Predicting response to neoadjuvant therapy in rectal cancer with multi-omics and AI

Establishing a Strategy for Selecting Watchful Waiting and Determining the Optimal Timing for Surgery Following Neoadjuvant Therapy

This project will try to use imaging, gut microbiome, pathology, and blood-based markers to predict how adults with stage II–III rectal cancer respond to neoadjuvant therapy.

Quick facts

• Study type: Observational
• Enrollment: 869 (estimated)
• Ages: 18 Years and up
• Sex: All
• Sponsor: Beijing Friendship Hospital (Academic / other)
• Drugs / interventions: immunotherapy
• Locations: 1 site (Beijing, Beijing Municipality)
• Trial ID: NCT07566559 on ClinicalTrials.gov

What this trial studies

This multicenter real-world observational project will integrate radiomics, gut microbiota profiling, pathological quality control, and liquid biopsy data to build a deep-learning model that predicts response to neoadjuvant therapy in rectal cancer. The study will analyse pre- and post-treatment imaging and molecular data to train and validate a multidimensional predictive system and to identify patients who might safely follow a 'watch-and-wait' approach instead of immediate surgery. It will also document outcomes of immunotherapy combined with neoadjuvant chemoradiotherapy versus traditional regimens in routine practice to inform protocol optimization. Model performance and clinical feasibility are planned to be validated through subsequent prospective clinical trials.

Who should consider this trial

Good fit: Adults (≥18 years) with biopsy-confirmed rectal adenocarcinoma, clinical stage II–III (cT1-4aN0-2M0), tumor ≤10 cm from the anal verge on MRI, no distant metastasis, and scheduled for neoadjuvant therapy.

Not a fit: Patients with distant organ metastasis, multiple primary colorectal cancers, a recent other malignancy (except certain cured skin/cervical in situ cancers), or those not receiving neoadjuvant therapy are unlikely to benefit from this study's predictive model.

Why it matters

Potential benefit: If successful, the model could help doctors identify patients who can avoid surgery and preserve bowel function by guiding personalized organ-preserving strategies.

How similar studies have performed: Previous studies show imaging and molecular markers can partially predict response and that watch-and-wait is safe in select complete responders, but combining radiomics, microbiome, liquid biopsy and AI in a validated real-world model is largely novel.

Eligibility criteria

Inclusion Criteria:

* The patient is informed and has provided written informed consent;
* Rectal adenocarcinoma confirmed by colonoscopic biopsy and pathology, meeting the following criteria:

  1. Clinical stage II/III locally advanced rectal cancer (LARC): cT1-4aN0-2M0;
  2. The distal edge of the tumor is ≤ 10 cm from the anal verge (measured by MRI);
  3. No distant metastasis;
  4. Scheduled to receive neoadjuvant therapy;
* Age ≥ 18 years, male or female。

Exclusion Criteria:

* Presence of distant organ metastasis;
* Multiple primary colorectal cancers;
* History of prior malignancy (except completely cured carcinoma in situ of the cervix, basal cell carcinoma, or squamous cell carcinoma of the skin).

Where this trial is running

Beijing, Beijing Municipality

• Beijing Friendship Hospital — Beijing, Beijing Municipality, China (Recruiting)

Study contacts

• Study coordinator: hongwei Yao, Professor, Doctoral Degree

Conditions Rectal Cancer; Rectal cancer; Real world study; Neoadjuvant Therapy