For patients with early-stage disease, surgery can be curative, but it does not always eliminate the risk of relapse. Conventional scans provide an important view of anatomy, yet they may not detect a small population of cancer cells before it grows into a visible lesion. MRD testing approaches the problem from a different direction. Instead of searching for a mass, it looks for fragments of tumor DNA circulating in the blood. If those fragments persist after resection, they may indicate that microscopic disease remains, even when imaging appears clear.

The study’s focus on EGFR-mutated lung cancer is particularly significant. EGFR mutations can drive the uncontrolled growth of tumor cells and are found in a substantial proportion of lung adenocarcinomas, especially among people who have never smoked or have smoked lightly. These alterations also create an opportunity for precision medicine because they can be targeted by drugs known as EGFR tyrosine kinase inhibitors. The challenge is determining which patients need additional treatment after surgery and which may be spared months or years of therapy and its potential side effects.

Molecular residual disease detection is designed to make that decision more precise. After a tumor is removed, researchers can analyze its genetic profile and identify mutations or other molecular features unique to that cancer. Highly sensitive sequencing methods can then search for matching fragments in subsequent blood samples. The technical difficulty is considerable: tumor DNA may represent only a tiny fraction of all cell-free DNA in the bloodstream, while normal tissues continuously release their own genetic material. A reliable test must therefore distinguish a genuine cancer signal from background noise and laboratory artifacts.

The clinical value of MRD does not rest solely on whether a test can detect DNA. The crucial question is whether the result changes what doctors do and improves outcomes for patients. A positive result might identify people at particularly high risk of recurrence, supporting closer surveillance or consideration of adjuvant targeted treatment. A negative result could help define a group with a lower immediate risk, although it cannot guarantee that a relapse will never occur. The timing of blood collection, the depth of sequencing, the mutation selected for tracking and the duration of follow-up all influence the meaning of a result.

In EGFR-mutated disease, the stakes are amplified by the availability of effective targeted therapies. Drugs such as osimertinib have demonstrated benefits in the postoperative setting, but treatment decisions still require a balance between reducing recurrence risk and avoiding unnecessary exposure. MRD could eventually provide a dynamic measure of disease status, allowing care to become more responsive than a one-time decision based only on tumor stage and pathology. A rising molecular signal might prompt further investigation, while sustained clearance could help doctors assess whether treatment is suppressing residual disease.

The research also highlights why a blood-based test should be interpreted as part of a broader clinical framework rather than as an isolated verdict. A negative result may reflect the biological limits of detection, particularly when a tumor sheds little DNA into the bloodstream. A positive result may require confirmation, because technical contamination or clonal changes in non-cancerous cells can complicate genetic analysis. For this reason, the practical adoption of MRD testing depends on standardized laboratory methods, carefully defined thresholds and prospective evidence connecting test results with treatment decisions and long-term survival.

As precision oncology moves beyond matching drugs to mutations, it is increasingly turning toward the continuous monitoring of disease. The work by Zhou and colleagues places EGFR-mutated early-stage lung cancer within that wider transformation, where molecular information collected after surgery may help reveal what conventional scans cannot yet see. The promise is substantial: earlier recognition of recurrence, more individualized use of targeted therapy and a clearer understanding of who remains at risk. The field’s next challenge is ensuring that molecular signals translate into decisions that are not only technically accurate, but demonstrably better for patients.

Subject of Research: Molecular residual disease detection in early-stage resected EGFR-mutated non-small cell lung cancer

Article Title: Clinical utility of molecular residual disease detection in early-stage resected EGFR-mutated non-small cell lung cancer

Article References: Zhou, F., Su, C., Liang, W. et al. Clinical utility of molecular residual disease detection in early-stage resected EGFR-mutated non-small cell lung cancer. Nature Communications (2026). https://doi.org/10.1038/s41467-026-76392-9

Image Credits: AI Generated

DOI: 10.1038/s41467-026-76392-9

Keywords: Molecular residual disease, MRD, EGFR mutation, non-small cell lung cancer, lung cancer, liquid biopsy, circulating tumor DNA, precision oncology, cancer recurrence, targeted therapy

Instead of relying on new biomarkers or additional tissue sampling, SÉMIL learns from two sources that pathology already produces: digital slide images and accompanying textual descriptions. This “semantic” integration allows the algorithm to interpret tumour features while preserving clinical context, turning everyday diagnostic materials into quantitative risk estimates.

In the study, researchers analysed more than 1,600 pathology slides and validated the model on 1,220 stage-two patients across three independent cohorts. The validation extended beyond a single site, drawing on data from multiple Australian institutions, a step aimed at testing robustness in real-world clinical variability.

Technically, SÉMIL evaluates tumour architecture at the invasive front—the boundary where cancer spreads into surrounding tissue. This region is considered prognostically important, yet difficult to label consistently between pathologists, partly because subtle growth patterns can be subjective.

The results show that each tumour can be assigned to a higher- or lower-risk group, supporting triage decisions after surgery. Importantly, when AI-based assessments aligned with evaluations by expert pathologists, the model produced the most accurate risk ratings.

Such performance matters because current Australian guidelines generally reserve chemotherapy for high-risk stage-two patients. If relapse risk can be determined more precisely, more patients could receive timely escalation—or spared when escalation is unlikely to help.

The researchers stress that the tool is intended to support clinical decision-making rather than replace it. By adding an “additional layer of information,” AI could help clinicians balance the benefits of chemotherapy against its side effects.

Beyond bowel cancer, the team envisions future systems that combine AI-derived pathology features with other emerging biomarkers to refine risk stratification and treatment planning in precision medicine.

Subject of Research: Human tissue samples
Article Title: AI-Assisted Risk Stratification in Stage II Colorectal Cancer: Multi-Institutional Validation of Semantically-Enhanced Deep Learning
News Publication Date: 28-Jul-2026
Web References: https://www.gastrojournal.org/article/S0016-5085(26)07069-1/fulltext; https://doi.org/10.1053/j.gastro.2026.07.009
References: Gastroenterology (published article)
Image Credits:

Keywords: artificial intelligence, colorectal cancer, digital pathology, risk stratification, deep learning, multiple instance learning, invasive front, precision medicine, clinical decision support

CAR T cell therapy has emerged as one of the most significant advances in the treatment of blood cancers, providing hope for many patients whose disease has not responded to conventional therapies. However, a concerning reality is that more than half of NHL patients who do not respond favorably to initial treatments end up relapsing or progressing shortly after receiving CAR T therapy. This high rate of treatment failure has underscored the need for advanced predictive tools that can identify which patients are most likely to benefit from such innovative therapies.

The researchers behind InflaMix have utilized machine learning methodologies to analyze the profiles of inflammation in the blood of 149 NHL patients. The significance of this tool lies in its ability to assess various blood biomarkers related to inflammation, which has been implicated as a contributing factor to CAR T therapy failure. Traditional clinical practices have not typically employed these biomarkers, which InflaMix has now identified as critical in forecasting treatment outcomes.

The model operates on an unsupervised basis, meaning that it was trained without any prior knowledge of patient outcomes. By detecting an inflammatory biomarker through a set of unique blood tests, InflaMix can illuminate the inflammatory signatures associated with a heightened risk of CAR T treatment failure, encompassing risks of disease relapse as well as increased mortality. This novel approach allows for a more nuanced understanding of the biological mechanisms at play during CAR T therapy.

Dr. Marcel van den Brink, one of the leading authors of the study and a prominent figure at City of Hope, expressed optimism about the potential of InflaMix. He emphasized that this tool could serve as a universal asset for oncologists everywhere, enabling them to evaluate the risks associated with CAR T therapy on an individual basis, ultimately leading to a more personalized treatment journey for each patient. This ability to tailor treatment strategies based on empirical evidence could revolutionize how oncologists approach CAR T therapy and similar innovative treatments.

Furthermore, the impressiveness of InflaMix is accentuated by its flexibility. The model performed well even when evaluated with only six commonly used blood tests, all of which are typically assessed in lymphoma patients. This flexibility signifies that the test could be broadly accessible, making it feasible for most NHL patients to benefit from its predictive capabilities, regardless of their specific clinical background or treatment history.

Oncologist Dr. Sandeep Raj, who led the study at MSK, affirmed that prior studies had hinted at inflammation being a risk factor for diminishing the efficacy of CAR T cell therapies. The team’s endeavor to refine this understanding and create a robust clinical tool has culminated in the development of InflaMix, which not only characterizes inflammation in blood but also predicts the likelihood of successful CAR T therapy outcomes among patients.

Validation of the model was established through studies that included three independent cohorts comprising 688 NHL patients. This diversified group exhibited various clinical characteristics and disease subtypes while having received different CAR T products. The array of clinical data reinforces the reliability of the InflaMix tool in diverse patient profiles, enhancing its utility as a standard part of clinical assessments.

Looking forward, researchers at City of Hope and MSK are poised to investigate further the relationship between the blood inflammation patterns identified by InflaMix and their impact on CAR T cell function. By exploring the underlying sources of this inflammation, the team aims to deepen the understanding of factors that influence treatment efficacy in NHL patients treated with CAR T therapy.

The potential applications for InflaMix extend beyond mere prediction. By effectively identifying patients with a high risk of treatment failure, there is an opportunity for clinicians to modify treatment plans. This could involve designing new clinical trials that integrate additional therapeutic strategies aimed at improving CAR T effectiveness—a prospect that holds promise for transforming the landscape of blood cancer treatment.

Currently, City of Hope stands as a leader in CAR T cell therapies, having treated over 1,700 patients since launching their CAR T program in the late 1990s. Their commitment to clinical excellence is reflected in their expansive array of ongoing clinical trials, including 70 studies focused on immune cell products, primarily CAR T therapies, that address various forms of blood and solid tumor cancers. Their efforts not only elevate patient care but also contribute to the overall advancement of cancer research.

Support for the team’s studies has stemmed from notable institutions, including the National Institutes of Health and the National Cancer Institute. With Dr. Van den Brink’s recent transition to City of Hope after two decades at MSK, the collaboration promises to yield innovative discoveries and further establish the institution’s role as a pioneer in CAR T cell therapy research and treatment.

As the cancer research community anticipates the broader implications of this work, InflaMix stands as a beacon of hope for NHL patients and a testament to the potential of integrating advanced technologies like machine learning in clinical settings. The move towards personalized medicine, guided by precise predictors of treatment outcomes, heralds a new era in the fight against cancer, making strides in the quest for more effective and individualized care.

Subject of Research: Machine Learning Tool for Predicting Response to CAR T Cell Therapy in Non-Hodgkin Lymphoma Patients
Article Title: InflaMix: A Machine Learning Approach to Predict CAR T Cell Therapy Outcomes
News Publication Date: 1-Apr-2025
Web References: City of Hope, Nature Medicine
References: NIH, NCI
Image Credits: City of Hope
Keywords: CAR T Cell Therapy, Non-Hodgkin Lymphoma, Machine Learning, InflaMix, Inflammation Biomarkers, Predictive Analytics, Personalized Medicine, Oncology Research, Blood Cancer Treatment, Clinical Trials.