AI Tool Predicts Treatment Response in Small Cell Lung Cancer
A computational pathology tool called PhenopyCell can predict whether patients with extensive-stage small cell lung cancer will respond to platinum-based chemotherapy before treatment begins, according to a study published in npj Precision Oncology. The prediction requires only existing diagnostic slides - no additional biopsies or procedures.
Researchers from Roswell Park Comprehensive Cancer Center, Winship Cancer Institute of Emory University, and University Hospitals Cleveland Medical Center validated the tool's accuracy across 281 patients. The team analyzed immune cell patterns visible in tissue samples to forecast treatment outcomes.
The Clinical Problem
Extensive-stage small cell lung cancer affects 70% of patients at diagnosis. The disease spreads rapidly, with median survival around 12 to 13 months. All patients currently receive the same standard treatment - platinum-based chemotherapy plus immunotherapy - regardless of whether they will respond.
By the time doctors confirm a treatment isn't working, the disease has often progressed too far to try alternatives. Newer SCLC treatments have recently won FDA approval or show promise in trials, but they work in only a subset of patients. Unlike many other cancers, small cell lung cancer has no identified biomarkers to guide treatment selection.
How PhenopyCell Works
The tool analyzes immune cell arrangements in diagnostic biopsy slides. Patients with better outcomes had tumors containing more immune cells organized in structured groups surrounding tumor clusters - a sign of stronger immune response. Patients with poor outcomes had fewer immune cells arranged in disorganized patterns distant from the tumor.
These immune cell patterns were only visible through the AI-powered analysis. Manual examination of the same slides did not reveal them as reliably.
PhenopyCell combines data from pathology slides and medical records to identify patterns linked to patient outcomes. The system requires no new tissue collection and no additional cost, since every small cell lung cancer patient already has a diagnostic pathology slide on file.
Potential Impact
The tool could help patients avoid treatments unlikely to help them and enroll earlier in clinical trials of newer drugs. It may also provide clearer prognosis information before treatment starts.
"Every patient with small cell lung cancer already has a pathology slide from their diagnostic biopsy," said Dr. Prantesh Jain, a thoracic oncologist at Roswell Park who co-led the research. "In a disease where survival is measured in months and re-biopsy is rarely possible, this has the potential to become a uniquely powerful tool."
The study was co-led by Dr. Jain and Dr. Anant Madabhushi of Winship Cancer Institute.
Learn more about AI for Healthcare and AI for Science & Research.
Your membership also unlocks:
