QOCA image Smart CXR Image Processing System
K221868Quanta Computer Inc. · cleared 2023-01-27 · product code QFM · Radiology
Premarket evidence — what FDA accepted
Device typesamd
source quote (p.4)
“QOCA® image Smart CXR Image Processing System is a software as medical device (SaMD) used, through artificial intelligence/deep learning technology, to analyze chest X-ray images of adult patient, and then identify cases with suspected pneumothorax.”
Algorithmartificial intelligence/deep learning technology; locked artificial intelligence algorithm
source quote (p.4)
“QOCA® image Smart CXR Image Processing System is a software as medical device (SaMD) used, through artificial intelligence/deep learning technology, to analyze chest X-ray images of adult patient, and then identify cases with suspected pneumothorax. This product, QOCA® image Smart CXR Image Processing System, is a web-based medical device using a locked artificial intelligence algorithm.”
Adaptive (vs locked)No
source quote (p.7)
“This product, QOCA® image Smart CXR Image Processing System, is a web-based medical device using a locked artificial intelligence algorithm.”
PCCPFDA source did not state this
Cybersecurity addressedFDA source did not state this
Validation studies (2)
Retrospective clinical
n=3,105 images
endpoints: AUC; Sensitivity; Specificity
Retrospective clinical
n=2,947 images
endpoints: AUC; Sensitivity; Specificity
Reported performance (9 observations)
sensitivity0.925CI [90.5%, 94.2%]
source quote (p.12)
“the sensitivity and specificity achieves 92.5% (95% CI: [90.5%, 94.2%])”
specificity0.94CI [93.9%, 94.6%]
source quote (p.12)
“94.0% (95% CI: [93.9%, 94.6%]) respectively”
aurocas written: “auc”0.978CI [97.0%, 98.5%]
source quote (p.12)
“the performance assessment dataset achieves an area under the curve (AUC) of 97.8% (95% CI: [97.0%, 98.5%]”
aurocas written: “AUC (MIMIC dataset)”0.977CI [96.5%, 98.8%]
source quote (p.12)
“The performance of the subject device to the MIMIC dataset is AUC of 97.7% (95% CI: [96.5%, 98.8%])”
sensitivityas written: “Sensitivity (MIMIC dataset)”0.937CI [90.6%, 96.0%]
source quote (p.12)
“the sensitivity and specificity is 93.7% (95% CI: [90.6%, 96.0%])”
specificityas written: “Specificity (MIMIC dataset)”0.933CI [92.3%, 94.2%]
source quote (p.12)
“and 93.3% (95% CI: [92.3%, 94.2%]), respectively.”
aurocas written: “AUC (Taiwanese dataset)”0.974CI [96.9%, 98.7%]
source quote (p.13)
“The performance of the subject device to the Taiwanese dataset is AUC of 97.4% (95% CI: [96.9%, 98.7%])”
sensitivityas written: “Sensitivity (Taiwanese dataset)”0.917CI [88.8%, 94.0%]
source quote (p.13)
“the sensitivity and specificity is 91.7% (95% CI: [88.8%, 94.0%])”
specificityas written: “Specificity (Taiwanese dataset)”0.949CI [93.9%, 95.7%]
source quote (p.13)
“and 94.9% (95% CI: [93.9%, 95.7%]), respectively.”
Each value carries its own analysis unit and task — never compare or pool across devices. Source: 510(k) summary PDF.
Predicate network
Postmarket — what happened after clearance
0
recalls in product code, 24mo
0
MAUDE reports in code, 12mo
—
vs code's own 3-yr baseline
0
drift signals on this device
Recall and MAUDE counts are product-code-level (reports aren't reliably attributable to one device). Signals are descriptive observables with sources — never a judgment that the device is unsafe or drifting. Snapshot 2026-07-08.
Reimbursement — how devices like this got paid
Not yet tracked — no payment pathway indexed for this clearance (the reimbursement corpus is a growing seed set).