Agent-Guided Relational Concept Discovery: Toward Interpretable Surgical Margin Assessment

Source

Source name: arXiv
Source kind: paper
Authors / organization: Nooshin Maghsoodi, Amoon Jamzad, Robert Policelli, Mohammad Farahmand, Dilakshan Srikanthan, Martin Kaufmann, Kevin Y. M. Ren, Shaila Merchant, Sonal Varma, Ross Walker, Doug McKay, John Rudan, Gabor Fichtinger, Parvin Mousavi
Venue: arXiv: cs.AI
Published: 2026-07-23
Collected: 2026-07-24
URL: http://arxiv.org/abs/2607.21437v1
PDF: https://arxiv.org/pdf/2607.21437v1
DOI: UNKNOWN

Plain-Language Brief

Researchers propose a new method called Agent-Guided Concept Discovery to help deep learning models assess surgical margins more effectively, particularly in noisy and unlabeled data acquired during surgery. The proposed framework learns meaningful concepts directly from the data without requiring predefined concept labels. It also ensures consistency with known metabolic relationships through a biochemical knowledge graph.

What the Document Says

Deep learning models can be effective for surgical margin assessment using Rapid Evaporative Ionization Mass Spectrometry (REIMS) data, but their clinical adoption is challenging due to limited generalization to operating room conditions.
The challenge arises because deep learning models are typically trained on labeled spectra from resected tissue samples while they must operate on noisy, unlabeled data acquired during surgery.
The black-box nature of deep learning makes it difficult to understand and improve their behavior.
Concept-based learning offers a promising solution, but supervised concept-based approaches rely on concept annotations which are hard to obtain in complex mass spectrometry workflows.
Agent-Guided Concept Discovery is proposed as a framework that learns meaningful concepts directly from data without requiring predefined concept labels.
During training, a reasoning agent refines semantic descriptions of the learned concepts and adaptively adjusts their weight based on diagnostic relevance.
These concepts are grounded using a biochemical knowledge graph to ensure consistency with known metabolic relationships.
The model was tested across Skin and Breast Cancer datasets, showing improved balanced accuracy and sensitivity over the baseline.
In a representative intraoperative case, it showed fewer false positives, indicating better generalization to surgical conditions.

Key Concepts Mentioned

Deep learning
Rapid Evaporative Ionization Mass Spectrometry (REIMS)
Surgical margin assessment
Concept-based learning
Agent-Guided Concept Discovery
Reasoning agent
Biochemical knowledge graph
Balanced accuracy
Sensitivity
False positives

Claims Made by the Document

Deep learning models can effectively use REIMS data for surgical margin assessment but their clinical adoption is challenging.
The proposed Agent-Guided Concept Discovery framework learns meaningful concepts directly from data without requiring predefined concept labels.
The model improves balanced accuracy and sensitivity over the baseline.
In a representative intraoperative case, it shows fewer false positives.

Evidence Used

Deep learning models can effectively use REIMS data for surgical margin assessment. However, their clinical adoption remains challenging due to limited generalization to operating room conditions. In addition, the black-box nature of deep learning makes it difficult to understand and systematically improve their behavior.
We propose Agent-Guided Concept Discovery, a framework that learns meaningful concepts directly from data without requiring predefined concept labels.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Deep learning models can effectively use Rapid Evaporative Ionization Mass Spectrometry (REIMS) data for surgical margin assessment. However, their clinical adoption remains challenging due to limited generalization to operating room conditions. This difficulty arises because models are typically trained on labeled spectra collected from resected tissue samples, while they must operate on noisy, unlabeled data acquired directly during surgery. In addition, the black-box nature of deep learning models makes it difficult to understand and systematically improve their behavior. Concept-based learning offers a promising way to address these challenges by mapping raw measurements to human-understandable concepts. However, supervised concept-based approaches rely on concept annotations, which are difficult to obtain in complex mass spectrometry workflows. We propose Agent-Guided Concept Discovery, a framework that learns meaningful concepts directly from data without requiring predefined concept labels. During training, a reasoning agent refines semantic descriptions of the learned concepts and adaptively adjusts their weight based on diagnostic relevance. These concepts are further grounded using a biochemical knowledge graph to ensure consistency with known metabolic relationships. Across Skin and Breast Cancer datasets, our model improves balanced accuracy and sensitivity over the baseline. In a representative intraoperative case, it shows fewer false positives, indicating better generalization to surgical conditions.