A significant new study published in Clinical Nutrition offers a powerful, host-centered method for predicting outcomes in cancer patients, revealing five universal patient clusters based on an early, comprehensive evaluation of their nutritional and inflammatory status. This innovative approach provides crucial prognostic insights independent of the patient’s specific tumor origin or whether their cancer has metastasized.

The research was driven by the recognition that despite their profound impact on patient outcomes, conditions like cancer-associated cachexia and systemic inflammation are frequently underestimated in oncology.

Malnutrition is highly prevalent among cancer patients, affecting an estimated 30% to 80% of individuals, varying with tumor stage, site, and treatment. Cancer-related cachexia, a complex, multifactorial syndrome, is particularly concerning as it leads to continuous depletion of skeletal muscle mass and is often irreversible with standard nutritional support alone.

To address this, researchers at Cochin University Hospital in Paris, France, conducted a study involving 1370 oncology outpatients who underwent a systematic pre-treatment assessment (PTA) between 2017 and 2023. This program, part of routine care, involved a multidisciplinary evaluation including anthropometric measurements, resting energy expenditure, estimated protein and calorie intake, and detailed biological assessments. While the data itself was prospectively collected as part of this routine program, the study conducted retrospective analyses on this routinely gathered clinical and biological information.
Using an advanced unsupervised clustering analysis across 53 clinical and biological features, the team identified seven principal components reflecting various aspects of patient health, such as inflammation, renal/vascular disease, and metabolism.

Based on these components, five distinct patient groups emerged:

• “Fit patients”, comprising 47% of the cohort, represented the healthier baseline.
• “Older comorbid patients”, making up 26%, were characterized by advanced age and multiple comorbidities, particularly cardiac and vascular issues.
• “Dysmetabolic patients”, accounting for 10%, displayed specific nutritional and metabolic imbalances, including an inability to meet their protein and calorie needs, often being overweight, with less than 5% weight loss, and impaired glucose tolerance. These patients frequently align with the criteria for pre-cachexia, a stage of high risk for developing full-blown cachexia.
• A smaller group of “digestive/liver injured patients” constituted 2% of the cohort, indicating specific gastrointestinal and hepatic impairments.
• Finally, “critically impaired/inflamed patients”, representing 15% of the cohort, exhibited characteristics such as higher weight loss, significantly decreased albumin and pre-albumin levels, and markedly elevated inflammatory markers like C-reactive protein. Inflammation in this group was found to profoundly contribute to sarcopenia and catabolism, often despite not meeting the traditional 5% weight loss criterion for cachexia, challenging existing diagnostic definitions.

Crucially, these newly identified patient clusters demonstrated a strong and clinically meaningful association with overall survival. Patients in the “critically impaired/inflamed” and “dysmetabolic” groups faced significantly poorer prognoses compared to the “fit” patients.
For instance, after accounting for tumor origin and metastatic status, patients in the “critically impaired/inflamed” cluster had a nearly three-fold increased hazard ratio for death, while “dysmetabolic” patients had a 1.7-fold increased risk. This classification system proved to be as effective as standard prognostic scores in predicting 5-year overall survival, and even significantly improved prediction when combined with existing scores.

These findings underscore that nutritional status and systemic inflammation are critical, yet often underappreciated, determinants of prognosis in cancer patients. Moreover, early management of these factors can significantly influence overall clinical outcomes.

While the study’s retrospective nature and cohort heterogeneity pose some limitations, these also strengthen the proposed clusters’ universal applicability across diverse cancer populations. Ultimately, this research provides a robust framework for clinicians to move beyond solely tumor-specific characteristics and integrate a patient’s systemic response into more effective, individualized cancer care strategies.

Source.