Researchers have developed a new biomarker called VMD that uses artificial intelligence to predict gastric cancer survival by analyzing fat and muscle density in CT scans. According to the São Paulo Research Foundation, patients with higher VMD values face a significantly worse prognosis, with a median survival of 13.8 months compared to 58.5 months for those with lower values.
How are AI-driven biomarkers changing cancer staging?
Current gastric cancer treatment typically relies on disease staging, which focuses on tumor size and the presence of metastases. However, this method often fails to explain why two patients with the same stage progress differently.
The VMD biomarker shifts the focus from the tumor to the patient. By evaluating the metabolic and inflammatory state of the individual, doctors can better predict how a disease will behave. Professor José Barreto Campello Carvalheira, a leader in the study and professor of clinical oncology, says the goal is to expand staging to incorporate a full assessment of the patient’s physical condition.
This evolution toward precision medicine allows for “therapeutic stratification.” In the future, this could help clinicians decide which patients require aggressive chemotherapy and which can avoid toxic treatments following surgery.
Traditional cancer staging looks at the “enemy” (the tumor). New biomarkers like VMD look at the “terrain” (the patient’s body) to predict the outcome of the battle.
Why does the relationship between fat and muscle density matter?
The VMD marker identifies a specific “integrated patient phenotype.” This phenotype combines data from two different types of tissue to create a clearer picture of clinical risk.
According to researcher Mendes, the two tissues react differently to disease progression:
- Adipose (Fat) Tissue: Higher radiodensity values are linked to a worse prognosis and may indicate active inflammation.
- Muscle Tissue: Lower radiodensity values indicate a worse prognosis.
By measuring the difference between these two densities, researchers can capture a more complete view of a patient’s metabolic health than they could by looking at a single variable.
Survival Rate Comparison
The impact of these density values on patient longevity is stark. The study observed a massive gap in median survival times based on VMD levels:
| VMD Indicator | Median Survival Time |
|---|---|
| Worst Indicators (High VMD) | 13.8 months |
| Better Indicators (Low VMD) | 58.5 months |
How does machine learning improve medical imaging accuracy?
Developing the VMD formula required moving beyond manual analysis. Researchers Maria Emília Seren Takahashi and her team utilized machine learning to process massive volumes of imaging, clinical, and laboratory data simultaneously.
“I taught the machine to look in the same direction that experts were already looking, but with greater speed and scale,” Takahashi explained. The AI tested various combinations of data until it identified the formula that most accurately flagged high-risk patients.
A major technical hurdle in radiology is that CT scans can vary slightly depending on the machine’s calibration. To solve this, the team used the difference between fat and muscle radiodensity. This method effectively “cancels out” technical variations between different scanners, making the marker more reliable across different hospitals.
What are the next steps for VMD and precision oncology?
While the results are promising, the researchers note that this was a retrospective study. To confirm if VMD can officially guide clinical decisions, it must undergo external validation through prospective, multicenter studies involving larger, more diverse populations.
There are also unanswered questions regarding treatment. Mendes noted that the team does not yet know if a patient’s body profile can be modified. While nutritional therapy might improve a patient’s metabolic state, the study did not evaluate whether changing that profile could directly improve survival rates.
The research is already expanding. Early results suggest that this AI-driven approach to analyzing tissue density may work for other types of cancer beyond gastric cancer.
Frequently Asked Questions
What is the VMD biomarker?
VMD stands for the difference in radiodensity between adipose (fat) tissue and muscle tissue, used to predict cancer prognosis.
Can this marker be used with existing hospital equipment?
Yes. Because the marker is derived from standard CT scans that are already part of routine cancer care, it does not require new or expensive diagnostic tests.
Does VMD only apply to gastric cancer?
The current study focused on gastric cancer, but researchers are already testing the marker in other cancer types with early signs of success.
Want to stay updated on the latest in medical AI and oncology? Subscribe to our newsletter or leave a comment below with your thoughts on the future of precision medicine.