AI-Assisted Treatment Planning for Gastrointestinal Cancer

Recruiting at 6 trial locations

Overseen By Di Dong, PhD

Age: 18+

Sex: Any

Trial Phase: Academic

Sponsor: Chinese Academy of Sciences

Disqualifiers: Non-GI cancers, Multiple hospitals, others

No Placebo Group All trial participants will receive the active study treatment (no placebo)

Compensation: Varies

Number of Visits: Unknown

400 Participants Needed

What You Need to Know Before You Apply

What is the purpose of this trial?

This trial explores the usefulness of ChatGPT in planning treatments for certain gastrointestinal cancers, such as stomach or colon cancer. It compares treatment plans created by doctors alone with those developed using ChatGPT (AI-assisted treatment planning). Individuals diagnosed with these cancers and possessing detailed medical records may be suitable candidates. The trial aims to determine if AI can improve or streamline treatment planning. As an unphased trial, it offers patients the chance to contribute to innovative research that could enhance future cancer treatment planning.

Will I have to stop taking my current medications?

The trial information does not specify whether you need to stop taking your current medications.

Is there any evidence suggesting that this trial's treatments are likely to be safe?

Research shows that using ChatGPT to help plan treatment for gastrointestinal cancer is generally safe. Studies have found that ChatGPT can assist doctors in creating treatment plans that align well with those made by human doctors, indicating that ChatGPT's suggestions are safe and suitable. No reports of harmful side effects or major issues have emerged when doctors use ChatGPT for assistance.

It's important to understand that ChatGPT doesn't provide medical treatments directly. Instead, it helps doctors decide which treatments might be best. So far, studies have not identified any specific safety concerns with using AI tools like ChatGPT in this way. ^{1 2 3 4 5}

Why are researchers excited about this trial?

Researchers are excited about this trial because it explores the integration of AI, specifically ChatGPT, into the treatment planning process for gastrointestinal cancer. Unlike traditional methods where clinicians create treatment plans based solely on their expertise, the experimental approach uses AI to assist and potentially enhance decision-making. This could lead to more personalized and efficient treatment plans by analyzing vast amounts of medical data quickly. By comparing AI-assisted plans with those made by clinicians alone, researchers hope to understand if AI can improve outcomes for patients with gastrointestinal cancer.

What evidence suggests that this trial's treatments could be effective for gastrointestinal cancer?

This trial will compare AI-assisted treatment planning with traditional methods for gastrointestinal cancer. In the GPT-Assisted Group, participants receive treatment plans from clinicians using ChatGPT. Research has shown that AI tools like ChatGPT can help create treatment plans for stomach and intestinal cancer. One study found that ChatGPT's suggested treatment plans often matched those made by human medical teams, indicating that ChatGPT can offer advice similar to expert recommendations. Another study highlighted that it could help reduce the workload for cancer care providers, suggesting that AI can support busy healthcare workers. While these findings are promising, ChatGPT is meant to assist, not replace, human judgment in planning treatments. Meanwhile, the Control Group will receive treatment plans directly from clinicians without ChatGPT's assistance. ^{1 5 6 7 8}

Who Is on the Research Team?

Di Dong, PhD

Principal Investigator

Institute of Automation, Chinese Academy of Sciences

Are You a Good Fit for This Trial?

This trial is for adults over 18 with a confirmed diagnosis of gastrointestinal cancer, specifically gastric or colorectal cancer. They must be treated at the participating hospitals and have detailed medical records available. It's not open to those treated across multiple hospitals or with different types of cancers.

Inclusion Criteria

I have been diagnosed with stomach or colorectal cancer.

I have all my medical records available from before starting treatment.

I will receive all my treatment at a participating hospital.

Exclusion Criteria

Participants who receive treatment in multiple hospitals

My cancer is not a gastrointestinal type.

Timeline for a Trial Participant

Screening

Participants are screened for eligibility to participate in the trial

2-4 weeks

Retrospective Data Collection

Collection of medical records from over 1,200 patients with gastrointestinal cancers for exploratory and validation datasets

Not specified

Prospective Data Collection

Random allocation of 400 participants into control and ChatGPT-Assisted groups, followed by treatment plan recommendations and adjustments

Not specified

Follow-up

Participants are monitored for 3-year progression-free survival (PFS) and 5-year overall survival (OS) rates

5 years

What Are the Treatments Tested in This Trial?

Interventions

• ChatGPT-Assisted Treatment Plan
• Clinician-Directed Treatment Plan

Trial Overview The study is testing how well ChatGPT can help doctors create treatment plans for gastrointestinal cancers compared to standard clinician-directed methods. It will use both past and current patient data to assess the effectiveness.

How Is the Trial Designed?

Group I: GPT-Assisted Group Experimental Treatment 1 Intervention

In this arm, participants receive treatment plans from clinicians with the assistance of ChatGPT.

Group II: Control group Active Control 1 Intervention

In this arm, participants receive treatment plans directly from clinicians without the assistance of ChatGPT.

ChatGPT-Assisted Treatment Plan is already approved in European Union, United States for the following indications:

Approved in European Union as Lynparza for:

• Breast cancer
• Ovarian cancer
• Fallopian tube cancer
• Peritoneal cancer
• Pancreatic cancer
• Prostate cancer
• Endometrial cancer

Approved in United States as Lynparza for:

• Ovarian, fallopian tube, and primary peritoneal cancer
• Breast cancer
• Prostate cancer
• Pancreatic cancer

Who Is Running the Clinical Trial?

Chinese Academy of Sciences

Lead Sponsor

Trials

103

Recruited

34,050,000+

University Hospital Magdeburg, Germany

Collaborator

Trials

1

Recruited

400+

City of Hope Medical Center

Collaborator

Trials

614

Recruited

1,924,000+

San Raffaele University Hospital, Italy

Collaborator

Trials

34

Recruited

26,600+

Peking University Cancer Hospital (Inner Mongolia Campus)

Collaborator

Trials

1

Recruited

400+

Jiangmen Central Hospital

Collaborator

Trials

25

Recruited

28,500+

ZhuHai Hospital

Collaborator

Trials

23

Recruited

46,400+

Fifth Affiliated Hospital, Sun Yat-Sen University

Collaborator

Trials

102

Recruited

1,120,000+

Published Research Related to This Trial

ChatGPT demonstrated a strong ability to identify appropriate systemic therapies for advanced solid cancers, achieving a valid therapy quotient (VTQ) of 0.77 when compared to the National Comprehensive Cancer Network (NCCN) guidelines, based on 51 distinct cancer diagnoses.

While ChatGPT consistently provided at least one recommended treatment for each cancer type, its exact role in aiding oncologists and patients in treatment decisions is still unclear, suggesting the need for further research to evaluate its effectiveness in clinical settings.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Capacity of ChatGPT to Identify Guideline-Based Treatments for Advanced Solid Tumors. Schulte, B.[2023]

In a pilot study involving 10 breast cancer patients, ChatGPT provided general recommendations for treatment options but struggled with specific therapy recommendations, indicating its current limitations in clinical decision-making.

While ChatGPT could identify risk factors and suggest evaluations for treatment, it made errors in therapy recommendations, such as incorrectly suggesting antibody therapy for certain Her2 statuses, highlighting the need for caution in using AI for personalized patient care.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Evaluating ChatGPT as an adjunct for the multidisciplinary tumor board decision-making in primary breast cancer cases. Lukac, S., Dayan, D., Fink, V., et al.[2023]

Precision medicine in oncology faces challenges due to the complexity of decision-making, as human cognitive capacity is limited to considering only a few variables at a time, while there are many biomarkers and treatment options available.

Multifactorial decision support systems (DSSs) that utilize artificial intelligence can help overcome these limitations by integrating diverse data sources to predict outcomes and optimize treatment decisions, thereby improving efficacy and cost-effectiveness in cancer care.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Decision Support Systems in Oncology. Walsh, S., de Jong, EEC., van Timmeren, JE., et al.[2020]

The disease management program at M.D. Anderson Cancer Center involved nearly 1,000 patients and aimed to standardize care practices by developing evidence-based guidelines and collaborative care paths for different disease sites.

The program utilizes integrated information systems to implement guidelines in real-time patient care, which is expected to enhance care quality, support outcomes research, and reduce healthcare costs.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Implementation of guidelines and paths in oncology. Morris, M.[2007]

The Western Canadian Gastrointestinal Cancer Consensus Conference brought together a diverse group of healthcare professionals to discuss and develop recommendations for systemic therapy in patients with resectable metastatic colorectal cancer.

The conference aimed to optimize the sequence of treatments, highlighting the importance of collaboration among oncologists, pathologists, and other specialists in improving patient care for gastrointestinal cancers.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Systemic Therapy and Its Surgical Implications in Patients with Resectable Liver Colorectal Cancer Metastases. A Report from the Western Canadian Gastrointestinal Cancer Consensus Conference. Ahmed, S., Bosma, N., Moser, M., et al.[2022]

The implementation of multidisciplinary team (MDT) decisions significantly improved overall survival (OS) for patients with advanced gastrointestinal cancer, with an OS of 24 months for those whose treatment followed MDT recommendations compared to 17 months for those who did not.

The study, which included 455 patients and was conducted across 13 medical centers in China, found that MDT discussions particularly benefited patients with colorectal cancer, highlighting the importance of timely follow-up discussions when patient conditions change.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

China special issue on gastrointestinal tumors-Improved survival after multidisciplinary team decision for patients with advanced gastrointestinal cancer: A multicenter, noninterventional, controlled study. Zhang, Q., Zhou, Y., Song, L., et al.[2023]

In a preliminary analysis of 30 colorectal cancer cases, chatGPT™ demonstrated an 86.7% concordance rate with multidisciplinary tumor board recommendations, indicating its potential effectiveness in formulating oncological management plans.

The AI's recommendations adhered to oncological principles and considered patients' overall health and functional status, suggesting it could be a valuable tool in clinical decision-making for complex cancer cases.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Conversational artificial intelligence (chatGPT™) in the management of complex colorectal cancer patients: early experience. Choo, JM., Ryu, HS., Kim, JS., et al.[2023]

The study analyzed treatment recommendations for 302 breast cancer patients using IBM's Watson for Oncology (WFO), finding a 77% concordance rate for postoperative adjuvant therapy, indicating WFO's potential effectiveness in guiding treatment decisions.

However, the concordance dropped to 27.5% for metastatic cases, primarily due to differing chemotherapy regimen choices, highlighting the need for further adaptation of WFO to better suit the complexities of metastatic breast cancer treatment in Chinese patients.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Concordance between treatment recommendations provided by IBM Watson for Oncology and a multidisciplinary tumor board for breast cancer in China. Zhao, X., Zhang, Y., Ma, X., et al.[2020]

In a study involving 250 colorectal cancer patients, the treatment recommendations made by IBM's Watson for Oncology (WFO) showed a high concordance rate of 91% with those made by a multidisciplinary team, indicating strong agreement in treatment strategies.

Factors such as patient age, cancer stage, and previous therapies significantly influenced the concordance, suggesting that refining these aspects in AI systems could enhance their effectiveness in personalized cancer care.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Artificial Intelligence in Decision-Making for Colorectal Cancer Treatment Strategy: An Observational Study of Implementing Watson for Oncology in a 250-Case Cohort. Aikemu, B., Xue, P., Hong, H., et al.[2022]

In a study of 153 patients with potentially curable upper gastrointestinal cancers, 41.8% did not adhere to multidisciplinary tumor board (MDT) recommendations, primarily due to patient refusal and treatment complications.

Non-adherent patients had significantly worse median overall survival (19.5 months) compared to adherent patients, indicating that following MDT recommendations may be crucial for improving oncological outcomes.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Adherence to Multidisciplinary Tumor Board Recommendations in Patients With Curable Esophageal and Gastric Cancers. Soon, JJY., Zhao, Y., Shannon, NB., et al.[2023]

In a study of 330 patients with locally advanced oesophageal cancer, the 2-year survival rate for those receiving potentially curative treatment was 49%, with surgery alone showing the highest survival rates compared to other treatment combinations.

Definitive chemoradiotherapy (CRT) provided comparable overall survival to surgery alone for patients unfit for surgery, suggesting it is a viable alternative, although non-surgical therapies had a higher incidence of local relapses.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

A prospective comparison of multidisciplinary treatment of oesophageal cancer with curative intent in a UK cancer network. Adams, R., Morgan, M., Mukherjee, S., et al.[2022]

Citations

1. withpower.com withpower.com/trial/phase-gastrointestinal-neoplasms-7-2023-120d7

AI-Assisted Treatment Planning for Gastrointestinal Cancer This study will evaluate the utility of ChatGPT in recommending treatment plans for patients with gastrointestinal cancers, using both retrospective and ...

2. pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov/articles/PMC11763811/

AI-Enhanced Healthcare: Integrating ChatGPT-4 in ePROs ... This study offers preliminary insights into the feasibility of integrating AI tools like ChatGPT-4 into ePRO cancer care, highlighting its potential to reduce ...

3. mdpi.com mdpi.com/2227-9032/13/18/2254

Chat GPT Performance in Multi-Disciplinary Boards ... This document was uploaded to Chat-GPT 4o, and the model was asked to evaluate the patient and propose a treatment plan based solely on the provided data.

4. pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov/articles/PMC12470819/

Evaluating the Concordance Between ChatGPT and ... This retrospective study evaluates treatment recommendations generated by ChatGPT-4.0 and compares them with multidisciplinary team (MDT) ...

5. ai.jmir.org ai.jmir.org/2024/1/e50442/

Assessment of ChatGPT-3.5's Knowledge in Oncology - JMIR AI These data might have assisted ChatGPT in responding to USMLE ... Indeed, ChatGPT-3.5's lowest score was achieved in gastrointestinal cancer ...

6. pmc.ncbi.nlm.nih.gov pmc.ncbi.nlm.nih.gov/articles/PMC11797205/

Letter to the Editor: ChatGPT's contribution to cancer study ... data, AI models can assist in tailoring treatment plans to individual patients. They can help predict treatment outcomes, identify the most suitable ...

7. ejcancer.com ejcancer.com/article/S0959-8049(24)00756-1/fulltext

ChatGPT's Gastrointestinal Tumor Board Tango: A limping ... This study aimed to assess the consistency and replicability of treatment recommendations provided by ChatGPT 3.5 compared to gastrointestinal tumor cases.

8. mdpi.com mdpi.com/1718-7729/32/1/7

AI-Enhanced Healthcare: Integrating ChatGPT-4 in ePROs ... This study offers preliminary insights into the feasibility of integrating AI tools like ChatGPT-4 into ePRO cancer care, highlighting its potential to reduce ...