor cells on our 281 chip with continuous perfusion and we were monitoring the drug 282 response in real time through the sensors that we are integrating 283 and also that time through the gene expression and so on.
284 And we could discuss these manners, 285 how the cells are reacting with that partner. Instead of waiting days 286 for the results, we could have in 48 hours, 287 72 hours, the primary results that we wanted. And that is 288 gonna change the decision making, that is gonna change the personalized 289 medicine, what we want. I am gonna say it so often in every 290 interview till we see a change. We want more collaborations from 291 hospitals in Germany to go through personalized medicine. That is not 292 a one man show. We have to work all together. We receive samples from Sweden, 293 from Romania, from all the Europe. But when it comes to 294 Germany, they are like this protection of the data, even though 295 we don't need the data. And that is the place that we really can 296 develop the personalized medicine. So we have to keep it together.
297 You were talking about a 48 hour time frame here 298 with current methodology, not your tool. How long does it take. 299 In general, when we are integrating models that is a bit dependent. When we have 300 a patient drive and they are really well biopsied 301 and we see that there are cancerous cells inside, they can directly 302 go to the chips. And in 42 hours to 72 303 hours, we have the base that we wanted to have it. 304 But sometimes we may need longer because some cells are not growing 305 in the speed that we want them to grow. But I always say we 306 will have a two week time since the patient 307 is operated and the tumor is out to the time that they are ready 308 and recovered for the therapy. And in these two weeks we can
309 answer so many questions that can help them. For personalized 310 medicine. We're here on 311 StartupRadio where more than 90% according to our 312 audience survey, at least listen for professional reasons. So let's drive 313 a little bit into your strategies. What 314 growth methodologies did you adapt in your early 315 startup phase to integrate science, engineering and startup execution? 316 Because. Many note that 317 they're saying in startups, yeah, move fast and break things, but that is 318 not really working in cancer treatment. So 319 what can a biotech founder learn from that? What 320 did you learn? I mean, I have to say a 321 very, very early point is that the startup founder should be 322 brave enough to be loud about what they are doing. Because if you are developing 323 the coolest product but no one knows you, there is no worth
324 for that product. And I learned it in a hard way because our investor was 325 like, you have to go out and tell their story what you are developing. 326 So being loud is the first thing that I always say to the 327 founders as then is to be fast. You know, you cannot be perfect. 328 You have to get out the first prototype, let the end users try it, 329 even in a pilot study. Then you understand what is the issues, what is 330 the challenges and you can make it. And of course the co-development 331 with users. You know, I always say that it's not a one man show. You 332 need collaborators, you need people in academia, in other industry 333 to help you understand what is the need of the market even. 334 And of course a very good business developer to get in touch with potential
335 customer because you can develop something for them and that is 336 not what they wanted. But if you can ask them what is 337 your need, what do you want? Then the product that you're going to develop develop 338 is the thing that they're going to be interested at the end of this story. 339 And now dive a little bit into your strategic decisions. 340 For example, for you it's a strategy 341 decisions. What cancer types to target first. 342 I think high on your wish list was of course a breast cancer. I'm very 343 virtual about this. Also setting your peace strategy, 344 prioritize co cultures and so on. 345 What decisions did influence your roadmap from 2023 346 to let's say next year, 2026. And beyond for 347 us was the choosing of the cancer type and 348 targeting was really dependent on the unfortunately I
349 have to name it trend. That's a very sad story of the 350 sicknesses and cancer types that they are getting more and more and they need really 351 more studies. That is colorectal cancer that is coming to younger generations generation 352 pancreas cancer and the cancers that they are mostly in the pipeline 353 for treatment for pharma and biotech companies. And that was for us the point 354 that we said okay, we will start from that point. The IP strategy for us, 355 I mean we already filed the first patent in 2024 because that 356 is also a thing that we are doing is novel. So why shouldn't we 357 protect it? And the plan or roadmap that we have for 358 2023-2026 for us was influenced by 359 really making something big that different companies, if they 360 are smaller, if they are big, if they are universities, that they can benefit
361 from that and that the goal is to really get away from 362 animal trials. And we are also influenced by that mindset as 363 well. What 364 customers or partner wins stand out for you Like University 365 Labs here. Oswald Farmer and how do you quantify 366 the value your tumor on a ship platform 367 delivers to those collaborations? The partners 368 that we have till now really valued our models because 369 they are less variable and they are more 370 predictable. And the reason behind that is that we have customized 371 models. So we ask them what is your need? And then we will design 372 the platform based on their need. And that is what a 373 bit all the standards and one group like reduce the time 374 that they had to put on the animal trials by around 30%. The other
375 one gained functional readout that they couldn't do. And they are the things 376 that we can quantify the value of the tumor and should be the 377 feedbacks that they gave us. They said I could understand a problem in my 378 drug development that I am struggling with since years with animal 379 models and I couldn't see where the drug is going. So they are 380 the ways that we understand that we are in the right pathway. 381 When I was putting this interview together, I had to 382 admit I have not the slightest idea of what people 383 you need for such a, such a 384 cutting edge technology. So I'm considering funding and team building. 385 What are key criteria and talent profiles you're 386 prioritizing now? Especially given that 387 specialized nature of microfoods, biotech and 388 oncology. I'm very sure you won't find a lot of
389 bachelors and masters people with a specialized, 390 with a specialized track for this specialties, right? 391 Yes, of course. I mean that is an 392 area that is really needed like education 393 in that area and also lab work in that area. So when we 394 are, I mean we are also planning or expansion and our team is going to 395 definitely be people that they have already worked with tissue 396 models. So if they don't have experience in microfluidic, that is not an issue. We 397 can bring it for them. And research is that they are 398 in the translational model. How can we translate these in 399 vitro models, the models that they are happening outside of the human body 400 to the clinical models that what is happening inside human body without 401 directly testing it. So these are the areas that we mostly
402 look for our expansion and of course business developers. Because as I 403 said, if you are developing the coolest product, no one out there knows 404 what you are doing. So you cannot really bring it to the market. 405 Understood. Let's go a little bit in the future 406 outlook. How do you see the tumor 407 on a chip and organ on chip markets evolving by something 408 like 2030? Where does can chip position itself in that 409 landscape? I have to say that 410 organ on chip will mature itself from innovation to 411 infrastructure. That is where the regulatory affairs are also going. 412 So I think and I hope that by 2030 413 it will be a standard in early drug discovery because since 414 2024 so many things happened from the day 415 one that the FDA said I am going to accept these models and we position
416 Can chip as a leader in realistic tumor on chip models 417 ready for precise oncology and ready to have customized 418 solution for pharma and biotech companies by 2030 we are 419 a leader. I'm 420 always trying to tease out a little bit of contrarian view here. 421 Many believe larger animal based preclinic models will always 422 dominate due to regulatory inertia. Why do you 423 think that belief is broken and what evidence do you have to 424 challenge it? I mean the regulatory inertia is 425 always real but the science is stronger and we don't have to forget that 426 the science is moving very fast lately. Animal models 427 simply do not replicate human body anymore. They do not 428 replicate patient response enough. The field is moving 429 toward human relevant system and regulators are also 430 acknowledging that. And when we think about combining the AI
431 data that we can have around patients, all models and 432 genetic models all together then none of the animal models 433 can really compete with that. That is my point of view on 434 this case. 435 For our listeners tuning in, I would be 436 curious what is one immediate question 437 you would challenge yourself to act on this week? 438 If you would be in Cazalus Seed, be bold. Leave 439 comment here. Let's talk a little bit about 440 advice for founders. If you're listening to 441 scientist, entrepreneur or deep tech founder, what are three 442 mistakes they should avoid and 443 three practices you should adopt to navigate 444 translation from a lab to commercialism 445 to commercialization? 446 I mean there's three points that I would really really 447 would tell them to avoid is to wait for the perfect data. There is no
448 perfect data. To not overbuild the tech before 449 talking to users. Sometimes there are some platforms that they are 450 so complicated that they cannot even work with it and 451 underestimate the complexity of the manufacturing and the steps that 452 you have to take. That is always even if it's about time or if it's 453 about cost, make it a bit more than you think it's enough 454 and points to adopt is early pilots to bring your 455 product or prototype out to see what is going on. The 456 IP strategy of course to think about it and really 457 secure it. That is going to save you and prioritize the 458 things that are important for the company without thinking about what is 459 going to happen if and if this prioritization is really 460 gonna make a huge difference. 461 We getting very close to the end of our interview and
462 I would be interested if Can chip over 463 achieves what is the one breakthrough in 464 personalized oncology we look back 465 on in 2030 and say ah, that came from cant 466 chips. Yes. That would be the drug response 467 prediction that matches the patient. That is what I always say. That is 468 matching the patient. That is not matching the animal. That is not matching a 2D 469 model. That's how personalized oncology become real. That we 470 can say this patient needs this specific drug. They get 471 it, they will get so healthy and the cancer will never come 472 back again. That will be great because a lot of people are 473 still dealing with 474 with, with cancer coming back and coming back and coming back. That's. 475 Really a tough spot to be in because it always gets harder to fight back.
476 That was depressing Note. I'm so sorry. We usually close our 477 interviews with two standard questions. So number one, 478 are you open to talk to new investors? Yes. 479 I learned it also that I should not close any doors, you know, because that 480 is not just about money. That is about the network. That is about the 481 knowledge that the investors also bring. So for a talk I am always 482 open. Uh huh. And of course the 483 usual question. Are you looking for talented employees? I 484 have to say that I have enough cvs that I can hire 485 minimum 10 people by tomorrow and they are highly educated. And 486 the point is that because we are an English speaking company 487 that makes it easier for them to approach to us because this language barrier 488 is real here. And I am, I mean we are
489 expanding but I already picked the people that I want to hire. 490 Aha. And we learned a pretty good secret. Pretty 491 big secret. You get good talent here in Germany if you are 492 an English speaking company. Aha. Aha. 493 Aha. So we finally teased something really cool out. 494 Thank you very much. Best of luck for your future mission for can 495 chips and and congratulations again for being 496 one of the winner of the German Startup Awards 2025. Thank you 497 very much. Thanks a lot. 498 That's all folks. Find more news streams, 499 events and 500 interviews@www.Startupradio.IO. 501 remember, sharing is caring. 502 Sam.
Partner with Startuprad.io
Startuprad.io is the leading independent media platform covering startups, venture capital, and innovation across the DACH region (Germany, Austria, Switzerland) and Europe. We offer B2B partnership opportunities for companies looking to reach startup decision-makers, founders, and investors.
Become a Partner — Learn about sponsorship and partnership opportunities
Contact us: partnerships@startuprad.io
Editor-in-Chief: Jörn "Joe" Menninger on LinkedIn
Subscribe to the Podcast
All podcast links: https://linktr.ee/startupradio
1. What is a tumor on a chip?
A microfluidic device that recreates human tumor environments for drug testing.
2. How does it differ from 2D models?
It uses 3D architecture, perfusion flow, and co-cultures to mimic real tissue.
3. Can it replace animal testing?
Yes — significantly. Many trials see 30% reduction.
4. Is the FDA accepting these models?
Yes, since 2024 FDA recognizes non-animal models.
5. How fast are results?
48–72 hours for drug response predictions.
6. What cancers can be modeled?
Colorectal, breast, pancreatic, and more.
7. Are chips customizable?
Yes — each platform can be tailored to partner needs.
8. How does perfusion work?
Fluid flow mimics blood circulation in tumors.
9. Why do animal models fail?
Their physiology and genetics differ from humans.
10. Can hospitals use these models?
Yes — for personalized oncology and faster decisions.
11. Are patient biopsies required?
Yes, for personalized tumor chips.
12. How reproducible are the chips?
Highly reproducible drug response curves.
13. What sensors are used?
Viability, metabolic, flow, and gene-expression sensors.
14. Is this scalable?
Yes — manufactured in modular chip formats.
15. What startups work in this space?
CanChip among the emerging leaders.
16. Does this help reduce clinical trial failure?
Yes — by improving preclinical predictiveness.
17. Can immune-tumor interaction be studied?
Yes — chips support immune cell co-cultures.
18. What industries benefit?
Pharma, biotech, hospitals, CROs.
19. Is data real-time?
Yes — sensors monitor tumor behavior continuously.
20. What’s the future by 2030?
Organ-on-chip becomes standard oncology infrastructure. Internal & External Linking This Month in DACH Startups - November 2025 | Deep DiveThis Month in German, Swiss and Austrian Startups - November 2025 (Top News)Synthflow’s Voice AI With Memory: The Contact Center Breakthrough Authority Sources FDA Modernization Acthttps://canchip.org/about-us/ The video is available up to 24 hours before to our channel members in what we call the Entrepreneur’s Vault. The Host & Guest The host in this interview is Jörn “Joe” Menninger, startup scout, founder, and host of Startuprad.io. And guest is Ghazaleh Madani | CEO and Co-Founder | CanChip Joe on LinkedIn Ghazaleh On LinkedIn
About the Host
Joern "Joe" Menninger is the host of the Startuprad.io podcast and covers founders, investors, and policy developments across the DACH startup ecosystem. Through more than 1,300 interviews and nearly a decade of reporting, he documents the evolution of the European startup landscape. Follow Joern on LinkedIn.
Support Startuprad.io
Tumor-on-chip technology represents a frontier in deep tech innovation with direct clinical impact. Companies building in deep tech and life sciences use Startuprad.io to reach founders, operators, and decision-makers across the DACH ecosystem. If that fits your goals, explore partnerships here: Partner with Startuprad.io
Comments