New AI tool accurately predicts how cancer patients will respond to treatment

Predicting how cancer patients will respond to treatment remains a critical challenge. Immune checkpoint inhibitors (ICIs), a type of immunotherapy, can offer significant benefits to some individuals with advanced-stage cancers. However, most patients undergoing this costly and often risky treatment see limited results.

Addressing this gap, researchers have developed a new tool, SCORPIO, that uses routine blood tests and clinical data to predict the efficacy of ICIs. This innovative approach may revolutionize cancer care by making treatment decisions more precise, accessible, and affordable.

Existing biomarkers, such as tumor mutational burden (TMB) and PD-L1 expression, are approved by the U.S. Food and Drug Administration (FDA) to predict ICI effectiveness.

Despite their promise, these methods face significant limitations. TMB requires advanced genomic sequencing, while PD-L1 testing lacks standardization and consistency. Both methods demand tumor samples, which are not always easily obtained, and require costly infrastructure that is unavailable in many healthcare settings.

These limitations highlight the urgent need for an alternative that is not only reliable but also widely accessible. The ideal solution would integrate routinely available clinical variables and standard laboratory tests, enabling physicians worldwide to assess ICI suitability quickly and cost-effectively.

Researchers at Memorial Sloan Kettering Cancer Center (MSK) and Mount Sinai have introduced SCORPIO, a machine learning model designed to fill this gap. SCORPIO leverages data from routine blood tests, such as complete blood counts and comprehensive metabolic profiles, alongside clinical characteristics to predict patient outcomes with ICIs. This approach eliminates the need for advanced genomic testing, making the technology accessible in a variety of healthcare settings.

“Immune checkpoint inhibitors are a powerful tool against cancer, but they don’t yet work for most patients,” says Dr. Luc Morris, co-senior author and surgeon at MSK. “These drugs are expensive and can have serious side effects, so the key is selecting the right patients.”

Published in the journal, Nature Medicine, SCORPIO’s development involved ensemble machine learning, which combines multiple analytical tools to identify patterns in clinical data.

The model was trained on a rich dataset from MSK, encompassing over 1,600 patients across 17 cancer types. It was then tested on additional MSK data, as well as external datasets from 10 global clinical trials and Mount Sinai. In total, the study included nearly 10,000 patients across 21 cancer types—the largest dataset in cancer immunotherapy to date.

SCORPIO outperformed existing biomarkers in predicting overall survival and clinical benefit from ICIs. During internal testing on over 2,500 patients, the model achieved a median area under the receiver operating characteristic curve (AUC) of 0.76 for predicting survival at intervals up to 30 months. In comparison, TMB achieved median AUC values below 0.55.

External validation demonstrated similar results. SCORPIO consistently surpassed PD-L1 immunostaining and TMB in predicting ICI outcomes across diverse patient populations. For instance, in data from 4,500 patients across six cancer types involved in global phase 3 trials, SCORPIO maintained robust predictive accuracy. An additional cohort of nearly 1,200 patients from Mount Sinai further confirmed its reliability.

“Our goal was not just to develop a model but to create one that would work across different locations and patient populations,” says Dr. Morris. This wide applicability is crucial for ensuring equitable access to precision medicine worldwide.

SCORPIO’s accuracy also translates to significant cost savings. By utilizing standard blood tests, the model avoids expensive genomic analyses and the logistical challenges of tumor sampling. This affordability ensures broader adoption, particularly in low-resource settings. The focus on equity aligns with global efforts to make advanced medical care universally available.

The simplicity of SCORPIO lies in its reliance on standard blood tests performed in clinics worldwide. These tests, combined with clinical variables such as body mass index and neutrophil-to-lymphocyte ratio, provide a comprehensive yet cost-effective method for predicting treatment outcomes. By eliminating the need for advanced genomic assays, SCORPIO offers a more equitable approach to cancer care.

“The model’s simplicity and affordability could help reduce costs and ensure that patients receive treatments most likely to benefit them,” says Dr. Morris. This adaptability is particularly important in resource-limited settings where advanced testing is unavailable.

Checkpoint inhibitors work by enabling immune cells to attack cancer more effectively. However, these drugs do not directly target cancer, which makes it critical to identify patients whose immune systems are primed to respond. SCORPIO addresses this challenge by offering a practical, data-driven solution.

The machine learning model also provides real-time predictions, making it suitable for use during initial consultations. Physicians can integrate SCORPIO’s insights with other diagnostic information to create personalized treatment plans. This streamlined approach reduces delays in treatment initiation, which is critical for improving patient outcomes.

To further refine SCORPIO, researchers are collaborating with hospitals and cancer centers worldwide to gather additional data. Feedback from diverse clinical settings will help optimize the model for broader use. Efforts are also underway to develop an accessible interface that clinicians can easily integrate into their practice.

“We envision a future where doctors anywhere can use this tool to make informed treatment decisions,” says Dr. Diego Chowell, co-senior author and assistant professor at Mount Sinai. With SCORPIO, the dream of universal access to precision oncology is becoming a reality.

Long-term goals include incorporating SCORPIO into clinical trials to evaluate its impact on survival rates and quality of life. By guiding treatment decisions, the model could reduce unnecessary exposure to ineffective therapies, sparing patients from potential side effects. This strategic application of AI aligns with the broader trend of integrating machine learning into everyday medical practice.

The potential impact of SCORPIO extends beyond cancer treatment. Machine learning models like this one represent a shift toward more personalized, data-driven healthcare. By leveraging routine clinical data, SCORPIO exemplifies how artificial intelligence can democratize access to cutting-edge medical tools.

As SCORPIO continues to evolve, its success underscores the importance of combining innovation with accessibility. For patients and physicians alike, this model represents a step forward in making cancer treatment both effective and equitable. Its widespread adoption could redefine how immunotherapy is administered, ensuring that more patients receive the right treatment at the right time.

The insights gained from SCORPIO’s development could also inspire similar applications in other areas of medicine. For example, predictive models could aid in managing chronic diseases or optimizing surgical outcomes. The overarching goal remains consistent: improving healthcare through data-driven decision-making and innovative technology.

With its demonstrated effectiveness and accessibility, SCORPIO has the potential to transform global cancer care. By reducing costs, improving outcomes, and ensuring equitable access, this AI-powered tool is a beacon of hope for patients and healthcare providers worldwide.

Note: Materials provided above by The Brighter Side of News. Content may be edited for style and length.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.