Search Results

Ice Cooling IPL Hair Removal Device with sapphire treatment window is indicated for the removal of unwanted hair. The device is also indicated for the permanent reduction in hair regrowth, defined as the long-term, stable reduction in the number of hairs regrowing when measured at 6, 9 and 12 months after the completion of a treatment regime.

Ice Cooling IPL Hair Removal Device is an over-the-counter, home-use and personal device for hair reduction by using Intense Pulsed Light (IPL) with safety and efficacy. It is designed with a lamp that can emit continuously double or triple pulses per shot. It works below the skin's surface and does not involve any cutting or pulling, reducing hair growth with nearly painless pain and nearly heatless.
The device is only powered by the external power adapter and its IPL emission activation is by finger switch. This product adopts sapphire treatment window that is suitable for multiple hair removal areas. It contains a skin sensor to detect appropriate skin contact, if the device is not in full contact with the skin, the device cannot emit the treatment light pulses. Besides, the device has the ice cooling function that will be activated throughout the whole hair removal process to provide users with a more comfortable and safer experience.

The provided FDA 510(k) clearance letter and summary for the "Ice Cooling IPL Hair Removal Device" describe a medical device submission and its review. However, it does not contain the specific information required to describe acceptance criteria and associated study results for an AI/ML-based medical device.

The document primarily focuses on:

• Regulatory classification and product codes: Identifying the device as Class II, Product Code OHT, under 21 CFR 878.4810.
• Intended use: Hair removal and permanent reduction in hair regrowth.
• Comparison to predicate devices: Demonstrating substantial equivalence in terms of intended use, design, specifications, and performance (e.g., wavelength range, energy density, spot size).
• Performance data (Non-Clinical): Referring to biocompatibility testing, electrical safety (EMC), light safety, software verification and validation (not AI-specific performance), and usability testing, all against established industry standards (IEC, ISO).

Crucially, there is no mention of:

• AI/ML components: The device is described as an Intense Pulsed Light (IPL) device, and its operation does not inherently involve AI/ML. "Software Verification and Validation" is a standard requirement for electronic medical devices and does not imply AI.
• Clinical study data for performance metrics: The performance data section refers only to non-clinical tests (biocompatibility, electrical safety, light safety, software V&V, usability). It does not provide details on clinical efficacy (hair reduction) studies with specific performance metrics, sample sizes, or ground truth establishment relevant to the device's hair removal claim. The "permanent reduction in hair regrowth" indication implies clinical testing, but the details are not present in this summary.
• Expert consensus, MRMC studies, or specific AI performance metrics like sensitivity, specificity, AUC: These are typical elements of a study proving an AI device meets acceptance criteria.

Therefore, it is not possible to fill in the requested table and answer the questions based solely on the provided text, as the device described is not an AI/ML medical device, and the document focuses on non-clinical substantial equivalence rather than detailed clinical performance of the hair removal efficacy.

Hypothetical Example (if this were an AI/ML device and the text provided the necessary details):

If, for instance, the device had an AI component to detect skin type or predict hair regrowth, and the document detailed a study on this AI component, the information could be extracted like this (this is purely illustrative and not based on the provided text):

Hypothetical Acceptance Criteria and Study for an AI-Powered Hair Removal Device

Let's imagine this device also had an AI feature, for example, an integrated AI system that analyzes skin pigmentation to recommend optimal IPL settings to minimize adverse events and maximize efficacy.

1. A table of acceptance criteria and the reported device performance

2. Sample size used for the test set and the data provenance

• AI Skin Type Classification Test Set: 1500 images/cases (1000 for training, 500 for validation/testing).
• Adverse Event/Efficacy Test Set (Clinical Trial): 300 participants.
• Data Provenance:
  • Skin images for AI classification: Retrospective dataset collected from dermatology clinics in North America (USA, Canada) and Europe (UK, Germany).
  • Clinical trial for adverse events/efficacy: Prospective, multi-center, randomized controlled trial conducted in the USA (5 sites) and China (3 sites).

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts

• Skin Type Ground Truth: 3 board-certified dermatologists, each with a minimum of 10 years of experience in aesthetic dermatology and laser/IPL treatments. All were trained to consistently apply Fitzpatrick Skin Type scale.
• Adverse Event/Efficacy Ground Truth: Clinical investigators (dermatologists) at each trial site, with at least 5 years of experience in IPL/laser treatments, reviewed and graded adverse events and hair reduction independently at follow-up visits.

4. Adjudication method for the test set

• Skin Type Ground Truth: For the AI classification test set, initial skin type labels were provided by one expert. For any ambiguous cases (e.g., initial disagrement or boundary cases), a 3-expert consensus (2+1 majority rule) was used. If a consensus was not reached (e.g., 1-1-1 split), the case was excluded from the ground truth set.
• Adverse Event/Efficacy Ground Truth: For clinical trial outcome adjudication, two independent, unblinded dermatologists graded outcomes (hair reduction percentage, adverse events) at each follow-up visit. In case of discrepancy, a third blinded dermatologist acted as an adjudicator to reach a consensus.

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

• Yes, an MRMC study was performed to assess the impact of AI-recommended settings vs. manual-expert settings on clinical outcomes.
• Study Design: 10 IPL therapists/dermatologists (readers) were recruited. They each reviewed 50 simulated patient profiles (cases) and recommended IPL settings.
  • Group A (without AI): Readers used standard clinical guidelines and their experience.
  • Group B (with AI assistance): Readers were provided with the AI system's recommended settings and could choose to accept, modify, or reject them.
• Effect Size: The AI-assisted group (Group B) demonstrated a statistically significant improvement in the rate of optimal setting selection (leading to good outcomes without adverse events) by 15% (Cohen's d = 0.65) compared to the unassisted group (Group A). Specifically, the rate of selecting optimal, safe, and effective settings increased from 70% (unassisted) to 85% (AI-assisted).

6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done

• Yes, a standalone performance evaluation of the AI skin type classification module was conducted prior to its integration into the device and the MRMC study. This included the accuracy, sensitivity, and specificity metrics mentioned in section 1.

7. The type of ground truth used

• Expert Consensus: For skin type classification (AI module training and testing).
• Clinical Outcomes Data: For validating the safety and efficacy of AI-recommended settings (adverse event rates, hair reduction percentage from prospective clinical trial data).

8. The sample size for the training set

• AI Skin Type Classification: 15,000 unique skin images with associated Fitzpatrick Skin Type labels.

9. How the ground truth for the training set was established

• The ground truth for the training set was established by three experienced board-certified dermatologists, similar to the test set experts. Each image was independently reviewed and labeled by all three. If discrepancies occurred, a consensus process involving discussion and re-evaluation was used to arrive at a final label. This iterative process ensured high-quality, reliable ground truth data for training the AI model.

Ask a specific question about this device