Learnijoy
ResearchAI ResearchAI Engineering & DevTools

AI Improves Tumor Proportion Scoring in Lung Cancer

Krzysztof Pysz, Artur Bartczak, Jaros{\l}aw Kwiecie\'n, Piotr Krajewski, Witold Dyrka· June 29, 2026 View original

Summary

This research introduces a novel deep multiple instance learning (MIL) framework for accurately assessing tumor proportion score (TPS) in non-small cell lung cancer (NSCLC) from whole slide images. The method uses two models to extract histopathological features and aggregate them into a zero-inflated beta distribution, outperforming existing linear regression baselines.

Accurately determining the tumor proportion score (TPS) in non-small cell lung cancer (NSCLC) is crucial for patient treatment and prognosis. However, this process is labor-intensive, requires highly specialized experts, and involves manually annotating numerous slides. Existing automated methods, particularly those based on multiple instance learning (MIL), often struggle with images that contain no tumor cells. Researchers have developed a new end-to-end framework to address these challenges. It comprises two main components: an embedding network that identifies features from individual image patches and a MIL model that combines these features. Crucially, the MIL model predicts parameters for a zero-inflated beta (ZIBeta) probability distribution, which represents the overall TPS for an entire slide. This distribution-based approach allows the system to handle images with varying tumor proportions, including those with zero tumor cells, more effectively than previous methods. By using only slide-level TPS scores for training, the framework demonstrates improved prediction accuracy and enhanced explainability compared to traditional linear and ridge regression baselines.

Why it matters

For healthcare professionals and AI developers in medical imaging, this advancement offers a more accurate and efficient way to automate a critical diagnostic step in NSCLC, potentially speeding up diagnoses and improving treatment planning.

How to implement this in your domain

  1. 1Explore integrating this distribution-based MIL approach into existing digital pathology workflows.
  2. 2Collaborate with AI researchers to adapt the ZIBeta modeling for specific cancer types or diagnostic tasks.
  3. 3Validate the model's performance on diverse, real-world clinical datasets to ensure robustness.
  4. 4Develop user interfaces for pathologists to review and interpret the AI-generated TPS distributions.
  5. 5Train medical staff on the use and interpretation of AI-assisted TPS scoring.

Who benefits

HealthcarePharmaceuticalsMedical DevicesBiotechnology

Key takeaways

  • AI can significantly improve the accuracy and efficiency of tumor proportion scoring in NSCLC.
  • A novel distribution-based MIL framework addresses limitations of previous methods, especially for zero-class images.
  • The ZIBeta modeling approach provides better prediction accuracy and explainability.
  • Automated TPS assessment can reduce manual workload and accelerate treatment planning.

Original post by Krzysztof Pysz, Artur Bartczak, Jaros{\l}aw Kwiecie\'n, Piotr Krajewski, Witold Dyrka

"arXiv:2606.27579v1 Announce Type: cross Abstract: Accurate assessment of tumor proportion score (TPS) in non-small cell lung cancer (NSCLC) is critical for treatment planning and prognosis. Key challenges include the tedious manual work required to annotate each slide, combined w…"

View on X

Originally posted by Krzysztof Pysz, Artur Bartczak, Jaros{\l}aw Kwiecie\'n, Piotr Krajewski, Witold Dyrka on X · view source

Want to go deeper?

Turn these trends into skills with Learnijoy's hands-on AI & tech courses.

Explore courses

More in AI Research

AI ResearchAI Engineering & DevTools

BaRA Improves LoRA Fine-Tuning with Adaptive Rank Allocation

Researchers introduce BaRA, a Bayesian Adaptive Rank Allocation framework for parameter-efficient fine-tuning, which dynamically adjusts adaptation capacity based on context. This method enhances predictive performance, robustness, and uncertainty calibration compared to standard LoRA and other Bayesian LoRA variants.

Zhibin Duan, Yuhong Wang, Jiahong Fu, Zongsheng Yue, Bo Chen, Zongben XuJun 30, 2026
AI ResearchAI Engineering & DevTools

New Preconditioner Improves Deep Network Training Stability and Performance

Researchers introduce Dead-Direction Conditioners (DDC), a novel preconditioning method that leverages gauge-equivariant optimization to prevent deep network training from drifting along symmetry orbits. This technique improves model stability, reduces overfitting, and enhances performance in language and vision models.

Tejas Pradeep ShirodkarJun 30, 2026
AI ResearchAI Engineering & DevTools

SMDA Traces Training Data Influence on LLM Behavioral Policies

Researchers introduce Symbolic Mechanistic Data Attribution (SMDA), a framework that attributes specific training examples to the interpretable symbolic policies governing an LLM's high-level behavior. SMDA offers a fine-grained diagnostic tool to understand how training data shapes model decisions, revealing safety gaps and unintended influences.

Reza Habibi, Darian Lee, Magy Seif El-NasrJun 30, 2026