LDCT lung-cancer screening is a delivery system, not a single scan: why mortality benefit depends on adherence and follow-up closure

Lung-cancer screening is often narrated as a technology decision: do one low-dose CT, find nodules earlier, save lives. The trial record says something more operational. Mortality benefit appears when the whole chain holds: high-risk selection, repeated annual attendance, controlled follow-up for indeterminate findings, and treatment access for curable-stage disease.

The practical consequence is straightforward. LDCT is not a one-time diagnostic upgrade. It behaves like a delivery system with failure points at every handoff.

The mechanism chain: risk enrichment → stage shift → mortality effect

The strongest evidence comes from randomized trials in heavy-smoking, older cohorts rather than broad low-risk populations.

In NLST, 53,454 high-risk participants were randomized to three annual rounds of LDCT versus chest radiography. Lung-cancer mortality was 247 vs 309 deaths per 100,000 person-years, a 20.0% relative reduction with LDCT, and all-cause mortality was 6.7% lower in the LDCT arm.[1]

NELSON, which used volume-based CT protocols, reported a similar directional signal. Among men, lung-cancer mortality at 10 years was 2.50 vs 3.30 deaths per 1,000 person-years (rate ratio 0.76); screening adherence averaged 90.0%.[2]

Mechanistically, this is a denominator-and-timing effect:

Risk enrichment concentrates screening in people with higher baseline event probability.
Stage shift moves a fraction of cancers into a treatment window where surgery or curative-intent therapy is realistic.
Time compression reduces delay between detectable lesion and intervention.
Cumulative repetition through annual rounds captures incident disease that a single baseline scan would miss.

When these steps stay connected, mortality curves separate. When one step breaks, the benefit attenuates.

Why false positives do not invalidate screening, but force pathway discipline

A common misunderstanding is to treat false-positive burden as proof that screening fails. The trial data are more nuanced.

In NLST, positive screens were frequent in LDCT (24.2% across rounds), and most positive results were ultimately false positives (96.4% in the LDCT arm).[1] This does not erase mortality benefit; it explains why workflow design matters.

The key operational variable is not the existence of indeterminate findings. It is how the program resolves them:

serial imaging when appropriate,
clear nodule management protocols,
escalation only when probability thresholds are crossed.

NELSON’s lower downstream burden illustrates this pathway effect: average additional CT scans for initially indeterminate findings were 9.2%, and referral for suspicious nodules was 2.1%.[2]

In other words, screening quality is less about “zero false positives” and more about preventing false positives from becoming uncontrolled invasive cascades.

Eligibility and interval are the first implementation gate

USPSTF’s current recommendation targets adults 50–80 years, with at least 20 pack-years, current smoking or quit within 15 years, and annual LDCT until stopping criteria are met.[3] The guideline logic is risk-concentration: net benefit depends on keeping screening within populations where event probability can justify harms.

Program leakage starts early when eligibility capture is weak. USPSTF’s evidence review notes that uptake has remained low; one multi-state estimate reported only 14.4% of people eligible under prior criteria received screening in the prior 12 months.[3]

Low uptake means trial-level efficacy cannot be translated into population impact, even when scanners are available.

The under-discussed multiplier: smoking cessation inside screening workflows

LDCT and cessation are often run as parallel tracks. Evidence-based policy treats them as one pathway. USPSTF and CDC both place cessation intervention inside screening encounters for current smokers.[3][4]

This changes the mechanism from “detect earlier” to “detect earlier while reducing future incidence pressure.”

Without cessation integration, programs can still generate stage-shift gains. With integration, they also reduce the incoming risk pool over time. That is a structurally larger effect, especially in systems planning over 5–10 years.

Where uncertainty still sits

The core mortality signal is robust in high-risk trial populations, but boundaries remain explicit:

Trial cohorts were selected and relatively healthy enough to complete repeated screening and potential curative care.[1][2]
Real-world systems with weak follow-up capacity may reproduce harms faster than benefits.
Overdiagnosis and radiation exposure remain non-zero risks and must stay in informed consent and audit loops.[4]

An actionable invalidation condition is clear: if a program cannot maintain annual adherence, nodule-workup quality, and treatment closure, expected mortality benefit should be discounted rather than assumed.

A 90-day implementation frame for health systems

A practical first quarter can be organized around closure metrics:

Eligibility capture: build structured smoking-history intake (pack-years + quit year) in primary care templates.
Shared decision workflow: standardize benefits/harms counseling before first scan.
Nodule pathway governance: adopt one explicit protocol for repeat imaging and escalation thresholds.
Cessation bundle: auto-trigger counseling/pharmacotherapy referral for current smokers.
Closure dashboard: track four rates monthly—enrollment among eligible patients, annual return adherence, diagnostic-resolution interval, and treatment initiation for screen-detected cancers.

If those four rates improve, screening behaves like the trial mechanism. If they drift, the program becomes expensive imaging without proportional survival return.

Bottom line

LDCT lung-cancer screening works in evidence terms, but it does not work as an isolated scan. It works as a managed sequence.

The durable policy question for 2026 is not whether to buy scanners. It is whether health systems can run a high-fidelity chain from eligibility through follow-up and treatment, while embedding cessation support at the point of contact. That sequence, not the machine alone, is where mortality reduction is won or lost.

cronfeed.work