Key Takeaways

• Molecular profiling is evolving from a one-time assessment into a longitudinal strategy, with circulating tumor DNA (ctDNA) testing offering opportunities to identify actionable alterations, monitor treatment response, and detect emerging resistance.
• Minimal residual disease (MRD) testing provides valuable prognostic information across several cancer types, but oncology APPs must distinguish between established clinical applications and investigational approaches to MRD-guided treatment decisions.
• Oncology APPs play a critical role in interpreting molecular reports, recognizing the limitations of tissue- and blood-based testing, coordinating multidisciplinary care, and educating patients about the implications of biomarker results.

Raymond DeMatteo, PharmD, BCOP, a clinical pharmacy specialist in the Early Drug Development Department at Memorial Sloan Kettering Cancer Center, discusses advances in precision oncology highlighted at the 2026 American Society of Clinical Oncology (ASCO) Annual Meeting and their implications for oncology advanced practice providers (APPs). He examines the expanding role of circulating tumor DNA (ctDNA), minimal residual disease (MRD) testing, and molecular profiling in identifying actionable biomarkers, monitoring disease, and detecting treatment resistance. For oncology APPs, these developments reinforce the importance of genomic literacy, collaboration with oncologists and pathologists, and clear patient communication to support informed treatment decisions and appropriate integration of emerging precision oncology strategies into clinical practice.

Transcript

My name is Raymond DeMatteo. I am a clinical pharmacy specialist in the Early Drug Development Department at Memorial Sloan Kettering Cancer Center, and I've been practicing clinically now for about 6 years.

Major precision oncology themes emerged from ASCO that highlight the need for adequate testing for patients to ensure both predictive and prognostic biomarkers are found to help inform patients and clinicians on potential therapies and guide them through their treatment trajectory.

Themes included biomarkers becoming longitudinal rather than episodic, meaning that we are looking at biomarkers now before treatment, during treatment, as well as after treatment; assessing disease status using molecular techniques to identify patients at risk of recurrence; biomarkers being used to monitor response and not just predict response, which is a shift from how we have classically thought about these findings; as well as using biomarkers to determine resistance patterns for patients in hopes of providing further options for them.

I think APPs play a crucial role in ordering and facilitating testing for patients. Every cancer patient should be getting appropriate molecular testing for their disease at some point in their journey. And since APPs will see patients throughout their treatment course, it is essential for them to be able to understand molecular findings and bring potential issues up to medical oncologists or pathologists when appropriate.

The expansion of molecular testing over the last few years has made it a part of every APP workflow, so they know how to educate patients on these findings and inform them of potential therapies that are being made available based on these assays. Knowing how to utilize and analyze these reports is really vital for oncology APPs that are currently in practice.

Several testing methods can be conducted to help guide treatment decisions and determine prognosis for patients either undergoing active treatment or contemplating additional treatment for their disease.

ASCO actually updated their guidelines right after the annual meeting that highlighted how we should be using powerful new molecular techniques, such as liquid biopsies, and reviewing when they are appropriate and standard of care versus, say, a clinical trial.

I think for this session, we'll review the different studies that assessed molecular techniques and describe the purpose of these tests while focusing on the general understanding of the results.

Now, management strategies following interpretation may vary depending on the strength of data, and that will be important when APPs are translating these findings to their patients and explaining if it warrants any changes in therapy, as these are common questions that will be brought up to you when patients look things up on their own or see these molecular results.

I think overall, APPs are going to be doing a great job of bridging that gap in the lack of understanding on the patient front and ensuring that patients feel comfortable with these results, which can oftentimes be confusing for them. I think that APPs are really doing a great job of educating these patients and going through exactly what's being tested. How is it going to impact their outcomes? And I think that we're going to see a lot more emerging data come out of this.

ctDNA has several advantages over tissue-based genomic testing. I think at first there was a lot of skepticism in terms of concordance between tissue-based testing and liquid testing, but we know now that the ctDNA assays are getting better at detecting alterations that will show up on classic tissue-based platforms.

Insufficient tissue attained during routine biopsies is one of the pitfalls of tissue-based molecular testing. Furthermore, there are instances where next-generation sequencing might fail due to insufficient DNA yield from tissues that were biopsied, and this can result in major delays of care since molecular testing for actionable mutations can predict response to certain therapies and is really essential in the workup of certain types of cancer, like non-small cell lung cancer, for example.

ctDNA also has the advantage of detecting tumor heterogeneity, meaning that it can pick up other mutations that may not be present in the tissue that was biopsied. This is a great way to understand the full molecular picture for a patient's disease, since all parts of a patient's cancer may not have the same molecular profile, and this could result in the possibility of certain alterations being missed, some of which could be actionable.

Furthermore, using ctDNA and assessing resistance has really become a huge advantage in determining next steps of therapy for patients. Some patients may have on-target versus off-target resistance on certain therapies that can be picked up by a simple blood test now, and the results of these tests can really potentially give patients more options.

There was a bunch of data that came out of ASCO. Impossible to go through it all, but we'll go through some that I think are prevalent.

One poster that I saw evaluated matched ctDNA to tumor samples for patients enrolled in the NCI-MATCH trial, which was a precision medicine clinical trial for a variety of cancers, and it showed a really high concordance rate of about 90% between tumor tissue and blood sampling, displaying that this strategy is viable for some patients with more common solid tumors, such as breast, lung, and prostate. We're really seeing that concordance rate between tissue-based molecular profiling and blood-based molecular profiling matching up very well.

Now, in terms of tumor heterogeneity, there was an abstract that assessed 10 patients with advanced solid tumors who underwent tissue and blood testing using commercial next-generation sequencing platforms, and they found a concordance rate of about 88%, plus 4 blood-only findings where 1 was actually actionable.

This really highlights that ctDNA can pick up a potential actionable mutation that is missed on tissue based on the next-generation sequencing results. This really supports the need to assess ctDNA when possible with tissue-based NGS and the need to expand the sampling size of these studies to really drive home the fact that patients should really routinely be getting both tissue sampling and blood-based sampling.

Now, in terms of emerging resistance, there was an interesting phase 2 study that came out of China that evaluated savolitinib, which is a MET inhibitor, in patients with MET-amplified gastric or gastroesophageal junction (GEJ) adenocarcinomas. It showed a pretty decent overall response rate of about 32%. But what was interesting is that they actually had an exploratory endpoint where ctDNA biomarker analysis showed that patients harboring baseline alterations in other oncogenes, such as FGFR2, EGFR, BRAF, or KRAS, had a lower overall response rate and a shorter progression-free survival. Additionally, patients who provided biopsies at progression developed alterations in these genes that were identified as acquired resistance mechanisms, and actually 4 patients had acquired on-target mechanisms.

This really highlights that ctDNA can be used to infer resistance based on tumor heterogeneity. And it's really more applicable to the ctDNA since we don't actually know what's going on if we just got it from one particular part of the patient's disease. In terms of longitudinal monitoring, this is an ongoing strategy that's utilizing a lot of ctDNA-based assays.

One poster highlighted that patients with advanced solid tumors who were treated with immune checkpoint inhibitor regimens had received serial plasma sampling with ctDNA, and that sampling of ctDNA demonstrated superior prognostic performance as opposed to imaging. They also showed that molecular progression preceded radiographic and clinical progression by a median of 63 and 77 days. This study really shows that ctDNA can detect the emergence of early immune checkpoint inhibitor-induced resistance in patients treated with these therapies who might have stable imaging.

I do think that further ongoing research will determine how we should act on these findings, and larger sample sizes are also needed just because these studies were limited by low sample sizes. But I do think the data are very promising, and it may help us guide how we can incorporate ctDNA into clinical care.

I think APPs need to know what a positive result means in the context of the clinical picture. It's vital that the entire molecular report is read in detail so you are aware of how to interpret a positive result, since some of these will give you an overview of the findings and provide you with useful information. For example, they will tell you if something is potentially actionable or not. It is also useful to do some of your own research. I'm a big proponent of this and figuring out what these alterations mean in the context of your patients. There are several databases that exist, such as OncoKB, that can give you a more comprehensive report on specific alterations.

In terms of a negative result, it's important to highlight that this does not mean there are no relevant biomarkers for patients who undergo liquid biopsies. ctDNA can lead to false negatives, especially in patients who have a small amount of disease. It makes sense because ctDNA is measuring the DNA shedding from tumors. If patients don't have a high burden of disease or disease is very localized, they might not actually be picking up anything on a blood-based assay. Furthermore, ctDNA is not as sensitive as tissue-based NGS assays in picking up certain alterations. You may need a tissue biopsy to get the complete picture and ensure nothing was missed if a ctDNA result is negative.

Minimal residual disease refers to the small number of cancer cells that persist after or during treatment for a patient with cancer. These patients may be in remission, actually, with no clinical symptoms, which is why MRD is so powerful since it gives us an idea if there is any latent disease that isn't showing up by our traditional exams, for example, by imaging, lab testing, or clinical symptoms.

MRD testing utilizes ctDNA for determination, and we now have several different assays that are used to assess MRD status with a level of detection that makes it very sensitive. Furthermore, it can be personalized to each patient now, which makes the test extremely specific for tracking someone's disease over a long period of time.

The first one I do want to mention is the CIRCULATE-AIO/ABCSG trial, which was a European trial that assessed patients with stage II colon cancer without high-risk features who were assessed for ctDNA positivity after surgery.

Those who were positive were randomized to get adjuvant chemotherapy versus observation with no chemotherapy. They first found that ctDNA was highly prognostic, with 3-year disease-free survival being about 87% in ctDNA-negative patients compared to 38% in ctDNA-positive patients. And they actually found in the per-protocol analysis that there was a lower recurrence rate in patients who were ctDNA-positive and received chemotherapy versus those who were observed. However, it is worth noting that in the intent-to-treat cohort, it did show a 3-year disease-free survival of about 61% versus 38%, but it was not statistically significant. This highlights how this may not be clinically applicable yet in the real-world setting, but still promising.

Another study was the CIRCULATE-Japan GALAXY study. This also assessed early-stage colorectal cancer patients who received adjuvant chemotherapy with 2 or more ctDNA results within 6 months postsurgery. ctDNA was assessed at several different time points, and patients were divided into 2 groups. It was either a short adjuvant chemotherapy group, and that was denoted as adjuvant chemotherapy being less than 3 months, or long adjuvant chemotherapy of 3 months or longer. They actually found no difference in the short versus long adjuvant chemotherapy duration among patients with sustained ctDNA negativity or ctDNA clearance.

What was also interesting is they found that patients with rising ctDNA experienced no benefit from longer adjuvant chemotherapy, indicating patients may need alternative treatments. This may help guide the de-escalation of patients to receive shorter adjuvant chemotherapy with MRD negativity. However, it is worth noting this study was observational, so extrapolating into the clinical setting may not be appropriate at this time.

Now, for lung cancer, there was the TRACERx study, and this looked at patients with perioperative stage IA to IIIB resected non-small cell lung cancer and showed preoperative ctDNA positivity was independently associated with worse recurrence-free survival and overall survival. Furthermore, it showed postoperative positivity predicted a higher risk of recurrence or death, with levels above 10 parts per million having poor outcomes. This actually gave a quantitative value from the ctDNA assay to determine who might actually do worse, which I thought was pretty interesting.

There were definitely some ongoing studies that were presented at ASCO, and there were no results yet, but I do think they're worth mentioning because they will be presented, I think, at future meetings, specifically in the early breast cancer realm.

There was the phase 2 ASPRIA trial, and this is including patients who have triple-negative disease who have residual disease following curative-intent therapy and detectable ctDNA. ctDNA-positive patients will receive atezolizumab, which is an immune checkpoint inhibitor, as well as sacituzumab govitecan, which is an antibody-drug conjugate against TROP2, with ctDNA assessment occurring periodically over 3 years to determine the rate of ctDNA clearance, with recurrence risks and survival outcomes being assessed.

Similarly, there's the PREDICT-RD study. This is assessing MRD status postneoadjuvant treatment with chemoimmunotherapy in early-stage triple-negative disease and escalating treatment for MRD-positive patients with datopotamab deruxtecan, which is another TROP2 antibody-drug conjugate.

And then lastly, I think one to be on the lookout for that was a poster presentation was the SIGNAL ER+ 101 trial. This is an ongoing phase 2 trial that will enroll patients with intermediate-risk stage II hormone receptor-positive, HER2-negative breast cancer who have undergone surgery, and patients who are deemed MRD-positive at baseline will receive endocrine therapy plus a CDK4/6 inhibitor for 2 years and then undergo ctDNA testing every 3 months. And if patients are negative, they will receive endocrine therapy alone with serial monitoring. Now, what's interesting is that for patients in the endocrine therapy-alone group, if they actually convert to MRD-positive, they will be escalated to a CDK4/6 inhibitor with endocrine therapy.

The primary outcome to be assessed is looking at disease-free survival with long-term follow-up over 9 years. I think that this is a lot of promising prospective data that will help us make treatment decisions based on MRD positivity.

A lot of times, what we've seen is that a lot of the data coming from the MRD space are saying these patients may have poorer outcomes, which means we may need to monitor them more closely. I think the utility of switching therapies for patients who are MRD-positive remains somewhat elusive at this point, and we need more prospective randomized data to ensure changing treatment based on MRD status is appropriate. From a predictive standpoint, I still think the data are needed, but I think future studies are going to help answer if we can use MRD as a marker for switching therapies.

I'm excited to see the results at future conferences, but I think a lot of the time we're looking at this more from a prognostic standpoint. We have to explain to patients that it can be prognostic and show their cancer may come back or a treatment may not be working as well, but we don't use it to dictate future treatment.

And sometimes patients may not have radiographic progression following MRD positivity, and this is especially prevalent with patients in the metastatic setting. This actually means that patients may get more time on a certain therapy, and this can be a good thing, especially if they're tolerating it.

The reasons why it may not be incorporated into routine surveillance could be that the data are not supporting it in their specific malignancy, cost. I think there's still a lot of data that need to come out in terms of financial toxicity of using these new assays over a long period of time.

And also, we don't always know what to do with that information. While it can stress patients out, I think having APPs speak to patients about these concerns and educating them will help them understand where we are as a field when it comes to monitoring these biomarkers.

There's a lot of different terminology on these molecular reports, and knowing the difference between certain alterations is really important since it may dictate whether patients are eligible to receive certain biomarker-driven therapies.

First, you might see something like, what's a somatic mutation versus a germline mutation? Somatic mutations refer to mutations that occur spontaneously and are not inherited. You can think of these being more environmentally driven, whereas germline mutations are mutations that are inherited from a parent and can be passed on to offspring.

Regarding terminology for specific alterations, mutations are defined as a change in a nucleotide that will result in an amino acid change that can result in a change of a protein that can result in it being either constitutively active or constantly activated, or it can result in a protein losing its function. And those will, we have large mutations in our oncogenes as well as our tumor suppressors. An example with this would be an EGFR mutation, say, in non-small cell lung cancer, and that's causing the driving of that non-small cell lung cancer and causing those cells to proliferate.

Amplifications refer to deletions or gains of the number of certain genes that are involved in cellular regulation. And again, these can be in oncogenes or tumor suppressors. Those extra copy numbers or deleted copy numbers result in uncontrolled growth. A good example of this would be HER2, where we'll look at HER2 amplification for patients with breast cancer to see if that's a main cause, if that's driving their cancer.

And then also we have fusions, which involve chromosomal rearrangements. This can actually result in pro-growth pathways being constantly activated. A classic example of this would be the KIF5B-RET fusion, where a part of that KIF5B gene has now joined with the RET proto-oncogene, and it's causing uncontrolled growth.

Now, it's worth noting that sometimes you can have alterations that are drivers, meaning that we know they are causing the cancer to grow, and you can also have ones that are passengers, where they actually might not necessarily be responsible for cancer growth. That can actually be denoted on some of these reports.

It's not always as straightforward as this, however. There could be findings that are considered variants of uncertain significance, or VUSs, where there's insufficient data to really know the clinical consequences of that finding, and more data are really needed to say if it's a driver or not.

And then making things even more complicated are variant allele frequencies, which refers to the fraction of alleles carrying a specific alteration. In a very broad way, you can think of this as higher VAF values suggesting a high proportion of tumor cells that harbor that specific alteration, whereas a low VAF suggests that that finding may not be as prevalent in a person's disease. This is important, especially in the realm of ctDNA, since some alterations that are found may have very low VAFs, in which case you aren't sure if that's actually something that's real or if it's background noise. This is why it's really important that we always are in collaboration with our medical oncologists as well as our pathologists.

There are several different panels that exist that test different numbers of genes for biomarker analysis, and ASCO ctDNA guidelines suggest that ctDNA testing should not necessarily replace any other standard-of-care testing, but it is useful for certain diseases and helping direct therapy given circumstances. Basically, if a patient doesn't have a safe area to biopsy or if there's going to be a long turnaround time, you might want to use ctDNA over tissue-based molecular profiling.

It's also worth mentioning that certain aberrations are more sensitive on tissue-based assays versus ctDNA. An example of this would be gene amplification. Gene amplifications are actually very hard to pick up using liquid biopsies or with ctDNA-based NGS platforms. For that reason, it's really good to have that complementary tissue-based platform to make sure that you're not missing anything. I think we see that these typically go hand in hand now, and we're always checking one and the other if we can.

Several studies were presented at ASCO that highlighted this movement. The first I want to review was a prospective multicenter longitudinal study that assessed 143 treatment-naive patients with rhabdomyosarcoma. Authors utilized integrated plasma and cerebrospinal fluid, or CSF, ctDNA analysis to track tumor evolution and therapeutic resistance over time. They actually found that ctDNA dynamics mirrored disease trajectory, with persistent ctDNA positivity predicting inferior outcomes.

Furthermore, the analysis actually paired samples at disease relapse, and they revealed molecular markers that may have been responsible for the progression with acquired resistance alterations in tumor suppressors like TP53, as well as other markers that are responsible for DNA damage repair pathways. What was even more interesting, though, about this study was that there was additional paired CSF sampling, and they actually showed that the CNS was a sanctuary site harboring distinct molecular driver mutations that differed from plasma. This really shows that longitudinal monitoring of ctDNA from both plasma and other specimens, such as CSF, can really help characterize a patient's cancer, and we can look at them through the trajectory of their disease as they're receiving treatment.

Another poster that was presented highlighted the use of serial monitoring in biliary tract cancer patients with alterations in the ERBB2 gene. ERBB2 is the gene that encodes for HER2, right? And the analysis included 21 patients who received 2 or more serial samples for molecular profiling. The authors found that over the course of systemic treatment, 7 patients who received anti-HER2-based therapy lost that ERBB2 amplification over time, showing that this biomarker wasn't really static. It changed posttreatment and probably induced some sort of resistance for these patients.

Longitudinal profiling also reported alterations in bypass pathways, as well as MET amplifications in patients who experienced disease progression, supporting the potential value of this longitudinal profiling to better understand resistance mechanisms and possibly help select future treatments, mostly within the realm of a clinical trial.

And then lastly, there was an update to SERENA-6. This is a published trial now. This is a phase 3 trial for patients who have hormone receptor-positive and HER2-negative advanced breast cancer who were receiving first-line aromatase inhibitors with a CDK4/6 inhibitor. Patients were monitored with ctDNA every 2 to 3 months, and upon detection of an ESR1 mutation—and ESR1 mutations induce resistance to aromatase inhibitors—if they saw that the ctDNA had this ESR1 mutation and patients had no radiographic progression, they were randomized to continue that aromatase inhibitor or switch to camizestrant. 

Camizestrant is an oral selective estrogen receptor degrader, or SERD, that has activity against ESR1. Upon switching to camizestrant, the updated ASCO data showed that there was a benefit for these patients for progression-free survival 2, or PFS2, which is defined as the time from randomization to earliest disease progression after first subsequent therapy, and it's almost thought to be a surrogate marker for overall survival. This actually led to the FDA approval of camizestrant in this setting, and it is probably the best data that we have for longitudinal ctDNA monitoring and this influencing treatment decisions in oncology. Very exciting.

APPs can educate patients on what the results of the ctDNA biomarker analysis are in the context of their disease by breaking down exactly what is found on the report and making sure the patient understands these findings. It's important that patients know that there are biomarkers that will show up that may not hold any clinical significance. And it's crucial to let patients know that because even if that report states that something is actionable, it might not necessarily be appropriate given the primary site of their disease or concurrent alterations that are picked up on these assays.

Again, it's really important that you collaborate with your pathologist, making sure you're reading them appropriately, but just making sure that you're talking in very patient-friendly language and giving them all the information possible about their report. Breaking this down and ensuring patients know where we are in the field is essential when discussing prognostic biomarkers, especially in the context of changing treatment, letting patients know that a lot of the data regarding ctDNA and MRD assessment are more a way to see how closely patients need to be monitored rather than prematurely stopping a treatment without sufficient evidence to do so. It also gives us ideas on how we may treat them when and if their cancer comes back. Having this information may not be beneficial now, but it could be beneficial later on.

And I also think stressing that a lot of this is still investigational. This lets them know that we are still providing the best care possible for their disease and that we're following guidelines. But we should note that these tests are aiming to improve their outcomes, and ongoing studies are happening that will give us a better idea of how to interpret these findings and eventually incorporate them into clinical care.

One of the biggest ones is assuming a genomic alteration is actionable when in fact it might not be. A classic example of this is that we do have some biomarker-approved agents, such as NTRK inhibitors. It's a pan-solid tumor approval for a specific genetic abnormality. There are 3 small-molecule inhibitors that are approved for solid tumors that harbor an NTRK fusion, not a mutation. Sometimes genomic reports may pick up an NTRK mutation that can be mistaken as a fusion, and its clinical relevance is unknown. Those patients who don't harbor that fusion may not be candidates for this targeted NTRK inhibitor, given that they have a different alteration. And this actually holds true for some fusions that maybe we just don't have enough data yet to elucidate if they're drivers or not. It really stresses the need for more collaboration among health care professionals.

And knowing if certain biomarkers induce resistance to targeted therapies is also important. I think that can be really important in terms of a patient-specific disease. These concepts are really important when we're reviewing molecular tests, ensuring that we are giving patients the correct information about their disease and why it may or may not work. Going over examples like the one I just provided may help a patient understand misconceptions in precision medicine, just because precision medicine isn't necessarily one-size-fits-all.

APPs can help patients understand a lot of these concepts by letting them know that cancer is very dynamic. And sometimes we need more than one test to figure out what the best options are for their disease. Repeat testing can be valuable if prior results were negative to ensure nothing is being missed.

Furthermore, certain molecular tests only look at germline alterations, and some of these findings can actually help dictate treatment and give patients an idea if genetic counseling is indicated for them or their family members.

Lastly, clinical trials are constantly evolving, and many are enrolling both treatment-naive and treatment-experienced patients, so letting these patients know that these tests are on the forefront of oncology care and that clinical trials are changing outcomes for them every day is extremely important.

And APPs, I think, are on the front lines having these conversations with patients very frequently. I know in my phase 1 clinic where I work, APPs have informational visits with patients, and they actually review their pathology and explain why they may be eligible for certain clinical trials based on their molecular findings. This really shows that they play a vital role in educating patients about their options.

What biomarker information do we need for this patient at this point in the disease course? This is important to note that biomarkers are dynamic. They may change. You have to make sure that you're asking yourself this question constantly. Are tissue, plasma, or both appropriate? And this matters. Just because something's not being picked up on a tissue-based NGS platform doesn't mean nothing will show up on plasma. It's always good to recognize that.

Is the result prognostic, predictive, or directly actionable? Sometimes this requires collaboration with fellow health care professionals because sometimes this is not very straightforward.

Always good to read your molecular report as well as collaborate with your fellow medical oncologists and pathologists.

Does this result change treatment today, or is this clinical utility still being investigated? This is important, as you will likely get asked this by patients, and they want to know what the meaning of the results is. Just knowing that a lot of these data are still premature, and we're still getting more prospective data out of this. And how can I explain the result and its limitations to the patient in understandable terms? Always remember to use simple words, slowly explain to them their results.

I really think that APPs are doing a great job and have the compassion and knowledge to relay these complicated concepts to patients that will make them feel both comfortable and informed about their results.