What makes a skin cancer high risk: the UK feature lists

Overview

• The same feature has two different thresholds depending on what it is being used for, and the UK dataset publishes a table comparing them. For squamous cell carcinoma, thickness over 4 mm makes a tumour high risk for clinical and multidisciplinary team management, while it takes over 6 mm to upstage a T1 or T2 to T3 in the staging system; level of invasion is 'at or beyond the subcutaneous fat' for risk and 'beyond the subcutaneous fat' for staging; and grade, lymphovascular invasion, subtype and margin status count for risk and not at all for staging (RCPath G124, appendix F). The dataset gives this mismatch as one of the reasons risk status caused confusion in practice, noting that 'particular confusion was generated by different risk factors being used for TNM upstaging compared with the tumour itself'. A tumour can therefore be high risk and low stage at the same time without either label being wrong.
• Risk status used to be a core item in the UK report and was removed. The dataset records that the core item 'caused numerous practical and clinical difficulties, reflected in the low level of acceptance and usage identified in the joint BAD-RCPath audit on NMSC', that binary low and high risk 'at times oversimplified a more complex clinicopathological situation', and that it was therefore moved out of the core section. What replaced it is a judgement made by the treating clinician or the skin cancer multidisciplinary team with the clinical facts in front of them. So an absent risk label on a report is the system working as designed, not an omission.
• The British Association of Dermatologists adds the things a slide cannot show. Its 2021 basal cell carcinoma guideline stratifies low against high risk on twelve criteria, of which only about half are pathological: alongside growth pattern, basosquamous differentiation, level of invasion, depth, perineural invasion, pathological T stage and margins, it counts the anatomical area and size together, whether the borders are well or poorly defined, whether the tumour is primary or recurrent, whether the person is immunosuppressed, and whether the site has had radiotherapy before. Its anatomical areas run from area L, the trunk and limbs, through area M, including the cheeks, forehead, scalp and neck, to area H, the mask areas of the face together with the hands, feet, ankles and genitals, where a tumour of any size is high risk. That is why the same 8 mm tumour is low risk on a back and high risk on a nose.
• The margin rule is stricter than most people expect, and it is not the same as an involved margin. A margin that is clear but under 1 millimetre is itself a defined high-risk pathological feature in both datasets. NICE and the national quality surveillance programme do not make a clear margin under 1 mm a mandatory reason for multidisciplinary team referral, and the datasets say the decision is individual or locally agreed, with a low threshold for asking the team. So a letter saying the cancer was completely removed and the case is still going to the team is not a contradiction.
• Counting the features matters as well as having one. The squamous dataset notes that the evidence strongly suggests that the number of high-risk factors present has significant clinical importance, and the greater the number the greater the overall risk. That is the same insight the Brigham and Women's Hospital system is built on, arriving at the same conclusion from the other direction.