John Paul Shen retweeted
Tl;dr it's well established how to search across and systematically catalog protein sequences/structures. But this has been incredibly hard, until now, for noncoding sequences without heavy use of simple heuristics
We're excited about our contribution here to decode the dark genome with AI models of DNA and investigative AI agents. Check out the preprint here: biorxiv.org/content/10.648…
We're excited about our contribution here to decode the dark genome with AI models of DNA and investigative AI agents. Check out the preprint here: biorxiv.org/content/10.648…
1
8
49
4,544
John Paul Shen retweeted
⭐️ Open Medicine Algorithm Spotlight ⭐️
Daraxonrasib Toxicity Prevention and Management in Pancreatic Cancer from @DrShubhamPant and
@GIcancerDoc:
myopenmedicine.com/algorithms/dar…
Daraxonrasib Toxicity Prevention and Management in Pancreatic Cancer from @DrShubhamPant and
@GIcancerDoc:
myopenmedicine.com/algorithms/dar…
5
21
36
23,266
John Paul Shen retweeted
Our paper is out in @Cancer_Cell!
We started this project almost 5 years ago with the idea that to understand why some cancer patients respond to immunotherapy and others do not, we may need to look beyond the tumor before treatment. Immunotherapy is not just a test of a is a perturbation. It changes the tumor immune microenvironment.
So the question became: can the direction of that change explain response?
authors.elsevier.com/a/1niWC5TA51p1…
We started this project almost 5 years ago with the idea that to understand why some cancer patients respond to immunotherapy and others do not, we may need to look beyond the tumor before treatment. Immunotherapy is not just a test of a is a perturbation. It changes the tumor immune microenvironment.
So the question became: can the direction of that change explain response?
authors.elsevier.com/a/1niWC5TA51p1…
11
99
415
24,359
John Paul Shen retweeted
🚨 NEST is a big deal, and now fully published.
24 patients with localized CRC. BOT + BAL (CTLA4 + PD1) before surgery. 60% with major responses.
We’ve been inching toward neoadjuvant therapy in colon cancer for years.
FOxTROT tried it with chemotherapy and showed there was something there.
NICHE then showed us that neoadjuvant immunotherapy can be wildly effective in dMMR/MSI-H colon cancer (I'm still pushing to rename this to Chalabi's disease).
The real question has always been: can we make this work in pMMR/MSS disease?
NEST says yes.
Among the 22 pMMR tumors:
🔥 41% major pathologic response
🔥 32% pCR
🔥 59% had ≥50% tumor regression
And no treatment-related delays to surgery.
Chalabi's tumors (dMMR)? 100% MPR.
It’s only a phase II, but a 32% pCR rate in pMMR colon cancer with immunotherapy is kind of insane.
NEST is telling us MSS is be coming through the IO door next.
And now we have ROBBIN (the phase III) coming.
Get ready. Neoadjuvant therapy is going to become a much bigger part of how we treat colon cancer.
This is where CRC is going. 🧬🔥
aacrjournals.org/clincancerres/…
@OncoAlert @oncodaily @Onco_Nexus @TheGutOncLab
24 patients with localized CRC. BOT + BAL (CTLA4 + PD1) before surgery. 60% with major responses.
We’ve been inching toward neoadjuvant therapy in colon cancer for years.
FOxTROT tried it with chemotherapy and showed there was something there.
NICHE then showed us that neoadjuvant immunotherapy can be wildly effective in dMMR/MSI-H colon cancer (I'm still pushing to rename this to Chalabi's disease).
The real question has always been: can we make this work in pMMR/MSS disease?
NEST says yes.
Among the 22 pMMR tumors:
🔥 41% major pathologic response
🔥 32% pCR
🔥 59% had ≥50% tumor regression
And no treatment-related delays to surgery.
Chalabi's tumors (dMMR)? 100% MPR.
It’s only a phase II, but a 32% pCR rate in pMMR colon cancer with immunotherapy is kind of insane.
NEST is telling us MSS is be coming through the IO door next.
And now we have ROBBIN (the phase III) coming.
Get ready. Neoadjuvant therapy is going to become a much bigger part of how we treat colon cancer.
This is where CRC is going. 🧬🔥
aacrjournals.org/clincancerres/…
@OncoAlert @oncodaily @Onco_Nexus @TheGutOncLab
2
24
60
4,216
Exciting times!
🚨 This may be the most consequential oncology study of 2026.
Yes, on par with daraxonrasib.
At #ASCO26 we highlighted the 5-year KEYNOTE-942 data for the individualized mRNA cancer vaccine intismeran + pembrolizumab in resected high-risk melanoma.
And the numbers were 👀
RFS HR 0.51
5-year RFS: 68.8% vs 49.1%
DMFS HR 0.41
OS HR 0.47 (still immature)
Amazing data. But 157 patients. Phase IIb.
The big question: does it hold up in phase III?
Apparently, yes. 🔥
INTerpath-001 randomized 1,137 patients with resected stage IIB-IV melanoma to intismeran + pembrolizumab vs pembrolizumab alone.
At the prespecified interim analysis:
✅ RFS met
✅ DMFS met
✅ Statistically significant AND
✅ No new safety signals
We don’t have the HRs yet. So some restraint until we see the actual data (also this is a press release, so grain of salt)
But why is this such a big deal?
Because this isn’t really about melanoma.
🧬 Sequence YOUR tumor
🎯 Identify YOUR neoantigens
💉 Manufacture a vaccine specifically for YOUR cancer
🦠 Train YOUR immune system to recognize it
And now that approach has succeeded in a randomized phase III trial.
This is actually personalized.
And if this platform works across tumor types, the implications are enormous.
Daraxonrasib may change how we treat RAS-driven cancers.
INTerpath-001 may change how we think about cancer vaccines altogether.
2026 has been an amazing year for oncology advances. 🔥
merck.com/news/merck-and…
@OncoAlert @Onco_Nexus @TheGutOncLab
0
0
4
425
John Paul Shen retweeted
‼️ A tumour isn’t necessarily the same all the way through.
Different populations of cancer cells can exist within it and evolve over time, including under the pressure of treatment.
A new Nature Cancer review looks at how understanding and tracking that evolution could lead to , monitor and treat cancer.
#LungCancer research has been central to building that understanding. #LCSM
nature.com/articles/s4301…
Different populations of cancer cells can exist within it and evolve over time, including under the pressure of treatment.
A new Nature Cancer review looks at how understanding and tracking that evolution could lead to , monitor and treat cancer.
#LungCancer research has been central to building that understanding. #LCSM
nature.com/articles/s4301…
2
49
175
10,685
John Paul Shen retweeted
August 5 issue of @Nature:
3 foundational papers that substantially expand patient-derived 3D cancer models and show what they can add beyond traditional 2D cell lines.
1️⃣ HCMI: 665 models from 637 patients across 25 cancer types, including rare cancers and more diverse mor-model genomic concordance was remarkably high (97.8%). Importantly, the work also shows where models can diverge from the original tumor—through stromal loss, clonal selection, and culture-driven epigenetic changes, some of which were reversible. These models are publicly available!
2️⃣ @sangerinstitute : 256 clinically annotated organoids across 5 cancers, with genome-wide CRISPR screens in 162. The screens uncovered dependencies that may be missed in conventional models. A nice example: KRAS dependency in colorectal cancer differed by allele—G12X tumors retained EGFR dependence, whereas Q61H did not. Treatment also reshaped dependencies, highlighting how cancer evolution can create new vulnerabilities.
3️⃣ @broadinstitute DepMap: 314 NextGen models across 10 cancers, including 147 CRISPR screens. These models reproduced tumor lineage far better than traditional cell lines—particularly striking in CNS tumors (93% vs 11%). They also uncovered context-specific dependencies in PDAC and GBM and showed something important: culture conditions themselves can change which genes appear essential (@DanaFarber_GU authors) with @boehmjesse team!
Taken together, these studies move cancer dependency mapping closer to the biology of actual human tumors. The model matters—and so does the environment in which we study it.
🔅
3 foundational papers that substantially expand patient-derived 3D cancer models and show what they can add beyond traditional 2D cell lines.
1️⃣ HCMI: 665 models from 637 patients across 25 cancer types, including rare cancers and more diverse mor-model genomic concordance was remarkably high (97.8%). Importantly, the work also shows where models can diverge from the original tumor—through stromal loss, clonal selection, and culture-driven epigenetic changes, some of which were reversible. These models are publicly available!
2️⃣ @sangerinstitute : 256 clinically annotated organoids across 5 cancers, with genome-wide CRISPR screens in 162. The screens uncovered dependencies that may be missed in conventional models. A nice example: KRAS dependency in colorectal cancer differed by allele—G12X tumors retained EGFR dependence, whereas Q61H did not. Treatment also reshaped dependencies, highlighting how cancer evolution can create new vulnerabilities.
3️⃣ @broadinstitute DepMap: 314 NextGen models across 10 cancers, including 147 CRISPR screens. These models reproduced tumor lineage far better than traditional cell lines—particularly striking in CNS tumors (93% vs 11%). They also uncovered context-specific dependencies in PDAC and GBM and showed something important: culture conditions themselves can change which genes appear essential (@DanaFarber_GU authors) with @boehmjesse team!
Taken together, these studies move cancer dependency mapping closer to the biology of actual human tumors. The model matters—and so does the environment in which we study it.
🔅
5
62
207
16,187
This is big - pan KRAS coming soon!
This is the daraxonrasib news I’ve been refreshing the RevMed site for!
The NDA has been accepted for FDA review.
And before anyone says it: no, this is not “late.”
NDA packets are enormous. Clinical data, CMC, safety, labeling, subgroup analyses, QA, regulatory back-and-forth. It takes time to package a drug for FDA review, especially when the goal is not “nice phase 1 signal” but actual approval.
Now the timer starts. ⏱️
Why does this matter?
An oral RAS inhibitor with phase 3 PFS and OS benefit is a big deal. I think we've all said that enough by now.
Now the FDA gets to do what it has said it wants to do:
Get life-saving drugs into patients’ hands as quickly as possible (outside of an EAP program).
FDA clock officially started. ⏱️
ir.revmed.com/node/13261/pdf
@TheGutOncLab @OncoAlert @oncodaily @Onco_Nexus
0
0
6
761
John Paul Shen retweeted
#GI26
Maybe I’m a little biased here, given Incyte has actually moved the needle before in a rare GI cancer (and Wainberg is a former mentor). But this signal is very clear in pancreatic cancer, and it’s worth paying attention 👀🧬
Updated phase 1 data for INCB161734, a KRAS reatic ductal adenocarcinoma only.
Study design matters:
🧪 KRAS G12D mutated, locally advanced or metastatic PDAC
📊 Monotherapy dose expansion at 600 mg and 1200 mg QD
💊 Combination cohorts with gem/nab-paclitaxel and mFOLFIRINOX using INCB161734 1200 mg QD
📍 Heavily pretreated population in the monotherapy expansion
What stands out:
📈 Monotherapy ORR ~37% at 1200 mg
🧬 ctDNA responses in ~66% at 1200 mg, defined as ≥90% molecular reduction
💊 Combination appears feasible with both gem/nabP and mFOLFIRINOX
⚠️ KRAS copy number gain emerging as a resistance signal
Bottom line:
KRAS-targeted therapies will be standard of care in PDAC. The only real question is which agents get us there first. Based on early efficacy, molecular depth, and combination feasibility, this could be best-in-class. Still early phase 1, durability and sequencing will matter, but this is one I’ll be watching very closely as it moves into phase 2.
Maybe I’m a little biased here, given Incyte has actually moved the needle before in a rare GI cancer (and Wainberg is a former mentor). But this signal is very clear in pancreatic cancer, and it’s worth paying attention 👀🧬
Updated phase 1 data for INCB161734, a KRAS reatic ductal adenocarcinoma only.
Study design matters:
🧪 KRAS G12D mutated, locally advanced or metastatic PDAC
📊 Monotherapy dose expansion at 600 mg and 1200 mg QD
💊 Combination cohorts with gem/nab-paclitaxel and mFOLFIRINOX using INCB161734 1200 mg QD
📍 Heavily pretreated population in the monotherapy expansion
What stands out:
📈 Monotherapy ORR ~37% at 1200 mg
🧬 ctDNA responses in ~66% at 1200 mg, defined as ≥90% molecular reduction
💊 Combination appears feasible with both gem/nabP and mFOLFIRINOX
⚠️ KRAS copy number gain emerging as a resistance signal
Bottom line:
KRAS-targeted therapies will be standard of care in PDAC. The only real question is which agents get us there first. Based on early efficacy, molecular depth, and combination feasibility, this could be best-in-class. Still early phase 1, durability and sequencing will matter, but this is one I’ll be watching very closely as it moves into phase 2.
5
19
64
10,455
John Paul Shen retweeted
Had the chance to give the colorectal cancer update at Best of ASCO Los Angeles.
Always fun to summarize a year of CRC data in ~20 minutes while pretending that is a normal thing to ask of a human being.
4 takeaways (without getting into some of the really exciting drugs that :
🔹 MSI-H CRC in complete clinical response on PD-1
Stopping at confirmed cCR looks safe. Maintenance may add toxicity without clear benefit.
🔹 BRAF V600E MSS mCRC
Encorafenib + cetuximab + chemotherapy is now a 1L standard. FOLFIRI is a great option, and FOLFOX is reasonable too. Either way, the targeted therapy backbone is here.
🔹 ctDNA / MRD
Not fully in NCCN yet, but increasingly actionable. I escalate select stage II patients, use it to frame de-escalation discussions, and strongly favor trials when possible.
🔹 Exercise
Prescribe it like we prescribe oxaliplatin.
30 min, 3x/week, ~80% max HR improves DFS/OS and is cost-saving.
CRC is moving fast. The challenge now is not just knowing the data, but figuring out how to apply it thoughtfully in clinic.
Grateful to be part of the conversation.
@TheGutOncLab @Onco_Nexus @OncoAlert @TotalHealthConf
Always fun to summarize a year of CRC data in ~20 minutes while pretending that is a normal thing to ask of a human being.
4 takeaways (without getting into some of the really exciting drugs that :
🔹 MSI-H CRC in complete clinical response on PD-1
Stopping at confirmed cCR looks safe. Maintenance may add toxicity without clear benefit.
🔹 BRAF V600E MSS mCRC
Encorafenib + cetuximab + chemotherapy is now a 1L standard. FOLFIRI is a great option, and FOLFOX is reasonable too. Either way, the targeted therapy backbone is here.
🔹 ctDNA / MRD
Not fully in NCCN yet, but increasingly actionable. I escalate select stage II patients, use it to frame de-escalation discussions, and strongly favor trials when possible.
🔹 Exercise
Prescribe it like we prescribe oxaliplatin.
30 min, 3x/week, ~80% max HR improves DFS/OS and is cost-saving.
CRC is moving fast. The challenge now is not just knowing the data, but figuring out how to apply it thoughtfully in clinic.
Grateful to be part of the conversation.
@TheGutOncLab @Onco_Nexus @OncoAlert @TotalHealthConf
1
19
56
3,761
John Paul Shen retweeted
Check out the latest Editors' Pick: Myofibroblasts Induce Neuroplasticity to Promote Pancreatic Inflammation and Cancer Progression doi.org/10.1158/2159-8…
May’s Editors’ Picks from AACR’s 10 peer-reviewed journals include new insights into cancer progression, disparities in pediatric leukemia outcomes, and more. Visit the #AACRBlog to read this month’s featured studies: brnw.ch/21x33CF
0
4
11
1,967
John Paul Shen retweeted
Imagine this:
If CROs and academic centers would allow clinical trials activation and enrollment in the US to be modernized, the KRAS plenary would have been presented in #asco25 and we would have three more RAS trial readouts by today.
If CROs and academic centers would allow clinical trials activation and enrollment in the US to be modernized, the KRAS plenary would have been presented in #asco25 and we would have three more RAS trial readouts by today.
Accelerating time to trial activation to answer important clinical questions and improve standard of care to benefit patients more quickly - @AACI_Cancer - @asco stakeholders meeting with @FDAOncology @NCIDirector and many more! Lots of interest, time to turn that into action!
5
8
21
4,764
John Paul Shen retweeted
For readers interested in cyclin-dependent kinase inhibitors in cancer, here's a comprehensive review discussing their progression beyond CDK4/6 inhibitors for breast cancer nature.com/articles/s4157…
rdcu.be/flRPL
#ASCO26
rdcu.be/flRPL
#ASCO26
0
15
46
3,585
John Paul Shen retweeted
Presented by Dr. Pavlos Msaouel at #ASCO26, a collaborative multi-institutional study identified EGFR as a key therapeutic vulnerability in SMARCB1-deficient renal medullary carcinoma, a rare and highly aggressive kidney cancer.
Panitumumab-based therapy produced a 53.9% g complete responses, in heavily pretreated patients, supporting a promising new treatment approach for this disease. @PavlosMsaouel #EndCancer
Panitumumab-based therapy produced a 53.9% g complete responses, in heavily pretreated patients, supporting a promising new treatment approach for this disease. @PavlosMsaouel #EndCancer
0
11
28
2,757
John Paul Shen retweeted
#ASCO26
LBA3501: Optimal duration of adjuvant chemotherapy in resected stage I-IV CRC using GALAXY Japan ctDNA dynamics
GALAXY Japan is the longitudinal dataset that keeps on giving.
Now, it gets directly at the question we face all the time in clinic:
How much adjuvant
GALAXY looked at ctDNA dynamics using a personalized tumor-informed assay from the post-surgery/pre-ACT time point to mid- or post-ACT, grouping patients into sustained negative, clearance, decrease but still detectable, or rising ctDNA.
Quick hits:
📌 ctDNA dynamics were strongly prognostic
Sustained negative did best
Clearance did next best
Decrease but detectable and rising ctDNA did poorly
📈 Median DFS
Sustained negative: NR
Clearance: NR
Decrease but detectable: 5.3 months
Rising: 1.1 months
🧬 Clearance matters
If ctDNA clears, outcomes look much better than if ctDNA remains detectable.
🚨 But detectable is detectable
A falling ctDNA that remains positive is not reassuring.
Rising ctDNA at 3 months is worse.
⏱️ Duration question
Among patients with ctDNA clearance after 3 months, longer ACT did not appear to improve DFS.
✅ Sustained ctDNA-negative patients
Longer ACT also did not appear to improve DFS in patients who stayed ctDNA negative.
This is a major step toward using ctDNA dynamics as a tool during ACT, not just a binary post-op risk marker.
For my practice, ctDNA dynamics are a key piece of information to be used in concert with the complete clinical picture during adjuvant therapy discussions.
The biggest takeaway:
Post-op ctDNA is important.
But the story may matter more than the snapshot.
(Credit to @CathyEngMD who I cribbed this nice side by side image from)
@TheGutonclab @UGrewalMD @TimothyJBrownMD @OncoAlert @Onco_Nexus @ASCO
@NazliDizman @LauraAlderMD @DVAraujoMD @DrBarbiOnc @LauraEsfeller @FunchainMD @YGaritaonaindia @DrSAHaddad @jgong15 @iandresmeraz @SakditadMD @RamilaShilpakar @RohitBanwar @lungoncdoc
LBA3501: Optimal duration of adjuvant chemotherapy in resected stage I-IV CRC using GALAXY Japan ctDNA dynamics
GALAXY Japan is the longitudinal dataset that keeps on giving.
Now, it gets directly at the question we face all the time in clinic:
How much adjuvant
GALAXY looked at ctDNA dynamics using a personalized tumor-informed assay from the post-surgery/pre-ACT time point to mid- or post-ACT, grouping patients into sustained negative, clearance, decrease but still detectable, or rising ctDNA.
Quick hits:
📌 ctDNA dynamics were strongly prognostic
Sustained negative did best
Clearance did next best
Decrease but detectable and rising ctDNA did poorly
📈 Median DFS
Sustained negative: NR
Clearance: NR
Decrease but detectable: 5.3 months
Rising: 1.1 months
🧬 Clearance matters
If ctDNA clears, outcomes look much better than if ctDNA remains detectable.
🚨 But detectable is detectable
A falling ctDNA that remains positive is not reassuring.
Rising ctDNA at 3 months is worse.
⏱️ Duration question
Among patients with ctDNA clearance after 3 months, longer ACT did not appear to improve DFS.
✅ Sustained ctDNA-negative patients
Longer ACT also did not appear to improve DFS in patients who stayed ctDNA negative.
This is a major step toward using ctDNA dynamics as a tool during ACT, not just a binary post-op risk marker.
For my practice, ctDNA dynamics are a key piece of information to be used in concert with the complete clinical picture during adjuvant therapy discussions.
The biggest takeaway:
Post-op ctDNA is important.
But the story may matter more than the snapshot.
(Credit to @CathyEngMD who I cribbed this nice side by side image from)
@TheGutonclab @UGrewalMD @TimothyJBrownMD @OncoAlert @Onco_Nexus @ASCO
@NazliDizman @LauraAlderMD @DVAraujoMD @DrBarbiOnc @LauraEsfeller @FunchainMD @YGaritaonaindia @DrSAHaddad @jgong15 @iandresmeraz @SakditadMD @RamilaShilpakar @RohitBanwar @lungoncdoc
0
15
28
3,097
John Paul Shen retweeted
#ASCO26
Observation vs maintenance PD-1 after cCR in dMMR/MSI-H CRC managed nonoperatively
Abstract #3502
Presentation: May 31, 2026
This is absolutely essential data.
MSI-H CRC has rapidly evolved from aggressive and hard to treat to curable without surgery.
We are suddenly dealing with questions that would have sounded absurd a few years ago:
If a patient with dMMR/MSI-H CRC has a clinical complete response after PD-1 therapy, do we need to keep treating?
This Chinese multicenter cohort looked at exactly that.
Quick hits:
📌 Study design
318 patients with dMMR/MSI-H CRC treated with PD-1 therapy
191 achieved cCR in the presented flow 129 entered non-operative management
👀 Groups
Observation: n=66, stopped PD-1 after cCR
Maintenance: n=63, received ≥2 post-cCR PD-1 cycles
🧬 Included both rectal and colon cancer
Rectal: 58
Colon: 61
Synchronous dual primaries: 10
🔎 cCR assessment was not casual
Endoscopy + imaging: 100%
Confirmatory biopsy: 80/129, 62%
Rectal cancer: DRE 100%, pelvic MRI 89.7%
⏱️ Time to cCR
Median cycles to cCR: 8 in both groups
Median treatment duration to cCR: 6 months
69.8% reached cCR within 8 cycles
93.8% reached cCR within 16 cycles
📈 3-year DFS
Observation: 96.4%
Maintenance: 98.2%
P=0.560
🙌 3-year OS
Observation: 100%
Maintenance: 98.4%
P=0.313
🔥 Consistent sensitivity analyses
Overlap-weighted 3-year DFS: 98.5% vs 99.6%
6-week landmark 3-year DFS: 96.0% vs 98.2%
⚠️ The toxicity tradeoff matters
Maintenance did not clearly improve outcomes, but continuing PD-1 means continuing exposure, cost, visits, and immune toxicity risk.
This is the kind of dataset that changes conversations in clinic.
Not definitive. But incredibly useful.
For a patient with dMMR/MSI-H CRC who achieves a high-quality cCR and is being followed closely, observation after stopping PD-1 looks very reasonable.
The remaining question is duration.
Does it depend on colon vs rectum? Localized vs metastatic? Disease burden? Quality of cCR assessment? Patient anxiety?
Almost certainly yes.
But the big message is clear:
In dMMR/MSI-H CRC with cCR after PD-1, more treatment is not automatically better.
Sometimes the win is knowing when to stop.
Plenary-worthy.
@TheGutonclab @UGrewalMD @TimothyJBrownMD @OncoAlert @Onco_Nexus @ASCO
@NazliDizman @LauraAlderMD @DVAraujoMD @DrBarbiOnc @LauraEsfeller @FunchainMD @YGaritaonaindia @DrSAHaddad @jgong15 @iandresmeraz @SakditadMD @RamilaShilpakar @RohitBanwar @lungoncdoc
Observation vs maintenance PD-1 after cCR in dMMR/MSI-H CRC managed nonoperatively
Abstract #3502
Presentation: May 31, 2026
This is absolutely essential data.
MSI-H CRC has rapidly evolved from aggressive and hard to treat to curable without surgery.
We are suddenly dealing with questions that would have sounded absurd a few years ago:
If a patient with dMMR/MSI-H CRC has a clinical complete response after PD-1 therapy, do we need to keep treating?
This Chinese multicenter cohort looked at exactly that.
Quick hits:
📌 Study design
318 patients with dMMR/MSI-H CRC treated with PD-1 therapy
191 achieved cCR in the presented flow 129 entered non-operative management
👀 Groups
Observation: n=66, stopped PD-1 after cCR
Maintenance: n=63, received ≥2 post-cCR PD-1 cycles
🧬 Included both rectal and colon cancer
Rectal: 58
Colon: 61
Synchronous dual primaries: 10
🔎 cCR assessment was not casual
Endoscopy + imaging: 100%
Confirmatory biopsy: 80/129, 62%
Rectal cancer: DRE 100%, pelvic MRI 89.7%
⏱️ Time to cCR
Median cycles to cCR: 8 in both groups
Median treatment duration to cCR: 6 months
69.8% reached cCR within 8 cycles
93.8% reached cCR within 16 cycles
📈 3-year DFS
Observation: 96.4%
Maintenance: 98.2%
P=0.560
🙌 3-year OS
Observation: 100%
Maintenance: 98.4%
P=0.313
🔥 Consistent sensitivity analyses
Overlap-weighted 3-year DFS: 98.5% vs 99.6%
6-week landmark 3-year DFS: 96.0% vs 98.2%
⚠️ The toxicity tradeoff matters
Maintenance did not clearly improve outcomes, but continuing PD-1 means continuing exposure, cost, visits, and immune toxicity risk.
This is the kind of dataset that changes conversations in clinic.
Not definitive. But incredibly useful.
For a patient with dMMR/MSI-H CRC who achieves a high-quality cCR and is being followed closely, observation after stopping PD-1 looks very reasonable.
The remaining question is duration.
Does it depend on colon vs rectum? Localized vs metastatic? Disease burden? Quality of cCR assessment? Patient anxiety?
Almost certainly yes.
But the big message is clear:
In dMMR/MSI-H CRC with cCR after PD-1, more treatment is not automatically better.
Sometimes the win is knowing when to stop.
Plenary-worthy.
@TheGutonclab @UGrewalMD @TimothyJBrownMD @OncoAlert @Onco_Nexus @ASCO
@NazliDizman @LauraAlderMD @DVAraujoMD @DrBarbiOnc @LauraEsfeller @FunchainMD @YGaritaonaindia @DrSAHaddad @jgong15 @iandresmeraz @SakditadMD @RamilaShilpakar @RohitBanwar @lungoncdoc
0
12
21
1,581
John Paul Shen retweeted
#ASCO26
BREAKWATER: Encorafenib + cetuximab + FOLFIRI in 1L BRAF V600E-mutant mCRC
Abstract LBA3503
Presentation: May 31, 2026
BRAF V600E-mutant CRC has always been different.
More aggressive biology. Worse prognosis. Less forgiving clinical course.
For years, the could move targeted therapy earlier and make a real dent in outcomes, not just response.
BREAKWATER keeps answering that question.
This cohort evaluated encorafenib + cetuximab with FOLFIRI in the first-line setting for BRAF V600E-mutant mCRC.
Quick hits:
📌 1L BRAF V600E-mutant mCRC
Encorafenib + cetuximab + FOLFIRI vs FOLFIRI ± bevacizumab
🎯 Response
64.4% vs 39.2%
📈 Median PFS
15.2 vs 8.3 months
HR 0.44
📊 Overall survival
HR 0.56
🔥 18-month OS
72.0% vs 54.5%
This is the kind of result that matters because it is not just one endpoint moving.
Response matters in symptomatic disease.
PFS matters in aggressive biology.
OS matters always.
And here, all three are positive.
The practical piece is also important: FOLFIRI gives clinicians flexibility. For patients with neuropathy, prior oxaliplatin exposure, or concern about cumulative oxaliplatin toxicity, this helps support a fluoropyrimidine/irinotecan backbone with EGFR/BRAF targeting.
My take: for BRAF V600E-mutant mCRC, the treatment frame has changed.
This is no longer a disease where we wait for later-line targeted therapy.
The targeted backbone belongs up front.
@TheGutonclab @UGrewalMD @TimothyJBrownMD @OncoAlert @Onco_Nexus @ASCO
@NazliDizman @LauraAlderMD @DVAraujoMD @DrBarbiOnc @LauraEsfeller @FunchainMD @YGaritaonaindia @DrSAHaddad @jgong15 @iandresmeraz @SakditadMD @RamilaShilpakar @RohitBanwar @lungoncdoc
BREAKWATER: Encorafenib + cetuximab + FOLFIRI in 1L BRAF V600E-mutant mCRC
Abstract LBA3503
Presentation: May 31, 2026
BRAF V600E-mutant CRC has always been different.
More aggressive biology. Worse prognosis. Less forgiving clinical course.
For years, the could move targeted therapy earlier and make a real dent in outcomes, not just response.
BREAKWATER keeps answering that question.
This cohort evaluated encorafenib + cetuximab with FOLFIRI in the first-line setting for BRAF V600E-mutant mCRC.
Quick hits:
📌 1L BRAF V600E-mutant mCRC
Encorafenib + cetuximab + FOLFIRI vs FOLFIRI ± bevacizumab
🎯 Response
64.4% vs 39.2%
📈 Median PFS
15.2 vs 8.3 months
HR 0.44
📊 Overall survival
HR 0.56
🔥 18-month OS
72.0% vs 54.5%
This is the kind of result that matters because it is not just one endpoint moving.
Response matters in symptomatic disease.
PFS matters in aggressive biology.
OS matters always.
And here, all three are positive.
The practical piece is also important: FOLFIRI gives clinicians flexibility. For patients with neuropathy, prior oxaliplatin exposure, or concern about cumulative oxaliplatin toxicity, this helps support a fluoropyrimidine/irinotecan backbone with EGFR/BRAF targeting.
My take: for BRAF V600E-mutant mCRC, the treatment frame has changed.
This is no longer a disease where we wait for later-line targeted therapy.
The targeted backbone belongs up front.
@TheGutonclab @UGrewalMD @TimothyJBrownMD @OncoAlert @Onco_Nexus @ASCO
@NazliDizman @LauraAlderMD @DVAraujoMD @DrBarbiOnc @LauraEsfeller @FunchainMD @YGaritaonaindia @DrSAHaddad @jgong15 @iandresmeraz @SakditadMD @RamilaShilpakar @RohitBanwar @lungoncdoc
0
39
68
9,410