CONTACT

American Association for Cancer Research Congress 2025

Key Takeaways from AACR Annual Meeting 2025

The American Association for Cancer Research (AACR) Annual Meeting remains a cornerstone event, and the 2025 gathering, attended by over 20,000 individuals with more than 7,000 abstracts submitted, underscored the dynamic and rapidly evolving landscape of cancer science and medicine. The meeting highlighted significant advancements across the spectrum of cancer research, from innovative prevention strategies and sophisticated early detection methods to groundbreaking therapeutic approaches.

Clinical Studies Updates

Immunotherapy Advances into Earlier Disease Stages

A prominent trend was the expanding role of IO in treating earlier stages of cancer. Noteworthy studies demonstrated the significant potential of neoadjuvant PD-1 blockade with ICIs such as dostarlimab, (Jemperli from GSK)) in mismatch repair-deficient (dMMR) solid tumors, achieving impressive complete response rates that could potentially spare patients from surgery. Furthermore, the KEYNOTE-689 trial provided further evidence for the benefit of IO beyond metastatic disease, confirming that perioperative pembrolizumab substantially improves outcomes for patients with locally advanced head and neck squamous cell carcinoma (HNSCC).

Immunotherapy in mismatch repair-deficient solid tumors

Two presentations from Memorial Sloan Kettering at AACR 2025 highlighted the promise of PD-1 immunotherapy in patients with DNA mismatch repair-deficient (dMMR) solid tumors. Dr. Andrea Cercek reported that all 49 rectal cancer patients treated with dostarlimab achieved complete responses, avoiding surgery. Notably, 65% of patients with non-rectal dMMR tumors also had complete responses, with 85% remaining recurrence-free at two years (data published at NEJM).

Dr. Yelena Janjigian presented a trial showing that pembrolizumab cleared ctDNA, in 11 of 13 post-surgical dMMR patients, with most staying relapse-free. These studies suggest PD-1 blockade may enable clearance of MRD before clinical recurrence highlighting ctDNA as a dynamic tool to guide treatment in early-stage dMMR cancers.

Combining PARP and IO inhibitors

Dr. Timothy A. Yap of MD Anderson Cancer Center presented results from the phase II KEYLYNK-007 trial, which evaluated pembrolizumab plus the PARP inhibitor olaparib (Lynparza) in in 332 patients with advanced solid tumors across 30 cancer types, with tumors harboring homologous recombination repair mutations (HRRm) or homologous recombination deficiency (HRD). This tumor-agnostic study showed encouraging antitumor activity, especially in patients with BRCA1/2 mutations, by targeting DNA repair pathways critical to these cancers. In response to a question on underdiagnosis of HRRm/HRD, Yap emphasized the need for improved assays that better reflect current biological understanding to ensure accurate patient identification.

Need for better predictive biomarkers in TNBC

Dr. Marleen Kok from the Netherlands Cancer Institute highlighted the clinical challenges of precision immunotherapy in triple-negative breast cancer (TNBC), the most inflamed breast cancer subtype. While adding pembrolizumab to chemotherapy modestly (5%) improves survival, lack of predictive biomarkers limits clinicians’ ability to identify which patients will benefit. Early trial findings presented, suggest that high tumor-infiltrating lymphocytes (TILs) predict better responses to anti-PD-1 therapy, even without chemotherapy. However, in metastatic TNBC, most patients who initially respond eventually develop secondary resistance. Dr Kok emphasized the need for better biomarker studies, noting that most trials only collect baseline biopsies, making it difficult to study mechanisms of resistance in real-world patients.

Runimotamab + Trastuzumab Shows Promise in HER2+ Breast Cancer

Dr. Shanu Modi (MSKCC) shared results from a Phase Ia/b trial of runimotamab, a bispecific antibody targeting HER2 and CD3, in HER2-positive breast cancer. While monotherapy showed no activity, combination with trastuzumab led to tumor regression in ~30% of patients and showed improved safety and tolerability.

Targeting KRAS G12D

Phase I trial has shown promising activity of zoldonrasib against KRAS G12D mutations, particularly in NSCLC. It is a selective, covalent inhibitor targeting KRAS G12D in its active (RAS[ON]) state. This mutation is found in ~40% of pancreatic cancers, ~15% of colorectal cancers, and ~4% of NSCLCs (often in nonsmokers) with low TMB and PD-L1 expression. In the trial (n=211), 18 NSCLC patients showed a 61% objective response rate and 89% disease control rate. For pancreatic cancer patients the response rate was 30%, with an 80% disease control rate.

Zongertinib Shows Durable Responses in HER2-Mutant NSCLC

Data was presented the Phase Ia/Ib Beamion LUNG-1 trial evaluating zongertinib, a HER2-selective oral tyrosine kinase inhibitor, in patients with advanced NSCLC harboring HER2 mutations (~2–4% of cases). In the primary cohort (n=75) with HER2 tyrosine kinase domain (TKD) mutations, zongertinib achieved an objective response rate (ORR) of 71%, a disease control rate (DCR) of 96%, a median progression-free survival (PFS) of 12.4 months, and a median duration of response (DoR) of 14.1 months. In exploratory cohorts, ORRs were 30% (DCR 65%) in patients with non-TKD HER2 mutations (n=20) and 48% (DCR 97%) in patients with TKD mutations previously treated with HER2-targeted ADCs such as trastuzumab deruxtecan (T-DXd) (n=31). Zongertinib showed activity even in those who had progressed on prior HER2-ADC therapy. Importantly, it demonstrated a favorable safety profile, with grade ≥3 adverse events reported in 17%, 25%, and 3% of patients in cohorts 1, 3, and 5, respectively, and no cases of interstitial lung disease. Zongertinib is currently under FDA review with Fast Track and Breakthrough Therapy designations.

Liquid Biopsies Demonstrate Increasing Clinical Utility in Diagnosis and Monitoring.

The VICTORI trial (Personalis) that used the tumor-informed NeXT Personal® ctDNA assay to detect MRD in 71 CRC patients after surgery, showed ctDNA detected 100% of recurrences before imaging, with 87% of recurrent cases testing positive post-surgery. Lower ctDNA levels (<100 ppm) were linked to longer lead times to recurrence, while ctDNA negativity correlated with no recurrence to date. Week 4 post-surgery emerged as the optimal time for ctDNA testing due to cfDNA fluctuations.

ROME trial results demonstrated that patients with advanced solid tumors had significantly better outcomes when treated with therapies guided by concordant genomic alterations detected in both tissue and liquid biopsies. Among 1,794 enrolled patients, 400 had actionable mutations, with 49.2% (n=197) showing concordant results across both biopsy types (T+L group). In this group, tailored therapy led to superior outcomes compared to standard of care, with a median overall survival (OS) of 11.05 vs. 7.7 months and median progression-free survival (PFS) of 4.93 vs. 2.8 months. The 12-month OS rate was 47.8% vs. 38.8%, and objective response rates were 20% vs. 11.8%, respectively. Outcomes were less favorable in patients with discordant biopsy findings, particularly those with alterations identified by liquid biopsy alone (OS: 4.05 months). The study underscores the value of integrating both biopsy modalities to better capture tumor heterogeneity and guide personalized treatment strategies. Discordance was attributed to biological variability, detection limitations, and test failures. Further validation using serial, integrated profiling is underway to refine diagnostic pathways in precision oncology.

New drugs

Precision KRAS Inhibitors

RMC-5127 (Revolution Medicines): KRAS G12V-selective ON inhibitor showed durable tumor regression in preclinical models; moving toward clinical trials.
AZD0022 (AstraZeneca): KRAS G12D ON/OFF inhibitor in Phase I/IIa (ALAFOSS-01) for pancreatic, CRC, and NSCLC, with EGFR combo cohort.
AMG 410 (Amgen): Pan-KRAS ON/OFF inhibitor active in preclinical pancreatic models; IND submitted, clinical trial planned.

Reprogramming the Tumor Microenvironment

ABP-102/CT-P72 (Abpro): HER2/CD3 bispecific T-cell engager with strong preclinical activity and low toxicity.
M0324 (Merck): Conditional CD40 agonist activated by MUC1; superior to other CD40 antibodies in preclinical testing.
BMS-986449 (BMS): Small molecule degrades Helios/Eos to reprogram Tregs; in Phase I with nivolumab.
FXX489 (Novartis): FAP-targeting radioligand therapy; in Phase I for multiple solid tumors.
RO7567132 (Roche): Bispecific antibody (FAP + LTBR) turns “cold” tumors “hot”; in Phase I with atezolizumab combo.

Prostate Cancer Dual-Targeting Therapies

GDC-2992 (Genentech): Bifunctional androgen receptor inhibitor; Phase I trial underway.
ABBV-969 (AbbVie): Antibody-drug conjugate targets PSMA and STEAP1; Phase I in metastatic castrate-resistant prostate cancer.

Novel Approaches in MSI and Liver Cancer

GSK4418959 (GSK): WRN inhibitor for MSI-high tumors; active in resistant models with minimal toxicity to MSS cells.
BAY 3547926 (Bayer): Alpha-emitting radiotherapy for GPC3-positive liver cancer; Phase I BANTAM-01 enrolling.

Industry announcements during AACR 2025

Tempus presented 18 abstracts, including an oral presentation, highlighting the transformative impact of AI on oncology treatment and patient outcomes. Their research emphasized the integration of AI in advancing precision medicine.

Natera presented in total data from 8 studies at AACR 2025, focusing on the utility of their Signatera™ MRD test across various cancers, including colorectal, breast, and gynecological cancers. The studies underscored Signatera’s role in immunotherapy response prediction and genomic landscaping. Backed by a large clinical validation study across five cancer types (to be presented at ASCO 2025), the assay identified ctDNA well before clinical recurrence and stratified postsurgical patients by recurrence risk. Additionally, Natera has launched its ultra-sensitive Signatera Genome assay nationwide in the U.S. for detecting MRD across multiple cancers. Using whole-genome tumor profiling and mPCR-NGS technology, the test detects ctDNA at levels as low as 1 part per million.

Personalis highlighted data from the VICTORI Trial aiming to determine the optimal timepoint for detecting MRD using their ultrasensitive ctDNA-based assay after surgery in patients with resectable colorectal cancer. 100% of treatment-naive patients with stage II or higher disease had detectable ctDNA before surgery, demonstrating that ctDNA can reliably detect MRD and predict recurrence earlier than imaging, even in hard-to-detect metastatic sites (e.g., lungs). Prognostic value of ctDNA was evident as early as 2 weeks post-surgery, thereby generating more findings supporting early postoperative MRD testing to guide adjuvant therapy decisions.

Guardant Health presented 18 posters and among those, data on their Shield™ multi-cancer detection (MCD) test, demonstrating high specificity and clinically meaningful sensitivity across 10 tumor types. The test’s performance informed its selection for the National Cancer Institute’s Vanguard Study.

GRAIL presented updated results from the PATHFINDER 2 study, reinforcing the clinical feasibility of MCED in real-world settings and has shared new data from a real-world study involving over 100,000 participants using their Galleri® test. The findings affirmed the test’s capability to detect cancer signals across a wide range of cancer types, including those without recommended screening protocols.

Precede Biosciences presented compelling data demonstrating their platform’s ability to resolve gene expression for thousands of tumor-specific genes from a simple blood test. Their comprehensive epigenomic platform enables systematic and quantitative prediction of tumor gene expression from cell-free DNA, including key targets for next-generation precision medicines and master regulators of transcription. These findings underscore the platform’s potential to inform therapeutic decisions and elucidate mechanisms of response and resistance from only 1mL of plasma.

Other companies that caught my attention were:

1Cell.Ai a precision oncology company specializing in liquid biopsy diagnostics. It is developing first-in-class, circulating tumor cell (CTC)-based single-cell multi-omics lab-developed tests (LDTs) designed for use in clinical cancer care and early-phase clinical trials in collaboration with biotech and pharmaceutical partners.

Imagene presented an AI-based model that predicts early distant recurrence in breast cancer patients using H&E-stained pathology slides. Validated on TCGA data (n=1,038), the model stratified patients into high- and low-risk groups, revealing 299 differentially expressed genes. Low-risk patients showed enrichment in estrogen signaling pathways, and PTPRT emerged as a novel prognostic marker, with lower expression linked to poorer survival outcomes. These findings highlight the potential of integrating AI-driven pathology with transcriptomics to improve breast cancer risk stratification and guide personalized treatment.

Quantum-Si who is developing next-generation protein sequencing (NGPS) using a chip-based, single-molecule approach. Their Platinum sequencer enables digital, label-free protein analysis with high sensitivity, detecting isoforms and post-translational modifications. The goal is to make proteomics more accessible and scalable, with applications in biomarker discovery and precision medicine.

At the end of this document you’ll find selected posters from companies mentioned here and few others.

Technological advancements

Minicolons

Matthias P. Lütolf, PhD, from the Roche Institute of Human Biology, has developed engineered organoid models called «minicolons» that structurally mimic intestinal crypts, enabling them to function like real intestines. Using optogenetics, his team introduced cancer mutations to study tumor development in real time. These patient-derived minicolons have also been used for drug screening.

Enhanced CAR-T cells therapy

Dr. Marcela V. Maus of Mass General Hospital presented innovative strategies to enhance CAR T-cell therapy by using genetic editing. Her team knocked out the IFNγ receptor in CAR T cells to reduce inflammation and immune checkpoint activation, leading to stronger antitumor responses and improved survival in mice.

Slowing down ctDNA clearance

Limited ctDNA levels are one of the major technical challenges faced and in addition to assay optimization (increasing the number of SNVs, type of alterations assessed), there are limits to what assay technologies on its own can acthive. Data was presented on priming agents developed by Amplifyer Bio which attenuates (slow down) ctDNA clearance and thereby increase cell-free DNA (cfDNA).

Cell free Fragmentomics

Predicine presented data demonstrating the use of cfDNA fragmentomics for noninvasive cancer subtype classification in prostate and breast cancers (Poster #5884). Using low-pass whole-genome sequencing and nucleosome profiling at transcription factor binding sites (TFBS), the study applied the Griffin framework to distinguish between tumor subtypes such as AR-positive vs. neuroendocrine prostate cancer and ER-positive vs. triple-negative breast cancer. The minimum tumor fraction required for reliable detection was 10% for AR and ER markers, and up to 20% for ASCL1. Integrating DNA methylation data further enhanced sensitivity and reduced the limit of detection.

DELFI (Poster #1961) and collaborators presented data showing how cfDNA fragmentome analysis can classify lung cancer subtypes and predict clinical outcomes in both U.S. and European patient cohorts. Using the LEMA cohort, the study evaluated over 570 lung cancer patients across multiple histologies (ADC, SCC, SCLC) and stages. Key findings showed that DELFI scores, which quantify genome-wide fragmentation patterns (and higher scores are observed with increasing disease burden) were significantly lower in adenocarcinoma (ADC) compared to squamous cell carcinoma (SCC) and small-cell lung cancer (SCLC), and increased with disease stage. Importantly, DELFI scores were independent of mutation status, comorbidities, and therapies, supporting their robustness as prognostic biomarkers. Machine learning models trained on TCGA-derived chromosomal alterations achieved high accuracy in distinguishing lung cancer subtypes from cfDNA, with an AUC of 0.91 for classifying ADC vs. SCC. The study concludes that cfDNA fragmentome profiling offers a powerful, noninvasive tool for subtyping, prognosis, and potentially guiding personalized treatment in lung cancer.

ctDNA Detection Using Low-Volume Blood Collection

Natera presented a study (Poster #5877) evaluating the feasibility of detecting ctDNA using the Tasso+ low-volume blood collection device, which allows for at-home sampling with just 250–500 µL of capillary blood. The aim was to assess whether ctDNA analysis from small-volume samples could match standard venipuncture-based methods. The technical feasibility study demonstrated that the Signatera™ RUO assay could detect SNVs down to the level of a single mutant molecule, even in volumes as low as 250 µL. In the clinical study, blood samples from 7 patients with advanced solid tumors were collected via both venipuncture and Tasso+ devices. ctDNA positivity was observed in 6 of 7 patients (86%) using all volumes, and in 5 of 7 (71.4%) using Tasso+ samples, with consistent mutation detection across plasma input volumes.

Predicting resistance to chemotherapy

Geoff Macintyre (CNIO, Madrid) showed how genomic signatures such as chromosomal instability (CIN) can predict resistance to cytotoxic chemotherapy across various cancer types, including NSCLC. The study introduced a set of CIN-based biomarkers capable of identifying patients less likely to respond to platinum- and taxane-based chemotherapies. These biomarkers, derived from WGS data, offer a potential tool for tailoring treatment strategies and improving patient outcomes.

Tumor Biology

Tumor heterogeneity

Dr. Yardena Samuels (Weizmann Institute) presented data on how tumor heterogeneity impacts immune response and therapy resistance. Her team found that homogeneous tumors are more resistant and identified migration inhibitory factor (MIF) as a driver of immunosuppression, even when expressed in only some clones. She also explored targeting immunopeptidome heterogeneity and introduced a strategy using TYW2 knockout to induce aberrant proteins that trigger strong antitumor immunity. Patient data supports TYW2 as a potential immunotherapy target.

Extrachromosomal DNA Emerges as a Key Driver of Oncogene Amplification and Poor Outcomes.

Extrachromosomal DNA (ecDNA) plays a significant role in cancer development and progression. Tumors containing ecDNA are frequently linked to treatment resistance, metastatic potential, and worse clinical outcomes. UK’s Gel-programme showed that up to 17% of all cancers harbor ecDNA, which acts as a key driver of oncogene amplification across multiple tumor types.

Dormancy and Reactivation: The Role of Fibrotic Scars in Shaping Tumor Behavior Within the TME

Johanna Joyce presented compelling findings on the role of fibrotic scars that develop following radiation therapy and their strong association with tumor recurrence. These scars tend to form at sites where recurrences eventually occur. Tumor cells residing within these fibrotic regions often remain dormant, highlighting a complex and dynamic interplay within the tumor microenvironment (TME). When these cells exit the fibrotic niche, they may undergo epithelial-mesenchymal transition (EMT), which can reignite proliferative activity.

This observation underscores the adaptive nature of the TME and points to fibrotic scars as potential reservoirs of dormant cancer cells. From a therapeutic standpoint, these scars appear to form immune-privileged niches that shelter residual tumor cells. Early preclinical data suggest that targeting and blocking fibrosis can improve treatment outcomes.

However, many critical questions remain. What are the molecular mechanisms that maintain cancer cell dormancy within fibrotic niches? And what triggers the switch from dormancy to active proliferation? Addressing these questions could open new avenues for therapeutic intervention aimed at preventing tumor relapse.

Minimal Residual Disease

A dominant trend was the continued reliance on tumor-informed MRD approaches, which leverage patient-specific genomic alterations to monitor ctDNA with high sensitivity. These methods, such as WES / WGS-guided bespoke panels and fixed mutation panels, were highlighted for their ability to detect MRD at ultra-low variant allele frequencies (<0.01%).

Simultaneously, tumor-uninformed, methylation-based and fragmentomic MRD approaches are gaining momentum. Adela’s blood-only methylome-enrichment assay in early-stage NSCLC (Poster #3249) demonstrated recurrence prediction up to 36 months in advance, with a mean lead time of 16.6 months—supporting broader accessibility and pan-cancer applicability.

Technological innovation continues to enhance MRD sensitivity and clinical value. AI-enhanced platforms integrating SNVs, CNVs, INDELs, and methylation signatures were shown to improve detection accuracy. The MAESTRO platform (Poster #3251) demonstrated detection of ultralow ctDNA levels post-surgery and strong correlation between MRD positivity and both relapse risk and overall survival, neck squamous cell carcinoma (HNSCC), often outperforming conventional imaging.

Radio-genomics

Max Diehn presented interesting data on integration of ctDNA analysis and radiomics for dynamic risk assessment in localized lung cancer study of 418 NSCLC patients undergoing chemoradiotherapy (CRT), where they developed and validated a dynamic risk model that integrates mid-treatment ctDNA levels, radiomic features, and other clinical risk factors to predict progression-free survival. While ctDNA is a known prognostic marker post-CRT, this study highlights its value during treatment, showing that mid-CRT ctDNA levels are predictive of disease progression. To enable tissue-free monitoring, the team optimized variant calling from plasma while accounting for clonal hematopoiesis. When combined with pre-CRT imaging data, the integrated model improved risk stratification over individual biomarkers alone. This study provides one of the first evidence that combining liquid biopsy and radiomic analysis improves prognostication in CRT-treated NSCLC. The model could support personalized, response-adapted therapies, potentially reducing toxicity and enhancing patient outcomes.

Researchers from Columbia University Irving Medical Center presented a study highlighting the potential of non-invasive radiogenomic phenotyping to enhance the prediction of tyrosine kinase inhibitor (TKI) response in NSCLC. By integrating advanced imaging features with genomic data, the team developed a multimodal machine learning model capable of predicting resistance to osimertinib, a third-generation EGFR-TKI, in late-stage NSCLC patients harboring activating EGFR mutations. The model achieved a concordance index (c-index) of 0.82, outperforming single-modality models and underscoring the value of combining radiomic and genomic data for personalized treatment. The findings from Columbia University suggest that radiogenomic phenotyping could serve as a valuable tool in clinical decision-making, enabling more accurate predictions of TKI response and potentially guiding more effective treatment plans for NSCLC patients.

Selected posters from some of the leading cancer genomics providers

Company	Assay Used	Poster Title & Number	Key Findings
AccuraGen Inc.	AccuScan	AccuScan: An Ultra-sensitive and Robust MRD Detection using WGS with Single-read Error Correction Poster #2023	Uses WGS with single-read error correction. Demonstrated high analytical sensitivity and specificity at low VAFs (10⁻⁴ to 10⁻⁶). Kaplan-Meier analysis showed MRD status predicts disease-free survival. Can function WBC-free with preserved sensitivity.
BGI Genomics / National Cancer Center China	WES-based personalized MRD monitoring	MRD as a predictive biomarker in patients with small cell lung cancer Poster #5909	MRD status predicted outcomes more accurately than traditional biomarkers; supported early intervention.
Burning Rock Dx & Fudan University	ctDNA-guided ACT (no specific name)	Circulating tumor DNA-based MRD guides treatment in mCRC with no evidence of disease Poster #4567	In 104 mCRC patients, ctDNA positivity post-treatment was strongly associated with lower recurrence-free survival (HR = 4.06). ctDNA detection preceded imaging-confirmed recurrence by a median of 7.8 months.
Burning Rock Dx & Peking University Hospital	STELLA	Validation of a Tumor-Agnostic DNA Methylation-Based Assay for Minimal Residual Disease Detection in Lung Cancer Poster #3236	Uses methylation-based ctDNA detection. LOD 0.02%, Sensitivity 32.6%. Outperformed tumor-naive mutation assay. Post-op MRD positivity associated with shorter survival (HR=3.98, p<0.001). Strong correlation with CanCatch Bespoke.
Exact Sciences	Systematic literature review (SLR) on ctDNA assays	Improving Detection of Colorectal Cancer Recurrence using serial ctDNA Measurements Poster #1933	Serial ctDNA improves sensitivity and enables earlier recurrence detection, with a lead time of up to 11.5 months compared to imaging.
Geneseeq & Kunming Medical University	Not named (targeted NGS)	Assessing MRD via Early Post-op Plasma ctDNA for Recurrence Risk in NSCLC Poster #3263	In NSCLC cohort, post-op ctDNA positivity highly predictive of recurrence (HR=2.25, p<0.001). Stratification highlighted regional mutation patterns (Xuanwei vs other areas). Positive MRD had high specificity but moderate sensitivity.
GRAIL	Galleri® MCED test	Real-world data and clinical experience from over 100,000 multi-cancer early detection testsPoster #7202	Demonstrated the test’s ability to detect cancer signals across a wide range of cancer types, including those without recommended screening, with high accuracy in predicting the cancer signal origin.
GRAIL	Galleri® MCED test	Estimated Post-Test Probabilities of Cancers For Individuals Receiving Multi-Cancer Early Detection (MCED) TestsPoster #7132	Modeling analysis indicated that individuals with a «no cancer signal detected» MCED test result have a reduced risk of cancer diagnosis for one year post-blood draw, emphasizing the importance of annual screening.
GRAIL	ctDNA-based targeted methylation assay	Promoter Methylation as a Cancer Biomarker: Insights From ctDNA-Based Targeted Methylation DataPoster #1943	Early proof-of-concept study highlighted the potential of GRAIL’s assay for detecting clinically informative promoter methylation signals in plasma samples.
GRAIL	ctDNA-based targeted methylation assay	Assessment of Cancer Subtypes Across Multiple Cancer Types Using a Circulating Tumor DNA (ctDNA)-Based Targeted Methylation AssayPoster #1947	Demonstrated that GRAIL’s assay enables subtyping across cancers using a single blood draw, eliminating the need for invasive tissue biopsies.
Guardant Health	Guardant Reveal (Guardant360 + Infinity)	Detection of Clinically Actionable Somatic Variants in Post-Operative Samples with MRD Poster #1953	In >4,000 MRD+ cases, 97.4% had actionable variants. 30%+ of samples with TF ≥0.05% had clinically actionable findings. Common variants: ESR1, PIK3CA, KRAS. Useful in CRC, breast, and lung cancers.
Guardant Health	Guardant Reveal™	Analytical validation of a tissue-free epigenomic assay for circulating tumor DNA (ctDNA)-based molecular residual disease (MRD) detection in early-stage cancer Poster #692	Validated ultra-sensitive MRD detection assay for early-stage cancer using a tissue-free, blood-based methylation approach.
Guardant Health	Guardant Infinity	Development and characterization of a negative prediction algorithm for actionable mutations utilizing genomic and epigenomic profiling in cfDNA Poster #733	Improves prediction of absence of actionable mutations; reduces false negatives.
Guardant Health	Guardant Infinity Screening	A novel methylation-based classifier to identify cancer signal of origin using blood-based testing Poster #6365	Accurately identifies tissue of origin for 12 cancer types from blood.
Guardant Health	Guardant Infinity	Inferring immune and tissue cell type contributions to cell-free DNA (cfDNA) with a DNA methylation assay Poster #1951	Deconvolutes cfDNA to determine tissue and immune cell origin.
Guardant Health	Guardant Reveal	Detection of clinically actionable somatic variants in post-operative samples from patients with molecular residual disease Poster #1953	Identifies post-op mutations to guide targeted therapy.
Guardant Health	Guardant Infinity	Blood-based mapping of the personalized tumor epigenomic landscape Poster #1955	Enables personalized, non-invasive tumor epigenetic mapping.
Guardant Health	Guardant Infinity	Analytical validation of cancer gene promoter methylation detection in cfDNA liquid biopsy assay Poster #3255	Validated promoter methylation detection in liquid biopsy.
Guardant Health	Shield MCD Test	Evaluation of a plasma cell-free DNA methylation-based multi-cancer detection test Poster #6425	Demonstrated high specificity and sensitivity for early cancer detection.
Guardant Health	Guardant Infinity	A tumor-specific, methylation-based algorithm to identify MTAP gene deletions via tissue-free circulating tumor DNA Poster #5924	Detects MTAP deletions non-invasively in blood.
Guardant Health	Guardant Infinity	Analytical validation of a robust, integrated multiomics tissue assay powered by the Guardant Infinity platform Poster #5935	High accuracy in detecting molecular alterations in tissue.
Guardant Health	Guardant Infinity	Landscape of epigenomic tumor fraction in large pan-cancer cfDNA cohort Poster #5936	Reveals tumor cfDNA fractions across 20+ cancers.
Guardant Health	Guardant Infinity	Analytical validation of a new plasma-only bioinformatics classifier to identify variants from clonal hematopoiesis (CH) in cfDNA liquid biopsy assays Poster #6603	Distinguishes tumor from clonal hematopoiesis mutations.
Guardant Health	Guardant Infinity	Non-invasive cell free DNA (cfDNA) methylation profiling for accurate proportional quantification of lung cancer subtypes Poster #1142	Accurately quantifies lung cancer subtypes non-invasively.
Guardant Health	Guardant360®	EGFR PACC mutations occur more frequently as compound mutations with better responses to EGFR TKI Poster #4594	Compound PACC mutations predict better response to EGFR TKIs.
Guardant Health	Guardant Inform™	Unveiling the potent anti-tumor activity and underlying mechanism of action of the novel pan-RAF inhibitor exarafenib in BRAF-mutated NSCLC Poster #6389	Shows strong anti-tumor activity; supports clinical trials.
Guardant Health	Guardant Reveal	Pre- and post-operative analysis of circulating tumor DNA in patients with breast cancer treated with neoadjuvant therapy using a tissue-free, methylation-based approach Poster #3360	Demonstrated utility of methylation profiling in tracking MRD dynamics and treatment response pre/post-surgery in breast cancer.
Guardant Health	Guardant Infinity	Liquid based methylation profiling for quantification of breast cancer subtypes Poster #3247	Enabled quantification and classification of breast cancer subtypes from cfDNA.
Guardant Health	Guardant Inform	Employing joint modeling to support clinical interpretation of ctDNA dynamics during the treatment of advanced colorectal cancer Poster #2481	Applied joint statistical modeling to interpret treatment response and disease progression using serial ctDNA measurements.
Illumina	WGS MRD assay	Highly sensitive detection of circulating-tumor DNA in plasma using whole genome sequencing Poster #6292	Integrates SNVs, CNVs, and indels to enhance detection; CNVs improve detection in low burden tumors; model improves LoD (~0.0006%).
IMBdx & Seoul National University Hospital	CancerDetect™	Analytical Validation of CancerDetect™ RUO Assay Poster #3240	Integrates actionable and personal mutations. LOD: 0.001% (personal) and 0.1% (actionable). High specificity. Breast cancer MRD positivity ~82–87%, gastric ~45%. Robust against interfering substances.
IMBDx / Korean institutions	CancerDetect (IMBDx); targeted sequencing	Postoperative Minimal Residual Disease Detection Using Tumor-Informed Circulating Tumor DNA Analysis in Stage II–III Gastric Cancer Poster #4566	ctDNA positivity post-op was significantly associated with recurrence-free survival; 29.6% overall ctDNA positivity; persistent positivity indicated poorest prognosis
IMBDx / Samsung Medical Center	CancerDetect™ personalized MRD assay	Perioperative Tumor-Informed ctDNA Monitoring in TNBC and HER2+ Breast Cancer Poster #3260	ctDNA positivity at surgery predicted recurrence; ctDNA-detected mutations had high AF and were cancer-relevant.
Invivoscribe, Inc.	LymphoTrack® (NGS-based)	LymphoTrack® Enterprise Software v3.0.1: A High-Throughput Software Solution for Immunogenomic Analysis Poster #6304	14x speed increase in processing; 98.6% CDR3 detection rate; supports automation and large-scale MRD analysis
LGC Clinical Diagnostics	Digital PCR, NGS	Cell-based Reference Standards for the Development and Validation of Assays for the Assessment of Minimal Residual Disease in Acute Myeloid Leukemia Poster #1904	Established reference standards for NPM1 and FLT3-ITD variants; potential to improve consistency and sensitivity of MRD assay validation
Myriad Genetics	Hybrid-capture WGS-based Tumor-informed Assay	Inclusion of INDEL somatic variants in MRD panels improves confidence in ctDNA residual disease detection Poster #6641	INDELs reduce error rates by 60–90% vs SNVs; improved MRD call confidence, especially in low TF samples
Myriad Genetics	Tumor-informed hybrid-capture NGS with optimized panel	Optimized Selection of Tumor-Informed MRD Panels Enhances Sensitivity of ctDNA Detection Poster #6639	Target prioritization improved sensitivity; top targets had 60% lower LoD than lowest-ranked targets.
Nanjing Geneseeq Technology Inc.	ShieldingUltra; ultra-deep NGS with UMI	ShieldingUltra: A novel approach for enhanced minimal residual disease detection through the integration of mutation, copy number variation, and fragmentomics Poster #4561	70–90% pre-surgical detection in pan-cancer; 96% sensitivity in longitudinal NSCLC monitoring; >99% specificity in healthy plasma
Natera	Signatera™ MRD test	Congruence of tumor-specific ctDNA and radiographic recurrence in high-risk sarcoma Poster #3342/12	Demonstrated high sensitivity and specificity of the assay in detecting recurrence in high-risk sarcoma patients.
Natera	Signatera RUO	Feasibility of ctDNA detection with a low-volume blood collection device Poster 5877	Signatera™ RUO assay could detect SNVs down to the level of a single mutant molecule, even in volumes as low as 250 µL
NYU Langone / Adela	cfMeDIP-seq (Methylome enrichment)	The development of a tissue-agnostic genome-wide methylome enrichment assay for lung cancer Poster #1949	MRD detection via cfMeDIP-seq correlates with recurrence-free survival; promising for NSCLC post-treatment monitoring.
Personalis	NeXT Personal® MRD assay	Ultrasensitive ctDNA detection with NeXT Personal® in patients with colorectal cancerPoster #1292/15	Showcased the assay’s ability to detect cancer recurrence earlier than standard methods in patients with resectable colorectal cancer, aiding in timely treatment decisions.
Personalis	NeXT Personal® MRD assay	Detection of MRD assessment with the Personalis NeXT Personal assay using MATRIX plasma-in-plasma contrived samplesPoster #6094	Analytical validation demonstrated ultra-sensitivity down to 1 PPM of ctDNA and high specificity, reinforcing the potential to detect recurrent cancer earlier.
Personalis	Next Personal Dx (NPDx)	Impact of an ultrasensitive ctDNA test on MRD detection in real world clinical patient testing Poster #4549	Achieves detection down to 1-13 ppm (ultra-low VAF); 46.6% MRD positivity; real-world application showed high sensitivity even in difficult sample contexts.
Precede Biosciences	Epigenomic platform for cfDNA analysis	Systematic and robust prediction of gene expression using comprehensive epigenomic profiling of plasma ctDNA Poster #668	Showcased the platform’s capability to predict tumor gene expression from cell-free DNA, informing therapeutic decisions and elucidating mechanisms of response and resistance from only 1mL of plasma.
Predicine	PredicineALERT	A Non-Invasive, Methylation-Based NGS Assay for Sensitive Detection and Monitoring of Bladder Cancer in Urine Poster #5905	Identified 182 bladder cancer-specific methylation markers. Achieved 97% accuracy distinguishing cancer from benign/healthy. TF estimates strongly correlated with tumor-informed MRD assay (R² > 0.80). Enabled non-invasive treatment monitoring.
Predicine & BeiGene	PredicineBEACON & WES+	Comparative Analysis of Blood- and Tissue-Informed MRD Studies in SCLC Poster #4558	62 LS-SCLC patients studied. 75.9% concordance between blood- and tissue-informed assays. MRD status concordant in 61/62 at C3D1. Strong correlation between TFs in both sample types (R² = 0.96).
Predicine, Inc.	Methylation-based: PredicineEPIC™ & PredicineALERT™	Leveraging Cell-Free DNA Methylation Biomarkers for Pancreatic Cancer Detection and Minimal Residual Disease Monitoring Poster #5880	PredicineALERT™ achieved 85–96% sensitivity in PDAC MRD detection; improved performance over traditional methods
Roche / UNC	Whole genome sequencing (WGS)	WGS of ctDNA enables MRD detection in head and neck cancer patients Poster #3262	Pre-treatment ctDNA detection at 100%; post-treatment ctDNA status linked to recurrence and pathological response.
Shenzhen People’s Hospital & Burning Rock Dx	PROPHET_lite	Comprehensive Genomic Profiling (CGP)-Informed ctDNA Analysis for Postoperative Monitoring in NSCLC (BEACON study) Poster #3322	In 122 early-stage NSCLC patients, MRD positivity predicted poorer DFS (median lead time: 295 days). PROPHET_lite had LoD of 0.01%. Longitudinal MRD monitoring provided improved stratification over landmark testing.
Tempus	AI-driven precision oncology platform	Stratification Based on PRAME Gene Expression Shows Inverse Survival Associations Among Histology Subtypes in a First-Line Non-Small Cell Lung Cancer Real-World Cohort Poster #734	This study evaluates PRAME expression as a prognostic biomarker in NSCLC patients treated with immunotherapy. Using Tempus xR RNA-seq data from 4,519 patients, PRAME levels were quantified and stratified into high and low expression groups. PRAME showed a bimodal distribution, with high expression more common in LUSC than LUAD. LUSC patients had worse survival overall, but high PRAME expressors in the IO+chemo group had improved outcomes. In LUAD, low PRAME expressors fared better. Findings suggest PRAME is a prognostic biomarker in NSCLC with histology-dependent implications.
Tempus	NA	Enhancing Clonal Hematopoiesis Variant Detection in Tumor-Normal Matched Sequencing: Using Machine Learning Poster 2388	This study presents an algorithm to improve detection of clonal hematopoiesis (CH) variants in tumor-normal matched sequencing. Using copy number data and tumor purity estimates, the Tempus xT platform distinguishes CH from germline and technical artifacts. Validation with high-depth sequencing confirmed gains in sensitivity, precision, and specificity. This approach enhances CH variant identification and may support better clinical interpretation and outcomes.
Tempus	NA	Genetic and clinical landscape of NUTM1 structural variants Poster 3389	Using Tempus’ multimodal real-world database, researchers identified 59 patients with NUT carcinoma, 81% of whom had NUTM1 fusions. An additional 106 patients had NUTM1 fusions without a NUT carcinoma diagnosis, indicating potential underdiagnosis. Fusion partners varied by cancer type, with some showing strong enrichment. Given the poor median survival (~5 months), the study suggests universal NGS testing in high-risk cancers may improve diagnostic accuracy and outcomes.
Tempus	NA	Genomic and transcriptomic mediators of resistance to antibody-drug conjugates (ADCs) in metastatic breast cancer (MBC): a comprehensive multi-center study Poster 1711	This study investigated resistance mechanisms to antibody-drug conjugates (ADCs) in metastatic breast cancer using Tempus’ multimodal real-world database. Genomic and transcriptomic data from patients treated with FDA-approved ADCs revealed no major changes in target antigens in those resistant to sacituzumab govitecan (SG) or trastuzumab deruxtecan (T-DXd). However, ERBB2 amplification and expression decreased post-trastuzumab emtansine (T-DM1). Increased efflux pump gene expression was observed in SG and T-DXd resistance. Findings suggest ADC-specific resistance patterns involving target antigen loss and efflux pump upregulation.
Tempus	NA	Metabolic and tumor immune cell landscapes are significantly different amongst KRAS mutational variants in non-small cell lung cancer Poster 2766	Using data from 5,925 NSCLC patients with KRAS alterations in the Tempus multimodal real-world database, researchers found that specific KRAS variants are linked to distinct tumor biology. KRAS G12D was associated with lower tumor mutational burden, reduced CD8+ T cell infiltration, and altered lipid metabolism—suggesting a less immunogenic tumor microenvironment. These insights may help explain variable responses to immunotherapy and underscore the need to tailor treatment by KRAS variant subtype.
Tempus	NA	ROS1 single nucleotide variants predict favorable survival outcomes on immunotherapy regimens in non-small cell lung cancer Poster 2042	This study used real-world data to identify genetic markers linked to immune checkpoint inhibitor (ICI) response in NSCLC. Across two cohorts—including Tempus’ multimodal database—researchers found that ROS1 SNVs were significantly associated with longer progression-free survival in patients receiving ICIs, but not in those on non-ICI therapies. These findings suggest ROS1 SNVs may serve as predictive biomarkers for ICI benefit in first-line NSCLC treatment.
Tempus	NA	Visium HD combined with deep-learning-based cell segmentation on H&E images yield accurate cell annotation at single-cell resolution Poster 2085	In this study, researchers trained a neural network for cell segmentation on H&E-stained NSCLC samples and mapped high-resolution (2 μm) Visium HD spatial transcriptomics data to single-cell gene counts. Cell clusters were annotated using a large language model (LLM), which showed strong concordance with pathologist-labeled cell types, including lymphocytes, cancer cells, and benign epithelium. This integrated approach demonstrates the potential of combining AI-driven cell annotation with spatial transcriptomics for precise biomarker discovery and deeper insights into the tumor microenvironment.
Tempus	NA	A longitudinal, circulating tumor molecular response biomarker as a predictor of clinical outcomes in a real-world cohort of patients with advanced solid tumors treated with tyrosine kinase inhibitors Poster 3300	This study analyzed 109 advanced cancer patients from Tempus’ multimodal real-world database to assess the prognostic value of changes in circulating tumor DNA tumor fraction (ctDNA TF) during tyrosine kinase inhibitor (TKI) therapy. Patients with decreasing ctDNA TF—defined as molecular responders—had significantly longer real-world overall survival compared to non-responders across multiple cancer types. These results suggest ctDNA TF dynamics may serve as a predictive biomarker for TKI response, aiding treatment decisions in real-world clinical settings.
Tempus	NA	A novel approach to define ctDNA molecular response to immunotherapy Poster 4630	In a Tempus pan-cancer cohort of 71 patients on ICI therapy, researchers used ctDNA tumor fraction (TF) to classify molecular responders (MRs) and non-responders (nMRs). MRs showed significantly longer real-world overall survival. Notably, patients who maintained low ctDNA TF—regardless of baseline—also had improved survival, suggesting they represent a favorable prognostic subgroup and supporting prospective validation of this threshold for early intervention.
Tempus	NA	Functional precision medicine: Uncovering high actionability in rare cancers beyond genomics Poster 4716	This study evaluated ex-vivo drug sensitivity testing using the PARIS® test in 86 patients with rare cancers. Patient-derived tumor cells were tested against customized drug panels, achieving an 85% actionability rate. One LGSOC patient had 29 months of disease stabilization with PARIS®-guided therapy. The results highlight the test’s potential to identify effective treatments, especially in rare cancers lacking standard options.
Tempus	NA	Identification of predictive biomarkers of response to ADCs by HTS of highly molecularly characterized panels of patient-derived organoids (PDOs) Poster 5993	This study developed a platform to screen antibody-drug conjugates (ADCs) using 60 patient-derived organoid (PDO) models across 10 solid tumor types. ADC-specific responses were observed, with target expression as a key driver of sensitivity. Gene expression analysis revealed additional biomarkers, highlighting the molecular complexity of ADC response and the utility of PDOs in guiding precision oncology.
Tempus	NA	Molecular subtypes identified in lung adenocarcinoma using a large-scale, real-world dataset Poster 4589	This study used Tempus’ multimodal real-world data to analyze the molecular diversity of lung adenocarcinoma (LUAD). Transcriptomic profiling revealed six distinct LUAD subtypes (C1–C6), each with unique mutations, tumor microenvironment features, and progression-free survival outcomes. The findings highlight LUAD’s heterogeneity and offer potential for subtype-specific treatment strategies.
Tempus	NA	Multi-modal Large Language Models for Metastatic Breast Cancer Prognosis Poster 6620	This study presents a novel method for cancer prognosis by converting structured data into clinical note-style narratives for analysis by large language models (LLMs). Using the Patient Chronological Note (PCN) algorithm and DistilBERT, researchers predicted overall survival in metastatic breast cancer more accurately than traditional Cox models. The LLM-Cox model identified 10 patient clusters with distinct prognostic risks, enabling improved stratification and potential discovery of treatment-specific biomarkers.
Tempus	NA	Association between LAG3 expression and immune checkpoint inhibitor (ICI) efficacy in advanced melanoma Poster 7287	This study explored LAG3 expression as a predictive biomarker for first-line ICI therapy in advanced melanoma using real-world data from the Tempus multimodal database. High LAG3 mRNA expression was associated with PD-L1 positivity and increased adaptive immune cells. LAG3 levels impacted response to aPD1 monotherapy but not to aCTLA4/PD1 combination therapy. These findings suggest LAG3 may help guide ICI selection in melanoma and warrant prospective validation.
Tempus	NA	Comprehensive Whole Genome Sequencing (WGS) Assay Provides Diagnostic Insight into Clinically Relevant Genomic Alterations Across Myeloid Malignancies Poster 7172	This study assessed whole-genome sequencing (WGS) for diagnosing myeloid malignancies in 230 patients. WGS reliably detected key genomic alterations with high concordance to standard methods. The results support WGS as a cost-effective, comprehensive diagnostic tool that could enhance personalized care, particularly in settings lacking cytogenetic testing infrastructure.
Tempus	NA	Low interferon expression coupled with 9p21 loss in non-small cell lung cancer (NSCLC) patients associated with distinct somatic landscape, altered immune population, and poorer response to immune checkpoint inhibitor (ICI) therapies Poster 7287	This study analyzed 16,947 NSCLC patients using Tempus’ multimodal real-world database to examine the impact of 9p21 deletions. Patients with CDKN2A/MTAP deletions were more likely to have low IFNK expression, which was associated with higher rates of progression on immune checkpoint inhibitors. These deletions also correlated with EGFR and ALK driver mutations, suggesting a potential link between 9p21 loss, immune resistance, and oncogenic pathways.
Tempus	NA	Stratification of Cell Therapies in Solid Tumor Organoids Using Deep Learning-Derived Imaging Metrics Poster 7446	This study used deep learning to predict the viability of patient-derived organoids (PDOs) from brightfield images, evaluating responses to 27 cell therapies across 60 PDOs from 9 cancer types. Trained on 11,000+ time-lapse images, the model showed strong concordance with viability assays, eliminating the need for fluorescent stains. It also identified interpretable phenotypic features that grouped therapies by mechanisms like infiltration and apoptosis. This scalable platform enables high-throughput, non-invasive screening of cell therapies and stratifies them by biological impact.
UCLA / EarlyDx	cfTrack-methyl	cfTrack-methyl: A Personalized Approach Using cfDNA Methylomes for Ultra-Sensitive MRD Detection Poster # 3244	Combines personal and population methylation data. Achieved AUC ~99.6% at 0.01% tumor fraction. Works in challenging contexts (e.g., liver disease). Outperforms general marker models. Enables robust detection at ultra-low ctDNA levels.
University Medical Center Utrecht / PGDx / Labcorp	WGS-based ctDNA test (Labcorp® Plasma Detect™)	MEDOCC-CrEATE trial update: Feasibility of measuring ctDNA post-surgery to guide adjuvant chemotherapy in stage II colon cancer patients Poster #9948	Feasibility and early ctDNA-guided ACT decisions demonstrated; ongoing accrual and follow-up.

Disclaimer: This table is based on AI-generated extraction and has been manually validated against AACR 2025 poster content. While care has been taken to ensure accuracy, minor errors or omissions may still be present. For confirmation and full context, please refer to the original posters or contact the respective study authors.

MRD-related posters

Poster Number	Title	Company / Institute	Assay	MRD Approach	Key Findings
2023	AccuScan: An Ultra-sensitive and Robust MRD Detection using WGS with Single-read Error Correction	AccuraGen Inc.	AccuScan	Tumor-informed	Uses WGS with single-read error correction. Demonstrated high analytical sensitivity and specificity at low VAFs (10⁻⁴ to 10⁻⁶). Kaplan-Meier analysis showed MRD status predicts disease-free survival. Can function WBC-free with preserved sensitivity.
5909	MRD as a predictive biomarker in patients with small cell lung cancer	BGI Genomics / National Cancer Center China	WES-based personalized MRD monitoring	Tumor-informed ctDNA via WES	MRD status predicted outcomes more accurately than traditional biomarkers; supported early intervention.
3561	A methylation-based, tissue-independent blood test for recurrence monitoring and treatment assessment	Burning Rock Dx	OverC™ Monitor	Tumor-uninformed	Validated a methylation-based, tissue-independent assay for recurrence monitoring across multiple cancer types. TMF correlated with disease progression. Sensitivity: LC (69.2%), CRC (81.3%), HCC (95.9%). Effective even in ultra-low ctDNA conditions.
4567	Circulating tumor DNA-based MRD guides treatment in mCRC with no evidence of disease	Burning Rock Dx & Fudan University	ctDNA-guided ACT (no specific name)	Tumor-informed	In 104 mCRC patients, ctDNA positivity post-treatment was strongly associated with lower recurrence-free survival (HR = 4.06). ctDNA detection preceded imaging-confirmed recurrence by a median of 7.8 months.
3236	Validation of a Tumor-Agnostic DNA Methylation-Based Assay for Minimal Residual Disease Detection in Lung Cancer	Burning Rock Dx & Peking University Hospital	STELLA	Tumor-uninformed	Uses methylation-based ctDNA detection. LOD 0.02%, Sensitivity 32.6%. Outperformed tumor-naive mutation assay. Post-op MRD positivity associated with shorter survival (HR=3.98, p<0.001). Strong correlation with CanCatch Bespoke.
1946	Postoperative Lymphatic Exudate in HPV-Negative Head and Neck Cancer Detects Recurrence	Droplet Biosciences	Ultra-deep sequencing of lymph cfDNA	Lymphatic cfDNA sampled 24h post-surgery	Lymph cfDNA more predictive than plasma (Se 71%, Sp 64%); detected recurrence earlier; useful for adjuvant treatment decision-making.
1933	Improving Detection of Colorectal Cancer Recurrence using serial ctDNA Measurements	Exact Sciences	Systematic literature review (SLR) on ctDNA assays	Mixed (Review)	Serial ctDNA improves sensitivity and enables earlier recurrence detection, with a lead time of up to 11.5 months compared to imaging.
3263	Assessing MRD via Early Post-op Plasma ctDNA for Recurrence Risk in NSCLC	Geneseeq & Kunming Medical University	Not named (targeted NGS)	Tumor-informed	In NSCLC cohort, post-op ctDNA positivity highly predictive of recurrence (HR=2.25, p<0.001). Stratification highlighted regional mutation patterns (Xuanwei vs other areas). Positive MRD had high specificity but moderate sensitivity.
1953	Detection of Clinically Actionable Somatic Variants in Post-Operative Samples with MRD	Guardant Health	Guardant Reveal (Guardant360 + Infinity)	Tumor-uninformed (epigenomic assay)	In >4,000 MRD+ cases, 97.4% had actionable variants. 30%+ of samples with TF ≥0.05% had clinically actionable findings. Common variants: ESR1, PIK3CA, KRAS. Useful in CRC, breast, and lung cancers.
5487	Analytical Validation of a Tissue-Free Epigenomic Assay for Circulating Tumor DNA (ctDNA)-Based Molecular Residual Disease Detection	Guardant Health	Guardant Reveal (tissue-free methylation-based)	Tumor-uninformed	Demonstrated 0.005% LoD and 100% specificity in early-stage CRC, lung, and breast cancer. Enables MRD detection without need for tumor tissue.
6292	Highly sensitive detection of circulating-tumor DNA in plasma using whole genome sequencing	Illumina	WGS MRD assay	Tumor-informed	Integrates SNVs, CNVs, and indels to enhance detection; CNVs improve detection in low burden tumors; model improves LoD (~0.0006%).
3240	Analytical Validation of CancerDetect™ RUO Assay	IMBdx & Seoul National University Hospital	CancerDetect™	Hybrid (Tumor-informed + Tumor-agnostic)	Integrates actionable and personal mutations. LOD: 0.001% (personal) and 0.1% (actionable). High specificity. Breast cancer MRD positivity ~82–87%, gastric ~45%. Robust against interfering substances.
4566	Postoperative Minimal Residual Disease Detection Using Tumor-Informed Circulating Tumor DNA Analysis in Stage II–III Gastric Cancer	IMBDx / Korean institutions	CancerDetect (IMBDx); targeted sequencing	Tumor-informed bespoke panel with serial plasma sampling post-op	ctDNA positivity post-op was significantly associated with recurrence-free survival; 29.6% overall ctDNA positivity; persistent positivity indicated poorest prognosis
3260	Perioperative Tumor-Informed ctDNA Monitoring in TNBC and HER2+ Breast Cancer	IMBDx / Samsung Medical Center	CancerDetect™ personalized MRD assay	Longitudinal ctDNA monitoring at perioperative stages	ctDNA positivity at surgery predicted recurrence; ctDNA-detected mutations had high AF and were cancer-relevant.
6304	LymphoTrack® Enterprise Software v3.0.1: A High-Throughput Software Solution for Immunogenomic Analysis	Invivoscribe, Inc.	LymphoTrack® (NGS-based)	Clonality tracking and MRD via T-cell and B-cell receptor sequencing	14x speed increase in processing; 98.6% CDR3 detection rate; supports automation and large-scale MRD analysis
3705	In silico modeling of ctDNA shedding using a novel computational pipeline which models tumorigenesis, progression, and metastasis	Jefferson University / Stanford University	Computational Simulation Model	Simulated ctDNA shedding (tumor-informed surrogate)	Simulated ctDNA VAFs statistically matched real-world data; enables in silico benchmarking and tumor modeling
1904	Cell-based Reference Standards for the Development and Validation of Assays for the Assessment of Minimal Residual Disease in Acute Myeloid Leukemia	LGC Clinical Diagnostics	Digital PCR, NGS	Cell-based standards using modified THP-1 cells with known mutations	Established reference standards for NPM1 and FLT3-ITD variants; potential to improve consistency and sensitivity of MRD assay validation
3251	Immediate postoperative minimal residual disease detection with MAESTRO predicts recurrence and survival in head and neck cancer patients	Mass General Brigham / Broad Inst.	MAESTRO-Pool	Tumor-informed	Enables early MRD detection; post-op positivity correlated with poorer survival; MRD guided risk stratification in the first 10-day post-op window.
7158	Molecular Residual Disease Detection via CAPP-Seq in PTCL Patients Undergoing Stem Cell Transplantation	MGH / Stanford Medicine	CAPP-Seq with lineage-informed MRD (LI-MRD)	Lineage-informed MRD in pre/post-SCT	LI-MRD improved specificity over standard MRD; correlated with relapse risk.
6641	Inclusion of INDEL somatic variants in MRD panels improves confidence in ctDNA residual disease detection	Myriad Genetics	Hybrid-capture WGS-based Tumor-informed Assay	Tumor-informed MRD using SNVs and INDELs	INDELs reduce error rates by 60–90% vs SNVs; improved MRD call confidence, especially in low TF samples
6639	Optimized Selection of Tumor-Informed MRD Panels Enhances Sensitivity of ctDNA Detection	Myriad Genetics	Tumor-informed hybrid-capture NGS with optimized panel	Ranked variant targets by tumor relevance, error profile, and technical features	Target prioritization improved sensitivity; top targets had 60% lower LoD than lowest-ranked targets.
4561	ShieldingUltra: A novel approach for enhanced minimal residual disease detection through the integration of mutation, copy number variation, and fragmentomics	Nanjing Geneseeq Technology Inc.	ShieldingUltra; ultra-deep NGS with UMI	Fixed-panel assay with AI-driven mutation/CNV/fragmentomics integration	70–90% pre-surgical detection in pan-cancer; 96% sensitivity in longitudinal NSCLC monitoring; >99% specificity in healthy plasma
N/A	WGS-based detection of MRD using individual-specific features	Not specified	WGS	SNV and fusion detection with statistical model	Cloud-based WGS pipeline detected MRD with LoD <1:12,800; applicable across various tumor types.
1949	The development of a tissue-agnostic genome-wide methylome enrichment assay for lung cancer	NYU Langone / Adela	cfMeDIP-seq (Methylome enrichment)	Methylation-based ctDNA detection	MRD detection via cfMeDIP-seq correlates with recurrence-free survival; promising for NSCLC post-treatment monitoring.
4549	Impact of an ultrasensitive ctDNA test on MRD detection in real world clinical patient testing	Personalis	Next Personal Dx (NPDx)	Tumor-informed	Achieves detection down to 1-13 ppm (ultra-low VAF); 46.6% MRD positivity; real-world application showed high sensitivity even in difficult sample contexts.
5905	A Non-Invasive, Methylation-Based NGS Assay for Sensitive Detection and Monitoring of Bladder Cancer in Urine	Predicine	PredicineALERT	Tumor-uninformed	Identified 182 bladder cancer-specific methylation markers. Achieved 97% accuracy distinguishing cancer from benign/healthy. TF estimates strongly correlated with tumor-informed MRD assay (R² > 0.80). Enabled non-invasive treatment monitoring.
4558	Comparative Analysis of Blood- and Tissue-Informed MRD Studies in SCLC	Predicine & BeiGene	PredicineBEACON & WES+	Both tumor-informed & blood-informed	62 LS-SCLC patients studied. 75.9% concordance between blood- and tissue-informed assays. MRD status concordant in 61/62 at C3D1. Strong correlation between TFs in both sample types (R² = 0.96).
5880	Leveraging Cell-Free DNA Methylation Biomarkers for Pancreatic Cancer Detection and Minimal Residual Disease Monitoring	Predicine, Inc.	Methylation-based: PredicineEPIC™ & PredicineALERT™	Tumor-agnostic methylation biomarker detection	PredicineALERT™ achieved 85–96% sensitivity in PDAC MRD detection; improved performance over traditional methods
1988	Detection of minimal residual disease using cell free plasma DNA to guide maintenance therapy in high grade serous ovarian cancer	Princess Margaret Cancer Centre, Toronto	Custom TP53 mutation panel	Tumor-informed	MRD detection in HGSOC using cfDNA was feasible; preceded imaging relapse in most cases, suggesting value in guiding maintenance therapy decisions.
3262	WGS of ctDNA enables MRD detection in head and neck cancer patients	Roche / UNC	Whole genome sequencing (WGS)	Mutation tracking (patient-specific variants)	Pre-treatment ctDNA detection at 100%; post-treatment ctDNA status linked to recurrence and pathological response.
2516	A qualitative research study on hematology and oncology consultants’ experiences and opinions on using MRD as a fast-track approval endpoint in early-phase clinical trials in MM	Saudi Food & Drug Authority	N/A (Qualitative study)	Not applicable	Study of 30 hematology/oncology specialists in Saudi Arabia. Found strong expert consensus supporting MRD as a valid surrogate for treatment efficacy in early-phase multiple myeloma trials. Flow cytometry favored due to accessibility; NGS favored for sensitivity.
3322	Comprehensive Genomic Profiling (CGP)-Informed ctDNA Analysis for Postoperative Monitoring in NSCLC (BEACON study)	Shenzhen People’s Hospital & Burning Rock Dx	PROPHET_lite	Tumor-informed	In 122 early-stage NSCLC patients, MRD positivity predicted poorer DFS (median lead time: 295 days). PROPHET_lite had LoD of 0.01%. Longitudinal MRD monitoring provided improved stratification over landmark testing.
3314	AML Patients Characterized By BH3 Profiling And The Response To Venetoclax-based Regimen	SUNY Downstate & MD Anderson	BH3 Profiling (not an MRD assay)	Not applicable	BH3 profiling reveals AML cell dependencies on anti-apoptotic proteins. Responders had higher cytochrome C release. Co-dependencies identified (e.g., BCL-2 & MCL-1). Supports treatment stratification rather than MRD tracking.
3244	cfTrack-methyl: A Personalized Approach Using cfDNA Methylomes for Ultra-Sensitive MRD Detection	UCLA / EarlyDx	cfTrack-methyl	Hybrid (Personal methylation + Tumor-agnostic)	Combines personal and population methylation data. Achieved AUC ~99.6% at 0.01% tumor fraction. Works in challenging contexts (e.g., liver disease). Outperforms general marker models. Enables robust detection at ultra-low ctDNA levels.
9948	MEDOCC-CrEATE trial update: Feasibility of measuring ctDNA post-surgery to guide adjuvant chemotherapy in stage II colon cancer patients	University Medical Center Utrecht / PGDx / Labcorp	WGS-based ctDNA test (Labcorp® Plasma Detect™)	Tumor-informed, post-surgical ctDNA detection	Feasibility and early ctDNA-guided ACT decisions demonstrated; ongoing accrual and follow-up.
3250	Tumor-independent monitoring of MRD in multiple myeloma using cfDNA fragmentomics	University of Toronto / UHN	WGS + fragmentomics	Fragment length, chromatin accessibility	Non-invasive monitoring; MM-specific fragmentomic signals predicted MRD status with high specificity.
1916	Oncogenic Fusion Detection for Leukemia Patients Using Droplet Digital PCR	Washington University School of Medicine	Droplet Digital PCR (ddPCR)	Personalized fusion detection via ddPCR for KMT2A rearrangements	Detected treatment failure up to 100 days earlier than standard; LOD of 1:1,000,000.