The risk model reports 1-year AUC 0.86–0.92 and 6-year C-index 0.75–0.81, validated on the NLST, Massachusetts General and Chang Gung cohorts. Requiring only the image, with no clinical variables, is what makes it deployable in a screening programme.
Nodule detection is the older task, at CPM around 0.85 on the standard benchmark. It answers a different question — where is the nodule, rather than what is this person’s risk.
Every model on this page runs as part of a managed AI pipeline in our GPU clusters, with a dedicated private cluster in our cloud or an on-premise installation where medical governance requires it. Output is decision support for a qualified professional to review, not a diagnosis.
Models in this group take a chest CT, typically a single low-dose screening study. Each table gives three hardware tiers — Minimum, the smallest configuration on which the model runs correctly; Medium, the usual production configuration; and High, a configuration sized for peak volume. The rate is what one server of that tier processes per hour. Use these figures for initial sizing only. Before production we benchmark your own data to confirm accuracy, latency, throughput and cost.
Sybil
Vendor: MIT CSAIL / Mass General
What it does: estimates one-to-six-year lung cancer risk from a single low-dose CT with no clinical variables at all, validated on three independent cohorts — the practical basis for risk-based screening intervals.
1-year AUC 0.86–0.92, 6-year C-index 0.75–0.81, validated on NLST, MGH and Chang Gung cohorts (independent, JCO 2023).
Requirement
Minimum
Medium
High
GPU type
RTX 3090
A100 80 GB
H100 80 GB
VRAM
24 GB
80 GB
80 GB
vCPUs
12
24
48
RAM
64 GB
128 GB
256 GB
Server
1× RTX 3090 24 GB
1× A100 SXM 80 GB
2× H100 SXM 80 GB
Rate (volumes/hour)
≈ 60
≈ 210
≈ 540
LUNA / grt123-style nodule detectors
Vendor: Diagnostic Image Analysis Group (Radboud UMC)
What it does: finds nodule candidates and scores malignancy at CPM around 0.85, answering the workup question rather than the risk one.
CPM about 0.85 on LUNA16 nodule detection (independent).
Requirement
Minimum
Medium
High
GPU type
RTX 3090
A100 80 GB
H100 80 GB
VRAM
24 GB
80 GB
80 GB
vCPUs
12
24
48
RAM
64 GB
128 GB
256 GB
Server
1× RTX 3090 24 GB
1× A100 SXM 80 GB
2× H100 SXM 80 GB
Rate (volumes/hour)
≈ 60
≈ 210
≈ 540
Choosing between them
For programme-level triage and interval setting, the risk model is the one with the evidence. For workup support on a specific scan, the detectors are the right shape. Both are sized for 3D CT volumes here.
Accuracy figures above are those the producers and independent evaluations report, on their own test sets. They are a shortlist tool, not a prediction of what you will see. At the start of a project we run a short proof of concept on a sample of your own data, which replaces them with real figures — so the cost and the schedule for the full engagement are known before anything is committed.
Send a representative sample, your expected volume and your latency target for lung cancer screening. We benchmark the shortlisted models, recommend the lowest-cost GPU configuration that meets the target, and scale it from pilot capacity to a dedicated production cluster — with the cost per unit of work known before you commit.