In the high-stakes world of oncology research, the quest for a surrogate endpoint that reliably predicts overall survival (OS) has long been the "holy grail." While traditional benchmarks like the Response Evaluation Criteria in Solid Tumours (RECIST) have served the medical community for decades, they often fail to capture the nuanced, multifactorial reality of a patient’s response to complex drug regimens. A groundbreaking study presented at the World Conference on Lung Cancer (WCLC) 2026 in Seoul, Republic of Korea, suggests that the future of cancer treatment assessment may lie in artificial intelligence.
Altis Labs has unveiled data demonstrating that its AI-powered imaging endpoint, IPRO, can detect early, clinically meaningful treatment effects more effectively than standard RECIST-based objective response rates (ORR). By moving beyond simple lesion measurement, this technology promises to transform clinical trial design and, more importantly, patient care.
The MAIRPOSA Study: A New Lens for Lung Cancer Treatment
The data in question stems from Johnson & Johnson’s pivotal Phase III MARIPOSA study (NCT04487080). This trial investigated the efficacy of the combination therapy Rybrevant (amivantamab) and Lazcluze (lazertinib) in patients diagnosed with epidermal growth factor receptor (EGFR)-mutated advanced non-small-cell lung cancer (NSCLC).
As part of the research effort, Altis Labs’ IPRO software performed a comprehensive analysis of approximately 10,000 radiology scans generated throughout the study. The software created longitudinal prognostic outcome measures for every participant, beginning at baseline and continuing through subsequent on-treatment assessments.
Unlike traditional methods that rely on human radiologist measurement of tumor diameter—a process susceptible to inter-observer variability—IPRO is a fully automated AI system. It predicts patient survival outcomes by extracting data directly from CT scans, utilizing a holistic view of the patient’s physiological state.
Chronology of a Diagnostic Shift
The integration of IPRO into the MARIPOSA trial provides a clear timeline for how this technology outperforms legacy systems.
- Baseline: IPRO establishes a prognostic baseline for each patient, integrating data that spans beyond the primary tumor to include body composition and organ health.
- The 16-Week Threshold: In the MARIPOSA study, IPRO identified a significant favorability toward the investigational treatment arm starting as early as week 16. This early signal persisted throughout the duration of the study.
- The Discrepancy: During the same assessment windows, standard RECIST-based ORR failed to anticipate the long-term survival benefit that was eventually observed. This indicates that while tumors might appear similar in size under traditional criteria, the underlying biological reality—as interpreted by IPRO—revealed that patients on the investigational arm were experiencing a more robust, survival-linked benefit.
- Long-Term Validation: The study found that pooled, patient-level IPRO trajectories were consistently linked to survival outcomes. Specifically, "IPRO improvement" served as a reliable proxy for overall survival benefit, while "IPRO deterioration" signaled a detriment to survival, allowing for a more granular understanding of patient response than binary RECIST classifications.
Beyond the Ruler: How IPRO Functions
The core innovation of IPRO is its departure from the linear measurement of tumor burden. RECIST, the industry standard, focuses primarily on the change in size of "target lesions." While useful, this approach is inherently reductionist, ignoring the systemic impact of cancer and the collateral health of the patient.
IPRO, conversely, analyzes three-dimensional scans to capture a far more complex picture. By identifying prognostic imaging biomarkers, the software assesses:
- Tumor Burden: An automated calculation of total tumor volume and distribution.
- Body Composition: Monitoring muscle and fat distribution, which can be critical indicators of cachexia and overall health status.
- Organ Health: Evaluating the systemic impact of the disease and the treatment on vital organs, providing a "whole-body" view of the patient’s response.
This multifactorial approach allows the software to generate an "IPRO-α" score. A response is defined by the study as a 50% or greater improvement in this score relative to the baseline. This provides a quantifiable, high-fidelity metric that is far more sensitive to early physiological changes than the subjective assessments of tumor shrinkage.
Official Perspectives: Bridging the Gap
The implications of these findings have not been lost on the leadership at Altis Labs. Felix Baldauf-Lenschen, Founder and CEO of Altis Labs, noted the significance of the WCLC presentation: "This readout proves that AI can anticipate meaningful clinical benefit that traditional imaging endpoints like ORR may fail to detect."
This shift is timely. The US Food and Drug Administration (FDA) has increasingly emphasized the importance of Overall Survival (OS) as the definitive measure of success in oncology. In fact, the agency released draft guidance last year underscoring OS as a key pre-specified endpoint for cancer clinical trials. However, OS can take years to mature, delaying the approval of life-saving therapies. By providing an early, validated surrogate endpoint that tracks with OS, IPRO could theoretically accelerate the drug development cycle without sacrificing regulatory rigor.
Implications for Clinical Practice and Patient Outcomes
The most profound impact of the IPRO technology will likely be felt in the clinic, where the ability to "course-correct" treatment is the difference between life and death.
1. Early Intervention for Treatment Failure
Currently, physicians often wait for physical symptoms or significant tumor growth to conclude that a treatment regimen is failing. By the time this failure is detected via traditional imaging, the patient’s disease may have progressed significantly, and their overall health may have declined. IPRO’s ability to detect potential treatment failure at earlier time points allows physicians to amend treatment plans proactively, potentially switching patients to second-line therapies before the disease becomes unmanageable.
2. Streamlining Clinical Trials
For pharmaceutical companies, the cost and duration of clinical trials are significant barriers to innovation. If regulatory bodies accept AI-derived endpoints like IPRO as valid early indicators of survival benefit, it could lead to more efficient trials. Researchers could identify the efficacy of a drug much earlier, reducing the time patients spend on ineffective therapies during trials and allowing successful drugs to reach the market faster.
3. Precision Medicine at Scale
The scalability of a fully automated system like IPRO cannot be overstated. By removing the burden of manual measurement from radiologists, healthcare systems can ensure consistent, high-quality analysis of imaging data across multiple trial sites. This ensures that the data is uniform, reducing the "noise" that often complicates large-scale clinical trials.
The Road Ahead
While the results from the MARIPOSA study are promising, the integration of AI into oncology will require continued validation. The regulatory landscape is slowly evolving to accommodate these digital tools, but the medical community remains cautious, prioritizing clinical safety and accuracy above all else.
As Altis Labs continues to refine IPRO and gather data from additional cohorts, the prospect of "AI-guided oncology" moves from the realm of science fiction into the standard of care. If the ability to predict survival outcomes from CT scans becomes a routine component of cancer treatment, we may be witnessing the beginning of a new era—one where the physician’s intuition is supported by a level of data-driven insight that was previously impossible to attain.
In the future, the "gold standard" may no longer be a ruler used to measure a tumor, but a sophisticated AI algorithm capable of seeing the broader story of a patient’s health, one scan at a time. The work presented in Seoul in 2026 marks a definitive milestone in that journey, providing a clearer path toward more effective, personalized, and rapid cancer treatment strategies.
About the Author
