Pulmonary Subsolid NodulesVF-PNNCT07900815
Virtual Follow-up Prediction of Subsolid Pulmonary Nodules: A Prospective Silent Reader Study
An observational study of Pulmonary Subsolid Nodules and Ground-Glass Nodules, run by Union Hospital, Tongji Medical College, Huazhong University of Science and Technology at 1 site in China.
Trial summary
The goal of this prospective, non-interventional study is to evaluate the predictive performance and potential clinical utility of an artificial intelligence (AI)-based virtual follow-up framework for incidentally detected pulmonary subsolid nodules (SSNs), including pure ground-glass and part-solid nodules. Read more about this study
- Who is it for?
- Ages 18 years and older
- Conditions the study is for
- Pulmonary Subsolid Nodules, Ground-Glass Nodules
- How many people
- 200 participants
- Where
- 1 siteChina
- Study dates
- 1 Oct 2024to 15 Jun 2026
- How long will I take part?
- 12 Months
- Phase
- Observational
- 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 years and older
- All sexes
- No healthy volunteers
- Adult, Older adult
Study population: Eligible patients with an index pulmonary subsolid nodule and two historical clinically acquired CT examinations; all receive ordinary clinical care.
Who may be able to join
- Adult patients with at least one persistent pure ground-glass or part-solid pulmonary nodule on T1 CT.
- At least two clinically obtained thin-section chest CT examinations (T0, T1), with adequate longitudinal lesion matching and image quality.
- Relevant clinical variables available at T1, and no T2 or pathology outcome known to the predictor or reader at lock time.
- Eligible for ordinary clinical surveillance and outcome collection according to the ethics-approved protocol.
Who cannot join
- Target nodule previously resected, ablated or otherwise treated before T1; clearly non-target pulmonary disease or uninterpretable CT.
- No valid paired scans, nodule matching or model inference due to prespecified image-quality failure.
- Reference outcome missing: retain in cohort flow and label missing rather than classify as stable.
- Intercurrent treatment before T2: exclude from primary natural-evolution imaging analysis, but document separately.
Trial locations and contacts
1 site in 1 country.
China1 site
Union Hospital, Tongji Medical College, Huazhong University of Science and Technology
Wuhan, Hubei, 430000
About this study
Official title: Prospective Multicenter Validation of AI-Generated Future CT Images for Subsolid Pulmonary Nodule Progression and a Nested Silent Reader Evaluation of Simulated Surveillance Decisions
The goal of this prospective, non-interventional study is to evaluate the predictive performance and potential clinical utility of an artificial intelligence (AI)-based virtual follow-up framework for incidentally detected pulmonary subsolid nodules (SSNs), including pure ground-glass and part-solid nodules.
The main questions it aims to answer are:
Read more
- Can the AI model accurately predict future CT appearances, quantitative growth characteristics, and morphological progression of pulmonary SSNs at specified follow-up intervals?
- Does AI-assisted virtual follow-up improve radiologists' assessment of nodule progression and simulated surveillance management decisions compared with assessment without AI assistance?
Researchers will compare radiologists' assessments with and without access to AI-generated future CT images and quantitative predictions. An additional reader evaluation will examine whether generated future images provide incremental value beyond quantitative predictions alone.
Participants will undergo clinically indicated CT examinations and receive routine care according to standard clinical practice. The research team will:
- Collect previously acquired CT images and clinical information.
- Generate and lock AI predictions before the corresponding follow-up CT examinations occur.
- Obtain and lock radiologists' assessments and simulated management recommendations before actual follow-up outcomes become available.
- Compare predicted nodule progression with observed follow-up CT findings.
AI predictions and research-related reader assessments will not be disclosed to treating clinicians or used to alter patients' actual surveillance or treatment decisions.
Read the detailed description
Prospective, multicenter, non-interventional prognostic-model validation with a nested, blinded, repeated-reader simulated decision experiment. At the T1 index scan, the frozen VF-PN model is applied to T0 and T1 CT plus prespecified clinical variables. Predictions and reader responses are recorded and time-stamped before true follow-up CT (T2) or subsequent outcome information becomes available. A clinically obtained T2 scan is used for independent validation. Study results are not returned to the clinical team. No CT is ordered or delayed solely for research.
Main 200-participant reader study (Cohort A): radiologists first review T0/T1 and available clinical information without AI, lock assessments, and after a four-week interval re-review in a randomized case order with full VF-PN output. Readers remain blinded to actual T2 and prior responses. This fixed condition order introduces possible period/re-reading effects; report it transparently and use a repeat-reading control if prospectively implemented. The number of readers and allocations must match the actual completed experiment.
Prospectively specified intervals for reader display: 3, 6 and 12 months after T1; an optional 24-month horizon is included in the source experimental plan but must be confirmed for final registration. For objective follow-up evaluation, a true T2 is matched to one prespecified time window: 3 months (2-4), 6 months (5-8), 12 months (9-15), or, if retained, 24 months (18-30). Resolve overlap by closest target month.
Ancillary study of image authenticity: 100 independent retrospective cases, candidate true T2 versus generated Tpred image (balanced 1:1 when implemented); four independent radiologists classify origin and score technical quality, morphological authenticity, longitudinal plausibility and usability. Identical display/postprocessing should be applied to both sources. This experiment is ancillary, not a prospective clinical outcome trial.
Optional incremental-value reader comparison: no AI versus quantitative AI only versus complete AI (quantitative output plus generated future CT). Include only if approved and actually undertaken, with the final sample size, reader count and allocation scheme predeclared in a protocol amendment. Do not assume it is already completed.
- Conditions
- Pulmonary Subsolid Nodules, Ground-Glass Nodules
- Keywords
- future CT image forecasting, pulmonary subsolid nodules, artificial intelligence, reader study, simulated surveillance management, visual authenticity
What is being measured
The primary outcome is the main question the study is designed to answer.
Change in radiologists' future nodule growth discrimination (AUC)
Within-reader difference between no-AI and complete-AI assessments using a prespecified 1-6 growth-likelihood score, compared against actual T2 growth (mass increase >30% from T1). Unit: AUC difference with 95% CI.
Measured: Reader assessments are recorded and locked before the corresponding follow-up CT outcome is available; reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.
Secondary outcomes (9)
Simulated management agreement / appropriateness
Proportion of reader recommendations consistent with independently expert-adjudicated acceptable management reference, compared between modes. Specify whether expert assessment is outcome-blinded or outcome-informed; do not call this objective clinical ground truth.
Measured: After the index CT examination and before the corresponding follow-up CT outcome is available at the matched 3-, 6-, or 12-month follow-up window.
Potential under- and over-management
Proportion of recommendations judged potentially insufficient or excessive; AI-induced favorable and unfavorable decision changes, determined with prespecified blinded review.
Measured: After the index CT examination and before the corresponding follow-up CT outcome is available at the matched 3-, 6-, or 12-month follow-up window.
Quantitative growth forecasting
MAE and relative error of predicted maximal diameter, volume, mass, solid-component maximal diameter, volume/mass growth rate vs true T2.
Measured: Reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.
Growth classification
Sensitivity, specificity, F1, accuracy, PPV, NPV and kappa; prespecified mass-growth threshold >30%.
Measured: Reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.
Radiologist assessment
Radiologist performance is evaluated by comparing reader assessments with the corresponding reference outcomes using AUC and accuracy. Prediction of pathological invasiveness is evaluated only in the subgroup with pathologically verified diagnoses. Reader confidence scores and interpretation time are also recorded to characterize assessment confidence and efficiency.
Measured: Reader assessments are recorded and locked before the corresponding follow-up CT outcome is available; reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.
Generation quality
The quality of AI-generated follow-up CT images is evaluated against the corresponding actual follow-up CT images at the matched prediction interval. ROI MAE quantifies absolute image-intensity differences. SSIM assesses structural similarity and PSNR assesses image reconstruction fidelity. The Dice similarity coefficient evaluates spatial overlap between the corresponding segmentation masks. Each metric is calculated separately according to its defined measurement scale, and composite image-quality score is calculated.
Measured: Reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.
Prediction uncertainty
Prediction uncertainty is evaluated by assessing the calibration and informativeness of AI-generated prediction intervals and their relationship with observed prediction errors. Interval coverage is defined as the proportion of observed follow-up measurements falling within the corresponding predicted intervals. Interval width quantifies the range of each predicted interval. The association between estimated uncertainty and observed prediction error is assessed to determine whether greater predicted uncertainty corresponds to larger prediction errors.
Measured: Reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.
Image authenticity
In the independent 100-case ancillary study: source recognition accuracy, synthetic-as-real rate, rating of quality, authenticity, morphological plausibility and clinical usability.
Measured: After enrollment and before the follow-up CT outcome is available at the matched 3-, 6-, or 12-month follow-up window.
Incremental-value reader study
If performed: compare complete AI vs quantitative-only AI, with no-AI baseline; prespecify its primary contrast and sample separately.
Measured: After the index CT examination and before the corresponding follow-up CT outcome is available at the matched 3-, 6-, or 12-month follow-up window.
Timeline
From the start of the study to its completion; the mark is today. Estimated dates are the sponsor’s plan.
- First posted28 Oct 2026
- Study start1 Oct 2024
- Primary completion24 Feb 2026
- Study completion15 Jun 2026
Resources
- Official record on ClinicalTrials.govThe registry entry this page is built from, with its full change history.
- Other trials for Pulmonary Subsolid NodulesStudies of the same condition, open and under way.
- More trials run by Union Hospital, Tongji Medical College, Huazhong University of Science and TechnologyThe sponsor's other registered studies.
- 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
- Observational
- Observational model
- Cohort
- Time perspective
- Prospective
- Enrolment
- 200 participants (actual)
- Target duration
- 12 Months
Sponsor and investigators
- Sponsor type
- University, hospital or other
- Collaborators
- Hunan Shaoyang Central Hospital, Sun Yat-Sen Memorial Hospital of Sun Yat-Sen University, Tianyou Hospital Affiliated to Wuhan University of Science and Technology, Xiangya Hospital of Central South University, Jinhua Central Hospital, The First Affiliated Hospital of Zhengzhou University, Yingcheng People's Hospital, Henan Provincial People's Hospital, Shanghai Pulmonary Hospital, Shanghai, China
- Responsible party
- Sponsor
IDs, dates and publications
- ClinicalTrials.gov ID
- NCT07900815
- Sponsor's protocol ID
- WHUH-VFPN-SILENT-2026-01
- Acronym
- VF-PN
- Status last verified
- Oct 2026
- First submitted
- 12 Oct 2026
- Last updated
- 28 Oct 2026
- Results posted
- No
- Individual data sharing
- Yes: Consider sharing deidentified tabular outcomes and analysis code on controlled request; raw CT images subject to separate governance.
Study at a glance
From the registry record, last updated by the sponsor on 28 Oct 2026. Recruitment status is as the sponsor reported it: confirm with the study team.
- Official title
- Prospective Multicenter Validation of AI-Generated Future CT Images for Subsolid Pulmonary Nodule Progression and a Nested Silent Reader Evaluation of Simulated Surveillance Decisions
- ClinicalTrials.gov ID
- NCT07900815
- Acronym / sponsor ID
- VF-PN / WHUH-VFPN-SILENT-2026-01
- Conditions
- Pulmonary Subsolid Nodules, Ground-Glass Nodules
- Keywords
- future CT image forecasting, pulmonary subsolid nodules, artificial intelligence, reader study, simulated surveillance management, visual authenticity
- Study type
- Observational
- Recruitment status
- Completed
- Enrolment
- 200 participants (actual)
- Lead sponsor
- Union Hospital, Tongji Medical College, Huazhong University of Science and TechnologyUniversity, hospital or other
- Collaborators
- Hunan Shaoyang Central Hospital, Sun Yat-Sen Memorial Hospital of Sun Yat-Sen University, Tianyou Hospital Affiliated to Wuhan University of Science and Technology, Xiangya Hospital of Central South University, Jinhua Central Hospital, The First Affiliated Hospital of Zhengzhou University, Yingcheng People's Hospital, Henan Provincial People's Hospital, Shanghai Pulmonary Hospital, Shanghai, China
- Study start
- 1 Oct 2024 (actual)
- Primary completion
- 24 Feb 2026 (actual)
- Study completion
- 15 Jun 2026 (actual)
- First posted
- 28 Oct 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.
The goal of this prospective, non-interventional study is to evaluate the predictive performance and potential clinical utility of an artificial intelligence (AI)-based virtual follow-up framework for incidentally detected pulmonary subsolid nodules (SSNs), including pure ground-glass and part-solid nodules.
The main questions it aims to answer are:
- Can the AI model accurately predict future CT appearances, quantitative growth characteristics, and morphological progression of pulmonary SSNs at specified follow-up intervals?
- Does AI-assisted virtual follow-up improve radiologists' assessment of nodule progression and simulated surveillance management decisions compared with assessment without AI assistance?
Researchers will compare radiologists' assessments with and without access to AI-generated future CT images and quantitative predictions. An additional reader evaluation will examine whether generated future images provide incremental value beyond quantitative predictions alone.
Participants will undergo clinically indicated CT examinations and receive routine care according to standard clinical practice. The research team will:
- Collect previously acquired CT images and clinical information.
- Generate and lock AI predictions before the corresponding follow-up CT examinations occur.
- Obtain and lock radiologists' assessments and simulated management recommendations before actual follow-up outcomes become available.
- Compare predicted nodule progression with observed follow-up CT findings.
AI predictions and research-related reader assessments will not be disclosed to treating clinicians or used to alter patients' actual surveillance or treatment decisions.
Detailed description
Prospective, multicenter, non-interventional prognostic-model validation with a nested, blinded, repeated-reader simulated decision experiment. At the T1 index scan, the frozen VF-PN model is applied to T0 and T1 CT plus prespecified clinical variables. Predictions and reader responses are recorded and time-stamped before true follow-up CT (T2) or subsequent outcome information becomes available. A clinically obtained T2 scan is used for independent validation. Study results are not returned to the clinical team. No CT is ordered or delayed solely for research.
Main 200-participant reader study (Cohort A): radiologists first review T0/T1 and available clinical information without AI, lock assessments, and after a four-week interval re-review in a randomized case order with full VF-PN output. Readers remain blinded to actual T2 and prior responses. This fixed condition order introduces possible period/re-reading effects; report it transparently and use a repeat-reading control if prospectively implemented. The number of readers and allocations must match the actual completed experiment.
Prospectively specified intervals for reader display: 3, 6 and 12 months after T1; an optional 24-month horizon is included in the source experimental plan but must be confirmed for final registration. For objective follow-up evaluation, a true T2 is matched to one prespecified time window: 3 months (2-4), 6 months (5-8), 12 months (9-15), or, if retained, 24 months (18-30). Resolve overlap by closest target month.
Ancillary study of image authenticity: 100 independent retrospective cases, candidate true T2 versus generated Tpred image (balanced 1:1 when implemented); four independent radiologists classify origin and score technical quality, morphological authenticity, longitudinal plausibility and usability. Identical display/postprocessing should be applied to both sources. This experiment is ancillary, not a prospective clinical outcome trial.
Optional incremental-value reader comparison: no AI versus quantitative AI only versus complete AI (quantitative output plus generated future CT). Include only if approved and actually undertaken, with the final sample size, reader count and allocation scheme predeclared in a protocol amendment. Do not assume it is already completed.
Study design
- Observational model
- Cohort
- Time perspective
- Prospective
- Target follow-up
- 12 Months
Who may qualify
- Ages
- 18 years and older
- Sex
- All
- Healthy volunteers
- Not accepted
- Age groups
- Adult, Older adult
- Study population
- Eligible patients with an index pulmonary subsolid nodule and two historical clinically acquired CT examinations; all receive ordinary clinical care.
Key inclusion criteria (4 listed)
- Adult patients with at least one persistent pure ground-glass or part-solid pulmonary nodule on T1 CT.
- At least two clinically obtained thin-section chest CT examinations (T0, T1), with adequate longitudinal lesion matching and image quality.
- Relevant clinical variables available at T1, and no T2 or pathology outcome known to the predictor or reader at lock time.
- Eligible for ordinary clinical surveillance and outcome collection according to the ethics-approved protocol.
Key exclusion criteria (4 listed)
- Target nodule previously resected, ablated or otherwise treated before T1; clearly non-target pulmonary disease or uninterpretable CT.
- No valid paired scans, nodule matching or model inference due to prespecified image-quality failure.
- Reference outcome missing: retain in cohort flow and label missing rather than classify as stable.
- Intercurrent treatment before T2: exclude from primary natural-evolution imaging analysis, but document separately.
Eligibility is decided by the study team after screening; this overview does not establish that a patient qualifies. Full inclusion and exclusion criteria.
Primary outcome
Change in radiologists' future nodule growth discrimination (AUC)
Time frame: Reader assessments are recorded and locked before the corresponding follow-up CT outcome is available; reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.
Within-reader difference between no-AI and complete-AI assessments using a prespecified 1-6 growth-likelihood score, compared against actual T2 growth (mass increase >30% from T1). Unit: AUC difference with 95% CI.
9 secondary outcomes
Simulated management agreement / appropriateness
Time frame: After the index CT examination and before the corresponding follow-up CT outcome is available at the matched 3-, 6-, or 12-month follow-up window.
Potential under- and over-management
Time frame: After the index CT examination and before the corresponding follow-up CT outcome is available at the matched 3-, 6-, or 12-month follow-up window.
Quantitative growth forecasting
Time frame: Reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.
Growth classification
Time frame: Reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.
Radiologist assessment
Time frame: Reader assessments are recorded and locked before the corresponding follow-up CT outcome is available; reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.
Generation quality
Time frame: Reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.
Prediction uncertainty
Time frame: Reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.
Image authenticity
Time frame: After enrollment and before the follow-up CT outcome is available at the matched 3-, 6-, or 12-month follow-up window.
Incremental-value reader study
Time frame: After the index CT examination and before the corresponding follow-up CT outcome is available at the matched 3-, 6-, or 12-month follow-up window.
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 28 Oct 2026.
- Individual participant dataSharing planned: Yes. Consider sharing deidentified tabular outcomes and analysis code on controlled request; raw CT images subject to separate governance.
Related trials for Pulmonary Subsolid Nodules
Not yet recruitingObservational study
A Prospective, Non-interventional Cohort Study of Subsolid Pulmonary Nodules
Solitary Lung Nodule, Multiple Pulmonary Nodules
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?
An observational study of Pulmonary Subsolid Nodules and Ground-Glass Nodules, run by Union Hospital, Tongji Medical College, Huazhong University of Science and Technology 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 4 inclusion and 4 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 28 Oct 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?
Change in radiologists' future nodule growth discrimination (AUC) (time frame: Reader assessments are recorded and locked before the corresponding follow-up CT outcome is available; reference outcomes are assessed using follow-up CT obtained at the matched 3-, 6-, or 12-month follow-up window after the index CT examination.). 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 28 Oct 2026.