Machine learning methods applied to audit of surgical margins after curative surgery for facial (non-melanoma) skin cancer. Issue 1 (January 2023)

• Record Type: Journal Article
• Authors: Tighe, David; Tekeli, Kemal; Gouk, Tara; Smith, Jennifer; Ho, Michael; Moody, Andrew; Walsh, Stephen; Provost, Simon; Freitas, Alex
• Is Part Of: British journal of oral and maxillofacial surgery. Volume 61:Issue 1(2023)
• Page Start: 94
• Page End: 100
• Publication Date: 2023-01
• Subjects: Audit -- Non-melanoma skin cancer -- Outcomes -- Surgical margin; Mouth -- Surgery -- Periodicals; Maxilla -- Surgery -- Periodicals; Face -- Surgery -- Periodicals; Surgery, Plastic -- Periodicals; Dentistry, Operative -- Periodicals; Face -- surgery -- Periodicals; Mouth -- surgery -- Periodicals; Maxilla -- surgery -- Periodicals; Surgery, Oral -- Periodicals; Oral Surgical Procedures -- Periodicals; Dentistry, Operative -- Periodicals; Bouche -- Chirurgie -- Périodiques; Maxillaire supérieur -- Chirurgie -- Périodiques; Face -- Chirurgie -- Périodiques; Chirurgie dentaire -- Périodiques; Dentistry, Operative; Face -- Surgery; Maxilla -- Surgery; Mouth -- Surgery; Surgery, Plastic; Electronic journals; Periodicals; 617.52059
• DOI: 10.1016/j.bjoms.2022.11.280
• Languages: English
• ISSNs: 0266-4356

Abstract

Abstract: We aimed to build a model to predict positive margin status after curative excision of facial non-melanoma skin cancer based on known risk factors that contribute to the complexity of the case mix. A pathology output of consecutive histology reports was requested from three oral and maxillofacial units in the south east of England. The dependent variable was a deep margin with peripheral margin clearance at a 0.5 mm threshold. A total of 3354 cases were analysed. Positivity of either the peripheral or deep margin for both squamous cell carcinoma (SCC) and basal cell carcinoma (BCC) was 15.4% at Unit 1, 21.1% at Unit 2, and 15.4% at Unit 3. Predictive models accounting for patient and tumour factors were developed using automated machine learning methods. The champion models demonstrated good discrimination for predicting margin status after excision of BCCs (AUROC = 0.67) and SCCs (AUROC = 0.71). We demonstrate that rates of positive excision margins of facial non-melanoma skin cancer (fNMSC), when adjusted by the risk prediction model, can be used to compare unit performance fairly once variations in tumour factors and patient factors are accounted for.