Rayvolve

K240845

AZmed SAS · cleared 2024-07-17 · product code QBS · Radiology

Premarket evidence — what FDA accepted

Device typesamd

source quote (p.7)

“The medical device is called Rayvolve. It is a standalone software that uses deep learning techniques to detect and localize fractures on osteoarticular X-rays.”

Algorithmstandalone software that uses deep learning techniques, object detection model, supervised deep learning

source quote (p.7)

“The medical device is called Rayvolve. It is a standalone software that uses deep learning techniques to detect and localize fractures on osteoarticular X-rays. Machine learning technology: Supervised Deep learning. The core design of the Rayvolve algorithm, including the object detection model, remains unchanged from the predicate device (Rayvolve K220164).”

Adaptive (vs locked)No

source quote (p.15)

“The core design of the Rayvolve algorithm, including the object detection model, remains unchanged from the predicate device (Rayvolve K220164). The architecture and key components are consistent with those previously described.”

PCCPNo

Cybersecurity addressedYes

source quote (p.9)

“HIPAA Compliant”

Validation studies (3)

Standalone

n=3,016 images

endpoints: sensitivity; specificity; AUC

Standalone

n=2,626 images

endpoints: non-inferiority of AUCs

Reader study (MRMC)

n=186 cases

endpoints: diagnostic accuracy (AUC) of readers aided by Rayvolve; sensitivity of Rayvolve-aided and unaided reads; specificity of Rayvolve-aided and unaided reads

Reported performance (27 observations)

sensitivity0.9611CI 0.9480; 0.9710

source quote (p.11)

“The results of standalone testing demonstrated that Rayvolve detects fractures of the musculoskeletal system radiographs with high sensitivity (0.9611, 95% Wilson's Confidence Interval (CI): 0.9480; 0.9710)”

specificity0.8597CI 0.8434; 0.8745

source quote (p.11)

“high specificity (0.8597; 95% Wilson's CI: 0.8434; 0.8745)”

aurocas written: “auc”0.9399CI 0.9330; 0.9470

source quote (p.11)

“and high Area Under The Curve (AUC) of the Receiver Operating Characteristic (ROC) (0.9399; 95% Bootstrap CI: 0.9330; 0.9470).”

aurocas written: “AUC by Anatomic Area (Pediatric) - Ankle”0.9489CI 0.9257; 0.9694

source quote (p.12)

“Ankle 0.9489 (0.9257; 0.9694)”

aurocas written: “AUC by Anatomic Area (Pediatric) - Clavicle”0.9263CI 0.8846; 0.9645

source quote (p.12)

“Clavicle 0.9263 (0.8846; 0.9645)”

aurocas written: “AUC by Anatomic Area (Pediatric) - Elbow”0.9308CI 0.8932; 0.9629

source quote (p.12)

“Elbow 0.9308 (0.8932; 0.9629)”

aurocas written: “AUC by Anatomic Area (Pediatric) - Forearm”0.936CI 0.9012; 0.9666

source quote (p.12)

“Foreram 0.936 (0.9012; 0.9666)”

aurocas written: “AUC by Anatomic Area (Pediatric) - Humerus”0.9568CI 0.9268; 0.9818

source quote (p.12)

“Humerus 0.9568 (0.9268; 0.9818)”

aurocas written: “AUC by Anatomic Area (Pediatric) - Hip”0.947CI 0.922; 0.9681

source quote (p.12)

“Hip 0.947 (0.922; 0.9681)”

aurocas written: “AUC by Anatomic Area (Pediatric) - Knee”0.9624CI 0.9472; 0.9756

source quote (p.12)

“Knee 0.9624 (0.9472; 0.9756)”

aurocas written: “AUC by Anatomic Area (Pediatric) - Pelvis”0.9263CI 0.8947; 0.9559

source quote (p.12)

“Pelvis 0.9263 (0.8947; 0.9559)”

aurocas written: “AUC by Anatomic Area (Pediatric) - Shoulder”0.9372CI 0.9037; 0.9664

source quote (p.12)

“Shoulder 0.9372 (0.9037; 0.9664)”

aurocas written: “AUC by Anatomic Area (Pediatric) - Tibia/Fibula”0.9616CI 0.9362; 0.9824

source quote (p.12)

“Tibia/Fibula 0.9616 (0.9362; 0.9824)”

aurocas written: “AUC by Anatomic Area (Pediatric) - Wrist”0.9484CI 0.9258; 0.9688

source quote (p.12)

“Wrist 0.9484 (0.9258; 0.9688)”

aurocas written: “AUC by Anatomic Area (Pediatric) - Hand”0.9485CI 0.9306; 0.9654

source quote (p.12)

“Hand 0.9485 (0.9306; 0.9654)”

aurocas written: “AUC by Anatomic Area (Pediatric) - Foot”0.9404CI 0.9211; 0.9581

source quote (p.12)

“Foot 0.9404 (0.9211; 0.9581)”

aurocas written: “AUC by Ethnicity (Pediatric) - Caucasian”0.944CI 0.9331; 0.9548

source quote (p.13)

“Caucasian 0.944 (0.9331; 0.9548)”

aurocas written: “AUC by Ethnicity (Pediatric) - Hispanic”0.948CI 0.9362; 0.9589

source quote (p.13)

“Hispanic 0.948 (0.9362; 0.9589)”

aurocas written: “AUC by Ethnicity (Pediatric) - African-American”0.9542CI 0.9335; 0.9724

source quote (p.13)

“African-American 0.9542 (0.9335; 0.9724)”

aurocas written: “AUC by Ethnicity (Pediatric) - Asian”0.9272CI 0.8932; 0.9588

source quote (p.13)

“Asian 0.9272 (0.8932; 0.9588)”

aurocas written: “AUC by Ethnicity (Pediatric) - Others”0.9308CI 0.9087; 0.9503

source quote (p.13)

“Others 0.9308 (0.9087; 0.9503)”

aurocas written: “Aided Reader AUC”0.89327

source quote (p.18)

“Reader AUC was significantly improved from 0.84602 to 0.89327, a difference of 0.04725 (95% CI: 0.03376; 0.061542)”

aurocas written: “Unaided Reader AUC”0.84602

source quote (p.18)

“Reader AUC was significantly improved from 0.84602 to 0.89327”

sensitivityas written: “Aided Reader Sensitivity”0.9554CI 0.94453, 0.96422

source quote (p.18)

“Reader sensitivity was significantly improved from 0.86561 (95% Wilson's CI: 0.84859, 0.88099) to 0.9554 (95% Wilson's CI: 0.94453, 0.96422)”

sensitivityas written: “Unaided Reader Sensitivity”0.86561CI 0.84859, 0.88099

source quote (p.18)

“Reader sensitivity was significantly improved from 0.86561 (95% Wilson's CI: 0.84859, 0.88099)”

specificityas written: “Aided Reader Specificity”0.83116CI 0.81673, 0.84467

source quote (p.18)

“Reader specificity was improved from 0.82645 (95% Wilson's CI: 0.81187, 0.84012) to 0.83116 (95% Wilson's CI: 0.81673, 0.84467)”

specificityas written: “Unaided Reader Specificity”0.82645CI 0.81187, 0.84012

source quote (p.18)

“Reader specificity was improved from 0.82645 (95% Wilson's CI: 0.81187, 0.84012)”

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

recalls in product code, 24mo

MAUDE reports in code, 12mo

—

vs code's own 3-yr baseline

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).