Predicting breast cancer spread with AI-generated spatial pathology maps

A Study on Predicting the Risk of Distant Metastasis in Breast Cancer Using AI-Generated Spatial Pathological Maps

This study will try an AI model that uses routine pathology images and clinical data to predict long-term risk of distant metastasis in adult women who had surgery for primary invasive breast cancer.

Quick facts

• Study type: Observational
• Enrollment: 400 (estimated)
• Ages: 18 Years to 95 Years
• Sex: Female
• Sponsor: Second Affiliated Hospital, School of Medicine, Zhejiang University (other)
• Locations: 4 sites (Changchun, Jilin and 3 other locations)
• Trial ID: NCT07244094 on ClinicalTrials.gov

What this trial studies

This observational study will train and validate a multimodal AI model using archived H&E whole-slide images and consecutive multiplex IHC sections from surgically resected primary breast tumors. Clinical and pathologic data (TNM stage, grade, receptor status, treatments) and at least five years of follow-up including distant metastasis-free survival will be used to label and tune the model. The AI approach focuses on spatial pathomics features (e.g., Pan-CK, CD3, CD20 patterns) derived from digitized slides to predict long-term metastasis risk. Data are drawn from cases treated between 2015 and 2025 at participating Chinese cancer centers and processed retrospectively for model development and validation.

Who should consider this trial

Good fit: Adult women (≥18 years) with histologically confirmed primary invasive breast cancer who underwent curative surgery between 2015 and 2025 and have high-quality digitizable H&E slides, consecutive mIHC sections, complete clinicopathologic data, and at least five years of follow-up.

Not a fit: Patients without archived or high-quality digitizable tissue, lacking consecutive mIHC sections or complete follow-up data, or those with noninvasive tumors or incomplete clinical records are unlikely to benefit from this analysis.

Why it matters

Potential benefit: If successful, the tool could help identify patients at higher risk of distant metastasis so clinicians can personalize surveillance and adjuvant therapy decisions.

How similar studies have performed: Retrospective studies applying AI to pathology images have shown promising prognostic signals, but multimodal spatial pathomic models for long-term metastasis prediction remain relatively novel and require broader validation.

Eligibility criteria

Inclusion Criteria:

1. Female patients aged 18 years or older.
2. Histologically confirmed primary invasive breast carcinoma.
3. Underwent curative surgical resection (mastectomy or breast-conserving surgery) between January 2015 and December 2025.
4. Before initiating the neoadjuvant therapy, there was a retention of the primary tumor specimen.
5. Availability of high-quality, digitizable Hematoxylin and Eosin (H\&E) stained whole-slide images (WSIs).
6. Availability of consecutive tissue sections from the same tumor block for multiplex immunohistochemistry (mIHC) staining (including markers such as Pan-CK, CD3, CD20).
7. Complete clinicopathological data and follow-up information must be available, including but not limited to: TNM stage, histological grade, molecular subtype (ER, PR, HER2 status), adjuvant treatment records, and clearly documented distant metastasis-free survival (DMFS) data.
8. A minimum follow-up of 5 years for patients with detailed information for distant metastasis events.

Exclusion Criteria:

1. Pure ductal carcinoma in situ (DCIS) without an invasive component.
2. Special histological subtypes of invasive carcinoma (e.g., metaplastic carcinoma) with distinct biological behaviors.
3. No original lesion samples were retained before neoadjuvant therapy.
4. Presence of contralateral breast cancer or a history of any other prior malignancy (except for cured non-melanoma skin cancer or carcinoma in situ of the cervix).
5. H\&E or IHC slides with significant technical artifacts (e.g., fading, folds, heavy knife marks, tissue tearing, uneven staining) that preclude reliable image analysis.
6. Low tumor cellularity (e.g., tumor area \< 10% in the scanned field of view).
7. Unavailable or unalignable consecutive tissue sections, preventing spatial registration of H\&E and mIHC images.
8. Lack of essential clinicopathological or follow-up data required for model training or validation.

Where this trial is running

Changchun, Jilin and 3 other locations

• Jilin Cancer Hospital — Changchun, Jilin, China (NOT_YET_RECRUITING)
• Cancer Institute and Hospital, Tianjin Medical University, China — Tianjin, Tianjin Municipality, China (NOT_YET_RECRUITING)
• 2nd Affiliated Hospital, School of Medicine, Zhejiang University, China — Hangzhou, Zhejiang, China (RECRUITING)
• The Fourth Affiliated Hospital of Zhejiang University School of Medicine — Hangzhou, Zhejiang, China (NOT_YET_RECRUITING)

Study contacts

• Study coordinator: Jiaojiao Zhou

Conditions: Breast Cancer, Artificial Intelligence, Distant Metastasis, Prediction