We Presented Seven Research Abstracts at AACR Annual Meeting 2026, Showcasing Breakthroughs Across Genomics, AI Diagnostics, and Liquid Biopsy 

From late-breaking TP53 research to next-generation MRD detection, our science took centre stage in San Diego 

SAN DIEGO & FOSTER CITY, CA | April 2026 :  We are proud to announce that 1Cell.ai presented seven research abstracts at the American Association for Cancer Research (AACR) Annual Meeting 2026 in San Diego, California–  one of the world’s most prestigious and widely attended oncology research conferences. Spanning genomic profiling, pharmacogenomics, AI-powered digital pathology, liquid biopsy, and minimal residual disease detection, our research presentations reflect the extraordinary breadth and depth of science being generated across our platform. 

A Statement of Our Scientific Leadership 

The AACR Annual Meeting brings together more than 20,000 cancer researchers, clinicians, and industry leaders from across the globe. Being represented with seven abstracts;  including one late-breaking clinical research presentation – is a powerful signal of our growing scientific reputation and our unwavering commitment to generating rigorous, peer-reviewed evidence that advances the field of precision oncology. 

“Our presence at AACR 2026 with seven research abstracts reflects the depth of clinical and scientific work happening across our platform,” said Mohan Uttarwar, our CEO and Co-founder. “Each abstract addresses a real, unmet clinical need- from reducing uncertainty in genomic interpretation to detecting cancer recurrence before it becomes visible. Together, they represent our commitment to making precision oncology more precise, more accessible, and more actionable for patients and oncologists worldwide.” 

Highlights From Our AACR 2026 Research Presentations 

Reducing Uncertainty in Clinical Genomics- Smarter Variant Classification 
One of the most persistent challenges in comprehensive genomic profiling is the interpretation of Variants of Uncertain Significance (VUS) – genetic mutations that cannot be immediately classified as harmful or benign. Our team presented a proprietary machine learning-based automated classification system that integrates multiple lines of evidence, including in silico analysis, evolutionary conservation scores, functional RNA studies, and genomic co-findings to systematically reclassify VUS with high clinical accuracy. In multiple cases, variants initially flagged as uncertain were reclassified as likely pathogenic- directly changing clinical management decisions. 
Read the full Abstract


Pharmacogenomics – Predicting Who Will Be Harmed by Chemotherapy 
Our retrospective pharmacogenomic analysis of 124 pan-cancer patients demonstrated that our OncoIndx® CGP panel can reliably identify patients at high risk of severe chemotherapy toxicity before treatment begins – including critical DPYD, TYMS, and UGT1A1 variants that predict adverse reactions to 5-FU and irinotecan-based regimens. Notably, our study also identified a significant prevalence of the DPYD IVS14+1G; A variant in Uzbek patient populations – a finding with major implications for pre-treatment risk stratification in underserved patient groups. 
Read the full Abstract


Late-Breaking Research- Precision Profiling of TP53 in 1,187 Patients 
In one of the most eagerly anticipated sessions of the meeting, our team presented late-breaking clinical research on the largest real-world TP53 mutation analysis in our dataset, spanning 1,187 patients with breast, colorectal, and lung cancers. Our findings demonstrated that not all TP53 mutations behave the same- gain-of-function mutations were associated with higher TMB, greater PD-L1 expression, and enriched metastatic patterns specific to each tumour lineage. This work establishes a new framework for understanding TP53 mutation class as a clinically actionable distinction with direct implications for immunotherapy selection and prognosis. 
Read the full Abstract  


Expanding Precision Oncology to Underserved Populations in Uzbekistan 
In collaboration with ILBOZA Med Hospital and Tashkent Medical Park LLC, we presented one of the first comprehensive genomic profiling studies in the Uzbek cancer patient population, identifying 331 pathogenic variants across 124 patients spanning multiple tumour types. From TP53 and PIK3CA to KRAS and BRCA1/2, our findings revealed a rich landscape of actionable mutations in a population where access to precision oncology has historically been extremely limited; demonstrating that our platform can meaningfully expand the reach of targeted therapy to patients who need it most. 
Read the full Abstract


MIRAGE:  Detecting Minimal Residual Disease With DNA Methylation 
Our team introduced MIRAGE: Minimal Residual Assessment using Genome-wide Epigenomics- a proprietary computational algorithm that detects residual cancer at tumour fractions as low as 0.001%, using genome-wide DNA methylation analysis. Validated across 156 samples with 96.8% specificity, MIRAGE addresses one of the most critical unmet needs in oncology- detecting cancer that remains invisible to conventional diagnostics after surgery or treatment, before it progresses to clinical recurrence. 
Read the full Abstract


OncoPredikt-  AI-Powered Digital Pathology for HER2-Ultralow Breast Cancer 
Our deep-learning model OncoPredikt was presented as a breakthrough solution for automated tumour detection and biomarker quantification in breast cancer IHC whole-slide images. The model achieved Dice Similarity Coefficient scores above 0.8 and demonstrated the ability to detect HER2-ultralow positivity that visual pathologist assessment had missed, directly addressing the diagnostic gap that currently prevents eligible patients from accessing life-saving Antibody-Drug Conjugate therapies. 
Read the full Abstract



OncoAlibrex: Ultrasensitive cfDNA Fragmentomics for Treatment Response 
Our team presented OncoAlibrex;  an NGS-independent cfDNA fragmentomics assay that detects tumour-specific DNA fragmentation patterns with exceptional sensitivity and 100% specificity, without relying on mutation detection. Analytically validated with intra-assay variability of less than 1% and linearity of R²≥0.99, OncoAlibrex enables early treatment response monitoring and MRD detection:-  from a simple blood draw, even at tumour fractions too low for conventional ctDNA assays to detect. 
Read the full Abstract


Seven Abstracts. One Vision.
 
Taken together, our seven AACR 2026 presentations tell a unified story of a team that is not just building diagnostic tools, but generating the clinical evidence to prove they work. From reducing genomic uncertainty and predicting chemotherapy toxicity, to detecting residual cancer at the molecular level and enabling smarter immunotherapy decisions, every piece of research we presented in San Diego reflects our core belief: that better data, better science, and better AI can deliver better outcomes for cancer patients everywhere. 
We are grateful to our scientific collaborators, clinical partners, and research teams across India and the United States whose tireless work made these presentations possible. And we look forward to continuing to bring this science and its clinical impact: to oncologists and patients around the world.