Precision Oncology in Practice: Evidence-Based CGP Matching & The Clinical Value of MRD by Dr Subotheni Thavaneswaran

This is a verbatim transcript of an interview conducted with Dr Dr Subotheni Thavaneswaran in July 2026. The transcript has been lightly edited for clarity.

Introduction

I’m Subo, I’m a medical oncologist who’s working in metropolitan Sydney, Australia. I treat colorectal and brain cancers and I am also involved in the relevant tumor type specific multidisciplinary team meetings. Alongside my clinical work, I’m also a researcher with an interest as well as expertise in cancer genomics and precision oncology. And I partake in a molecular tumor board.

How has Comprehensive Genomic Profiling (CGP) integration evolved in Australia recently, and how are precision oncology initiatives transforming routine practice?

Over the recent years we’ve made significant progress in Australia with regards to precision oncology initiatives. So if we look back 5 or 10 years, we started off with single gene testing. In a particular cancer type like lung cancer, we knew about EGFR mutations, but not much more. Then we evolved. The more we looked the more we found. And so CGP played a big role in what we knew about cancer types, and not just in lung cancer, but over a range of cancer types. In addition, CGP gave us information about microsatellite instability (MSI), tumor mutational burden (TMB) [1]. Biomarkers that we weren’t even aware of previously. Assays that include an RNA base gave us additional information about fusions. There was a dramatic evolution between single gene testing to comprehensive genomic profiling.

The question arises as to how do patients access this testing? Under the initiative of Professor David Thomas, we began the MoST program, which stands for the Molecular Screening and Therapeutics program. And that was the first infrastructure that was built in Australia for advanced cancer patients to access comprehensive genomic profiling across all cancer types [2]. It ran between 2017 and 2025. Close to 5000 patients gained access to comprehensive genomic profiling and relevant therapies that were linked to their molecular profile [3].

Then the second phase was the PrOSPeCT study. So PrOSPeCT stands for Precision Oncology Screening Platform Enabling Clinical Trials. Its main purpose was to increase the number of patients that could access comprehensive genomic profiling. So it went from 5000 patients to 25,000 patients, it was supported by good funding from the government and support, as well as industry partnerships [4]. The difference between MoST and PrOSPeCT was that prospect brought CGP earlier in the patient’s disease trajectory. We started to realize that patients who were refractory to standard treatments and so late in their disease trajectory, perhaps weren’t the ideal candidates for clinical trials. And so PrOSPeCT brought this access earlier in the patient’s disease process and potentially allowed us to identify biomarkers that were relevant to trials, industry led trials that were already in place and active across the country.

Going forward the best way is to have a partnership obviously to fulfill all the initiatives that all the partners want. But to then be able to provide patients with the best potential recommendations for treatment and actual access to those treatments.

From a pan-cancer perspective, how does broad CGP fundamentally change how we identify patients and enroll them into targeted clinical trials?

From my perspective, the more we look, the more we find, and knowing the full molecular profile of a tumor is always going to be beneficial. Single gene testing may identify a key driver. But knowing the whole co-occurring mutational profile of a cancer is always important in the role that it plays in prognosis, as well as the depth or durability of a response that we might expect to see with a targeted therapy.

So we have seen in the ROME trial, for instance, that tiering the evidence supporting actionability requires a look at the molecular variant in the context of the rest of the tumor and importantly, requires an incorporation of all of the histotype-specific nuances [5].

A brain cancer with a BRAF mutation doesn’t behave the same way as a melanoma with a BRAF mutation. So while there are a few genomic biomarkers that have proven themselves as pan cancer, most other genomic biomarkers confer varying biological potential based on the tumor histology [5]. So while I think that CGP is important across cancer types, the yield and the actionability varies between cancers, the ability for us to improve outcomes does require us to take into account the histotype-based differences.

From a patient perspective I think that it gives patients hope. We’re providing them with the opportunity to have their tumor individually profiled and knowing the full landscape of their particular tumor and its properties. It opens up additional avenues, potentially for therapy. So they might receive first line treatment. And access to molecular testing. And when those results come back, potentially that opens up another avenue for therapy, depending on what the genomics reveals or identifies. And I think patients are all about precision oncology now. They understand that every tumor is different. And so understanding the molecular landscape and understanding that there might be other options for them that can potentially improve their outcomes is really important. It is a great thing that we can provide for our patients.

How critical is a structured, evidence-based framework for matching genomic biomarkers to therapies in rare and refractory cancers, and how does it impact patient survival?

I can’t emphasize the importance enough of a structured, evidence based framework. I think it’s key. We saw it in the ROME trial, for instance, where they use the ESCAT framework to match the evidence to the genomic biomarker that was identified and the level of evidence corresponded well to how much benefit patients derived from that therapy [5].

So similar to that we develop TOPOGRAPH and this was recently published in JAMA Oncology earlier this year. And essentially what we identified is that the evidence supporting the strength of a molecular biomarker and our ability to target that with a drug is really important in the amount and amount of benefit that patients derive from that targeted therapy [2].

So Tiers 1 to 3 in TOPOGRAPH align with clinically active targeted therapies those levels of evidence Tiers 1 to 3 were associated with significant improvements in patient outcomes. I think it’s really important to look more closely at the Tier 3B group, because that’s where we’re sort of just simply inferring data from another tumor type and applying it to another cancer type and hoping that it works. Off label access to certain agents, just purely based on that molecular alteration or genomic variant is probably not safe based on our TOPOGRAPH publication, we’ve identified that the Tier B group, simply based on histotype-based differences, but the same genomic variation, do not realise any survival gains by accessing the matched therapy.

So Tiers 1 to 3 in TOPOGRAPH align with clinically active targeted therapies those levels of evidence Tiers 1 to 3 were associated with significant improvements in patient outcomes. I think it’s really important to look more closely at the Tier 3B group, because that’s where we’re sort of just simply inferring data from another tumor type and applying it to another cancer type and hoping that it works. Off label access to certain agents, just purely based on that molecular alteration or genomic variant is probably not safe based on our TOPOGRAPH publication, we’ve identified that the Tier B group, simply based on histotype-based differences, but the same genomic variation, do not realise any survival gains by accessing the matched therapy.

What are the limitations of current cancer recurrence monitoring, and how does Molecular Residual Disease (MRD) testing address these gaps?

A lot of my patients do sort of inquire about whether there are any other ways for us to be able to monitor their cancer and whether it truly isn’t coming back, I guess, as someone who treats colon cancer, there is sort of guidelines to tell us that after completion of their adjuvant treatment, that we should continue to monitor them on a three monthly basis, essentially with a blood test and a clinical review and imaging, like CT scans or PET scans [6]. For some patients, they have relevant tumor markers, such as CEA that might have been elevated at baseline. And so that’s something that reassures us during surveillance [7]. More than 50% of patients won’t have a biomarker such as CEA to monitor over time over their surveillance period [7]. So I could see something like MRD or ctDNA testing playing a role in perhaps identifying microscopic disease at an earlier time point in their surveillance period.

Once MRD status is identified, what are the key clinical use cases and how should this technology complement standard imaging?

The intensity of our surveillance can be altered accordingly. We know from the DYNAMIC studies that looked at stage two as well as stage three colon cancer, those studies demonstrated that we could appropriately risk stratify patients into a higher risk and lower risk population, and potentially de-escalate the amount of treatment that we gave patients [8, 9]. And so the ability to identify lower risk patients and to spare them from the toxicities of those adjuvant treatments is also potentially a benefit.

The other role that I see for my patients is in the TNT setting. So total neoadjuvant therapy for rectal cancer, for instance, they have intense upfront chemotherapy plus or minus radiotherapy [10]. And then there’s a period of time where we’re closely watching them to see what their radiological response has been and whether or not they weren’t further surgery or not. I can see ctDNA or MRD playing a key role in that, particularly because the tumor is in situ at baseline. And we could do we could potentially access a tumor informed assay that would identify the relevant genomic findings and be able to track that over time, even potentially quantify the amount of ctDNA that the tumor is shedding and whether that changes over time really will help guide the intensity of the of the MRIs and the colonoscopy or sigmoid copies that we do to continue to surveillance those patients over time.

So I think it’s important for us to understand that MRD isn’t a replacement for what we do currently. Standard surveillance still includes scans, clinical reviews and blood tests, which include tumor markers if it’s relevant to that particular patient [6, 7].