AI Tool for Skin Cancer Triage Receives Conditional NHS Approval

AI-Powered Skin Cancer Triage Tool Gains Conditional NHS Approval

The National Institute for Health and Care Excellence (NICE) has conditionally recommended an artificial intelligence (AI) system called DERM (Deep Ensemble for Recognition of Malignancy) for use in the NHS in England. This innovative tool is designed to assess suspicious skin lesions and support teledermatology services, potentially revolutionizing skin cancer detection and triage 1

1

2

How DERM Works

DERM utilizes a smartphone with a dermoscopic lens attachment to capture high-quality images of suspicious skin lesions. These images are then uploaded to a secure online platform where a non-learning AI algorithm analyzes them. The system compares each image to a fixed database of known skin conditions, providing a suspected diagnosis and triaging the case 1

1

Potential Benefits and Impact

NICE suggests that DERM could significantly reduce delays for patients referred to the urgent suspected skin cancer pathway. Early evidence indicates that the tool might be as accurate as face-to-face or teledermatology assessments while easing pressure on dermatology services 1

1

Dr. Anastasia Chalkidou, director of NICE's Centre for Health Technology Evaluation, stated, "DERM has shown promising results in its ability to accurately distinguish between cancerous and non-cancerous skin lesions, with evidence suggesting it could halve the number of referrals to dermatologists within the urgent skin cancer pathway while maintaining patient safety" 2

2

Addressing NHS Challenges

The implementation of DERM comes at a crucial time for the NHS:

1. High referral rates: NHS dermatology services receive over one million referrals annually from primary care, with about 60% being urgent referrals for suspected skin cancer 1

   1

   .
2. Low cancer confirmation rate: Only 6% of urgent referrals are confirmed as cancer 1

   1

   .
3. Consultant shortage: Some NHS trusts report having no dermatology consultants at all 1

   1

   .
4. Appointment backlog: As of July 2024, there was a backlog of 441,000 elective dermatology appointments, with only 63% meeting the 18-week treatment target 1

   1

   .

Trial Period and Further Evaluation

DERM will be used for the next three years while the NHS collects further evidence on its impact. NICE will reconsider the evidence once the evaluation is complete and issue updated guidance 1

1

2

Related Stories

AI Model Achieves 94.49% Accuracy in Skin Cancer Detection, Promising Early Diagnosis Breakthrough

AI-Powered 'DeepMerkel' Predicts Outcomes for Aggressive Skin Cancers

UK Launches World's Largest AI Trial for Breast Cancer Screening

AI Model Rivals Dermatologists in Assessing Skin Cancer Aggressiveness

Limitations and Considerations

While promising, there are some limitations to consider:

1. Performance on diverse skin tones: Current evidence is based largely on patients with white skin. Its performance in people with black or brown skin is less certain 1

   1

   .
2. Additional review: To address this limitation, patients from these groups will receive an additional review by a healthcare professional during the trial period 1

   1

   .
3. Impact on non-cancer cases: It remains unclear whether the tool can free up capacity for patients with non-cancer, non-urgent inflammatory skin conditions that still require face-to-face assessment 1

   1

   .

Conclusion

The conditional approval of DERM represents a significant step forward in leveraging AI technology to improve healthcare delivery and patient outcomes. As Ashley Dalton, minister for public health and prevention, noted, "By embracing the power of AI, this exciting technology could help us slash waiting times, meaning that people with suspected skin cancer get the help they need, or peace of mind, faster" 2

2