Germany's national lung cancer screening program launched in April 2026 with a requirement that no other European screening program has: AI-assisted detection is mandatory at every reading stage. But what does that actually mean in practice? Here's where AI fits — step by step.
The Screening Pathway at a Glance
Germany's program follows a structured pathway defined by the G-BA and covered by 8 new EBM billing codes:
GP Qualification → LDCT Scan → Primary Reading (incl. AI) → Second Reading if Lung-RADS ≥ 3 (incl. AI) → Consensus → Follow-up
AI is not a separate step in this workflow. It supports the reading process at both the primary and secondary reading stages — a distinction that matters for how institutions plan their infrastructure.
AI Touchpoint #1: Primary Reading
Every LDCT screening scan must be interpreted with qualified CAD software. This is not a recommendation — it is a legal requirement under the Lung Cancer Early Detection Ordinance (LuKrFrühErkV).
In practice, this means the radiologist's reading is supported by AI-assisted detection that provides:
•
Automated nodule detection: Identification of pulmonary nodules across the full scan
•
Volumetric measurement: 3D volume calculation for each detected nodule — not just diameter
•
Classification: Categorization by size, density, morphology, and risk characteristics
•
Structured output: Standardized analysis results ready for reporting
The radiologist retains full clinical judgment. AI provides a quantitative, reproducible baseline that supports — not replaces — the reading process. The result is a standardized starting point regardless of which site or which reader performs the primary reading.
This is where
AVIEW LCS Plus operates: installed at the primary reading site, it delivers AI-assisted detection as part of the first reading workflow.
AI Touchpoint #2: Dual-Reading Support
Not every case requires a second reading. Under the German framework, a second reading is triggered only when findings are classified as Lung-RADS ≥ 3 — meaning the scan shows findings that require monitoring or further workup.
When triggered, the case moves to an independent second reader at a certified lung cancer center. This second reading must also be performed with CAD software support.
Here is where the operational challenge emerges: primary and secondary readers are often at different institutions, with different IT environments, different workstation configurations, and potentially different detection tools. Without standardization, two readers looking at the same scan may detect different nodules, measure differently, and reach different conclusions.
AI solves this by providing the same analysis baseline to both readers. Both the primary and secondary reader work from the same AI-generated detection and measurement data — ensuring that the independent second opinion is built on consistent, quantitative evidence rather than subjective visual assessment alone.
The second reading takes place on AVIEW:HUB, installed at the certified lung cancer center. The case — including the AI analysis baseline generated at the primary site — is transmitted to the second reading site, where the certified reader accesses it through AVIEW:HUB's web-based viewer. First readers work exclusively on AVIEW; second readers work exclusively on AVIEW:HUB. This separation simplifies user management and access control across institutions.
After the second reading, AVIEW:HUB routes cases to a consensus stage for final verification, including cases with discrepant results, with full audit documentation.
AI Touchpoint #3: Longitudinal Tracking
Lung cancer screening is not a one-time event. Patients return for follow-up LDCT scans — typically at 12-month intervals — and the clinical question becomes: have the nodules changed?
This is covered by billing code GOP 01872 and requires comparison with baseline findings. Manually matching nodules across scans, recalculating volumes, and assessing growth rates is time-consuming and subject to inter-reader variability — especially at national scale across hundreds of sites.
AI automates this process:
•
Automatic nodule matching: Previously identified nodules are matched across the baseline and follow-up scans
•
Volumetric change calculation: Growth or regression is quantified objectively over time
•
Clinical decision support: The output supports the radiologist's decision on whether to continue monitoring or initiate further workup
This longitudinal tracking capability is a core function of AVIEW LCS Plus and becomes increasingly critical as screening programs scale — with millions of scans requiring consistent follow-up year after year.
The Scale Challenge: Why AI Is the Standardization Layer
Germany's screening program is not a pilot. It is national infrastructure designed to operate across hundreds of sites with different CT scanners, different radiologists, different protocols, and different levels of experience.
Reimbursement is set. The regulatory framework is set. The operational question is: how do you deliver consistent quality across all of these variables?
AI provides the standardization layer:
| What varies by site |
What AI standardizes |
| CT scanner brand and model |
Nodule detection criteria |
| Acquisition protocols |
Measurement methodology |
| Reader experience level |
Volume calculation |
| Workstation configuration |
Growth rate assessment |
| Local reporting habits |
Structured output format |
Without this layer, every site is a variable. With it, every site operates on the same quantitative baseline — which is what the national program's dual-reading and consensus requirements demand.
This Has Already Been Proven
The workflows described above are not theoretical. They have been validated in practice:
•
HANSE Pilot Program (2022): AVIEW LCS Plus was the sole AI software deployed across 3 certified lung cancer centers in the government-led screening study that directly informed the G-BA's decision to establish the national program. 5,000 patients over 2 years.
•
Production deployments across Germany: Charité – Universitätsmedizin Berlin, MHH Hannover, LMU Munich, Cologne University Hospital, and more — representing over 60% of Germany’s top 10 hospitals.
•
Pan-European screening programs: Italy (RISP), France (IMPULSION), the EU 4-IN THE LUNG RUN trial, and expansion into national lung cancer screening programs.
What This Means for Your Institution
If your institution is planning to participate in Germany's national LDCT screening program, three questions matter:
1. Does your AI solution support both reading stages?
The LuKrFrühErkV requires CAD software at primary and secondary readings. You need infrastructure that covers both.
2.
Can your workflow scale across sites?
If you're part of a hospital group or radiology network, consistency across locations isn't optional — it's a reimbursement requirement.
3. Is longitudinal tracking built in?
Screening is a recurring program. Your infrastructure needs to handle not just the first scan, but every follow-up — with objective comparison over time.
For answers to the most common questions about AI infrastructure for Germany's screening program, see our
FAQ: What Hospitals Need to Know About AI Software.
For the policy background, see our analysis: Germany's Lung Cancer Screening Is Now Fully Reimbursed — And AI Is Built Into the System.
• Contact: Coreline Europe GmbH
• E: cle@corelinesoft.com
List
