November 2026—This conversation is part of a series of one-on-one virtual roundtables in which CAP TODAY publisher Bob McGonnagle speaks with one vendor and one laboratory expert to spotlight a company and a customer for their laboratory solutions and work. He spoke on Aug. 6 with Emmanuel Gorospe, MD, MPH, GI medical director for Castle Biosciences, and Daniel Jondle, MD, gastrointestinal pathologist and director of the Arizona Digestive Health pathology laboratory. Dr. Jondle is a paid speaker of Castle and was compensated for his participation in this roundtable session.
Risk stratification has been increasingly essential for pathologists and laboratories, particularly as we try to optimize treatment and follow the course of patients. Dr. Jondle, what does risk stratification mean in the context of gastrointestinal pathology, particularly Barrett’s esophagus?
Dr. Jondle (Arizona Digestive Health): Within the context of Barrett’s esophagus, it’s easy to understand. About 95 percent of Barrett’s biopsies don’t show evidence of dysplasia. Patients with dysplasia are at a higher risk of developing esophageal cancer and therefore need to be more aggressively surveilled or treated. For patients without dysplasia, we can surveil more conservatively and keep an eye on them.
While they are considered low risk, a subset of the 95 percent of patients who are diagnosed as having Barrett’s without dysplasia do progress to dysplasia and adenocarcinoma. I want to identify the patients who are going to progress so that information can help guide the surveillance and treatment decisions their treating gastroenterologist will make, but I don’t want patients who aren’t going to progress to be over-surveilled or over-treated.
Up until recent years, we’ve only had a few different stains and our eyes to determine if dysplasia is present. Some people use immunohistochemistry, the p53 stain in particular, to help determine if dysplasia is present, but it hasn’t proven itself good enough to use for risk stratification.
For decades, I have wanted a test that could help identify which patients with Barrett’s esophagus are at higher risk of progressing to cancer. Fortunately, we now have TissueCypher that does this.
Dr. Gorospe, explain how the TissueCypher test came into existence, what its characteristics are, and where it stands today.
Dr. Gorospe (Castle Biosciences): TissueCypher is an AI-derived technology that pulls together nine protein biomarkers, including p53. But it’s more sophisticated than that because we’re not just measuring the intensity of a single biomarker; it brings together nine in the context of the seven cellular structures. From multiplex immunofluorescence slide staining, using AI we are able to have image analysis extracting those features. We’re looking at not just biomarkers but also what we call spatialomics. The test uses spatialomics technology to give a patient an individualized five-year risk of progression.
About 4 million people are living with Barrett’s esophagus in the U.S. We don’t want to miss the needle in the haystack because we’re dealing with esophageal cancer, which, next to pancreatic cancer, has the second-lowest five-year survival rate. Knowing a patient’s risk allows you to tailor their treatment.
Give us insight into TissueCypher’s development. The algorithms involve intensive work, including algorithms directed against the digital slide and in this case with an immunofluorescence element, which is particularly challenging in digital pathology.
Dr. Gorospe (Castle Biosciences): Validation studies were conducted to establish the evidence supporting these features. The features were derived and then validated in case-control studies involving more than 600 patients, making this one of the largest studies assembled for this purpose. Beyond validating the test, researchers also evaluated its clinical utility, demonstrating that the use of a prognostic score like TissueCypher can lead to changes in patient management.
Dr. Jondle (Arizona Digestive Health): To be clear, a pathologist is required to make the diagnosis of Barrett’s. Only after that does the tissue get sent to Castle Biosciences.
AI in medicine seems to be strengthening the role of the physician, in all specialties but particularly in pathology because the initial pathologic examination is the first step.
Dr. Gorospe (Castle Biosciences): It strengthens our work and partnership with pathology. When we have a patient with Barrett’s esophagus, our colleagues in pathology help us determine the presence of disease. The next question is how to tailor management for someone with nondysplastic Barrett’s. What is the chance this patient will progress? TissueCypher assists pathologists and endoscopists by providing prognostic information that goes beyond histology.
To clarify for our readers, this is a laboratory-developed test. The slides are sent to your laboratory, where the evaluations are made and a report is then issued to people like Dr. Jondle to supplement his findings in the initial case.
Dr. Gorospe (Castle Biosciences): Correct. It is not only an LDT but also one of the few ADLTs [advanced diagnostic laboratory test], which means there is no other test or aggregative test that can give the same results this test does.
Dr. Jondle, I’m sure many people over the years have wanted to show you the latest and greatest aid to maximize the quality of your diagnosis. What led you to TissueCypher after so many years of experience?
Dr. Jondle (Arizona Digestive Health): TissueCypher grabbed my attention because of how many different biomarkers it looks at, including p53. It is more robust than any other test I’ve seen. They do multiplex immunofluorescence staining and whole slide scanning and end up with millions of data points for each specimen. The AI algorithm they created allows them to generate an individualized risk score that gives the patient and gastroenterologist data to help assess that patient’s risk of progression.
Tell us how you became acquainted with TissueCypher.
Dr. Jondle (Arizona Digestive Health): The Castle Biosciences representative came to me and, after just a five-minute introduction of the test, I realized there’s something more to it than most of the other tests offer. I was then invited to tour their lab, which made a big difference, especially seeing how they process specimens. The lab is impressive and they have an impressive crew performing the tests.
Dr. Gorospe, it’s easy to think of advanced LDTs as a black box, but there’s nothing like seeing the operation with your own eyes. Do you typically sponsor lab tours?
Dr. Gorospe (Castle Biosciences): We welcome everyone interested in our technology. My role as medical director is to interface with clinicians, answer their questions, and provide the scientific background to help them care for their patients using this test.
Let’s return to the question of risk surveillance and why it’s important. What is the cost of a false-positive dysplasia in Barrett’s, in terms of morbidity and mortality?
Dr. Gorospe (Castle Biosciences): Esophageal cancer has a very low survival rate, so we don’t want to miss progression. In the past, when pathologists were deciding whether to call high-grade dysplasia or early cancer, I’m sure they struggled because the treatment was esophagectomy. Now, with advances in technology, we can offer minimally invasive endoscopic eradication treatments, such as radiofrequency ablation or endoscopic mucosal resection.
I was a practicing GI before I joined Castle, and I’ve seen two things change: We’re moving beyond histology, and although we have many treatment options for this disease, we’ve lacked a tool to assess the risk of progression. We’re addressing that unmet need.
Dr. Jondle, you serve many physicians within your large group. Have they been receptive to this additional information you can provide?
Dr. Jondle (Arizona Digestive Health): They’ve been very receptive, which is not common because we don’t want to blindly accept any new test that comes about.
One interesting anecdote I’ll share: A few weeks ago, a gastroenterologist called me and said, “I know I have Barrett’s. I’ve been screened twice before. I lose sleep at night because I know some people with Barrett’s progress, but we don’t always know which ones will.” He asked me how he can make sure that TissueCypher would be performed on his biopsies. It was very telling that he and many of my GI partners are embracing this test.
Dr. Gorospe, what do you say to gastroenterologists and GI pathologists who think p53 is a golden tool?
Dr. Gorospe (Castle Biosciences): First and foremost, we’re not just looking for an aberrant or wild-type p53. Instead, we measure p53 in the context of spatialomics. For example, some of the factors we look for are p53 with some nuclear intensity, p53 mean intensity in the nuclear clusters, or the nuclear area in which there’s a positive p53 but p16-negative cells. Those are specific markers the human eye cannot decipher. Second, before TissueCypher, trying to convey to a patient that they have a wild type or an aberrant type was more of a qualitative discussion—you’re high risk or low risk. With this test I can say, for example, “Your risk of progression is 20 percent in five years. Is that a risk you want to take? Should we consider treatment at this point?” Especially when you benchmark it against the risk for progression: Population-based is three percent in five years. If I have a patient and they see a 15 percent progression risk, that’s fivefold. So the conversation changes; it becomes more specific, more meaningful.
Dr. Jondle, did this make sense as you acquainted yourself with the TissueCypher test? That is, how TissueCypher is a complex assay that requires algorithms?
Dr. Jondle (Arizona Digestive Health): Absolutely. Any of us who use p53 IHC realize that there is tremendous variability, not only in the staining but also in the interpretation. Having tens of thousands of data points gathered with the p53 data in combination with millions of other data points with the other markers they use makes much more sense than one stain with one interpreter.
Dr. Gorospe, spatialomics is becoming an ever-hotter topic, one that will need to be well understood particularly by pathologists and physicians, because this is a relatively new entrant into the clinical arena. Do you agree?
Dr. Gorospe (Castle Biosciences): Even as a relative newcomer, we have performed more than 100,000 tests. The utility is not only in finding that needle in the haystack, the high-risk patient, but also in finding the vast majority of patients who are low risk. Imagine being told you have a premalignant lesion, and your gastroenterologist says it’s nondysplastic; come back in three to five years. You’d ask for more information, or you might not return. By giving a patient their risk score, you can properly risk-stratify them according to what the guidelines recommend. For the high-risk patients, it’s closer surveillance or consideration for endoscopic eradication therapy. For low-risk patients, you have something objective that can allay their concerns.
Tell us about the difference between precancerous and risk stratification.
Dr. Jondle (Arizona Digestive Health): We want to know what the chances are that something will turn into cancer and how quickly. Most of our guidelines are based on big data sets that say patients with “this” level of change are “this” likely to progress to cancer within “this” time frame. In the past five to 10 years, we’ve been developing more molecular tests that lead to more exact and precise guidelines, with recommendations based on much more data. Instead of saying you have Barrett’s without dysplasia and we’ll see you in three to five years, we’re saying based on this molecular test called TissueCypher, you have a 50 percent chance of progressing in the next five years. Because of this, we will bring you back next year—not wait three to five years as we used to—and take more biopsies because we don’t want you to develop cancer, which you might do if you wait five years before you are seen again.
Dr. Gorospe, each time there’s a new report from the WHO or other group, it seems some of the molecular information coming to the fore is changing our basic understanding of these diseases. Is it not?
Dr. Gorospe (Castle Biosciences): Correct. In fact, American Gastroenterological Association guidelines say that molecular testing such as TissueCypher can help risk-stratify a subset of patients who are nondysplastic, those whom we have thought of as low risk. We traditionally believed that high-risk patients were male, Caucasian, older than 50, with history of gastroesophageal reflux disease or family history of Barrett’s esophagus. We did an analysis of more than 8,000 patients, and you see high-risk patients not only in males but also in females. They can be an ethnicity other than Caucasian, and they don’t necessarily have symptoms. Mayo Clinic performed a pooled analysis and found that TissueCypher was the strongest predictor of progression beyond clinicopathologic risk factors.
If you have good risk stratification, do you agree that the classification of precancerous condition falls by the wayside?
Dr. Gorospe (Castle Biosciences): Yes, because when you say “precancerous” or that Barrett’s esophagus is the known origin of esophageal adenocarcinoma, patients get scared. But if you’re able to say you have Barrett’s esophagus, our pathologist has confirmed it is nondysplastic, and then you add a personalized progression rate, you could allay those fears. In our own data, we see the vast majority are low risk, but when you have a patient who is high risk, it reassures them that you can take action before histologic changes become more apparent.
Dr. Jondle, it seems to me the test makes the pathologist ever more prominent because increasingly we believe pathologists will not only be talking to their colleagues in medicine but also to patients and consumers and will have a big role in explaining this new science.
Dr. Jondle (Arizona Digestive Health): Absolutely. I have fielded dozens of questions from my gastroenterologist colleagues about TissueCypher, and they’ve appreciated my understanding of the test and ability to explain its utility. I occasionally discuss results with patients as well. Patients are increasingly recognizing the important role pathologists play as members of their health care team.
I’d like to add that based on our data locally, from Arizona, Colorado, and Utah, about 10 percent of the diagnoses of Barrett’s with no dysplasia are coming back from TissueCypher as intermediate or high risk. It’s highlighting that subset of patients who will likely progress. Thanks to TissueCypher, those patients are going to get the therapy they need. That’s reassuring to me and my GI colleagues.
Dr. Gorospe (Castle Biosciences): I was wondering if Dr. Jondle could give us a hint of how TissueCypher works in his daily workflow.
Dr. Jondle (Arizona Digestive Health): It does not affect my workflow. It’s wonderful. From the pathologist side, there is no additional work. It all takes place in the background.
Do you regularly hold case conferences with your colleagues about Barrett’s and the role TissueCypher results have played in patient management?
Dr. Jondle (Arizona Digestive Health): We haven’t yet. We’ve been using the test for about a year, and in the next six to 12 months we will gather all that data and incorporate it into a report and likely start sending out a periodic report to my GI colleagues.
Dr. Gorospe (Castle Biosciences): With our commercial group, the sales representatives serve as partners and resources for our clinicians. They bring me out as medical director, or bring in their medical science liaison, to sit down with a group of physicians to review the test results, get an overview of the demographics and results, and allow the physicians to validate among themselves the clinical utility of having such tests in their practice. So far, the response has been positive.
Dr. Jondle, do you expect there to be a stream of publications in the field?
Dr. Jondle (Arizona Digestive Health): There are already more than 17 peer-reviewed utility and validation articles in the literature, and more are coming [https://bit.ly/4yvJgUM].
What would you like to see next from Castle Biosciences?
Dr. Jondle (Arizona Digestive Health): A big issue many GI pathologists are talking about is intestinal metaplasia in the stomach. Just as intestinal metaplasia in the esophagus is precancerous, some intestinal metaplasia in the stomach is precancerous, but we don’t know which cases are precancerous. We need a way to predict which patients will progress. I have talked to the Castle Biosciences team, explaining this is a burgeoning area and that it would be helpful if we had a test to help us predict progressors for gastric adenocarcinoma.
Dr. Gorospe, I can tell from your enthusiasm and activity that you have not ignored this question as you continue the Castle Bio story.
Dr. Gorospe (Castle Biosciences): Spatialomics has many applications. At Castle, we are always looking to answer an unmet need. In GI now, it’s Barrett’s esophagus, but we’re looking to partner with our colleagues like Dr. Jondle, and all the clinicians and gastroenterologists facing questions, on how to provide more individualized precision medicine.
As we wrap up, I invite you to share closing thoughts.
Dr. Jondle (Arizona Digestive Health): I was talking recently to a group in Wisconsin about TissueCypher. When they have tough Barrett’s cases, they oftentimes will send those out for an external consultation with an expert GI pathologist. They are excited that they can now send these cases for TissueCypher and save the expense of a costly consultation for the hospital.
Dr. Gorospe (Castle Biosciences): TissueCypher has allowed us, as gastroenterologists, to work closely with our pathologists in Barrett’s esophagus. Our discussion now is no longer putting pressure on them to answer if it is nondysplastic, low-grade dysplasia, et cetera. Prognostication is individualized and beyond histology. Our own GI guidelines suggest having two GI pathologists review samples; it’s hard enough for some rural places to find one. TissueCypher could help augment that service and help foster a better working relationship between GI and pathologists, not just in establishing the diagnosis of Barrett’s but also providing prognostication.
CAP TODAY does not endorse named products in commercial roundtables, and the publisher’s comments are his own and not necessarily reflective of CAP TODAY or College of American Pathologists policy.