(86 days)
The STARlight is intended for medical purposes for use on infants from 3 to 18 months of age, with moderate-to-severe non-synostotic positional plagiocephaly, including infants with plagiocephalic-, brachycephalic- and scaphocephalic-shaped heads by applying mild pressure to prominent regions of the infant's cranium in order to improve cranial symmetry and/or shape. The device is also indicated for adjunctive use for infants from 3 to 18 months of age whose synostosis has been surgically corrected, but who still have moderate-to-severe cranial deformities including plagiocephalic-, brachycephalicand scaphocephalic-shaped heads.
The STARlight redirects the head growth to improve proportion and symmetry. The practitioner takes a plaster impression or a scan of the infant's head to acquire the existing shape. The mold is sealed and filled with plaster or the scanned shape is carved from a rigid polyurethane foam blank to create a positive model of the head shape. The positive model is modified to obtain greater symmetry and space in the areas of flattening. The STARlight provides total contact over the prominent or bossed areas of the baby's head to discourage growth there. Over the course of treatment, the inside of the band is further modified by the practitioner to provide space for growth to occur in the flat or depressed areas. The shape of the STARlight directs growth into the areas of least resistance and creates a precise pathway for the head shape to improve in symmetry and proportion.
The STARlight product family as it was released in K090587 are essentially still the same devices. The STARlight Side Opening design and the STARlight Bi-Valve design are made of a plastic shell of 5/32" - 1/4" clear Surlyn or 1/8" - 7/32" Clear Co-Polyester. The STARlight PRO (Post-operative Remolding Orthosis) design is made of 1/4'' to 3/8" clear Surlyn. The STARband Bi-Valve design is made with an outer shell of 5/32" polyethylene-polypropylene copolymer plastic and an inner liner of 1/2" pelite polyethylene foam. Optional Aliplast (closed cell polyethylene) padding is available for the clear plastic bands and in addition optional Reston (polyurethane - 3M Medical Product) foam is available for the STARlight PRO design.
The STARlight Side Opening design has a top opening and a side opening. The band is held in place by a Velcro strap across the side opening. The STARlight PRO has two side openings, no top opening, and is held in place by a Velcro strap across each side opening. The STARlight Bi-Valve design and the STARband Bi-Valve design consist of two plastic shells that overlap with a superior sliding mechanism. The right and left overlap tabs are connected via a Velcro strap with chafe and loop.
The proposed device modification is the addition of a new shape capture method. specifically the scanGogh-11" by Vorum Research, Inc. This scanner uses one laser, either Class 1 or Class 2 depending on the option, and one camera to capture shape data. The Class I laser option of the scanGogh-II is as safe as other scanner devices previously cleared for the STARlight and will not require any extra safety precautions. The Class 2 laser option is safe to use on infant patients when used in conjunction with eye protection. The Class I laser option of the scanGogh-II shall be the recommended option for scanning infants. If the Class 2 option of the scanGogh-II is used, the infant should wear eye protection as instructed in the proposed labeling.
Here's a breakdown of the acceptance criteria and study information for the STARlight® Cranial Orthosis, based on the provided text:
Acceptance Criteria and Device Performance
The core of this submission is the addition of a new shape capture method, the scanGogh-II, for the STARlight Cranial Orthosis. The acceptance criteria and performance are tied to the accuracy of this new scanner compared to existing methods.
Table of Acceptance Criteria and Reported Device Performance
| Acceptance Criteria / Performance Metric | Requirement (Implicit) | Reported Device Performance (scanGogh-II vs. Cast) |
|---|---|---|
| Cranial Shape Capture Accuracy Study | Maintain or improve accuracy compared to Cast method. | Pass (performed superior for all listed metrics) |
| Proximal Radius | Substantially equivalent/superior to cast. | Pass |
| Proximal Anterior Radius | Substantially equivalent/superior to cast. | Pass |
| Anterior Radius | Substantially equivalent/superior to cast. | Pass |
| Anterior Posterior Length | Substantially equivalent/superior to cast. | Pass |
| M-L Width Anterior | Substantially equivalent/superior to cast. | Pass |
| M-L Width Posterior | Substantially equivalent/superior to cast. | Pass |
| Posterior Panel Flatness | Substantially equivalent/superior to cast. | Pass |
| Lateral Panel Flatness | Substantially equivalent/superior to cast. | Pass |
| L-P Panels Square | Substantially equivalent/superior to cast. | Pass |
| Medial Panels Flatness | Substantially equivalent/superior to cast. | Pass |
| M-P Panels Squares | Substantially equivalent/superior to cast. | Pass |
| Repeatability and Reproducibility (R&R) Analysis | Acceptable level of repeatability and reproducibility. | Determined to be acceptable. |
| Laser Safety (Class 1 Laser option) | Safe for infant use without extra protection. | Safe. |
| Laser Safety (Class 2 Laser option) | Safe for infant use with proper eye protection. | Safe when used with eye protection. |
| Eye Shield Fit Assessment (for Class 2 Laser option) | Proper fit, complete coverage, effective protection. | Pass (acceptable fit, completely covered the eye). |
| Material Biocompatibility (Surlyn, Copolymer + Pelite, Copolymer + Aliplast) | Non-sensitizing, negligible dermal response, non-cytotoxic. | All materials passed tests as "Not a Sensitizer," "Negligible Dermal Response," and "Non-cytotoxic" (with one "slight reactivity" for Copolymer + Aliplast in Cytotoxicity). |
Study Information
2. Sample size used for the test set and the data provenance
- Sample Size:
- For the Cranial Shape Capture Accuracy Study: "a representative cranial shape that possesses a predefined shape with known dimensions" was utilized. This implies a single, standardized phantom or model rather than a dataset of patient scans.
- For the Repeatability and Reproducibility (R&R) Analysis: "uniform shapes with known dimensions that represent various sizes of pediatric patients between ages 3 to 18 months of age" were utilized. No specific number is given, but it implies a set of phantoms or models.
- For the Eye Shield Fit Assessment: "infants" were assessed. No specific number is provided.
- Data Provenance: The studies appear to be prospective in nature, conducted specifically for this submission. The "country of origin" for this data is not explicitly stated, but given the company's location (Orlando, FL) and FDA submission, it's presumed to be within the US or compliant with US standards.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts
- The ground truth for the Cranial Shape Capture Accuracy Study and R&R Analysis was established using "uniform shapes with known dimensions" and "a representative cranial shape that possesses a predefined shape with known dimensions". This suggests that the ground truth was based on pre-defined, precisely measured physical models or phantoms, rather than relying on human expert consensus or clinical data. Therefore, no human experts were explicitly mentioned as establishing the ground truth for the performance testing of the scanner itself.
4. Adjudication method for the test set
- Not applicable as the ground truth was established by known physical dimensions of models/phantoms, not by human review requiring adjudication.
5. If a multi-reader multi-case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance
- No MRMC comparative effectiveness study was performed or needed. This device is a cranial orthosis and a new scanning method. There is no "human reader" component in the sense of image interpretation for diagnosis that would benefit from AI assistance. The scanning method replaces manual measurement or other digital scanning methods.
6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done
- Yes, the performance of the scanGogh-II (the algorithm/hardware for shape capture) was evaluated in a standalone manner against known physical dimensions without a human-in-the-loop performance scenario in the context of the Cranial Shape Capture Accuracy Study and R&R Analysis.
7. The type of ground truth used
- Known physical dimensions of uniform and representative cranial shapes/phantoms. This is a form of engineered or precisely measured ground truth.
8. The sample size for the training set
- The document does not describe any "training set" in the context of machine learning or AI. This submission is for a physical medical device (cranial orthosis) and a new shape capture tool. The scanGogh-II is a scanner, not an AI algorithm that would typically require a training set to learn.
9. How the ground truth for the training set was established
- Not applicable, as no training set for an AI/ML algorithm is mentioned.
§ 882.5970 Cranial orthosis.
(a)
Identification. A cranial orthosis is a device that is intended for medical purposes to apply pressure to prominent regions of an infant's cranium in order to improve cranial symmetry and/or shape in infants from 3 to 18 months of age, with moderate to severe nonsynostotic positional plagiocephaly, including infants with plagiocephalic-, brachycephalic-, and scaphocephalic-shaped heads.(b)
Classification. Class II (special controls) (prescription use in accordance with § 801.109 of this chapter, biocompatibility testing, and labeling (contraindications, warnings, precautions, adverse events, instructions for physicians and parents)).