Overview
Patients with indeterminate lung nodules often face a choice between continued CT surveillance and invasive investigation. This trial compares management with or without access to an AI report that matches nodules across real serial CT scans, measures change, and estimates validated malignancy or invasiveness risks. The principal question is whether offering the report reduces lung resections that yield benign pathology over 24 months. Diagnostic delay, cancer stage, treatment opportunity, complications, and longer-term cancer outcomes are also assessed.
Description
This pragmatic superiority trial uses individual 1:1 randomization, concealed allocation, and blinded central outcome assessment. Eligible patients have persistent 6 to 30 mm nodules and at least two already available CT examinations. They are enrolled at an unresolved management decision before an invasive procedure or definitive treatment. A second CT must never be delayed or ordered solely to make an otherwise urgent case eligible.
Both arms receive structured longitudinal review, usual quantitative tools, the same nodule-management pathway, and the same appointment reminders. The intervention arm additionally receives the locked AI report. Clinicians retain responsibility for decisions and document patient preferences. Thus the comparison estimates the added effect of the AI-assisted strategy over an organized usual-care pathway, rather than over absent follow-up.
All randomized participants contribute to the primary analysis. Benign procedures are assessed per randomized patient rather than only among patients eventually classified as benign. This avoids making the primary denominator depend on postrandomization diagnostic work-up. Long-term surveillance also addresses the possibility that an apparent early reduction in surgery merely postpones diagnosis.
Eligibility
Inclusion Criteria:
\- 1. At least one persistent indeterminate pulmonary nodule measuring 6 to 30 mm on the latest clinically obtained CT; solid, part-solid, and pure ground-glass nodules are eligible. 2. Two CT examinations in total are already available: the latest within 30 days before randomization and an earlier examination 3 to 24 months before it. Both meet the frozen image-quality requirements for longitudinal comparison. 3. The current decision between continued surveillance and further investigation remains unresolved, with no mandatory urgent procedure or committed definitive treatment before randomization.
4\. The patient is potentially eligible for curative treatment if cancer is diagnosed, with an estimated life expectancy of at least five years independent of the study nodule. 5. Written informed consent and willingness to permit record retrieval and five-year contact.
Exclusion Criteria:
\- 1.Definitive diagnosis or previous treatment of the target nodule; overt metastatic disease, suspicious adenopathy requiring urgent work-up, or another indication making delayed evaluation inappropriate. 2.Active malignancy or cancer treatment within the preceding five years, except definitively treated nonmelanoma skin cancer or in situ cervical cancer.
3\. Clinically important immunocompromise or active pulmonary infection requiring a separate diagnostic pathway. 4. Current participation in an organized lung-cancer screening program with a mandated management pathway incompatible with this trial. 5. CT quality or acquisition differences prevent valid longitudinal evaluation; a clearly benign lesion does not require a diagnostic management decision. 6. Known pregnancy; a conflicting interventional trial; prior randomization in this study; or consent cannot be obtained.
