The acquired K211678 file contains product-specific indications and users, design and operating principle, non-clinical evidence actually described sections. Compare these dated manufacturer statements with your own indication and design before commissioning evidence work. They describe this submission; they do not prescribe your testing, validate performance independently or establish a suitable predicate.
Evidence retrieved 2026-10-07. Source versions and topic-specific limits are listed below.
Computed exact-code research
This decision beside its recorded product-code cohort
The 2021–2025 complete-year cohort contains 23 selected decisions under primary code QDQ. 23 have valid reported receipt/decision pairs; 0 pairs are missing or invalid.
Calendar elapsed days: lower quartile 116.0, median 168.0, upper quartile 204.5. These are recorded calendar differences, not active FDA review time.
The table shows the named decision and up to four latest exact-code research records through 2026-09-27. It can include partial-year records outside the complete-year cohort. The quartiles above use the full stated cohort, not this displayed selection.
| Research record | Recorded device | Receipt → decision | Calendar elapsed days | Recorded submission type |
|---|
| K211678 ↗ · this page | Lunit INSIGHT MMG | 2021-06-01 → 2021-11-17 | 169 | Traditional |
|---|
| K261517 ↗ | qCT Lung (v1.0) | 2026-05-07 → 2026-09-14 | 130 | Traditional |
|---|
| K260898 ↗ | Transpara (2.1.0) | 2026-03-18 → 2026-08-14 | 149 | Special |
|---|
| K260086 ↗ | CadAI-B Dx | 2026-01-12 → 2026-08-12 | 212 | Traditional |
|---|
| K260303 ↗ | See-Mode Augmented Reporting Tool, Breast (SMART-B) | 2026-01-30 → 2026-07-28 | 179 | Traditional |
|---|
Use this comparison: select a record to investigate, read its actual indications and design, and document differences. A shared code or elapsed period cannot establish its relevance to your device.
Calendar elapsed days are not active FDA review time, a promised schedule or a success probability. The selection includes only the stated recorded decisions, not all applications. Latest records can postdate the named historical decision and are contextual research, not its predicates. Shared codes, summaries and observed records do not prove equivalence, current market availability, buyer applicability or predicate suitability.
Calculation, selection and source versions
Reuse the complete-endpoint exact-primary-code, documented SE-decision cohort and its stated cutoff/window. Retain its type-7 quartiles only when at least 20 valid date pairs are present. Independently subtract the named official record’s receipt date from its decision date; missing or negative pairs remain unknown. Show up to four latest exact-code research records, excluding the named ID. Parent analysis hash, source snapshots and calculation versions are retained.
Calculation: fda-named-decision-cohort-context-1; cutoff 2026-09-27; 0 parent records excluded; units record counts; calendar elapsed days.
Source snapshots: 5a86994ba49a37e71e09be96803fd51981c6c43eaa96caab41aea2b188bfe002.
Read beyond the decision row
Recorded evidence and differences to investigate
The summary is manufacturer-submitted context hosted by FDA. Negative/no-study statements are preserved. Evidence themes are not binding legal requirements.
Short opening excerpts identify the source sections. Read the linked original for the full study context, negative statements, tables and limitations. Public quotations are limited to 120 words per acquired summary; the internal evidence packet retains exact section ranges for review.
Indications and users
K211678 · PDF page 6 ↗ Source 2
Lunit INSIGHT MMG is a radiological Computer-Assisted Detection and Diagnosis (CADe/x) software
device based on an artificial intelligence algorithm intended to aid in the detection, localization, and
characterization of suspicious areas for breast cancer on mammograms from compatible FFDM
systems.
The excerpt is shortened at a complete statement. Read the complete section in the original PDF before interpreting the evidence.
Compare with your product: Compare the recorded indication, population, body site and use environment with your proposed label. Record each change instead of assuming the primary code settles intended use.
Design and operating principle
K211678 · PDF page 5 ↗ Source 3
Lunit INSIGHT MMG is a radiological Computer-Assisted Detection and Diagnosis (CADe/x) software
for aiding interpreting physicians with the detection, localization, and characterization of suspicious
areas for breast cancer on mammograms from compatible FFDM (full-field digital mammography)
systems.
Compare with your product: Compare physical design, materials, energy, accessories and operating principle with your design. Ask which documented differences change the evidence scope.
Non-clinical evidence actually described
K211678 · PDF page 7 ↗ Source 4
The Lunit INSIGHT MMG is considered as a “moderate” level of concern, since a failure or a latent
design flaw of the software device could result in Minor injury, either to a patient or to a user of the
device.
The excerpt is shortened at a complete statement. Read the complete section in the original PDF before interpreting the evidence.
Compare with your product: Treat these as documented evidence themes in this case, including any limitations or negative statements. Ask which protocols, design inputs and acceptance rationale would be justified for your product; these excerpts are not a mandatory test panel.
Extraction selection and limits
Select observed whole-line section headings or explicit colon-delimited inline labels on readable PDF pages, including the explicitly documented heading-spacing variants in this parser version. Preserve the original heading, character ranges and layout; do not repair excerpt text. Exclude letter boilerplate and uncertain pages. Stop at observed table headings; do not infer table columns or scalar results from layout text. Multi-column or suspicious split-numeric excerpts are excluded until reviewed alternate extraction is available. Original layout records remain retained; any alternate text has its own extractor-version locator and requires original-page review. For each dimension, prefer usable versioned alternate text where available, selecting its longest usable excerpt; otherwise select the longest usable original-layout excerpt. Ties preserve source order. Bounded to 64 PDF pages and 3,500 characters per section. This is a selection, not a complete dossier analysis.
Excluded extraction items: {}.
