Large Language ModelsNCT07654036
Preliminary Evaluation of a Large Language Model-Based Tool for Complex Surgical Decision Support in Lung Cancer
A study testing GAPS-Agent and LLM in people with Large Language Models and Lung Cancer (NSCLC), run by Peking University People's Hospital at 1 site in China.
Trial summary
This study is an exploratory effect-size estimation study, with the following specific objectives: ① to estimate the point estimate and 95% confidence interval of the Win Ratio for the experimental group (GAPS-Agent) versus the control group (large language model) in blinded pairwise preference judgments by thoracic… Read more about this study
- Who is it for?
- Ages 18 to 65 years
- Conditions the study is for
- Large Language Models, Lung Cancer (NSCLC)
- What is being tested?
- GAPS-Agent and LLM
- Could I receive a placebo?
- No, everyone receives an active treatment
- How many people
- 8 participants
- Where
- 1 siteChina
- Study dates
- 10 Jun 2026to 21 Jun 2026
- Phase
- Interventional
- Are there results?
- None posted in the registry
Key requirements
Who can take part, as listed by the sponsor. Only the study team can confirm whether you are eligible, so take this list to your doctor.
- Ages 18 to 65 years
- All sexes
- No healthy volunteers
- Adult, Older adult
Who may be able to join
- Resident Physician Subjects:
- Holds a valid and legally effective Physician Practice License of the People's Republic of China;
- Currently holds the rank of resident physician in a thoracic surgery department at a tertiary Class A (3A) hospital;
- Agrees to complete all assessment tasks of the main study phase in accordance with the study protocol;
- Can guarantee the time and effort required to complete all assessment tasks of the main study.
- Study Cases:
Show all 15Show fewer
- The case was discussed at the Thoracic Oncology Multidisciplinary Team (MDT) conference of Peking University People's Hospital between January 2025 and May 2026;
- The current version of the NCCN guidelines does not provide an explicit recommendation covering the management of the case;
- Does not overlap with the GAPS evaluation set;
- The case is presented in pure text in a structured format, with all direct and indirect identifiers removed and complete de-identification performed prior to inclusion;
- From the pool of eligible cases, 12 cases will be randomly drawn using Python (numpy.random, with a fixed and archived seed) to serve as the main study cases. The cases will cover 6 themes (chest mass of undetermined diagnosis, early-stage lung cancer, locally advanced lung cancer, oligometastatic/oligoprogressive disease, special intraoperative situations, and tumor recurrence), with 1 - 4 cases per theme.
- Adjudication Expert Panel:
- Holds a valid and legally effective Physician Practice License of the People's Republic of China;
- Currently holds the rank of attending physician or above in a thoracic surgery department at a tertiary Class A hospital;
- Chairs or regularly participates in lung cancer multidisciplinary team (MDT) work in their department.
Who cannot join
- Resident Physician Subjects:
- Has previously participated in the construction of the GAPS evaluation set or the development of GAPS-Agent;
- Unable to complete the tasks of the study phase.
- Study Cases:
- Key case information is missing, such as text-form data on pathology (including IHC/NGS), imaging, laboratory tests, prior medical history, comorbidities, or PS score;
- Decision-making for the case is strictly dependent on non-text information.
Show all 9Show fewer
- Adjudication Expert Panel:
- Participated in the construction of the GAPS evaluation set, the content validity verification, or the development of GAPS-Agent for this study;
- Has a direct conflict of interest with any specific product among the two-arm tools of this study.
Trial locations and contacts
1 site in 1 country.
China1 site
Peking University People's Hospital
Beijing, Beijing Municipality, 100044
About this study
This study is an exploratory effect-size estimation study, with the following specific objectives: ① to estimate the point estimate and 95% confidence interval of the Win Ratio for the experimental group (GAPS-Agent) versus the control group (large language model) in blinded pairwise preference judgments by thoracic surgery expert adjudicators, to serve as a sample size planning parameter for subsequent multicenter confirmatory clinical trials; ② to preliminarily evaluate the value of GAPS-Agent within clinical workflows.The hypothesis of this study is as follows: compared with a general-purpose large language model without medical enhancement (control group), a structured agentic workflow optimized on the basis of the GAPS evaluation framework (GAPS-Agent, experimental group) can help junior resident physicians generate clinical decision plans for complex lung cancer cases that are more strongly preferred by senior thoracic surgery expert adjudicators.
- Conditions
- Large Language Models, Lung Cancer (NSCLC)
- Being tested
- GAPS-Agent, LLM
- Keywords
- Large Language Models, Lung Cancer, Multidisciplinary Team, Benchmark Test, Clinical Trial
- Medical subject headings
- Lung Neoplasms, Carcinoma, Non-Small-Cell Lung
What participants receive
What participants receive. In a randomised study each person is assigned to one group by chance.
GAPS-AgentOther
The research group has previously developed the GAPS evaluation framework for complex clinical decision-making in lung cancer. In this framework, G (Grounding) characterizes the cognitive depth of decision-making (ranging from knowledge retrieval to decisions that go beyond clinical guidelines), A (Authority) corresponds to the grading of evidence strength, P (Perturbation) describes the identification and management of real-world clinical confounding factors, and S (Strength) corresponds to the calibration of recommendation strength. Within this framework, the research group has completed the construction of a 100-item complex lung cancer decision-making evaluation set along with its corresponding rubrics, and has invited multiple thoracic oncology experts to complete content validity validation. Based on this, the research group developed GAPS-Agent, which uses an open-source large language model as its foundation and integrates functional modules such as guideline and evidence retri
LLMOther
Open source large language model that is not specifically enhanced in medical field.
| Study group | Role | What they receive |
|---|---|---|
| test arm | Experimental | GAPS-Agent Receives: Other: GAPS-Agent |
| control arm | Active comparator | LLM Receives: Other: LLM |
What is being measured
The primary outcome is the main question the study is designed to answer.
Overall plan Win Ratio
A total of 10 blinded expert judges made Win/Tie/Loss ternary preference judgments on 192 paired scheme comparisons in terms of overall scheme quality. The win ratio was calculated as Wins ÷ Losses, and the 95% confidence interval was estimated using a two-level (physician × case) cluster bootstrap resampling method (B = 10,000, quantile method on the log scale).
Measured: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
Secondary outcomes (11)
Inter-rater agreement
For the ternary preference judgment results of 10 expert judges across 192 paired comparisons and 6 evaluation domains, Fleiss' kappa was used to assess inter-rater agreement. The kappa value and its 95% confidence interval are reported for each evaluation domain.
Measured: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
Redundancy Win Ratio
A total of 10 blinded expert judges made Win/Tie/Loss ternary preference judgments on 192 paired scheme comparisons in terms of overall scheme quality. The win ratio was calculated as Wins ÷ Losses, and the 95% confidence interval was estimated using a two-level (physician × case) cluster bootstrap resampling method (B = 10,000, quantile method on the log scale).
Measured: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
Evidence-based medicine adherence Win Ratio
A total of 10 blinded expert judges made Win/Tie/Loss ternary preference judgments on 192 paired scheme comparisons in terms of overall scheme quality. The win ratio was calculated as Wins ÷ Losses, and the 95% confidence interval was estimated using a two-level (physician × case) cluster bootstrap resampling method (B = 10,000, quantile method on the log scale).
Measured: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
Actionability Win Ratio
A total of 10 blinded expert judges made Win/Tie/Loss ternary preference judgments on 192 paired scheme comparisons in terms of overall scheme quality. The win ratio was calculated as Wins ÷ Losses, and the 95% confidence interval was estimated using a two-level (physician × case) cluster bootstrap resampling method (B = 10,000, quantile method on the log scale).
Measured: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
Completeness Win Ratio
A total of 10 blinded expert judges made Win/Tie/Loss ternary preference judgments on 192 paired scheme comparisons in terms of overall scheme quality. The win ratio was calculated as Wins ÷ Losses, and the 95% confidence interval was estimated using a two-level (physician × case) cluster bootstrap resampling method (B = 10,000, quantile method on the log scale).
Measured: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
Safety Win Ratio
A total of 10 blinded expert judges made Win/Tie/Loss ternary preference judgments on 192 paired scheme comparisons in terms of overall scheme quality. The win ratio was calculated as Wins ÷ Losses, and the 95% confidence interval was estimated using a two-level (physician × case) cluster bootstrap resampling method (B = 10,000, quantile method on the log scale).
Measured: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
GAPS automated rubric score
A third-party large language model, independent of the two study arms' base models, served as the judge model and automatically scored all 96 plans according to the GAPS rubric.
Measured: Generated up to 3 weeks after residents finished their plan generation.
Subject physician's self-confidence score
After submitting each case plan, the participating physicians self-rated their confidence in their own plan using a 1-5 point Likert scale.
Measured: Completed at the time when residents submitted their plans. Calculated up to 3 weeks after the submission.
Tool satisfaction score
After submitting each case plan, the participating physicians rated their satisfaction with the tool using a 1-5 point Likert scale.
Measured: Completed at the time when residents submitted their plans. Calculated up to 3 weeks after the submission.
Tool trustworthiness score
After submitting each case plan, the participating physicians rated the tool's credibility using a 1-5 point Likert scale.
Measured: Completed at the time when residents submitted their plans. Calculated up to 3 weeks after the submission.
Decision-making time
The time taken (in minutes) by each participating physician to complete the production of each case plan was automatically recorded by the evaluation platform. Differences between groups were analyzed using a linear mixed-effects model.
Measured: Completed at the time when residents submitted their plans. Calculated up to 3 weeks after the submission.
Timeline
From the start of the study to its completion; the mark is today. Estimated dates are the sponsor’s plan.
- First posted17 Jun 2026
- Study start10 Jun 2026
- Primary completion20 Jun 2026
- Study completion21 Jun 2026
Resources
- Official record on ClinicalTrials.govThe registry entry this page is built from, with its full change history.
- Other trials for Large Language ModelsStudies of the same condition, open and under way.
- What the trial phases meanHow a treatment moves from first tests in people to approval.
More details
For doctors and researchers: how the study is designed, who runs it, and its registry IDs.
Study design
- Study type
- Interventional
- Phase
- Phase N/A
- Primary purpose
- Other
- Allocation
- Randomised
- Intervention model
- Parallel groups
- Masking
- Single blind
- Who is masked
- Outcomes assessor
- Enrolment
- 8 participants (actual)
Sponsor and investigators
- Lead sponsor
- Peking University People's Hospital
- Sponsor type
- University, hospital or other
- Responsible party
- Principal investigator, XiuYuan Chen, Peking University People's Hospital
IDs, dates and publications
- ClinicalTrials.gov ID
- NCT07654036
- Sponsor's protocol ID
- 2026PHB458-001
- Status last verified
- Jul 2026
- First submitted
- 10 Jun 2026
- Last updated
- 30 Jul 2026
- Results posted
- No
- Data monitoring committee
- No
- FDA-regulated drug
- No
- FDA-regulated device
- No
- Individual data sharing
- No
Study at a glance
From the registry record, last updated by the sponsor on 30 Jul 2026. Recruitment status is as the sponsor reported it: confirm with the study team.
- Official title
- Preliminary Evaluation of a Large Language Model-Based Tool for Complex Surgical Decision Support in Lung Cancer
- ClinicalTrials.gov ID
- NCT07654036
- Acronym / sponsor ID
- 2026PHB458-001
- Conditions
- Large Language Models, Lung Cancer (NSCLC)
- Keywords
- Large Language Models, Lung Cancer, Multidisciplinary Team, Benchmark Test, Clinical Trial
- Study type
- Interventional
- Recruitment status
- Completed
- Enrolment
- 8 participants (actual)
- Lead sponsor
- Peking University People's HospitalUniversity, hospital or other
- Study start
- 10 Jun 2026 (actual)
- Primary completion
- 20 Jun 2026 (actual)
- Study completion
- 21 Jun 2026 (actual)
- First posted
- 17 Jun 2026 (actual)
- Results
- Not posted in the source record
Clinical summary
The research question as the sponsor registered it. This describes what the study plans to test, not what it found.
This study is an exploratory effect-size estimation study, with the following specific objectives: ① to estimate the point estimate and 95% confidence interval of the Win Ratio for the experimental group (GAPS-Agent) versus the control group (large language model) in blinded pairwise preference judgments by thoracic surgery expert adjudicators, to serve as a sample size planning parameter for subsequent multicenter confirmatory clinical trials; ② to preliminarily evaluate the value of GAPS-Agent within clinical workflows.The hypothesis of this study is as follows: compared with a general-purpose large language model without medical enhancement (control group), a structured agentic workflow optimized on the basis of the GAPS evaluation framework (GAPS-Agent, experimental group) can help junior resident physicians generate clinical decision plans for complex lung cancer cases that are more strongly preferred by senior thoracic surgery expert adjudicators.
Study design
- Primary purpose
- Other
- Allocation
- Randomised
- Intervention model
- Parallel groups
- Masking
- Single blind (outcomes assessor)
- Number of arms
- 2
| Arm | Type | Receives |
|---|---|---|
| test arm | Experimental | Other: GAPS-AgentGAPS-Agent |
| control arm | Active comparator | Other: LLMLLM |
Who may qualify
- Ages
- 18 to 65 years
- Sex
- All
- Healthy volunteers
- Not accepted
- Age groups
- Adult, Older adult
Key inclusion criteria (15 listed)
- Resident Physician Subjects:
- Study Cases:
- Adjudication Expert Panel:
Key exclusion criteria (9 listed)
- Resident Physician Subjects:
- Study Cases:
- Adjudication Expert Panel:
Eligibility is decided by the study team after screening; this overview does not establish that a patient qualifies. Full inclusion and exclusion criteria.
Interventions
| Intervention | Type | Arms |
|---|---|---|
| GAPS-AgentThe research group has previously developed the GAPS evaluation framework for complex clinical decision-making in lung cancer. In this framework, G (Grounding) characterizes the cognitive depth of decision-making (ranging from knowledge retrieval to decisions that go beyond clinical guidelines), A (Authority) corresponds to the grading of evidence strength, P (Perturbation) describes the identification and management of real-world clinical confounding factors, and S (Strength) corresponds to the calibration of recommendation strength. Within this framework, the research group has completed the construction of a 100-item complex lung cancer decision-making evaluation set along with its corresponding rubrics, and has invited multiple thoracic oncology experts to complete content validity validation. Based on this, the research group developed GAPS-Agent, which uses an open-source large language model as its foundation and integrates functional modules such as guideline and evidence retri | Other | test arm |
| LLMOpen source large language model that is not specifically enhanced in medical field. | Other | control arm |
Primary outcome
Overall plan Win Ratio
Time frame: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
A total of 10 blinded expert judges made Win/Tie/Loss ternary preference judgments on 192 paired scheme comparisons in terms of overall scheme quality. The win ratio was calculated as Wins ÷ Losses, and the 95% confidence interval was estimated using a two-level (physician × case) cluster bootstrap resampling method (B = 10,000, quantile method on the log scale).
11 secondary outcomes
Inter-rater agreement
Time frame: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
Redundancy Win Ratio
Time frame: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
Evidence-based medicine adherence Win Ratio
Time frame: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
Actionability Win Ratio
Time frame: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
Completeness Win Ratio
Time frame: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
Safety Win Ratio
Time frame: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.
GAPS automated rubric score
Time frame: Generated up to 3 weeks after residents finished their plan generation.
Subject physician's self-confidence score
Time frame: Completed at the time when residents submitted their plans. Calculated up to 3 weeks after the submission.
Tool satisfaction score
Time frame: Completed at the time when residents submitted their plans. Calculated up to 3 weeks after the submission.
Tool trustworthiness score
Time frame: Completed at the time when residents submitted their plans. Calculated up to 3 weeks after the submission.
Decision-making time
Time frame: Completed at the time when residents submitted their plans. Calculated up to 3 weeks after the submission.
Results and safety
The study is completed, but results are not available in the source record. That does not mean the treatment did or did not work.
Research links
- Official record on ClinicalTrials.govLast updated by the sponsor 30 Jul 2026.
- Individual participant dataSharing planned: No
Related trials for Large Language Models
Not yet recruiting
CTEPH Differential Diagnosis Assisted by Enhanced Large Language Model Support
Chronic Thromboembolic Pulmonary Hypertension (CTEPH), Acute Pulmonary Embolism (PE)
Recruiting
Vision-Language-Based Large-Scale Model for Emergency Screening
Acute Thoraco-Abdominal Emergency Lesions
Not yet recruiting
Conversational AI for Adult Flu Vaccination Decisions
Influenza, Human, Vaccine Hesitancy
RecruitingObservational study
AI-generated Treatment Recommendations in Precision Oncology
Cancer
Not yet recruitingObservational study
Human-LLM Conversations as a Digital Signature of Cognitive Performance in Adults
Dementia
Referral and contact
AzTrials does not enrol participants or take referrals. Contact the study team or a site listed below; they decide who can take part. This study is not recruiting at the moment.
The record lists no contact. Use the official record to reach the sponsor.
1 site in 1 country: see every site and its status.
Questions about this trial
What is this clinical trial studying?
A study testing GAPS-Agent and LLM in people with Large Language Models and Lung Cancer (NSCLC), run by Peking University People's Hospital at 1 site in China.
The full registered description is in the About this study section and the official record.
Who may be eligible to take part?
The sponsor lists 15 inclusion and 9 exclusion criteria. Eligibility is decided by the study team after screening, so read the full criteria and contact a listed site.
Is this clinical trial currently recruiting?
The record says: finished, as of the sponsor's last update on 30 Jul 2026. Status can change between updates, so confirm with the study contact before making plans.
Where is the study taking place?
The record lists 1 site in China. Whether a site is taking part right now should be confirmed with that site. See every site.
What are the possible benefits and risks?
They depend on the study and the treatment, and no study can promise a benefit. The study team explains the known risks during informed consent, before anyone joins; the record itself does not list them.
Are results available for this study?
The study is completed, but results are not available in the source record. That does not mean the treatment did or did not work.
How can a doctor refer a patient or contact the research team?
The record lists no contact. The official record names the sponsor, who can be asked directly. AzTrials does not enrol participants or take referrals.
What does the primary outcome measure?
Overall plan Win Ratio (time frame: Measured at the time when experts completed their preference judgements. Calculated up to 3 weeks after the preference judgements.). It is what the study was designed to measure, not a finding.
Where can I read the original study information?
Does a listing mean the treatment is approved or proven to work?
No. A listing means the study is registered. It does not show that the treatment is safe, works, or is approved by a regulator.
How often is this information updated?
AzTrials copies changed records from ClinicalTrials.gov every night and reloads the whole registry every week. The sponsor last updated this record on 30 Jul 2026.